Forest Ecology and Management 144 (2001) 91±99 Impacts on soils from cable-logging steep slopes in northeastern Tasmania, Australia Mike Laffana,*, Greg Jordanb, Nathan Duhiga a Forest Practices Board, Perth, Tasmania 7300, Australia University of Tasmania, G.P.O. Box 252-55 Hobart, Tasmania 7001, Australia b Received 14 August 1999; accepted 5 February 2000 Abstract Cable-logging is a method commonly used to harvest native forests on steep slopes in Tasmania, Australia. A study was carried out to determine impacts on site disturbance and selected soil properties resulting from cable-logging wet eucalypt forest developed on granite substrates in northeastern Tasmania. The results show that both the area and depth of soil surface disturbance are relatively minor, with ca. 11% of the study coupe affected by moderate (litter and part topsoil removed) or severe (subsoil exposed) disturbance. Cable draglines accounted for just over 7% of the soil disturbance (7% moderate and 0.2% severe), whereas tree uprooting associated with the logging accounted for a further 3% moderate disturbance and 1% severe disturbance. Measurements of soil properties in the surface layer (0±10 cm) show that bulk density is ca. 20% higher and organic carbon content (kg/ha) is 15% lower on cable draglines compared to undisturbed sites. However, the dominant type of soil disturbance caused by cable-logging appears to have been displacement of the upper 10 cm of topsoil from the centre to the outside edges of draglines rather than compaction in situ. Comparison of the data with studies of ground-based logging of wet native forests elsewhere in southeastern Australia shows that cable-logging has resulted in signi®cantly less impact on both area of soil disturbed and soil properties. The results are also discussed in relation to relevant sustainability indicators (soil erosion, organic matter and compaction) speci®ed in Criterion 4 of the Montreal Process together with recently proposed threshold values. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Timber harvesting; Sustainable soil management; Sustainable forest management; Soil properties; Soil compaction; Soil erosion; Montreal criteria and indicators 1. Introduction Logging operations can cause signi®cant and widespread soil disturbance, including removal, mixing and compaction of the various soil layers. Disturbance can adversely affect both soil physical properties and soil nutrient levels to such an extent that severely * Corresponding author. Present address: Division of Forest Research and Development, Forestry Tasmania, 15962 Midland Highway, Perth, Tasmania 7300, Australia. diminished growth of subsequent tree rotations as well as a signi®cant increase in runoff and sediment loads may result. The extent and severity of logging impacts depend on a number of factors such as inherent soil properties, soil moisture content, topography, type of operation (whether selectively logged or clearfell, ground-based or cable-logged), type of machinery used, the number of machine passes, and forest conservation practices (for example, the use of slash as matting on snig tracks). 0378-1127/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 2 7 ( 0 0 ) 0 0 3 6 0 - 1 92 M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 Ground-based logging using rubber-tyred or tracked machinery to skid logs to a landing point for loading onto trucks is conventionally carried out where the terrain is ¯at, undulating or rolling (slopes <208). Steeper terrain invariably requires side cutting of slopes and construction of benches to allow traf®cability of ground-based machinery. Such practices can result in signi®cant exposure and displacement of soil and lead to increased runoff and erosion. In cable-logging systems logs are attached to an aerially suspended cable and pulled to a landing site. Depending on the degree of de¯ection of the cable, logs may be either fully suspended above the ground or only partly suspended so that logs are dragged over the surface of the ground. Data from North America and New Zealand (Krag et al., 1986; McMahon, 1995a) show that these systems generally result in much less site and soil disturbance than ground-based logging. They may be used on any type of terrain but are more common for harvesting steep slopes. In Tasmania cable-logging is the preferred