Effects of future crop tree flagging and skid trail planning on conventional diameter-limit logging in a Bolivian tropical forest

Transcription

1 Available online at w w w.sciencedirect.com SCIENCE N C E ^ D I R E C T «ELSEVIER Forest Ecology and Management 188 (2004) Forest Ecology and Management Effects of future crop tree flagging and skid trail planning on conventional diameter-limit logging in a Bolivian tropical forest Werner Krueger* International Resource Management Program, School of Forestry, University of Montana, 32 Campus Drive 0576, Missoula, MT 59812, USA Received 15 May 2003; received in revised form 7 June 2003; accepted 5 August 2003 Abstract Sustainable natural forest management initiatives in the lowland tropical forests of Bolivia require that cost-effective measures be identified to help maintain timber stand quality and reduce damage to advance regeneration during logging operations. This study considered the potential costs and benefits of including two best management practices the pre harvest flagging of future crop trees (FCTs) and improved planning of skid trails in adjacent 45 ha plots, in conjunction with conventional planned diameter-limit logging in a lowland tropical forest in the Department of Santa Cruz, Bolivia. The study also included a general assessment of FCT damage and timber waste associated with standard felling operations. The additional costs of flagging FCTs >30 cm DBH were estimated to be US$ 0.38/ha. FCT flagging significantly reduced damage by 20% in felling gaps and by 10% along skid trails that received 2-10 skidder passes. A mean directional error of ±35 per tree felled was observed for felling operations, indicating that damage to FCTs during felling could be reduced through implementation of directional felling techniques. Improved felling techniques could increase mean timber utilization by as much as 0.16 m^ or US$ 3.20 per tree felled through reduced wood waste associated with stump pull, a common problem observed during felling operations. Pre-harvest skid trail planning increased costs an estimated US$ 1.46/ha, but reduced mean skidder search times (min/m^ harvested) by 23% and unnecessary skid trails by 96%. Compared to normal harvesting operations, overall operational efficiency increased by approximately US$ 3.00/ha. Pre-harvest skid trail planning significantly reduced FCT damage along skid trails that received 2-10 skidder passes. The results of this study indicate that pre-harvest skid trail planning and improved timber felling techniques yield short-term financial gains relative to conventional management as well as potential long-term gains through FCT conservation and, therefore, warrant implementation. Forest certification groups should consider FCT flagging for inclusion in timber certification criteria for the region given its low cost and proven utility for maintaining timber stand quality. (C) 2003 Elsevier B.V. All rights reserved. Keywords: Bolivia; Sustainable forest management; Logging damage; Reduced-impact logging 1. Introduction sustainable forest management in the lowland tropical and subtropical forests of Bolivia began in earnest After decades of low-intensity high-grade logging less than 10 years ago with the adoption of the for one or two choice species, attempts to introduce 1996 forestry law (see ITTO, 1996a; Fredericksen, 2000 for a review). Bolivian law now requires the Present address: Grant Creek Road, Missoula, development of forest management plans that seek MT 59808, USA. Tel.: sustainable timber harvesting on all of Bolivia s address: wernerkusa@yahoo.com (W. Krueger). permanent production forestlands (Fredericksen and /$ - see front matter 2003 Elsevier B.V. All rights reserved. doi: /j.foreco

2 382 W. Krueger/Forest Ecology and Management 188 (2004) Mostacedo, 2000). Adherence to the new law with recovery of tropical logging operations (Gerwig only minor modihcation fulhls the requirements set et ah, 1996; Barreto et ah, 1998; Boltz et ah, 2001), forth by the Forest Stewardship Council (FSC) for thus making the adoption of RIL potentially attractive timber certihcation and thus provides access to pre to commercial forest managers. Specihc RIL practices mium markets (Darby, 1999; Nittler and Nash, 1999; that have been incorporated into Bolivian tropical Fredericksen, 2000). In response, many of Bolivia s forest management include the following: pre-harvest industrial-scale timber producers have shifted man inventory, mapping, and marking of all commercial agement from traditional high-grade logging toward and seed trees; pre-harvest vine cutting on or near the a low-intensity selection cut that more closely re boles of all commercial trees at the time of inventory; sembles sustainable forest management (Putz and pre-harvest planning and demarcation of roads, land Romero, 2001). ings, and primary skid trails; the training of logging Today, Bolivian forest management plans are based crews in RIL techniques; and the use of rubber-tired on the removal of isolated mature trees on a sustained skidders for ground-based yarding (T. Fredericksen, yield basis, with the goal of improving the overall personal communication). commercial value of the forest (Fredericksen et ah, Despite noteworthy advances, damage to FCTs 2001). Minimum diameter-at-breast-height (DBH) during logging is thought to be one of the largest cutting limits guide felling operations, and approximately 20% of trees above the diameter limit are left agement in the neotropics (Fredericksen, 1998) and silvicultural challenges facing sustainable forest man as seed trees to provide for subsequent, natural regen Bolivia (Fredericksen, 2000). Recent data reported by eration (Fredericksen et ah, 2001). Bolivian law Jackson et al. (2002) from the FSC certihed La Chonta requires a minimum 20-year cutting cycle for subse timber concession, for example, indicate that logging quent logging entries (Fredericksen et ah, 2001). It operations kill or severely damage an average of has been suggested that polycyclic selection logging 22 trees greater than 10 cm DBH per tree extracted, systems such as this one can provide a competitive 6 of them commercial species. While some damage is flow of economic benehts while, at the same time, unavoidable during logging operations and accepta protecting