Bayesian Network Application to Land Suitability Classification in the Sewage Sludge Amendment of Agricultural Soils

摘要

Sewage sludge (SS) amendment of agricultural soils has recently become an issue of great interest because improvements in wastewater treatment systems have increased production of this type of waste. This practice is known to benefit soil and crops. However, the potential contamination of the environmental matrices has been rarely assessed because suitable models and data regarding the diffuse contamination of SS in agricultural fields are limited. By using land classification tools, we can evaluate the suitability of each soil parcel, the nature of the local soil, and the landscape parameters, and thus take more confident decisions regarding the management of this waste. In this study, we describe the development of a land classification tool to determine whether SS is suitable for organically-amending a given agricultural soil. Bayesian networks (BNs) were used to assess the suitability of agricultural soils located in Catalonia (northeastern Spain). The case study includes a description of the area and the procedure used to develop the BN model. The results of this evaluation are presented in innovative ways, with spatial representation of the probability results to give a broader understanding of the classification problem.