Estimating the Soil Water Retention Curve Using Different Empirical Models and A Piecewise Regression Method

Authors : Azize DOĞAN DEMİR, Hasan ER, Yasin DEMİR, Ramazan MERAL

Pages : 565-575

Doi:10.29133/yyutbd.1101682

View : 19 | Download : 5

Publication Date : 2022-09-30

Article Type : Research Paper

Abstract :In this study, the aim was to compare experimental and empirical methods used for estimation the soil water retention under different soil conditions. Soil samples were chosen to represent examples of heavy, medium and light soil structures. Water retention curves were obtained in the laboratory using the standard method. The van Genuchten insert ignore into journalissuearticles values(1980); insert ignore into journalissuearticles values(vG);, and the Brooks and Corey insert ignore into journalissuearticles values(1964); insert ignore into journalissuearticles values(BC); methods were used empirically. Model parameters were determined by artificial neural networks and Solver optimization methods. In addition, soil water retention SWR curves were obtained by using a piecewise regression insert ignore into journalissuearticles values(PR); method. As a result of the study, determination coefficient R2 values from 0.8946 to 0.9879 were obtained for the vG model, while the Solver method gave better results. R2 values from 0.8914 to 0.9267 were obtained for the BC method and finally from 0.9598 to 0.9717 for the PR method. No clear differences were observed for different soil structures. Finally, the use of PR has been suggested for water retention curves where breakpoints are to be included, and it is also easy to use. In addition, the vG and BC models gave reasonable results for different soil groups. It is understood that the Rosetta method provided with the HYDRUS software program can be used in the case of limited data to determine model parameters. However, the Solver method provided more reliable results and was easy to use with both models.
Keywords : Soil moisture, Water retention curve, Piecewise regression, van Genuchten, Brooks and Corey