method for harvesting slopes above 208 (Forestry Commission Tasmania, 1993). In southeast Australia numerous studies have been made of ground-based logging of cool temperate forests (Incerti et al., 1987; Williamson, 1990; Lacey et al., 1994; Rab, 1994, 1996), but no reports appear to have been published on the effects of cable-logging. To address this, a coupe proposed for cable-harvesting (SF 119H) near Scottsdale in northeastern Tasmania was selected to study the relative proportions of the soil surface disturbed by cable-logging and to measure any resultant changes in soil physical and chemical properties. The results of the study can be used as a basis for comparing the impacts of cable-logging with those resulting from ground-based logging on similar soils. Recently, the sustainable management of temperate forests has assumed an international perspective through the development of a comprehensive set of sustainability criteria and indicators as part of the Montreal Process (UN, 1995). The conservation and maintenance of forest soil and water resources are speci®ed in Criterion 4 of the Montreal Process with special reference to soil erosion, soil organic matter and soil compaction. The results of the cable-logging study are discussed in relation to these issues. 2. Site data 2.1. Location, geology and topography SF 119H is an elongated coupe on the southern and eastern sides of Cockatoo Hill (grid reference; 500 384, sheet 8415) ca. 8 km southeast of Scottsdale. Altitude ranges from ca. 150 m on the lower southern coupe boundary to 400 m near the crest of Cockatoo Hill. The underlying bedrock is Devonian granite forming mainly steep (20±268) southwest facing slopes. At elevations between ca. 250 and 300 m, very steep (>268) slopes are associated with a line of prominent rocky bluffs. East of Cockatoo Hill the coupe is mainly strongly rolling (12±198). 2.2. Vegetation and soils Native vegetation is dominated by wet eucalypt forest with an overstorey of Eucalyptus obliqua and E. regnans over Pomaderris apetala, Acacia melanoxylon, A. dealbata, Olearia argophylla, Coprosma quadri®da and Dicksonia antarctica. At elevations below ca. 300 m on the major spur forming the southwestern boundary of the coupe, the native vegetation is damp eucalypt forest rather than wet forest. The canopy typically comprises E. obliqua and E. amygdalina overlying an array of drier-tolerant shrubs and ground cover including Bedfordia spp., Monotoca glauca, Pteridium esculentum, Polystichum proliferum with minor Pomaderris apetala and Gahnia grandis. The soils are mainly deep (>80 cm) and well drained with gradational texture-pro®les characterised by relatively thick (>30 cm) dark greyish brown coarse sandy clay loam topsoils overlying brown coarse sandy clays. They are correlated with `Stronach soil pro®le class' described by Grant et al. (1995a) and described and mapped at 1:50 000 scale in the Forester sheet (Grant et al., 1995b). The soils have relatively high levels of nutrients (total N and P, organic C) and, because of high proportions of water-stable aggregates together with other favourable physical properties, their rating for erodibility (Laffan et al., 1996) is low. In the southwestern part of the coupe where damp eucalypt forest predominates, the soils have coarser textures in upper layers forming texture-contrast pro®les. They are characterised by gravelly coarse sandy M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 loams with pale-coloured A2 subsurface horizons overlying brown gravelly light clay subsoils. Nutrient levels in upper layers are low. The soils were not sampled for aggregate stability analysis, but, because of higher coarse sand content in surface and subsurface layers they are likely to have lower proportions of water-stable aggregates and a higher rating of erodibility than `Stronach' soils. 2.3. Logging The coupe was logged in several stages with the ®rst stage comprising the southwestern half below the bluffs occurring at elevations of ca. 250±300 m. The second stage covered the higher elevation area east of Cockatoo Hill. The study of cable-logging impacts was con®ned to the area harvested ®rst (ca. 10 ha) and to the slopes above the landing. The area was logged during February and March 1997 using conventional ground-based skidding on undulating and easy rolling slopes, and downhill cabling of the higher steeper slopes onto a landing near the southern boundary. The area studied is planned for a controlled burn in the autumn of 2000 followed by planting to radiata pine several months later. 