Management of Victoria s Publicly-owned Native Forests for Wood Production

Transcription

1 SUSTAINABILITY & RESOURCES PROJECT NUMBER: PRC FEBRUARY 2011 Management of Victoria s Publicly-owned Native Forests for Wood Production This report can also be viewed on the FWPA website FWPA Level 4, Queen Street, Melbourne VIC 3000, Australia T +61 (0) F +61 (0) E info@fwpa.com.au W

2 Management of Victoria s Publicly-owned Native Forests for Wood Production: A Review of the Science Underpinning their Management. Prepared for Forest & Wood Products Australia by J. Turner, D. Flinn, M. Lambert, K. Wareing and S. Murphy

3 Publication: Management off Victoria s Publicly-owned Native Forests for Wood Production: A Review of the Science Underpinning their Management. Project No: PRC Forest & Wood Products Australiaa Limited. All rights reserved. Forest & Wood Products Australia Limited (FWPA) makes no warranties w orr assurancess with respect to this publication including merchantability, fitness for purpose or otherwise. FWPA and all persons associated with it exclude all liability (including liabilityy for negligence) in relation to any opinion, advice or information contained in this publication or for any consequences arising from the use of such opinion, advice or information. This work is copyright and protected under the Copyright Act (Cth). All material except the FWPA logo may be reproduced in whole or in part, provided that it is nott sold or used for commercial benefit and its source (Forestt & Wood Products Australia Limited) ) is acknowledged. Reproduction or copying for other purposes, which is strictly reserved only y for the owner or licensee of copyright under the Copyrightt Act, is prohibited without the prior written consent of Forest & Wood Products Australia Limited. This work is supported by funding provided to FWPA by the Department of Agriculture, Fisheries and Forestry (DAFF). ISBN: Principal Researcher: John Turner, David Flinn, Marcia Lambertt Forsci Pty Ltd forsci@fluoroseal.com.au Kevin Wareing Kevin Wareing & Associates Pty Ltd Simon Murphy Department of Sustainability & Environment, Victoria Simon.murphy@dse.vic.gov.au Final report received by FWPA in November, 2010 Forest & Wood Products Australia Limited Levell 4, Queen St, Melbourne, Victoria, 3000 T F E info@fwpa.com.au W

4 TABLE OF CONTENTS 1. INTRODUCTION AND OBJECTIVES 1 Page No. 2. HISTORICAL BACKGROUND TO VICTORIA S NATIVE FORESTS Forests Prior to European Settlement Impacts of European Settlement Clearing of Forests for Agriculture, Mining and Settlement Forest Regulation Fire Regimes Timber Harvesting Current Tenure of Native Forests Discussion and Conclusions 9 3. NATIVE FOREST MANAGEMENT IN VICTORIA Statutory, Institutional and Policy Framework Forest Management Planning Timber Harvesting Introduction Determination of Appropriate Sawlog Harvesting Levels Establishment of VicForests Separation of Commercial and Policy, Monitoring and Regulatory Roles Victoria s Timber industry Strategy Towards Sustainable Forest Management in Victoria Discussion and Conclusions 22 i

5 4. COMMERCIALLY IMPORTANT FOREST TYPES AND THEIR SILVICULTURAL 23 MANAGEMENT 4.1 Forest Types Introduction to Silvicultural Principles Silvicultural Research and Practice Alpine Ash Forests Introduction Silviculture of Alpine Ash in NSW and Tasmania Tasmania New South Wales Silviculture of Alpine Ash in Victoria A Detailed Case Study Flowering, Fruit Development and Seeding Germination Seedling Establishment Silvicultural Systems Alpine Ash SFM Case Study Alpine Ash Regeneration Following Wildfire Future Seed Management Thinning Mountain Ash Forests Introduction Silvics of Mountain Ash Forests Stocking Following Harvesting of Mountain Ash in Victoria Mountain Ash Silviculture Mountain Ash Regeneration Mountain Ash Growth and Development High Elevation Mixed Species Forests Introduction Silvics of HEMS Forests Regeneration of HEMS Forests HEMS Silviculture HEMS Seed-tree and Clearfell HEMS Selection System HEMS Shelterwood HEMS Variable Retention HEMS Reforestation and Re-treatment Thinning HEMS Forests Overwood Competition Seed Management Low Elevation Mixed Species Forests Introduction Silvics of LEMS Forests Stocking Following Harvesting of LEMS in Victoria LEMS Silviculture LEMS Seed-tree System 61 ii

6 LEMS Clearfell System LEMS Selection System LEMS Shelterwood System LEMS Variable Retention System LEMSD Reforestation Thinning Growth Response Regeneration Regrowth Development Thinning and Fertilizing Overwood Competition Impacts of Overwood on Regeneration Box-Ironbark Forests Introduction Silvics and Regeneration of Box-Ironbark Forests Box-Ironbark Silviculture Selection Thinning Ecological Thinning River Red Gum Forests Introduction Site Characteristics of River Red Gum Silviculture of River Red Gum Effect of River Regulation Discussion and Conclusions Introduction Alpine Ash Forests Mountain Ash Forests High Elevation Mixed Species Forests Low Elevation Mixed Species Forests Box-Ironbark Forests River Red Gum Forests ASSESSING THE ECONOMIC VALUE OF SUSTAINABLE FOREST 85 MANAGEMENT 5.1 Introduction Native Forest Products and Marketing in Victoria Products and Markets Non-Wood Forest Products Introduction Historical Considerations Recent Developments Discussion and Conclusions 92 iii

7 6. MANAGEMENT OF FIRE, PESTS AND DISEASES IN FORESTS Introduction Wildfires Damage to Trees and Stands Community Issues and Government Enquiries Environmental Impacts Use of Fire Retardants Fuel Reduction Burning Introduction Knowledge of Fuel Accumulation and Fire Behaviour Effectiveness of Fuel Reduction Burning in Wildfire Control Ecological Impacts of Fuel Reduction Burning A New Bushfire Strategy Fire and Sustainable Management Pests and Diseases Introduction Insect Pests Mountain Ash Psyllid Gum Leaf Skeletoniser Phasmatids Pathogens Target Spot or Corky Leaf Spot Armillaria Root Rot Cinnamon fungus Forest Health Surveillance Discussion and Conclusions PRODUCTIVITY AND YIELD Introduction Estimation of Growth and Biomass General Considerations Assessment of Victoria s Forest Resources Estimates of Biomass Productivity of Forest Types Alpine Ash Forests Mountain Ash Forests High Elevation Mixed Species Forests Low Elevation Mixed Species Forests 124 iv

8 7.6 Growth Estimates for the Major Forest Types Discussion and Conclusions ENVIRONMENTAL ASPECTS OF SUSTAINABLE FOREST MANAGEMENT Introduction Soils and Soil Management Erosion Compaction Nutrient Loss and Productive Capacity Water Values Water Yield Introduction Harvesting and Wildfire Effects Thinning Effects Water Quality Introduction Fire Effects Harvesting Effects Aquatic Values Biological Diversity Introduction Wet/Montane Forests Flora Fauna Current Research Directions Lowland/Dry/Damp Forests Flora Fauna Thinning Greenhouse Gases and Carbon Storage Greenhouse Gases Carbon Storage Greenhouse gases and carbon storage Discussion and Conclusions Soils and Soil Management Water Values Flora and Fauna Values Wet/Montane Forests Lowland/Dry/Damp Forests Greenhouse Gases and Carbon Storage 168 v

9 9. DISCUSSION AND OVERALL CONCLUSIONS Overview and Discussion The Concept of Sustainable Forest Management Management of Fire, Pests and Diseases Pests and Diseases Wildfires Fire Retardants Fuel Reduction Burning Silvicultural Practices Alpine Ash Forests Mountain Ash Forests High Elevation Mixed Species Forests Low Elevation Mixed Species Forests Box-Ironbark Forests River Red Gum Forests Standard of Timber Harvesting Productivity and Yield Environmental Impacts Soil Values Water Values Flora and Fauna Values Future Directions Research, Development and Monitoring Community Awareness Extent and Availability of Victorian Forest Research BIBLIOGRAPHY ACKNOWLEDGEMENTS APPENDICES 214 I. List of Acronyms 214 II. List of Species 215 III. Listing of Native Forest Silvicultural Guidelines 216 vi

10 1. INTRODUCTION AND OBJECTIVES Natural forests have been used extensively to produce timber and a range of non-wood forest products for more than 150 years in Victoria, and for considerably longer periods in countries such as Finland, Sweden, Germany, France, Canada and the United States. In Australia, the scientific basis for the employed forest management practices (including silvicultural systems, harvesting technologies, protection measures from damaging agents such as wildfire, pests and diseases, and environmental impacts of these practices) is arguably poorly understood by the full range of stakeholders and particularly the broader community. Despite a long history of relatively intensive stakeholder consultation by relevant Victorian Government Departments, addressing a wide range of forest management issues, during processes including the development of Forest Management Plans, Regional Forest Agreements and Codes of Practice and in the development of forest management policies, it is considered that the broader community in particular is ill-informed on the science that underpins the management of native forests in Victoria and beyond. This is well illustrated in the print and electronic media where simple concepts such as commitments to regeneration following timber harvesting are clearly not widely acknowledged. Another common example is the confusion surrounding the terms back burning and fuel-reduction burning which are frequently used inter-changeably by the press and public. It is also arguable that some of the information in the public domain on many key forest management practices lacks scientific credibility yet continues to have an influence on popular opinion. Furthermore, some of the more credible scientific and technical information is often difficult to access (for example, information in internal Departmental research reports which have a restricted distribution) and has not been consolidated into one or more up-to-date documents. Victoria's forest industries are involved in the growing, tending, protecting, harvesting and processing of logs from plantations and native forests. There are significant differences between the species, the size and wood properties of logs produced from plantations and native forests and consequently, most forest companies operating in Victoria have dedicated plants to process logs from these two sources. The products produced by the individual processors reflect the inherent properties of the timber being processed, while the volume of logs now processed from native forests reflects the area of State forest currently suitable and available for timber harvesting. This area has been determined via exhaustive land-use evaluations and regulatory processes that have taken place over the past t In 2005, the Victorian Department of Primary Industries (DPI) commissioned Forsci Pty Ltd to review and document the science that underpins the management of Victoria s publicly-owned commercial native forests for the information of policy makers, environmental and industry groups, the scientific community, relevant Unions and all other stakeholders including the broader community. At that time, DPI had a charter to provide services that drive sustainable development of the State's agricultural, fishing, minerals, forestry and petroleum industries for the benefit of the Victorian community, both now and in the future. The Department's vision for Victorian forest industries was that they will be: 1. Sustainably managed to meet the needs of current and future generations. 2. Internationally competitive. 3. Operating in a market environment that optimises economic, social and environmental outcomes, and 4. Recognised by the broader community as an important part of the Victorian economy by providing goods from a renewable resource, protecting water supplies and enhancing biodiversity. 1

11 The Brief for the review required a focus on key management issues including silvicultural and associated management aspects of native forest management. Consideration of most social, economic and cultural aspects including indigenous issues of the management of native forests was beyond the scope of the review, except for contextual purposes. In order to achieve the objectives of the review, some important background information is initially provided about the influence of European settlement on the present-day native forest estate before considering a series of strategic forest management issues. Some of these issues were broadly based while others were highly focused. The importance of this historical information is best illustrated by the 1939 wildfires. If these fires had not occurred, some of the regrowth forests currently available for timber production would now be old-growth forests with a far higher conservation value than the 1939 regrowth forests. Space limitations set by the Brief for the review prevented all the issues surrounding the native forest logging debate being addressed. It is also important to appreciate that a complete review of all the relevant science, and the quality of this science, was likewise beyond the scope of the review. Rather, the review attempts to focus on what the authors regarded as key research papers and technical reports from a Victorian perspective. Priority was given to published literature and unpublished reports of the past two decades, though reference to earlier studies has been essential for some aspects of the review. The final chapter (essentially an Executive Summary) presents the major findings of the review. The review was completed in October 2005, but was not published by DPI. Forsci Pty Ltd therefore sought alternative funding to update the review and make it readily available to all stakeholders. Forest and Wood Products Australia (FWPA) provided such funding in April In the intervening period, there have been some significant policy initiatives and research achievements. Many of the more recent policy initiatives leading up to the publication date for this review of July 2010, including a Sustainability Charter and a Timber Industry Strategy present new scientific challenges in the management of native forests. The review is therefore timely; a similar review for Victorian forests was last undertaken in the early 1980s (Campbell et al. 1984). 2

12 2. HISTORICAL BACKGROUND TO VICTORIA S NATIVE FORESTS 2.1 Forests Prior to European Settlement The first human beings were thought to have arrived in Australia from southeast Asia more than 60,000 years ago, and Indigenous occupation of the land has probably been continuous since then (BRS 2003). In addition to naturally occurring fires from lightning, Aboriginal peoples applied fire as a land management and hunting tool, although its use varied considerably across the country. Over tens of thousands of years, regular burning/igniting by Aboriginal peoples probably had a major effect on vegetation structure and composition, but the extent and implications of this are still debated. However, thousands of years of regular burning by Aboriginal peoples created a landscape of grassy woodlands and open forests on the plains and foothills, and this was very attractive to the early squatters (Hallam 1979, Williams and Gill 1995). The patterns and structure of forests at the time of European settlement are a product of the environment and the combination of regular burning by Aboriginal peoples, interspersed with periodic high intensity fires caused primarily by lightning. A conclusion to be drawn from this is that changes in fire regimes, either by long-term total fire exclusion and/or an increase in high intensity fires, will lead to changes away from existing long-term forest processes and to consequent changes in forest structure and species composition. 2.2 Impacts of European Settlement on the Forests Clearing of Forests for Agriculture, Mining and Settlement Immediately prior to European settlement, Victoria's 22.7 million hectares were mainly forested (including woodlands and shrublands) and remained so until the late 1860s (TIS 1986, Woodgate and Black 1988). The first European settlers arrived in the mid 1830s with their sheep and cattle in what is the State of Victoria today. The previous long-term fire history had made many of these forests and woodlands suitable and attractive for grazing. Grasslands were mainly confined to parts of southwest and northwest Victoria, with shrubland (mainly Mallee scrub) covering the far northwest. The balance of the State was forested (forested being defined as areas having a tree cover with greater than 10% crown projection). Many of these forests in the western half of the State in particular would have had a woodland-type structure. By the mid 1840s, only the Mallee, the eastern ranges and parts of Gippsland remained largely unoccupied by European settlers. 3

13 One hundred and seventy years of predominantly European settlement have had a significant impact on the total area, structure and floristic composition of Victoria's native forests (Woodgate and Black 1988, BRS 2003). Events that have played a role in determining the extent and nature of the presentday forests include: Cessation of frequent low-intensity fire following settlement. Discovery of gold and the consequent rapid expansion of both Victoria's population and economy. Alienation and clearing of large areas of Crown land for agricultural development. Use of wood as a major energy resource for homes, mining and industry from the 1850s up to the 1940s. Use of native forests as the major domestic source of structural timber for housing and other buildings until the 1970s, after which increasing supplies of plantation-grown softwood became available. Major wildfire events, particularly those in 1926, 1939, 1983, 2003, 2006/07 and 2009 that killed extensive stands of Alpine Ash (Eucalyptus delegatensis) and/or Mountain Ash (E. regnans), and Increasing concern about the conservation of Victoria's forests and increased direct public participation in decision-making processes. Generally, the use and management of Victoria's native forests have reflected contemporary community values varying from the general indifference of the early settlers, more effective controls over the alienation and use of forests to provide for future timber requirements in the latter part of the 19 th and early part of the 20 th Centuries, and the expansion of National Parks and other conservation reserves in the latter half of the 20 th Century and the first decade of the 21 st Century. Widespread clearing of Victoria's forests and woodlands did not occur until the gold rushes of the 1850s. The gold miners stripped the vegetation to access shallow alluvial gold deposits and felled nearby forests for mining timbers and fuel. The forests were regarded almost solely as a useful adjunct to the then booming gold-mining industry and were heavily exploited (Ferguson 1965). The influx of gold seekers rapidly increased the State's population and created a pressing demand for farming land, not only to grow food but also to encourage agricultural settlement as a means of providing a basis for permanence and prosperity. Large areas of Crown land were made available for selection in the 1860s, 1870s and 1880s, and selectors set about clearing and developing their blocks. The magnitude of this task varied. Clearing of the sparsely timbered plains and woodlands posed few problems compared with the Herculean task of clearing tall forests in Gippsland (for example, in the Strzelecki Ranges) and in the Otway Ranges. 4

14 2.2.2 Forest Regulation Community concern about the vast quantities of timber being felled and wasted on land that had been alienated for agriculture together with the lack of controls over timber cutting in the remaining forests eventually resulted in legislation in 1907 to establish a State Forests Department. Although the establishment of the Department reflected a gradual realisation that the welfare of the community would suffer if the protection and management of the State's forests continued to be disregarded, it was ineffective in stemming the tide of forest destruction as there was a prevailing attitude that only land unsuitable for agriculture should be reserved as State forest. The Forests Act 1919 constituted the Forests Commission Victoria, and gradually, forest areas were identified and dedicated as reserved forest. The management and protection of State forests then progressively improved Fire Regimes Prior to the establishment of the Victorian Forests Commission, broadscale forest clearing created a great deal of debris which was usually burnt in summer to ensure a good burn. Many of these so called clearing fires escaped, causing damage to adjacent uncleared forest (McKinty 1969). Fire was also used by graziers to stimulate palatable feed on the forest floor. Again, many of these fires were lit at dangerous times of the year and escapes were common (McKinty 1969). It is generally accepted that fire regimes (season, intensity and frequency of fire in a specific area over time) have changed since European settlement (Banks 1990, BRS 2003, AFS 2007). Fires now tend to be less frequent and more intense, but this has been confounded in more recent times by the extensive use of low-intensity fire for fuel-hazard reduction. These changed fire regimes have impacted in various ways (see Section 6.3.4). Few studies have attempted to determine the historical frequency of forest fires. Banks (1990) used a mixture of dendrochronology, fire scars and age of regeneration to study fire frequency in the Glenbog State Forest in NSW that comprised both wet and dry sclerophyll forests. This forest is comparable with Victoria's High Elevation Mixed Species forests. Banks (1990) found that fire frequency in the dry sclerophyll forest was twice that in the wet sclerophyll forest, and there was a pattern of a higher frequency of severe fires following European settlement. Widfires are an important part of the history of Victoria's native forests. In one catastrophic week in January 1939, one of the most damaging fires since European settlement resulted in the loss of 71 lives and thousands of cattle, sheep, horses and other livestock along with the destruction of townships, 69 sawmills, bridges, tramways and the incineration of 1,364,000 hectares of forest including most of the State's prime Mountain Ash forests (Griffiths 2001). The effects of the 1939 fires are still very much evident today. Amongst other things, they are reflected in the age class distribution of Victoria's Mountain Ash forests, the reduced water yield from many catchments, and a reduction in habitat diversity. The 2003 Alpine fires and the 2006/07 Great Divide fires also killed extensive areas of both mature and regrowth Mountain Ash and Alpine Ash (Wareing and Flinn 2003, Flinn et al. 2008). Apart from the unprecedented loss of human lives, the 2009 Black Saturday fires also killed or damaged large tracts of State forests, including commercially valuable regrowth and mature Mountain Ash in the Central Highlands. These three more recent fires will have significant long-term economic and ecological impacts. As a result of the destruction of extensive areas of Mountain Ash forests in the 1939 fires, a massive operation was immediately mounted to salvage wood from the fire-killed forests. The salvage 5

15 operations, which were finally suspended around 1950, produced some 4.5 million cubic metres of sawlogs and made a major contribution to meeting Victoria's timber requirements during World War II and in the immediate post-war years (TIS 1986). Significant salvage operations were also mounted for the three more recent major fires. In the case of the 2003 Alpine fires, Theobald and Lawlor (2006) documented a salvage operation focused on economically suitable and accessible Alpine Ash stands that had been assessed as being fire-killed. Utilisation Plans were developed to salvage an estimated 600,000 cubic metres of sawlog, involving amongst other things, a major upgrade of local roading networks to cater for a substantial increase in log truck traffic. By December 2004, around one third of the sawlogs had been harvested along with more than 300,000 cubic metres of residual logs. Salvage operations continued for the next two years, with care being taken to minimise damage to fire-induced Alpine Ash regeneration Timber Harvesting Victoria s forests have played a critical role in the development of the State, particularly during the latter half of the 19 th century and the first half of the 20 th century, when they provided vast quantities of timber for housing and other building, bridge construction, rail sleepers, boiler fuel to generate power for gold mining and manufacturing, and firewood for domestic heating and cooking. After World War II, returning servicemen and women, together with the acceptance of refugees from displaced persons camps in Europe and the Government's initiative to boost immigration, contributed to a post- War housing shortage. In 1946 for example, it was estimated that Victoria would need to construct about 90,000 homes to satisfy community needs (TIS 1986). To meet this increased demand for timber for housing and to offset the dwindling supplies from the salvage of timber from forests devastated by the 1939 fires, the sawmilling industry was encouraged to expand into the mountain regions of northeast Victoria and north-central Gippsland, and into the extensive foothill and coastal forests of east Gippsland. The production of hardwood sawlogs from State forests then rapidly increased and peaked at around 1.3 million cubic metres per annum in the mid-1950s. Production has since declined, with annual production of wood products (mainly sawlogs and pulpwood) being around 2 million cubic metres in 2005/06 of which 25% was sawlogs (DSE 2009a). Indeed, sawlog production hovered at around 0.5 million cubic metres between 2001/02 and 2005/06. The east Gippsland forests have played an important long-term role in Victoria's development, and this commenced well before the housing boom. Despite the remoteness of these forests from Melbourne, the construction of the Orbost railway in 1916 provided a cost-effective opportunity to supply Melbourne and other areas with timbers with unique qualities that were not readily available from other forest types, particularly durable timbers for a range of uses including railway sleepers and infrastructure (such as bridges and wharves) (McKinty 1969). Some of the more unusual uses for the durable timbers were River Red Gum paving blocks for the streets of Melbourne and other durables for the manufacture of Victorian Railways rolling stock. McKinty (1969) noted that sleeper production from east Gippsland foothill and coastal forests increased from around 250,000 to 350,000 sleepers per annum pre-war to 500,000 to 600,000 sleepers per annum during the extension of the standard gauge rail line from Albury to Melbourne. While the rate of harvesting in native forests was increased to meet the needs for additional supplies of building timber, it was recognised that the existing State forests would be unable to meet Victoria's long-term timber requirements. In 1960, as part of an initiative to become self-sufficient in wood products, Victoria commenced a program to expand the area of softwood plantations. This continued until the early 1990s. Plantation softwood from Victoria and interstate/overseas has now replaced 6

16 native forest hardwood in many structural applications. Hardwood producers have responded to this development by re-positioning their production to higher value products where the appearance, hardness and strength properties of native forest hardwoods provide a competitive advantage and satisfy an important niche market that would otherwise be met totally from imported timbers (including tropical timbers sourced from rainforests). It is noteworthy that the only significant clearing carried out by successive Government Departments responsible for State forests has been the conversion of native forest to softwood plantations, though the area involved has more or less been balanced by reforestation of degraded mine sites and abandoned farmland at a number of locations including the Otway and Strzelecki Ranges. Some of the softwood plantations established on denuded mine sites in northeast and north-central Victoria are now into their third rotation. Following European settlement, modifications occurred due to clearing (complete or partial removal of trees) for agriculture and, to a lesser extent, mining. Many of the forests decimated by the miners were able to recover, mainly from coppice and lignotubers and some of those resultant regrowth forests now form part of Victoria's National and State Park system. Victoria ceased clearing of public native forests for plantations soon after the release of the 1986 Timber Industry Strategy (TIS 1986). Increasing concern about the conservation of Victoria's natural resources and the controversy over a proposal to alienate public land in the Little Desert for agricultural development led to the formation of the Land Conservation Council in Its role was to make balanced decisions on the use of public land. The Council encouraged direct public involvement in its decision-making processes. The recommendations of the Council and its successors have led, inter alia, to the cessation of the alienation of public land for agricultural development and to a greatly expanded area of National Parks and other conservation reserves where timber harvesting has been excluded. As noted above, a Timber Industry Strategy was released by the Victorian Government in 1986 which set new directions for the industry based on a balance between timber production and environmental protection (TIS 1986). Among the initiatives that were implemented as part of the Strategy were the development of a Code of Practice for Timber Production [the first Code was ratified by the Victorian Parliament in 1989; this was revised in 1996 and again in 2007 (DSE 2007a)], reduction of hardwood sawlog harvesting to more sustainable levels on a regional basis, a major silvicultural research program to evaluate alternatives to clearfelling and public participation in the development of Forest Management Plans. The Our Forests, Our Future Policy Statement released by the Victorian Government in 2002 reinforced the Government's commitment to a more sustainable timber industry and provided the basis for the future sustainable management of the native forests (DNRE 2002a). Subsequent key elements of the framework for the sustainable management of State forests in Victoria include a Sustainability Charter, a Thinning Policy Statement and an Environmental Policy (certification to ISO requirements). However, one of the most significant developments in the past few years has been the release of a new Timber Industry Strategy (TIS) (DPI 2009) 23 years after Victoria s first in Since 1986, Victoria s timber industry has faced rapid change and continues to operate in a challenging environment. Greater resource security on public and private land is considered essential for industry development and the 2009 Timber Industry Strategy focuses on key policy changes to deliver resource security in a sustainable framework. 7

17 2.3 Current Tenure of Native Forests Victoria has approximately 7.8 million hectares of native forest (public and private) equivalent to almost 35% of the total area of the State (Table 2.1). Within the publicly-owned native forest area, around 51% is in formal conservation reserves where timber harvesting is not permitted. Within the remaining area (46.4%) of public native forest (excluding the Other public land category), the harvesting and sale of timber may be carried out by VicForests in accordance with the Allocation Order made under the Sustainable Forests (Timber) Act The Allocation Order is for a period of 15 years divided into five-year periods and may be extended. The Allocation to VicForests (Further Amendment) Order 2010 (VGP 2010b) specifies both the total gross area available for timber harvesting and the maximum gross area that VicForests can harvest in each five-year period. The gross area of State forest available for timber harvesting is million hectares (VGP 2010b). Further information on the determination of the gross areas available for timber harvesting and the determination of sustainable harvesting levels is provided in Section 3. Table 2.1 Areas of native forests according to land tenure in Victoria. Land Tenure Use Area % of % of Total % of Total (ha) Total Area Native Forest Public Forest Total in the State of Victoria 22,742, Total Native Forest (public & private) 7,837, Native Forest Public Land Multiple-use State forest 3,163, Conservation Reserve 3,505, Other Public land including 144, Leased & Unresolved Tenure Sub-total 6,812, Native Forest Private Land 1,025, Source: (Montreal Process Implementation Group (MPIG) 2008) 8

18 2.4 Discussion and Conclusions 1. The extent, structure, floristic composition and age class distributions of Victoria's native forests have been determined by an interaction between environment (rainfall, soils and altitude in particular) and human input, initially by Aboriginal peoples. 2. Clearing for agriculture continued unabated and unrestricted until the early 1900s despite concerns by some legislators. As a result, only around 7.8 million hectares of native forest remain in Victoria including 6.8 million hectares on public land. Around 3.5 million hectares are in formal conservation reserves, and a further 3.2 million hectares is State forest. The gross area of State forest available for timber harvesting is 2.2 million hectares, equivalent to about 67% of the total State forest area or 31% of Victoria s public native forest area. 3. Fire regimes (season, intensity and frequency of fire in a specific area over time) have changed since European settlement. Pre-European fire regimes, characterised by frequent low-intensity fire, were replaced by regimes that encouraged less frequent but more intense fire. This has affected the structure and species composition of present-day native forests (BRS 2003, AFS 2007). 4. Victoria has a long history of devastating wildfires, and this situation is likely to be an on-going challenge for forest management organisations. In relation to timber resources, the salvage operation mounted in the Mountain Ash forests killed by the 1939 wildfires is a milestone event in Victorian forestry. 5. Victoria has developed two Timber Industry Strategies (TISs) to provide policy guidance for the management of Victoria s forests. The first TIS in 1986 covered both native forests and plantations, while the focus of the second TIS released in 2009 focuses on policy changes to deliver resource security in a sustainable framework. 9

19 3. NATIVE FOREST MANAGEMENT IN VICTORIA 3.1 Statutory, Institutional and Policy Framework Of the 7.84 million hectares of native forest in Victoria, 3.51 million hectares is in National Parks and other conservation reserves, 3.16 million hectares is multiple-use State forest and 1.03 million hectares is private land. The number of Victorian government agencies with management and/or regulatory responsibilities for the State s native forests together with their detailed strategies and policies reflect the response of successive governments to the debate on the use and management of native forests. This has continued at both the global, national and State level over the last 40 years. The following is a summary of current management arrangements as at 1 July The total area of National Parks and conservation reserves has expanded dramatically over the last 40 years largely as a result of successive Governments acting on the studies and recommendations on balanced land use by the Land Conservation Council (between ), the Environment Conservation Council ( ) and the Victorian Environmental Assessment Council (VEAC since 2001). The Department of Sustainability and Environment (DSE) provides policy guidance and is responsible for fire prevention and suppression in State forests, forested National Parks, other parks and conservation reserves. Parks Victoria is responsible for all other activities in National Parks, other parks and conservation reserves. Victoria s State forests are managed in accordance with the provisions of the Forests Act 1958, Conservation, Forests and Land Act 1987, Flora and Fauna Guarantee Act 1988 and Sustainable Forests (Timber) Act 2004, together with related regulations, codes of practice, management plans and policy initiatives, including Growing Victoria Together (2005), Our Forests, Our Future (2002) and Our Environment, Our Future Victoria s Environmental Sustainability Framework (2005) and the Timber Industry Strategy (2009). DSE manages State forests; VicForests is the state-owned enterprise that is responsible for the sustainable harvest and commercial sale of timber from State forests. The Minister for Agriculture, and therefore, DPI provides guidance to VicForests on operating in a framework that is consistent with government policy and priorities. The Sustainability Charter for Victoria s State Forests (DSE 2006a) sets out the Government s Vision for Victoria s State forests and through this Charter, DSE, DPI and VicForests carry out their relevant responsibilities in accordance with the following objectives: 1. To maintain and conserve biodiversity in State forests; 2. To maintain and improve the capacity of forest ecosystems to produce wood and non-wood products; 3. To promote healthy forests by actively managing disturbance; 4. To maintain and conserve the soil and water resources of State forests; 5. To maintain and better understand the role of Victoria s State forests in global carbon cycles; 6. To maintain and enhance the socio-economic benefits of State forests to Victorian communities; and 7. To ensure Victoria s legal, institutional and economic frameworks effectively support the sustainable management of State forests. 10

20 To meet the requirements of the Sustainable Forests (Timber) Act 2004, Criteria and Indicators (C&I) have been developed (DSE 2007b) for the sustainable management of Victoria s State forests that are consistent with national and Montreal Process C&I (see Section 3.5). Victoria s seven Criteria are directly compatible with the seven objectives in the Sustainability Charter for Victoria s State forests (DSE 2006a). Forty five Indicators covering social, economic, environmental and cultural values have been identified to inform Victorians on progress towards sustainable forest management (SFM). However, given that the risk of devastating wildfires in Victoria is higher than anywhere else in the world with the possible exception of California, the extent and severity of major wildfires during the period 2003 to 2009 inclusive and the impact of these fires on biodiversity, on the production of wood and non-wood forest products, on soil and water resources, on carbon emissions and on the socioeconomics of many Victorian communities, it appears that the scant attention and/or the omission of any specific reference to fire prevention and suppression in a number of appropriate Indicators is a major oversight that needs to be addressed in the future. This issue is given further consideration in Section 6.5. Victoria s private native forests are managed by a relatively large number of individual owners or their nominees. DPI is responsible for private forestry and the productive use of plantations and native forest on private land. DPI also provides policy guidance relating to forestry and the forest products sector in public and private forests and was responsible for developing the new Timber Industry Strategy for Victoria (DPI 2009) which sets out key action areas to ensure that Victoria has a productive, competitive and sustainable timber industry. 3.2 Forest Management Planning The 1986 Timber Industry Strategy (TIS 1986) provided for the adoption of an integrated forest management planning approach and the preparation of Forest Management Plans (FMPs) that: Apply to Forest Management Areas (FMAs) (see Table 3.1) which are units for planning integrated management. Address the full range of values and uses of the forest. Apply for ten years with provision for revision after five years, and Are produced with extensive opportunity for public consultation and participation. To provide an agreed national response to the management of native forests and to avoid potential conflicts between different levels of government, the Commonwealth, and all State and Territory Governments subsequently came together to develop a strategy (National Forest Policy Statement) for the ecologically sustainable management of Australia s forests (NFPS 1992). A key element of the Statement was the preparation of Regional Forest Agreements (RFAs) that apply for up to 20 years and have three main objectives: To protect environmental values in a Comprehensive, Adequate and Representative (CAR) Reserve System based on nationally agreed criteria. To encourage job creation and growth in forest-based industries, including wood products, tourism and minerals, and To manage all native forests in an ecologically sustainable way. 11

21 In Victoria, RFAs were prepared for each of five identified RFA Regions (East Gippsland 1997, Central Highlands 1998, North East 1999, Gippsland 2000 and West 2000) and signed off by both the Prime Minister and the Premier. In broad terms, the parties to the Agreements: Agreed that the primary function of the CAR Reserve System is to ensure the conservation and protection of heritage values, and that any change in that component of the CAR Reserve system in State forest will be in accordance with the terms of the Agreement and will not lead to any deterioration in the protection of identified CAR values; Re-affirmed their commitment made in the National Forest Policy Statement to the conservation and management of the private forest estate and noted that under the Planning and Environment Act 1987, Victoria has native vegetation retention controls to regulate the clearance of native forest on private land; and Agreed that State forest outside the CAR Reserve System is to be available for timber harvesting in accordance with the Victorian forest management system. The sustainable yields for forests will be reviewed based on new resource information. Adoption of the CAR Reserve System for the five Victorian RFAs increased the area of State forest reserved for conservation purposes by 900,000 ha (DNRE 2002a). While expansion of the area of conservation reserves involved some changes in public land tenure from State forest to National Park or other tenure categories, it was largely accomplished by dividing the State forest into management zones in Forest Management Plans (FMPs). The integration of the forest planning initiatives set out in the Timber Industry Strategy in 1986 and the formulation of the CAR Reserve System set out in the National Forest Policy Statement in 1992 provided a comprehensive and integrated approach to the sustainable management of Victoria s native forests. To illustrate this, the FMP for Gippsland (DSE 2004b) is consistent with the Gippsland RFA and provides a framework for the sustainable management of State forest within the region. The legislative and policy framework underpinning this and other FMPs is multi-layered and complex and has been described by DSE (2004b) as follows: The Plan has been developed to conform with the Victorian land and natural resources legislation including the Forests Act 1958, Land Act 1958, National Parks Act 1975, Reference Areas Act 1978, Flora and Fauna Guarantee Act 1988, Heritage Rivers Act 1992 and the Catchment and Land Protection Act Protection of species listed under the Commonwealth Environment Protection and Biodiversity Conservation Act 1999 is also provided for in this Plan. The Plan has also been developed to conform with the cultural heritage legislation such as the Archaeological and Aboriginal Relics Preservation Act 1972 (Vic.), Aboriginal and Torres Strait Islander Heritage Protection Act 1984 (Commonwealth) and Heritage Act 1992 (Vic.). The Plan fulfils a requirement of the Code of Forest Practices for Timber Production (the Code) (DNRE 1996). This plan also conforms with land use decisions made in accordance with the Land Conservation Act The principal strategy to achieve the aims of FMPs, has been to divide the State forests into three management zones. The Special Protection Zone will be managed for conservation; the Special Management Zone will be managed to conserve specific features and the General Management Zone will be managed for a range of uses and values, with sustainable timber production being a major use. In addition to the management zones, the Plans contain aims, management guidelines, and proposed actions for various forest uses. 12

22 3.3 Timber Harvesting Introduction Following the completion of the RFA process, the Victorian Government announced the Our Forests, Our Future (DNRE 2002a) policy initiative that focused on the sustainable management of State forests not included in the CAR Reserve System, and the separation of the Government s commercial and policy, monitoring and regulatory roles in forest management. Among the initiatives contained in the package were: The determination of appropriate sawlog harvesting levels in each Forest Management Area. The establishment of VicForests as a separate fully commercial entity, reporting through an independent Board, to manage the commercial harvesting and sale of forest produce from State forests with industry. A forest industry structural adjustment package to assist with industry restructuring, and The implementation of a statutory framework for the separation of the Government s commercial and policy, monitoring and regulatory roles in forest management Determination of Appropriate Sawlog Harvesting Levels The Department of Natural Resources and Environment prepared and published Estimate of Sawlog Resource reports for each FMA (DNRE 2002b). The area of what is referred to as the economically accessible resource is tabulated in the reports together with the annual volume available from that resource (see Table 3.1). Table 3.1 Estimates of the area and sawlog yield for each Forest Management Area. Forest Total area of Economically Accessible Resource Management State Area % of Total Sawlog Implied Area Forest (ha) (ha) Area Yield (m 3 /yr) Sawlog MAI a, b East Gippsland 636, , , Tambo 418,908 82, , North East 530,971 38, , Central Gippsland 521,917 89, , Dandenong 56,325 22, , Central 203,244 71, , Benalla-Mansfield 171,530 24, , Sub-total East 2,539, , , Midlands 114,300 26, , Otways 92,900 35, , Horsham 103,300 4, Portland 113,346 20, , Mildura 348,000 4, Mid-Murray 61,170 46, , Sub-total West 833,016 46, , Total 3,372, , , Source: DNRE (2002b). a Mean Annual Increment (m 3 /ha/yr) b Calculated by the authors as sawlog yield divided by economically accessible area. 13

23 The estimates of annual sawlog production are based on approximately 20% of the total area of State forest being used for sawlog production. The respective estimates have been adopted as the appropriate sawlog harvesting level for each Forest Management Area. Following the publication of estimates of sawlog resources, the Government announced that logging would be phased out in the Otways FMA. As a result of this decision, the area of economically accessible resource was reduced from 676,371 hectares to 640,916 hectares. The Sustainable Forest (Timber) Act 2004, addressed in Section 3.3.4, provides for the ongoing review of sustainable harvesting levels. These reviews which take account of changes in the land use and the impacts of major fires are reflected in the amendments to the 2004 Allocation Order in 2007 and Establishment of VicForests VicForests is a State body, established by Order in Council dated 28 October 2003, under the provisions of the State Owned Enterprises Act 1992 (VGP 2003). The Order in Council states that the functions of VicForests are to: Undertake the sale and supply of timber resources in Victorian State forests, and related management activities, as agreed by the Treasurer and the Minister (of Agriculture), on a commercial basis. Develop and manage an open and competitive sales system for timber resources, and Pursue other commercial activities as agreed by the Treasurer and the Minister. VicForests became operational on 1 August 2004 and is responsible to a Board of Directors who report to the Treasurer of Victoria. Initially, VicForests was only required to carry out its functions in eastern Victoria where the higher sawlog yielding forests and the bulk of the economically accessible sawlog resources are located (see Table 3.1), while the Department of Natural Resources and Environment and its successor the Department of Sustainability and Environment continued to manage commercial timber harvesting in western Victoria. The 2009 Timber Industry Strategy which is addressed in Section 3.3.5, details VicForests current role Separation of Commercial and Policy, Monitoring and Regulatory Roles The Sustainable Forests (Timber) Act 2004 provides the framework for separation of the forest stewardship and commercial timber harvesting roles in State forest. The Act, inter alia, addresses the determination of Criteria, Indicators and reporting requirements for Sustainable Forest Management, allocation of timber to VicForests, the transfer of existing licences, management of timber resources by VicForests, management of timber harvesting, conduct of timber harvesting operations and fire prevention and suppression. In accordance with the Sustainable Forests (Timber) Act 2004, the Minister of Environment issued an Allocation Order, dated 29 July 2004 (VGP 2004a). The Allocation Order specifies the maximum area of each forest stand (in hectares) that can be harvested for timber or thinned in the 15-year time frame from 1 August 2004 to 31 July 2019, The Order, divided into three five-year time periods, specifies the maximum area of State forest available to VicForests for harvesting to at least meet existing commitments under licences in the period from 1 August 2004 to 31 July 2019, the activities that VicForests is authorised to undertake, and the conditions and standards that apply to VicForests in 14

