عنوان مقاله [English]
نویسندگان [English]چکیده [English]
Soil water retention curve plays an important role in water and soil sciences and laboratory methods of its determination are expensive and time consuming. Therefore many researchers have modeled the soil water retention curve. One of these models based on the similarity between soil particle size distribution curve and soil water retention curve has been presented by Arya and Paris . The most important parameter of this model is called scaling parameter which can be assumed as a constant value for any soil texture or can be determined by linear and logistic methods. The goal of this research is to develop a solution for the improvement of the linear method in determining of scaling parameter (α). This improvement yields more accurate determination of soil water retention curve. In the present research, instead of one linear fit used in the Arya and Paris method, three linear fits were utilized. In this direction, pressure heads were divided into three classes of 0-100, 100-1000 and more than 1000 cm and for each class one line was fitted. In this research, according to Arya and Paris studies, UNSODA data were used for five soil textures which included sand, sandy loam, loam, loam silt and clay. For comparison of the proposed methods with the two methods proposed by Arya and Paris, the root mean square error (RMSE) was calculated. Comparing of the linear and improved methods showed reductions in RMSE values. The minimum and maximum reductions of RMSE dependent on the soil texture were 0.09 and 1.793, respectively. These reductions were repeated in comparing of the logistic and improved linear methods and the minimum and maximum reductions were 0.108 and 1.366, respectively. The reason of these reductions in the improved linear method is the accurate determination of the soil water retention curve in the moderate and high head pressures.