3. Methods 3.1. Pre-logging surveys and sampling Pre-logging surveys were carried out during January and February 1997 to determine soil characteristics and variability. A single transect running along the contour was established approximately midslope in the area to be cable-logged. Small inspection pits were excavated to depths of 50±60 cm at intervals varying between 10 and 25 m. In total, 28 pits were excavated and brie¯y described to record soil horizon nomenclature, thickness, texture and colour. Slope angle and aspect were also recorded at each site. Twelve sites with gradational soils (Stronach soil pro®le class) were sampled at depths of 0±10 cm for laboratory analysis of bulk density, total and available phosphorus (P), total nitrogen (N), organic carbon (C) and pH. Five of these sites were also sampled at depths of 10±20 and 20±30 cm for laboratory analysis. Two further sites with Stronach soils were sampled at 93 0±10 cm for bulk density only. Carbon was analysed according to the Walkley and Black method, nitrogen and total P by FIA analysis following a Kjeldahl digest, and available P was extracted with Olsen bicarbonate (Rayment and Higginson, 1992). 3.2. Post-logging surveys and sampling Post-logging surveys using a point transect method (McMahon, 1995b) were carried out in August 1997 to assess the proportions of disturbed and undisturbed sites. Three transects were established on the contour covering the upper, middle and lower sections of the cable-logged area. According to McMahon (1995b) a certain number of site observations is required within a setting or coupe to ensure a speci®ed error limit on the results. For example, McMahon (1995b) states that 10 000 observations are required for an absolute error of 1%, 1111 observations for an error of 3% and 400 observations for an error of 5%. For the study in SF 119H it was decided that an absolute error of 3% would be suf®ciently precise. Consequently, a total of 1139 observations was completed at intervals of 1 m along the three transects. At each site the degree and type of soil disturbance was assessed visually and recorded into one of the following disturbance classes:  Class 1: Undisturbed (litter layer and mineral topsoil intact, and no sign of compaction or other disturbance),  Class 2: Slightly disturbed (litter and topsoil intact, but evidence of minor compaction, or litter partially removed),  Class 3a: Moderately disturbed (litter and part topsoil removed on cable dragline),  Class 3b: Moderately disturbed (litter and part topsoil removed by tree uprooting associated with logging operations),  Class 4a: Severely disturbed (litter and topsoil removed to expose subsoil on cable dragline),  Class 4b: Severely disturbed (subsoil exposed by tree uprooting associated with logging operations),  Class 5: Slash cover (vegetation debris including tree heads and understorey species overturned during logging). The degree of soil disturbance beneath the slash cover was difficult to ascertain, but, in all cases appeared to be undisturbed or only slightly disturbed. 94 M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 Cable draglines are the natural lines of movement along the ground surface of partially suspended logs attached to a cable extending from the edges of the coupe to the cable-machine on the landing. Surface (0±10 cm) soil layers of seven cable draglines on Stronach soils were sampled for bulk density and nutrients (total and available P, total N organic C and pH). Samples for four draglines were taken about the mid-slope position, and, for the three other draglines sites were sampled in both the upper part of the coupe and at lower elevation nearer the landing. At each of the 10 sites 3±5 bulk density cores and a single sample for soil nutrients (bulked from three separate cores) were collected for laboratory analysis. Because nearly all cable-draglines were moderately disturbed only Class 3a sites were sampled for laboratory analysis. Sample sites were all located in the bottom of draglines where disturbance and compaction was assumed to be severest. Low ridges of displaced topsoil mixed with slash occur along the sides of draglines and these sites are likely to have lower bulk densities than undisturbed sites. However, they were not sampled because of difficulties in obtaining representative cores. The effects of moderate disturbance on soil properties were tested with unpaired, two-tailed t-tests between 12 and 14 randomly selected undisturbed sites and 10 randomly selected dragline sites. Similar tests compared soil properties of two subsurface layers in ®ve undisturbed sites. 4. Results 4.1. Pre-logging transect study The pre-logging transect revealed that six of the observations were texture-contrast soils and 22 were gradational soils (Stronach soil pro®le class). The results for mean thickness of topsoil (A1, A11, A12, AB horizons) and depth to B horizon for Stronach soils are presented in Table 1 together with mean slope angle for all 28 sites. These results con®rm the relatively thick (>30 cm) A1 horizons in Stronach soils and the occurrence of steep (20±268) slopes in the central part of the coupe. Table 1 Means and standard deviations (standard errors in parentheses) for slope angle and topsoil thickness for gradational soils in unlogged forest Mean Mean Mean Mean thickness of A11 or A1 (cm) thickness of A12 or AB (cm) depth to B horizon (cm) slope angle (8) 21.0  9.7 10.1  7.6 31.0  9.9 25.2  3.7 (2.1) (1.6) (2.1) (0.7) 4.2. Soil disturbance A photograph of the western part of the coupe (Fig. 1) shows cable draglines and typical surface disturbance following logging. The results of the post-logging survey showing frequency and percentage occurrence of the various disturbance classes are summarised in Table 2. Table 2 shows that most observations (>70%) were slash cover (47%) and undisturbed soils (25%). Disturbed soils accounted for nearly 30% of observations with most comprising slightly (17%) and moderately (10%) disturbed classes. Severely disturbed (subsoil exposed) soils accounted for only just over 1% with nearly all attributable to tree uprooting during logging operations. Soil disturbance due to gouging along draglines accounted for 7% of observations, nearly all of which were moderately disturbed (litter and part topsoil removed). The width of cable draglines varied from ca. 2 m in upper slope positions to 3±4 m in lower slope sites nearer the landing. In cross-section, draglines are concave in shape with shallow central ruts and low ridges of displaced topsoil on each side. The maximum depth of draglines below the original ground surface was dif®cult to determine because of the ridges of topsoil and logging slash accumulated along both sides of draglines. However, it was estimated that for most draglines somewhere between ca. 5 and 15 cm of topsoil had been removed from the central ruts and deposited to the sides. Exposed subsoil was observed at only one site in the survey lines and covered a very small area (<0.2 m2). 4.3. Soil properties Results for bulk density and chemical analysis from undisturbed (pre-logging) sites and moderately disturbed (post-logging) sites on draglines are given in M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 95 Fig. 1. Western part of cable coupe (SF 119H) 6 months after completion of logging showing draglines radiating out from landing (where vehicle is parked). Further draglines extend out into the eastern part of the coupe to the right of the photo. Table 3. The samples are highly variable for each mineral nutrient (total P and N, available P) and organic carbon, but less so for pH and bulk density. The mean nutrient concentrations and organic C% are each higher in the undisturbed soils than in the disturbed soils by 30±40%. Bulk density is ca. 20% higher in the disturbed sites compared to undisturbed sites. However, calculation of organic C and total N contents (kg/ha) using mean concentrations and bulk density data show that differences between disturbed Table 2 Frequency of soil disturbance classes following logging Soil disturbance class Frequency % Undisturbed (Class 1) Slightly disturbed (Class 2) 288 192 25 17 Moderately disturbed Cable dragline (Class 3a) Tree uprooting (Class 3b) 77 32 7 3 Severe disturbance Cable dragline (Class 4a) Tree uprooting (Class 4b) 2 <0.2 10 1 Slash cover (Class 5) Total (%) 538 1139 47 100 and undisturbed sites are only ca. 15 and 10% for carbon and nitrogen, respectively. The t-test shows