Quantitative site and soil descriptors to improve the utility of forest soil surveys

Forested landscapes frequently have a higher degree of heterogeneity than a gricultural landscapes at the site and hillslope scales. The terrain may ha ve a high proportion of rock outcrop, standing trees, fallen logs, and vari ous forms of biotic disturbance. Aspects of the soil and land surface are n ot well characterised in current description systems even though they influ ence landscape processes, productivity and management. Explicit methods for locating observations and describing components of the soil and land have been developed and tested in a forested granitic landscape in southeast NSW , Australia. Potential sampling patches were generated by computer prior to fieldwork using terrain analysis methods in a Geographic Information Syste m. Areal density and height of rock outcrop was measured along two 10 m tra nsects using a line intercept method. Estimation of area occupied by sub-su rface boulders was made using a soil probe to a depth of 0.5 m at every met re along the transects. All of the rock fragments with an axis >0.05 m were excavated from the soil pit and three axes measured with a ruler. The dime nsions of the soil pit were also recorded, allowing calculation of volumetr ic rock content in various size classes. Rock fragments 0.005-0.05 m were s ieved and weighed from disturbed horizon samples to estimate mass percentag e of different grades of rock fragment. The volume of dead wood per ha was estimated using a line intercept method, while the volume of live wood was estimated using basal area wedges. Surface cover and disturbance were estim ated as percentages and types at metre intervals along the transects. Litte r was collected, dried and weighed, and expressed on a mass per ha basis. T he methods were developed and used in conjunction with conventional soil de scription methods (McDonald R.C., Isbell, R.F., Speight, J.G., Walker, J., Hopkins, M.S., 1990. Australian Soil and Land Field Handbook, 2nd Edition. Inkata Press, Melbourne, 197 pp.), an expanded set of morphological descrip tions (McKenzie, N,J., Jacquier, D.W., 1997. Aust. J. Soil Res. 35, 803-826 ), and laboratory measurements to provide quantitative estimates of soil an d hydraulic properties across forested landscapes. Statistical models were developed to predict the distribution of individual attributes across the c atchments with specified levels of confidence. Some properties (such as roc k outcrop, percent clay in the B horizon) could be predicted with some reli ability across the catchments, while others (such as litter loads, surface cover) could not. Many of the predictable attributes are directly useful to forest managers and also enable estimation of more complex properties (e.g . plant-exploitable soil volume or hydraulic properties). Selected results clearly demonstrate the utility of the additional site and soil descriptors presented in this paper. (C) 2000 Elsevier Science B.V. All rights reserve d.