24 undertaking these activities. When an Allocation Order is made, VicForests is required to prepare a Timber Release Plan in respect of any area that it proposes and/or sell timber resources from or undertake management activities within. This Plan, and any subsequent amendments, must be submitted for approval to the Secretary, Department of Sustainability and Environment. The Timber Release Plan must be consistent with the Allocation Order and must include a schedule of coupes to be harvested, associated road access requirements and details of the location and timing of timber harvesting in the proposed coupes. All proposed timber harvesting and roading must comply with any relevant Codes of Practice. The VicForests Timber Release Plan for 2004 to 2009 was approved in August 2004 (VGP 2004b). The Allocation Order dated 29 July 2004 (VGP 2004a) provides quite detailed information and identifies the extent and location of 16 forest stand types which VicForests can access in each of the three five-year periods. The maximum areas allocated in each of the three periods for harvesting in eastern Victoria (that is, the East Gippsland, Tambo, Central Gippsland, North East, Central, Dandenong and Benalla-Mansfield Forest Management Areas) is summarised in Table 3.2. Table 3.2 Areas allocated for harvesting in eastern Victoria for three five-year periods commencing in Period Five-Year period Maximum Area allocated for Harvesting (ha) No. Ash Species Mixed Species Total 1 August 2004 to July ,240 18,860 29,100 2 August 2009 to July ,930 17,400 26,330 3 August 2014 to July ,000 17,440 25,440 It can be concluded from the data in Table 3.2, that VicForests will be permitted to access, harvest, rehabilitate and regenerate about 2,050 hectares of ash species and 3,750 hectares of mixed species per annum in the first five-year period, and this will decline to about 1,600 hectares of ash species and 3,500 hectares of mixed species per annum in the third five-year period. It is expected that about 500,000 m 3 /yr of grade D and better sawlogs and up to 1,500,000 m 3 /yr of residual logs will be produced from the areas allocated to VicForests in eastern Victoria. With respect to sawlog production, this expectation is consistent with DNRE s Estimates of Sawlog Resource (DNRE 2002b). The Allocation Order made on 29 July 2004 was amended on 21 March 2007 following the Great Divide Fires, to take into account the impact of those fires on timber resources and facilitate the salvage harvesting of trees killed by the fires (VGP 2007). In accordance with the Sustainable Forests (Timber) Act 2004, the Minister for Environment and Climate Change completed a five-year review of the Allocation Order in 2009 (DSE 2010) and determined that it be amended to address: The impacts of major forest fires in and 2009 on the structure and condition of large 15

25 areas of State forest and therefore on the availability of timber resources. Reductions in the area of State forest as a result of additions to the area of conservation reserves in East Gippsland. Changes to forest management zoning and harvesting prescriptions resulting from new and revised action statements, and The addition of a new five-year period (Period 4) to the Allocation Order. Following this review, the Allocation to VicForests (Amendment) Order 2010 (VGP 2010a) was made to include the following amendments to the area allocated to VicForests: The inclusion of areas of State forest in western Victoria as well as those in eastern Victoria. While DSE will continue to meet legal timber supply obligations from western Victoria, the decision to conduct sustainable timber harvesting operations in these areas rests with VicForests. A reduction from 16 to 3 in the number of forest stand types identified, namely ash species, mixed species and durable species. A state based allocation according to stand type with no reference to Forest Management Areas. The allocated areas are expressed in terms of gross area of State forest that is available for timber harvesting (previously, DSE used net area), and modelled merchantability, productivity and sustainable harvest levels in allocating an area to VicForests. The exclusion of the area of new and expanded National Parks and conservation reserves from the Allocation Order. The maximum area that VicForests can harvest in each forest stand type and in each five-year period of the amended 2010 Allocation Order are shown in Table 3.3, Table 3.3 shows the gross areas available for timber harvesting in Period 2 (1 August 2009 to 31 July 2014) of the Allocation Order. The gross areas available for timber harvesting in Period 3 (1 August 2014 to 31 July 2019) and Period 4 (1 August 2019 to 31 July 2024) are identical to those for Period 2. Table 3.3 Area available for timber harvesting in Period 2 (1 August 2009 to 31 July 2014) of the Allocation Order. Forest Stand Total Available Nominal Gross Area available Type Gross Area Rotation for Harvesting in each (ha) (yrs) five-year period (ha) Ash 239, ,400 Mixed Species 1,720, ,100 Durable Species 148, ,700 Total 2, ,200 Source: VGP (2010a) 16

26 From the data in Table 3.2, VicForests will be permitted to access, harvest, rehabilitate and regenerate about 8,930 hectares of ash species and 17,400 hectares of mixed species in Period 2. Furthermore it was noted that these areas were expected to yield about 500,000 m 3 /yr of grade D and better sawlogs and up to 1,500,000 m 3 /yr of residual logs. With respect to sawlog production this expectation was consistent with DNRE s Estimates of Sawlog Resource (DNRE 2002b). Because of changes to the method for defining the area available for timber harvesting, it is not possible to compare the data in Tables 3.2 and 3.4 or to draw conclusions about the annual production of sawlogs or other products. 3.4 Victoria s Timber Industry Strategy In 2009, the Minister for Agriculture in Victoria released a new Timber Industry Strategy to provide a framework and long-term direction for the timber industry for the next 20 years (DPI 2009). The Strategy identifies thirteen key action areas under four priorities as follows: Priority 1. A productive, competitive and sustainable timber industry. Action 1. Provide greater certainty of access to public native forest timber resources Action 2. Improve estimation and communication of sustainable harvest levels from public native forests Action 3. Improve the sales system for native hardwood logs from public native forests Action 4. Sustainably develop timber plantations Action 5. Assist the timber industry to adapt to climate change Priority 2. Develop and support efficient timber markets. 4. Action 6. Improve freight infrastructure and logistics to support the timber supply chain 5. Action 7. Support the commercial development of new and emerging markets for timber and timber related products 6. Action 8. Support market access and improve biosecurity for sustainable timber production 7. Action 9. Strengthen governance arrangements for forests and timber production Priority 3. Innovative forestry science, technology and practice changes. Action 10. Encourage industry innovation and research and development Action 11. Improve industry occupational health and safety Priority 4. Strong timber industry communities. Action 12. Build a skilled workforce Action 13. Enhance community understanding of the benefits of Victoria s forests The 2009 Timber Industry Strategy does not specifically nominate which Government agencies will take the lead role in implementing each of the 13 specified actions. It does, however, clarify management and governance arrangements for sustainable forest management and timber 17

27 production in Victoria (DSE 2010) as follows: VicForests is responsible for determining long term sustainable harvest levels from the stands to which it has access to under the Allocation Order. VicForests must undertake strategic, tactical and operational planning for the management of timber harvesting operations to ensure the long term productivity of the forest is maintained and the regulatory framework for sustainable forest management is complied with. This includes taking into account the impacts of fire on the structure and condition of State forests, and therefore the availability of timber resource. The relevant Minister (the Minister for Agriculture) and the Department of Primary Industries is responsible for the oversight of VicForests operations, including its commercial charter and customer relationships. Furthermore, the recently developed Sustainability Charter for Victoria s State forests (DSE 2006a) requires VicForests to respond to the Government s sustainability agenda by developing initiatives and targets to progress objectives of the Charter. The Treasurer and the Department of Treasury and Finance retain their responsibilities legislated in the State Owned Enterprises Act The Minister for Environment and Climate Change and the Department of Sustainability and Environment is responsible for land management, environmental policy, establishing the regulatory framework for sustainable forest management and for auditing VicForests determination of long term sustainable harvest levels and its timber harvesting operations to ensure compliance with that regulatory framework. 3.5 Towards Sustainable Forest Management in Victoria At a global level, forestry was one of the first primary industries to adopt the modern day concept of sustainable management. The concept of what is now commonly termed Sustainable Forest Management (SFM) can be traced back to the 1992 United Nations Conference on Environment and Development (UNCED) which focused world attention on the importance for forestry to adopt sustainability principles, although the 1987 World Commission on Environment and Development Report first drew broad community awareness to the concept of Ecologically Sustainable Development (ESD). This was the stimulus for a number of regional and international processes that progressively identified a set of social, cultural, heritage, economic and environmental Criteria and Indicators (C&I) that could be used to characterize the state of a nation s forests and to monitor trends in land use changes (for example, Montreal Process, Pan European Process, Tarapoto Proposal - see ISCI 1996, Rametsteiner 2001). C&I can be applied at various spatial scales ranging from national and sub-national (State) levels to regional and Forest Management Unit (FMU)* levels. Initial emphasis was on the development of national level Indicators for, amongst other things, the purpose of raising awareness, of gaining commitment (to SFM) and to assist in measuring broad progress towards achieving SFM (Raison et al 2001a). Clearly, national level Indicators are, as noted by Raison et al (2001a), often not sufficiently sensitive to be useful at the FMU level. * In this review, the term Forest Management Unit (FMU) is loosely defined as either a (large) coupe, a compartment or a (small) Working Circle, in recognition (see Turner et al. 2003) that while some Indicators are only appropriate to administrative boundaries such as Forest Management Areas or 18

28 Bioregions, others are more site specific (for example, at the coupe level). This flexible distinction is important in the cost-effective monitoring of forest management outcomes. Australia became a signatory to the Montreal Process and has regularly reported progress towards addressing most of the 67 Indicators that are embodied in the Process (MPIG 1997, MPLO 2000, BRS 2003, MPIG 2008). Victoria has released two State of the Forests Reports for 2003 and 2008 (DSE 2005a, DSE 2009a). The focus of these reports, however, is on State and national reporting rather than on reporting a range of appropriate scales including at the Forest Management Unit (FMU) level. Such broadscale reporting is not directly related to forest management activities. Furthermore, baseline values for many of the Indicators are not generally available so that reporting on forest status in absolute terms is problematical. Also, the Indicator system so far has not adequately addressed the interaction between the various ecological processes. A comprehensive National Forest Policy Statement (NFPS) was released inter alia in response to the UNCED conference (NFPS 1992). The NFPS was developed by the Australian Forestry Council, and committed State, Territory and Federal Governments to work towards ESD. The NFPS acknowledged that there is no common definition of ESD, but noted that there are three main requirements, namely: Maintaining the ecological processes within the forests (the formation of soil, energy flows, and the carbon, nutrient and water cycles). Maintaining the biological diversity of forests, and Optimising the benefits to the community from all uses of forests within ecological constraints. The statement contains national goals that address conservation, wood production and industry development, integrated and coordinated decision-making, private native forests, plantations, water supply and catchment management, tourism and other economic and social opportunities, employment and workforce education and training, public awareness and education, research and development, and international responsibilities. SFM is an evolving concept and it can be argued that there is no unique definition because those stakeholders with a legitimate interest in a particular forest estate need to negotiate outcomes that reflect local goals and issues and in doing so, achieve a balance between social, economic, environmental and cultural considerations (Raison et al. 2001a). Such goals and considerations change over time so that a static approach is inappropriate. Importantly, SFM must be underpinned by the principle of adaptive management. This involves planning (setting goals and identifying Indicators), implementation, monitoring and evaluation (against Indicators), and review leading to adapted plans or guidelines. Use of this model will ensure that the forests progressively become better managed. Raison et al. (2001a) do not imply that stakeholders (including the broader community) have a role in approving on-ground performance against SFM Indicators. Indeed, this would not only be totally impracticable but also totally inappropriate. They merely stress the need for ownership of the C&I developed through appropriate consultative processes such as those used in the development of the Victorian Sustainability Charter (see below). A clear expectation of this process, however, is that progress towards meeting SFM objectives and requirements through adaptive management is regularly reported to stakeholders in a transparent manner. Victoria's Sustainable Forests (Timber) Act 2004 (Version incorporating amendments as at 1 January 2010) provides a framework for SFM and sustainable timber harvesting in State forests. The Act defines ESD as development that improves the total quality of life, both now and in the future, in a 19

29 way that maintains the ecological processes on which life depends, and amongst other things identifies three objectives of ESD and seven guiding principles. Section 6 of the Act requires the responsible Minister to determine Criteria and Indicators (C&I) for SFM, and these may take account of any nationally or internationally agreed C&I of SFM. The Minister must also determine reporting requirements relating to each Indicator, while the Secretary of DSE must report to the Minister on the status, performance or achievement in relation to the Indicators determined by the Minister within the specified time. To this end, the Secretary of DSE may require VicForests to provide information relating to specific Indicators. Under the Act, VicForests must develop initiatives and associated targets that respond to requirements under the Sustainability Charter, including some on-coupe measurements and monitoring (for example, regeneration surveys).. The reporting frequency must be a period not less than every five years. As noted earlier, and consistent with a requirement under the Sustainable Forests (Timber) Act 2004, Victoria has also developed a set of C&I that are intended to provide a framework for the monitoring and reporting on sustainability outcomes in State forests used for commercial wood production (DSE 2007b). The framework is closely aligned with national and Montreal C&I. Forty five Indicators were identified under the seven Montreal Criteria (DSE (2007b) with the reporting unit for each C&I being State forest, though potential sub-indicators were also identified, some of which could be applied at the FMU level. Consistent with the MLPO (2000) approach, Victoria adopted three Indicator Categories as follows: Category A, Category B. Indicators that can be reported against immediately for many areas; Indicators that can be measured for some areas of forest, but where there remains a methodological or resourcing issue; and Category C. Indicators where further R&D is required to determine if they can be implemented. For each Indicator, DSE (2007b) identifies potential sub-indicators and provides a template for reporting (including rationales for Indicator, issues to consider in reporting on the Indicator, data sources, potential data collection methods, and guidance on interpretation and implications of data). As already noted, the Act requires that reporting against C&I be done at regular intervals. To date, this is being accomplished through State of the Forests Reports. As previously discussed, in order for a forest estate to be well managed, there is a broad range of requirements that must be met within an acceptable framework such as the Montreal Process. At the present time, a full suite of SFM Indicators covering social, cultural, heritage, economic and environmental issues that can be reliably used to judge performance at appropriate scales including the FMU level, has not been developed. In the Australian context, one interim approach is to develop a modified sub-set of the Montreal Process Indicators and associated requirements in order to monitor performance at the regional level. Development of such Indicators has been the subject of intensive research (Turner et al. 2003). For a sub-set of Indicators to be credible, it is important that they are developed within an appropriate framework, that they are scientifically based and meet broad stakeholder approval (see Raison et al. 2001a), that they are flexible to accommodate adaptive management, and that they consider all land tenures within a defined FMU. In this regard, it is noteworthy that the Australian Montreal Process Implementation Group (MPIG 1997, MLPO 2000) accepted 30 of the 67 national Indicators as being relevant at the regional level, with a further 25 national-level Indicators being revised to better reflect regional issues. Twelve new Indicators were also developed. From this new set of 67 regional Indicators, MIG identified three sub-sets of Indicators comprising 12 Indicators that are largely implementable now, eight Indicators that require some development and 13 Indicators that require longer-term R&D. This and other processes have 20

30 clearly demonstrated the complexity of developing and monitoring/measuring sustainability Indicators, especially those that apply to regional or lesser scales. Indeed, there are numerous challenges that continue to be faced in progressing the successful implementation of C&I. The first ever Victorian State of the Forests Report was compiled in 2003 and published soon after (DSE 2005a). It was structured using Montreal Criteria and was intended to provide baseline information to assess and evaluate future performance and progress toward the achievement of SFM. A second report (Victoria s State of the Forests Report 2008) published in 2009 (DSE 2009a) complemented Australia s 2008 State of the Forests Report (MPIG 2008) and drew on the C&Is developed for Victorian conditions (DSE 2007b). However, these collective Reports address relatively broad Indicators that are best suited to reporting trends at the forest rather than (where relevant) at regional and FMU levels. Development of a full set of Indicators within each Criterion which can be reliably applied at the appropriate scale including the FMU level to gauge management performance from an SFM perspective, however, remains a major scientific challenge, and this is not acknowledged by the Sustainable Forests (Timber) Act Indeed, the Act has committed Victoria to an exceedingly ambitious undertaking in developing such Indicators given that Australian and global experience (for example, Turner et al. 2003) shows that much more research needs to be done to identify a broad set of Indicators that can be reliably applied at appropriate scales including at the FMU level. The identification of a coarse set of C&I (DSE 2007b) should not, therefore, be viewed as fulfilling the requirements of the Act. Furthermore, the application of Indicators will require an increased commitment to monitoring in order to meet the conditions specified in Sections 6 (3) (a) and (b) of the Act. As noted in Section 3.1, the Sustainable Forests (Timber) Act 2004 provides for the development of a Sustainability Charter. The Charter, which was published in 2006 (DSE 2006a), requires DSE, DPI and VicForests to manage Victoria s State forests in accordance with seven objectives that are aligned to the DSE Criteria (DSE 2007b) and the national principles of ESD. The overall aim of the Charter is to provide for the sustainability of forests and the sustainability of the timber harvesting industry. It will be reviewed at 5-yearly intervals. Progress towards achieving the objectives of the Charter will be communicated through a number of channels including regular State of the Forests Reports and third party auditing. A flow on from international SFM processes has been certification. To this end, an Australian Forestry Standard (AFS) was developed under the auspices of Standards Australia and within the Montreal and ISO frameworks by AFS Ltd with the assistance of a Technical Reference Committee comprising a broad range of stakeholders (AFS 2007). The AFS provides Forest Management Organizations with economic, social, environmental, cultural and heritage Criteria and Requirements that need to be met for well-managed forests. The AFS has international recognition under the Programme for the Endorsement of Forest Certification Schemes. VicForests is currently certified under the AFS. It is significant to note that the AFS focuses on continuous improvement and makes no claim that meeting all of it s Requirements translates to a sustainably managed forests, despite the demanding Requirements that must be met at both the forest and FMU levels to gain certification. 21

31 3.6 Discussion and Conclusions 1. In 2002, DSE estimated that the area of economically accessible sawlog in Victoria was 696,000 hectares of which 539,000 hectares were in eastern Victoria and the remaining 137,000 hectares in western Victoria. The separation of the Government s commercial and policy, monitoring and regulatory roles resulted in VicForests taking responsibility for managing commercial timber harvesting in eastern Victoria and commercial timber harvesting virtually ceasing in western Victoria. Since 2002, extensions to the area of National Parks in East Gippsland and major fires in 2003, and 2009 have killed large areas of mature and regrowth mountain ash and alpine ash forests. The impact of the events on the long term harvest level is unclear at this stage. 2. The impact of various silvicultural systems on future productivity is poorly understood. This particularly applies to any systems (for example, those that prescribe high levels of overwood retention) that have a strong bias towards above-normal protection of a range of forest values that are arguably already largely catered for in reserves and special management zones identified in the forest management planning process (refer also to Section 7). 3. The policy, planning and regulatory framework for native forest management is very complex. 4. The most recent State of the Forests Report for Victoria (DSE 2009a) only provides a very broad overview of the main findings for reporting progress against the 45 Indicators being used to assess SFM in State forests. Whilst some extremely useful and informative data are presented in the Report, the data are incomplete (due to a variety of reasons including lack of required monitoring and management systems which clearly must be intensified as a matter of priority if the pursuit of SFM is a genuine objective of Government). Furthermore, reporting is primarily at the State or State forest levels rather than (where appropriate) at the FMU level (for example, an individual coupe would be the finest scale), so that the information is not sufficiently detailed to gauge whether or not the forests are being sustainably managed. However, Vicforests are certified under the AFS which indicates that appropriate information is available to demonstrate that State forests in eastern Victoria are being well-managed. 5. The concept of sustainable management of Australia's forests has gained momentum in the past decade, initially through Australia's formal involvement in the Montreal Process. This has been a positive development. As would be expected from the complex nature of measuring and monitoring of SFM Indicators, reporting against the full suite of Indicators (applicable to Australia) embodied in this and similar global processes at appropriate scales is proving to be a major challenge. As an example, Australia's latest State of the Forests Report (see MPIG 2008) shows that much work needs to be done before meaningful and valuable reporting against all relevant Indicators is achievable at the State and FMU levels (see also DSE 2009a). The situation at the FMU level is particularly complex and will require significantly more research before even a sub-set of Indicators can be used with confidence to judge on-ground performance from an SFM perspective. For example, in the case of Protection of Soil Values (just one of many SFM requirements), the authors found no evidence that statistically valid programs are in place to monitor soil compaction, soil disturbance, soil fertility/productive capacity and soil erosion at acceptable spatial (coupe) and temporal scales. Such programs are not possible given current knowledge. In the meantime, therefore, processes need to be put in place to monitor, evaluate and review practices as part of an adaptive forest management approach. 22

32 4. COMMERCIALLY IMPORTANT FOREST TYPES AND THEIR SILVICULTURAL MANAGEMENT 4.1 Forest Types Several systems are used to classify forest types and floristic communities in Victoria, the variation depending on purpose and level of information available. As an example, Flinn and Bales (1990) present a stylised diagram to show the influence of rainfall and elevation on broad floristic communities (including montane forests and lowland, dry, damp and wet sclerophyll forests) and major tree species [including Silvertop Ash (E. seiberi), Messmate (E. obliqua), Cut-tail (E. fastigata), Mountain Ash and Alpine Ash] in east Gippsland (Figure 4.1). At a more refined scale, Ecological Vegetation Communities can be used to describe the floristics of Victoria's native forests. At an even more detailed floristic level, Mueck (1990) provides a comprehensive insight into the many floristic sub-communities of Alpine Ash forests and Mountain Ash forests. ELEVATION Figure 4.1 Broad relationship of forest communities in East Gippsland based on moisture and elevation (Y-axis) (after Flinn and Bales 1990). Various classification systems have been used to describe these forests on the basis of: forest structure or form (NFI 2003); type of timber (HARIS used by Grierson et al. 1992); silvicultural type (for example, Bennett and Adams 2004a); and dominant eucalypt species groups (that is,. 'EUCGROUP' as used in SFRI; DSE 2003a). The diversity of these forests is illustrated by a comparison of three classification systems (Table 4.1). 23

33 Table 4.1 Comparison of three forest vegetation classification systems used in Victoria. HARIS 1 Dominant overstorey species 1 Broad Forest Types or Silvicultural 2 Alpine Ash (AA) Mountain Ash (MA) Mountain Mixed Species (MMS) Shining Gum (SHG) Alpine Mixed Species (AMS) Foothill mixed species (FMS) Coastal mixed species (CMS) Box-Ironbark (BIB) River Red Gum (RRG) Predominantly Alpine Ash in pure stands. Other species may include Mountain Gum Predominantly Mountain Ash in pure stands. Other species may include Shining Gum and Alpine Ash at higher elevations or Messmate, Mountain Grey Gum and Manna Gum at lower elevations. Predominantly Cut-tail, Messmate, Mountain Grey Gum either in pure stands or in mixtures. Other species include Manna Gum, Blue Gum, and Candlebark where mature stand height is generally greater than 40 m Predominantly Shining Gum in pure stands, also in mixed stands. Generally peppermint/ gum stands at high elevations (e.g. Mountain Gum) Predominantly Messmate, Manna Gum with some or all of Candlebark, Broadleaved Peppermint, Silvertop Ash and Narrow-leaved Peppermint where mature stand height is generally greater than 28 m Predominantly Silvertop Ash and White Stringybark in a mixture with other species Stands of box or ironbark eucalypts (e.g. Red Ironbark, Red Box, Yellow Gum, Grey Box), also on occasion in association with lesser species such as Red Stringybark, Yellow Box, White box and Long-leaved Box Predominantly River Red Gum in pure stands, with box species in close association. Vegetation form 3 Alpine Ash Tall open forest 4 Mountain Ash Tall open forest 4 High elevation mixed species (HEMS) Tall open forest 4 HEMS Tall open forest 4 HEMS Open forest 5 Low elevation mixed species (LEMS) Open forest 5 LEMS Open forest 5 Box-Ironbark (BIB) River Red Gum (RRG) Open forest 5 Open forest 5 or Woodlands 6 1. Hardwood Resource Information System ('HARIS'; Grierson et al. 1992) Mountain Ash is E. regnans, Shining Gum is E. nitens and E. denticulata, Alpine Ash is E. delegatensis, Messmate is E. obliqua, Mountain Grey Gum is E. cypellocarpa, Manna Gum is E. viminalis, Cut-tail is E. fastigata, Blue Gum is predominantly E. bicostata, Candlebark is E. rubida, Mountain Gum is E. dalrympleana, Broad-leaved Peppermint is predominantly E. dives, Silvertop Ash is E. sieberi, Narrow-leaved Peppermint is E. radiata and White Stringybark is E. globoidea, Ironbark is predominantly (Red) E.tricarpa but also includes (Mugga) E.sideroxylon, Red Box is E. polyanthemos, Yellow Gum is E. leucoxylon, Grey Box is E. microcarpa, Red Stringybark is E. macrorhyncha, Yellow Box is E. melliodora, White Box is E. albens and Long-leaved Box is E. goniocalyx. 2. Victorian silvicultural forest types (Bennett and Adams 2004a). 3. Vegetation Growth Forms (NFI 2003) 4. Tall open forest: projected foliage cover of tallest stratum 30-70%, trees > 30 m (formerly Wet Sclerophyll Forest). 5. Open forest: projected foliage cover of tallest stratum 30-70%, trees m (formerly Dry Sclerophyll Forest) 6. Woodland: projected foliage cover of tallest stratum 10-30%, trees m 24

34 In the present review, commonly used broad forest types have been adopted largely on the basis of silvicultural considerations. Four main forest types account for the majority of the timber production in Victoria, with a further two forest types having significant silvicultural or ecological interest. The four main forest types are Alpine Ash, Mountain Ash, High Elevation Mixed Species (HEMS) and Low Elevation Mixed Species (LEMS) while the other two forest types are Box-Ironbark and River Red Gum. Essentially, HEMS and LEMS forest types are, as noted above, largely subdivided on the basis of rainfall, altitude and speciation. This contrasts with NSW systems which are based essentially on species and structure of forest types (for example, Anon. 1965, Keith and Sanders 1990) and hence any comparison between States needs to be based on species composition and structure. As noted in Section 2, the Box-Ironbark forests and the River Red Gum forests while economically minor now, made a significant contribution in Victoria's development in the earlier years following European settlement. The ecological characteristics of these six broad forest types have previously been described in detail (for example, Campbell et al. 1984). The Mountain Ash forests occur at elevations from as low as 120 metres to more than 1,100 metres in the central and eastern Highlands and the Otway and Strzelecki Ranges. In contrast, the Alpine Ash forests occur at higher altitudes where snow is a dominant feature in winter and spring, extending from the lower reaches of alpine vegetation down to around 1,000 metres. A feature of the Alpine Ash forests is their scattered distribution, occurring at a number of discrete locations across the eastern Highlands (for example, Marysville, Mt Stirling, Baw Baw, Mitta Mitta, Nunniong Plateau and the Errinundra Plateau in the far east). The HEMS forests occupy extensive upland areas of eastern and southern Victoria. The LEMS forests, however, are by far the most extensive and varied forest type used for commercial wood production. In this review, LEMS forests include the extensive foothill and coastal forests south of the Great Dividing Range in east Gippsland (also commonly known as the Silvertop Ash/Stringbark forests) and the Messmate/Peppermint forests of central Victoria (for example, Mt Cole), northeast Victoria (for example, Strathbogie Ranges) and southwest Victoria (for example, Heywood). Box-Ironbark forests, comprising Red Ironbark (E. tricarpa) and a range of other species that tolerate low rainfall conditions {for example, Grey Box (E. microcarpa) and Yellow Gum (E. leucoxylon)} were once quite extensive but are now largely confined to around 240,000 hectares in north-central Victoria (Lutze et al. 1999a). Victoria s River Red Gum forests are largely associated with the Murray, Goulburn and Ovens Rivers. The forests are strongly influenced by the flooding frequency and vary from treeless wetlands to Black Box (E. largiflorens) and Murray Pine (Callitris) woodlands to highly productive grassy riverine River Red Gum forests. In addition, a non-riverine occurrence of River Red Gum can be found in a forest formation near Woohlpooer in western Victoria. The current silvicultural review is largely confined to the forests of eastern Victoria with the exception of the Box-Ironbark forests and the River Red Gum forests, as eastern Victoria accounts for most of the wood production in the State. It is important to note, however, that the silviculture of mixed species forests in western Victoria has been the subject of significant research, including detailed silvical studies (see for example Kellas 1994). 25

35 4.2 Introduction to Silvicultural Principles Silviculture involves the establishment, tending, maintenance and harvesting of the forest crop. In the current review, the main focus is directed to regeneration requirements and early stand development of the main commercial species, along with systems used to harvest the forests. Silvicultural practice in any forest is dictated by the structure and function of the forest ecosystem. Identification of an appropriate silvicultural system or systems for a given forest type requires taking a number of factors into account in addition to management objectives. They include: The silvics of the eucalypt species present on the harvested area (for example, flowering and seeding cycles, seed germination requirements such as stratification, and tolerance to shade). Occupational Health & Safety issues in the knowledge that felling trees in tall forests is an extremely hazardous occupation, particularly for single tree or small gap selection systems. [NB. WorkSafe Victoria, in partnership with forest industry stakeholders including the CFMEU, DSE and VicForests, developed an Industry Standard for Safety in Forestry Operations: Harvesting and Haulage (WorkSafe Victoria 2007). This Standard recognised that workers in harvesting and haulage experience a high proportion of fatalities considering their relatively small-sized workforce]. Wood quality considerations such as damage to retained trees during felling and extraction and stimulation of epicormic shoot development following exposure. Current market conditions and cost-effectiveness of the preferred system, and Minimisation of environmental impacts and maintenance of site productive capacity. Essentially, these factors represent four management issues: What is required for any forest type to obtain adequate regeneration and acceptable growth? What activities should be undertaken to maintain long-term health and productivity and protect the environment? What type of thinning can be undertaken and what are the impacts? Which systems and harvest technologies may be used at final harvest? Ryan (1997) provides an alternative but similar framework for choosing a silvicultural system. He notes that many of the factors to be considered in identifying potential systems need to be determined by pre-harvest assessment. The AFS also addresses this issue in some detail (AFS 2007). Irrespective of which silvicultural system is adopted, there are three fundamental requirements that must be met to achieve successful eucalypt regeneration by seeding and satisfy a key SFM indicator - a receptive seedbed, an adequate supply of viable seeds, and favourable conditions for seedling establishment (for example, prevention from browsing). The seed sources (for example, local provenance) may be specified where supplementary seeding is required. 26

36 4.3 Silvicultural Research and Practice The silvicultural management of Australia s east coast forests has been the subject of on-going debate since large-scale integrated harvesting commenced in the forests of southeast NSW (Routley and Routley 1975). The main focus of debate has been on harvesting and regeneration rather than issues such as rotation lengths, thinning regimes and forest protection. Public and scientific attitudes to forestry activities vary greatly, and clearfelling has been a major issue. While it has been a long held view by some stakeholders that clearfelling and slash burning in montane forests mimicked wildfire, others (Lindenmayer et al. 1990, Ough and Ross 1992 and Ough and Murphy 1999) present convincing ecological evidence to the contrary, particularly in relation to impacts on the spatial arrangement, abundance and longevity of hollow-bearing trees and tree ferns. The need to continuously develop and improve silvicultural practices to satisfy evolving socioeconomic requirements (Campbell 1997a, b, Bauhus 1999) is becoming more complex due to a combination of a reduction in areas available for timber production (putting more pressure on forests set aside for wood production) and rapidly improving silvicultural and ecological knowledge which needs to be accommodated. Forest research has a relatively short history in Victoria. A formal Research Section was established within the Forests Commission Victoria (FCV) in the mid 1950s. By the mid 1960s, a Forestry Education and Research Division was in place and the focus of research was on silviculture, pathology and entomology (Ferguson 1965). Research on these and other disciplines including genetics, hydrology, nutrition and physiology has subsequently been responsible for hundreds of published and unpublished technical reports, scientific papers and bulletins being produced (see for example, Murphy et al. 1988). Valuable research has also been conducted at the University of Melbourne, often in collaboration with FCV. The Victorian Timber Industry Strategy of 1986 made a major commitment to research and development on silvicultural systems alternative to clearfelling and to explore the potential for commercial thinning of regrowth stands to improve future sawlog production (Flinn and Mamers 1991). This strategy led to the Silvicultural Systems Project (SSP) which involved the establishment of major research sites in Mountain Ash forests near Tanjil Bren (Squire 1987, Campbell 1997a, b) and in Low Elevation Mixed Species forests at Cabbage Tree Creek in east Gippsland (Squire et al. 2006). It is noteworthy that Tasmania initiated a similar study some years later at its Warra Long Term Ecological Research Site (Brown et al. 2001). In 2002, a Round Table meeting was held to provide a forum for a targeted number of key stakeholders to engage in dialogue on forest management issues in the Victorian Central Highlands (Lindenmayer et al. 2004). There were several important outcomes from the meeting, but the dominant issue of relevance to the present review was new silvicultural systems for montane ash forests. The meeting was critical of the RFA (Regional Forest Agreement) process for failing to recognise the ongoing need to investigate and adopt silvicultural systems that have greater congruence with natural disturbance regimes, as alternatives or improvements to clearfelling. At the very least, the meeting believed that systems for Alpine Ash and Mountain Ash that are more ecologically sensitive in parts of the logged forest are essential to achieve more balanced economic and environmental outcomes, with better protection for special values (for example, ecological, landscape, water). New studies were commenced in the region within three months of the meeting (Lindenmayer et al. 2004), and initial progress was reported nearly four years later (Lindenmayer 2007). The round table meeting is just one of many Victorian examples of gaining broad stakeholder input to identify practical and scientifically-sound management options that work towards SFM. 27

37 As noted by Campbell (1997a, b), silvicultural systems are often named to reflect the spatial and temporal distribution of harvested trees. Whilst this terminology implies a separation of systems, most systems essentially fall on one of two continuums that relate to either the size of the gap (small to large) or the level of retained overwood (none to high). Studies such as SSP in Victoria and Warra in Tasmania have provided managers with alternative approaches to the traditional clearfell system involving large and often contiguous coupes in Mountain Ash and LEMS forests and old-growth Tasmanian forests. For example, the Tasmanian research proposed the adoption of a mixed silvicultural approach with variable levels of retained overwood, including retention of unharvested aggregates within coupes, clearfelling in steep country (but using smaller coupes), and group or single tree selection in special areas providing it is safe and cost-effective. SSP findings follow a similar pattern (Campbell 1997a, b). An account of the development of silvicultural practices in native State forests in Victoria was provided by Lutze et al. (1999). For essentially the same six commercial forest types as considered in the current review, they provided detailed data for 1997/98 on areas treated for regeneration and/or thinning, and the proportion of area harvested by different silvicultural systems. The analysis showed that of the 12,080 hectares treated, nearly 50% of this was attributed to LEMS and HEMS forests where clearfelling and seed-tree systems dominated. This contrasted with River Red Gum and Box- Ironbark forests where 1,230 and 3,500 hectares respectively were subjected to either harvesting (by either single tree or group selection systems) or thinning (50% and 70% respectively). Victoria recently developed a Policy Statement on thinning in State forests (DSE 2008a). Thinning for commercial purposes has a long history in Victoria, and both DSE and VicForests are working towards expanding the thinning of regrowth forests originating from either wildfire or past logging to improve long-term forest productivity. This claim, however, is misleading. Thinning does not affect long-term productivity per se, but by capturing growth otherwise lost to mortality of suppressed stems, thinning can increase total merchantable volume. Commercial thinning is usually conducted to concentrate growth on retained stems, thereby producing larger diameter logs for a given rotation length and minimising merchantable productivity loss resulting from natural mortality of mainly suppressed or co-dominant trees. Thinning also provides a yield of timber early in a rotation, though, (as stressed in the Policy Statement) the aim of the practice should be to maximise total economic benefit. This objective can only be realised, however, by adopting practices that minimise damage to retained stems. As noted in the Policy Statement, thinning can be undertaken for a variety of other purposes including increased water yield and the provision of habitat. Thinning, however, may be associated with adverse environmental impacts, particularly if operations are poorly conducted (see for example, Flinn and Mamers 1991, Peacock 2006). Thinning Guidelines have therefore been developed (see Appendix 1) to minimise any adverse impacts of thinning, including environmental effects and damage to retained trees. A workshop was held in 2002 to review current knowledge and experience with thinning programs being conducted in State forests (DSE 2006b). At that time, thinning was primarily being undertaken in mixed species forests, namely Mountain Ash forests, River Red Gum forests and Box-Ironbark forests where significant areas were being thinned on an annual basis. Specific issues addressed at the workshop included growth responses to thinning, the extent of damage being sustained during routine thinning programs in the various forest types, and harvesting systems (recognising that topography and harvest techniques have a strong influence on damage which is an important SFM consideration). 28

38 4.4 Alpine Ash Forests Introduction The Victorian Alpine Ash forests are widely distributed in numerous separate areas of the eastern Highlands where they generally occur as pure stands following past wildfires or logging. At the margins of its natural range, Alpine Ash can be found in mixtures with Mountain Ash, Shining Gum (E. nitens or denticulata), Mountain Gum (E. dalrympleana), Cut-tail, Mountain Grey Gum (E. cypellocarpa), Narrow-leaved Peppermint (E. radiata) and Snow Gum (E. pauciflora) (Campbell et al. 1984). It is sensitive to fire and is generally killed by even moderately intense fire (Grose 1957, Wareing and Flinn 2003, Flinn et al. 2008). The stringy bark extends well up the main stem and adds to the overall fuel hazard. Providing there is a source of viable seed in the tree crowns, regeneration following wildfire is generally prolific leading to even-aged stands. However, multi-aged stands are occasionally found if low intensity ground fires stimulate regeneration in over-mature or old stands without killing all of the overwood, as noted by Chesterfield (1978) in forests surrounding the Benisons Plains. Alpine Ash neither coppices readily from cut stumps nor produces lignotubers. Significant harvesting activity in the Alpine Ash forests did not commence until the end of salvage logging of the 1939 fire-killed Mountain Ash forests and an increasing demand for timber products. The initial silvicultural system took the form of clearcutting with the retention of single trees, or groups or strips of trees variously distributed, as seed sources for regeneration (Grose 1957). The system, however, failed to deliver consistently satisfactory regeneration from both density and distribution viewpoints. This precipitated an intensive study of Alpine Ash silviculture (Grose 1957, 1960a, 1963). The first formal field study commenced in 1954 in northeast Victoria, and this was followed by extensive laboratory studies and complementary field studies. A detailed review of this research provides an excellent case study demonstrating the complexities of developing reliable silvicultural systems for the native forests of Victoria Silviculture of Alpine Ash in NSW and Tasmania Before considering the results from Victorian research, it is instructive to briefly review research on Alpine Ash in Tasmania and NSW, primarily because silvicultural systems alternative to clearfelling have been successfully employed in some of these forests. Furthermore, Alpine Ash is being used as a case study in the present review, thereby warranting the broader treatment of its silviculture in the other States. There are differences in the silvics of the species between Tasmania and the mainland and this is reflected in the differences in provenance characteristics (Boland and Dunn 1985). Boland (1985) divided the species into E. delegatensis subspecies delegatensis on the mainland and E. delegatensis subspecies tasmaniensis in Tasmania Tasmania Tasmanian research was mainly initiated in response to limited regeneration associated with clearfelling, slash burning and aerial seeding. There were two main problem areas (Keenan 1986). The first, in northern Tasmania, was characterised by a grass or shrub understorey where inadequate regeneration was common. Furthermore, where regeneration was obtained, it often exhibited poor form. The second problem area was on the Central dolerite Plateau where, according to Bowman and Kirkpatrick (1984), moisture supply appeared to be an important factor in determining understorey vegetation. Many of the Alpine Ash stands had been selectively logged, heavily burnt and subjected to grazing, giving rise to multi-aged stands and it was concluded that a shelterwood system may be more appropriate than clearfelling for these problem sites (Keenan 1986). Burning practices by Aboriginal peoples would have also contributed to the multi-aged structure (Ellis et al. 1987). 29

