Using expert systems and simulation modelling for land evaluation at farm level: a case study from New York State

Publisher Website
Google Scholar
Abstract
A land evaluation at farm level was made for a toposequence of three New York soils, using the FAO framework and the LEACHN simulation model. The land utilization type (LUT) was continuous corn, using three varieties with different growing periods. Ten experts identified the same land qualities but three different types of land characteristics, as derived from soil survey data, field observations and simulation modelling. Differences could be explained by the backgrounds of the experts. A land evaluation procedure according to framework criteria using decision trees rather than matching tables was more accessible and transparent. Simulation of the soil‐water regime provided quantitative data for all land qualities considered, and this procedure was preferred to qualitative estimates based on poorly defined procedures. When some land qualities cannot be characterized by simulation, however, a mix of qualitative and quantitative decision tree branches may be considered. The standard scheme of land evaluation, which is scale independent, provides irrelevant results at farm level because farmers’ questions are different from those of regional planners who are well served by the traditional land evaluation procedure. For the farmer, development of decision support systems should be focused on optimal management decisions.