the ecological integrity of tropical forests ble levels of damage have yet to be dehned for natural (Anderson, 1990; Gomez-Pampa et ah, 1991; Johnson forest management in Bolivia, these hndings suggest and Carbarle, 1993; Dickinson et ah, 1996; Howard that current RIL practices may not sustain timber et ah, 1996; Whitman et ah, 1997; Jackson et ah, yields through a series of cycles given the low density 2002). of commercial and potentially commercial trees in As with all polycyclic harvesting systems, the Bolivia s production forests (Fredericksen, 2000). sustainability of the Bolivian model is largely depen For this reason, there is a perceived need among dent on advance regeneration. In other words, forests forest managers to identify additional RIL practices must be adequately stocked with well-formed future that can reduce FCT damage in a cost-effective crop trees (FCTs) prior to logging, and residual stand manner. It has been suggested that the marking, damage must be kept acceptably low during logging or flagging, of FCTs and careful planning and demar operations (Gullison and Hardner, 1993; Whitmore, cation of all skid trails prior to logging can help 1998). In keeping with this tenet of polycyclic silvi reduce damage to residual trees (Dykstra and culture, Bolivian forest managers have adopted a Heinrich, 1996), but these practices are not currently number of reduced-impact logging (RIL) practices. employed in most Bolivian forest management The term RIL embraces a host of improved road- operations (T. Fredericksen, personal communica building, felling, and extraction techniques known tion). The main objectives of this investigation were to increase the post-harvest forest value and the to: (1) quantify current harvest costs and benehts long-term potential for sustainable forest management associated with FCT hagging and improved skid trail (Dykstra and Heinrich, 1996; ITTO, 1996b; Johns planning relative to conventional diameter-limit log et ah, 1996; Boltz et ah, 2001; Holmes et ah, 2002; ging, (2) estimate the extent to which FCT damage Putz et ah, 2001). RIL practices have also been shown can be mitigated during logging operations, and (3) to increase the operational efhciency and timber identify cost-effective best management options that

3 This study was conducted in the 100,000 ha La Chonta timber concession located in the Guarayos Forest Reserve (15 45'S, 62 60'W) in the province of Guarayos, Department of Santa Cruz, Bolivia. The site is classihed as tropical humid forest with a mean annual temperature of 25.3 C and mean annual rainfall of approximately 1560 mm (Gil, 1997), the majority of which falls between November and March. Logging activities cease almost entirely during December and January, the peak of the rainy season. The concession was selectively logged for mahogany {Swietenia macwphylla King) and Spanish-cedar {Cedrela odomta L.) during the 1970s and 1980s (Gil, 1997). Over 100 tree species have been identihed on the site, and 18 tree species are currently considered commercially valuable (BOLFOR, 2000). The annual harvest area is approxi mately 2370 ha, with harvest volumes ranging from 7 to 22 m^/ha (BOLFOR, 2000). Subsequent logging entries are scheduled at 30-year intervals (Gil, 1997). The company s management was certihed by the FSC in Comparisons between normal logging operations and two alternative timber harvesting systems were conducted in adjacent 45 ha blocks, all three of which were located within one 852-ha harvest unit (4 km X 2 km) known locally as AAA One main north-south spur road (4 km long) was planned prior to harvest according to contour maps and soil prohles, roughly bisecting the harvest unit. Logs from the three experimental blocks were skidded to this main road, but nowhere did the road enter the blocks. Primary skid trails leaving the main road were con structed at 150 m intervals and ran 1000 m east and west from the road; it was not uncommon for logs to be skidded 1000 m or more in the study. Each experi mental block was served by three primary skid trails. All three blocks had little relief elevation ranged from 240 to 280 m above sea level, and slopes were less than 5%. Harvest volumes ranged from 8 to 10 m^/ha in the three blocks. All harvesting occurred between April and June 2002, and the same sawyers W. Krueger/Forest Ecology and Management 188 (2004) might increase the timber value of post-harvested and skidder operators, using the same logging equip stands. ment, performed the harvesting in all three experi mental blocks. A pre-harvest inventory of FCTs cm DBH was conducted along three 1000 m 2. Site description and research approach transects in each block; FCT density was estimated to be 6, 10, and 12FCTs/ha for Blocks 1, 2, and 3, respectively Normal harvesting operations Block 1 The harvesting system employed in Block 1 was representative of La Chonta s normal management. Harvesting was based on a minimum diameter cutting limit of 50 cm DBH for all species except Hura crepitans L. and Ficus boliviana C.C. Berg, (histori cally referred to as Ficus glabrata Kunth.) for which the minimum diameter limit was 70 cm DBH. Twenty percent of all commercial stems above the diameter limit were left as seed trees for subsequent natural regeneration. Trees were selected, marked, and mapped for harvest approximately 1-year prior to felling using a systematic pre-harvest inventory; vines on or near the boles of trees marked for harvest were cut at this time. Felling operations were conducted by two trained sawyers who operated independently of one another to fell and top trees marked for harvest using Husqvarna 288 chainsaws with 81 cm saw bars. Trees were topped one cut made to separate the bole from the crown at a predetermined merchantable diameter of 40 cm. Both sawyers received formal training in chainsaw use and directional felling techniques during a 5 day workshop sponsored by the Bolivian Sustain able Forest Management Project (Proyecto BOLFOR) in 1997, and both have been employed as sawyers since that time. Their skill level is considered to be similar to other experienced sawyers in the region, and both have provided training to novice sawyers. Typical of logging practices in the region, felling wedges are not used in the concession. Sawyers receive a flat salary plus a bonus based on the total number of logs felled and yarded each month. Two drivers operating Caterpillar 5 18C rubber-tired skidders independently of one another performed all skidding operations in the study. Both operators received training in RIL considerations prior to this study. Tree-length logs were skidded individually or, on rare occasions, in bunches of two or three to - - -

