كاربرد مدل هاي تجربي و آماري در پيش بيني ميزان نفوذ آب در آبياري جويچه اي

Prediction of Depth of Infiltration in Furrow Irrigation Using Tentative and i3 Statistical Models

Infiltration is very important in management of surface irrigation. Mathematical models that have been developed for infiltration are generally functioning of infiltration opportunity Time. In this study prediction of average depth of infiltration evaluated is using ANN, ANFIS and MLRM methods using initial soil moisture content and furrows inflow rate. Field studies were conducted in five different cities during 1998-2007, under various soil textures. To develop coefficients of Kostiakoff Luis infiltration volumetric water balance method was used. Results showed that regression models were more accurate in heavier soils. Neural Network models were suitable for medium textured soils. These models tend to overestimate infiltration in heavy soils and under estimate in Lighter soils. However ANFIS method was capable of estimating infiltration in any situation with high accuracy. Keywords: Infiltration, Surface Irrigation, ANFIS, ANN, MLRM