Prospecting for Commercial and Environmental Opportunities for Agroforestry in Australia. Abstract

Transcription

1 122 Prospecting for Commercial and Environmental Opportunities for Agroforestry in Australia Phil Polglase 1, Keryn Paul 1, Charlie Hawkins 1, Anders Siggins 1, James Turner 2, Trevor Booth 1, Debbie Crawford 1, Tom Jovanovic 1, Trevor Hobbs 3, Kimberley Opie 1, Auro Almeida 1, Jenny Carter 1 1. CSIRO Sustainable Ecosystems 2. Scion, NZ 3. SA DWLBC and Future Farm Industries CRC Phil Polglase CSIRO Sustainable Ecosystems PO Box E4008 Kingston ACT Philip.Polglase@csiro.au Abstract This project, funded by the Joint Venture Agroforestry Program, assessed the potential economic outcomes and environmental impacts of new agroforestry and plantation developments across Australia within a spatial framework. In particular, it compared outcomes across: (i) regions, (ii) agroforestry systems, and (iii) time frames (now, and in the future with assumed changes in policy and infrastructure). Ten agroforestry or plantation systems were examined: Sawlog systems Short-rotation systems Carbon plantings Hardwoods Softwoods Pulpwood Bioenergy Integrated Tree Processing Fodder (in-situ) Environmental plantings Hardwood plantations Softwood plantations Mallee plantings The project developed spatial layers for outcomes for productivity, economics and environmental impacts and included the opportunity costs of planting forests compared with the preceding agricultural enterprise. The project has shown how the information and tools developed can be used to prospect for regional opportunities for the various types of agroforestry systems available.

2 123 Introduction The opportunities for large-scale expansion of agroforestry and plantation systems are many and varied and will be driven by the imagination of the diverse range of investors. Environmental benefits would be realised mostly in the lower rainfall zones (for example, mm mean annual rainfall) but a major impediment to expansion of agroforestry in these regions has often been the inability to quantify and present the economic case. Investment in agroforestry, as with any type of investment, often comes down to risk identification and management. Because trees growing in the more marginal environments have been less well tested, the risks to success are relatively greater than in traditional forestry areas. A key challenge is to quantify the expected economic and environmental outcomes from agroforestry plantings but a difficulty is that trees often take many years after being established before their impacts are manifested. Impact assessment therefore needs to be based on best available knowledge and often rely on predictive modelling to extrapolate in time and into regions for which data are presently limited. This project brings together in a spatial framework a range of information related to growth, product options and economics of agroforestry systems. The objectives have been to: (i) (ii) Construct a framework that can be used to prospect for opportunities given the many assumptions in the scenarios we developed. It is not intended to predict where new agroforestry or plantation developments will be established, and Provide a spatial data base that can be used for further scenario development and exploration of potential agroforestry opportunities with different user-defined inputs and assumptions. Methods The general processes and logic followed in the project are outlined in Figure 1. Scenarios: 10 different agroforestry production systems With or without carbon payments Variable or fixed transport distances and costs Water interception + Biodiversity benefit Data interrogation and synthesis: Sum of all positive NAER values across Australia Sum of all positive NAER values within CMA regions Intersection of profitability with environmental benefits Uncertainty and sensitivity analyses + Economic outcome (NAER) Regional priorities for agroforestry investment: Review of CMA plans Surveys of PFDCs, AFG Generic economic spreadsheet Economic model (AER) minus Profit at full equity of agriculture ( opportunity cost ) Data collation for: Costs of establishment, maintenance and harvesting Prices of products Thinning and harvest volumes Transport distances and costs to processing facilities Growth layers 3-PG2 growth model: Calibration and validation Climate, soils layers Growth data collation and compilation Figure 1. Schematic flow diagram of processes and logic followed during the study.

