Loughborough University
Browse

Pseudospectral methods provide fast and accurate solutions for the horizontal infiltration equation

Download (323.97 kB)
journal contribution
posted on 2021-11-18, 10:13 authored by Simon A Mathias, Graham SanderGraham Sander
An extremely fast and accurate pseudospectral numerical method is presented, which can be used in inverse methods for estimating soil hydraulic parameters from horizontal infiltration or desorption experiments. Chebyshev polynomial differentiation in conjunction with the flux concentration formulation of Philip (1973) results in a numerical solution of high order accuracy that is directly dependent on the number of Chebyshev nodes used. The level of accuracy (<0.01% for 100 nodes) is confirmed through a comparison with two different, but numerically demanding, exact closed-form solutions where an infinite derivative occurs at either the wetting front or the soil surface. Application of our computationally efficient method to estimate soil hydraulic parameters is found to take less than one second using modest laptop computer resources. The pseudospectral method can also be applied to evaluate analytical approximations, and in particular, those of Parlange and Braddock (1980) and Parlange et al. (1994) are chosen. It is shown that both these approximations produce excellent estimates of both the sorptivity and moisture content profile across a wide range of initial and boundary conditions and numerous physically realistic diffusivity functions.

History

School

  • Architecture, Building and Civil Engineering

Research Unit

  • Water, Engineering and Development Centre (WEDC)

Published in

Journal of Hydrology

Volume

598

Publisher

Elsevier

Version

  • AM (Accepted Manuscript)

Rights holder

© Elsevier

Publisher statement

This paper was accepted for publication in the journal Journal of Hydrology and the definitive published version is available at https://doi.org/10.1016/j.jhydrol.2021.126407.

Acceptance date

2021-04-28

Publication date

2021-06-01

Copyright date

2021

ISSN

0022-1694

Language

  • en

Depositor

Prof Graham Sander. Deposit date: 18 November 2021

Article number

126407

Usage metrics

    Loughborough Publications

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC