Estimating catchment-scale annual soil loss in managed native eucalypt forests, Australia

Abstract

Conservation of soil and water resources is one of the key criteria underpinning sustainable forest management. While soil and water resources are important determinants of forest productivity, without appropriate assessment of soil erosion risk and the application of best management practices (BMPs), some forest management activities can adversely affect hillslope erosion rates with detrimental consequences for aquatic environments and downstream water users. In the multiple-use native eucalypt forests of New South Wales (NSW), Australia, hazard matrix tables are currently used to identify soil erosion risk based upon rainfall erosivity, soil regolith stability and slope classes at the compartment scale prior to undertaking forestry activities. Resultant "inherent hazard levels" (IHLs) direct the BMPs to be used, such as riparian buffer widths, during harvesting and roading operations. The IHL model, being an ordinal classification system, only provides a relative indication of erosion potential without any quantitative estimate of possible post-harvest erosion rates. To potentially better identify erosion risk and quantify likely soil erosion under a range of forest management and climatic scenarios at the hillslope and/or catchment scale, in this paper we utilised an alternative approach by modelling soil erosion using the empirically-derived Revised Universal Soil Loss Equation in combination with a GIS-based spatially distributed raster analysis. In four case study catchments in Kangaroo River State forest, two of which were subjected to single-tree selection harvesting operations, mean annual changes in soil loss were estimated at a grid cell level. Potential differences in soil loss estimates were assessed before, during and after selective logging. Vegetation cover and soil samples were recorded in a 500 x 1000 m rectangular network laid out across the catchments. Slope gradient was found to contribute substantially to the spatial variability of soil loss estimation across the catchments. However, between-year differences demonstrate that the highest estimated annual rates of soil loss occurred on steep hillslopes when high levels of rainfall were recorded, while the values on those same areas remained considerably lower during low rainfall periods. The major effect of the rainfall component in generating soil erosion overshadows the modest impacts of selective logging operations.