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 (K_s). 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 K_s-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.

用钙质条件标尺对区域尺度的水入渗速率进行建模对于更好地了解入渗过程和入渗建模发展至关重要。本研究的目的是推导和评估Philip模型参数（吸附度和水力传导率）的数字土壤图，以利用伊朗西北钙质土壤中的环境数据预测水的渗透。渗透数据在田间规模的92个位置进行，重复3次。在每个位置都测量了各种基本的土壤特性，并且从属性获得的环境数据是从数字高程模型和遥感数据中得出的。前馈多层感知器人工神经网络（ANN）模型用于估算吸附性（S）和水力传导率（K _s）。对于第一种类型的预测模型，测量的基本土壤特性，对于第二种类型，基于环境数据的测量的基本土壤特性和主成分（PC）用作输入数据。结果表明，基于环境数据（土壤加PCs数据）开发的ANN模型比仅基于土壤数据预测Philip入渗模型参数的ANN模型具有更高的性能。对于S参数，R ²标准提高了0.18（从39到57），对于K _s提高了0.15（从0.44到0.59）使用基于环境数据开发的ANN模型进行参数预测。结论是，从DEM和遥感信息中获得的属性可能是潜在的环境数据，可用于改善Philip渗透模型参数和开发高质量的渗透数据图。这将是针对特定地点的土壤利用，管理和环境保护的第一步。