4 384 W. Krueger/Forest Ecology and Management 188 (2004) roadside landings. The rubber-tired skidders were equipped with a winch, arch, and 15 m of cable. Typical of logging practices in the region, the skidder operators used the main line cable to hook logs neither chokers nor choker setters were employed by the concession during the time of this study. Skidder operators are paid a flat salary plus a bonus based on the total number of logs yarded each month FCT flagging Block 2 The FAO Model Code o f Forest Harvesting Practice suggests that damage to residual trees can be reduced by marking those trees that are expected to form part of the FCT population, and thereby provide visual reinforcement for logging crews in the held (Dykstra and Heinrich, 1996). Typical Bolivian forest manage ment, however, does not include FCT marking or inventory. The harvesting system employed in Block 2 included all the operations described in Section 2.1, normal harvesting operations plus the hagging of all FCTs prior to harvest for the purpose of facilitating FCT identihcation during logging operations. FCTs were dehned in this study as trees of those species listed in Table 1 greater than 30 cm DBH with wellformed boles and crowns. FCTs were hagged with a single ring of blue spray paint (the same color used to mark seed trees in the concession) applied approxi mately 170 cm above ground level Improved skid trail planning Block 3 The residual stand damage along secondary and tertiary skid trails and high density of unnecessary skid trails in La Chonta reported by Jackson et al. (2002) is likely the result of limited investment in skid trail planning. While roads and primary skid trails are planned and demarcated prior to harvest, secondary and tertiary skid trails are not. Instead, they are established on the hy by skidder operators to maxi mize the ease of extraction of each log on a case-by- case basis. Sawyers typically give skidder operators a clue as to the location of felled boles often a brief verbal description but the exact location and lay is left largely to conjecture. This hit or miss style of yarding likely results in unnecessary damage to the residual stand, the formation and use of sub-optimal angles between skid trail intersections, and the high Table 1 Species included as FCTs in this study (modified from BOLFOR (2000)) Family and species name Anacardiaceae Spondias mombin L. Apocynaceae Aspidosperma cylindrocarpon Muell. Arg. Bignoniaceae Tabebuia serratifolia (Vahl) G. Nicholson Bombacaceae Ceiba pentandra (L.) R Gaertner Boraginaceae Cordia alliodora Cham. Common name Ocorocillo Jichituriqui Tajibo amarillo Hoja de yuca Picana negra Caesalpiniaceae Caesalpinia pluviosa DC. Momoqui Copaifera chodatiana Hassl. Sirari Hymenaea courbaril L. Paquio S. parahyba (Veil. Cone.) S.F. Blake Serebo Combretaceae Terminalia amazonica (J.F. Gmel.) Exell Verdolago Euphorbiaceae H. crepitans L. Ochoo Eabaceae Centrolobium microchaete (Mart, ex Benth) Lima ex G.P. Lewis Sweetia fruticosa Spreng. Tarara amarilla Mam Lecythidaceae Cariniana estrellensis (Raddi) Kuntze Yesquero negro C. ianeirensis R. Knuth Yesquero bianco Meliaceae Cedrela odorata L. Cedro (Spanish-cedar) Swietenia macrophylla King Mara (mahogany) Mimosaceae Inga edulis Mart. Pacay Moraceae Clarisia racemosa Ruiz & Pavon Murure F. boliviana C.C. Berg. Bibosi Colorado Pseudolmedia laevis (Ruiz & Pavon) Ojoso Colorado J.E. Macbr. Rubiaceae Calycophyllum spruceanum (Benth) Guayabochi Hook f. ex Schumann Sapotaceae Pouteria nemorosa Baehni Coquino Ulmaceae Ampelocera ruizii Klotzsch Blanquillo

5 W. Krueger/Forest Ecology and Management 188 (2004) density of superfluous skid trails reported in the con cession by Jackson et al. (2002). In an attempt to maximize the efhciency of labor, machine, and forest resource use, the harvesting sys tem employed in Block 3 included all operations described in Section 2.1, normal harvesting opera tions, plus the planning of all secondary and tertiary skid trails. Following felling operations, but prior to log skidding, a forest laborer demarcated optimal log extraction routes using orange hagging tape. When necessary, the forest laborer had sawyers cut down- and-dead logs into sections to facilitate skidder access. The forest laborer was instructed to make straight skid trails that intersected one another at obtuse angles and to avoid well-formed advance regeneration when possible (but FCTs were not marked prior to harvest). 3. Methods 3.1. General assessment o f felling technique The implementation of directional felling requires that sawyers are fully capable of felling trees in any predetermined direction (ITTO, 1996c). To assess the technical ability of the concession s timber fallers, both were observed in the felling of approximately 40 trees selected at random. They were asked to indicate the specihc direction in which they intended to fell the tree, and that azimuth was recorded using a compass prior to the tree s fall. After felling, the azimuth of the tree s actual fall was recorded and notes were taken at the stump regarding the placement and dimensions of the face cut, back cut, and holding wood (the cut and uncut sections of wood that control the speed and direction of a tree s fall). The difference between the intended and actual direction of the tree s fall was then calculated for each observation, and an average margin of felling error was calculated for each sawyer, diameter class, and species observed Assessing FCT damage associated with felling operations Approximately 30 single tree fall sites were selected at random from both the normal harvesting and FCT flagging blocks, and the FCT damage incurred at each site was tallied (methods modihed from Johns et al. (1996)). Damage to FCTs was recorded according to the cause of damage (i.e., fell ing or skidding) and the location and severity of the wound (Table 2, modihed from Jackson et al. (2002)). F-tests