USDA Logo
ARS Logo

  Soil Physics & Pesticide Research
Printer FriendlyPrintable version     Email this pageEmail this page
 
Search
 
 
This site only
  Advanced Search
 
Research
  Programs and Projects
 
 
  Display category headings
Research
Research >
Research Project: Decision Support Tools and Databases for Optimal Management of Chemically-Affected Soils

Location: U.S. Salinity Laboratory
Soil Physics and Pesticide Research

Title: Analysis of Soil Water Retention Data Using Artificial Neural Networks

Authors
item Jain, Sharad - NATL I. HYDROL, INDIA
item Singh, Vijay - LSU, BATON ROUGE, LA
item Van Genuchten, Martinus - rien

Submitted to: Journal Hydrologic Engineering
Publication Acceptance Date: January 7, 2004
Publication Date: September 10, 2004
Citation: Jain, S.K., Singh, V.P., Van Genuchten, M.T. 2004. Analysis Of Soil Water Retention Data Using Artificial Neural Networks. Journal Hydrologic Engineering. 9(5):415-420.

Interpretive Summary: The soil water retention curve (WRC) describes the ability of a soil to store water. This curve, also known as the soil moisture characteristic, is one of the most basic water properties of a soil by relating soil suction (or the pressure head) with the volumetric water content. As the suction increases, progressively smaller pores lose their water and hence the water content decreases. The WRC of a soil depends upon both soil texture and soil structure. Coarse-textured (sandy) soils generally release their water much quicker than fine-textured (clay) soils. Soil water retention data are often conveniently described using analytical expressions. One disadvantage of analytical expressions is that they are essentially empirical and incorporate certain assumptions about the shape of the WRC. For this purpose several investigators have used alternative approaches to mathematically describe observed retention data. An emerging modeling technique that appears well suited for this purpose is the use artificial neural networks (ANNs) which are now increasingly used in the soil, hydrologic and environmental sciences. The ANN approach is generally faster compared to its conventional counterparts, and more flexible in that they can be made as precise as needed to describe the data, while not requiring any assumption about the relationship between input (suction) and output (water content) data. The WRCs of several soils were modeled with ANNs using the measured data of soil moisture content and suction. We found that the ANNs with only suction data as input were able to describe the WRC better than several existing analytical models. Equally important, the ANN approach was able to better describe the hysteretic nature of WRCs (i.e., the curve following different paths depending upon the wetting or drying history of the soil). Our results indicate that artificial neural networks are very useful for fitting retention data, and as such may find application in a variety of modeling studies requiring information about soil moisture.

Technical Abstract: Many studies of water flow and solute transport in the vadose zone require estimates of the unsaturated soil hydraulic properties, including the soil water retention curve (WRC) describing the relationship between soil suction and water content. An artificial neural network (ANN) approach was developed to describe the WRC using observed data from several soils. The ANN approach was found to produce equally or more accurate descriptions of the retention data as compared to several analytical retention functions popularly used in the vadose zone hydrology literature. Given sufficient input data, the ANN approach was also found to closely describe the hysteretic behavior of a soil, including observed scanning wetting and drying curves.

 
Project Team
Skaggs, Todd
Van Genuchten, Martinus - Rien
Shouse, Peter - Pete

Publications

Related National Programs
  Water Quality & Management (201)
  Soil Resource Management (202)

Related Projects
   Development and Testing of Software to Predict the Subsurface Transport of Agricultural Chemicals
   Model Abstraction Techniques for Soil Water Flow and Transport
   Technology Transfer of Computer Software for Modeling Subsurface Contaminant Transport

 
ARS Home |  USDA |  Home | About Us | Research | Products & Services | People & Places  | News & Events | Partnering | Careers | Contact Us | Help |
Site Map |  Freedom of Information Act |  Statements & Disclaimers |  Employee Resources |  FirstGov |  White House