Soil–landscape resource assessment for plantations — a conceptual framework towards an explicit multi-scale approach

Soil survey is a major component of forest land resource assessment. Conceptual and operational problems arise from employing the conventional methods of survey in forest lands, namely: implicit methods of landscape interpretation (lack of explicit procedures), transfer of data by analogy to unsampled landscapes by inferences which are scale-, and interpreter-dependent, variability of intuitive surveyor judgement, and poor expression of soil variation within map units. These issues are being addressed through the forestland resource assessment and modeling study (FRAMS). This study redefines the conceptual process of resource assessment, and applies soil–landscape modeling (developed here as regolith–terrain modeling) by developing explicit relationships between soil–landscape attributes within a digital, spatial geographic information system (GIS) framework. Soil survey (advanced here as regolith–terrain modeling) is the science and art of predicting soil attribute patterns in the 3D landscape. The FRAMS attempts to overcome some scale and procedural issues related to soil mapping in forest site assessment by adopting a multi-scale and explicit landscape modeling approach. The conceptual aspects of the method presented here aim to predict the ranges in variation of soil–geomorphic attributes that are relevant to forest plantation management. Soil–landscape analysis is adapted in this study to encompass regolith–terrain analysis (i.e. the complete regolith within an understanding of geomorphic systems) employed at three environmental scales: ‘hillslope’, ‘catenary’, and ‘landscape’. There is no linear relationship of data resolution and expression of regolith–terrain attributes between these scales. Each scale is a scale-dependent system linked by an explicit multi-scale method. When combined with geological and climatic data analysis the resultant model provides an advanced, stratified sampling scheme for subsequent field survey procedures in forestland resource assessment. The field analysis, remote-sensing and digital terrain model (DTM) analyses are managed in a raster GIS and can then be effectively classified, a posteriori, according to ‘fuzzy logic’ rules. In the FRAMS, we investigate the scale effect on both the regolith–terrain parameters and their notional relationships to forestland management by investigation at finer scales: hillslope and catenary scales (in southeast Queensland for planted native hoop pine ( Araucaria cunninghamii)), and at a