were used to test the homogeneity of variances, and equal variance f-tests were used to test for differ ences in FCT damage between harvesting systems. Differences were considered statistically signihcant at P < Assessing FCT damage associated with skidding operations All skid trails in each of the three 45-ha experi mental blocks were mapped and delineated into four classihcations: (1) primary skid trails, where more than 10 trees had been skidded, (2) secondary skid trails, where 2-10 trees had been skidded, (3) tertiary skid trails, where only one tree had been skidded, and (4) other skid trails, where skidders had traveled but no logs had been skidded i.e. dead ends and low-inten sity access routes (from Jackson et al. (2002)). The lengths of individual skid trails were measured, and FCT damage was tallied along the entire length of all primary, secondary, tertiary, and other skid trails according to the cause of damage (i.e., felling or Table 2 Classification of damage sustained by FCTs along skid trails and in felling gaps (modified from Jackson et al. (2002)) Damage type Bole Root Crown Severe Moderate Minor Snapped at base, bent, or severely leaning Exposed and damaged cambial tissue Exposed cambial tissue but no damage, bark scrape Uprooted Exposed and damaged cambial tissue Exposed cambial tissue but no damage, root scrape Loss of entire crown; loss of less than entire but more than two-thirds of crown Loss of less than two-thirds but more than one-third of crown Loss of less than one-third of crown

6 386 W. Krueger/Forest Ecology and Management 188 (2004) skidding) and the location and severity of the wound (Table 2, modihed from Jackson et al. (2002)). To adjust for differences in pre-harvest FCT density among stands and individual skid trail lengths, the percentage of FCTs damaged per unit length of skid trail constructed was calculated. Differences between treatments were tested for signihcance using Kruskal - Wallis tests and, when appropriate, subsequent f-tests. Differences were considered statistically signihcant at P < assess the quality of all logs left in the woods. Logs left behind were scaled following standard practices used in the concession Additional costs o f FCT flagging and improved skid trail planning Prior to logging, the cost of FCT hagging was estimated by summing the costs of labor and paint per unit area marked for harvest ($/ha). During log ging, the average cost of improved skid trail planning 3.4. Skidding efficiency was estimated by summing the costs of labor and hagging tape per unit area logged ($/ha). Mean cycle times and production rates (excluding delays) in m^ yarded per hour of machine operation were estimated for both normal and improved skid trail planning operations by measuring log 4. Results and discussion volume and round trip times and distances from land 4.1. General assessment o f felling operations ings to felled boles over a 5-day period (methods modihed from Stokes et al. (1993), Barreto et al. (1998)). Mean cycle times were broken down into component parts as follows: search time was dehned as the time required for empty skidders to travel from the landing to the felled bole in the woods; maneuver ing time was the time required by skidder operators to push and maneuver logs in an attempt to elevate the butt end prior to hooking; hook time was the time spent by skidder operators attaching the main-line cable to logs; and haul time was the time required to haul logs from the stump zone to the primary skid trail. In order to make comparisons between treat ments, mean cycle times and their component parts were adjusted for a standard travel distance. Mean volume hauled per skidder cycle was 4.49 m^ in Block A mean tree felling error of ±35.2 (S.E. = 4.1, n = 89) was observed for felling operations in La Chonta (mean DBH = 96 cm for trees in this study). Thus, concession sawyers are not capable of felling trees in precise predetermined directions. Student s f tests revealed no signihcant difference in mean felling error between the two timber fallers employed. Mean felling error increased with each DBH class observed, but regression analysis showed that felling error was not strongly correlated to DBH (R^ = 0.24; Fig. 1). Mean felling error was also found to vary by species (Fig. 2). Mean felling error associated with the three most commonly harvested species in La Chonta was found to be 30.3 for Cariniana ianeirensis R. Knuth (S.E. = 4.6, n = 45), 52.6 for//, crepitans 1 and 4.41 m^ in Block 3; thus, it was assumed that log volume did not contribute to observed differences (S.E. = (S.E. = 13.8, n = 8), and 75.9 for F. boliviana 20.2, n = 7). The exceptionally large mean between treatments in mean cycle times. F-tests were employed to test the assumption of homogeneity of variances, and equal- or unequal-variance f-tests were felling error associated with H. crepitans and F. boliviana is likely due to the larger diameters of these trees and their abnormally large crowns, some used to determine the signihcance of differences times 1600 per tree (Jackson et al., 2002), which observed between treatments in mean cycle times and their components. Differences were considered statistically signihcant at P < To assess the timber recovery of skidding opera tions, a complete post-harvest inventory was con ducted in both the normal and improved skid trail planning harvest blocks. An experienced log grader and timber scaler employed by the concession helped make it difhcult for sawyers to gauge the natural weight distribution and lean of these trees. The most common error in timber falling technique observed during the study involved mismatched cuts on the face cut (Eig. 3). Typically, horizontal cuts were sawn well past sloping cuts, effectively establishing a face cut within the face cut, a common error in timber falling described by Dent (1974). As a tree begins to

7 W. Krueger/Forest Ecology and Management 188 (2004) Ul O) 0* o species # C. ianei F. i>o(ia X H. c r s f l! other ^ S. pafah Linear R egression with 95.00% Mean Prediction interval 01 ca _E Stump diameter (cm) Fig. 1. Felling error as a function of stump diameter Error Bars show 95.0% Cl of Mean in 0) 0) 03 "O V 03 c C. ianeirensis F. boliviana H. crepitans other species s.parahyba Fig. 2. Felling error by species groups.

8 W. Krueger/ Forest Ecology and Management 188 (2004) / i Mismatched horizontaland sloping cuts Concentration of pressure in holding wood Stump pull Fig. 3. The most common error observed in timber falling technique. move in the direction of fall, the horizontal cut (which is only the width of the saw kerf) closes immediately, causing the committed tree to momentarily hesitate. This creates an extreme concentration of pressure in the holding wood (Fig. 3) that causes the tree to jump from the stump or, in extreme cases, to split vertically up the bole (the so-called barber-chair effect). In either case, the result is a loss of control over the felling direction of the tree and unnecessary wood waste in the form of stump pull (Fig. 3) large slivers of wood that are pulled from the butt of the falling tree and remain attached to the stump (as dehned by Conway, 1973). A random sample of 28 C. ianeirensis stumps evaluated in the normal harvesting block revealed a mean volume of 0.16 m^ (S.E. = 0.04 m^) of wood wasted per tree due to stump pull (deductions were estimated by calculating the area of an imaginary cylinder encompassing the defect). Assuming an aver age log volume of 3.66 m^ for C. ianeirensis (Empresa Agroindustrial La Chonta Ltda., unpublished data), waste due to stump pull may account for as much as 4.5% of total log volume, or nearly 13% of the total waste associated with log processing operations for this species at the sawmill (Harausz, 2002). The hawed timber falling technique observed in this study is likely the result of inadequate equipment and an incentive program that rewards rapid sawing over careful sawing. Sawyers make excessively deep hor izontal cuts, usually 1/3-1/2 the tree s diameter, to ensure the tree will commit to the general felling direction they intend. In the absence of felling wedges or hydraulic tree jacks, this effectively prevents the tree s weight from sitting down on the saw bar when the back cut is made. Sloping cuts made in proportion to excessively deep horizontal cuts on large diameter trees, however, require arduous sawing when only 81 cm saw bars are used. Sawyers may hnd it more expedient to make one shallow disproportionate sloping cut because they are provided an incentive based on the total number of logs felled with no deductions made for stump pull, breakage, or damage to ECTs incurred during felling operations FCTs killed by felling operations Despite the inability of sawyers to accurately and consistently fell trees in the desired directions during this study, severe damage to ECTs due to felling was signihcantly reduced in the ECT flagging harvest block compared to the normal harvest block (Student s f-test; P = 0.015). Normal felling operations killed or severely wounded nearly 30% of all ECTs present, while felling with ECT flagging killed or severely wounded only 10% of all ECTs present. Assuming a mean density of 4-6 ECTs/ha in the concession (Gil, 1997), these results suggest that the flagging of ECTs prior to logging could save 1 ECT/ha and more than 2000 ECTs each year in the concession FCT damage, operational efficiency, and timber recovery o f skidding operations Relative to normal harvesting operations, both ECT flagging and improved skid trail planning signihcantly reduced ECT damage along secondary skid trails (Table 3). This may have important implications for management in La Chonta, given the high level of residual stand damage reported along secondary skid trails by Jackson et al. (2002). ECT damage along

9 W. Krueger/Forest Ecology and Management 188 (2004) Table 3 Comparisons of the percentage of FCTs moderately or severely damaged per 100 m of skid trail constructed, by treatment type and skid trail classification Skid trail classification Block 1 (normal harvesting) Block 2 (FCT flagging) Block 3 (improved skid trail planning) Tertiary skid trails 31.0 a (±14.3, n = 55) 13.3 a (±9.3, n = 51) 22.2 a (±18.2, n = 47) Secondary skid trails 10.9 a (±4.3, n = 23) 0.8 b (±0.8, n = 16) 0.0 b (±0.0, n = 13) Primary skid trails 8.9 a (±5.2, n = 3) 2.8 a (±2.2, n = 3) Not applicable Other skid trails 0.0 a (±0.0, n = 16) 0.0 a (±0.0, n = 15) Not applicable Mean percentages are presented with S.E. and number of replicates (individual skid trails) in parentheses. Means with the same letter within rows are not signihcantly different (Kruskal-Wallis test, P > 0.05). Primary skid trails were not planned differently in the improved skid trail planning harvest block compared to the normal harvest block. Only one skid trail classihed as other was created in the improved skid trail planning harvest block. tertiary and primary skid trails, however, was not significantly reduced by either the FCT flagging or improved skid trail planning treatments (Table 3), suggesting that, as long as tree-length logs are skidded, the FCT damage associated with normal harvesting operations along these types of skid trails may be unavoidable. Alternatively, tree-length logs, which average approximately 18 m in length, could be bucked to length in the woods. They are currently bucked to 2.3 m lengths at the landing prior to loading for all species except Schizolobium parahyba (Veil. Cone.) S.F. Blake. This would likely cause less damage along skid trails and, if pre-set chokers were used as part of the skidding operation, it would not necessarily reduce skidding efhciency. FCT damage along unnecessary access routes and/ or dead ends that were not used to haul logs was found to be negligible in both the normal and FCT hagging harvest blocks no moderate or severe damage to FCTs was observed (Table 3). Improved skid trail planning, however, successfully eliminated this cate gory of skid trail from the concession. In the normal harvest block, 16 unnecessary skid trails were created and accounted for nearly 8% (440 m) of the total skid trail network (5765 m), while only one unnecessary skid trail was created in the improved skid trail plan ning harvest block accounting for less than 0.5% (20 m) of the total skid trail network (5241 m). Although superhuous skid trails likely have little consequence in terms of FCT damage, they represent unnecessary soil disturbance and a substantial cost to forest managers in terms of operational efhciency. Mean search times (the time empty skidders spend searching for the next log to be hauled after leaving the landing) were signihcantly reduced in the improved skid trail planning harvest block (Table 4; Student s f-test; P = 0.032), adjusted for a standard travel dis tance. This suggests that the hagging of secondary and tertiary skid trails prior to yarding operations allowed skidder operators to move without hesitation and make fewer mistakes when entering the forest in search of felled boles. Mean cycle (round trip) times were reduced, but not signihcantly so (Table 4; Student s f-test; P = 0.055), in the improved skid trail planning harvest block compared with the nor mal harvesting block. No signihcant differences were Table 4 Comparison of skidding efficiency data between treatments Treatment Mean cycle time components (minis) Mean cycle distance (m) Search Maneuvering Hook time Haul time Mean cycle time time time Mean cycle volume (m ) Replication Normal 10:47 a 1:56 a No data 2:41 a 43:08 a Improved skid 8:02 b 2:00 a 2:20 1:14 a 31:46 a trail planning Means with the same letter within columns are not significantly different (Student s t-test; P > 0.05).

10 390 W. Krueger/Forest Ecology and Management 188 (2004) observed in adjusted mean maneuvering, hook, or haul times between treatments (Table 4; Student s f-test; P > 0.05). In terms of timber recovery, no differences were observed between the normal harvesting and improved skid trail planning blocks. Logs were left in the woods in both operations, but not because skidder operators did not hnd them. In the case of C. ianeirensis, approximately 0.24 m^/ha were left because of timber defect caused by wood-staining fungi in the standing trees that rendered the logs unmerchantable. In the case of H. crepitans, approximately 0.21 m^/ha were not skidded because some logs contained insufhcient volumes of heartwood, the merchantable portion of this species. In the case of both species, experienced timber cruisers and sawyers were unable to recognize unmerchantable individuals prior to felling. The main difference between normal harvesting and improved skid trail planning was not in terms of timber recovery, but rather operational efhciency ^unnecessary skid trails leading to unmerchantable logs were common in the normal harvest block, but rare in the improved skid trail planning harvest block Economic considerations Economic considerations from this study are sum marized in Table 5. Improved felling techniques in La Chonta could yield substantial current harvest gains in timber recovery. Assuming a mean stumpage value of approximately US$ 20.00/m^ for species harvested in La Chonta (R. Keating, personal communication), US$ 3.20 per tree could be recovered on average for C. ianeirensis by eliminating stump pull during felling operations. This will require that forest managers provide sawyers with additional tools (i.e., felling wedges and/or hydraulic jacks) and training. A bonus based on some fraction of wood recovery as well as felling productivity may also be necessary to encou rage improved felling techniques. More careful saw ing could also yield economic benehts in the next harvest by reducing FCT damage in felling gaps. Gains from improved felling techniques will likely be in excess of the perceived felling productivity costs associated with more careful sawing. FCT hagging also holds great potential for mitigat ing damage caused by felling operations in this study area. FCT hagging adds approximately US$ 0.38/ha to the concession s normal management costs. If the costs of vine cutting on FCTs were also included, a treatment that has been shown to enhance FCT growth in this forest (T. Fredericksen, personal communica tion), the cost would be approximately US$ 0.47/ha (T. Fredericksen, unpublished data). This is a small cost, but may be difhcult for forest managers to justify given that it will yield economic returns only in the next cutting cycle. The additional costs of improved skid trail planning were estimated to be approximately US$ 1.46/ha. Improved skid trail planning reduced the average time spent by skidder operators searching for the next log to haul by 2 min and 45 s per turn, or 23%. Assuming a total operating cost of approximately US$ 30.00/h for skidding operations (from estimates by Holmes et al. (2002) for skidding operations in Brazil), improved skid trail planning will likely yield current harvest hnancial returns in excess of US$ 3.00/ha over normal skidding operations in La Chonta. Improved skid trail planning can also be expected to yield hnancial benehts in the next harvest through an 11 % reduction Table 5 Summary of economic considerations associated with modified logging practices in La Chonta Activity Estimated cost Estimated current harvest return Potential long-term benefits (US$/m^) (US$/ha) (US$/m^) (US$/ha) Improved felling No data No data No data techniques FCT flagging % reduction in FCT damage in felling gaps; 10% reduction in FCT damage along secondary skid trails Improved skid trail planning % reduction in soil disturbance caused by unnecessary skid trails; 11% reduction in FCT damage along secondary skid trails