that only organic C%, total N% and bulk density are signi®cantly different (p<0.05) between disturbed and undisturbed sites. Pearson correlations between the variables show that the six measured parameters are not independent of each other. Organic C% is strongly correlated with bulk density, total N% and total phosphorus. Likewise total N% is strongly correlated with total P%, and total N% is also strongly negatively correlated with bulk density. These correlations can be simply explained by increases in the proportion of organic C% reducing bulk density and increasing nutrient levels. Results of laboratory analysis of subsurface (10±20 and 20±30 cm) soil samples from undisturbed sites are given in Table 4. As expected, organic carbon and nutrient concentrations decrease with depth whereas bulk density increases with depth below the soil surface. Comparison of Tables 3 and 4 shows that data for bulk density and total N for disturbed sites (0±10 cm depth) is almost identical to that for undisturbed sites at depths of 10±20 cm. The other nutrients, particularly organic C and total P are also similar at these depths suggesting that cable-logging has caused displacement of the surface layer of topsoil from the 96 M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 Table 3 Means and standard deviations (standard errors in parentheses) for soil properties (0±10 cm) for undisturbed and moderately disturbed sites and results of t-tests pH Bulk density (Mg/m3) Organic C (%) Organic C (kg/ha) Total N (%) Total N (kg/ha) Total P (ppm) Available P (mg/kg) Undisturbed sites Moderately disturbed sites (cable-draglines) t p 5.44  0.43 (0.12) 0.96  0.12 (0.03) 4.33  1.47 (0.43) 40,900  12,200 (3500) 0.32  0.09 (0.03) 3060  750 (220) 314  120 (35) 12.2  4.8 (1.5) 5.11  0.33 (0.10) 1.17  0.15 (0.05) 3.07  0.91 (0.29) 35,000  7590 (2400) 0.24  0.07 (0.02) 2760  620 (200) 247  117 (37) 9.0  3.7 (1.6) 2.03 3.83 2.46 1.38 2.32 1.03 1.35 1.72 0.055 0.001 0.024 0.184 0.031 0.314 0.19 0.10 centre to the outside edges of draglines and exposed the underlying layer below 10 cm. 5. Discussion 5.1. Comparison with studies of ground-based logging The results of studies of ground-based logging of wet forests in southeastern Australia together with those from SF 119H are summarised in Table 5. Williamson's (Williamson, 1990) grid line intercept study of conventionally-logged coupes under wet forests in Tasmania covered a range of substrates including dolerite, granite, basalt, mudstone/sandstone and quartzite/mudstone sequences. Severe soil disturbance where topsoils were wholly displaced or removed covered between 5 and 11%, and moderate disturbance where topsoils were partially removed occupied from 8 to 19% of these coupes, respectively (Williamson, 1990). Rab (Rab, 1994, 1996) investigated the impacts of conventional logging on site disturbance and soil properties in the Victorian Central Highlands. The six coupes studied were located at elevations between ca. 600 and 980 m with a cool temperate climate and a cover of wet eucalypt forest dominated by Eucalyptus regnans. Landforms were mainly steep to rolling hills on granitic and metamorphic substrates. Red gradational soils predominated and pro®les were characterised by relatively thick (25 cm) clay loam topsoils with up to 40 cm along drainage lines. Most of the disturbance recorded in the later study (Rab, 1996) was con®ned to the topsoil. Taking into account all disturbed sites including snig tracks, landings and general logging areas outside snig tracks it was found that topsoils were disturbed across 60±80% of the coupe area. Incerti et al. (1987) studied the impacts on soil physical properties resulting from clear-falling Table 4 Means and standard deviations (standard errors in parentheses) for subsurface (10±20 and 20±30 cm) soil properties from undisturbed sites and results of t-tests pH Bulk density (Mg/m3) Organic C (%) Organic C (kg/ha) Total N (%) Total N (kg/ha) Total P (ppm) Available P (mg/kg) 10±20 cm 20±30 cm t p 5.5  0.5 (0.2) 1.17  0.13 (0.06) 2.7  1.0 (0.5) 30,300  9,100 (4100) 0.23  0.07 (0.03) 2630  670 (300) 295  135 (61) 5.4  1.8 (0.8) 5.6  0.3 (0.1) 1.35  0.15 (0.07) 1.9  0.9 (0.4) 25,100  10,900 (5500) 0.17  0.08 (0.04) 2070  900 (450) 254  140 (63) 3.7  2.0 (0.9) 0.48 1.93 1.24 1.71 1.23 1.05 0.46 1.44 0.65 0.09 0.25 0.12 0.25 0.34 0.66 0.19 