39 A review similar to that of Keenan (1986) of silvicultural systems for the Central Plateau was undertaken by Bowman (1986). Based on the work of Bowman and Kirkpatrick (1986), he noted that Alpine Ash in Tasmania was far more resistant to fire than Alpine Ash in Victoria and NSW, giving rise to the previously noted multi-aged stands (Keenan 1986). Bowman (1986) proposed a model to explain the ecology of Alpine Ash on the dolerite Plateau. The model postulated that survival and early growth of dense regeneration following fire was determined by the degree of intra-specific competition. Rapid early growth protected the better stems from fire through bark development which insulated the cambium. If the stems were defoliated, they then resumed growth from epicormic shoots while if they were killed by fire, they could regenerate from basal sprouts, unlike suppressed stems. Consistent with Keenan (1986), Bowman (1986) also reported that shelterwood systems may be appropriate and need to be researched, particularly given that frosts often severely damage regeneration on these sites. Shelterwood systems have been implemented. Ellis et al. (1987) noted that, given the multi-aged structure of many Alpine Ash stands on the Central Plateau, clearfelling and burning not only often failed to obtain adequate regeneration (or resulted in regeneration that exhibited so-called growth check, comprising severe growth retardation and poor form which can persist for over a decade), but the method also sacrificed smaller regrowth stems which had future sawlog potential. On the other hand, they observed that small gap group selection on these high altitude sites usually resulted in good regeneration without the troublesome snow grass or shrub/herb development which appeared to be associated with growth check. The cause or causes of growth check are not apparent from the more recent work of Neyland and Cunningham (2004). Given this background, Ellis et al. (1987) investigated a partial cutting treatment for the multiaged and often degraded stands (mainly due to selective logging, wildfires and epicormic shoot development). This treatment removed most trees which had no sawlog potential ensuring that the multi-aged structure was retained. Four years after the partial-cut logging, it was found that trees with healthy crowns were more responsive to release than those with unhealthy crowns, irrespective of size class. In relation to tree size, smaller trees were found to be more responsive than larger trees to release and more susceptible to suppression. Whilst clearfelling, slash burning and seeding continued to be a common method for successfully regenerating the tall wet Alpine Ash forests of Tasmania, research on alternative systems commenced in the highland dry forests (Neyland and Cunningham 2004) following on from work of Ellis et al. (1987). These systems included shelterwood retention, shelterwood removal, retention of potential sawlogs (as proposed by Ellis et al. 1987) and retention of advanced regrowth on the basis that partial canopy retention is essential to prevent growth check. The benefits of partial harvesting systems compared with the clearfell and slash burn method for Alpine Ash were discussed by Neyland and Cunningham (2004). Such systems became increasingly accepted in the 1980s and 1990s, but monitoring of regeneration success was inadequate and this led to several problems (such as failure to remove culls, failure to undertake the second shelterwood cut, and excessive retention of mature trees). A new monitoring system (termed the uneven-aged treatment procedure or UAT) was developed, involving assessing the pre-harvest forest structure, guiding the development of the harvesting prescription, monitoring the harvesting operation and providing feedback to the harvesting contractors within a continuous improvement framework (Forestry Commission Tasmania 1990). The proposed system had strong scientific merit but required an on-going commitment to detailed monitoring of forest operations. Hickey and Wilkinson (1999) provided an historical account of silvicultural treatments in Tasmania and a review of silvicultural systems for State and private lands, recognising that unlike Victoria, large tracts of native forests in Tasmania are in private ownership. They found that for a 10-year period from 1988/89 onwards, the mean annual percentage of native forest harvesting in State forests was 30

40 45% partial logging systems (including advanced growth retention, seed-tree, shelterwood, and group and single tree selection) and 39% clearfell method, with the balance converted to plantations. On private land, the partial systems were dominant New South Wales In NSW, little research has been published on Alpine Ash silviculture, which is primarily located in Bago and Maragle State Forests. One possible reason is that few problems have been experienced in obtaining regeneration. Both group selection and seed-tree systems have been effective in management of this forest type (Horne and Robinson 1990). In spacing studies (considered to be a first thinning) in 28-year-old fire regrowth where nominal spacings of 188, 269, 416 and 770 stems per hectare were tested, it was found that only small gains were made in stand growth (Horne and Robinson 1990). This was because the stands were fast growing leading to rapid natural self-thinning in both thinned and unthinned stands. Regeneration of Alpine Ash in Bago, NSW, unlike Victorian Alpine Ash, often occurred in small groups which reflected the selective nature of early logging operations in these forests (FCNSW 1986). The presence of these small, even-aged patches has shaped present-day silvicultural practices in the locality, namely group selection with logging disturbance providing the seedbed and retained trees adjacent to the gaps being the seed source. Successful regeneration has been the norm with this approach. However, for the Maragle State Forest (on the southern fall of Bago State Forest) which comprises virgin stands in much steeper country than the Bago Plateau and more closely resembles the Victorian situation, a seed-tree system and slash burning has been adopted. Thus, in NSW, Alpine Ash regeneration is obtained by either group selection or by seed-tree systems, with the choice of system being largely dictated by the history of the stands (particularly logging activity) and the nature of the topography. Figure 4.2 Alpine Ash in NSW about 1914 (see Dalrymple Hay 1915). 31

41 4.4.3 Silviculture of Alpine Ash in Victoria A Detailed Case Study In Victoria, early attempts at regeneration using a seed-tree system and mechanical disturbance were often unsuccessful and this led to wide ranging and detailed studies (Grose 1957, Grose 1960a, b, Grose 1963, Grose et al. 1964, Grose 1965). These studies have been reviewed by Campbell et al. (1984) and FCNSW (1986). The field studies by Grose (1957) were undertaken in approximately 105- year-old stands (probably originating from the 1851 fires) and some significant-sized patches of approximately 58-year-old stands (possibly originating from the 1898 fires) at Delatite in northeast Victoria. The study area generally has permanent snow cover in winter. Major timber harvesting did not commence in the area until Initial regeneration surveys indicated that seedbed and seed supply may have been critical but not exclusive factors influencing regeneration Flowering, Fruit Development and Seeding Inflorescence buds generally appear in the leaf axils from October to December. Their bracts are shed about a year later as the flower buds continue to develop for a further months. As a consequence, flowering generally occurs from January to March, or around two years after bud initiation. Flowering is not uniform between years, and heavy flowering is irregular. There are several other important features of seeding that are highly relevant to regeneration: Based on extensive observations of the heads of felled trees, little seed is provided by suppressed or even intermediate trees in mature stands. Seed-tray studies showed that over a 2-year period ( ) in mature stands, total seeds cast on northwesterly (exposed) aspects were 4.72 million seeds per acre compared with 2.13 million seeds per acre on south-easterly (sheltered) aspects. This was attributed to differences in capsule abundance and in number of seeds per capsule between aspects and has obvious implications in terms of the number of trees to be retained if a seed-tree system is adopted. In a mature stand aged about 110 years, the average annual seed cast was found to be one to two million seeds per hectare which was around six times that cast from stands aged 55 to 60 years. Over a 2-year study period, more seeds were usually cast from capsules (that is, free seed) than were cast in capsules. Seeds cast in capsules are either eaten by insects (mainly ants) or lose their viability. Peak seed cast occurs in late summer and early autumn, with most seed from a particular flowering event being shed in the third year following flowering, though rapid seed cast can be induced by ringbarking (sap-ringing), stem injection with a herbicide or a hot burn (including wildfire). Natural seed shed does not usually commence until about two years after flowering, although a small fraction of the seed crop can be cast on branchlets following wind break or bird damage much earlier. There is significant variation in the number of fertile seeds per capsule between trees. For 40 randomly selected trees over a range of sites across Victoria, the maximum mean count per capsule was 7.1 while the minimum mean count was 1.7. The overall mean number of fertile seeds per capsule was 3.7. Significant seed is lost to small wasps which eat the embryos of viable seeds in the capsules before they are cast. Seedfall can extend over a period of up to three or four years so that significant seed reserves can be held in the tree crowns through accumulation from at least three successive years of flowering. This is important for the regeneration of Alpine Ash after wildfire as it may safeguard against a fire occurring immediately following a poor seed year. 32

42 The effective distance of seed dissemination from seed-trees appears to be equal to around one tree height, and Regeneration relies on free seed (non-capsule seed). However, 50% to 75% of the free seed can be harvested by insects [particularly the lygacid bug (Dieuchis notatus) and an ant (Tridomyrmex foctans)], while those that fail to germinate in the first spring following seedfall become non-viable. Therefore, seed is not stored on the forest floor and this has important implications for regeneration. Clearly, the previous features are critical in identifying regeneration systems for Alpine Ash (that rely on seed rather than planting) as part of their overall silviculture Germination Provision of optimal conditions for the germination of seed is fundamental to identifying silvicultural systems that reliably lead to effective regeneration. Major findings from laboratory studies by Grose (1963, 1965) were: Most Alpine Ash seeds are dormant (that is, they have a primary dormancy), irrespective of their provenance (location), and this acts as a survival mechanism by preventing autumn and early winter germination and hence massive seedling losses in the first winter due primarily to snow. The primary dormancy can be readily broken by moist stratification for a minimum of four weeks at constant temperatures ranging from 1 o C-9 o C, with 3 o C-5 o C being the optimum range. This treatment can also extend the range of temperatures over which the stratified seeds will germinate. Ideally, stratification needs to take place over an 8 to 10 week period to ensure a high proportion of germination (seeds stratified for a long period will germinate better over a wider range of temperatures compared with those stratified for short periods). Secondary dormancy can be induced when after-ripened (stratified) seeds are stored at temperatures and moisture contents unfavourable for germination. Stoneman (1994) undertook a detailed review of the ecology and physiology of establishment of eucalypt seedlings from seed. Harvesting (mostly by ants), soil moisture conditions, air humidity, and seedbed conditions were generally found to be the most important factors dictating germination and emergence (depending on the species), while the factors of most importance in terms of mortality of germinants were water, temperature and light (all of which are influenced by the density of overstorey and understorey vegetation) and nutrients. Seed longevity is an important issue in relation to seed banks for the raising of seedlings and for aerial seeding. Seed longevity was found to be substantially reduced by afterripening, whereas air dry, non-stratified seeds retain their viability under proper storage conditions for at least seven years. Seed size appears to influence early vegetative growth. Large seeds not only have a weaker primary dormancy but also larger cotyledons, giving them a better chance of producing seedlings with enhanced prospects of field survival, and As previously noted, under field conditions, Alpine Ash seeds are shed in late summer/early autumn. Most of these seeds do not germinate until mid to late spring after the over-winter period has broken the primary dormancy. 33

43 Seedling Establishment Like many eucalypts, Alpine Ash requires a receptive seedbed for germination of seed. A heavy grass sward does not constitute such a seedbed, nor does compacted bare soil and undisturbed litter. Grose (1957) compared six seedbed types including ashbeds produced by slash burning and various unburnt, loosely cultivated seedbeds. After 13 to 14 months, percentage survival on ashbeds was significantly higher than on the other five seedbeds though on disturbed soils, seedbeds still resulted in satisfactory seedling survival. Despite favourable seedbed conditions, losses of spring germinants in the cotyledon and two leaf stages can be as high as 60%, mainly due to root exposure by water action and frost heave, decapitation by ice action, and desiccation. Nonetheless, sufficient germinants are able to survive and develop into established seedlings by their first winter, and the ability of these seedlings to survive the first winter is largely determined by the interacting factors of seedbed type and site exposure. In contrast, dense germination can occur on undisturbed seedbeds, but this does not lead to any established seedlings (Grose 1960b). Aspect is also an important influencing factor because percentage of germination on sheltered sites can be at least double that on exposed sites. This in effect compensates for the differential seed production between sheltered and exposed aspects. Deaths of seedlings in their first winter appear to be primarily due to prolonged saturation of the intercellular spaces of the leaves with water from melted snow. The first 3 or 4 leaf pairs tend to be horizontally aligned and non-glaucous whereas subsequent leaves hang more or less vertically and are glaucous so that they more readily shed water. Seedling size therefore is important because larger seedlings, as produced by ashbeds, are less likely to be flattened by snow and in turn exposed to prolonged saturation leading to water injection. The end result of these interacting factors is that seedlings on ashbeds are better able to survive their first winter compared with seedlings on loose bare soil. Furthermore, ashbed seedlings are taller and equally glaucous irrespective of aspect influences whereas seedlings on loose bare soil seedbeds under sheltered conditions (due to aspect or shade from vegetation) are considerably less glaucous than those grown on exposed sites. This interaction between seedbed and exposure was important in developing successful silvicultural systems for the species. Grose et al. (1964) stated that an ashbed is the only effective seedbed for sheltered sites. A more recent study in a sub-alpine stand of Alpine Ash in the Snowy River National Park (Victoria) also re-confirmed that receptive seedbed conditions were required for successful regeneration (Florentine et al. 2008). Under the conditions of their study, soil disturbance resulting from severe windthrow followed by intense wildfire provided favourable conditions for both seed germination and enhanced early seedling development Silvicultural Systems Prior to the above pioneering work by Grose (1957, 1960a, b, 1963, 1965), little success was achieved in consistently obtaining regeneration following logging of mature Alpine Ash stands. A seed-tree system with around 15 retained trees per hectare, however, soon became commonly adopted with seedbeds being prepared by slash burning and seedfall induced by either burning, sapringing or stem injection. The seed-trees were then removed. But this two-stage operation presented operational challenges similar to those experienced in Tasmania before the introduction of the UAT (Uneven Aged Treatment) procedure. Victoria therefore moved to clearfelling, slash burning wherever possible (and mechanical disturbance when a good burn could not be achieved) followed by hand or aerial application of around 125,000 viable seeds per hectare. Aerial seeding provided the opportunity to sow large areas in a short timeframe before seedbeds lost their receptivity, and to obtain a more uniform distribution of seed compared with hand sowing techniques (Grose et al. 1964). 34

44 Subsequent work led to a refinement of sowing rates according to seedbed type and distribution (proportion of a coupe with a receptive seedbed) and other relevant factors detailed above such as aspect influences. Seed coating was also introduced to improve seed distribution and reduce harvest losses from insects, although this practice is no longer adopted. A continuous improvement approach has been adopted in Victoria to ensure a high standard of regeneration. The process involves monitoring of regeneration and evaluating the results against stocking standards that are consistent with SFM principles. Where required, regeneration practices are then reviewed to take account of findings from the evaluation and any new research information. Stocking survey results for the period 1996/97 to 2000/01 have recently been reported (Fagg et al. 2008). For Alpine Ash, 3,075 ha were harvested (primarily by clearfelling) across seven FMAs, with 81% successfully regenerated at the first attempt (compared with 91% for both Mountain Ash and LEMS forests). The report notes that limited use of the seed-tree system resulted in satisfactory regeneration. Figure 4.3 Unthinned Alpine Ash, Matlock (Photo: P. Fagg, DSE). 35

45 Alpine Ash SFM Case Study One of the first case studies to evaluate the application of Montreal Process indicators under Australian conditions at a forest management unit level was undertaken in the Alpine Ash forests of the Bago/Maragle State Forests (Turner 1996). These forests have a long history of multiple-use management including significant timber production. In a number of reports, evaluation considered forest productivity and yield, insect attack (specifically Phasmatids), soil properties and water quality assessments. The analysis found that a lack of relevant data (particularly historical or baseline data which the Montreal Process frequently refers to) was the main limitation in applying many of the indicators at the FMU level (Turner, 1996). Hatich et al. (1996) undertook a study in the same area, that focused on the development of a set of indicators that address Criterion 2 (Maintenance of productive capacity of forest ecosystems) of the Montreal Process. They found that Indicators (see below) could be largely satisfied by the available data at the FMU level: 2(a) - Area of forest land and net area of forest land available for timber production, 2(b) - Total growing stock of both merchantable and non-merchantable tree species on forest land available for timber production, and 2(d) - Annual removal of wood products compared with the volume determined to be sustainable. Victoria should be in a similar position in the case of Alpine Ash, because unlike the uneven-aged forests of southern NSW, Victorian Alpine Ash is managed in Victoria on an even-aged basis which simplifies requirements to meet Indicator 2(d) in particular. Figure 4.4 Uneven aged Alpine Ash, Bago State Forest. 36

46 Alpine Ash Regeneration Following Wildfire As previously noted, fire-killed Mountain Ash and Alpine Ash rely largely on seed shed from the crowns immediately after fire for self regeneration. However, it is known that seed crops tend to be smaller and more variable in younger ash-type forests than in mature ones (see Bassett 2009a). Landscape-scale assessments of flowering and seed-crops are therefore essential to provide reliable predictions on the capacity of fire-killed forests to self regenerate. Approximately 40,000 ha of Alpine Ash forests were killed or severely damaged in the 2003 Alpine fires and the 2006/07 Great Divide fires. As part of the Victorian Government s formal fire recovery response, a co-operative project involving DSE and VicForests was initiated to regenerate State forests (including Alpine Ash forests) impacted by the Great Divide fires. Results for the 2007 calendar year have been reported (DSE 2008b). This report notes that aerial flowering assessments of Alpine Ash in the four years prior to the Great Divide fires (also see Bassett et al. 2010) predicted that seed crops would be generally light across the fire area, though there was considerable variation between FMAs. Overall, it was predicted that seed crops would be too low for complete natural or self-regeneration of burnt stands (Bassett 2005, Bassett et al. 2010). Following some preliminary seed trap and soil studies, DSE and VicForests embarked on the largest seed crop assessment program ever undertaken in Victoria involving 665 post-fire plots and nearly 4,000 trees that were assessed for seed crops. Outcomes of this 2007 strategic seed crop assessment indicated a poor distribution of seed at the time of the Great Divide fires, with many sites recording no seed. These assessments confirmed the forecasts that Alpine Ash had limited capacity to self regenerate following the 2006/07 fires. About 2.5 tonne of seed remained available for sowing in 2007 from collections made after the 2003 Alpine fires. Allocation of this seed was directed to the highest priority sites (viz. young firekilled stands and stands that were carrying very light seed crops). The aerial seeding program was undertaken in July 2007 across approx. 1,700 ha and germination and survival plots were then established at four sites using standard procedures detailed in Silvicultural Guideline No 10 (Appendix No. 1). Germination and seedlings were 2.3% and 1.6% respectively (DSE 2008b) up to February 2008; the values are relatively low and may not result in acceptable stocking standards. Germinant and seedling survival on sites with low fire intensity and well developed grass and scrub competition are of particular concern and significantly increase the risk of failure. Whilst the four sites received heavy snowfalls in August 2007 (those probably adequate for stratification), the following two months were dry and this may have delayed or reduced the level of spring germination (DSE 2008b). Lutze et al. (2005) found that as would be expected under normal seasonal conditions, Alpine Ash germination following the 2003 Alpine fires was almost completed before the onset of the next summer. Possible reasons were explored by Fagg (see DSE 2008b) for the poor germination and germinant survival measured at the four sites sown in early winter 2007, and also observed across broad operational areas sown at the same time,. Sowing rates (which were deliberately reduced to maximise the area that could be treated) were discounted as a major contributing factor, particularly given the highly receptive seedbeds. Early frost heave and insufficient initial stratification were rejected as reasons for the failures. In the latter case, as already noted, heavy snowfalls occurred soon after sowing in many locations, and this should have broken primary dormancy. On the other hand, as noted by Grose (1960b), seed harvesting insects can have a major impact on the number of seeds that are available for germination and combined with low-levels of fire-induced seedfall, may have been contributing factors to lowering the expected regeneration success. Other possible reasons put forward by Fagg (see DSE 2008b) were poor seed viability, low soil moisture levels soon after snow thaw, and induced secondary dormancy which as noted in Section can occur if 37

47 stratified seed is kept at temperatures or moisture levels unfavourable for regeneration. The latter two reasons, however, could not be further investigated due to the lack of data for soil temperature and moisture. Whilst not discounting the above-mentioned factors (O. Bassett, pers. comm. May 2010, Forest Solutions Pty Ltd, Benalla, Victoria) considers that poor seed vitality (or germinative energy) of sown seed collected in 2003 was the most likely major contributor to poor field performance. The seed appeared to have lost germinative energy during storage and it was further believed that this was compounded by seed from natural seedfall being slightly immature which could also result in poor vitality of seedfall seed (see also Section ). Strategic seed-crop assessments were also made following the Black Saturday fires of February 2009 (Bassett 2009a). Consistent with seed-crop predictions based on the flowering events of 2007 and 2008 (Bassett et al. 2010), the assessments indicated that seed supply in 2009 should be widespread and significant in quantity. Bassett (2009a) concluded therefore that (with the exception of the Bunyip fire area) the Kinglake and Dargo fire areas should self-regenerate without any supplementary artificial sowing. This work underlines the critical importance of flowering and seed crop assessments that provide fundamental knowledge to guide regeneration practices including the cost-effective use of stored seed. The assessments also make a significant contribution towards meeting sustainability indicators relevant to the regeneration of disturbed forests. An interesting finding by Bassett (2009a) was that sufficient germination of Alpine Ash occurred at some sites in the Marysville district in autumn 2009 to adequately regenerate the forests, assuming that germinants survived the winter period. Whilst it is known that up to 10% of Alpine Ash seed may not require stratification to remove any dormancy (Grose 1960b), the extent of the autumn regeneration event at specific sites was unusually high. Bassett (2009a) put forward two possible explanations for this unusual event; viz. a provenance effect and/or a seed-crop which was so large that the sheer number of non-dormant seeds was adequate to ensure sufficient germinants. The latter explanation appears most plausible based on the work of Grose (1963, 1965) indicating that most Alpine Ash seeds are dormant irrespective of provenance Future Seed Management Whilst the pioneering works of Grose, Fagg and Bassett (see Bibliography for individual citations) have provided a strong scientific basis for the regeneration silviculture of Alpine Ash in Victoria. Recent findings indicate that the knowledge base on some aspects of seed management and seed biology is incomplete. The 2003 Alpine fires and the 2006/07 Great Divide fires provide contrasting scenarios in terms of seed management issues. Following the 2003 fires, 5 tonne of Alpine Ash seed were collected, of which half was used to sow fire-killed stands in This seed had been collected mainly from the Mansfield District in autumn 2003 where heavy flowering had occurred in January 2002 (it had been aerially mapped which demonstrated the value of aerial flowering assessments). As already noted, the remaining half of the 2003 collection was used to treat areas devastated by the Great Divide fires. By then, the seed was four-years-old, and seed tests showed that percentage germination had declined sharply, indicating that immature crops (one season old) do not store very well and consistent with this reduced germinative capacity, the field performance of the seed was exceedingly poor. Furthermore, Bassett s findings (reported in DSE 2009b) reinforce those of DSE (2008b), viz. the 2006/07 seed crop was mostly either absent or immature where light crops were present, so that stockings were well below acceptable standards in stands that relied solely on fireinduced seedfall. Indeed, Bassett (pers. comm., May 2010) expected that in the absence of silvicultural intervention, significant areas of Alpine Ash forest may not recover from the 2006/07 fires or will recover with a markedly different forest structure. 38

48 One of the most significant studies on Alpine Ash seed biology since that undertaken by Grose (1957; 1960a, b; 1963; 1965) has recently been reported (Bassett et al. 2010). The report deserves special consideration in the current review because some new understandings of seed production have been identified. It has a focus not only on forecasting flowering and seed-crops for up to three years ahead, but also on identifying factors that may influence flowering periodicity and seed-crop production based on extensive long-term data for ash-type species. Whilst Alpine Ash in particular has long been regarded as a reasonably dependable seed producer (Grose 1960a, Campbell et al. 1984), Bassett et al. (2010) found that the frequency of flowering at the landscape-level (that is, across the geographical range of both Alpine Ash and Mountain Ash forests) has declined to as little as 1 or 2 per decade over the past 15 years compared with 3 to 4 times a decade in the previous eight and fifteen-year periods for Alpine Ash and Mountain Ash respectively. Based on data from the annual Aerial Flowering Assessment program for Alpine Ash and Mountain Ash between 2005 and 2009, the flowering intensity for both species was found to be high to very high in 2007 and 2008 with good to excellent distribution at the landscape-level (the first such significant flowering since 2003 for Alpine Ash and low elevation Mountain Ash, and 1995 for high elevation ash), while limited floral data from four traps at a research site near Matlock showed flowering intensity of Alpine Ash to be high in 2008 (the first flowering event at this site for ten years). Bassett (pers. comm., May 2010) noted that the extensive stand of Alpine Ash at Matlock is somewhat unique in that it appears to have different site characteristics to most other stands across the eastern highlands, further influencing flowering outcomes. Two consecutive seasons of flowering in 2007 and 2008 at the landscape-level has been recorded only one other time in the past 50 years (Bassett et al. 2010). Using Keetch-Byram Drought 1 Index (KBDI) values as an indicator of drought, Bassett et al. (2010) suggested that poor flowering may be linked to prolonged drought, whilst the burst of flowering in 2007and 2008 could be a response to little or no flowering in the preceding five years. The technique used by Bassett et al. (2010) to forecast flowering and seed-crops involves drought analysis, floral component sampling for bract fall and operculum counts, aerial flowering assessment, and seed-crop assessment in the year of seed production. Importantly, the forecasting recognizes that seed-crops arising from one flowering season remain available for collection or natural regeneration in the event of a wildfire for two seasons that start one year after flowering. Their analysis predicted that there will be no seed available after winter 2010 (before then, an opportunity exists to collect seed resulting from the consecutive flowering events in 2007 and 2008), and no seed is expected in 2011 and Accordingly, Bassett et al. (2010) recommend that during 2010 to 2012, the annual Aerial Flowering Assessment program focus on identifying any outliers of isolated flowering (as detected at Mansfield in 2002) so that these stands can be targeted for future seed collection. Annual Aerial Flowering Assessments are clearly a critical component of seed management strategies for Alpine Ash and Mountain Ash forests in Victoria. The impacts of wildfire and drought on seed supply are emerging as major operational and research challenges in the regeneration and perpetuation of ash-type forests. Ferguson (2009) has recently drawn attention to the risks posed by major wildfires to the perpetuation of both Mountain Ash and Alpine Ash forests in Victoria. He notes that these two forest types are not only fire-sensitive but are also now considered to be irregular seeders. Furthermore, significant seed crop production usually does not commence until stands reach around 20 years of age. 1 KBDI is essentially a numerical expression of soil moisture deficiency or drought. KBDI values can also be used as a guide to expected fire behavior 39

49 Based on simulation studies, Ferguson (2009) concluded that if Alpine Ash seed collection and storage is restricted to relatively low levels for the next four decades (irrespective of climate change considerations), there is a high probability that there will be insufficient seed to ensure artificial regeneration through post-fire intervention of those stands that are killed by wildfire and do not regenerate naturally. The clear long-term implication of this simulation study is the potential demise of ash-type forests currently available for commercial wood production. Furthermore, a store of seed is also required by VicForests for routine post-harvest regeneration and by DSE for backlog regeneration. The conclusions of Ferguson (2009) are consistent with those of Shugart (1984) who summarised work on the simulated pattern of successional dynamics for Alpine Ash in the Australian Capital Territory using a gap model approach. In the absence of fire, Alpine ash dominates the sites while periodic fire with a 50-year return frequency, results in a systematic reduction of the species over a 600-year simulation period, with Snow Gum (E. pauciflora) and Mountain Gum progressively becoming more dominant. Continuous monitoring and analysis of flowering and seed-crops over the past 15 years indicates an interaction between climate (as indicated by drought index) and flowering incidence in Mountain Ash and Alpine Ash (O. Bassett, pers. comm., May 2009). Although further research is required, this monitoring work indicates that flowering of high elevation Mountain Ash is susceptible to periodic drought whereas the response to flowering of low elevation Alpine Ash and Mountain Ash is relatively tolerant of intermittent drought. However, all stands irrespective of elevation or species appear to be susceptible to inter-seasonal cumulative drought like that experienced over the past decade. Native Forest Silviculture Guideline No 17 (Appendix 1) outlines the action sequence that should be followed after wildfire to ensure that fire-killed stands are regenerated. The rapid response techniques that underpin the actions have been progressively developed over the past decade (see for example Bassett 2005, 2009b) Thinning A large trial was established in northeast Victoria in 1964 to examine the response of Alpine Ash to uniform thinning of a regrowth stand resulting from the 1928 wildfires that burnt large tracts of mountain forests (DSE 2006b). Four thinning intensities {37%, 46%, 60% and 100% of basal area (BA) retained} were tested in a replicated experiment. Results from the most recent measurement in 1998 showed that the light and moderate thinning treatments were associated with the highest mean annual increments. However, diameter class data indicated that the 46% BA-retained treatment produced more trees with a diameter at breast height of 80 cm or more. The current prescription (50% BA retained) is more or less in line with this finding. These stands were more responsive than the results reported by Horne and Robinson (1990). 40

50 Mountain Ash Forests Introduction Mountain Ash dominated forests occupy a prime position in eucalypt forest development in southern Australia. It forms a largely mono-specific tall open forest canopy, with a dense and complex understorey of mesomorphic trees, shrubs and ferns (Ashton 1981). In Victoria, Mountain Ash forests occupy the high quality mountain land in the central and eastern Highlands, Otway Ranges and South Gippsland. They occur over an altitudinal range from about 120 to 1,100 metres, and a mean rainfall range of about 750 to 1700 millimetres with a winter maximum, but without any severe dry period. Their best development occurs on sheltered aspects with annual rainfall greater than 1,100 millimetres on deep, friable clay loams. The deep well-structured soils of these forests provide the foundation for outstanding growth, and are usually characterised by thick layers of litter and humus that strongly influence soil fertility and water availability. Mountain Ash forests typically form pure even-aged stands that have regenerated following intense wildfire, although some multi-aged stands occur naturally, following less severe fires. Many regrowth stands have arisen from the catastrophic wildfires of 1919, 1926, 1932, 1939 and They contain occasional trees or stands of Shining Gum and at the limits of their range, they may be found in association with Alpine Ash, Messmate, Mountain Grey Gum and Manna Gum. The understorey vegetation may be present in one or more of three strata, including trees (for example, wattles, sassafras and myrtle beech), shrubs (for example, hazel, musk, blanket leaf and tree ferns) and ground flora (for example, ferns, wire grass and mosses). A more detailed listing of the flora of Mountain Ash forests is contained in the reports by Ough and Ross (1992) and the LCC (1977). Figure 4.5 Remnant stags of Mountain Ash from the 1939 fire (photo DSE Victoria). 41

51 Commercially, harvesting of virgin stands of Mountain Ash began late last century and continued until the catastrophic fires of 1939 when the majority of remaining stands available for wood production, were killed. The harvesting and regeneration of virgin stands was initially on a sawmiller selection basis (Squire et al. 1991a, b), but evolved into an even-aged seed-tree system as loggers found that the great height and bulk of the trees made selection logging difficult and dangerous. Regeneration was assisted using the soil disturbance provided by logging as a seedbed. The development of highlead cable systems created the capacity and the need for logging of the Mountain Ash forests to be more intensive from the 1920s, whilst the use of steam-heated kilns enabled development of larger markets for dried Mountain Ash timber. By 1931, it was estimated that 80% of the flooring in Melbourne was kiln dried Mountain Ash (Moulds 1991). The establishment of the Maryvale pulp and paper mill in 1936 was based on the high quality fibre products that could be produced from Mountain Ash and the non-sawlog waste that was generated during sawlog harvesting operations. Large-scale salvage operations, assisted by increasing mechanisation, were carried out following the 1939 wildfires in the period 1939 to 1951 (Moulds 1991), helped in supplying timber for the post-war housing boom. Following completion of salvage logging, harvesting activities within the Mountain Ash forests were reduced to a relatively minor scale, and utilisation switched largely to the Alpine Ash forests. However, where harvesting continued to be conducted, clearfelling silviculture was used (Cunningham 1960). Harvesting moved in the late 1970s into regrowth stands of pre-1939 origin and into the 1939 regrowth in the early 1980s as part of salvage-logging following the 1983 wildfires. Thinning operations in young 1939 regrowth occurred during the period 1962 to 1974 and also in logging regrowth and plantations in the period 1988 to In 2002, thinning operations commenced in logging regrowth and in the 1983 fire regrowth. The high productivity of these forests has supported the development of more sophisticated harvesting systems. High-lead logging, supported by steam winches, was used up to the 1939 wildfires, but was largely replaced by ground-based systems using bulldozers to extract logs. This also coincided with the development and broader introduction of the modern chainsaw. In the late 1970s, rubber-tyred skidders were replacing bulldozers, while on log landings from the early 1980s onwards, rubber-tyred loaders and bulldozers were being replaced by excavator-based crab-grab machines. By the mid-1990s, excavator-based felling machines were being used to harvest 1939 regrowth and to reduce the extent of hand-felling. These machine developments have enabled greater flexibility in harvesting operations and the development of new and safer approaches to harvesting to provide improved environmental outcomes. For example, the adoption of cording and matting of extraction tracks and the use of shovel logging during wetter periods to reduce soil compaction and disturbance on coupes (Wilkinson 2000, Campbell 2003) has only become possible with comparatively recent machine developments. Environmental issues relevant to the utilisation of Mountain Ash forests are governed by the Code of Practice for Timber Production (DSE 2007a) and associated Management Procedures (DSE 2005b). 42

52 4.5.2 Silvics of Mountain Ash Forests An overview of the ecology of these forests was provided by Ashton and Attiwill (1999), including soils, regeneration and growth, and succession. In the absence of ecosystem disturbance, regeneration of Mountain Ash is rare (Ashton and Willis 1982). The main barriers to seedling establishment in the undisturbed forests are the lack of receptive seedbeds (Cunningham 1960, King et al. 1993), microbial antagonism (Ashton and Willis 1982), insect predation (Ashton 1979) and the low light intensities observed beneath the understorey (Ashton and Turner 1979). Natural regeneration normally results from the destruction of seed-bearing trees following severe wildfire. High intensity crown fires kill the overstorey and understorey competition, prepare a receptive seedbed and induce seedfall from the crowns of standing trees. Germination is rapid with the onset of cool moist conditions, and results in a dense stratum of even-aged regeneration (Ashton 1976). Under lower intensity surface fires, the understorey may be killed but the overstorey may survive or sustain only partial death. In such circumstances, a regenerating stratum may establish under a partial canopy, leading to the development of a dual or multi-aged forest structure (Gilbert 1959, Ashton 1976). Extensive research by many scientists has provided a wealth of information on the silvics of Mountain Ash. As the species does not produce lignotubers or coppice, regeneration therefore must occur from seed. The seed is small and thin coated (Grose and Zimmer 1958), susceptible to fungi and insect predation (Cunningham 1960, Cremer 1966, Ashton 1979, Neumann and Kassaby 1986), and does not persist in the soil for longer than one year (Cunningham 1960). Seed storage is therefore in the capsules in the canopy. Seed is shed naturally from the capsules as they die and dry out, a process which may take one to three years (Cremer 1960). By contrast, fire may result in rapid capsule death and shedding of seed may commence within hours (Cremer 1965). Mountain Ash is generally considered a light but variable seed producer, with heavier bud initiation occurring every two to four years (Ashton 1975) and its reproductive capacity increases with age. Saplings rarely flower before six years of age and are unlikely to carry sufficient quantities of viable seed for regeneration until 10 to 20 years of age. Ashton (1975) found fruit set in mature forest (200years-old) was 1.6 times as great as spar-stage forest (50 years) and 3.5 times as great as pole-stage forest (25 years). Typical seed crops in 1939 regrowth stands (the predominant age class currently managed for wood production) were 0.5 to 5 million seeds per hectare (Campbell et al. 1990). The pattern of seed supply is often irregular, with monitoring of seed crops near Tanjil Bren indicating a moderate-heavy crop in 6 out of 16 years (Flint and Fagg 2007). Relatively poor flowering of Mountain Ash appears to occur in about 2 years following dry years, as flowering occurs about 2.5 years after bud initiation and appears to be strongly linked to moisture availability. Poor flowering over successive seasons can cause severe shortages of the collectable seed-crops required for the artificial sowing of coupes. Mountain Ash seeds may exhibit a small degree of primary dormancy (up to 15% of viable seeds), but o exposure of imbibed seeds to a germination temperature above 24 C can induce secondary dormancy (Flint and Fagg 2007). Field germination is not likely to be significantly influenced by the above processes. Cold moist stratification removes both primary and induced dormancy of Mountain Ash. 43

53 4.5.3 Stocking Following Harvesting of Mountain Ash in Victoria Recently, Fagg et al. (2008) reported on 5,402 hectares of Mountain Ash forest type that was treated and surveyed on public land between 1996/97 and 2000/01, an average of nearly 1,100 hectares per annum. Over these five years, silvicultural systems producing even-aged regrowth were employed on 100% of this area, with the majority of this concentrated in Central Gippsland (54%), Central (24%), Dandenong (16%) and Otway (4%) FMAs. Analysis of the silvicultural systems used in Mountain Ash forests for the period 1996/97 to 2000/01 at the first attempt at regeneration (Tables 14 and 15 in Fagg et al. 2008), showed the following percentage of treated coupe area that was satisfactorily stocked: Clearfelling - 90% (4,925 hectares treated and surveyed). Seed-tree - 97% (197 hectares). Overall, this equated to 91% of 5,122 hectares meeting the minimum stocking standard and is generally better than Alpine Ash (81%) and HEMS (70%), and on a par with LEMS (91%) (Table 2 & Figure 2 in Fagg et al. 2008). All sources of regeneration are counted in stocking surveys, provided they meet acceptable criteria (Dignan and Fagg 1997). In Mountain Ash forests, the vast majority of regeneration is from seed. The consistently high stocking results for Mountain Ash indicates that the silviculture applied is generally very appropriate. The higher stocking of the area treated with seed-tree silviculture may be indicative, but the area treated is insufficient to draw any firmer conclusions. Site factors are also likely to be important to the generally high stocking outcome, including: low levels of browsing, good seed supply (natural or artifical), and elevations that are less subjected to frosts Mountain Ash Silviculture For Mountain Ash dominated forests, Campbell et al. (1984) and more recently Lutze et al. (1999a), provided instructive accounts of past and present approaches to their silvicultural management in Victoria. Key research which has been critical to the silvicultural management of Mountain Ash has been a study in 1939 regrowth at Tanjil Bren in the Central Highlands. This study was initiated to determine whether a better balance between economic and environmental benefits could be achieved through alternative silvicultural systems to clearfelling (Squire 1990). Results up to 1994 are reported in Campbell (1997a). He concluded that clearfelling and improved seed-tree systems remain the best systems to employ in final harvesting where timber production is the main objective. Shelterwood and group selection systems may be an option where flora and fauna values need special consideration, though longer-term data are needed before more definitive conclusions can be drawn on these alternative systems. As noted in Section 4.3, Lindenmayer (2007) has reported initial results from a study, established in 2003 in the Victorian Central Highlands, to investigate a system known as the Variable Retention Harvesting System (VRHS). VRHS involves the retention for at least one rotation of strategic elements of the forest from a biological diversity perspective. Maintaining stand structural complexity, including multi-aged stands, large dead and living trees, large logs on the forest floor, and thickets of understorey vegetation, is an over-riding objective of VRHS. The System is under-pinned by the concept of adaptive management (see Raison et al. 2001a) and has the potential to achieve the structural complexity objective and simultaneously cater for (reduced) timber harvesting at the landscape-level. The System is not suited to every logging coupe for a variety of practical considerations including OH & S issues. Importantly, VRHS must provide for reasonable spatial distribution of retained structural components of the ecosystem so that vegetation retention 44

54 associated with Code of Practice requirements (for example, riparian vegetation to protect water values) does not satisfy the VRHS requirements. Lindenmayer (2007) provided details of the treatments being evaluated, viz. natural control, existing clearfelling, and two retained vegetation treatments configured as one 1.5 ha island and three 0.5 ha islands dispersed across the coupes. As at mid-2007, 16 coupes and six control sites had been established and surveyed for vertebrates. Although the purpose of the retained islands is not to support viable populations of any animal groups, early results indicate that terrestrial small mammals can persist (albeit with reduced densities) within the retained vegetation. Lindenmayer (2007) stressed the need for long-term monitoring to provide forest managers with feedback as part of adaptive management. The selection and implementation of silviculture in Victoria s Mountain Ash forests is covered in detail in the Mountain Ash in Victoria s State Forests; Silviculture Reference Manual No.1 (Flint and Fagg 2007) Mountain Ash Regeneration The establishment and development of Mountain Ash regeneration following harvesting has been studied extensively. A detailed study of regeneration processes has widened the silvicultural knowledge provided by the earlier studies of Cunningham (1960), Ashton (1976) and Campbell and Bray (1987). This study, as part of the SSP project initiated in the late-1980s at Tanjil Bren, evaluated the influence of a range of gap sizes, overwood retention levels with fire and mechanical site preparation, on the establishment and development of Mountain Ash regeneration. The effect of a range of small gaps (nil to 30 metres by 30 metres) and large gaps (50 metres by 50 metres to 27 hectares clearfell) on Mountain Ash establishment at Tanjil Bren (Van Der Meer et al. 1999) were studied. Logging and regeneration treatments were undertaken over three consecutive years. Total seedling density was found to be significantly affected by both year of regeneration treatment and gap size (higher in the group of small gaps). The annual variation was attributed to annual climatic differences. Seedling height of the more dominant seedlings was positively related to gap size in the group of large gaps. For these gaps, seedling growth was better on burnt than on mechanically disturbed seedbeds. Seedling survival tended to increase with increasing gap size. Van Der Meer et al. (1999) concluded that, depending on management objectives, a gap-cutting system may be a viable silvicultural treatment in Mountain Ash forests, though regeneration success will be markedly affected by annual climatic variation and by individual site factors. Growth will also be compromised to an unknown extent. There are trade-offs, therefore, in regeneration establishment and early growth with both gap size and level of retained overwood. Frost heave, which physically lifts seedlings from the soil, exposing their roots, may result in an almost 100% loss of seedlings during the germinant stage (Campbell and Bray 1987). Dignan (2002) examined the effect of season of sowing of Mountain Ash on high elevation sites (above 900 metres) with different seedbeds. He found that the main period of germination for autumn sowings was midaugust to late-september, and a little later for winter sowings. This spring-germination resulted in satisfactory stockings of seedlings at the standard sowing rate of 200,000 viable seeds per hectare and largely avoided frost damage to germinants. 45

