Wetting and drying of soil in response to precipitation: Data analysis, modeling, and forecasting

Carnegie Mellon University
By: , and 

Links

Abstract

This paper investigates methods to analyze and forecast soil moisture time series. We extend an existing Antecedent Water Index (AWI) model, which expresses soil moisture as a function of time and rainfall. Unfortunately, the existing AWI model does not forecast effectively for time periods beyond a few hours. To overcome this limitation, we develop a novel AWI-based model. Our model accumulates rainfall over a time interval and can fit a diverse range of wetting and drying curves. In addition, parameters in our model reflect hydrologic redistribution processes of gravity and suction.We validate our models using experimental soil moisture and rainfall time series data collected from steep gradient post-wildfire sites in Southern California, where rapid landscape change was observed in response to small to moderate rain storms. We found that our novel model fits the data for three distinct soil textures, occurring at different depths below the ground surface (5, 15, and 30 cm). Our model also successfully forecasts soil moisture trends, such as drying and wetting rate.

Additional publication details

Publication type Conference Paper
Publication Subtype Conference Paper
Title Wetting and drying of soil in response to precipitation: Data analysis, modeling, and forecasting
Year Published 2016
Language English
Publisher Association for the Advancement of Artificial Intelligence (AAAI)
Contributing office(s) Geology, Minerals, Energy, and Geophysics Science Center
Conference Title 13th Conference of the Association for the Advancement of Artificial Intelligence
Conference Location Phoenix, Arizona
Conference Date February 12–17, 2016