Abstract
Modeling water infiltration rate at the regional scale with the ruler of calcareous conditions is important for a better understanding of infiltration processes and infiltration modeling development. The aim of the present study was to derive and evaluate a digital soil mapping of the Philip model parameters (sorptivity and hydraulic conductivity) to predict water infiltration using environmental data in calcareous soils in northwest Iran. The infiltration data was carried out at 92 locations at the field scale with three replications. At each location, the various basic soil properties were measured, and environmental data obtained from attributes was derived from digital elevation models and remote sensing data. The feed-forward multilayer perceptron artificial neural networks (ANN) model was used to estimate sorptivity (S) and hydraulic conductivity (Ks). For the first type of prediction models, the measured basic soil properties, and for the second type, the measured basic soil properties and principal components (PCs) based on the environmental data were used as input data. The results showed a higher performance of ANN models developed based on environmental data (soil plus PCs data) than that developed based on only soil data for predicting both Philip infiltration model parameters. The R2 criteria was improved by 0.18 (from 39 to 57) for S-parameter and by 0.15 (from 0.44 to 0.59) for Ks-parameter prediction using ANN models developed based on environmental data. It was concluded that attributes derived from DEMs and remotely sensed information could be a potential environmental data for improving Philip infiltration model parameters and developing high quality infiltration data maps. That would be a first step in site-specific soil utilization, management and protection of the environment.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have