55 Bassett (1996a) developed a Seed Crop Assessment Kit for Mountain Ash, based on detailed studies of capsule densities as related to tree size. It enables foresters to evaluate the adequacy of standing seed crops for either seed-tree systems or seed collection. Continued supply of seed needed for regeneration of clear-felled coupes is a major operational issue, but tools such as aerial mapping of flowering (for example, Bassett and Roberts 1995) and ground-based trapping of bracts (Roberts 2000) enable efficient targeting of areas where seed may be collected. Good seed storage facilities can buffer the natural variation in seed crops, allowing storage for 10 years in appropriate conditions (Boland et al. 1980), however, adequate seed inventories rely on a skilled seed collection workforce and sufficient resourcing Mountain Ash Growth and Development During the early stages of forest development, stands of Mountain Ash can reach rates of aboveground net primary production of 36 tonne/ha/yr (Attiwill 1992), and over a nominal 80-year rotation 3 are expected to yield mean merchantable increments of 7.5 m /ha/yr (Parkes 2001) with mean 3 sawlog increments of around 4 m /ha/yr (DNRE 2002b). This high biomass productivity combined with high wood utility is the basis for the commercial importance of this forest type. Figure 4.6 Regrowth mountain ash (photo from DSE). 46

56 As with most fast growing eucalypt species, Mountain Ash is considered to be intolerant of competition, as evidenced by very high initial seedling densities, thinning out exponentially to stand densities of about 2,000 trees per hectare at 20 years of age (pole stage), 400 at 40 years of age and 100 at 80 years of age. Half the final stand height of 55 to 75 metres is achieved by about 20 years of age, with rapid segregation into dominance classes and subsequent mortality of those in the intermediate and suppressed classes. The critical level of retained overwood density below which the regenerating stratum cannot survive, is unknown. Ashton and W illis (1982) observed successful eucalypt establishment on disturbed soil beneath a complete mature canopy with the understorey removed. They found seedlings continued to develop rapidly to the intermediate stage, but by 5 to 6 years of age, they began to stagnate and by 10 years, they died and were replaced by understorey species. A number of studies have shown that Mountain Ash regrowth responds to thinning with increased diameter and basal area (and thus volume) growth of the retained dominant and co-dominant trees. Most thinning of Mountain Ash has involved thinning from below (removal of the smallest and suppressed stems), increasing the availability of water, light and nutrients to the retained trees (Webb 1966). Removing suppressed trees during thinning in effect removes trees which would have died 3 through natural competition. Yields of up to 280 m /ha of pulpwood have recently been harvested during thinning of 26-year-old Mountain Ash logging regrowth in the Toolangi State forest. Using the growth model STANDSIM (Opie 1972, West 1991), responses to thinning regimes in Mountain Ash forest without undue loss of stand basal area increment could be expected at thinning intensities which remove less than 50% of the initial basal area. This growth model also indicates that using an intensive thinning regime, more sawlog volume is produced in 50 years than the unthinned regime in 80 years (Kerruish and Rawlins 1991). Figure 4.7 Thinning Mountain Ash, Otways 1991 (photo from P. Fagg, DSE). 47

57 High Elevation Mixed Species Forests Introduction The high elevation mixed (eucalypt) species forests of Victoria (HEMS) occupy extensive areas of high quality forest land in the eastern and southern parts of the State. The most important commercial forests are those of east Gippsland. In these forests the Messmate-Peppermint-Gum group of the Low Elevation Mixed Species (LEMS) forest often extends to high elevations (above approximately 700 metres) where low temperature is limiting to autumn/winter germination and survival of eucalypt seedlings. These forests are referred to as High Elevation Mixed Species (HEMS) forest, which occur extensively throughout the eastern Highlands (Figure 4.8). They are typically damp to wet forests varying from 35 to 50 metres in height, but height may exceed 75 metres. On some sites, other species may replace the low elevation species. Messmate, Mountain Grey Gum, Manna Gum, Mountain Gum, Broad and Narrow-leaved Peppermint are common throughout the range of HEMS forest. Cut-tail or Brown Barrel and Errinundra Shining Gum (E. denticulata) are common in east Gippsland, and Blue Gum (E. bicostata) is common in the northeast. There are two common stand structures: virgin stands, which are mainly uneven-aged, although large mature and over-mature trees tend to dominate, and pure regrowth stands resulting from post-1960 logging. The structure of virgin stands indicates that fire is a moderately common occurrence, and while young trees are fire sensitive, as they mature the rough fibrous bark provides protection from ground fires of moderate intensity. The coppicing ability of Cut-tail is relatively weak, but Messmate, Peppermint and Blue Gum are capable of producing coppice. The diversity of these forests is illustrated by a comparison of these classification systems below. HARIS 1 Dominant overstorey species 1 Silvicultural 2 Dominant form 3 Mountain Mixed Species (MMS) Predominantly Cut-tail, Messmate, Mountain Grey Gum either in pure stands or in mixtures. Other species include Manna Gum, Blue Gum, and Candlebark where mature stand height is generally greater than 40 m High elevation mixed species (HEMS) Alpine Mixed Species (AMS) Generally peppermint/ gum stands at high elevations (e.g. Mountain Gum) HEMS Open forest Shining Gum (SHG) Shining Gum HEMS Tall open forest 4 Tall open forest Hardwood Resource Information System ('HARIS'; Grierson et al. 1992) Cut-tail is E. fastigata, Messmate is E. obliqua, Mountain Grey Gum is E. cypellocarpa, Manna Gum is E. viminalis, Blue Gum is predominantly E. bicostata, Candlebark is E. rubida, Mountain Gum is E. dalrympleana and Shining Gum is E. nitens and E. denticulata 2. Victorian silvicultural forest types (Bennett and Adams 2004a) 3. Vegetation Growth Forms (NFI 2003) 4. Tall open forest: projected foliage cover of tallest stratum 30-70%, trees > 30 m (formally Wet Sclerophyll Forest) 5. Open forest: projected foliage cover of tallest stratum 30-70%, trees m (formally, Dry Sclerophyll Forest) 48

58 Figure 4.8 Location of HEMS forests (Map provided by P. Fagg, DSE). While selective harvesting was used extensively up until the early 1980s in northeast Victoria, its use in east Gippsland was limited due to accessibility. From the 1960s, when accessibility in east Gippsland improved, HEMS forest has been harvested using the system developed for Ash species (that is, clearfelling followed by slash-burning and aerial sowing). This seemed logical as the forest environment had many similarities to the Ash forests, and this type of silviculture had successfully been transferred to other forest types. However, the system proved to be unreliable and planting of mainly Shining Gum and Alpine Ash was also employed. More recently, seed-trees have been used with supplementary aerial sowing. Browsing by wallabies is sometimes a problem, and while controlled by poisoning prior to the 1980s, this is no longer an option. Issues relating to the assurance of regeneration and growth in the HEMS forest type prompted research into this area, with systematic work commencing in the early 1970s. Studies in the Wombat and Mt Cole forests focused on growth following release from competition, impact of fire, and the pathogen Armillaria. In east Gippsland, major work related to seedbed types and sowing times following harvesting (Fagg, 1981), with later studies focusing on the role of seed-trees as the seed supply for regeneration instead of artificially applied seed (for example, aerial sowing). More recently, studies into seed crop development and maturation have been completed for a number of key HEMS species (for example, Murray et al, 2004). While this silvicultural research has led to a higher level of understanding of the silvicultural requirements for HEMS forests, the dissemination of the information to field staff and incorporation of the knowledge into everyday practice are equally important. This was assisted through the 1990s and early 2000s by the activities of the HEMS Research and Development Action Group, which included preparation of guidelines and decision support systems, for example Ryan (1997) and Sebire (2000). Controversy regarding clearfell logging of these forests has also driven research into alternative harvesting systems, including the Value Adding Utilisation System (VAUS) studies from the early 1990s and the Variable Retention Trial that was recently conducted in the Bendoc area of east Gippsland. 49

59 4.6.2 Silvics of HEMS Forests Ecologically, these forests have a strong focus on population lifecycle adaptations, with fire being a fundamental factor influencing their regeneration, growth and evolution. Despite the majority of the eucalypt species present in these forests having the ability to re-grow from lignotubers and epicormics, it is more usual for regeneration to occur as a result of seedling growth (Sebire and Fagg 2009). The species are partially fire sensitive, with mortality of younger trees reasonably common and even of mature trees on occasions, particularly following high intensity fire. Where fire intensities result in partial crown scorch (about 50 percent), studies have shown there is a temporary reduction in basal area increment (Kellas et.al 1984). Normal increment is resumed within 3-4 years, coinciding with full restoration of foliage. HEMS species do not commence to carry seed crops before 15 years of age and do not carry significant amounts of seed until they are at least 20 years of age. There can be a large variation in the production and loss of components during the seed development cycle, for example for Errinundra Shining Gum and Cut-tail, (Murray et. al. 2004). Observation of Messmate and Mountain Grey Gum during seed-crop development studies by Murray and Lutze (2004) indicated that moderate to large seed-crops could be expected in one year in four. This periodic cycle results in operational issues, particularly in relation to the economics of collecting seed for artificial seeding and the suitability of seed-tree silviculture. However, at specific localities even during a poor seeding year it may be possible to locate trees and patches with larger than average seed-crops due to inter-tree variations. Jurskis and Grigg (1996) monitored Cut-tail seed production over 15 years. Flowering occurred in 70% of years and seed storage in tree canopies dampened the irregularities in flowering, so moderate quantities were usually shed in autumn and winter. They also reported that the resulting inadequate seed supply might result in sub-optimal regeneration in some years. Generally, the pattern of development of a seed-crop is the same across all HEMS species with the length and seasonality of the cycle being the major differences. Reduced rainfall conditions can impact on the flowering phase and subsequently on seed production. Studies have shown there is a strong relationship between tree diameter and the amount of seed carried in the crown. This relationship has been used to produce Seed Crop Assessment Kits for two HEMS species; Messmate and Mountain Grey Gum (Sebire 2001b and DSE 2003a). Tree size and health and capsule density on branchlets (approximately 2 centimetres diameter) are combined to provide an indication of the amount of seed in the tree crown and hence its suitability for use as a seed-tree or for seed collection. In mature HEMS forest, large amounts of seed are produced and shed naturally, even in the absence of fire or other disturbance. Murray et al. (2004) reported that annually a stand of Errinundra Shining Gum shed 2.5 million viable seeds per hectare and Cut-tail shed 1.2 million viable seeds per hectare. In both cases no seedlings were established from the seed. Natural seedfall (in the absence of fire) generally commences 9-12 months after the capsules mature in the crown. This timing appears to vary between species: seedfall from Errinundra Shining Gum and Mountain Grey Gum is spread throughout the year whereas Messmate seedfall tends to be in spring and autumn. The viabilities of s are variable across localities, so it is important that each is individually tested before usage (Native Forest Silviculture Guideline No. 4 Eucalypt Seed Sampling and Testing (Appendix 1)). HEMS species do not generally exhibit primary seed dormancy in the seed (Lutze et. al. 1998). Following seedfall, seed survival in soil is low with few seeds remaining viable and ungerminated on the forest floor for more than one year (Fagg 1981). Germination can be highly variable, with observed germinants as low as 0.8%-1.5% ranging up to 19.3%-40% (Fagg 1981 and Lutze et.al. 1999b). The reasons for low germination percentages are not fully understood but are likely to include fungi in the seedbed, harvesting by ants, low soil moisture (Sebire and Fagg 2009). 50

60 4.6.3 Regeneration of HEMS Forests The successful regeneration of harvested coupes can be highly variable (Fagg et al. 2008, Delbridge 1998), which is not altogether surprising given the wide distribution of HEMS forests. This distribution contributes to a broad range of site conditions and a varying history of disturbance (both utilisation and fire), which is reflected in the forest type s variable nature. Also, rainfall and temperature can be quite variable, both seasonally and annually (Sebire and Fagg 2009). Frosts are common in all HEMS forest areas with frequency dependent on site elevation. For example, at elevations of around 1,000 metres, 100 to 150 frosts per year may occur, while at elevations of 700 metres, 40 to 50 frosts may occur per year (Fagg 1981). Frost kill is common with autumn-germinated seedlings, so it is generally preferable to encourage spring germination to avoid the most severe frosts, particularly at higher elevations. Time of year for successful germination leading to regeneration provides a means of separating forest types into "high" or "low" elevation mixed species forest. The determining factor is whether the most successful germination occurs in autumn (LEMS) or spring (HEMS). The prime contributor to the success of regeneration is the environment, and relates to whether regeneration can establish sufficiently to survive extreme winters or hot summers. If severe frost and the associated frost heave, or snow damage in winter don t kill germinants, hot summer spells can easily kill germinants that are establishing in late-spring, especially in low elevation areas and on more exposed northern and western aspects (Ryan and Dore 1996). On this basis, mixed species forests generally above an elevation of around metres (but as low as 600 metres for wetter aspect forests such as Mt Disappointment) are for the purposes of regeneration, treated as HEMS forests. Fagg (1981) commenced an intensive field study on HEMS regeneration in 1974 in response to consistently poor regeneration outcomes following harvesting. A factorial design was used to investigate fencing/no fencing, burnt/unburnt seedbeds, and time of sowing at five contrasting sites. Plots were sown at heavier than normal sowing rates with a mixture of the two main species (Messmate and Cut-tail) for all experiments except for the 1976 trial where only Messmate was sown. Intensive records were maintained on germination, survival and subsequent development of germinants, browsing damage and seedling growth. The main findings were: Browsing was only a problem at one of the sites. Seedling percents (number of live seedlings as a percentage of number of viable seeds sown) were markedly greater on disturbed than on burnt seedbeds for most sites and sowing times. Satisfactory seedling percents were obtained for seed sown in February to May, and in August and September, though best results were associated with sowings on unburnt seedbeds in April, May and August, and Mean height growth within the first 3 to 4 years after sowing was generally slightly greater on burnt than on disturbed seedbeds. A great deal of effort went into determining the cause of death of 5,513 individually marked germinants and seedlings Fagg (1981) but the likely cause of death in 1,951 cases could not be determined, while in a further 1,051 cases there was no trace of the dead germinants/seedlings. For the remaining 2,511 cases, cause of death was attributed to ash-bed deaths (15% of cases), frost heave (9%), freezing (14%), high soil temperature or drought (15%), suppression by weeds (33%), browsing and trampling (12%) and insect damage (2%). It was concluded that for those coupes predicted at the coupe planning stage to be difficult to regenerate (based on past experience), a 51

61 rough heaping (preferably with a root rake)/soil disturbance regime may be the best regeneration strategy to adopt from the outset. Notwithstanding this, the silvicultural system continued to be clearfelling/seed-tree followed by slash-burning and aerial sowing in autumn. The rate of regeneration failure, not surprisingly, has continued to be higher than in any other forest type in Victoria (Fagg et. al. 2008). Delbridge (1998) undertook an analysis of HEMS regeneration records in the East Gippsland FMA for the harvesting and regeneration period 1987/88 to 1995/96. He found highly variable successes with frequent under-stocking, identified a range of issues (factors), but could not predict which combination of factors would result in regeneration failure in any one year. He concluded that it was possible to identify single contributing factors important to regeneration success in a particular year, but the same factor may not be significant at another site or in a different year. Year of regeneration was therefore highly significant in that study with many interacting factors, the significance of which will vary from site to site. The various seasonal growing conditions found in HEMS have significant influence, and they tend to mask the effects of most other factors (Delbridge 1998). The more extreme conditions are related to frost exposure during winter and drought during summer (Lutze et al. 1998). Since the early 1990s, HEMS coupes have been sown in winter to promote spring germination. Winter mortality of autumn germination is severe at high elevations, and winter sowing ensures germination will not occur until spring (Fagg 1981, Delbridge 1998, 1999, Lutze et al. 1998). However, mortality of spring germination will be high if spring and summer rainfalls are low (Lutze et al. 1998, 1999b) and this occurred in the east Gippsland region during 1993/ /95. Another important factor which reduces regeneration success in HEMS is browsing of seedlings by animals, primarily wallabies. Native Forest Silviculture Guideline No. 4 (Appendix 1) and Sebire (2001a) reported that browsing has moderately to severely affected 26% to 31% of HEMS coupes State-wide in approximately the last decade. Expensive fencing is the only current option to effectively control severe browsing, but is not always used due to the effort and expense (Native Forest Silviculture Guideline No. 7 Browsing Management (Appendix 1)). In an attempt to better manage this issue, an inter-state workshop on browsing in native forest was held in Victoria in 1999, and the Proceedings present findings of recent research, operational experience, and identify a range of different techniques that when applied together may assist to reduce browsing impacts (DNRE 2002c) HEMS Silviculture Recently, Fagg et al. (2008) reported on 5,529 hectares of HEMS forest type that was treated and surveyed on public land between 1996/97 and 2000/01, an average of just over 1,100 hectares per annum. This included areas that were classified as both even- and uneven-aged, re-treated and reforestation. Over the five years, silvicultural systems producing even-aged regrowth were employed on 97% of this area, with the majority of this concentrated in east Gippsland (44%) and Tambo (30%) FMAs, and the rest more broadly located on the Great Divide. The small area of regeneration from uneven-aged silvicultural systems (185 hectares over the five years) was concentrated in the northeast, mainly within the Benalla-Mansfield FMA. 52

62 Analysis of the silvicultural systems used in HEMS forests for the period 1996/97 to 2000/01 (Tables 14 and 15 in Fagg et al. 2008), showed, at the first attempt at regeneration, the following percentages of treated coupe area that were satisfactorily stocked: Seed-tree - 74% (3,116 hectares treated and surveyed) Clearfelling - 66% (1,406 hectares) Shelterwood - Group and single tree selection 56% (411 hectares) - 82% (185 hectares). At first attempt overall, this equated to 70% of 5,118 hectares meeting the minimum stocking standard, which is well below the average for other forest types (Figure 2, Fagg et. al. 2008). All sources of regeneration are counted in stocking surveys, provided they meet acceptable criteria (Dignan and Fagg 1997). In HEMS forests, the vast majority of regeneration is from seed. Fagg et al. (2008) suggested that seasonal conditions, such as rainfall and frost, seem to have the largest influence on successful regeneration in this forest type rather than the silvicultural system. However, seed-tree and selection systems do appear to be the more successful, albeit selection silviculture is very limited. The difficulties in regenerating the HEMS forest type combined with its importance as a source of high quality timber have led to significant ecological and operational research in an endeavor to better identify the causes of poor regeneration. This aspect is critical to any consideration of the silviculture that is appropriate for this forest type. For mixed eucalypt species, including HEMS, Campbell et al. (1984) and more recently Lutze et al. (1999a) provide instructive accounts of past and present approaches to their silvicultural management in Victoria. It appears that no simple silvicultural system can be uniformly applied in these forests. The condition and structure of each stand should determine the appropriate silvicultural decision, providing that the three key factors of seed supply, seedbed condition and establishment conditions are satisfactory to optimise regeneration success. To facilitate this, a general silvicultural strategy for HEMS has been developed, primarily resulting from more than two decades of research in east Gippsland. The extensive research and development program (DNRE 1999) in combination with extensive operation knowledge was summarised into a Regeneration Decision Tree and later further developed into a Decision Support System (DSS) (Sebire 2000, Fagg et al. 2008). Key elements are: Use of seed-trees. Autumn site preparation using slash-burning. Use of soil disturbance if harvesting finishes after the burning period to avoid the need to burn the following year. Winter sowing for spring germination. Good monitoring; including seed, seedbed and germination. This Decision Support System has produced better silvicultural outcomes and has been instrumental in improving regeneration results in HEMS, but it has not overcome the problems inherent when certain growing conditions prevail. Delbridge (1999) found that use of the DSS was most effective in years when rainfall is not limiting. 53

63 Recently reported research by Dooley et al. (2005), investigating the effect of seedbed conditions on regeneration and the adequacy of slash burning prescriptions using the Aerial Drip Torch, again demonstrated the complex nature of regeneration success on HEMS coupes. Topsoil disturbance without burning resulted in more eucalypt regeneration than on burnt areas with or without topsoil disturbance - a total topsoil disturbance of 60% or greater resulted in satisfactory stocking percents. McCarthy and Dooley (2004) reported that while the recommended HEMS prescriptions to achieve a good burn do not necessarily result in acceptable regeneration success, this is not to say that the variables included in the burn prescriptions are not important. Moderate to hot burns result in a number of positive factors for regeneration including: the removal of slash, competing vegetation and organic layers; and heat induced seedfall from seed-trees. However, regeneration success in HEMS coupes is also dependent on many other factors as outlined above. The selection and implementation of silviculture in Victoria s HEMS forests is covered in detail in HEMS in Victoria s State Forests; Silviculture Reference Manual No.2 (Flint and Fagg 2007) HEMS Seed-tree and Clearfell Clearfelling was extensively practiced in HEMS forest throughout the eastern half of Victoria from the 1960s until the mid-1990s, when it was largely superseded by the seed-tree system, although it is still preferred for Shining Gum and Alpine Ash stands, and where seed from retained trees is problematic. Clearfelling is different to seed-tree in that it relies on direct sowing as the principal source of seed. On seed-tree coupes, seed supply is largely determined by the number and proximity of seed-trees, tree size and the seed crop carried by those retained trees. Typically, between 5 and 11 retained trees per hectare are required (Sebire 2001b). To maintain proportional species presence on harvested sites, as required by the Code of Practice for Timber Production (DSE 2007a), supplementary sowing/planting may be required occasionally. It has been common practice in east Gippsland coupes to aerially sow 25-50% of the standard full rate to cover predicted shortfalls in the seed available from seed-trees (Sebire and Fagg 2009). Additionally, artificial sowing is often carried out and timed, so as to promote spring germination. Following harvesting, site preparation to allow for seed to come into contact with mineral soil, normally uses fire to prepare the seedbed, and to induce seedfall where seed-tree silviculture is used. The conditions under which slash burning should be carried out were investigated by McCarthy and Dooley (2004), with several important factors identified to achieve a successful burn, as follows: Exposed surface fuel moisture content of less than 13%. Weather conditions, particularly temperature greater than 18ºC. Percentage of cloud-cover less than 20%. Atmospheric stability that is neutral or unstable, and Soil moisture content at a level where some dust is raised when small heaps of soil are kicked with the foot. Where the success of slash burning is problematic, mechanical disturbance (preferably with a root rack) is a suitable alternative. Whilst this form of site preparation is expensive it can be cost-effective when it reduces the risk of regeneration failure. In this situation (absence of fire), seed can be induced to fall from seed-trees with the application of a sub-lethal dose of herbicide. 54

64 HEMS Selection System Whilst in the past, selection systems were being used throughout the HEMS forests of Victoria, this has resulted in many forest stands becoming degraded in terms of sawlog productivity because the better trees have been removed and all that remain are the unmerchantable trees. In more recent times, the general use of selection silviculture in HEMS has not been recommended, and its use has been restricted to meeting particular structural needs (Sebire and Fagg 2009). There are specific issues regarding safety, economics, and the establishment and growth of regeneration. In NSW, harvesting in HEMS forests is often undertaken using a selection or tree retention system where trees are retained for further growth, ecological purposes or seed-trees. Decisions on tree retention are often influenced by the lack of markets for residual logs, resulting in the retention of high numbers of cull trees on many coupes. As noted below, this situation leads to suppressed regeneration and ultimately reduced sawlog production HEMS Shelterwood In the period , the shelterwood system was used to regenerate what would be considered to be HEMS sites in the Wombat Forest. Similar to the selection system, the general use of shelterwood in HEMS has not been recommended in recent times (Sebire and Fagg 2009). There are specific issues regarding safety, economics, and damage to retained trees HEMS Variable Retention The use of the variable retention silvicultural system has been trialed in east Gippsland in older growth forest, but it has not been used operationally in Victoria. The general use of this system in HEMS has not been recommended (Sebire and Fagg 2009). It is a system best suited to areas where some timber production is desired but maintenance of the forest structure and biological values found in older forests is also important. Experience from Tasmania has indicated that it is best suited to more productive sites, where the reduced wood volumes and additional costs will not make coupes unviable (Forestry Tasmania 2009). The system however has potential application on some Mountain Ash sites HEMS Reforestation and Re-treatment The HEMS forest type is moderately resilient to wildfire and has largely avoided being cleared for agriculture. Consequently, the need for reforestation is an uncommon issue in HEMS forests. However, remedial regeneration treatment is required on approximately 30% of area following first regeneration attempt (Fagg et al. 2008). Normal regeneration techniques can be used to retreat these areas (Sebire and Fagg 2009). 55

65 Thinning HEMS Forests Generally, HEMS forest does not self thin as actively as Mountain Ash or Alpine Ash forest. Thinning in this forest type has been confined to thinning of advanced regrowth for sawlog markets in northeast Victoria. In east Gippsland, thinning trials in Cut-tail fire-regrowth have been conducted, mainly for pulpwood production. The potential for mechanised thinning of regrowth from post-1960s logging would need to be established, as it is likely that large quantities of debris from trees felled to waste in the absence of a residual log market would impede access. The prescribed intensity of thinning is a compromise between maximising growth per hectare compared with maximising growth per tree. Growth per hectare is maximised provided that less than 50% of the stand basal area is removed (Incoll and Webb 1970), while growth per tree is maximised only if more than 50% of the stand basal area is removed (Goodwin 1990). The Victorian thinning prescriptions are based on the maximisation of growth per hectare. Native Forest Silviculture Guideline No. 14 Thinning of Mixed Species Regrowth (Appendix 1) details a commercial thinning specification using basal area limits, so that growth is maximised whilst minimising epicormic development and windthrow. The specification is based on research, expert advice and commercial thinning experience mainly gained from LEMS forests, but it should also be applicable to HEMS. Whilst there are no specific pre- or non-commercial thinning (or early spacing) specifications for HEMS forests, there are specifications which were used operationally in Messmate-Peppermint LEMS forests in the Wombat, Mt Cole and Otways State forests during the period These specifications are found in Guideline No. 14 and could provide useful guidance for any HEMS precommercial thinning Overwood Competition In relation to overwood, many studies have shown that the retention of older trees (commonly referred to as overwood) can have significant suppressive effects on the development of regrowth for a range of mixed species forests (for example, Incoll 1979, Bauhus et al. 2000, Bassett and White 2001), with the degree of impact being in part dependent on the shade tolerance of individual species. Bassett and White (2001) concluded that the retained overwood, as may be found in some uneven-aged silviculture, will impose a suppressive zone of influence on a regenerating stand equivalent to at least 1.7 to 3.0 crown radii around each retained tree. Clearly, the overall impact is highly dependent on the density of old trees, while the ability of regrowth to respond to release will be species dependent. Harvesting in the HEMS forests of NSW is often undertaken using a selection or tree retention system. Such practices result in forest that may have three recognisable age classes: old, advanced growth and regrowth. For such irregular forests, there is little growth and yield information. In a study based on destructive sampling of plots and growth ring analysis, Bi and Jurskis (1996b) estimated growth rates of Cut-tail forests in southern NSW. They developed three productivity classes which at a nominal age of 80 years, had total volume mean annual increments (MAIs) of 7.8, 4.24 and m /ha/yr for the good, average and low sites respectively. The study evaluated the impact of the retained mature trees on stand growth. Curves were presented up to 80 years of age; comparative data at 60 years are shown in Table 4.2. The impact of retained trees was significant in terms of reduced overall productivity and the effect of an individual old tree ranged from one to two crown radii equivalents (Bi and Jurskis 1996a). 56

66 Table 4.2 Volume mean annual increment of Cut-tail regrowth at 60 years of age for three different productivity classes and with three levels of old trees present in the overstorey. Overstorey 3 Volume Mean Annual Increment at 60 years (m /ha/yr) (trees/ha) Most productive Site Average Site Least Productive Site Nil old trees present old trees present old trees present Source: Bi and Jurskis (1996b) Seed Management As detailed previously, there is an ongoing need for sound monitoring of seed production in the HEMS forest type, both in relation to retained seed-trees and for the collection of seed for sowing. Sound seed management (including collection and long-term storage) is essential, particularly during periods of poor seed-crops when seed-crops on retained trees may be marginal and higher rates of seed may need to be artificially sown more extensively. Low seed inventories occur due to limitations on funding for seed collection, natural variation in seed crops, and lack of an adequate seed collection workforce. Consequently, in years when little seed is available for collection, shortages of specific seedlots can be common and a wider transfer of seed is required. To help manage these seed issues, research into seed-crop development and the maturation process has been carried out for a number of HEMS species (Murray and Lutze 2004, Murray and Terrell 2004, Murray et al. 2004). As previously discussed, this has resulted in the development of seed-crop assessment kits for use by field practitioners to assist in identifying trees suitable for seed-tree retention and at the desirable density, or the suitability of seed-crops for picking (for example, Sebire 2001b). Additionally, the value of seed-trees on carryover coupes to supply seed was monitored and guidance was provided on their value over time (Dooley et al. 2005) Low Elevation Mixed Species Forests Introduction These forests, which are also referred to as dry sclerophyll or foothill and coastal mixed species forests, occur in every region except northwest Victoria and cover nearly 63% of the gross area of State forest (Lutze et al. 1999a). There are similar forests in southeast NSW and in Tasmania. Compared with HEMS forests (covered in the previous Section), the LEMS forest type is typically located on warmer, drier sites, at elevations less than metres, and winters are less severe with lower frost frequency. The species composition of LEMS forests is highly variable, being determined by a range of environmental factors (for example, elevation, topography and annual rainfall), management factors (for example, utilisation histories) and the severity and extent of past wildfires. In relation to eucalypt overstorey species, nearly 30 species could be considered to occur in LEMS, as described by Kellas and Hateley (1991). Two main groupings of species occur: 57

67 5. The stringybark-peppermint-gum group. This type is widely distributed across the State and the proportion of each component varies according to location. As a generalisation, stringybarks predominate in the west and peppermints in the northeast. 6. The silvertop-stringybark group. This group, which can form a mosaic with the first, occurs east of Melbourne and south of the Great Dividing Range. The LEMS forests often occur in low-moderate rainfall areas and on less fertile soils. Commercial forests are generally considered to occur on those sites with a potential tree height of at least 28 metres. Typically, trees on these more productive areas range between metres in height, with trunk diameters (DBHOB) of mature trees generally ranging from centimetres. Figure 4.9 Extent of low elevation mixed forests (LEMS) in Victoria (provided by Peter Fagg, DSE). There has been a long history of silvicultural research in LEMS forests, with a strong focus on east Gippsland and central Victoria since the early 1970s. This research presence culminated in the establishment of Silvicultural Systems Projects in the mid to late 1980s, with study areas at Bullarto (Kellas 1987, 1994) and Cabbage Tree Creek (Squire et al. 2006). Detailed silvical studies were initiated, particularly in relation to flowering, fruit development and seed supply. This was in recognition from a commercial forestry perspective, that regeneration from seed following harvesting is the principal mechanism used in LEMS forests. While these forests have a record of generally reliable regeneration (for example, Bridges 1983, Murphy and Fagg 1996, Faunt et al. 2006, Fagg et al. 2008) a critical factor is that the volume and regularity of seed production by the major species is well suited to seed-tree regeneration systems. Studies by Faunt et al. (2006) and Kellas (1994) have demonstrated that regeneration can be established under a wide range of conditions in LEMS forests following harvesting, and that given adequate site preparation, seed supply was the major 58

68 determinant of levels of germination and seedling density. However, back in the mid 1980s when SSP sites were being established, the seed-crop development process for many of the species that comprise these LEMS forests had not been adequately studied to guide regeneration techniques and ensure that the species composition of regenerated forests more or less approximated the original mixture. The forests have been periodically affected by insect attack such as cup moth (Doratifera spp.) and gumleaf skeletoniser (Uraba lugens). In the initial assessment in East Gippsland, average levels of 71% defoliation were found but the follow up assessment indicated this had declined to 34% (Collett and Fagg 2010). Forest health surveys are being undertaken and recommendations have been made for continued assessments Silvics of LEMS Forests These forests have fire adaptive traits rather than population life cycle adaptations found in wetter forests where regeneration is more dependent on seed (Gill 1981). There are four sources of regeneration in these forests: seedlings, coppice from stumps, lignotuberous (dormant or quiescent) seedlings and saplings remaining after disturbance. Most stands are uneven-aged but the extent to which this is so, depends on their disturbance history. Many of the otherwise even-aged stands also have a high residual tree component, including large trees retained for specific purposes, and advanced growth trees. Species composition varies from pure or near pure to mixed stands. Shade tolerance varies between species, and this may partly determine the response of seedlings and lignotuberous seedlings to disturbance, and therefore species composition in a local area. Radial growth in LEMS forests is usually reduced following fire; the loss in growth generally being related to the amount of crown killed (Hodgson and Heislers 1972). Despite LEMS seed production being reasonably regular in general terms, with trees usually carrying abundant seed-crops throughout the year, and most species accumulating seed in capsule crops throughout successive years, species differences in seed-crop development processes can be important. For example, Bassett (2002) studied these processes in Silvertop Ash and White Stringybark and found differences in the timing and reliability of flowering, and the period over which seed-crops could be held-over in crowns. Intensive monitoring was undertaken by Bassett (2002) of seed traps in a stand comprising 67% Silvertop Ash and 33% White Stringybark according to basal area. For Silvertop Ash, inflorescence buds first appeared in mid-to-late spring and continued through summer, while flowering generally occurred months after the first buds. Silvertop Ash also exhibited highly synchronised flowering, with operculum fall peaking in the same fortnight over a 4-year period. Dissemination of seed from the capsules generally occurred months after flowering, but can continue for up to four years. This means that Silvertop Ash trees that flower every year could hold up to six consecutive capsule crops at any time (Bassett 2002). For White Stringybark, inflorescence buds are initiated in early winter and unlike Silvertop Ash, the time of flowering is highly variable. For example, the flowering time of individual trees in the same stand can vary by several months, while flowering behaviour of individual trees can likewise be highly variable. Bassett (2002) attributed part of this variability to soil moisture deficits. Seed dissemination from capsules largely occurs in the autumn of the second year following flowering, so that only two capsule crops are usually ever present. 59

69 The contrasting seed crop development processes of the two species studied by Bassett (2002) have important silvicultural implications. His work clearly shows that Silvertop Ash is well suited to a seedtree system due to its capability to produce large seed-crops every year and to store some seed in the crowns of old trees for up to six years. This contrasts with White Stringybark which due to variation in seed production is less suited to this system unless special care is taken in selecting the trees to be retained using seed-crop assessment techniques developed by Bassett (1996b) for LEMS forests. Some supplementary seeding may be required depending on the outcomes of the seed-crop assessments. Lutze (1998b) investigated seed supply, germination and survival of the five most common eucalypt species over a 4-year period for a range of silvicultural treatments at the SSP Cabbage Tree Creek experimental area. In general, Silvertop Ash not only had the greatest germination percent but also had an over-representation in the more intensively harvested treatments including a 10-hectare clearfelled coupe. It was concluded that, in order to maintain pre-harvest species composition, the optimal management practice would be to remove all overwood (and hence a seed supply) and apply seed artificially, with Silvertop Ash being under-represented in the species composition of the applied seed. This is particularly important given the findings of Bassett (2002) and Geary and Fryar (1999). The latter workers found that for seed collected for artificial sowing, the viability within a species can vary markedly depending on provenance, season of collection, and extraction, cleaning and storage practices. As an example, the range in viability for east Gippsland seedlots of Silvertop Ash and White Stringybark was 82,000 to 136,500 viable seeds per kilogram and 32,500 to 364,000 viable seeds per kilogram respectively. Germination tests that are routine practice (Wallace and Fagg 1999) are therefore critical for not only the use of efficient seed but also to manipulate species composition Stocking Following Harvesting of LEMS in Victoria Past and present approaches to the silvicultural management of Victoria's LEMS forests have been outlined by Campbell et al. (1984) and more recently Lutze et al. (1999a). Recent reporting by Fagg et al. (2008) found that between 1996/97 and 2000/01 on public land, harvesting and regeneration of the LEMS forest type totaled about 15,300 hectares, an average of about 3,000 hectares per annum. Silvicultural systems producing even-aged regrowth were employed on 90% of this area, with the majority of this concentrated in east Gippsland (68%), and the rest more broadly located south of and on the Great Divide. The small area of regeneration from uneven-aged silvicultural systems (1,320 hectares over the five years) was concentrated mainly in the southwest, within the Portland FMA. Analysis of the silvicultural systems used in LEMS forests for the period 1996/97 to 2000/01 (Tables 14 and 15 in Fagg et al. 2008), showed, at the first attempt at regeneration, the following percentages of treated coupe area that were satisfactorily stocked: Clearfelling - 82% (2,370 hectares treated and surveyed) Seed-tree - 94% (11,480 hectares) Group and single tree selection - 99% (1,320 hectares) All sources of regeneration are counted in stocking surveys provided they meet acceptable criteria (Dignan and Fagg 1997). These percentages indicate that a range of silvicultural systems are appropriate for LEMS forests, and they are discussed below. 60

70 4.7.4 LEMS Silviculture LEMS Seed-tree System The seed-tree system has been used for decades in LEMS forests, particularly in east Gippsland and relies on there being an adequate on-coupe seed-crop (Fagg 2001). The number and proximity of seed-trees depends on tree size and the seed-crop carried by those retained trees and is typically 2 between 5 and 15 trees per hectare, or approximately 10% (normally about 5 m /ha) of the preharvesting stand basal area. Generally, supplementary sowing is not required, however, where a tree species is present on the coupe, but is carrying little or no mature seed capsules, then in the absence of adequate regeneration, supplementary sowing may be required to maintain proportional species presence on the site as required by the Code of Practice for Timber Production (DSE 2007a). Site preparation following harvesting, normally uses fire to prepare the seedbed and induce seedfall. It is not general practice to harvest the seed-trees following successful regeneration, although this can be an option. Suppression of regrowth from retained seed-trees is unlikely to be a serious problem in the regenerating forest unless a significant number remain (that is, greater than 10 seed-trees per hectare). In such a situation, the excess competing trees may be removed by felling or by herbicide treatment (for techniques, see Sebire 1999). The health and survival of seed-trees is critical to the success of seed-tree systems that retain, rather than remove, the seed-trees after site preparation by burning or mechanical disturbance. Five years after harvesting a mixed species forest in east Gippsland, the overall depletion rates of seed-trees due to windthrow or death partly as a result of the Cinnamon fungus (see Section 6) was only 5.4 %, indicating that there is good insurance against the risk of a wildfire destroying young regrowth before it has reached seed-bearing age (Featherston 1983) LEMS Clearfell System The clearfelling system is relatively uncommon in LEMS forest (15% of total harvested area), although the predominant LEMS silvicultural system, seed-tree, is sometimes regarded to be just a variation of clearfelling. However, unlike the seed-tree system, clearfelling relies on direct sowing as the principal source of seed. This system is only appropriate where regeneration is reliant on seed supply from retained trees and this is problematic or specific species manipulation is required LEMS Selection System Selection systems have been used extensively throughout LEMS forests, although they are not currently common. In earlier times, this was often in the form of sawmiller selection with little control over which trees are harvested and little attempt at achieving regeneration. Where this system is used, harvesting is generally of small groups (that is, group or gap selection), at relatively short time intervals (10 to 20 years), repeated indefinitely. This involves the frequent establishment of regeneration and an uneven-aged stand results. As noted earlier in this review, many of the LEMS forests, particularly those in east Gippsland, have been selectively logged for lengthy periods (see McKinty 1969, Featherston 1985) and been subjected to disease and periodic intense wildfires. McKinty (1969) commented that selection logging for sought-after species (for example, durable species) and good quality trees over an extended period had resulted in degeneration of the forest 61