97 M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 Table 5 Comparison of results from ground-based logging of wet forests in southeastern Australia with cable-logging of wet forest in northeast Tasmania (SF 119H) Ground-based logging Area occupied by snig tracks or cable draglines (%) Increase in bulk density on snig tracks or cable draglines (%) Decrease in organic C (%) on snig tracks or cable draglines (%) Decrease in organic C (kg/ha) on snig tracks or cable draglines (%) Area of severe disturbance (%) Area of moderate disturbance (%) Cable logging (SF 119H) Williamson (1990) Lacey et al. (1994) Rab (1994) Rab (1996) Incerti et al. (1987) 23±39 ± 25 18 10 7 ± 20 63 53 27 22 ± ± 52 40 ± 29 ± ± ± ± ± 15 5±11 8±19 ± ± ± ± 5±15 60±80 ± ± 1 10 E. regnans forest in the Otway Range in southern Victoria using ground-based machinery. Information on substrates was not provided, but, the soils were broadly classi®ed as brown gradational. These authors found that snig tracks and landings occupied ca. 10% of the coupe, and that bulk density in the upper 6 cm of topsoil increased from 0.96 to 1.22 Mg/m3 (27%) on snig tracks. Lacey et al. (1994) measured changes in soil physical properties following ground-based harvesting of wet eucalypt forests in southern New South Wales. Substrates were Devonian granodiorite with red gradational soils. Their sampling method was not intended to estimate relative site condition across the entire harvest area, but they found that the bulk density of snig tracks increased from 1.19 to 1.43 Mg/ m3 (20%) in the surface layer (0±10 cm). Comparison of results of soil disturbance between cable-logging in SF 119H and conventional-logging elsewhere in Tasmania (Williamson, 1990) and in Victoria (Rab, 1994, 1996; Incerti et al., 1987) show that cable-logging has caused signi®cantly less surface disturbance. Approximately just over 11% of the cable-logged coupe has been topsoil disturbed compared to between 60 and 80% of the coupes in the Victorian central highlands. Cable draglines covered just over 7% of SF 119H, whereas the area under snig tracks in ground-based logging ranged from 23 to 39% in Tasmania (Williamson, 1990), 18 to 25% in the Victorian Highlands (Rab, 1994, 1996), and ca. 10% in the Otway Ranges (Incerti et al., 1987). The degree of surface disturbance is also much less in the cable-logged coupe with only 1% severe disturbance compared to 5±11% under ground-based logging elsewhere in Tasmania (Williamson, 1990) and 5±15% in the Victorian Highlands (Rab, 1996). Impacts on soil properties are also generally less severe under cable-logging compared to ground-based logging. Increases in bulk density following cablelogging averaged 22% compared to increases following ground-based logging of 53 and 63% in the Victorian Highlands (Rab, 1994, 1996), 27% in the Otway Ranges (Incerti et al., 1987) and 20% in southern NSW (Lacey et al., 1994). Likewise, organic carbon concentrations decreased by 29% following cable-logging compared to 40 and 52% in the Victorian Highlands (Rab, 1994, 1996). However, comparisons of soil organic matter and nutrient levels should preferably be made on the basis of weight per unit area (kg/ha) to take account of differences in soil thickness and bulk density. Calculation of organic C content (kg/ ha) for the upper 10 cm shows that the reduction due to cable-logging is signi®cantly less (15%) compared to using values of organic C% (Table 5). Similarly, the decrease in soil nitrogen concentration (total N%) is 25%, whereas the amount of soil N (kg/ha) has decreased by only 10%. 98 M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 5.2. Soil sustainability indicators The maintenance of soil fertility is generally regarded as being crucial to the sustainable management of forests. However, the effects of changes in soil properties on subsequent regeneration and growth of trees, and on other forest values is an area which is not well understood. It is generally accepted that severe disturbance such as complete removal of topsoil will have signi®cant deleterious effects on soil fertility levels and forest productivity. On the other hand, the effect of slight disturbance on forest productivity is much more dif®cult to quantify. The criteria and indicators of the Montreal Process are only broadly de®ned and do not specify critical or threshold