3 124 The project essentially was a large modelling and GIS exercise, including stakeholder engagement to help develop scenarios and assess the attitudes to forestry within regions. It had the following key components: Stakeholder engagement that included: (i) Review of published Investment Plans for the 57 NRM Regions (mostly CMAs) for their (ii) interest in agroforestry, Survey of regional representatives of the AFG and the PFDCs to assess their views on community and industry attitudes to agroforestry in their region, and (iii) Collating information to develop the regional silvicultural scenarios, economic costs of production (establishment, management, harvesting, transport) and product prices. Scenarios for current economic and policy conditions for the agroforestry systems above. Scenarios for potential future markets: (i) Including sequestered carbon as a product, and (ii) Using a fixed transport distance (100 km) to simulate the establishment of new processing mills in areas of prospective plantation expansion. Growth modelling across Australia for species that typically could be grown in various geo-climatic zones and for which data were available for calibration and validation of the 3-PG2 model (Table 1). Economic modelling to calculate spatially the profitability (at mill gate) of agroforestry systems, in terms of Net Annual Equivalent Return (NAER, $ ha -1 yr -1 ) using discounted cash flow and that also included the opportunity cost of the preceding agricultural enterprise. Environmental impacts of agroforestry on: (i) (ii) Biodiversity, and Rainfall interception to identify areas where profitability coincided with biodiversity need and least water impact. Uncertainty and sensitivity (Monte Carlo) analyses to identify aspects of agroforestry to which profitability was sensitive. Results and Discussion One of the most important determinants of profitability is growth rate as this affects the yield of product, whether it is a harvested commodity or sequestered carbon. Figure 2 shows an example data set and model calibration. For all the species, the model calibrations and validations were generally reasonable. Collectively, the data sets begin to constrain expected rates of production. There are obviously large variations in growth rates for any given site, often reflecting specific differences in genetic stock used, soil type, and management impacts such as weed control and nutrition, many of which are difficult to incorporate accurately in predictive models. The aim for the grower is to maximize the rates of production, within a given rainfall zone, towards the upper limit expected and as shown in Figure 2 (a) for example. Following calibration, the 3-PG2 model was run spatially across Australia at 1 km scale using spatial layers for climatic and soil physical properties as inputs. Figure 3 shows the annual productivity surfaces for the hardwood species for annual increments in stem volume and total live carbon as an example. Table 1. Species and number of locations from which growth data were collated for calibration and validation of the 3-PG model. Some of these locations had multiple growth measurements from different times or treatments. Species/forest type Number of calibration sites E. globulus C. maculata/e. cladocalyx E. camaldulensis E. grandis Oil mallees 24 0 Pinus radiata P. pinaster P. caribaea, P. elliottii and their hybrids 14 6 Environmental plantings Number of validation sites

4 125 Observed stem volume (m 3 ha -1 ) 400 (a) y = 0.61x R 2 = 0.21 EF = 0.44 N=216 Predicted DBH (cm) 60 (b) Age (yrs) 20 Predicted 10 DBH (cm) 0 y = 0.94x R 2 = 0.76 EF = 0.79 N= Observed DBH (cm) Figure 2. Data set collated for combined C. maculata/e. cladocalyx plantations for: (a) estimated stem volume, and (b) Calibration of the 3-PG model against measured tree diameters (diameter at breast height, DBH). Hardwood carbon plantings, MAI (m 3 ha -1 yr -1 ) Hardwood carbon plantings, Total carbon (t C ha -1 yr -1 ) Figure 3. Predicted mean annual volume increment and total carbon in total live biomass (above- and below-ground) at 20 years of age. Results are for plantations of E. globulus, E. cladocalyx/c. maculata, E. grandis, and E. camaldulensis, depending on the geoclimatic zone modelled.

