Estimation of soil water retention in conservation agriculture using published and new pedotransfer functions

Conservation agriculture has been developed as a means to improve the sustainability of agricultural systems and reduce drawbacks conventional practices. Cropping practices can have large influence on soil properties such water retention. Proper tools are needed assess model effects conservation properties. As measuring retention is expensive time consuming, pedotransfer functions (PTFs) now commonly used predict them. The objectives this study were (i) present new dataset data, (ii) performances existing PTFs in predicting soils under (iii) develop specific more accurately. We data collected only fields France evaluate several published with three evaluation criteria (RMSE, prediction bias (ME) Nash-Sutcliffe Efficiency (EF)). then using methods ― multiple linear regression, regression tree random forest content at matric heads -100 (θ100, field capacity for sandy soils), -330 (θ330, other soils) -15 000 cm (θ15 000, wilting point). Soil tillage, presence cover crop, rotation length previous reduced/no tillage predictors addition basic trees forests. quality ME EF) was calculated each PTF cross-validation procedure. Generally, predictions point had lower absolute error than those sandy-soil (RMSE = 0.044 0.066 cm3/cm3, respectively). EF usually negative all contents. performance similar (RMSE: 0.028, ME: 0.000, EF: 0.34 θ100) 0.027, 0.36 θ100), generally worse (especially EF). Multiple that did not consider cropping performed well thus identify any major management predicted content. Future research developing should focus identifying relevant predictors.