values. Using results from the Victorian Central Highlands, Rab (1999) has outlined measures and operating standards for quantifying the Montreal Process soil indicators 4a (area and percent of forest land with signi®cant soil erosion), 4d (area and percent of forest land with signi®cantly diminished soil organic matter and/or changes in other soil chemical properties) and 4e (area and percent of forest land with signi®cant compaction or changes in other soil physical properties). For indicator 4a, Rab (1999) used the sum of the area occupied by access roads, snig tracks, ®rebreaks and harvested sites with subsoils disturbed. However, this approach ignores the soil properties resisting erosion (erodibility status) which have a highly important role in determining the degree and extent of erosion. A better method for quantifying indicator 4a would be to assess actual soil movement at some speci®ed time interval after logging, or to use measures of soil erodibility (for example, Laffan et al., 1996) in combination with slope angle and rainfall erosivity to arrive at estimates of erosion hazard of exposed mineral soil. Such an approach was proposed by Boyer and Dell (1980) for the Paci®c Northwest Region of USA, where operating standards were based on three ratings (low to moderate, high, very high) of erosion hazard. For indicators 4d and 4e, Rab (1999) used threshold values of 20% increase in bulk density and 30% decrease in organic matter content measured in the surface soil (0±10 cm) as indicating signi®cant impacts on soil fertility levels resulting from log- ging. Table 5 shows that in the 0±10 cm soil layer for the cable-logged coupe, mean bulk density increased by 22% and mean organic carbon content (kg/ha) decreased by 15%. Applying the standards proposed by Rab (1999), it is evident that the threshold value for increase in bulk density will be exceeded only on part of the area of SF 119H where moderate and severe soil disturbance occur (<11%). The threshold value for organic matter decrease is only likely to occur on the area (1%) of severe soil disturbance where subsoils have been exposed and on that part of the area (<3%) where deeper topsoil layers (below 20 cm depth) have been exposed by tree uprooting. The areas affected are much less than that for the Victorian Central Highlands where Rab (1999) found from 38 to 51% of coupes were affected by bulk density increases >20%, and 12 to 45% of coupes had >30% decrease in organic matter. However, Rab's data included ®rebreaks, landings and some areas scari®ed during site preparation for regeneration as well as snig tracks, and soils were sampled for organic matter after regeneration burns. Studies of the effects of compaction and loss of organic matter on tree growth quoted by Rab (1999) show that impacts on forest productivity will be highly dependent on inherent levels of organic matter, bulk density and other soil properties such as texture, existing prior to logging. For example, 30% decrease in organic matter is likely to have a large adverse effect on subsequent tree growth in sandy soils with inherently very low levels and poor reserves of organic matter, but conversely, may have negligible impact in loamy or peaty soils with large organic reserves. Clearly, further research is required into establishing threshold values for bulk density and organic matter on a wide range of soils. It is recommended that in addition to organic matter other measures of soil nutrient status such as nitrogen and phosphorus levels be included as indices of sustainability. For example, concentrations of total N and P are routinely used to assess soil nutrient availability and site suitability for plantations in Tasmania (Laffan, 1997). Also, because of relatively high proportions of charcoal commonly found in soils under native eucalypt forest, organic carbon content may be grossly over-estimated using normal laboratory methods. M. Laffan et al. / Forest Ecology and Management 144 (2001) 91±99 99 6. Conclusions References The cable-logging study in northeastern Tasmania clearly shows that this logging method has signi®cantly less impact on both area of soil disturbed and on soil properties compared to most other studies of ground-based logging in southeast Australia. These results are consistent with overseas studies which reveal similar relatively