71 stands because the best trees were removed leaving defective old trees and suppressed regrowth with poor form. In the 1970s, selection logging was replaced by more intensive logging to inter alia encourage regeneration and overcome potential disease (from the Cinnamon fungus) and species mix problems (Lutze et al. 1999a) LEMS Shelterwood System In the period , the shelterwood system was used widely in Wombat Forest, which contains significant areas of the LEMS forest type. It was introduced to replace thinning and selection systems that were creating a species shift from Messmate to the slightly more shade-tolerant peppermint and gum types (Kellas 1994). The shelterwood system was selected to provide a more balanced regeneration outcome appropriate for all species, using on-coupe seed, and to enable good growth outcomes for both the regeneration and the retained seed-trees LEMS Variable Retention System In LEMS forests, the variable retention silvicultural system has not been used operationally in Victoria. It is a system best suited to areas where some timber production is desired but maintenance of the forest structure and biological values found in older forests, is also important. Experience from Tasmania has indicated that it is best suited to more productive sites, where the reduced wood volumes and additional costs will not make coupes unviable (Forestry Tasmania 2009) LEMS Reforestation The need for reforestation, except in restricted circumstances, is an uncommon issue in LEMS forests. It is mainly restricted to situations where there has been a man-made vegetation change in the past and there is a desire to return the land to its previous condition. A recent example of reforestation is the Delatite Arm Re-vegetation Project, a large-scale rehabilitation project that has revegetated cleared Radiata Pine (Pinus radiata) plantation with native species to re-establish LEMS forest. The project, near Lake Eildon in northeast Victoria, commenced in 2002 and has produced a number of publications on site preparation, weed control, revegetation using native species, and monitoring and evaluation (Kasel and Bennett 2007, Bennett et al. 2008, Kasel 2008, Kasel et al. 2008). The feasibility of restoring dieback-affected forests in east and south Gippsland has received considerable attention. While planting is an option (Lutze 1998a), the work of Fagg (1987) showed that sowing is also a potential and cost-effective treatment. Three trials were established comprising seven sites on dieback sites in east Gippsland to investigate whether satisfactory regeneration could be obtained by sowing. Both burnt and cultivated seedbeds were tested. Satisfactory eucalypt regeneration was achieved in two of the trials. Key factors for this success, apart from seed supply, were the provision of a well distributed, receptive seedbed and favourable conditions for germination and rapid seedling growth. Results at one of the sites indicated that a partial cutting system using natural seedfall induced by slash burning should be successful where there is a sufficient number of older trees carrying adequate seed. Importantly, Fagg (1980) found that in addition to Silvertop Ash 62

72 and White Stringybark, species more tolerant to dieback could also be established by sowing. While a severe rainfall event that would be conducive to a dieback outbreak was not experienced over the period of the study, the results combined with planting trials give confidence that these dieback forests can be restored to commercial production Thinning While LEMS thinning, mainly for firewood, posts, poles and pulpwood, in the s encouraged the rapid development of regrowth, it was not until the mid-1980s that extensive routine operations began in earnest with the development of machine-based thinning. This was in response to reduced access to high yielding native forests in the 1980s and 1990s and a need to promote future sawlog production in even-aged LEMS regrowth which has resulted from wildfires and past logging (Flinn and Mamers 1991, Lutze at al. 1999a). A major focus was the Silvertop-Stringybark regrowth in southeast NSW, and east and central Gippsland using excavator-based felling heads which were developed from softwood thinning machinery (Kerruish and Rawlins 1991). Pole-age (20-30 plus years) regrowth stands of even-aged Messmate (Kellas et al. 1987, Brown 1993) and Silvertop Ash (Connell and Raison 1996, Fagg and Thomson 2001, Connell et al. 2004) are now being commercially thinned to accelerate the growth of retained future sawlog crop trees. Figure Thirty two-year old Silvertop ash forest recently thinned from below. 63

73 As for the HEMS forest type, the Victorian thinning prescriptions for LEMS are outlined in Native Forest Silviculture Guideline No. 14 Thinning of Mixed Species Regrowth (Appendix 1), which details the commercial thinning specification using basal area limits, so that growth is maximised whilst minimising epicormic development and windthrow. The specification is based on LEMS research, expert advice and commercial thinning experience. Pre- or non-commercial thinning (or early spacing) specifications for LEMS forests, which were used operationally in the Messmate-Peppermint areas in the Wombat, Mt Cole and Otways State forests during the period , are also found in Guideline No Growth Response Regeneration Kellas (1994) recorded seedling percents of 6.7, 2.7, 4.8, 1.9 and 2.5% under 0, 10, 15, 20 and 26 2 m /ha respectively of retained overwood in Messmate-Peppermint forest near Bullarto in the Wombat Forest. This represents a general pattern of survival decreasing with increasing residual basal area, that is, increasing shading and competition for moisture. Similarly, at the Cabbage Tree SSP site, patterns of mortality followed those of germination to 3 years of age, with total mortality decreasing as gap size increased and as levels of retained overwood decreased (Faunt et al. 2006). Reporting on the growth of regeneration in Silvertop-Stringybark forest to age 10 years, at the same site, Lutze and Faunt (2006) found that rates of diameter and height growth in the regenerating stands were reduced by increasing levels of retained overwood and decreasing gap size, and this was reflected in the 3 volume index (m /ha) at age 12 years (Figure 4.11). 180 V o lu m e in d ex (m 3 /h a ) ha gap 4 ha gap 10/0 (10 ha 7% overwood gap & 0% overwood) Volume from 1989 treatment 22% overwood 35 % overwood 10% overwood Volume from 1990 treatment Figure 4.11 Volume index of regeneration at age 12 years for each harvesting treatment applied in two separate years (after Lutze and Faunt 2006). In terms of early growth, clearfell and seed-tree systems are likely to be the most appropriate silvicultural systems on sites where wood production is a high priority. However, where ecological values have a higher than normal value, SSP findings indicate that regeneration can be achieved using a range of gap sizes and overwood densities, suggesting that other silviculture such as retained overwood systems may be successful. 64

74 Regrowth Development Regrowth following wildfire or developing after intensive logging is essentially even-aged due to a high seeding component. Bridges (1983) provided data on stand development of a LEMS forest in southeast NSW dominated by Silvertop Ash (Table 4.3) which showed that stocking levels following wildfire can be very high for prolonged periods. For example, the stocking level 14 years after a fire was 26,700 stems/ha. However, self thinning reduced this high initial stocking to 1,470 stems/ha by age 38 years. In the forests studied by Bridges (1983), stocking levels in logging regrowth were lower than for fire regrowth, but were more than adequate from a regeneration perspective. This lower initial stocking is reflected in greater mean diameter growth up until nine years of age. Table 4.3 Growth data from regrowth stands in the Silvertop Ash/Stringybark type forest (from Bridges 1983) Origin Logging Fire Fire Fire Fire Age (yrs) Stocking (trees/ha) Mean dbhob (cm) Mean dominant height (m) 2 Basal area (m /ha) 6,000 37,000 26,700 2,300 1, Volume under bark (m /ha) Figure 4.12 Regrowth LEMS, Orbost 1992 (photo from DSE, Victoria). 65

75 Thinning and fertilizing The magnitude and duration of LEMS growth responses can be dramatically affected by the timing of thinning treatments and the physiology of the species involved. In fast growing species, such as Silvertop Ash, first thinning may be within 2 to 4 years of crown closure, whereas the timing of thinning for more shade tolerant and persistent species, such as Messmate, would be less critical because the effects of competition on the tree condition are less apparent. Results from a collaborative study between CSIRO and the Department of Natural Resources and Environment demonstrated strong growth responses to both pre-commercial thinning and commercial thinning (Connell and Raison 1996). In the former instance, early spacing of 9-year-old Silvertop Ash regrowth resulted in a two-fold increase in basal area three years after treatment, while in the latter situation, thinning of 26- to 28-year-old stands that reduced basal area by around 40% also almost doubled growth rates of retained trees. Furthermore, Connell and Raison (1996) found additional but variable responses following fertilization of thinned stands with nitrogen and phosphorus. Similar studies have been conducted in LEMS forests in NSW. As an example, thinning of 25-yearold stands at Eden yielded sufficient timber to make the operation commercially viable (Bridges 1983). Growth responses were also significant. Additionally, applying fertilizer to stands immediately after thinning led to significant growth responses. As this growth is on retained trees developing into higher value sawlogs, such treatments are an option for forest managers depending on better information becoming available on the long-term ecological consequences of fertilization. Many of the soils supporting LEMS forests are low in nutrients so that fertilizer additions can be targeted to nutrients known to be in short supply (Table 4.4). After five years, the use of fertilizer led to an additional m /ha of sawlog. The effects of thinning and fertilizer on average diameter were: Unthinned control (all trees) 17.6 cm Unthinned control (best 230 trees) 24.1 cm Thinned control 30.5 cm Thinned plus NP fertilizer 32.7 cm 3 Table 4.4 Volume response (periodic annual increment (m /ha/yr) five years after fertilizer treatment) of Silvertop Ash to applications of fertilizer applied after thinning from 1,500 stems/ha to approximately 250 stems/ha. Application of Nitrogen Application of Phosphorus Nil 100 kg N/ha 200 kg N/ha Nil kg P/ha kg P/ha Source: V. Jurskis, Silviculturalist, Forests NSW pers. comm. Young (16- to 21-years-old) E. obliqua stands in Tasmania were pre-commercially thinned and fertilized (La Sala 2006). There was little gain in basal area increment from thinning but there were gains in larger tree sizes plus responses from fertilizer application. Pre-commercial thinning of 50% of basal area was recommended to enhance growth of final crop trees. 66

76 Research clearly indicates that there is considerable scope for intensive silviculture to increase yields and economic returns from east Gippsland LEMS forests (Connell and Raison 1996). They note, however, that intensive silviculture involving fertilization would need to be initially confined to the least environmentally contentious sites close to markets where the most benefits will be realised. Connell and Raison (1996) also stressed the need for more operational scale experience with intensive silviculture Overwood Competition Similar to the HEMS forest type, the retention of older trees can have significant suppressive effects on the development of regrowth for a range of mixed species forests, with the degree of impact being in part dependent on the shade tolerance of individual species and the density of older tree retention (Bassett and White 2001). The effect of retained overwood on dominant seedling height growth, three years after harvesting, compared with clearfelling is illustrated in Table 4.5. Table 4.5 Effect of retained overwood on dominant seedling growth, three years after harvesting, compared with clearfell for LEMS forest (source Bassett and White 2001). Retained Overwood Height Loss (% of Basal Area) (%) 10% 13% 30% 55% 50% 67% 100% 82% Impacts of Overwood on Regeneration Retention of increased numbers of habitat trees (non-commercial trees) and more attention to their spatial arrangement (for example, clumping) is increasingly being advocated to enhance fauna and other forest values. There is debate, however, about the level of impact of such retention strategies on future commercial production and is relevant to a range of forest types. Two aspects of noncommercial tree retention need to be considered, namely their competition with commercial trees in established stands and their zone of influence on regenerated coupes. In established stands, trees that have no commercial value still contribute to the overall productive capacity of a site in the same manner as trees with commercial value. On the assumption that tree volume is directly proportional to basal area, the extent of commercial stem volume lost through competition from non-commercial trees for a range of stocking levels was reported by Bassett and White (2001) (Table 4.6). These losses are substantial and are an under-estimate if the non-commercial trees are mainly in the dominant crown classes. 67

77 Table 4.6 Production loss due to retention of single trees with different diameters for a range of stocking levels (from Bassett and White 2001). Retained tree Production Loss (%) for a Given Stand Stocking Diameter Basal Area 50 m /ha 40 m /ha 25 m /ha 15 m /ha 2 (cm) (m /tree) Following harvesting, retained overwood will have a suppressive effect on the density and growth of the regeneration. This so-called zone of influence has been well studied. The impacts of retained trees on regeneration (when compared with a clearfell situation) increase markedly with increasing levels of overwood retention (Table 4.7). For example, three years after harvesting, volume losses in Mountain Ash can be as high as 45% with 30% overwood retention. Bassett and White (2001) noted that the zone of influence of retained trees is typically 2 to 3 times their crown radii, so that a relatively small number of scattered non-commercial trees can have a major impact (for example, 25% of a coupe can be affected by retention of only five trees per hectare). It is important that the productivity trade-offs are taken into account when identifying practices that result in balanced outcomes. Table 4.7 Effect of retained overwood on dominant seedling growth, 3 years after harvesting, compared with clearfell for LEMS forests, Mountain Ash mixed forests in Tasmania and Mountain Ash forests in Victoria. Source: Bassett and White (2001) Retained overwood Height loss (%) Volume loss (%) (% of Basal Area) LEMS Mountain Ash Mountain Ash (Tasmania) (Victoria) n.a

78 Box-Ironbark Forests Introduction As rainfall decreases inland, the LEMS forests give way to Box-Ironbark (eucalypt) forests. In Victoria, these forests once occupied more than 4.7 million hectares (Newman 1961). Today this is much reduced, with the main contiguous area in the country between the Northern Plains and the Great Dividing Range which is drained by the middle reaches of the Goulburn, Campaspe, Loddon, Avoca and Wimmera Rivers (DNRE 1998). Within State forest, there are about 240,000 hectares, mainly in north central Victoria, with isolated patches occurring near Chiltern, Heyfield and Airey's Inlet. In addition, there are parts of lowland forests in east Gippsland where Red Ironbark (E. tricarpa) is a component species (Figure 4.13). Figure 4.13 Location of Box-Ironbark forests in Victoria (provided by P. Fagg, DSE). Victoria's Box-Ironbark forests have been extensively disturbed since European settlement. With reports of fertile pastoral country in central Victoria in the early-1800s, much of the area was soon covered by extensive grazing leases. Initially, the environmental impacts of grazing were minimal, however, fertile forested land was progressively cleared for more intensive agricultural uses such as cropping and horticulture. The discovery of gold in the region, for example in Castlemaine in 1851, initiated dramatic long-term changes to these forests. Original stands of Box-Ironbark were clearfelled to provide timber and fuel for the mining industry and associated settlements. In the 1890s, the rapid expansion of the railway system across Victoria made additional demands for heavy construction and sleeper timbers. Much of this timber was provided from the Box-Ironbark forests because of the extremely hard, durable, and strong nature of wood from the species. These timbers were also in high demand during and after the two World Wars. 69

79 By the 1920s, all Box-Ironbark forests, especially those near population centres, had been selectively cut-over several times. This demand required forest agencies to develop forest utilisation controls. From 1928, Working Plans that set out the objects and methods of forest management were developed and implemented (Newman 1961). Early thinning efforts were greatly accelerated during the economic depression of the 1930s. Large workforces of unemployed men were involved in forest thinning and improvement works. A similar burst of activity occurred in the 1940s when licensed operators, in conjunction with alien and prisonerof-war labour, selectively harvested extensive areas for firewood and charcoal. This was necessary to sustain domestic industries, largely starved of coal and liquid fuel due to the demands of War (Newman 1961). The requirements for power and telephone poles and fencing timbers also increased greatly during this period. Forest management practices over this period were designed to encourage rapid growth of the regenerating forest, leading to the removal of many older trees that escaped harvesting during the gold rushes. Also, there was removal of trees with perceived 'defects', including hollows, as they were judged unlikely to produce future sawlogs. Today Box-Ironbark forests are only a small proportion of the Victorian public forested land (about 5%). Commercial timber productivity in these forests is low and management is on a multiple-use basis for a variety of forest products such as fencing timbers, firewood, honey and eucalyptus oil together with other values such as recreation, wildlife and grazing (DSE 2008d). Silviculture in recent decades has been largely regrowth thinning at regular intervals combined with single tree selection for sawlogs. Sawlog production is restricted, more because of slow growth than lack of demand, and is currently low. Annual rainfall is generally between 380 and 510 mm, characterised by long dry summers and frequent high temperatures and periodic drought, and sometimes with severe winter frost. Permanent streams are few and topography is undulating with an average elevation of 250 metres. Soils are characteristically poor, chiefly shallow sandy or clay loams over clay of Ordovician or Silurian origin and are low in fertility and permeability. Three sub-communities or species associations are recognised as shown in Table 4.7. Table 4.7 Forms of Box-Ironbark forests in relation to topography and growth (Kellas 1991). Species association * Topography Growth Maximum Stand Height (m) Red Ironbark/Red Stringybark/Red Box Steeper slopes Low Red Ironbark/Yellow Gum/Grey Box Lower slopes Moderate Grey Box/Yellow Gum Alluvial plains Higher > 24 The current forest structure is indicated by data from the Box-Ironbark Timber Assessment Project in the Bendigo Forest Management Area and Pyrenees Ranges, where there is an average of almost 500 stems per hectare, most having less than 25 cm diameter (DNRE 1998). However, there is considerable variation, as illustrated by the data in Table

80 Table 4.8 Number of stems per hectare by diameter class in each Working Circle. Data from recommended parks and reserves have been excluded (VEAC 2001). Stocking (stems per hectare) Working Circle < 20 cm cm cm > 60 cm Total St Arnaud Inglewood-Dunolly Avoca Maryborough Bendigo Castlemaine Rushworth-Heathcote In their final report on the Box-Ironbark investigation, VEAC (2001) suggested that prior to European settlement, Box-Ironbark forests in many areas carried fewer and larger trees. It is estimated that 20 to 30 large trees per hectare would have been reasonably common and this contrasts markedly with the present-day situation as depicted in Table Figure 4.14 Ironbark forest Dargile (photo from DSE). 71

81 4.8.2 Silvics and Regeneration of Box-Ironbark Forests Red Ironbark reproduces vigorously from coppice (sprouting from cut stumps) and regeneration from seed is much less frequent except on the moister sites. On these sites, successful establishment of Box-Ironbark seedling regeneration requires good site quality, large canopy openings, good ground coverage of grasses or other low vegetation, favorable rainfall and the exclusion of browsing or grazing. While stems of seedling origin may only comprise around 25 % of all stems, they can make a higher contribution to total stand basal areas. Regeneration from seed can be problematic, even though flowering can be fairly regular. Keatley et al. (2004) reported on the average flowering season between 1940 and 1970 for four key Box-Ironbark species. The findings were based on the records of the former Forests Commission of Victoria, which covered the flowering of Yellow Gum, Grey Gum, Red Box and Red Ironbark, and suggest that some overlap of flowering occurs for all four species. Flowering is regular, with Yellow Gum flowering every year, Grey Box and Red Ironbark on average missing one year in six, and Red Box missing four years in thirty. Despite this, Kellas (1991) found that good seed production was limited for Red Ironbark, with one satisfactory year followed by a number of years when little or no production occurred. He also noted that seed harvesting ants reduced the number of seeds available for subsequent germination. In an east Gippsland study, Lutze (1998c) found that regeneration by planting Red Ironbark seedlings would be most likely to be successful if overwood and understorey competition was minimised, and if fencing or tree guards were used to protect the seedlings from browsing animals. Browsing (by wallabies) can also inhibit successful coppice development, especially of Yellow Gum, with Grey Box and Red Ironbark affected only in a minor way (Vearing 2003). Coppice can also be retarded, or even killed, when there is heavy infestation by the native semi-parasites, Dodder-Laurel and Mistletoes of various species. Removal programs in the past have sought to control these plants (see Kellas 1991), but no active control programs are now employed. The heavy cutting during the latter half of the 19th Century has resulted in seedling and coppice regeneration over extensive areas of the central Victorian Box-Ironbark forests. Supervised harvesting and thinning produced forests containing essentially two size-classes, with various strata of regrowth beneath older and larger overwood trees. Typically, overwood stems are uniformly 2 distributed with a total basal area of about 11 m /ha, whereas regrowth occurs in clumps within which 2 basal area may be equivalent to about 10 m /ha, with individual stems often under intense competition (Kellas et al. 1998). Across the Bendigo FMA, there is an average stocking of about stems per hectare (ranging from 229 to 780) and an average basal area of 12.5 m /ha (ranging from 9 to 19) (DNRE 1998). Diameter growth in fully or over-stocked stands is very low and recruitment into larger size classes relies on reducing competition through death or removal of individual trees. Natural self-thinning in Box-Ironbark forests is slow because the trees are tolerant of extreme conditions such as drought and fire. Average diameter increment varies between 3.2 and 3.8 mm per year. Newman (1961) reported average mean annual increments for Box-Ironbark regrowth stands at 3 around 60 years of age as being about m /ha. 72

82 4.8.3 Box-Ironbark Silviculture Past and present approaches to the silvicultural management of Victoria's Box-Ironbark forests are outlined by Campbell et al. (1984) and more recently Lutze et al. (1999a). Recent reporting by Fagg et al. (2008) found that on public land, between 1996/97 and 2000/01, harvesting and regeneration of the Box-Ironbark forest type occurred mainly in the Bendigo FMA and totaled about 2,030 hectares, averaging about 500 hectares per annum. Selection silviculture producing uneven-aged regrowth was employed exclusively on this area. On average, 95% of area that was treated and surveyed was satisfactorily stocked. Surveys in these uneven-aged stands were conducted according to Native Forest Silviculture Guideline No. 10 Eucalypt Stocking Surveys (Appendix 1) Selection As outlined above, selection systems have been used exclusively throughout Box-Ironbark forests. This silvicultural system involves the selection-felling of marked trees, either individually or in small groups, with the objective of producing sawn timber products, whilst minimising impacts on species composition and forest structure. Harvesting is focused on previously-thinned sites which contain trees up to 59 cm dbhob, and involves the retention of at least 50% of the pre-thinning basal area, including all trees >60 cm dbhob and trees with identified habitat values (for example, hollows). Species composition is maintained (DSE 2008d). Sawlog and sleeper operations cut trees from 45 cm to 60 cm diameter. Post-cutters harvest trees up to 40 cm diameter, mostly for sawing into split posts and other fencing products, and cut smaller dimension wood, producing round posts. Firewood is produced as a by-product of the sawlog harvesting and post-cutting, from the heads of felled trees and thinning of small stems Thinning A large proportion of smaller diameter trees in the Box-Ironbark forests and woodlands of Victoria may be in the form of multi-stemmed regrowth (or coppice). Because of the persistent nature of BoxIronbark species, thinning operations are used to release selected trees by reducing competition. Studies on thinning have indicated that removal of competing coppice and spacing trees widely will lead to improved growth on individual trees (Kellas et al. 1982). The results showed that individual Red Ironbark trees retain a capacity to respond to reductions in competition in fully stocked stands. For regrowth (dbhob <20 cm), the response can be rapid with the removal of competing overwood and can be further enhanced by also reducing the regrowth competition. The response of overwood (dbhob >20 cm) to release is slower. For both regrowth and overwood trees, total competition from all competitors appears more important than that from overwood or regrowth alone. These studies show that while the Box-Ironbark forests have low productive capacity relative to forests in the higher rainfall zones, significant growth responses can be expected with appropriate thinning regimes but responses may be limited to less than 10 years, requiring periodic thinning for sustained responses. However, it has been suggested that impacts on other values (particularly flora and fauna) would not make this appropriate. 73

83 For comparison, in 1935 a stand of Red Ironbark in State forest near Heathcote, was uniformly planted at a density of about 1,700 stems per hectare. There was considerable mortality in the stand 2 and in 1997, the stocking had reduced to 617 stems/ha, with a basal area of 22.7 m /ha and a mean dominant height of 15.6 metres. At this time, Murphy and Forrester (2008) reported that a trial was established, using a number of thinning treatments, to better understand the thinning response of Red Ironbark in a small stand situation, particularly in the cm (dbhob) size class, and to also better understand the effect of the coppice on retained tree growth. On 40 m x 40 m (0.16 hectare) replicated plots, thinning from below was used to reduce basal areas down to 33%, 50% and 66% retention. A 33% treatment with coppice control was implemented. They found that ten years after treatment, only the 33% retention thinning increased the basal area or volume of the largest 100 selected crop trees per hectare. When coppice was retained, the thinning response was reduced, probably due to competition for water resources more than light or nutrients. Across the thinned treatments there was a 63% survival of cut stumps, with coppice diameters (OB) ranging from < 1 cm to 12.4 cm and the height of the tallest stem on each stump ranging from 2.1 m to 8.1 m. Coppice 2 basal area was 1.9 m /ha, making up 15% of the stand total. Over the life of a stand, coppice survival has been found to decline after overstorey canopy closure in Red Ironbark (Kellas et al. 1998), and it is likely that coppice competition will similarly decline. The Victorian thinning prescriptions for Box-Ironbark forests are outlined in Native Forest Silviculture Guideline (NFSG) No. 15 Thinning of Box-Ironbark Forests (Appendix 1), which details thinning specifications using basal area limits, so that growth is maximised whilst minimising epicormic development and windthrow. The specification is based on Box-Ironbark research, expert advice and thinning experience. Pre- or non-commercial thinning (or early spacing) for Box-Ironbark forests, is only conducted when labour and funds are available. In recent times this has been rarely Ecological thinning Ecological thinning is appropriate in Box-Ironbark forests where the main aim is to improve fauna (and understorey flora) rather than future timber production, by thinning to increase growth of selected trees. In ecological thinning, there are increased requirements to keep larger trees (regardless of form or health), retain all hollow-bearing trees, avoid damaging understorey elements, and retain trees containing mistletoes (Native Forest Silviculture Guideline No. 15). In 2003, Parks Victoria initiated an Ecological Thinning Trial in Box-Ironbark forests (Parks Victoria 2007, 2009), as part of a long-term field-based trial to evaluate different methods of ecological thinning and the effects they have on key habitat characteristics, with the broad aim of restoring a greater diversity of habitat types to the landscape. The trial was examining the retention of trees of all forms and sizes in a patchy distribution, with clumps at higher tree densities within a general mosaic of wider spaced trees. From these trials operational specifications have been developed and are outlined in Guideline No

84 River Red Gum Forests Introduction River Red Gum (Eucalyptus camaldulensis) has been a highly prized timber ever since European settlement, and the forests provide an excellent example of successful multiple-use management as evidenced by wetlands of international significance, natural history and cultural heritage, contribution to local, regional and State economies and popularity for tourism and a wide variety of recreational pursuits (Dexter and Poynter 2005). The species is one of the most widespread of naturally distributed eucalypts and is found in all States except Tasmania (Pryor 1979). River Red Gum is one of the most durable native timbers produced in appreciable quantities in Australia as well as being appreciated for appearance and strength. It has been commonly used for bridge and wharf piling, decking, house stumps, urban and rural fencing, railway sleepers, premium grade charcoal (up to 60 charcoal retorts were operating within the Barmah Forest in 1941), firewood and woodchips for garden mulch in addition to being highly regarded as a sawlog species for the manufacture of furniture and other value-added products. It is also a highly prized species for wood turning Site Characteristics of River Red Gum Victoria s main River Red Gum forests of commercial or potential significance are confined to immediate floodplains of the Murray River system. Dexter and Poynter (2005) identified seven blocks of forest namely, Mildura, Red Cliffs, Nyah-Robinvale, Gunbower, Goulburn, Barmah and Yarrawonga-Ovens. The Barmah floodplain area has a section of limited capacity known as the Barmah Choke which causes the river flows in the upper river to spill via a complex system of effluent creeks and branches in the forest. On the Victorian side, the water drains back into the river near the town of Barmah while in NSW, water passes through the Edwards and Wakool systems back to the Murray. Parsons et al. 1991) noted that the floodplains have led to a complex system of forests, woodlands, grasslands and wetlands. The riverine forests and associated biological communities of the central Murray Valley have developed in response to the following important environmental characteristics (Parsons et al. 1991): A semi-arid climate. Terrain with generally little relief. Heavy alluvial soils with swelling clays (which can prevent infiltration) overlying sand lenses. A deep (10 m to 20 m ) groundwater table. A water supply heavily dependent on surface flooding from the river system rather than on rainfall, and Alternate wetting and drying of the surface and near sub-soil environments. 75

85 There is strong correlation between plant communities and the natural flooding frequency (Table 4.9). Table 4.9 Relationship between natural flood frequency and plant communities (from Parsons et al. 1991). Flood Regime Community or Communities Annual prolonged (>4 months) inundation Rushlands, permanent wetlands Inundation for 3-4 months in 80% of years River Red Gum forest (>20 metres mature stand height) with various grass communities Inundation in <80% of years River Red Gum forest (<20 metres mature stand height) and woodland Brief inundation in 30% of years Black Box woodland and open woodland Seldom, if ever, flooded Yellow Box, Grey Box, Murray Pine woodland and open woodland Silviculture of River Red Gum The silvical characteristics of River Red Gum have been comprehensively studied by Dexter (1967, 1970), and his findings continue to provide the underlying science for the silvicultural management of these unique forests. Whilst Dexter (1970) focused his research on the Barmah Forest, the results are applicable to all River Red Gum forests in the central Murray River region. The extensive work of Dexter (1967, 1970) shows that River Red Gum flowers regularly from latespring to mid-summer, but poor flowering years can occasionally occur due to defoliating insects. In contrast to many eucalypt species, River Red Gum seeds mature relatively rapidly in the capsules (good germinative energy and capacity can occur within a few months of flowering), and natural seedfall occurs throughout the year starting about 9 months after flowering. However, peak seedfall is in spring. Dexter and Poynter (2005) noted that this may have adaptive significance as floods usually recede during this period, therefore providing optimal seedbed conditions for germination. Seedbed type and environmental conditions strongly influence seed losses which can be as high as 100%. This is a critical factor to consider in identifying successful regeneration systems. Dexter (1970) found that seedbeds which protected seed from desiccation (for example, seed covered by ash or soil) resulted in improved germination. Importantly, initial growth rates of seedlings, which are critical to their survival and development, were found to be higher on ashbeds and cultivated soil compared with compacted surfaces and grassy sites (Dexter 1967, 1970). Extensive surveys of stocking levels and relevant site factors were conducted in the 1960s on areas subject to repeated fellings over several decades using a single tree selection system (Dexter 1967). Detailed analyses of these data showed that natural regeneration had been highly variable. Dexter (1970, 1978) concluded that: River Red Gum naturally regenerates in irregular, even-aged patches, and Attainment of satisfactory regeneration (as judged by density and distribution) depends on many factors including seed supply, incidence of flooding, the time taken for floods to recede, seedbed type, availability of sub-soil moisture, distribution of summer rainfall, impacts of insects and domestic/non-domestic animals, duration and depth of flooding in the season following germination, and competing vegetation (overstorey, grasses and weeds). 76

86 These factors must be considered in identifying an appropriate silvicultural system for the River Red Gum forests. River Red Gum typically has weak apical dominance (Jacobs 1955), so there are compelling reasons to encourage dense clumps of regeneration and hence improve the form of saplings. The potential productivity of the forests has commonly been under-estimated by sections of the community. Baur (1983) reported provisional yield tables for even-aged stands of River Red Gum on the NSW side of the Murray River, indicating that SQ I stands subject to natural flood regimes could 3 achieve an MAI volume of 4.8 m /ha/yr over a 108-year rotation. However, this could be regarded as near the upper limit of productivity on high quality sites subjected to optimal flooding regimes. At the 3 other end of the growth range, MAIs of less than 1 m /ha/yr are common. th Despite the massive exploitation of the forests prior to the 20 Century, they still remain capable of producing significant volumes of highly prized timber products. The sustainable harvest from the Mid 3 Murray FMA is presently set at 5,200 m of sawlogs per annum and around 8,700 sleepers per annum. Dexter and Poynter (2005) provided an instructive account of the history of the central Murray River Red Gum forests. They noted, inter alia, that the health of the forests has been in slow decline due primarily to an altered forest flooding regime (frequency, timing and extent). There is no evidence that timber harvesting has significantly contributed to the decline. Based on the pioneering work of Dexter (1970) and subsequent experience, Dexter and Poynter (2005) have strongly recommended the adoption of an even-aged silvicultural system that is consistent with the regeneration requirements and growth habits of the species. Amongst other things, this will require: Adoption of a patch felling system at first harvest, creating gaps of 5 to 10 hectares. Effective regeneration of the patches as rapidly as possible after harvesting to provide evenaged stands. Continuation of regrowth thinning, and Identification and regeneration of areas of forest which are grossly under production as a result of past management practice. This requirement, however, needs to recognise that cull treatment will be limited by habitat prescriptions. The above recommendations appear to be more aligned with the silvical characteristics of River Red Gum than the single tree selection system that results in uneven-aged forests. Di Stefano (2002) noted that the single tree selection system is the most frequently used system in both NSW and Victoria. NSW has more recently adopted a patch felling system. Based on previous research, concern was expressed by Di Stefano (2002) about the impacts of mature River Red Gum trees on seedling survival and growth rates given the large zone of influence that such trees have on soil moisture and other site factors. The scientific basis for a single tree selection system appears to be seriously lacking. 77

87 Jacobs (1955) and Bren (1991) both noted that on high ground areas (involving only small differences in elevation) deprived of regular flooding, box species in a woodland structure are likely to replace River Red Gum forests. Di Stefano (2002) reviewed the future of the River Red Gum forests and noted that the forests have undergone major change in structure, distribution and growth patterns due to a wide range of factors including river regulation, silvicultural practices (past and present), recreational activities and introductions of exotic animals. He expressed concern with trends to regenerate former River Red Gum sites with other ( replacement ) species that are less dependent on flooding and more tolerant of dry-land conditions. The Forest Management Plans (Mid Murray and Mildura) that cover River Red Gum forests in Victoria recommend group or single tree selection systems, though clearfelling is permitted for salvage purposes (for example, after fire or dieback). These plans also outline the prescriptions that must be followed in terms of habitat tree retention. As previously noted, such prescriptions may be a constraint on broad-area use of alternative systems including a patch cutting system. On the other hand, the authors question the long-term use of selection systems in these forests for the following reasons: 1. There is likely to be a tendency for removal of the best trees, ultimately leading to a forest with a reduced potential for sawlog production. 2. A limitation on future silvicultural options (for example, thinning) due to very small regeneration patches, and 3. The impact of retained overwood on future productivity Effect of River Regulation The normal flooding of the forests and other vegetation communities has been affected in recent years by river regulation by which winter/spring flows are stored for release in summer/autumn. Up to 60% of the annual flow of the Murray River is diverted from its natural flow (Thompson 1992). Such effects reduce water for tree growth and there are subsequent impacts on growth and health such as defoliation from the Gum Leaf Skeletonizer moth. While trees can rapidly recover from a single defoliation, the major factor for control of the moth appears to be winter flooding. Effects on tree growth and regeneration and distribution are being detected. The change in pattern of flooding has been documented by Bren (1988b) (Table 4.10) for the Barmah and Gunbower Island forests. A changed pattern in flooding is evident and if maintained, there will be long term declines in the forest and woodland communities. 78

88 Table 4.10 Flood regimes in the natural and regulated environment (from Bren 1988b). Natural Flow Regulated Flow Regime Regime Barmah Forest Frequency of effective flood 79% 42% Longest number of consecutive years without flood 2 10 Average duration of flood 3 to 6 months 2 to 5 months 84% 39% Gunbower Island Frequency of effective flood Longest number of consecutive years without flood 2 11 Figure 4.15 Low flooding in Red Gum forest. In summary, from the studies by Dexter (1967), Dexter et al. (1986), Bren (1987, 1988a, 1988b, 1991, 1992), Bren and Gibbs (1986), Bren et al. (1987, 2010), river regulation has: Reduced the extent and depth of winter flooding. Reduced the frequency of flooding. Increased the duration of non-flood periods. Increased the occurrence and variability of summer floods. Increased river flow capacity due to de-snagging, and Decreased flow in the total Murray River system below Tocumwal. 79

89 There has been gradual encroachment of box species on to River Red Gum sites, indicating general drying out of former River Red Gum sites. Potential effects have been reported of modified flooding on tree growth (Table 4.11). Lack of flooding has led to declining tree health interacting with parasitism such as dwarf cherry (Exocarpus strictus) (Sinclair 2006). Table 4.11 Estimates of the effects of various flood regimes on production of merchantable timber (merchantable height of 10 m and trees greater than 44 cm diameter) in the Gulpa Island State Forest, NSW (from Bacon et al. 1993). 3 Indicative Forest Productivity (m /ha/yr) Forest in Floodwater Forest Areas up to Total Forest 75 m Distant from Flood No flood for 2 years Flood each winter Flood summer and winter Figure 4.16 Regrowth Red Gum on the Murray River. 80

90 Discussion and Conclusions Introduction The silviculture of Victoria's major forest types of commercial significance has been extensively studied over the last five decades. This has enabled forest managers to develop species-specific silvicultural systems that take account of social (including Occupational Health and Safety considerations), economic and environmental constraints. The focus has now moved to improving these systems to optimise environmental and economic outcomes in particular, in recognition that there may be significant trade-offs between reduced timber production and increased environmental protection over and above that required in Codes of Practice and strategic forest management plans. Such trade-offs (for example, increased overwood retention for habitat purposes versus reduced growth of regenerating stands) need to be quantified and clearly understood by all stakeholders. Obtaining satisfactory eucalypt regeneration following harvesting has been a particular focus of Victorian research, primarily because it is a fundamental requirement of Sustainable Forest Management. In more recent times, information on regeneration success in Victoria's commercially important forest types has become publicly available. Key findings for individual forest types are summarised below Alpine Ash Forests 1. Alpine Ash forests occur in disjunct even-aged stands in Victoria and are an important source of high quality timber suited to a broad spectrum of value added products. 2. The works of Grose, Fagg and Bassett (see Bibliography for individual citations) have collectively provided a strong scientific basis for the silvicultural procedures currently used to obtain successful early establishment and growth of Alpine Ash in Victoria, though this knowledge base has been found from recent studies and operational experience to be incomplete. The research findings are consistent with the ecological requirements of the species. 3. Key silvicultural issues relate to seed management, level of retained overwood (particularly on sheltered aspects), and seedbed conditions that optimise germination rates and the survival of subsequent seedlings. 4. Seed management has been brought into focus as a major issue for Alpine Ash (and also Mountain Ash) forests following three major wildfires in 2003, 2006/07 and The Round Table Meeting in Victoria in 2002 (Lindenmayer et al. 2004) unanimously agreed with improved systems (for example, small patch retention) being explored for montane forests (including Alpine Ash) that may lead to improved biodiversity outcomes. 6. There is convincing evidence that the silvics of Alpine Ash in Tasmania and NSW are very different from those of Victorian Alpine Ash. Extrapolating experience from Bago State Forest in NSW and Tasmania's Central Highlands is therefore not recommended. 81

91 7. The evidence from silvicultural research indicates that clearfell and slash burn systems should continue to be used for harvesting Victorian Alpine Ash at rotation age. However, if monitoring of relevant SFM indicators (once developed) shows that ecological objectives are not being met at the FMU level, then innovative methods (within the silvical limits set by research reported here) that may lead to improved environmental outcomes need to be thoroughly explored by forest managers in co-operation with scientists and other stakeholders. Seed-tree systems for instance have been shown from regular surveys to be effective in regenerating harvested sites where seed-crops have been assessed as being adequate Mountain Ash Forests 1. Mountain Ash forests occupy a prime position of eucalypt forest development in southern Australia. Mountain Ash forms a largely mono-specific tall, open forest canopy, and is fire sensitive (moderate to severe wildfires generally kill all trees irrespective of their age). The 1939 and 1983 wildfires have reshaped the present-day age class distribution of Mountain Ash forests and there will be further effects from more recent fires. 2. Mountain Ash is the most important commercial forest type in Victoria in terms of potential sawlog volume and the quality of sawn timber which is suitable for a very broad range of valueadded products. It is also an excellent pulpwood species, enabling residues from sawlog harvesting operations to be utilised. 3. Recent research (for example, SSP) has tended to focus on evaluating silvicultural systems alternative to clearfelling in order to obtain a better balance between economic and environmental outcomes. Improved clearfell and seed-tree systems (see Campbell 1997b) are the optimal systems on sites where wood production has the normal level of priority. 4. SSP and related research has amongst other things assisted in improving overall forest practices from ecological perspectives (for example, provision of understorey islands for flora conservation). Retention of overstorey trees or use of small gaps to further protect ecological values will be associated with reduced productivity of the regrowth forest. 5. Mountain Ash responds well to thinning, and thinning yields in young regrowth can be relatively 3 high (for example, 280 m /ha from 26-year-old regrowth). Thinnings are valued as a high quality pulpwood source. 6. More recently, an alternative research approach has been taken that builds on the clearfell system through the provision of one or more small clusters of retained overwood on clearfelled coupes (termed a Variable Retention Harvesting System or VRHS). Retention of overwood in this manner can be accomplished within both Occupational Health and Safety constraints and the silvical characteristics of the species. The system also minimises lost production through the impacts of retained overwood on regrowth development. 7. Long-term ecological monitoring of the Variable Retention Harvesting System (VRHS) treatments has recently commenced within an adaptive management framework. It is critical that this monitoring be funded. 8. A significant portion of the Mountain Ash forest type is reserved for conservation purposes, including strategic water supply catchments. 82

