Premium
A Dynamic Multidomain Green‐Ampt Infiltration Model
Author(s) -
Stewart Ryan D.
Publication year - 2018
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/2018wr023297
Subject(s) - ponding , infiltration (hvac) , surface runoff , soil science , soil water , environmental science , hydraulic conductivity , hydrology (agriculture) , geotechnical engineering , geology , materials science , ecology , drainage , composite material , biology
Shrink‐swell soils possess dynamic hydraulic properties, which may limit the applicability of traditional models for simulating infiltration and overland flow. This study incorporates Green‐Ampt infiltration concepts into a multidomain porosity framework to account for variations in pore size distributions and saturated hydraulic conductivities caused by soil shrinkage and swelling. The model requires three input variables (initial water content, rainfall rate, and time) and up to 15 parameters to simulate infiltration and overland flow, though most of the parameters are universal constants or can be estimated from auxiliary measurements. In comparison, the classic Green‐Ampt model, which assumes constant hydraulic properties and a single domain, requires the same three inputs and up to seven parameters to use. Performance of the proposed multidomain model was verified with two data sets. The first came from a study in Mexico where time to ponding and soil matrix infiltration were quantified under simulated rainfall, and the second came from a study in Chile where overland flow was measured during irrigation experiments on runoff plots. By tuning two (Chile) or three (Mexico) parameters, the multidomain model provided accurate estimations of infiltration/runoff partitioning at multiple scales. Compared to the classic single‐domain model, the multidomain model had lower root‐mean‐square deviations (reducing simulated infiltration errors by 2–3 times) and Akaike Information Criterion (AIC) scores (ΔAIC ~100), thus providing better simulations of infiltration, ponding, and runoff. These results demonstrate that modeling hydrological processes in shrink‐swell soils necessitates separating soil properties mediated by the matrix from those associated with interblock shrinkage cracks.