11 W. Krueger/Forest Ecology and Management 188 (2004) in FCT damage along secondary skid trails. These rudimentary estimates should provide sufficient incen tive to justify investing in improved skid trail planning Regional implications Inferences from this study are limited to this and similar forest types with comparable management schemes and market settings. The operational efh ciency of skidding operations, for example, depends largely on harvest unit layout and design considera tions, the number of workers employed to carry out the operations, the degree of supervision provided to logging crews in the held, and local environmental variables. The study presented here, however, may have relevance for several nearby production forests that have followed the La Chonta model for develop ing logging operations. For example, the 100,000 ha Lago Rey timber concession in northern Santa Cruz follows this model and may be experiencing similar efhciency issues during skidding operations (personal observation). That no difference was observed in timber recovery between normal skidding operations and improved skid trail planning in La Chonta indicates that only modest investments in training and skid trail planning are needed to entirely eliminate waste associated with trees that are felled but never yarded. This supports the idea that planned skidding can yield substantial economic gains over unplanned skidding in the neo tropics (Gerwig et al., 1996; Johns et al., 1996; Barreto et al., 1998; Boltz et al., 2001; Holmes et al., 2002). Unmerchantable timber that was felled but not yarded in La Chonta results in unnecessary residual stand damage and reduced productivity during felling opera tions. The felling of unmerchantable timber may be a forest management issue throughout the ranges of H. crepitans and C. ianeirensis in Bolivia. Measures should be taken to better understand and identify timber defects associated with these species during forest inventory operations. The inability of well-trained, experienced timber fallers in this concession to accurately and consis tently fell trees in precise predetermined directions may have important implications throughout Bolivia. Directional felling has been identihed as a key component for sustainable natural forest management initiatives in the neotropics, and studies have shown its utility for reducing FCT damage, wood waste, and inefhciencies during logging (Gerwig et al., 1996; Johns et al., 1996; Barreto et al., 1998). The results from this study support the idea that accurate directional felling of old-growth timber necessarily involves the use of felling wedges and/or hydraulic jacks (Dent, 1974; ITTO, 1996c). When such tools are not used, felling can be expected to result in unne cessary residual stand damage, wood waste, and operational inefhciencies regardless of sawyer train ing or experience. This is likely the case for the majority of Bolivia s permanent production forests. FCT hagging may also have important regional implications for forest management, as it was shown to signihcantly reduce damage in felling gaps and along secondary skid trails in the study area. The cost of FCT hagging seems a modest investment in light of the potential to maintain or enhance stand quality, but there is currently little capital available among Boli vian forest managers for investment in silvicultural treatments that will yield economic benehts only in the medium or long-term (BOLFOR, 2000). Given the high risk associated with silvicultural investment in general (i.e., hre, insects, and disease), and in the tropics in particular (i.e., high discount rates and political instability), it seems unlikely that Bolivian forest managers will invest in FCT hagging unless it were required as part of FSC certihcation criteria. The demonstrated utility and low cost of this treatment warrant its adoption by the FSC. 5. Conclusion Moderate and severe damage to FCTs caused by conventional planned diameter-limit logging could be reduced in La Chonta in a cost-effective manner. As a hrst step, forest managers should inventory and mark all FCTs >30 cm DBH as part of regular forest inventory work prior to logging. To ensure that this initial investment is not lost during logging operations, forest managers should require that sawyers strictly adhere to directional felling techniques by providing them with felling wedges and/or hydraulic jacks, additional training, and an incentive based on the quality and quantity of their work. The implementation of improved timber falling techniques would likely pay for itself in current operations through