92 High Elevation Mixed Species Forests 1. From the 1960s, HEMS forests in east Gippsland have been harvested and regenerated using systems developed for the Ash-type species, while selection harvesting was used extensively up until the early 1980s in the HEMS forests of northeast Victoria. Variable success has been achieved from a regeneration perspective. 2. Intensive research was initiated in east Gippsland in the 1970s to investigate reasons for failed regeneration. 3. Many factors that contribute to less than satisfactory stocking levels have been identified through research and observations by field foresters. They include browsing by native animals, poor seedbed preparation, insufficient attention to monitoring of seed crops on retained seedtrees and seasonal conditions and autumn sowing. 4. Further research is justified in HEMS forests to identify systems that will consistently result in adequate stocking levels Low Elevation Mixed Species Forests 1. The LEMS forests are found in all regions of Victoria except in the northwest of the State and account for nearly 63% of Victoria's State forests. 2. They have been heavily utilised over a long period of time, and have been subject to diseases (for example, Cinnamon fungus and Armillaria Root Rot) and intense wildfires. 3. Selection logging has altered the pre-european species composition of east Gippsland LEMS forests in particular by selectively harvesting the durable species without proper attention to their subsequent regeneration. 4. Research in the 1970s and 1980s along with more recent studies in the Cabbage Tree Creek SSP has led to an improved understanding of the silvics and effective silvicultural systems of the most common species, but more work is required on the other species including Red Ironbark. 5. Retention of overwood not only reduces the development of regrowth, but also influences the species composition of the resultant forest. This finding needs to be taken into account if retained overwood systems are adopted in State forests where ecological issues have higher than normal priority. 6. Improved clearfell and seed-tree systems (see Campbell et al. 1997b) are the optimal systems on sites where wood production has the normal level of priority. 7. Much research has been done on how best to restore forests affected by dieback associated with the Cinnamon fungus. This research has identified sowing or planting as options. Restoration of these dieback forests to enhance their future sawlog potential represents a resource (funding) challenge to forest managers. 83

93 Box-Ironbark Forests 1. The once extensive Box-Ironbark forests are now confined to around 240,000 hectares of State forest (mainly in northcentral Victoria). 2. The current forest structure reflects their past history, including exploitation to support the mining boom that commenced in the 1850s and selection logging for a wide range of forest products where wood durability was an important requirement. 3. Prior to European settlement, the Box-Ironbark forests generally carried fewer and larger trees than the present-day forests which can have stocking levels ranging from 200 to 700 trees/ha. The forests were heavily thinned in the 1930s and 1940s, with the thinnings being used for firewood and charcoal. Large and so-called defective trees were generally felled during these thinning operations. 4. Regeneration of the Box-Ironbark forests relies largely on coppice from cut stumps. 5. In a recent investigation by VEAC, it has been proposed that ecological thinning be now undertaken to in effect. reverse the trend of the cull removal programs by encouraging the development of large trees on areas where habitat values have been diminished River Red Gum Forests 1. The River Red Gum forests that occur largely along the Murray River and its major tributaries are commercially and ecologically valuable. Their distribution, structure and silviculture is directly related to the flooding regime and their long-term sustainability for any selected objective is dependent on appropriate flood management. 2. Despite heavy exploitation over an extended period and markedly altered flooding regimes due to river regulation, the current-day forests (probably resulting from extensive regeneration in the 1880s) retain their capacity to produce a wide range of valuable wood products. The forests also provide a good example of the long-term success of multiple-use management. 3. Altered flood regimes, however, arguably pose the biggest threat to the long-term health and productivity of the forests. There is already strong evidence that reduced flooding has been largely responsible for a marked decline in forest health and lack of regeneration in more recent times. 4. The silvicultural characteristics of River Red Gum have been comprehensively studied. However, in relation to silvicultural systems, there is divided opinion on whether even-aged or uneven-aged systems should be adopted. 5. The present review of Victorian research indicates that, consistent with current NSW practice, patch felling creating gaps of 5 to 10 hectares may be the most appropriate silvicultural system. The gaps should be regenerated as soon as possible after harvesting, with the aim of obtaining dense, even-aged regeneration that can be subject to later thinnings in order to optimise sawlog potential. 84

94 5. ASSESSING THE ECONOMIC VALUE OF SUSTAINABLE FOREST MANAGEMENT 5.1 Introduction Victorian wood production and processing generates direct employment for 19,500 people and directly accounts for a net value of production of $3,034 million per annum (Cameron 2005). The wood production and processing of timber from Victoria s native forests is substantial, accounting for almost 40% of the total employment in this sector and 33% of the net value of production. These estimates exclude the economic contribution from secondary processing industries using timber for making furniture, joinery, flooring, wall frames, and other products, or paper for making envelopes and other paper products. Alternative timber resources from softwood or hardwood plantations, either in Victoria or interstate, are generally unsuitable substitutes for resources currently available from native forests. In addition to sawlogs and pulp-logs and other wood products, native forests also provide additional commercial products, such as honey, and their value could be increased with appropriate management and research (see Section 5.5). 5.2 Native Forest Products and Marketing in Victoria Over the last two decades, there have been significant changes in timber harvesting arrangements on Public land in Victoria. These arrangements, which are outlined in Section 3.3 have largely been implemented although in western Victoria, the transition to new log supply arrangements is still in progress. The Allocation Order issued by the Minister for Environment, dated 29 July 2004 (VGP 2004a), specifies the maximum area of various forest stands in eastern Victoria that will be available to VicForests for harvesting to at least meet existing commitments under managed licences in the 15-year time frame from 1 August 2004 to 31 July The 15-year time frame is divided into three 5-year periods. No estimates are provided in the Allocation Order of the quantity of sawlogs and residual logs that are planned to be produced in each 5-year period. However, the maximum area that can be harvested in each 5-year period is specified (see Table 3.2). In 2007/08, the VicForests contracted volume for ash species totalled approximately 480,000 m3 (VicForests 2009). The grade and species mix of sawlogs in VicForests supply contracts in 2007/08 are shown in Table 5.1. Table 5.1 Sawlog grade and species mix VicForests supply contracts 2007/08. Species Percent of 2007/08 Supply Contracts by Sawlog Grade B Grade C Grade D Grade E Grade All Grades Ash Mixed species All species Source: VicForests (2009) 85

95 The above data indicate that VicForests sawlog supply contracts in 2007/08 were for a total volume of some 676,000 m3 (480,000 m3 ash species and 196,000 m3 mixed species) including some 108,000 m3 of inferior quality sawlogs or E grade logs. In addition to sawlogs, VicForests produced and sold about 1.3 million cubic metres of pulp-logs in 2007/08 to three customers, with about half sold to Paper Australia s Maryvale mill and the balance chipped for export by South East Fibre Exports at Eden and Midway at Geelong (VicForests 2009). In addition to sawlogs and pulp-logs, VicForests also supplies firewood logs to commercial firewood operators when demand or supply requires. VicForests does not supply firewood to domestic users (VicForests 2009). The data in Table 3.2 indicate that VicForests can harvest about 2,050 hectares of ash species and about 3,750 hectares of mixed species per annum in the first five-year period. The contract supply levels are consistent with the expected yields from type and area of forest available for harvesting and with respect to sawlog production. This expectation is consistent with DNRE s Estimates of Sawlog Resource (DNRE 2002b). Following the initial five-year review of the Allocation Order in 2009, the Allocation to VicForests (Amendment) Order 2010 was issued in May 2010 (VGP 2010a) and the allocation to VicForests (Further Amendment) Order 2010 was issued in September 2010 (VGP 2010b). These amendments include the following significant changes to the area allocated to VicForests: The identification of only two stand types (ash, mixed species) whereas 16 stand types were previously identified. A State-based allocation according to forest type with no sub-division into Forest Management Areas. The allocated area is a gross area that has not been adjusted for modelled exclusions based on the provisions of the 2007 Code of Practice for Timber Production and assumptions about the economic viability of harvesting operations, and The exclusion of areas of new and expanded National Parks and conservation reserves in East Gippsland and in the Murray River region. Table 3.3 shows the gross areas available for timber harvesting in Period 2 (1 August 2009 to 31 July 2014), Period 3 (1 August 2014 to 31 July 2019) and Period 4 (1 August 2019 to 31 July 2024) of the Further Amendment Allocation Order. Because the areas allocated to VicForests in Period 2 under the initial and amended Allocation Orders are not comparable due to the significant changes outlined above, it is not possible to draw any meaningful conclusions about the impact of changes in the land use and major fires on the sustainable production of sawlogs, pulp-logs or firewood logs. The major sawmills that are processing sawlogs supplied by VicForests are located at Cann River, Orbost, Nowa Nowa, Bairnsdale, Heyfield, Morwell, Erica, Noojee, Yarra Junction, Dandenong, Swifts Creek Benalla and Corryong. Most of these sawmills have kiln drying and further processing facilities; two thirds of the sawn timber produced in Victoria is kiln dried. There are also a number of smaller sawmills that participate in the competitive market auction system developed by VicForests. Whilst this review only considered published reports up to 1 July 2010, it was brought to the attention of the authors by DPI that an amended Allocation Order was issued in September 2010 that had significant changes to timber harvesting arrangements. These are now included in this review. 86

96 5.3 Products and Markets In order to compete with other wood and non-wood products in domestic and international markets, processors of logs from Victorian native forests need to produce products that have a competitive advantage over alternative timber and non-timber products with respect to price or inherent properties such as strength, hardness, appearance, durability and fibre length and diameter. Industries processing hardwood sawlogs from Victorian native forests have generally targeted markets where hardwood has a competitive advantage due to its superior appearance, strength and hardness and no longer competes with plantation softwood for house framing and other structural applications. In 2007/08 a survey by VicForests of their 18 largest customers, who account for about 95% of sawlog sales, indicated that the sawn timber produced was used evenly between high value appearance products, seasoned structural (F17) and unseasoned structural, and lower value fencing and pallet products (VicForests 2009). The results of the survey according to sawn product are shown in Table 5.2. Table 5.2 Sawn product output for 2007/08 of VicForests largest customers. Category Products % of Total Sawn Timber Produced High value appearance products Furniture and appearance Joinery Flooring 5% 20% 9% Seasoned structural F 17 38% Unseasoned F 8 F 14 Pailing, pallet and other 7% 21% Total 100% Source: VicForests (2009) The range of products that are produced reflect the range of grades of sawlogs processed (Table 5.1). The kiln dried timber is marketed predominantly to customers within the eastern States of Australia, either directly to consumers or to wholesalers, agents and retailers. Many of these customers are small to medium sized businesses manufacturing furniture, various types of joinery (for example, doors and door frames, windows, and kitchen cupboards), structural components (for example, beams, lintels, bearers and joists where the properties of hardwood, such as appearance, hardness, strength or stiffness, provide a competitive advantage), staircases, flooring components, frames and trusses. Although markets for green hardwood have declined due to competition from plantation softwood, it is still in demand for relatively low value products such as pallets, roofing battens, fencing and other structural uses where strength or superior nail holding properties are important. 87

97 Hardwood fibre is preferred to softwood fibre for the production of printing and writing papers as it produces papers with improved surface properties and higher opacity. Light coloured hardwood woodchips from residual logs and sawmill residues from Victorian native forests are used for the manufacture of printing and writing paper in Victoria and overseas. Victorian producers of sawn hardwood timber mainly supply domestic markets. Exports only account for about 5% of the total volume of sawn timber produced. Australia s production of sawn timber and the levels of imports and exports are shown in Table 5.3 in order to provide some perspective on the contribution of production from Victoria s native forests. Table 5.3 Australian production, imports and exports of sawn timber (thousands of cubic metres) for and Source Softwood Hardwood Total Domestic Production Imports Sub-total Exports Apparent Consumption 3,012 3,740 1, ,037 4, ,741 4,274 1,168 1,084 4,908 5, ,670 3,973 1,132 1,030 4,801 5,003 Source: ABARE (2005, 2010) Although Australia s apparent consumption of sawn timber has not markedly increased or decreased in recent years, the trend towards increasing supplies of sawlogs from softwood plantations and decreasing availability of sawlogs from native hardwood forests has continued. Australia s imports of both softwood and hardwood sawn timber appear to be declining, and while exports are increasing, Australia is still a net importer of sawn timber. It is unlikely that the current position will change significantly in the foreseeable future unless there is a sharp reduction in hardwood production from native forests in Australia (mainly Victoria, Tasmania, New South Wales and Western Australia) and a concomitant increase in imports of tropical timbers, often in the form of finished products including furniture. Adams and Attiwill (2005) noted that whilst the Commonwealth Government provides neighbouring developing countries with assistance to prevent unsustainable exploitation of their forests, Australian State Governments increase the pressure for (such) exploitation by reducing supply from Australian native forests. For obvious reasons, that message needs to be effectively communicated to all stakeholders including the broader community {as part of the requirement of Action 13 (enhance community understanding) of the 2009 Timber Industry Strategy}. Domestic production of softwood has increased by almost 24% from 3,012,000 cubic metres in to 3,740,000 cubic metres in while production of hardwood declined by about 4% from 1,026,000 cubic metres to 990,000 cubic metres during the corresponding period (Table 5.3). While the overall result was a modest increase in the apparent consumption of sawn timber, imports of sawn timber have declined from 18% to 8% of domestic consumption. 88

98 Provided there is no significant decline in apparent consumption, Australia is likely to remain a net importer of both softwood and hardwood sawn timber for at least the next decade. Progressive increases in softwood sawlog availability are expected to be modest and there remains considerable uncertainty about the potential of existing hardwood plantations to economically produce suitable sawlogs (Ferguson et al. 2003). Most hardwood plantations in Australia are presently being managed on pulpwood rather than sawlog regimes. The economic value of the timber industry based on native forests in Victoria is significant. Cameron (2005) estimated that the direct plus indirect impacts of the industry are in the order of 15,500 jobs and $2.3 billion in the value of output. Significantly, the industry is largely regionally-based where unemployment levels are often significant. Indicator 6.1a for both the national (MPIG 2008) and Victorian (DSE 2007b) C&Is essentially addresses the annual production and dollar value of wood products (mainly comprising sawlogs and pulpwood). For the period 2001/02 to 2005/06 (DSE 2007b), the value of wood production from native State forests increased by 6% from $137 million per annum to $147 million per annum despite a reduction in sustainable harvest levels. The value of log production was equivalent to 0.1% of the Gross State (Victoria) Product. Firewood has long been an important source of fuel for both domestic and industrial (particularly in the post-wwii era) uses in Victoria. It is estimated (DSE 2010b) that Victorian s currently consume approximately 600,000 cubic metres of firewood annually, with around 13% of this being sourced from Public land. In recognition for the need to ensure a balance between community needs for firewood and environmental values, a Firewood Strategy was recently finalized (DSE 2010b). The Strategy addresses inter alia options for making firewood available for householders, sustainability issues, health and safety imperatives, cultural heritage values, risk of accidental fire ignition during collection of firewood, and programs for improving household energy efficiency Non-Wood Forest Products Introduction Forests produce a wide range of timber and non-timber products that have many uses and values. The latter can be as simple such as for the provision of shade and shelter. With some notable exceptions, little attention has been given to the importance of non-wood forest products (NWFPs) in Victoria and even less to their sustainable management. This contrasts with many overseas forests where NWFPs can be equally or more important than wood products and where a conscious effort is made to better manage the level of harvesting of such products through joint forest management or similar arrangements. In the brief treatment of NWFPs below, users such as fir recreation and indigenous traditional uses are regarded as forest uses rather than NWFPs. Grazing, however, is included because the forests are providing an identified product in the form of fodder for the grazing animals. Use of forests as carbon sinks is not considered here, but is addressed in Section

99 5.4.2 Historical Considerations In Victoria, the three main NWFPs are considered in this review to be water, honey and grazing. The impacts of forest operations on water flows are covered in detail in Section 8.3. Honey production in Victoria was estimated to be worth $8.3 million in 2000 (Table 5.4). Honey has long been recognised as an important NWFP from Victoria's forests, woodlands, shrublands, agricultural crops and native grasslands. It is estimated that over 80% of honey production in the State relies on native flora (BRS 2003). Table 5.4 Honey production in the year 2000 in Victoria. a No. of Beekeepers 174 No. of Productive hives 57,000 No. of Unproductive hives 6,000 Total number of bee hives 63,000 Beeswax (tonnes) 76 Honey (tonnes) 4,971 Gross Value ($) $8.3 million Public land used in the last five years (%) 48% Source: BRS (2003) a Each beekeeper may have multiple apiary sites The role that native State forests play in the commercial beekeeping industry has recently been reviewed (Dooley 2004). The focus of the review was on the four major beekeeping regions in Victorian northeast forests and the Bendigo, Mildura and East Gippsland FMAs. Forest managers and beekeepers in these regions were interviewed to identify critical requirements and management issues. At the time of that review (2001), Victoria was producing about 5,100 tonnes of honey per annum with a value of around $10 million, equivalent to about 16 % of Australia's production of honey. Nearly 80% of the nectar is collected by bees from eucalypts, with the balance from coastal heath, weeds and agricultural crops. In 2000/01, the Department of Sustainability and Environment was administering around 650 permanent apiary sites and a further 2,000 temporary sites. Table 5.5 provides more detail on permanent and temporary apiary sites for the four regions. 90

100 Table 5.5 Location of licenced apiary sites in some forested areas in Victoria in 2001/02. Area Permanent Apiary Sites Temporary Apiary Sites (1.6 km radius) Northeast forests Bendigo FMA Mildura FMA East Gippsland Source: Dooley (2004) Grazing within forests has a long history in Victoria. Ferguson (1965) noted that grazing was traditionally largely focused on the northern lowland box and River Red Gum forests together with the alpine areas and was strictly controlled through annual license or agistment arrangements. Fire was used by graziers to provide better fodder on the forest floor. Grazing, however, has become an emotive issue in more recent times, and grazing management is given special attention in relevant FMPs (BRS 2003). The FMPs provide guidelines to ensure that the cultural and biodiversity impacts of grazing together with the introduction of pest plants are minimised. Whilst these guidelines appear to be soundly based, the sustainability concept requires that monitoring be undertaken and results evaluated to ensure that management objectives are being met. Other NWFPs harvested under strictly controlled conditions in Victoria's public native forests include gravel, sand and other rock materials, mined ore, salt, Eucalyptus oil, seed, tea-tree (for fencing), tree ferns (for landscaping), mushrooms, and game (for example, deer). Of these, gravel extraction has caused significant problems in the past. Past Codes of Forest Practice have provided guidelines for dedicated quarries to ensure that extraction is planned and designed in a manner that minimises soil erosion and adverse water quality and visual impacts. Rehabilitation of quarries at the end of their working life is mandatory Recent Developments Consistent with the Montreal Process, the Victorian 2008 State of the Forests Report has a dedicated section on NWFPs. However, as for other States and Territories (see MPIG 2008), data on the production and value of NWFPs and services remain generally poor in Victoria (DSE 2009a). The report notes that statistics on NWFPs are often difficult to collect due to the often large variability in annual production or in annual supply. As an example, seedcrops for seed collection can vary substantially between seasons due to the silvicultural characteristics of individual species and the impacts of seasonal weather conditions (especially drought). Wildfire can impact significantly on NWFPs. For example, honey production and water production are two key NWFPs that can be influenced by landscape-scale, high intensity wildfires. Any adverse impacts of such fires need to be recognised in the allocation of human and financial resources for fire management. The contribution of NWFPs could become increasingly important to local and regional economies in the future (for example, the discovery of native plants with medicinal properties). The management of NWFPs deserves increased attention. 91

101 5.5 Discussion and Conclusions 1. The native forest timber industry in Victoria is significant in terms of direct and indirect employment, and is estimated to have a mill door value of around $500 million per annum from initial processing of logs from eastern Victoria. Although markets for green hardwood timber have declined, this is not the case for kiln dried hardwoods which are used for a wide range of value-added purposes including furniture, joinery and flooring and for structural purposes. Hardwood mills in Victoria supply timber to both national and international markets. 2. The timber industry based on native forests in Victoria is significant in terms of direct and indirect regional employment and value of output. Australia continues to be a net importer of hardwood sawn timber, while exports account for only a small proportion of total production; 3. Processors of sawlogs from native forests have continued to produce products that have a competitive advantage by making best use of the inherent properties of hardwood timbers. This swan timber is marketed to a wide range of customers, many of whom are small to medium-sized businesses manufacturing value-added products such as joinery, structural components and flooring; and 4. It is possible that NWFPs from Victoria s native forests will gain increasing prominence in the foreseeable future, irrespective of wood production considerations. In the Victorian context, water and honey production are key NWFPs, and these can be impacted by landscape-scale wildfires. NWFPs deserve increased attention from a SFM perspective. 5. Current timber allocations are carried out on an area rather than volume basis. It can be argued that this does not sit comfortably with sustainability principles. 92

102 6. MANAGEMENT OF FIRE, PESTS AND DISEASES IN FORESTS 6.1 Introduction Wildfire arguably poses the greatest immediate and on-going threat in terms of stand damage and ecological impacts to Victoria s 7.9 million hectares of native forest but in the longer term, the potential for pests and/or diseases to cause serious and widespread damage cannot be under-estimated. The number of overseas visitors and container ships entering Australia continues to escalate and so too does the risk of accidental introductions of damaging agents despite Australia's strict and wellenforced quarantine laws. Another factor to consider is the potential impacts of any changes in climatic conditions which could both not only lead to more extensive and intensive wildfires but also to more favourable conditions for indigenous and exotic pests and diseases to flourish. This Section of the review addresses the threats from wildfire, and pests and diseases, to Victoria s commercial forest types and how such threats are best managed. In relation to fire, particular attention is given to the practice of prescribed fire for fuel management purposes (that is, fuel reduction burning or FRB). In 1995, the Department of Natural Resources and Environment released a comprehensive Code of Practice for Fire Management on Public Land, and this was re-issued in 2001 (DNRE 2001). The Code provided a framework for the integrated management of all fire and fire-related activities to ensure that the use of prescribed fire (slash burning, FRB and ecological burning) is consistent with sound environmental principles. A specific issue addressed by the Code is the use of fire retardants to assist suppression activities. The Code was further revised in February 2006 (DSE 2006c). This latest Code inter alia takes account of new scientific information on fire behaviour, fire ecology and improved processes for community engagement. The Code also addresses the use of prescribed fire to achieve specified ecological objectives, noting that the perpetuation of many floristic communities is fire-dependant. Prescribed fire has many potential ecological applications for the future including for example, accelerating tree hollow formation for many species of arboreal mammals and birds in eucalypt forests (Adkins 2006) and maintaining forest health (for example, Jurskis 2005a, b). 6.2 Wildfires Victoria is widely regarded as a fire-prone environment due primarily to dry and windy summer conditions, frequent droughts and ignition sources (particularly by lightning), and flammable forests. Flammable can be defined as the ease with which a substance is set on fire (DSE 2006c). New insights into the flammability of Australian forests, including forests dominated by eucalypt species, are provided by Gill and Zylstra (2005). On the basis of rates of fire spread, fire intensity and fire residence times, many eucalypt forests can be regarded as being flammable. It was noted, however, that the future flammability of forests could be impacted by climate change, social attitudes, rates of ignition and changing fuel types. Two important aspects of wildfire impacts in commercial forests are damage to trees/stands and environmental effects. 93

103 6.2.1 Damage to Trees and Stands Wildfires, by their very nature, usually show considerable variation in fire intensity across the burnt area (for example, see Wareing and Flinn 2003 and Flinn et al. 2008). The level of damage sustained by commercial forests in Victoria depends largely on the interaction between fire intensity and the sensitivity of individual forest types to fire. In fire-sensitive Mountain Ash and Alpine Ash forests, moderate to intense wildfire invariably kills the stands, whereas they often survive mild ground fires. Low Elevation Mixed Species (LEMS) forests on the other hand usually survive even intense wildfires through their ability to produce epicormic shoots. However, wood quality in these forests may be downgraded following intense wildfire through the formation of kino (gum) veins and the development of epicormic shoots. Where the Ash forests have reached a commercial size, the fire-killed stands can be salvage-logged as occurred after the 1939 fire and to a lesser extent, the 2003, 2006/07 and 2009 fires. Providing the stands carry a good seed crop at the time of the fire, adequate natural regeneration will usually follow. However, young regrowth stands (of logging or wildfire origin) that have not reached seed-bearing age (usually around 20 years) will not regenerate following an intense wildfire without expensive human intervention (for example, sowing of seed or hand planting of seedlings). Figure 6.1 Alpine Ash burnt in 2003 wildfire at Gibb Range (DSE 2008c). Whilst there have been some excellent accounts of past wildfires in Victorian forests in terms of their extent, loss of life, and property damage (Cheney 1976, Dexter et al. 1977, Rawson et al. 1983, Wareing and Flinn 2003, Flinn et al. 2008), timber losses have rarely been detailed except in Annual Reports of the former Victorian Forests Commission. Whilst not estimating timber losses, Wareing and Flinn (2003) did provide tabulated data for the 2003 Alpine fires showing that 159,000 hectares of Ashtype forests were burnt, with 85,600 hectares of this gross area being State forest. A high proportion of the burnt area suffered moderate to severe crown scorch. The burnt area of mixed species forests (including LEMS forests) was estimated to be 756,000 hectares across all land tenures including 94

104 National Parks and State forest. Over 50% of this area was moderately to severely burnt (Wareing and 3 Flinn 2003). Assuming a mid-range mean annual increment for Ash-type forests of 11 m /ha/yr (see Table 7.1), an average age of the fire-killed stands of 30 years (a combination of regrowth from logging over the past three decades and wildfire up to and including 1939), and 50,000 hectares of the 85,000 hectares being available for logging following exclusions by prescription, the volume lost could be as high as 17 million cubic metres. In the case of the 2006/07 Great Divide fires, nearly 100,000 hectares of ash-type forests were burnt, of which around 75% suffered either total, severe or moderate crown scorch (Flinn et al. 2008) Community Issues and Government Inquiries The 2003 Alpine fires invoked heated community debate surrounding the failure of relevant agencies [primarily DSE and Parks Victoria (PV)] to contain the fires at an early stage before they developed into larger fires, including the Alpine fire that burnt about 5% of Victoria. The criticism led to an investigation of fire prevention and preparedness by the Victorian Auditor General and a Government inquiry (Esplin et al. 2003) which received submissions and evidence from interested parties. One such party was Forest Fire Victoria (FFV) who subsequently published a comprehensive report entitled The facts behind the fire (Dexter and Hodgson 2005). This report was highly critical of the containment strategy adopted for the multiple outbreaks that occurred on 7 January Their major criticisms centred on the failure to undertake first attack on the initial outbreaks with sufficient priority and resources, a lack of resources with fire experience and local knowledge, and changes in natural resource management policies and practices (particularly, closure of tracks and a reduced commitment to FRB). The 2006/07 and 2009 fires reinforced community concern about fire management policies and practice in Victoria, and this led to further Government Inquiries. The outcomes of Federal and State inquiries into major wildfires that occurred in south-east Australia during the 2002/03 bushfire season, were reviewed by Kanowski et al. (2005). They found that in broad terms, the four inquiries, including the Victorian Bushfire Inquiry (Esplin et al. 2003), reached similar conclusions. Major issues included lack of knowledge on a range of fire management issues, including the role of landscape-scale FRB as an effective fire prevention measure. In response to the 2003 Alpine fires that burnt around 5% of Victoria s land area, including large tracts of valuable timber resources, a Group known as the Stretton Group was formed which was committed to ensuring that the public debate is conducted on a balanced and informed (factual and scientificallydefendable) basis. The subsequent fires in 2006/07 and 2009 strengthened the resolve of the Stretton Group, which has held a series of forums and seminars, proceedings of which have been published (for example, Fire-Flood-Mud-Water by R Gilder and Prof D Dunkerly, May 2008; Fire in Water Catchments by Prof M Adams, November 2008; and The Catastrophe Australia had to have which Crippled Victoria by R Underwood, March 2009). These presentations drew on all relevant science and have made an important contribution to the public and scientific debate on fire management in Victoria and beyond. In a similar response, so-called The People s Review (the Review) was initiated with a focus on both the 2003 and 2006/07 fires (Attiwill et al. 2009). In contrast to preceding inquiries, the Review claims that it focused on obtaining the views of people with the experience and knowledge on the full range of fire-related issues in an open and transparent manner. An overwhelming finding of the Review was that future fire protection (prevention and suppression) measures must not be driven by political or ecological ideologies. The Review stresses that fire prevention and readiness, including extensive FRB, road and track maintenance and appropriate first attack capability are critical to a credible fire management strategy for Victoria. Some of these issues were also canvassed by Wareing and Flinn (2003) and Flinn et al. (2008). The Review also strongly agreed with most of the 95

105 recommendations of a Parliamentary Inquiry into the impact of public land management practices in Victoria (Parliament of Victoria 2008). The Parliamentary Inquiry, which adopted an exhaustive in-depth process involving scientists, professionals and field practitioners, had a focus on aspects of prescribed burning, biodiversity and floods, bushfire suppression infrastructure, impact of traditional land uses, community and stakeholder engagement, and impacts of climate change. The People s Review, however, was critical of the Government s response to the recommendations of the Inquiry, with 13 of the 20 high-level recommendations being supported only in principle. Included among these, was a recommendation for DSE to increase its annual prescribed burning target from 130,000 hectares to 385,000 hectares. The Review questioned the precise meaning and basis for the term support in principle, given firstly the exhaustive process of the Parliamentary Inquiry and secondly the existence of what they regarded as overwhelming scientific and operational evidence to support most if not all of the 20 recommendations. Finally, the Victorian Government announced a Royal Commission in mid-february 2009 into the catastrophic 2009 Victorian Bushfires (Black Saturday fires) that resulted in the loss of 173 lives and extensive damage to plantation and natural forests, to private assets and to the environment in general. The Institute of Foresters of Australia dedicated a special edition of their Journal to reproduce in full the Institute s comprehensive submission to the Royal Commission (IFA 2009). The Commission is expected to hand down its final report in mid Flinn et al. (2008) stressed the dangers of having extensive (landscape-scale), long-duration fires burning over the summer period, particularly because of the increased probability of experiencing catastrophic fire weather. This in fact happened when three significant fire runs occurred about three weeks after the ignition of the 2003 Alpine fires. These fire runs were arguably of a similar intensity to the Black Saturday fires. The distinguishing feature of the 2003 fire runs, however, was that they were largely confined to remote forested country. If fire weather similar to that of Black Saturday had been encountered at the peak of the 69-day duration Great Divide fires that ultimately burnt over 1 million hectares, many rural and semi-rural communities could have been simultaneously and seriously impacted Environmental Impacts A number of detailed ecological studies on wildfire impacts have been undertaken, some broadly-based and others highly specific (for example, the study by Neumann (1992) of the impacts on ants). Ecological studies show that wildfire impacts are highly variable depending on a range of factors 1 including fire intensity and forest type in particular. Some vegetation communities actually benefit from periodic fire. This review, however, confines itself to a brief consideration of the impacts of wildfire on soil erosion in forests. Two local studies have been conducted on the subject. In a 35-hectare sub-catchment near Warburton severely burnt during the 1983 Ash Wednesday wildfires, Leitch et al. (1983) studied massive sheet erosion associated with an intense but short duration storm event six days after the fire. They estimated that in the order of 800 tonnes (22.9 t/ha) of ash plus loose soil containing 2,900 kilograms nitrogen (82.9 kg/ha) and 220 kilograms phosphorus (6.3 kg/ha) were washed from this small sub-catchment, all of which was directly deposited into the Yarra 1 Fire intensity is commonly defined in terms of the rate of heat output per length of fire-line expressed as kilowatts per metre (kw/m) and classified as low (less than 500 kw/m), moderate (500-3,000 kw/m), high (3,0007,000 kw/m) or very high (7,000-70,000 kw/m) (DSE 2004a). 96

106 River. It was observed that the soils remained hydrophobic for at least three months after this Ash Wednesday fire so that they remained in a high erosion state. At two hydrologic research catchments in northeast Victoria that were burnt in the 2003 Alpine fire, Lane et al. (2004b) found that sediment exports increased by factors of 8 to 9 in the first year following the fire, diminishing to 2- to 6-fold increases in the second year. From a strictly commercial forestry perspective, the important question is the impact of erosion losses (including nutrients) on the productive capacity of the affected area. This issue is considered further in Section 8.2. A key conclusion to be drawn on wildfire is that the impacts are long lasting. The length of time that the impact is discernible depends on the forest type and the ecosystem component being studied (for example, water flows and timber yields). The 1939 fires transformed large tracts of old-growth forest to regenerating forest, the impacts of which may be considered in terms of decades or centuries. If the 2003 and 2006/07 scenarios in particular are fully repeated in the Victorian Mountain Ash water supply catchments, the impacts on water quality and yield will be catastrophic for Melbourne (see also Adams and Attiwill 2005). The.question arising from such an outcome is whether the effects of more frequent wildfires are cumulative and what the effects are compared with other activities (such as thinning, for example). Few studies are available on the long-term frequency of wildfires. Woodgate et al. (1994) reported on a tree-ring analysis and regeneration study from a single compartment of a Silvertop Ash (E. sieberi) forest in east Gippsland, the objective being to determine the growth stages of trees and assist in defining Old-Growth forests. Their combined data indicated an average firefree period of 22.6 years over the last 300 years. Figure 6.2 Following the 2003 fires, the Macalister Valley was rapidly affected by floodwater and debris and massive loss of topsoil (Photo Aircraftman Warren Derment, RAAF Sale, Victoria). 97

107 6.2.4 Use of Fire Retardants There are two types of fire retardants. Short-term retardants (fire suppressant foams) contain wetting agents which enhance the extinguishing ability of water by reducing its evaporation and increasing its retention on fuel surfaces. They are less contentious than long-term retardants which are the focus of this review. Long-term retardants have chemicals that retard combustion, and have traditionally been used to slow the spread of fires caused by lightning in inaccessible terrain, allowing ground crews time to reach the outbreaks and more safely work on the strikes before the fire rapidly develops. Their use has been broadened over the past two decades (see Waring and Flinn 2003), and has resulted in some community concern about the use of long-term retardants. This concern was highlighted by the 1998 Caledonia fire that burnt 10,000 hectares of the Carey State forest and 22,000 hectares of the Alpine National Park which are located in two major water supply catchment areas. A fire retardant known as Phos-Chek D75-R was used on parts of the fire, particularly the steeper country, in order to assist ground crews. On 12 January 1998, a localised storm event occurred, and consistent with the findings of Leitch et al. (1983), it washed large amounts of ash and soil into the Caledonia and Macalister Rivers. Concern was expressed that the fire retardant would have also been washed into water supplies posing a threat to human health. A subsequent CSIRO-led investigation was undertaken into the effectiveness of fire retardants and environmental risks associated with their use (DNRE 2000), and it was found that the fire retardant used in Victoria (Phos-Chek D75-R) is effective if properly applied in front of low intensity fires. However, fires of higher intensity are likely, through the process of spotting, to breach fire retardant barriers like those depicted by Wareing and Flinn (2003). Figure 6.3 Aerial application of fire retardant in Extensive use of retardant in this manner (see Figure 6.2) may be in conflict with the underlying intent of SFM indicators relating to aquatic values and accumulation of persistent toxic substances related to public land. BRS (2003) states that fire retardants contain high concentrations of nutrients - nitrogen, phosphorus and sulphur as well as very small amounts of performance additives, while the Victorian 2003 State of the Forests Report (DSE 2005a) states that the retardant used in Victoria contains chemicals generally found in a broad range of agricultural fertilizers. These statements fail to recognise the potential toxicity of retardants as raised by CSIRO. CSIRO concluded that whilst the retardant used in Victoria mainly comprises ammonium salts of sulphate and phosphates, it also has some additive compounds that are particularly harmful to biological systems and human health. 98

108 Furthermore, they noted that little is known about the impacts of these additives in natural environments including their longevity or transport to aquatic systems. Clearly, retardants need to be used strategically and sensitively until further research is undertaken into their potential toxicity. Another concern with the use of fire retardants is their ecological impacts, particularly in floristic communities that may be sensitive to the application of nitrogen and phosphorus contained in the retardant. Given that invertebrates have a close association with the structure and composition of vegetation, Collett and Schoenborn (2005) undertook a study of the effects of fire retardant on surface-dwelling invertebrate communities in heathlands at two locations in Victoria. Using pitfall trap sampling over a 12-month period, more than 136,000 invertebrates were collected for taxonomic analysis after fire retardant treatment at three rates. It was concluded that any change observed in general insect diversity, taxon richness and community evenness at the ordinal and sub-ordinal taxonomic levels were due to site-related environmental factors rather than application of retardant. On balance, targeted use of fire retardants is important as part of an overall fire suppression strategy. Extensive use of fire retardant, however, remains an issue due to lack of short and long term toxological studies on human health and on terrestrial and aquatic ecosystems Fuel Reduction Burning Introduction Broadscale prescribed burning for fuel hazard reduction or fuel reduction burning (FRB) as it is more commonly known (other terms being burning off and cool burning ), involves the deliberate lighting of fires to burn a nominated area under pre-determined conditions (McCarthy and Tolhurst 2001). The aim is to reduce forest fuels so they are no longer available to be burnt by wildfires, and also to reduce the intensity and extent of wildfires so that they can be more readily and safely controlled. The main consideration in this section is to review the effectiveness of FRB and to examine impacts on ecological values. However, as previously noted, many of the forests have evolved with thousands of years of regular burning and whereas in a study situation, burning is considered to be the treatment and the unburnt situation is the control, in fact in many forests the control (or normal situation) should arguably be the regularly burnt situation and the unburnt situation is a treatment Knowledge of Fuel Accumulation and Fire Behaviour The theory and knowledge behind FRB and other prescribed fires have been thoroughly reviewed by Tolhurst and Cheney (1999). They concluded that broad area prescribed burning (FRB), changes a number of forest characteristics that will result in reduced rate of spread, reduced spotting, reduced flame heights, reduced fire intensity and increased ease of suppression for a period of time following the burn. These conclusions, based on underpinning science, provide a compelling reason for an appropriate annual FRB program. 99

109 Assessing the fuel hazard is clearly a vital aspect of fire management. There have been significant changes in how this has been assessed over the past decade or longer due to the work of Wilson (1992), Cheney et al. (1992) and McCarthy et al. (1999). Traditionally, fuel hazard was gauged by the amount of accumulated litter expressed in tonnes per hectare, however, it was subsequently shown that such an approach does not indicate the overall fuel hazard of a forest (for example, McCarthy et al. 1999). They defined overall fuel hazard as the sum of the influences of bark hazard plus elevated fuel hazard plus surface fine fuel hazard so that the entire fuel complex is considered. A guide produced by McCarthy et al. (1999) provides detailed information on assessing the overall fuel hazard. Similar guides were developed from research by Cheney et al. (1992) for coastal forests, namely Silvertop Ash in NSW, and this has relevance for east Gippsland. Whilst the importance of considering the overall fuel hazard cannot be understated, the surface fine fuel hazard along with coarse woody debris remain important from ecological, nutrient and burn-out time perspectives (see Tolhurst and Cheney 1999). Fuel accumulation or fuel dynamics has been reasonably well researched (for example, Flinn et al. 1983, Simmons and Adams 1986, Tolhurst and Kelly 2003). In some wet forests, fine fuels can continue to accumulate for 30 or more years before reaching a so-called steady or equilibrium state (Tolhurst and Cheney 1999) while in other forests, a steady state can be reached in a few years. One of the most comprehensive studies on fuel dynamics in Victoria was conducted in a mixed eucalypt foothill forest by Tolhurst and Kelly (2003). They noted that litter containing dead plant parts is consumed in the flaming zone of a fire and hence it is an important component in determining fire rates of spread and flame heights. The average equilibrium level for surface fine fuels in this forest type was found to be around 16 tonnes per hectare, but this fluctuated from 8 to 26 tonnes per hectare depending on seasonal conditions. It was concluded that such variation means that models of fuel dynamics need to be used with considerable caution. Whilst the impact of a low intensity fire can last for over a decade, Tolhurst and Kelly (2003) showed, however, that surface litter rapidly accumulates after a fire and can reach around 90% of its steady state within four years of a spring or autumn burn so that the impacts on surface fuels were relatively short-lived in these mixed species forests. Bridges (2004, 2005) reported on intensive sampling of fine fuels in the dry sclerophyll forests of southeastern NSW over a period of 16 years. Sampling was undertaken in six of the years during this period. In the multi-aged stands that had not been recently disturbed, the fine fuel load averaged 10.6 tonnes per hectare for the six years sampled. Consistent with the findings of Tolhurst and Kelly (2003), however, Bridges (2005) found that between years, the loads ranged from 8.9 to 16.6 tonnes per hectare. He concluded that fire management benefits could be achievable at the landscape level by FRB at 5 to 8 year intervals Effectiveness of Fuel Reduction Burning in Wildfire Control Apart from the above theoretical understandings and research studies, there are numerous experienced firefighters that can testify to the effectiveness of FRB in reducing rates of spread and fire intensity and in assisting with safer fire suppression activities. Despite this, however, sections of the broader community continue to express concerns about FRB from either an ecological or an effectiveness perspective. Lack of well-documented case histories was a contributing factor to such concerns in the 1970s, but since that time, a number of studies have been reported on the subject. Three of these studies have been selected to illustrate the practical benefits of FRB in assisting fire suppression. 100

