Modelling of the adsorption of urea herbicides by tropical soils with an Adaptive-Neural-based Fuzzy Inference System

Sorption of pesticides by soils holds a major consequence for their fate in the environment. As such, sorption coefficient (K-d/K-oc), which is derived from laboratory or field experiments is a fundamental parameter used in almost all screening tools to evaluate the fate or mobility of these compounds. The value of this coefficient is controlled by many soil and solute specific properties, as well as environmental variables. Soft computing techniques such as Adaptive Neuro-Fuzzy Inference System (ANFIS) have been successfully used to predict the equilibrium partitioning of many compounds in various engineered systems. Application of these techniques to natural systems such as soils is however lacking. Here, we present the use of ANFIS in predicting the sorption per unit mass of soil, Q(e), used in the calculation of K-d/K-oc of compounds in soils. In a previous study, we collected data associated to the adsorption of five phenylurea herbicides in 18 tropical soils. Here, we analysed such data and based on established correlations, nine variables were selected as potential input vectors (i.e., six soil properties, two herbicides molecular descriptors and initial solute concentrations). A total of 255 ANFIS models of one to eight input vectors were elaborated under 10-fold cross-validation. Multiple linear regression (MLR) models were similarly developed, and compared with the ANFIS in terms of mean absolute error (MAE), root-mean-square error (RMSE) and coefficient of determination (R-2). The best ANFIS model (M94) has an MAE(test), RMSEtest and R-test(2) of 3.43 +/- 0.43, 4.94 +/- 0.80 and 0.95 +/- 0.01, respectively, whereas the best MLR model (M13) returned an MAE, RMSE and R-2 of 7.71 +/- 0.13, 10.11 +/- 1.21 and 0.81 +/- 0.01, respectively. We observed that generally ANFIS performed better than MLR regarding both accuracy and interpretability. Accordingly, we recommend the use of ANFIS for predicting the sorption coefficients of phenylurea herbicides (PUHs) in soils.

Automated summary

This study utilized ANFIS to predict the sorption coefficients of phenylurea herbicides in soils, with results indicating that ANFIS models outperformed MLR models in terms of accuracy and interpretability. Therefore, the use of ANFIS for predicting compound sorption coefficients in soils is recommended.