12 392 W. Krueger/Forest Ecology and Management 188 (2004) increased timber recovery. Finally, a forest laborer should be employed to plan all secondary and tertiary skid trails prior to yarding operations. This will result in increased operational efhciency and reduced FCT damage along secondary skid trails. The cost effec tiveness of these treatments and their potential to maintain or improve timber stand quality should make their adoption attractive to both commercial forest managers and regional timber certihcation groups. Acknowledgements Special thanks to Todd Fredericksen and Marielos Pena for their critical comments, advice, and logistical support throughout the duration of this project. Thanks also to Jack Putz, Steve Siebert, Carl Fiedler, Carlos Baled, Geoff Elate, Fred Boltz, Calvin Kiehn, Kristen Ohlson, Juan Carlos Licona, Joaquin Justiniano, Andrea Harausz, Jeanenne Bergen, Bob Keating, Remigio Ancalle, Daniel Nash, Arturo Justiniano, Rob Schantz, Ryan Maroney, Kerri Rosenstein, and two anonymous reviewers for insightful comments and advice. Thanks to Ina Brokhof, Roberto Quevedo, and the entire La Chonta logging crew for logistical assistance, cooperation, and valuable comments while conducting this study. Thanks to Lucio Alvarez Negrete, Antonio Urapuka Apirezai, Angel (Wicho) Mendez, and Ricardo Mendez Salvatierra for held assistance. This study was funded by the United States Peace Corps and by BOLFOR, a sustainable forest management project of the Bolivian govern ment and the United States Agency for International Development. References Anderson, A.B. (Ed.), Alternatives to Deforestation: Steps Toward Sustainable Use of the Amazon Rain Forest. Columbia University Press, NY, 281 pp. Barreto, P., Amaral, P., Vidal, E., Uhl, C., Costs and benefits of forest management for timber production in eastern Amazonia. For. Fool. Manage. 180, BOFFOR, Study Plan: Fong-term Silvicultural Research Project (FTSRP) in Bolivian Tropical Forests. Proyecto BOFFOR, Santa Cruz, Bolivia. Boltz, R, Carter, D.R., Holmes, T P, Perreira Jr., R., Financial returns under uncertainty for conventional and reduced-impact logging in permanent production forests of the Brazilian Amazon. Ecol. Econ. 39, Conway, S., Timber Cutting Practices, 2nd ed. Miller Freeman Publications, San Francisco, CA, 192 pp. Darby, J., Fa certificacion y el manejo forestal sostenible en Bolivia. Technical Document No. 79. Proyecto BOFFOR, Santa Cruz, Bolivia. Dent, D.D., Professional Timber Falling: A Procedural Approach. Ryder Printing Co., Portland, OR, 181 pp. Dickinson, M.B., Dickinson, J.C., Putz, F.E., Natural forest management as a conservation tool in the tropics: divergent views on possibilities and alternatives. Commonweal. For. Rev. 75, Dykstra, D.P., Heinrich, R., FAO Model Code of Forest Harvesting Practice. Food and Agriculture Organization of the United Nations, Rome, 85 pp. Fredericksen, T.S., Fimitations of low-intensity selection and selective logging for sustainable tropical forestry. Common weal. For. Rev. 77, Fredericksen, T.S., Fogging and conservation of tropical forests in Bolivia. Int. For. Rev. 2 (4), Fredericksen, T.S., Mostacedo, B., Regeneration of sawtim- ber species following selective logging in a Bolivian tropical dry forest. For. Ecol. Manage. 131, Fredericksen, T.S., Contreras, R, Pariona, W., Guia de Silvicultura para Bosques Tropicales de Bolivia. Proyecto BOFFOR, Santa Cruz, Bolivia, 81 pp. Gerwig, J.J., Johns, J.S., Vidal, E., Reducing waste during logging and log processing: forest conservation in eastern Amazonia. Unasylva 187 (47), Gil, P., Plan General de Manejo Forestal. Empresa Agroindustrial Fa Chonta Ftda., Santa Cruz, Bolivia, 45 pp. Gomez-Pampa, A., Whitmore, T.C., Hadley, M. (Eds.), Rain Forest Regeneration and Management. UNESCO, Paris, pp. 8-17, Gullison, R.F., Hardner, J.J., The effects of road design and harvest intensity on forest damage caused by selective logging: empirical results and a simulation model from the Bosque Chimanes, Bolivia. For. Ecol. Manage. 59, Harausz, A., Herleitung lokaler Volumenfunktionen fur drei Baumarten in Bolivia. M.Sc. Thesis. University of Albert- Fudwigs, Freiburg, Germany. Holmes, T P, Blate, G.M., Zweede, J.C., Perreira Jr., R., Barreto, P., Boltz, R, Bauch, R., Financial and ecological indicators of reduced-impact logging performance in the eastern Amazon. For. Ecol. Manage. 163, Howard, A.R, Rice, R.F., Gullison, R.F., Simulated financial returns and selected environmental impacts from four alter native silvicultural prescriptions applied in the neotropics: a case study of the Chimanes Forest, Bolivia. For. Ecol. Manage. 89, ITTO (International Tropical Timber Organization), 1996a. Bolivia adopts new forest law. Tropical Forest Update, vol. 6, No. 3, p. 15. ITTO, 1996b. What foresters can do? Tropical Forest Update, vol. 6, No. 3, p. 1. ITTO, 1996c. Felling the right way: some hints on the art and science of directional felling. Tropical Forest Update, vol. 6, No. 3, pp. 5-7.

13 W. Krueger/Forest Ecology and Management 188 (2004) Jackson, S.M., Fredericksen, T.S., Malcolm, J.R., Area disturbed and residual stand damage following logging in a Bolivian tropical forest. For. Ecol. Manage. 166, Johns, J.S., Barreto, P., Uhl, C., Logging damage during planned and unplanned logging operations in the eastern Amazon. For. Ecol. Manage. 89, Johnson, N., Carbarle, B., Surviving the Cut: Natural Eorest Management in the Humid Tropics. World Resources Institute, Washington, DC, 71 pp. Nittler, J.B., Nash, D.W., The certification model for forestry in Bolivia. J. For. 97, Putz, F.E., Romero, C., Biologists and timber certification. Conserv. Biol. 15, Putz, F.E., Blate, G.M., Redford, K.H., Fimbel, R., Robinson, J., Tropical forest management and conservation of biodiversity: an overview. Conserv. Biol. 15 (1), Stokes, B.J., Kluender, R.A., Williams, R.A., Klepac, J.F., Assessment of costs and impacts for alternative harvesting methods in mixed stands. Paper Presented at the Seventh Biennial Southern Silvicultural Research Conference, Mobile, AL, USA, November 17-19, Whitman, A.A., Brokaw, N.V.L., Hagan, J.M., Eorest damage caused by selection logging of mahogany (Swietenia macro phylla) in northern Belize. Lor. Ecol. Manage. 92, Whitmore, T.C., An Introduction to Tropical Rainforests, 2nd ed. Oxford University Press, New York, 282 pp.