110 Study 1. Billing (1981) provided brief case histories for several forest-related fires including one at Daylesford and three in the Grampians. At Daylesford, a fire (known as the Barkstead fire) that ignited in logging slash, moved rapidly with the assistance of spotting, towards the Barkstead settlement. However, the 1.2 kilometre leading fire edge reached a fuel-reduced area late in the afternoon and this stopped the forward spread and no suppression action was required. The fuel-reduced area was also wide enough to catch all spotting activity. In the first Grampians fire, there was a rapidly spreading wildfire threatening the Mt Difficult pine plantation until it reached a recently fuel-reduced area and consequently, no further fire suppression activity was required on that sector of the fire. In the second Grampians fire, a fuel-reduced area prevented a fast moving fire from spreading into dense heath and swamps (where fire control is difficult) despite significant spotting. The third Grampians fire, a rapidly spreading fire in non-forested land, met a fuel-reduced strip which did not stop the fire but caused the head to become fragmented and reduced the rates of spread and fire intensity. This allowed suppression activities that prevented the fire burning into inaccessible forest terrain. Study 2. Rawson et al. (1985) reported on case studies of ten wildfires. Four of these were the same fires as reported by Billing (1981) while two (at Lorne and Mt Macedon) were covered in the report on the 1983 Ash Wednesday fires (Rawson et al. 1983). They found that for the Lorne fire, FRB around settlements assisted firefighting efforts under the extreme fire conditions that were characteristic of the Ash Wednesday blazes. Rawson et al. (1985) also presented graphic evidence to show the effectiveness of FRB which enabled Braemar College near Mt Macedon to be saved. It is one of the best illustrated examples of the important role that FRB plays in the protection of assets. Study 3. McCarthy and Tolhurst (2001) undertook a very exhaustive retrospective analysis of the general effectiveness of broadscale FRB by sampling 116 wildfires of variable size across Victoria where fuel-reduction had either assisted or failed to assist fire suppression activities (that is, they analysed whether or not the FRBs were helpful ). Using a modeling approach, they found that the suppression benefits of FRB are reduced as fire danger levels increase so that weather rather than forest conditions becomes the dominant factor in terms of suppression. A second model predicted that the benefits of broadscale FRB in assisting fire suppression are highest in the first four years after burning and become minimal by around ten years post-burning. Another key finding from this study was that, on average, only 1 in 20 of the wildfires in Fuel Management Zone (FMZ) 3 (that is, broad area fuel reduced mosaic, as defined in DNRE 2001) burnt into what they termed helpful fuel-reduced areas. This indicates that either more extensive broadscale FRB is needed or that FMZ-3 areas need to be burnt more frequently (particularly to prevent high to extreme fuel hazards being reached). This finding has clear implications for commercial forestry in Victoria. 101

111 6.3.4 Ecological Impacts of Fuel Reduction Burning A full review of ecological impacts of prescribed burning was limited by the initial review Brief. This section therefore only considers FRB, not prescribed fire for seedbed preparation and ecological burning for which guidelines were recently issued. Victoria was one of the first States to recognise the need for a well-designed, long-term integrated experiment to investigate season and frequency of burning on a wide range of ecological values. In that study, five areas were selected in the Wombat State forest to investigate the effects of burning season and frequency (Tolhurst and Flinn 1992, Tolhurst et al. 1992, DSE 2003a). Spring-burning treatments were first applied in 1985 while autumn-burns were first undertaken in Various ecological components of this study were reported in invertebrates by Collett and Neumann (2003), bird abundance by Loyn et al. (2003), reptile populations by Irvin et al. (2003a), insectivorous bat populations by Irvin et al. (2003b), structure and composition of understorey vegetation by Tolhurst (2003), terrestrial mammal populations by Irvin et al. (2003c), tree growth and bark thickness by Chatto et al. (2003), soil nutrients by Hopmans (2003). Summary findings of the reports from the study in Wombat State Forest (DSE 2003a) are: Surface fuels in this forest have an average steady state (dynamic equilibrium) of 16 tonnes per hectare with a range between 8 and 26 tonnes per hectare. Surface fine fuel loadings re-accumulate to within 90 % of the long-term unburnt state within four years, irrespective of season of burning. However, kindred studies show that other components of the fuel complex may be modified for 15 to 25 years. Bark loss due to burning depended on the frequency and season of burning, with the loss being greater for autumn burns. Tree mortality was largely unaffected by any of the treatments. A significant decline in levels of carbon and nitrogen in surface soils was detected in those treatments with repeated fires at 3-yearly intervals, and there was some evidence of a change in the quality of organic matter. Such impacts were not evident in less frequently burnt treatments. Over a 16-year period, no plant species were either lost or gained in any of the treatments including four successive spring burns. However, relative abundance of some species was either favoured or decreased depending on season and frequency of burn. For example, short-rotation spring burning favoured Austral Bracken (Pteridium esculentum), herbs and geophytes and disadvantaged Forest Wire-grass (Tetrarrhena juncea), rushes, legumes and climbers. Three low intensity prescribed fires within eight years had minimal impact on litter arthropods, though impacts of such fire regimes on the abundance of the Coleoptera and Diptera were confounded by significant changes in their activity levels in the control treatments. None of the reptile species studied was affected by the burning treatments. Unburnt microhabitats such as logs and deep litter layers along with unburnt patches provide important refuges. 102

112 The two terrestrial mammals studied, Brown Antechinus (Antechinus agilis) and Bush Rat (Rattus fuscipes), were not favoured by any of the burning treatments, but habitat preferences were observed and the survival and recovery of both species depended strongly on unburnt patches that are characteristic of low intensity, mosaic FRB. Populations of the Bush Rat took three breeding seasons to recover when greater than 50 % of its preferred habitat was burnt in a spring fire, but no such recovery occurred when all of the habitat was burnt, and Season of burning had similar effects on bird abundance. However, any beneficial effects of fire are more likely obtained with autumn rather than spring burns, probably because birds are nesting in spring. The changes in fuel quantities, fuel structure and vegetation structure have been studied in a number of burning trials (K. Tolhurst pers. comm.). In addition to gathering quantitative data, a series of fixed point photographs were taken in the series in the Eaton Block (Figures 6.4 to 6.12). They cover a period of 35 years commencing shortly after the first fire in 1977, and demonstrate the value of visual records in such studies. In terms of impacts of FRB on individual components of the ecosystem studied, it was found (DSE 2003a) that results from short-term studies can be misleading (for example, they may overlook underlying trends) and that study periods should exceed ten or more years to gain more reliable impact data. Recognising the limitations of the Wombat State forest study (for example, the size of the burnt areas was only 10 to 15 hectares), it was concluded that frequent burning (such as every three years) over a lengthy period will most likely result in irreversible changes to the structure (of the forest), fertility (of the soils) and relative abundance of fauna and flora species. Such FRB strategies should be avoided, and mosaic burning is strongly encouraged within the pre-determined area to be burnt. In relation to the season of burning, the results indicated that a combination of spring and autumn burning should be used to optimise outcomes for biological diversity at both the local and broad area forest levels. Procedures to minimise ecological impacts associated with these and other factors are provided in detail in the summary report (DSE 2003a). The Wombat State forest study shows the strength and value of long-term ecological monitoring of purpose-designed studies. Further guidance on optimal burning regimes will be obtained from an investigation of the interactions between harvesting and fire in the forests of south-eastern NSW (Binns and Bridges 2003). The study addresses stand structure and ecological impacts (including soils) at a compartment level. Six treatments (interaction between two harvesting and three burning regimes) are being evaluated in a study with three replicates. It is a significant long term study directly applicable to Victorian LEMS forests. The variability of the intensity and extent of fire is critically important. Some early research and retrospective wildfire studies have selected uniformly burnt areas for study whereas in practice, the proportion of area burnt and fire intensity across the burnt area varies considerably (that is, some habitats remain unaffected for a given fire). 103

113 Figure 6.4 Eatons Block burning study, one month after fire in 1977 (Photo: K. Tolhurst). Figure 6.5 Figure 6.6 Eatons Block burning study in 1978 (Photo: K. Tolhurst). Eatons Block burning study in 1979 (Photo: K. Tolhurst). 104

114 Figure 6.7 Figure 6.8 Eatons Block burning study in 1980 (Photo: K. Tolhurst). Eatons Block burn study in 1989 (Photo: K. Tolhurst). Figure 6.9 Eatons Block burn study in 1992 (Photo: K. Tolhurst). 105

115 Figure Eatons Block burn study in 1994 (Photo: K. Tolhurst). Figure 6.11 Eatons Block burn study in 1997 (Photo: K. Tolhurst). Figure 6.12 Eatons Block burn study in 2010 (Photo: K. Tolhurst) 106

116 There has been considerable debate about the potential role of both FRB and present day fire regimes in forest health and eucalypt die-back. Jurskis (2000; 2005a, b) for example, provides a wellreasoned hypothesis that present-day vegetation patterns in Australia are the interacting result of climate, fire, and variation in soil fertility. He notes that burning by Aboriginal peoples was understandably more frequent in woodlands and similar open grassy vegetation communities prior to European settlement, and that these grasslands were regularly burnt. Jurskis (2000) argued that whilst distribution of vegetation types has largely been determined by climatic and edaphic factors, human disturbance (particularly changed fire regimes) has been the main determinant of their floristic composition and structure. Turner and Lambert (2005a) noted that regular FRB may assist in maintaining the health and growth of eucalypt forests currently subject to crown die-back in the east coast region of Australia, while Jurskis (2005b) draws attention to forest health decline in North America caused by changes in fire regimes being implicated in adverse impacts on natural processes such as nutrient cycling, competition, mortality and recruitment. He advocates more frequent burning of native eucalypt forests to maintain forest health. The potential role of fire (along with a range of other factors) in the severity and extent of bell-miner-associated-dieback (BMAD) in NSW has been investigated by WardellJohnson et al. (2005). They note that arguments have been advanced in the scientific literature promoting either more frequent burning (for example, Jurskis 2005a, b) that mirror natural fire regimes or fewer prescribed fires (for example, Henderson and Keith 2002). BMAD is a complex issue, and Wardell-Johnson et al. (2005) concluded that many interacting factors, including altered fire regimes, most likely contribute to die-back at the landscape level. 6.4 A New Bushfire Strategy In December 2008, the Victorian Government released a first ever comprehensive bushfire strategy titled Living with Fire: Victoria s Bushfire Strategy (Vic. Govt. 2008). The Strategy aims to position Victoria s bushfire management agencies to effectively manage (fire) risk in partnership with the community. It is centred around six themes as follows: 1. Strategic directions (with a focus on increasing the annual prescribed burning program to meet fire prevention and fire ecology objectives). 2. Building community capacity to live with fire (with a focus on improving the community s understanding of the role of fire in the environment and increasing shared responsibility for all aspects of fire management). 3. Enhanced response and recovery. 4. Workforce/volunteer capability. 5. Planning for protection. 6. Risk and adaptive management (including investing in research). 107

117 Whilst the Strategy provides detailed and informative background data on all aspects of fire management in Victoria, it is the view of the authors that it fails to set quantifiable targets or commitments to support the identified future strategic directions. It is likely that the report of the Bushfires Royal Commission will be the catalyst for revision of the Strategy. In relation to theme 1, Johnston et al. (1983) provided baseline information on areas subject to FRB between 1972 and They recorded that an average of around 206,000 hectares (gross area) were burnt each year during this ten-year period, representing on average 2.8% per annum of the then total forested area of around 7.5 million hectares. 6.5 Fire and Sustainable Forest Management On the basis of (i) lives lost from periodic wildfires (including those in 1939, 1983 and 2009), (ii) the adverse impacts of high intensity, long duration (or landscape-scale) wildfires and changed fire regimes on soil and water resources and biological diversity, and (iii) the documented socio-economic impacts of large wildfires on regional communities, the risks of devastating wildfires is greater in Victoria than in the rest of Australia. It can therefore be argued that Victoria should give priority to all aspects of planned and unplanned fire when assessing SFM. For example, the Allocation to VicForests Order Review 2009 needed to inter alia take account of the impacts of the landscape scale fires (the Great Divide fires , and those of February 2009) on the structure and condition of large areas of native forest in assessing the future availability of timber resources. As noted in Section 3.5, Victoria has developed a set of C&I for SFM of the native forest estate as required under the Sustainable Forests (Timber) Act The C&Is, published in 2007 (DSE 2007b), are consistent with the Montreal Process and aligned to Australia s regional Indicators. DSE anticipate that monitoring and reporting against the C&I will amongst other things define sustainable forest management on public land in the Victorian context, allow credible performance reporting to the community, and highlight the forest sector s contribution to sustainable development in Victoria. The C&Is are also expected to provide the vehicle to monitor and report on progress towards the requirements of the Sustainability Charter. Above all, however, the prime purpose of the C&Is is to demonstrate through credible performance reporting that Victoria is committed to SFM. In order to achieve this goal, the Indicators must therefore be able to demonstrate that practices adopted in the management of State forests are sustainable from social, heritage, cultural, environmental and economic perspectives. Clearly, this requires fire to be considered in many of the Indicators developed under the Sustainability Charter. Under the national-level Montreal Process Implementation Group, Indicator 3.1b (area of forest burnt by planned and unplanned fire) is one of the 44 being used to report on progress towards SFM in Australia. It is the only Indicator that directly addresses fire. Fire, however, is relevant, to greater and lesser degrees, to all seven Criteria of the Victorian framework, but reference to fire in appropriate Elements and Indicators is often scant or non-existent as shown below: Criterion 1. Conservation of biodiversity Fire is not mentioned in two relevant Indicators of Element 1.2 (species diversity),viz. Indicator 1.2b (area of habitat available for forest dependent indicator species) and Indicator 1.2c (representative indicator species from a range of habitats monitored at scales relevant to regional management), despite the potential for significant and long lasting impacts of intense wildfires on habitat at both local and landscape scales. 108

118 Criterion 2. Productive capacity Fire clearly has the potential to significantly impact on Indicator 2.1 (area available and suitable for timber production). Whilst loss of area due to roads and quarries is mentioned, there is no reference to loss of area due to wildfire (for example, impacts of successive wildfires on Alpine Ash). Whilst it is assumed that fire impacts would be taken into account when reporting on this Indicator, there should be specific acknowledgement of fire. The same comment applies to Indicator 2.2 (total volume of wood available and suitable for timber production), though in this instance there is reference to the need to take account of growing stock lost to natural and anthropogenic disturbances. Criterion 3. Ecosystem health and vitality As for the national C&Is, this Criterion recognises that fire can impact on the health and vitality of forest ecosystems. Indicator 3.1 (ii) addresses the area of State forest burnt annually (by forest type and age class) by both planned and unplanned fire. The source of ignition (where known) is also considered. The focus, however, is on impacts of fire on forest ecosystem health. Indicator 3.2 considers prescribed burning for fuel-reduction, regeneration and ecological purposes, again from a forest health perspective only. Criterion 4. Conservation and maintenance of soil and water resources There is no reference to fire in relation to key soil attributes and to river health, despite extensive literature that demonstrates that fire can have adverse and long lasting impacts on soil and water values. There is, however, acknowledgement of the well-documented impacts of fire on runoff. Criterion 5. Carbon cycles Fire and harvesting are recognised in Indicator 5.1 as mechanisms for carbon loss, though it is noted that changes in carbon stocks due to fire are difficult to estimate. Flinn et al. (2008), however, provide such an estimate for the 2006/07 Great Divide fires using published data and specified assumptions. Criterion 6. Socio-economic benefits Despite a long history of Victorian wildfires in Victoria having major social and economic impacts at regional levels in particular, fire is not highlighted as an issue that needs to be considered when addressing this Criterion. The 2009 Black Saturday fires are a stark reminder of the massive social disruption that fires can inflict on rural and semi-rural communities. The economic impacts of these fires have likewise been of State and national significance. Clearly, fire impacts on Indicator 6.5a (employment in the forest sector), but this is not mentioned in the Victorian C&I document, but FRB is recognised in Indicator 6.2a (investment and expenditure in forest management). Criterion 7. Legal, institutional and economic framework Fire management is recognised as an important sub-indicator of Indicator 7.2 (extent to which the institutional framework supports the conservation and sustainable management of forests). From the above brief analysis, it is concluded that fire, and particularly wildfire or unplanned fire, is not given sufficient attention in the current set of Indicators being used to assess the status of SFM in Victoria and to report on trends in successive State of the Forests reports. There is voluminous scientific evidence supported by field observations and operational experience that wildfire in particular has the potential to have significant and often long-lasting socio-economic, environmental and economic impacts. This needs to be explicitly recognised in the next revision of the current Victorian C&I Guidance document (DSE 2007b). 109

119 Pests and Diseases Introduction Victorian forests have experienced significant damage from a range of insect pests and diseases ever since records have been kept and some of Victoria's more valuable commercial forests have been the target for some of these attacks. A brief outline of the more important damaging agents (not in any order of priority) is given below. Criterion 3 of the Montreal Process relating to ecosystem health and vitality uses Area and percent of forest affected by processes or agents beyond the range of historic variation as an indicator to assess impacts. This requires some understanding of the levels of impact and the spatial and temporal effects of such attacks. Having effective forest health surveillance systems in place is clearly integral to meeting the requirements of this increasingly important Criterion of SFM Insect Pests Mountain Ash Psyllid The Mountain Ash psyllid (Cardiaspina bilobata) is a sap feeding insect that caused extensive defoliation in several thousand hectares of Mountain Ash regrowth forest in the Central Highlands of Victoria in the early 1990s, with approximately 2,000 hectares of forest severely affected and a further 7,300 hectares moderately affected (Campbell 1997b). Extensive outbreaks also occurred in Mountain Ash forests around Tanjil Bren in 1986 within high quality 1939 regrowth stands (DCE 1992), while a review of available data also indicated it may have been responsible for defoliation observed in Mountain Ash forests in east Gippsland near the Rodger River in 1986 and 1994 (DCNR 1994). Damage is usually most prevalent in forests below an elevation of 800 metres on wetter sites such as drainage lines and easterly and southerly aspects. This is believed to be related to a preference by psyllids to feed on mature foliage which is not water stressed. The psyllid is well adapted to cold weather and thrives under the wet and cold conditions experienced in Mountain Ash forests. Mild, cool summers, lacking an extended period of hot, dry weather may also contribute to psyllid populations increasing (DCNR 1994). The psyllids, possibly in combination with the leaf pathogen Target Spot (Aulographina eucalypti), cause premature leaf fall resulting in relatively open crowns and premature death among some suppressed trees, thereby accelerating the natural thinning process (Coy 1996, Coy and Sweetman 1997). Lower parts of the crowns of trees are defoliated first, followed by the upper crowns in severe attacks. Some trees produce epicormic shoots in response to chronic attacks that may impact adversely on later wood quality (Campbell 1997a, b). Leaf and root pathogens are frequently associated with infestations and are therefore likely to be contributing factors to the defoliation and death of psyllid-attacked trees. There appears to be a high correlation between the severity of the Target Spot leaf pathogen and psyllid attack. This is possibly due to an increase in available nitrogen in foliage for the psyllid due to a breakdown in plant protein caused by the pathogen (DCNR 1994). Extensive areas of psyllid-affected Mountain Ash forest were salvage harvested, particularly in the Thomson catchment, near Erica. These areas, mainly in 1939 regrowth, were clearfelled rather than salvage thinned due to the difficulties associated with felling of the tall trees without damaging their neighbours. If younger stands had been severely affected by the psyllid, then salvage thinning may have been an option. 110

120 While no studies have examined any links between thinning and psyllid activity, it could be inferred from the above observations that thinning might assist in reducing psyllid populations by opening up a stand to sunlight and wind thus increasing ambient air temperatures, reducing moisture and consequently decreasing psyllid populations. Similarly, the risk of a Target Spot infestation is likely to reduce Gum Leaf Skeletoniser Periodically, the defoliation of extensive areas of River Red Gum (E. camaldulensis) forest has caused widespread public concern. Some members of the communities who see this widespread and almost total defoliation for the first time have thought they were witnessing the death of a forest. Such an event occurred in the Murray and Goulburn Valleys of Victoria in 1975 when 60,000 hectares were affected. However, a few months later the trees had recovered as they have done several times during their lifetime after previous attacks (DCFL 1986). The cause of the damage was the Gum Leaf Skeletoniser moth (Uraba lugens), whose larvae (caterpillars) feed voraciously on eucalypt foliage. The Gum Leaf Skeletoniser is widespread in Victoria and occurs in most Australian States. It attacks a wide range of eucalypt species and some ornamentals, but plague populations only develop during dry periods in areas where humidity is usually high for part of the year. In Victoria, such areas are found along the coast, in parts of the highlands and along the river systems of the warm inland regions. These severe defoliations result in lost growth and some mortality of suppressed trees. The few predators and parasites of the insect have low population densities and contribute little to its control. The insect's population crashes soon after the forest is completely defoliated because larvae starve and suitable egg-laying sites are not available. Along major rivers, another controlling factor is flooding, which increases humidity, allowing two species of Aspergillus fungi which kill the larvae, to flourish. Without flooding, most larvae reach maturity and large numbers of eggs are subsequently laid, greatly increasing the potential population of the next generation. From a silvicultural perspective, one possible way of reducing the severity of attack is to thin dense stands which suffer the most severe attack. This would reduce the amount of healthy foliage close to the ground and so reduce the availability of egg-laying sites. It would also restrict the direct movement of larvae between trees, which occurs when crowns overlap. Thinning of stands to a density of less than 750 stems per hectare appears to significantly reduce the severity of outbreaks which follow, provided the stumps of thinned trees are prevented from coppicing and debris is removed. Research in the 1970s has enabled an understanding of the biology of this insect and its relationship with River Red Gum forests and stand silviculture (Harris 1974, 1975, Harris et al. 1977). The east Gippsland forests have a long history of infrequent defoliation episodes caused by both biotic and abiotic agents. The Gum Leaf Skeletoniser and the Mottled Cup Moth (Doratifera vulnerans) have been implicated in some of these episodes (Collett and Fagg 2009). In October 2003 and 2005, large tracts of coastal and foothill mixed species forests were subjected to another significant defoliation event. Consistent with AFS requirements, ground surveys were conducted on five occasions between November 2003 to March 2006 to identify the primary causes of the defoliation, to determine the extent of the damage, and to monitor tree recovery and overall forest health. These surveys showed that the principle defoliating agent in the 2003 outbreak, which was more severe than the 2005 episode, was the Cup Moth (Doratifera spp.). The 2005 outbreak was mainly linked to the Gum Leaf Skeletoniser. A feature of the survey results was variation in susceptibility of the various eucalypt species. Potential predisposing factors to the separate outbreaks are discussed by Collett and Fagg (2009). They also stress the importance of Forest Health Surveillance, particularly if outbreak frequency is increased by prolonged drought. 111

121 Phasmatids The stick insect or phasmatid (Didymuria violescens) has long been known as a serious primary insect defoliator in specific forest types in southeast Australia. In Victoria, widespread and severe infestations of this phasmatid have been responsible for serious defoliation of Mountain Ash forests in the Central Highlands and Alpine Ash forests in northeast Victoria. The biology of the insect has been studied in detail (Neumann et al. 1977). Under Central Highlands conditions, the phasmatid generally has a biennial life cycle, so that peak populations rarely occur over successive years. The eggs of the insect usually incubate on the forest floor for around 20 months before hatching between spring and early summer. The newly hatched nymphs then stagger across the forest litter and debris, and climb any vertical object in response to strong anti-gravity instincts (Neumann et al. 1977), but they will only survive if the vertical object is a eucalypt because they need eucalypt foliage for food and water soon after hatching. The five nymphal instars and the adults then spend their life feeding on the eucalypt crowns. Eggs are generally laid (dropped to the forest floor) in late summer to early autumn, with the total number of fertile eggs being over 300 per adult female insect. There is therefore potential for rapid population increases. Forest litter sampling for viable egg counts is an effective predictor of phasmatid populations. Figure 6.13 Alpine Ash defoliated by Phasmatids in the late 1950s Bago, NSW (Photo: Reg Humphreys) 112

122 Research has shown that severe infestations can kill co-dominant trees. High mortality is associated with repeated biennial defoliation, while growth of surviving trees (mainly dominants) can be markedly reduced. FRB in dry sclerophyll forests is very effective in controlling the pest by destroying the eggs in the litter, but this technique is not suitable for ash-type forests. In the past, potential serious infestations have therefore been controlled by aerial spraying with the organophosphate, Malathion, following confirmation from egg sampling that a serious outbreak will occur (Table 6.1). These have been intervention treatments rather than corrective action following severe defoliation. The impacts of spraying this pesticide on non-target insects have received considerable research attention. A summary of this research along with details on the mode of action and toxicity of Malathion has been provided by Harris (1980). Table 6.1 Areas defoliated and those treated with intervention spraying to control phasmatids in the Central Highlands. The number of localities defoliated for the first time are shown in parenthesis. Year Areas Sprayed with Malathion No. of Area Localities Sprayed Sprayed (ha) 1964/65 2 (2) 1, /66 1 (1) /67 2 (1) /68 4 (2) /69 5 (1) 2, /70 15 (9) 1, /71 27 (11) 6, /72 12 (1) 1, /73 25 (9) 3, /74 7 (1) /75 5 (2) 1, /76 3 (0) / Source: Neumann et al. (1977) Pathogens Target Spot or Corky Leaf Spot Target Spot or Corky Leaf Spot kills variously aged eucalypt foliage in high rainfall areas during wet days of spring and summer. In the 1970s, it was associated with significant defoliation of Shining Gum (E. denticulata) on the Erinunderra Plateau in east Gippsland (Marks et al. 1982, Park et al. 2000). As noted above, in the late 1980s and early 1990s and in association with the Mountain Ash psyllid, it contributed to severe defoliation of Mountain Ash in the Central Highlands. 113

123 Armillaria Root Rot The fungus, Armillaria Root Rot (Armillaria luteobubalina,) is a native primary pathogen of many species of both native and introduced trees (Shaw and Kile 1991, Kile 2000). The conditions that trigger an outbreak are not clear but appear to be related to the presence of a food base (for example, stumps and roots) and/or soil moisture status and the physiological state of the host. Symptoms vary from the occurrence of scattered dead individual trees to distinct patches or infection centres up to 20 hectares in extent (Edgar et al. 1976). In Victoria, it occurs mainly in mixed species eucalypt forests and is an important disease in some damp and wet forest types, principally in western central Victoria. Historically, it has been mainly associated with selectively logged areas, with the disease impact being greatest in mature and over-mature stands, causing crown dieback, reduction in basal area and volume and eventually death. Armillaria Root Rot has damaged approximately 2,000 hectares of mixed species forest in the Mt Cole and Wombat State forests. Similar to the Cinnamon fungus (Phytophthora cinnamomi), this species kills trees and shrubs of any age through infection of the major roots and stem of the plant. It spreads between plants mainly through root-to-root contact. The disease has been the subject of long-term research in Mt Cole forest (Kellas et al. 1997) who reported 18-year results of a study aimed at determining the impact of a range of intensive site preparation treatments (including deep ripping and whole tree pushing) on tree establishment. These treatments, however, were not effective in reducing disease impacts including tree mortality on high disease hazard sites. Management of the disease remains problematical. One approach is to reduce the chance of contact between healthy trees and the fungus by adopting clearfelling (followed by regeneration burning) rather than selective harvesting techniques in areas prone to Armillaria infestation. The creation of an ashbed should promote dense, vigorous and healthy seedling regeneration that will assist in genetic selection for resistance to infection, and drying of the site, thus making conditions less favourable for Armillaria (Smith and Smith 2003). Single tree selection, group selection and shelterwood harvesting systems and thinning may exacerbate the disease on sites that are prone to infection, because live trees which may support growth of the fungus are left in the harvested areas (Kellas et al. 1997). However, in areas less prone to infection, the impact of Armillaria following partial harvesting such as thinning are not certain, with the findings of studies appearing contradictory. In some overseas studies, thinning has resulted in increased incidence of disease by providing a higher food base for the pathogen, whereas other studies have shown that reduction in tree stress has resulted in increased resistance to disease development (Hood et al. 1991). It is likely that the interaction between relevant site factors will be variable for this class of disease, and further work is needed to clarify under what conditions operations such as thinning can be undertaken without predisposing residual stands to disease attack. Other Armillaria species (A. hinnulea, A. novae-zealandiae, and A. fumosa) are generally secondary pathogens of eucalypts and may cause damage in mountain areas or areas with high rainfall (Marks et al. 1982). They only attack trees weakened by other causes such as drought, waterlogging, low light levels, nutrient deficiency or insect defoliation. These environmental and biological stresses may also play a role in disease development for Armillaria Root Rot. 114

124 Cinnamon Fungus The Cinnamon fungus has been responsible for significant ecological damage in both National Parks and State forests (including commercially-important forests) in Victoria. The disease has been the subject of intensive research in Western Australia and Victoria. Victorian research experience with this soil-borne fungus which infects roots is well documented, and Marks and Smith (1991) provided an excellent summary of this work including an account of the various hypotheses concerning its origin. The weight of evidence appears to favour the conclusion that the disease is a relatively recent introduction in Victoria (possibly in 1935 when it may have been responsible for a disease of Cricketbat Willow). The first probable occurrence of the disease causing forest dieback was in 1938 near Nowa Nowa in east Gippsland. Under favourable conditions (for example, disease-conducive soils), the disease can cause severe dieback and tree mortality, particularly if forests have been pre-disposed to the fungus by, for example, selection logging and climatic conditions. Eucalypts vary considerably in their tolerance to the disease, and lists of tolerant and susceptible species are available (for example, Marks and Smith 1991). Unfortunately, the fungus is readily transported in infested soil attached to heavy machinery (for example, logging equipment and bulldozers used to construct fire trails), in nursery seedlings, in infested gravel, and by water and animals. Its mode of action depends on diseased soils being saturated so that unsuberised fine roots can be infected with zoospores during warmer weather. In susceptible species, the fungus then progresses into the major roots and stem leading to crown dieback. It ultimately girdles the stem leading to tree death. Figure 6.2 Eucalypt dieback, South Gippsland (photo Ian Smith) 115

125 Marks and Smith (1991) provide detailed comment on management options for Victorian forests, assuming that the fungus is an introduced disease, within an integrated management approach. These management options, which relate mainly to rehabilitating sites severely affected by the disease, include increasing genetic resistance by encouraging vigorous regeneration using seed trees (Fagg 1987) or seed trees plus supplementary sowing onto well burnt seedbeds (Lutze 1999), direct sowing of mixtures of local eucalypt species (Lutze 1999), replanting with tolerant species on well prepared sites (for example, using mounding), chemical control, integrated control (that is, a combination of two or more of the above), and land hygiene and quarantine to slow the spread of the disease. The option of replanting in combination with intensive site preparation has received considerable attention, with research in east and south Gippsland commencing in Lutze (1998a) reported growth of six disease-tolerant eucalypt species plus Silvertop Ash to 21 years of age planted on light, moderate and severe dieback hazard sites in east Gippsland and a high hazard site in south Gippsland. As expected, Silvertop Ash experienced very high mortality, though Lutze (1998a) considers that factors other than the fungus may have contributed to this poor survival. Based on mortality and growth over the 21-year period, Spotted Gum (Corymbia maculata) and Sydney Blue Gum (E. saligna) were identified as species with high potential for good growth on high hazard sites. Reforestation with non-indigenous species in State forests is therefore one option in situations where disease issues mean that local species cannot be used. Tregonning and Fagg (1984) noted that epidemics occurred in east Gippsland in 1953, 1956, 1967 and 1971, and they have invariably been associated with 3 to 6 months of heavy summer rainfall followed in autumn by a similar period of low rainfall. They calculated that such conditions can be expected in the future (every 15 years on average for east Gippsland) so that parts of the coastal forests will continue to be at risk, particularly if past selection logging has made them more vulnerable. Rehabilitation of dieback sites with the full range of original species mixtures and limiting the spread of the Cinnamon fungus to areas not already infected remains a challenge in Victoria. Consistent with Montreal Process requirements relating to rehabilitation of degraded forests, and with requirements under the AFS, there is at the very least an obligation to identify affected areas and assess their potential for full rehabilitation. 6.7 Forest Health Surveillance Resources available for routine forest inspections have declined markedly in the past 2 to 3 decades despite an ever increasing risk of pest and/or disease introductions through tourism and other vectors. Both the Montreal Process and the AFS require that plans and procedures are in place for effective forest health surveillance, and that the activity is regularly undertaken. If a damage agent is detected, then there is a requirement to assess its potential for causing significant damage (ecological and/or commercial) and to review management procedures within an Integrated Pest Management framework. As damage agents do not respect land tenure boundaries, surveillance needs to cover all forested land. Under Montreal Process Criterion 3, the Wood and Paper Industry Strategy supported three projects focused on using an index of crown condition to indicate crown health (Turner et al. 2003). This is a remote sensing approach that has potential to provide substantial cost savings compared with ground-based assessments. Promising results were obtained from two of the projects that focused on native forests, including Mountain Ash at Tanjil Bren. The remote sensing technique, however, requires further development before it can be routinely used. 116

126 6.8 Discussion and Conclusions Conclusions on the management of fire, pests and diseases in forests are: 1. In recent times, Victoria has experienced extensive and damaging wildfires in 2003, 2006/07 and Amongst other things, these fires have impacted significantly on the long-term availability of timber sourced from the commercially-important ash-type forests. The 2003 and 2006/07 fires were the second and third largest landscape-scale fires ever recorded in Victoria, whilst the 2009 fires were the most devastating in terms of loss of life. 2. These and other fires in south-eastern Australia have led to a number of State and Federal Inquiries and intense public debate, culminating in a Royal Commission into the 2009 Victorian Black Saturday fires. 3. Extensive (landscape-scale), long-duration fires with long, active fire perimeters that burn for long periods over summer increase the probability of a catastrophic weather event being encountered (such as the 2003 and 2006/07 fires that collectively burnt the equivalent of 10% of Victoria s land area) leading to the likelihood of numerous rural and semi-rural communities being simultaneously and seriously impacted. Indeed, there were some major fire runs during the 2003 Alpine fires (fortunately these were largely confined to relatively remote forested land) that were arguably of a similar intensity to the Black Saturday fires. These fire runs coincided with the blow-up of the 2003 Canberra fires that experienced similar extreme fire weather. 4. Victoria has recently up-dated its Code of Practice for Fire Management on Public Land and developed a first-ever Fire Strategy for Victoria. 5. The impacts of wildfire on the various components of ecosystems can be severe and longlasting depending on the forest type and the extent and severity of the fire. FRB is an important management tool in those forest types not sensitive to fire. It is concluded that: 6. Evidence is overwhelming for the effectiveness of FRB in reducing the extent and severity of wildfires and assisting ground crews to more safely contain fires. Based on the findings of a range of Government Inquiries into the 2003 Alpine fires and 2006/07 Great Divide fires, a more substantial FRB program needs to be maintained in Victoria to better protect community assets including commercially important forests. This is particularly important if climate change progressively results in more prolonged fire seasons with increased numbers of extreme or catastrophic fire danger days, and A significant FRB program can be managed to minimise ecological impacts, though longer-term data are needed from on-going research and ecological monitoring before more definitive impact statements can be made. In the meantime, future FRB programs should take full account of the findings from the Wombat State forest study (DSE 2003a) and companion studies. On the basis of major wildfire events, particularly those from 1939 onwards, it is clear that there is a very strong case for a strengthening commitment by relevant agencies (particularly PV and DSE) to take all reasonable steps to limit the extent and severity of future wildfire events. This is best accomplished by embracing well-documented integrated strategies including FRB targeted at protecting not only communities but also valuable timber resources (plantations and natural forests) and water supply catchments, early detection and effective first attack of new outbreaks, and efficiently coordinated suppression activities that use the latest technology. 117

127 7. Fire retardants should continue to be used, providing such use recognises their limitations and potential environmental impacts if applied inappropriately (for example, they may potentially violate a number of Montreal Process indicators that relate to contamination of soils and water systems with chemicals). 8. Fire, and wildfire in particular, transcends all aspects of SFM. There is overwhelming evidence, based on loss of life and the adverse impacts of relatively frequent landscape-scale, high intensity fires on environmental values and timber resources, that the risk of devastating wildfires is greater in Victoria than in the rest of Australia. The C&I developed for Victorian State forests (DSE 2007b) does not explicitly recognise the potential for significant and longlasting socio-economic, environmental and economic impacts of unplanned fires. This should be rectified in the next revision of the C&I Guiding Document. 9. Annual FRB targets must recognise the influence of seasonal conditions (which dictate the number of suitable "burning days") and the fact that only a portion of Victoria's State forests are suitable for FRB due to the silvical and ecological characteristics of individual forest types. Ash-type forests for example are not suited to landscape-scale FRB. 10. Serious outbreaks of phasmatids in Victoria s Mountain Ash forests can be periodically expected in the foreseeable future. Under SFM guidelines, damage agents such as this stick insect need to be controlled if they pose a major threat to the health and vitality of commercial forests and if they will have significant economic impact. Community attitudes to chemical spraying have most likely changed significantly since the last spraying with Malathion in 1975/76. Contingency planning should therefore be instituted in readiness for the next serious outbreak. Serious and prolonged infestations of phasmatids leading to widespread and uniform tree mortality in water supply catchments dominated by Mountain Ash could lead to regrowth and hence have important implications for long-term water yields for the city of Melbourne. 11. Managing the Mountain Ash psyllid, Target Spot and the Gum Leaf Skeletoniser within an Integrated Pest Management framework is also challenging, with the use of thinning seen as a potential silvicultural approach to minimising the impacts of these defoliators. Clearfelling may have a role in managing Armillaria, but the role of thinning in managing this disease is less clear and needs further work. 12. There is a requirement under SFM guidelines to rehabilitate sites degraded by fire, pests, diseases and other agents. In the case of dieback sites associated with the Cinnamon fungus, the technology to restore affected areas is now reasonably advanced and well documented. 13. Monitoring of the entire forest estate for forest health, and limiting major wildfire events, will be a significant challenge in the future management of Victoria's forests. The interaction of climate change and forest health and wildfire frequency, severity and extent are additional yet important considerations. 14. The importance of forest health surveillance (FHS) cannot be over-stated. Consistent with AFS requirements, this is fully acknowledged by the plantation forestry sector who accord increasing priority to this critical monitoring activity. It is important that such surveillance and any necessary follow-up actions are given equal priority for all tenures of native forests where potential for large-scale ecological and commercial damage is significant. The relatively recent insect defoliation of mixed species forests in east Gippsland (Collett and Fagg 2009) is a good example of the benefits of an FHS approach. 118

128 7. PRODUCTIVITY AND YIELD 7.1 Introduction Forest productivity is basically the rate of production of biomass at a particular time. In commercial forests where the focus is on wood, production is considered in terms of total accumulated wood 3 3 (m /ha) or mean annual increment (MAI, expressed as m /ha/yr), the latter being age dependent, and thus should always be interpreted carefully. Yield, on the other hand, represents the quantity of wood actually harvested from a site and is therefore less than the estimates of total productivity due to losses in harvesting (for example, breakage), utilisation standards (for example, small-end diameter limit) and tree mortality. That is, yield is the net productivity of a tree or stand product of interest at a particular time. Actual productivity is a product of the interaction of forest type, environmental conditions, disturbance and management effects. In east-coast Australian forests, there is a broad relationship between forest type and site characteristics. Some species (such as Mountain Ash) are found on sites with better soils and higher rainfall, whereas species in the Low Elevation Mixed Species forests (such as Silvertop Ash and a variety of stringybark species) will be found on sites with lower rainfall and often nutritionally poor soils, resulting in differences in patterns of growth and overall productivity. 7.2 Estimation of Growth and Biomass - General Considerations Estimates of growth and yield are made in a number of ways: Growth is usually estimated using a series of fixed area, permanent plots on which periodic measurements of trees are undertaken. These plots provide data which can be used to provide information on rates of production and potential yields for given growth periods. They also provide a basis for developing or validating growth models. It is important that there are sufficient numbers of plots to cover site variation and thereby avoid any sampling bias. Inventory is often a single measurement in time of the estimation of characteristics of a large forested area and is therefore usually stratified and the intensity specified (West 2004). Research trials and plots provide some growth information, and depending on the type of trial, responses to treatments such as thinning, and Yield data are often provided as the volume or weight per hectare of timber removed from a coupe or an equivalent management unit. Additionally, the estimated products and size classes need to be stated. Some estimates only report sawlog production, while others are for either total merchantable production or total productivity. A number of growth and yield models have been developed which allow predictions to be made on growth and to estimate growth at the forest or estate level. The models, their utility or their accuracy are not reviewed here. Various methods used to estimate yields give differing results due in part to the size of the population being assessed and the intensity of the sampling. It is therefore important to identify the purpose of the yield data so that appropriate scales of management are adopted. The key questions in relation 119

