Development and evaluation of pedotransfer functions for saturated hydraulic conductivity of seasonally impounded clay soils

Data on hydraulic properties of soils are often not available. Hence, pedotransfer functions (PTFs) are used to estimate hydraulic properties of soils. PTFs formulate a relationship between basic soil properties and property of interest (mostly any hydraulic property). Little information is available on the hydraulic properties of clay soils that are impounded by water for the entire monsoon season (known as 'HaveJi') in large tracts of Madhya Pradesh. This study was conducted to calibrate and evaluate PTFs to predict saturated hydraulic conduCtivity (K,) using basic soil properties. Global PTF 'Rosetta' was also eva1uated fcir'its validity in predicting Ks of the 'Haveli' soils. Available data on the 'Haveli' sods was used for the analysis. It included information on particle-size distribution, bulk density and water retention characteristics. PTFs were calibrated using regression as well as artificial neural networks. Imprecise estimates of Ks indicated that the calibrated PTFs were not reliable. However, excluding the samples with unusually low field capacity «0.3 m 3 m· 3 ) despite clay content (> 50%) resulted in PTFs that performed with precision. Estimates of Ks obtained by implementing hierarchical rules in generic PTF 'Rosetta' were poor in precision but improved with inclusion of field capacity and permanent wilting point as predictor variables. The. study indicated limitations of calibrated as well as generic PTFs in predicting saturated hydraulic conductivity.