Quality evaluation of restored soils with a fuzzy logic expert system

Abstract Due to landscaping, mining and construction activities on previously cultivated land, more and more soils are excavated, translocated, deposited and restored. In many cases restored soils show signs of structural degradation such as overcompaction and water logging. There is a lack of methods to evaluate and assess the physical quality of restored soil. In this study a fuzzy logic expert system was developed which allows us to evaluate the potential plant productivity of restored soils based on measured physical soil parameters. The system is based on the statements of a group of soil experts relating physical soil quality to packing density, penetration resistance, air capacity and saturated hydraulic conductivity according to their personal experience or expertise. From these statements we derived fuzzy membership functions and inference rules. The expert system was applied to evaluate 10 restored sites in comparison to nearby non-restored reference soils. The physical soil quality had remained unchanged or decreased after restoration at most investigated sites. Only two horizons showed clearly improved soil conditions after restoration. The validity of the fuzzy logic expert system is demonstrated by comparing the results with evaluations of the same soils using two other indicators of the physical soil quality for plant production: the least limiting water range (LLWR) and the S-parameter (i.e. the slope of the water retention curve at the inflection point). The physical soil quality assessment with the fuzzy logic expert system was highly correlated with both the LLWR (r2 = 0.80) and the S-parameter (r2 = 0.70). These results show that fuzzy logic expert systems may provide a suitable tool to assess physical soil quality, taking proper account of the vagueness and ambiguity necessarily involved in this task.