129 to a forest type are the growth rates and ranges (across sites types), the size classes and how silvicultural treatments affect these growth rates. Land management agencies and research organisations have large quantities of information in relation to these questions, but only a small proportion is published. Also, it is often questionable how typical these published data are in relation to the entire accessible harvesting area. Some published volume data for common Victorian eucalypts have been reviewed and are presented in Table 7.1. They have been generally derived from unthinned stands, and thus would be modified by silvicultural treatments including thinning (La Sala et al. 2004). Table 7.1 Reported productivity estimates for different forest types relevant to Victoria at ages of 40, 60 and 80 years. The stands have mainly been unthinned. a b Species SI Mean Annual Volume Mean DBH Reference 3 (m) Increment (m /ha/yr) (cm) Approx. Age (yrs) Alpine Ash after Khanna (1998) n.a Grierson et al. (1992) n.a Cut-tail Cut-tail Cut-tail 40 Mixed forest n.a. Mountain Ash Mixed forest 50.8 Borough et al. (1984) Horne and Robinson (1990) Borough et al. (1984) West (1991) Grierson et al. (1992) 2.57 Bi and Jurskis (1996b) Bi and Jurskis (1996b) Bi and Jurskis (1996b) c d 8.3 Grierson et al. (1992) 10.4 La Sala et al. (2004) Messmate 40 Silvertop Ash Incoll (1974) Silvertop Ash Incoll (1974) Silvertop Ash Incoll (1974) Silvertop Ash 31 e Mixed forest n.a Murray River Redgum Q1 f Borough et al. (1984) Bridges (1983) 5.9 Grierson et al. (1992) 26.7 f 50.0 Baur (1983) Murrray River Redgum Q Baur (1983) a b c d e f Site Index, dominant height at age 50 years (20 years is used in Victoria) Diameter at breast height (1.3 metres) Regularly thinned site Reported at 65 years of age Reported at 38 years of age Site quality classes 1 and 2 n.a. Not available 120

130 7.3 Assessment of Victoria's Forest Resources Victoria has invested heavily since 1994 in a State-wide Forest Resource Inventory (SFRI) program with the aim of providing forest managers with a reliable, timely and complete set of forest resource information for making informed and consistent sustained yield forecasts, and decisions on forest land-use planning and resource allocation (DNRE 1997). SFRI has used state-of-the-art technology and inventory techniques that are more reliable and flexible than traditional methods. Both aerial photographic interpretation (API) and field inventory (using a model-based sampling approach to determine the distribution of inventory plots in the various forest stand classes identified by API) are employed. SFRI places Victoria in a strong position to set sustainable yields for a range of forest management scenarios. 7.4 Estimates of Biomass Published data together with Departmental utilisation and inventory data were used to estimate the biomass in above-ground components for 50-year-old Mountain Ash, a combined HEMS/LEMS forest of medium to high productivity, a typical Box-Ironbark forest and a mature River Red Gum forest (Flinn et al. 2001). These estimates (Table 7.2) show the large range in total standing biomass between these forests and reflect different soils, rainfall, elevation and other relevant site factors together with the characteristics of the individual tree species. In order to estimate the stemwood component of the biomass estimates shown in Table 7.2, estimates of growth rates for the four forest types were required (Flinn et al. (2001). These estimates (broad average growth rates) were obtained from the Department of Natural Resources and Environment (DNRE) and are provided in Table 7.3. It is noteworthy that the MAIs in Table 7.3 are higher than those deduced from Table 7.1. Flinn et al. (2001) considered that the DNRE estimates for ash-type forests and River Red Gum forests were conservatively low whereas estimates for the HEMS/LEMS forests would not be realised unless more attention was given in the future to thinning, fire protection and disease management. The authors of this review agree with the assessment. Furthermore, we believe that the ratios of sawlog to residual log deduced from Table 7.3 are highly optimistic from a sawlog production perspective. The data shown in Table 7.2 for the various above-ground biomass components are an essential prerequisite for determining the quantities of nutrients removed in harvested wood (with or without bark) as a proportion of the total quantities held in a given ecosystem (comprising above- and belowground biomass and soils). Such determinations when combined with other data (see Section 8.2.3) can be used to predict the likely impact of logging at varying intensities on long-term site productivity. Maintaining the productive capacity of a site is an important requirement of Sustainable Forest Management (SEM). 121

131 Table 7.2 Comparative biomass production for Victorian forests. Forest Type Biomass Component Dry Weight (t/ha) Mountain Ash Harvested logs Stembark Branches and leaves Residue and standing dead trees Litter and understorey Total HEMS/LEMS Harvested logs Stembark Branches and leaves Residue including stumps Litter and understorey Retained overwood Total Total Box-Ironbark River Red Gum Harvested products: Fuelwood Sawlogs/sleepers Fencing materials Stembark Understorey/litter/slash Retained overwood Total merchantable stemwood 1 80 Non-merchantable stemwood 25 Stembark 10 Understorey/litter/slash 20 Total 135 Source: Flinn et al. (2001) 1 Includes sawlogs, sleepers and posts. Table 7.3 Estimates by DNRE of broad average growth rates for four forest types. Forest Type Mean or Periodic Annual Increment Nominal 3 (m /ha/yr) Sawlog Rotation Sawlog Volume Gross Bole Volume (years) Ash-type forests (MAI) HEMS/LEMS (MAI) River Red Gum (PAI) Box-Ironbark (PAI) Source: DNRE (1997) 122

132 Productivity of Forest Types Alpine Ash Forests Productivity of Alpine Ash has been reported from growth studies in stands in NSW (Bago/Maragle) and it is assumed they have applicability to Victoria. For an unthinned stand with a site index of 40 3 metres (dominant height at 50 years), a total production of 763 m /ha at 60 years was achieved, 3 representing a mean annual volume increment of 12.7 m /ha/yr (Table 7.1). The average estimates 3 used by Grierson et al. (1992) indicated a total mean annual volume increment of about 10 m /ha/yr over 80 years, though the individual plot data indicated there was a large range in productivity. A more recent study (Chee 1999) estimated total aboveground biomass (using previously developed allometric regression equations) for five regrowth stands in Bago State Forests ranging in age from 42 to 91 years of age. The aboveground tree biomass ranged from 132 t/ha for a 43-year-old stand to 460 tonne /ha for the 91-year-old Alpine Ash stand. The rotation length for Alpine Ash forests is generally 60 to 80 years and beyond but could be reduced depending on available markets for thinnings and the growth response to thinning. While Horne and Robinson (1990) reported there was little to be gained from thinning Alpine Ash at a later age because of its heavy self thinning nature, Incoll (1972) indicated that thinning of regrowth was beneficial and more recent work (DSE 2006b) supports these findings. The effect of thinning was to produce a greater sawlog yield. Based on the growth curves of Borough et al. (1984), the mean annual volume increment culminates somewhere between 35 and 40 years, whereas the indications from Horne and Robinson (1990) were that it was still increasing in this period and culmination was somewhere beyond 50 years. This may primarily be a result of site quality differences, as the analyses from Ryan et al. (1996) at Bago/Maragle indicated significant soil-site variation which related to forest type and productivity. Based on inventory plots and harvesting estimates (standing volume plus growth minus yield), the 3 MAI of Alpine Ash sawlogs over the whole Bago/Maragle area was 2.43 m /ha/yr. The harvesting undertaken was on a selection system. If it was assumed that the ratio of sawlog to small log and 3 pulp was 1:3, the total mean annual volume increment would be in the range of 9 to 10 m /ha/yr Mountain Ash Forests Mountain Ash is considered to be a highly productive species based on high yields at harvesting and a number of regeneration studies undertaken post-wildfire. Yields from harvesting (assumed to be 3 total over-bark) in 80-year-old stands indicate average MAI at about 10 m /ha/yr. Productivity figures based on growth plots are, however, generally higher. For example, Borough et al. (1984) reported 3 14 m /ha/yr at 60 years. The average estimates used by Grierson et al. (1992), based on a selection 3 of inventory plots, indicated a mean annual volume increment of approximately 11 m /ha/yr at 80 years but the plotted information showed a large range in productivity (between about 7 and 15 3 m /ha/yr). There appears to be only limited published data on growth responses of Mountain Ash to commercial thinning in Victoria. Thinning prescriptions and expected responses are largely based on the comprehensive work of Webb (1966). More recently, as part of the Young Eucalypt Program, West (1991) used modelling techniques to also show that regrowth stands are highly responsive to thinning in terms of both improving the proportion of sawlogs and reducing overall rotation length. Compared 3 with unthinned control plots (sawlog production of 738 m /ha at age 80 years), the most heavy 3 thinning tested at age 20 years reduced rotation length to 50 years with a sawlog yield of 699 m /ha. One of the main reasons for this response was that thinning captured early mortality that occurred in 123

133 dense, unthinned stands. Responses to strip and uniform thinning in (relatively old) 31-year-old stands of 1939 Mountain Ash at two sites in the Central Highlands using un-replicated plots were reported by DSE (2006b). Thinning from below (60% basal area reduction) was found to be associated with the greatest percent basal area growth response. The nominal rotation length for this species is often proposed as 80 years, although harvesting of the 1939 regrowth commenced when trees were aged about 50 years. There does not appear to be a silvicultural reason for a rotation length of 80 years. Current thinning of 20- to 27-year-old logging regrowth may lead to the option of adopting shorter rotations for thinned stands (for example, 50 to 60 years) in the future. However, any such reduction may favour wood production objectives at the possible expense of ecological considerations, and therefore would need to be critically evaluated High Elevation Mixed Species Forests The productivity of the HEMS forests covers a wide range, as expected considering the variety of species and sites. They are generally productive but this is greatly affected by site factors. The effect of site differences was demonstrated by Bi and Jurskis (1996b) for Cut-tail (Eucalyptus fastigata) in 3 3 NSW covering three site qualities ranging from 3.4 m /ha/yr to 9.8 m /ha/yr at 60 years where there was no residual overstorey trees. However, retention of overstorey trees greatly impacted on this growth. The average data from Grierson et al. (1992) covering a limited number of mixed species plots, showed a similar range of productivity. Responses to thinning have been reported for young stands where there are significant increases in tree diameter (and hence sawlog production) at some expense of total productivity (Borough et al. 1984). The silvicultural prescriptions often include burning of thinning residues and hence stands need to be at an age where fire would have minimal impact on the trees Low Elevation Mixed Species Forests The LEMS forests are often on lower rainfall areas and less fertile soils. The estimates of productivity, especially anything equating to rotation length are limited. Most of the data available on productivity are from Silvertop Ash related forests. The small amount of data used by Grierson et al. (1992) for 3 Victorian forests indicate an MAI at 80 years of about 6 m /ha/yr with a peak MAI occurring quite early in the rotation (before 30 years of age). The estimates from Incoll (1974) were significantly greater than the estimates reported here. In the same forest type on the south coast of NSW, there have been a number of reports on productivity resulting from regrowth after integrated harvesting (Bridges 1983). While there were site quality differences, the estimated mean annual volume increments (under-bark) at 14, 25 and 38 3 years were 3.4, 4.5 and 5.4 m /ha/yr respectively. A spacing trial established at 4 years of age in Silvertop Ash and measured when 28- years-old, indicated significant responses in total productivity and in the production of higher quality logs (V. 3 Jurskis, Forests NSW, pers. comm.). The control plots had an MAI of 10 m /ha/yr in comparison with 3 lightly thinned plots with 15.1 m /ha/yr and actual production of sawlogs and increased total production. Heavier thinnings led to a progressive decline in total productivity. These results were supported by the work of Connell and Raison (1996) in Silvertop ash forests in east Gippsland. They measured a 61% increase in basal area for the largest 150 stems per hectare three years after a thinning treatment that removed 45% of the basal area. 124

134 A significant thinning program commenced in the east Gippsland LEMS forests in the early 1990s following a major multi-disciplinary research study which demonstrated that the thinning was both economically feasible and ecologically sound (see Flinn and Mamers 1991). A replicated study established in 1968 in a 21-year-old stand of Silvertop ash in the Boola Boola State forest involved six thinning treatments that could be grouped into four levels of retained basal area (30%, 69%, 81% and 100%) (DSE 2006b). The most recent measurement in 1994 (26 years post-thinning) showed that the highest gross bole volume was associated with the lightest thinning, but there was a clear trend for more gross volume in trees larger than 40 cm DBHOB with increasing intensity of thinning. Furthermore, the growth response by diameter class for the medium intensity thinning (69% retained) when compared with unthinned plots showed much more volume in larger diameter classes for the thinned plots. 7.6 Growth Estimates for the Major Forest Types Estimates of volume production based on plot estimates (Table 7.1) show high variation in growth between forest types together with differing patterns of growth. The estimates of yields from broader areas need to consider the typicality of the stands, however, while expecting differing proportions of sawlog, yields should reflect the differences noted in Table 7.1. That is, making the assumption that the mid-range productivity estimates are reasonably typical of the broader forest type and that sawlog production is 25% of total production and a rotation length of 60 years is adopted, the following sawlog MAIs can be deduced (from Table 7.1): 3 Alpine Ash 2.5 m /ha/yr Mountain Ash 3.1 m /ha/yr HEMS 2.4 m /ha/yr LEMS 1.6 m /ha/yr If the estimates are reasonably typical of these forests, and even assuming losses in harvesting, the implied sawlog MAIs in Table 3.1 appear to be highly conservative. However, the yields in Table 3.1 generally assume a nominal rotation length in excess of 60 years and have been carefully calculated on the basis of data from SFRI (DNRE 1997) together with extensive field experience comparing predicted yields with actual yields. 125

135 7.7 Discussion and Conclusions The published data available on productivity and yield of different forest types in Victoria are quite limited. However, based on published data available from Victoria and NSW, the following conclusions can be drawn: 1. Alpine Ash is a very site specific and productive species, and (like Mountain Ash) is strongly self thinning and is managed on rotations in excess of 60 years. Commercial thinning is practised on suitable sites. 2. Mountain Ash can be highly productive but there is high variation in its productivity levels. It is responsive to thinning. Longer rotations without thinning probably lead to significant timber losses due to mortality, pointing towards the value of shorter rotations with well-planned thinning regimes. 3. The HEMS forests are valuable and productive, though yields are highly variable depending on species and site index. Productivity data are very limited. There is evidence of a significant loss of total productivity if retained overstorey is common. 4. The LEMS forests have lower levels of productivity than HEMS or Ash forests. Many stands have a high density of regeneration after disturbance and there are strong responses to appropriate early thinning regimes. The suitable topography of many LEMS forests and their proximity to processing plants are additional advantages from a thinning perspective. 5. The available data indicate that the MAIs for Alpine Ash, Mountain Ash, HEMS and LEMS forests used to estimate sawlog yields from publicly-owned native forests are lower than many published growth plot estimates. The reasons for this need to be evaluated. 6. There is reliable information on above-ground biomass components for Victoria's main forest types. This is essential prerequisite information to determine quantities of nutrients removed in harvested wood and hence the potential impact of logging on long-term site productivity. 7. Silvicultural guidelines have been prepared to guide thinning programs in a range of forest types. They are based on research and inventory data supported by operational experience (see Section 4). 126

136 8. ENVIRONMENTAL ASPECTS OF SUSTAINABLE FOREST MANAGEMENT 8.1 Introduction A primary objective of forest management is that activities are undertaken in a sustainable manner. Timber harvesting causes visible soil and vegetation disturbance and is often considered to have significant on-site and off-site (downstream) effects. An objective of the development of Codes of Forest Practices is to provide guidelines for Best Management Practice (BMP) to minimise such effects. Several processes which relate to maintaining the productive capacity of sites and protecting environmental values were identified in the Montreal Process which provided Criteria and Indicators for the purpose of evaluating the impacts of timber harvesting. This section of the review focuses on the impacts of timber harvesting and associated activities on soil, water, flora and fauna values along with the role of forests in carbon storage Soils and Soil Management Erosion A low to high level of soil erosion and deposition occurs naturally in all forests. High levels of erosion in undisturbed forests can occur during high intensity storm events. Increased rates of soil erosion, however, are of concern in managed forests for two main reasons. Firstly, productive capacity may be reduced by loss of soil and secondly, there can be off-site impacts on water supplies and aquatic habitats. While it is recognised that significant loss of soil can lead to a reduction in productive capacity, a major focus on sediment has been on the impacts on the aquatic environment, essentially as measured in water quality. Increases in soil loss vary according to a wide range of factors including natural features (soil type, vegetation type, rainfall pattern) and management effects (harvesting system, roading system, types of retention/protection). The reasons and causes of increased erosion need to be considered in relation to the background level of erosion for that forest/soil type (Lacey 1993): Mass Wasting. Mass wasting in forest lands commonly occurs as sudden slope failures including deep-seated slumps, landslides, and debris avalanches on steep terrain. It is often initiated or accelerated by roading or poorly planned harvesting activities. Based on published reports, Victoria s forests are relatively free of significant wasting hazards but it is not an issue to be ignored. It has been noted in the Otway and Strzelecki Ranges. Surface Erosion. Surface erosion involves the detachment and transport of soil particles by raindrop impact and flowing water. Flow or runoff, occurs when water accumulates on the soil surface as a result of rainfall rate exceeding the infiltration rate. It has been argued that undisturbed forest floors are virtually erosion proof as infiltration rates far exceed usual rainfall rates, though this is not always the case. This situation changes dramatically after wildfire with loss of litter and possible development of hydrophobicity. Surface erosion occurs in forests not disturbed by human activities, albeit at low levels, and commonly occurs as stochastic events with low frequency and short duration initiated by some form of natural disturbance such as wildfire, mass wasting, or animal perturbation. This low level is often called the geological erosion rate, norm or benchmark (Lacey 1993). Reported 127

137 background erosion rates for Australian native forests show losses of less than 0.05 tonne/ha/yr (Table 8.1). Patric (1976) reported on 18 studies in the United States where background rates were between and 0.71 tonne/ha/yr. Those reported in Australia fall within this range. It is important to note that soil erosion is measured and reported in several different ways in relation to the specific issues being addressed. Results therefore need to be interpreted in relation to the system of assessment: The quantity of soil lost from an eroding area. The estimation of actual soil lost from specific areas provides an indication of potential degradation and relative erosion rates resulting from different activities. The quantity of sediment actually reaching streams. Not all eroded soil actually enters stream channels, due to deposition. The sediment delivery to streams depends on the proximity of the eroding area and riparian protection measures. Impacts on water quality. Water quality, especially suspended sediment, is measured directly and relates to impacts on aquatic systems and provides generalised erosion information. This is the main approach used in many Australian forest studies. Table 8.1 Comparison of sediment loss from undisturbed and harvested native forest research catchments. Buffer strips were intact in each case and there was no effect from roading. Costin et al. (1960) used soil traps and included mature and regrowth forest. Geology Forest Type Reference Rainfall in the Period (mm) Sediment Increase in Loss in Runoff (tonne/ha/yr) Sediment Loss due to Harvesting (tonne/ha) Adamellite Silvertop Ash Granite Silvertop Ash Granite HEMS Sediments LEMS 1,690 Basalt Alpine Ash Quartz diorite Dacite n.a (0.139 fire effect) n.a , n.a. 5. Alpine Ash 1, n.a. 6. Mountain Ash 1, Lacey (1993) 2. Olive and Rieger (1987) 3. Costin et al. (1960) 4. Hopmans et al. (1993) 5. B. George (Forests NSW), pers. comm. 6. Lane et al. (2009) 7. J. Dayson pers. comm. 128

138 Erosion most readily occurs on soils bared by roading, timber harvesting or site preparation and essentially involves unsealed roads, snig tracks, log landings and general harvest areas (Croke et al. 1999a, Sheridan and Noske 2007b). The quantity of sediment derived from harvesting operations is usually small (Table 8.1) especially when compared with effects of roads (Table 8.3) or other land uses. Erosion is greatly influenced within harvested areas by the patterns and type of disturbance. For example, ground skidding results in larger areas of disturbed soil than those associated with cable logging. However, the irregular pattern of logging leaves a mosaic of disturbed and undisturbed soil with varying degrees of erosion hazard and in addition to the undisturbed areas, creates slash piles and mounds of earth which may trap sediment. Primary sources of sediment in forestry operations are roads and log landings. Whether the sediment moves into the logging or vegetated areas or into creeks depends on the connectivity and location of the roading system. While the design of roads is one critical element, the systems protecting drainage systems are also important. Guidelines for widths of protection are outlined in Codes of Practice (for example, DNRE 1996) as shown below in Table 8.2, however, the basis from research for such widths (either as narrow or wide) is not evident in many cases nor is the primary objective (the total limitation of sediment movement or restriction of nutrient movement). Proposed widths of buffer strips for water quality protection vary widely. Theoretically, it is expected that the width would vary according to environmental conditions and topographical formation but this has not been evaluated. The relative effectiveness of the various streamside protection schemes is difficult to evaluate due to the variation in the environments, management and harvesting practices, and other site specific factors pertaining to research sites. A review by Castelle et al. (1994) found an undisturbed buffer of at least 15 metres was required under most conditions for adequate stream and wetland protection. This review considered forested and non-forested cases. Grayson et al. (1995) considered a 20 metre undisturbed buffer to be best practice, and Clinnick (1985) recommended 30 metres. Table 8.2 Minimum widths (m) for buffer strips (B) and filter strips (F) to be applied to various waterway categories in relation to water quality risk and slope. Waterway Class Sites with Low or Moderate Water Quality Risk Slope 0-30º Sites with High or Very High Water Quality Risk Slope 0-20º Slope 21-30º Pools, permanent streams & wetlands 20B 30B 40B Temporary streams 10F 10B + 10F 20B Drainage lines 10F 10F 15F Source: DSE (2007a) 129

139 Research has clearly shown that roading in some cases can be a major cause of stream sedimentation. Roads and tracks pose the greatest threat to water quality because they (1) are a ready sediment supply, (2) are highly compacted surfaces which generate high rates of runoff, and (3) have the potential to connect directly to stream networks when they are sited close to streams or where they cross streams. It is this connectivity that may lead to deleterious effects on water quality. Few studies have investigated the contribution of roading to catchment scale sediment and nutrient budgets. An experiment initiated in the Upper Tyers catchment in the Central Highlands of Victoria, aimed at measuring sediment and nutrient generation rates, erosion processes and catchment impacts (Sheridan and Noske 2006, 2007a, b, Sheridan et al. 2006). This catchment contains approximately 100 stream crossings but much of the road network is well buffered from the stream network. The study revealed that only 3.3% of the total sediment load and 1.4% of the total phosphorus load was derived from roads. The location of roads to maximise buffering potential is crucial for water quality protection. Tools have been developed to aid the location and drainage of roads to minimise runoff and sediment production and optimise buffering potential (for example, Croke and Mockler 2001, Hairsine et al. 2002, Sheridan and Noske 2005, Lane et al. 2006, Takken et al. 2008a, b). Stream crossings are particularly vulnerable and if poorly designed and maintained, will often constitute a significant point-source of sediment. Lane and Sheridan (2002) reported on a study conducted at a newly constructed, unsealed road and stream crossing (first-order cobble bed permanent stream) in a native forest catchment in the Central Highlands of Victoria with deep gradational soil. Continuous measurements of turbidity and estimates of total suspended sediments were taken upstream and downstream of the crossing (culvert) over a 5-month period. Significant decreases in downstream water quality were found, particularly after rainfall events, with the principle sediment sources being fill slope and the road verge which had not been metalled. The authors concluded that simple measures could be taken to reduce sediment delivery from these pointsources. The effect of traffic volume and road water-status (water content of the road) on water quality from forest roads in high rainfall Mountain Ash forests near Marysville in Victoria was investigated by Sheridan and Noske (2006). Three sections of metalled forest road were monitored for low and high levels of logging truck/quarry truck traffic during wet winter conditions and dry summer conditions. The median sediment concentration increased nearly four-fold when comparing low and high trucktraffic conditions. The water status of the road at the time of traffic did not affect the quality of water discharging from the road but the results were conditional upon the roads being of sufficient quality and being adequately maintained so that truck-traffic did not compromise the lateral drainage of the road profile. Sediment derived from roads may be from a number of components such as road surface, cut batter, fill batter and table drain, all of which are part of the road right of way. Soil loss may be measured from the entire right of way or individual components. Systems for reporting of erosion loss from roads are not standard, but some estimates have been made to allow comparison with general harvesting areas (Table 8.3). Croke et al. (1999c) reported unsealed roads produced an order of magnitude more sediment than snig tracks and they were an order more than the general harvesting area. While the sediment losses from sections of roads may be orders of magnitude greater than the undisturbed situation, the effect from the whole catchment may increase only marginally. The measured level of impact of roads depends on the size of the catchment used for measurement. Issues for roads and sediment loss are: 130

140 Topography and terrain position. Soil types. Croke et al. (1999d) reported that soils derived from granite produced 5 times more sediment than those derived from sedimentary material. Road construction and drainage techniques. Surface stability (gravelled roads suffer considerably lower soil loss). Time. Erosion is greatest in the first few years after road construction. Soil loss increases with level of traffic, and Level of road maintenance. Table 8.3 Sediment losses associated with roading in forests. Location / Reference Component Sediment Loss Road surfaces Melbourne catchments (Haydon et al. 1991) Gravel roads t/ha of road/yr East Victoria (Sheridan and Noske 2004) Unsealed roads 4-37 t/ha of road/yr Eden NSW (Chalmers 1979) Logging road on adamellite 1032 t/ha of road/yr Batters Karuah NSW Cut batters on sediments t/yr /m width (Riley 1988) Fill batters on sediments t/yr/m width It is reasonable to conclude that the dominant sources of sediment from newly constructed roads are the cut and fill batters whereas for established roads, it is the road surface. Good design (particularly at stream crossings) and regular maintenance are essential to avoid water quality problems from roads. Minimisation of sediment movement from roads into streams is achieved by effective planning of the road location (as noted above) and development of effective filter strips, especially those containing obstructions on the forest floor to slow down water movement. Croke et al. (1999a, b) reported that movement of 80 to 90% of sediment was stopped in a 10-metre buffer strip. Little systematic work has been undertaken on the effectiveness of different buffer/filter strip widths and their effectiveness on different soils types and topographies, probably leading to a more conservative approach being applied in setting minimum buffer widths. Their studies found that snig tracks showed no significant recovery over the five years since construction, however hydraulic conductivity increased and sediment-production rates declined over five years. The sediment production varied greatly with soil type. 131

141 1.2 Soil Loss (kg/m2) Age (years) Red granite soil Figure 8.3 Light granite soil Metasediment Sediment loss from snig tracks on three soil types (after Croke et al. 1999a, 2001). Soil movement is rarely assessed within a coupe. Mackay et al. (1985) used fixed pin studies to monitor soil surface change over time in harvested coupes of the Yambulla hydrology project to study small scale soil movement patterns. The results demonstrated that both erosion and deposition patterns were occurring, with erosion in the logged areas immediately after harvesting and slash burning and deposition becoming more significant from about the third year after disturbance. Procedures for assessing soil erosion in Australian forestry (that is, to evaluate land in terms of its propensity to erode under various land use treatments) have recently been thoroughly reviewed (Ryan et al. 2003). They noted that erosion hazard is a multi-factor concept comprising energy source (for example, wind, rain), terrain setting (that is, landform), resistance force (for example, vegetative cover) and management practices (for example, contour ripping). They also reviewed current hazard systems used in native forests and plantations across Australia from these perspectives, and found that the systems varied from rudimentary to sophisticated, depending on individual forest management organisations. It was concluded that a single system for assessing erosion hazard is neither sensible on scientific grounds nor efficient from a cost viewpoint. However, the researchers presented a carefully considered and well-argued strategy for improving current processes, with initial priority for climatic stratification, regolith stability classification, conservation of soil vegetative cover, effective measurement of soil erodibility, adoption of BMP prescriptions, and effective use of erosion models. Nonetheless, Sheridan and Noske (2006) considered that existing erosion hazard assessment schemes could be further improved by assessing the dominant properties of a coupe that will drive impacts on water quality (that is, the extent, arrangement and connection of snig tracks with streams and other compacted areas). It is a basic requirement of most certification schemes (for example, AFS 2007) for a candidate forest management organisation seeking certification to have a formal erosion hazard assessment system in place and to adhere to Code of Practice requirements for protection of water quality. 132

142 8.2.2 Compaction Compaction and redistribution of soils, primarily as a result of harvesting operations is a significant issue. The amended Montreal Process Indicator (by MIG) is Indicator 4.1e Proportion of harvested forest area with significant change in bulk density of any horizon of the surface (0-30 cm) soil. The rationale of the Indicator was to obtain a broad measure of soil physical properties important for soil fertility and hydrological processes. Assessment of compaction and disturbance may be undertaken in a number of ways but it is very expensive, especially for extensive areas. Large increases in soil bulk density have been shown to be confined to areas of high traffic density. Lacey et al. (2002) conducted studies in NSW (Silvertop Ash) and Victorian Mountain Ash forests. In NSW, retrospective studies were undertaken on low productivity sites to assess: (1) the effects of logging on soil properties, and (2) tree regeneration and soil strength in relation to disturbance. In the logging study, about 20% of each coupe was occupied by snig tracks, and there was very heavy disturbance on about 5 to 10% of the coupe. Only the most heavily disturbed classes showed evidence of significant increase in bulk density or soil strength (access tracks, log landings or major snig tracks). Matching traffic intensity with disturbance classes and soil physical properties did not reveal clear relationships. Effects of disturbance on regeneration growth were highly variable, but reduction occurred on snig tracks with evidence of a positive edge effect (Table 8.4). At the coupe level, there was a relationship between the level of disturbance and tree growth. The studies in Victoria assessed factors affecting degree of disturbance in 20 clearfelled coupes. The extent of operational categories of disturbance appeared to be independent of operational factors but was related to the volume of timber removed. Primary snig tracks had higher bulk density and lower organic matter content. Some differences were found between the three transect methods used by Lacey et al. (2002) to assess degree of disturbance, however, estimates of various disturbance classes were relatively insensitive to sample size (a lower sampling intensity could have therefore been used). In the Warra study in Tasmania, 3 3 the surface soils were compacted from 0.58 Mg/m to 0.70 Mg/m although no reports on subsequent growth were available (Pennington et al. 2001). 133

143 3 Table 8.4 The effects of disturbance on the total volume of eucalypt growth (m /ha). The percentage of growth compared with the control is shown in parentheses. Project Control Track Track Edge Logged (undisturbed) Silvertop Ash (Lacey et al. 2002) - Five largest trees at 8 years of age Log landing/ snig track (38%) 10 (83%) Access road (25%) 15.5 (111%) Major snig track (88%) 16 (128%) Minor snig track (87%) 38 (253%) Mountain Ash (Rab and Kelly 2002) Control - All trees at 10 years of age 22.5 Primary snig track Secondary snig track Tertiary snig track Log landing 14.5 (64%) 9.0 (40%) 7.5 (33%) 2 (9%) 7 (31%) 9 (40%) 7 (31%) Messmate (Pennington and Laffan 2001) - All trees 17 to 23 years of age Site (18%) 315 (129%) Site (43%) 375 (179%) Site (42%) 210 (210%) Nutrient Loss and Productive Capacity Strong associations exist between forest types and site characteristics (for example, Turner et al. 1979, Grant et al. 1995), site being defined here as the interaction between climate, topography, soil parent material and soil characteristics. At the broadest level, climate is the major determinant of forest type whereas at a smaller scale, topography and soils have greater significance. Relationships can be found between soil characteristics and the presence or absence of species and their growth rates, however, the key determining factors for each species will vary. Species with greater demands are found on better soils and this is reflected in growth rates and the nutrient status of plant tissues. Nutrients are lost from the site through a number of processes including erosion, fire and harvesting. While in absolute terms the quantities of nutrients removed in harvesting differ between sites, they are proportionally similar across most sites, however, results obtained for one forest site type cannot necessarily be applied to other site types. The soil/site characteristics of forests vary greatly and a selection that demonstrates the level of variation from published material is shown in Table 8.5. In addition, there is a wide range of soil rooting depths, textures and other characteristics which are important in any analysis. The quantities of nutrients available in the soil for tree growth are affected by a number of factors. A key issue is whether management activities (for example, harvesting or burning) can or will lead to a reduction in soil nutrient status and whether this may affect long-term health, productivity or other ecosystem processes. The typical study approach to this long-term issue has been to estimate the quantities of nutrients in the system together with nutrient fluxes and then use a simple input/output model over a number of rotations. Quantities of nutrients will include total and available nutrients in the soil and nutrients contained in each component of biomass. Inputs usually include precipitation inputs and nitrogen fixation while losses include those in runoff water, harvesting and due to fire. 134

144 A number of studies relevant to Victoria have been undertaken estimating tree biomass nutrient contents together with soil nutrients and movements of nutrients (Stewart et al. 1979, 1985, 1990, Attiwill 1980, Feller 1980, O Shaughnessy et al.1981, Baker and Attiwill 1985, Bek 1985, Turner and Lambert 1986, Hopmans et al. 1987, Mackay and Robinson 1987, Adams and Attiwill 1988, 1991, Ryan et al. 1989, together with data from the authors of this current report). There were differences in nutrient contents between forest types but they were reasonably consistent within a forest type (recognising differences in biomass). For example, for mature forests the above-ground phosphorus contents (in kg/ha) were 20 for Silvertop Ash (LEMS), 20 to 30 for Messmate and Cut-tail (HEMS), 48 for Alpine Ash and 49 for Mountain Ash (see Figure 8.2). The estimates for calcium (kg/ha) were 320 to 440 for Silvertop Ash (LEMS), 200 to 500 for Messmate and Cut-tail (HEMS), 1,300 for Alpine Ash and 1,020 for Mountain Ash. Table 8.5 Selection of soil surface characteristics reported from different forest types. Forest Type Rainfall (mm) Total Exchangeable C N C/N P Ca Mg K (%) (%) (mg/kg) (cmol+/kg) a Mountain Ash 1750 b Mountain Ash 1660 c Mountain Ash 1700 Parent Material ph Granite 5.02 Quartz dacite Granodiorite Basalt Granite b Quartz dacite Granite 4.65 Granite Basalt j Sediments Adamellite Sediments Rhyolite Sandstone Sandstone d Alpine Ash d Alpine Ash HEMS d HEMS g HEMS c HEMS LEMS f LEMS g LEMS h LEMS j LEMS Box-Ironbark k k River Red Gum 770 Colluvium a Goldsworthy (1975) g Turner and Kelly (1977) b Feller (1980) h Kelly and Turner (1978) c Sibley (1975) I Stewart and Flinn (1985) d Ryan et al. (1996) j Romanya et al. (1994) e f Ryan et al. (1989) Hopmans et al. (1993) k Murray and Mitchell (1962) 135

145 3500 Phosphorus (kg/ha) Mountain Ash Alpine Ash Soil reserve HEMS Messmate Soil available HEMS - Cuttail LEMS Silvertop Ash Biomass Figure 8.2. Quantities of phosphorus in soil reserves, available in soil and within the biomass for four forest types in Victoria. The biomass is only a very small proportion of the ecosystem total, but the absolute quantities vary according to forest type. Quantities of nutrients within the soil varied greatly but there were identifiable relationships with forest type. As an example of the variation when measured to a depth of one metre in the soil, nitrogen ranged from 250 kg/ha to 17,200 kg/ha, total phosphorus from 351 kg/ha to more than 4,400 kg/ha while exchangeable potassium ranged from 120 kg/ha to more than 1,540 kg/ha. Such variation makes generalisations difficult without being able to include analysis of stand and site conditions that is beyond the present Review Brief. Inputs, primarily in precipitation, and outputs of elements in runoff water (kg/ha) are available for a range of undisturbed forest types. Rainfall has been considered extensively in eastern Australia (Turner et al. 1996) but the number of estimates of losses, mainly from small catchment studies are much more limited in number (Feller 1981, Stewart and Flinn 1985, Hopmans et al. 1987, Mackay and Robinson 1987). Analyses of runoff from undisturbed forests indicated: Losses of phosphorus and cations are generally low both in concentrations and loadings and are largely determined according to geology; Sodium, chloride and boron inputs are correlated with distance to the coast; There are differences in runoff nutrients between forest types (related in part to geology); There is considerable annual variation in runoff water nutrient concentrations and loadings; Most nutrient losses are in close balance with inputs. Nitrogen is very variable between forest types and most losses (LEMS, HEMS) are low. Mountain Ash and Alpine Ash appear to be saturated with a high total loss and an imbalance of losses (excluding fixation); and Water quality baselines in terms of individual nutrients and suspended sediment (see later for water quality) will need to be established on a forest site type basis. 136

146 Nutrient distributions and balances were reviewed for five main forest types, namely Mountain Ash, Alpine Ash, HEMS (Messmate and Cut-tail) and LEMS by assessing estimated nutrient inputs and outputs over the rotation (see Figures 8.2, 8.3, 8.5 and 8.6). The quantities of nutrients contained within the different forest types varied greatly. The data for the Alpine Ash stand was compiled from Bago State Forest in NSW growing on basalts and as such may have been more fertile than those typically found in Victoria. In the case of phosphorus (Figure 8.1), the largest quantities are in the soil reserves and the biomass is a very small proportion of the total. Phosphorus is probably in balance over a full rotation with very low losses of this nutrient as a result of harvesting, particularly due to low phosphorus concentrations in the wood. The nitrogen inputs and outputs showed that at the end of an 80-year cycle in Mountain Ash and Alpine Ash forests, there may be a deficit, but this is primarily due to higher levels of nitrogen in runoff water because the soils probably have high nitrogen saturation (as evidenced by the low soil C/N ratio). The other systems were close to nitrogen balance. Issues arise with elements such as calcium where there are significantly higher removals over the rotation and a large difference between species. The differences are reflected in the water quality. Due to the leakage of such a nutrient as nitrogen, the use of water quality parameters to set a baseline for mature forests, is probably not the most appropriate approach. In a Mountain Ash catchment, Feller (1981) reported an average loss of 2.6 kg N/ha/yr. This is relatively high considering most forests are in the range of kg N/ha/yr (Young et al. 1996) Nitrogen (kg/ha/rotation) Mountain Ash Alpine Ash HEMS Messmate HEMS - Cuttail LEMS - Silvertop Ash Rotation inputs Rotation losses Soil available Net loss Biomass Figure 8.3. The quantities of nitrogen in soil and biomass together with estimated nutrient inputs and losses over a rotation (80 years). The Mountain Ash stand had a net loss over the rotation but this was not apparent in the other forest types. Regular burning will lead to some nutrient losses, especially nitrogen, but the evidence in the longer term is that levels of nitrogen are kept stable by regular burning whereas the unburnt forests accumulate nitrogen, in some cases causing nutrient imbalances (Jurskis and Turner 2002, Turner et al. 2008) and long-term forest health problems. In other systems, it may lead to nitrogen saturation and potential nitrogen leakage (Figure 8.2) into the aquatic system. Nutrient losses have been noted in both direct volatilisation and particulate matter (in smoke) and there are additional losses in 137

147 leaching and sediment movement. Some estimates of leaching and sediment losses have been made in small catchment studies but in general, while the wildfire losses are recognised as high, they have been difficult to quantify (Mackay and Cornish 1982, Mackay and Robinson 1987). After the 2003 wildfires, Lane et al. (2008) and Noske et al. (2010) found that nitrogen and phosphorus exports from catchments with severely burned Alpine Ash, increased 5-6 fold in the first year to 1.6 kg P/ha and 15.3 kg N/ha. The particulate load as opposed to the soluble losses, were found to be most important. Losses declined rapidly as the catchment re-vegetated. The nutrient losses from wildfire are high and while there will be general recovery, an increased frequency of wildfire will lead to significant nutrient depletion and long-term nutritional imbalance. Inputs of nitrogen into the system through fixation have generally been assumed to be low (1-5 kg N/ha/yr). However, more recent mass balance studies, at least in the LEMS forests, indicate that in the absence of fire, accumulation of nitrogen is in the range of kg N/ha/yr, the variation being related to the levels of other nutrients in the soil (Turner et al. 2008). It appears that, in the absence of fire, many forest systems are accumulating nitrogen and that where there are periodic, low intensity fires, the nitrogen levels maintain a reasonably constant level. This is abruptly altered where wildfires occur and nitrogen losses are much higher (losses of kg N/ha may be expected). Figure 8.4. Recent wildfire in LEMS forest showing the extent of removal of litter and understorey and some crown loss. It has been calculated that the level of nutrient removals in routine sawlog harvesting from most stands will have minimal impact on forest productivity for at least three to five rotations (for example, Turner and Lambert 1986, Hopmans et al. 1993). The most susceptible nutrient to change will be calcium, this being even more so in some soils with low reserves of this element. However, in analyses of nutrient removals there is no system for establishing critical nutrient removal levels or losses or for identifying potential impacts. Such results are discussed in detail by Turner and Lambert (1986) and Hopmans et al. (1993). In the case of calcium, all three forest types have shown (Figure 8.5) an apparent net loss of calcium over the full rotation, mainly as a result of the harvesting removals. 138