PySWR- A Python code for fitting soil water retention functions

• Published in: Computers & geosciences Vol. 156; p. 104897
• Main Authors: Memari, Sama S., Clement, T. Prabhakar
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2021
• Subjects: Brooks and Corey model; computer software; Curve fitting; data collection; dogs; Parameter estimation; Python; soil water; soil water retention; Soil water retention function; van Genuchten model; van Genuchten model; Parameter estimation; Soil water retention function; Brooks and Corey model; Curve fitting; Python
• ISSN: 0098-3004; 1873-7803
• DOI: 10.1016/j.cageo.2021.104897

Soil water retention (SWR) function is an important model that provides an empirical relationship between soil moisture and capillary pressure. We present a simple Python tool for fitting different types of SWR functions to laboratory-measured soil moisture data. Three different optimization methods including the Levenberg-Marquardt (LM) method, Trust Region Reflective (TR) method, and Dog Box (DB) method are considered. We used all three methods to fit the van Genuchten (VG) and Brooks and Corey (BC) models to ten soil moisture datasets. Our results show that the TR method, which allows the user to search for optimal parameter values within a constrained region, is the best approach for fitting these models. We developed a new graphical procedure for evaluating the guesstimates and bounds for different SWR model parameters. Overall, the TR method available in Python, together with the proposed graphical procedure, is an excellent approach for fitting both VG and BC models to soil moisture data. [Display omitted] •A Python code for fitting soil water retention models to experimental data is provided.•PySWR can fit both van Genuchten and Brooks and Corey water retention models.•PySWR supports two constrained and one unconstrained non-linear fitting methods.•A graphical approach has been developed to provide good initial guesses and parameter bounds.•Trust region reflective method is the best approach for fitting soil water retention models.