5 126 The productivity surfaces were then combined with other spatial data to calculate the potential economic return (net annual equivalent return, NAER) of the forestry enterprises. Figure 4 shows an example for hardwood sawlog systems. Generally, large areas were identified where there were potential opportunities to grow trees for a variety of products. The main conclusions can be summarized as: Agroforestry can be competitive with agriculture in some regions and for some agroforestry or plantation systems, as demonstrated by positive values of NAER (that is, the forestry system is more profitable than the preceding agriculture phase). Transport distances and product price are important in influencing profitability. This is well established but reinforces the fact that large-scale expansion of agroforestry systems will be constrained by distance to existing processing or handling facilities, or new ones will have to be built. Northern Australia and the east coast show promise for expansion of agroforestry systems and industries due to the often low profitability of agriculture and potential fast rates of growth. Dedicated bioenergy and Integrated Tree Processing (ITP) systems are not profitable at present unless they are very close to processing facilities. This is due to the high cost of production (harvesting and transport) relative to low product price for wood energy (market failure at present). Carbon farming appears to be promising due to the relatively low cost of production (no harvesting, transport) relative to a possibly high product price. This indicates that new forests can be grown in many locations and for multiple environmental outcomes. The profitability of harvested forestry systems can be significantly improved if carbon is included as an additional, saleable, product. Maximizing rates of forest growth remains one of the most important determinants of profitability and thus where agroforestry research can have high impact. The information was then further interrogated by running filters over the spatial layers for economic outcomes and environmental impacts. In one illustrative scenario (Figure 5), possible regions were identified where environmental-type tree plantings might be economically competitive with existing agricultural land use. Areas were restricted to environments where interception of rainfall by the forest was less than 150 mm/yr and where annual return (NAER) from the agroforestry operation was greater than $100 / ha/yr. Assessment was restricted to privately-owned, cleared, agricultural land. Under this scenario, the main opportunities were confined to south-eastern Australia (west of the Great Dividing Range and extending through Victoria and NSW up to the Queensland border), southern and south-eastern South Australia and parts of Tasmania and south-west Western Australia. A total of 9.1 M ha were identified with a combined value of NAER of $1.9 bill/yr (assuming a carbon payment of $20 /t CO 2 -e). The annual rate of carbon sequestration was also projected to be 39 million t C/yr at an average areal rate of 4.3 t C/ha/yr. This total annual increment is equivalent to about 25% of Australia s net 2005 greenhouse gas emissions. However, rates of carbon accumulation rates in these trees would eventually plateau, meaning they would no longer represent a significant on-going net sink for carbon. Net profitability> $100 ha/yr Water yield reduction: <150 mm/yr Plantable area: 9.1 Mill ha Total profitability: $1.9 Bill/yr Annual carbon sequestered: 39 mill t C/yr $0 to $150 $150 to $500 > $500 Total SQKM 104, ,826 61, ,893 Annual Return 739,200,000 3,259,460,000 5,696,940,000 9,695,600,000 Mean Annual/ha $350 Figure 4. Predicted regions of profitability for the hardwood sawlog systems. In this example it was assumed that only the existing infrastructure for sawlog and residual wood products was accessible and that there was no payment for sequestered carbon. Figure 5. Areas identified as potentially suitable for environmental-type forest plantings in which carbon sequestration (assuming $20 /t CO 2 -e) was competitive with current agricultural returns (calculated profit was $100 /ha/yr more than current agricultural profit at full equity) and rainfall interception was less than 150 mm /yr. The results from this project should be viewed as prospecting for opportunities in which the aim is to identify regions that may meet the criteria of a particular type of investor but which require greater groundtruthing to verify the assumptions used in the model and to assess the local social, economic and policy conditions as they pertain to agroforestry developments. The outputs from each agroforestry scenario represent

6 127 only one set of conditions within each of the broad geoclimatic regions. In reality, the assumptions used may vary greatly and thus affect the calculated profitability. A main output from this report is the spatial data set that has been compiled and underpinned by its various components such as the extensive calibration and independent validation of the growth model for several species. The spatial economic model developed has a user-friendly interface that enables rapid interrogation of the data base according to user-specified scenarios and to adjust the productivity predictions upwards or downwards if local knowledge suggests that the accuracy of the growth model predictions can be improved. Acknowledgments We gratefully acknowledge the help of the following people: The members of the Project Steering Committee and in particular Rosemary Lott for their guidance and management of the project Mark Kelly, Jade Thompson Duncan Macleod (URS forestry) for help with information on the development of scenarios, costs associated with the forestry systems and product prices. David Bush (CSIRO Plant Industry) for advice on species suitability Colin Mues, Walter Shafron and Mily Lubulwa (ABARE) for information relating to spatial statistics and sources of spatial data. Geoff Elklerton (ABS) for advice on the CPI adjustment of the Profit at Full Equity layer Stefan Hajkowicz, Brett Bryan (CSIRO) and Mike Young (University of Adelaide) for advice on the Profit at Full Equity layer Miles Prosser (A3P) and Nick O Brien (New Forests) for advice on potential carbon trading systems and costs David Freudenberger (Greening Australia) for advice on costs associated with biodiversity plantings and for assistance with development of the biodiversity score layer. The many organisations and landholders who contributed data John Bartle and Dan Huxtable (WA CALM and the CRC-FFI) for provision of mallee data for model calibration The many regional PFDC and AFG representatives who responded to the stakeholder surveys and helped define the agroforestry scenarios.