low impacts fromcable-logging. The similarity of data for bulk density and nutrient concentrations between draglines (0±10 cm) and undisturbed subsurface (10±20 cm) soil layers in SF 119H indicate that the dominant type of disturbance caused by cable-logging has been displacement of topsoil rather than compaction in situ. It appears that the process of dragging logs over the ground surface has moved the upper soil layer (0±10 cm) from the centre to the outside edges of draglines and exposed the underlying subsurface (10±20 cm) soil layer. Using the standards proposed by Rab (1999) for increase in bulk density (20%) and decrease in organic matter content (30%), it is conjectured that cablelogging impacts on soil fertility are likely to be relatively minor over the entire coupe (ca. 11% affected). They may be signi®cant on that part (1%) of SF 119H where severe disturbance has occurred, however, further research is required on a wide range of soil types and site disturbance classes to relate forest productivity to changes in bulk density and organic matter. Keeping in mind its relatively low impacts on soil and site disturbance, cable-logging should also be considered as an alternative to ground-based harvesting on easy slopes, particularly on sensitive sites such as karst and areas with poorly drained or highly erodible soils. Boyer, W.D., Dell, R., 1980. Fire effect on Pacific Northwest forest, R6 WM 040, USDA Forest Service, Portland, OR, 59 pp. Forestry Commission Tasmania, 1993. Forest Practices Code. Forestry Commission, Hobart, Tasmania. Grant, J.C., Laffan, M.D., Hill, R.B., Neilsen, W.A., 1995a. Forest Soils of Tasmania. A Handbook for Identification and Management, Forestry Tasmania. Grant, J.C., Laffan, M.D., Hill, R.B., 1995b. Soils of Tasmanian State Forests. 2. Forester Sheet, Northeast Tasmania. Soils Bulletin No. 2, Forestry Tasmania. Incerti, M., Clinnick, P.F., Willatt, S.T., 1987. Changes in the physical properties of a forest soil following logging. Aust. For. Res. 17, 91±98. Krag, R., Higginbotham, K., Rothwell, R., 1986. Logging and soil disturbance in southeast British Columbia. Can. J. For. Res. 16 (6), 1345±1354. Lacey, S.T., Ryan, P.J., Huang, J., Weiss, D.J., 1994. Soil Physical Property Change from Forest Harvesting in New South Wales. Research Paper No. 25, Research Division, State Forests of New South Wales. Laffan, M., Grant, J., Hill, R., 1996. A method for assessing the erodibility of Tasmanian forest soils. Aust. J. Soil Water Cons. 9 (4), 16±22. Laffan, M., 1997. Site selection for hardwood and softwood plantations in Tasmania. A methodology for assessing site productivity and suitability for plantations using land resource information, 2nd Edition. Soils Technical Report No. 3, Forestry Tasmania and Forest Practices Board. McMahon, S., 1995a. Cable logging disturbance on an unstable slope: A case study. LIRO Report, Vol. 20, No. 12. Logging Industry Research Organisation, Rotorua, 9 pp. McMahon, S., 1995b. Survey method for assessing site disturbance. LIRO Report 54, Logging Research Organisation, Rotorua, 16 pp. Rab, M.A., 1994. Changes in physical properties of a soil associated with logging of Eucalyptus regnans forest in southeastern Australia. For. Ecol. Manage. 70, 215±229. Rab, M.A., 1996. Soil physical and hydrological properties following logging and slash burning in the Eucalyptus regnans forest of southeastern Australia. For. Ecol. Manage. 84, 159± 176. Rab, M.A., 1999. Measures and operating standards for assessing Montreal process soil sustainability indicators with reference to Victorian Central Highlands forest, southeastern Australia. For. Ecol. Manage. 117, 53±73. Rayment, G.E., Higginson, F.R., 1992. Australian Laboratory Handbook of Soil and Water Chemical Methods. Inkata Press, Melbourne, Sydney. UN (United Nations), 1995. Santiago Declaration. United Nations Secretariat. Williamson, J.R., 1990. The Effects of Mechanised Forest Harvesting Operations on Soil Properties and Site Productivity. Research Report No. 5, Tasmanian Forest Research Council, Inc. Acknowledgements Mr Phillip Pennington, Experimental Scientist, CSIRO Forests and Forest Products, Hobart, is acknowledged for comments on interpretation of the data. Mr Lachie Clark, vacation worker with the Forest Practices Board, is thanked for ®eld assistance during the pre-logging survey.