A soil water model for two contrasting tillage systems


TR Number



Journal Title

Journal ISSN

Volume Title


Water Resources Research Center, Virginia Polytechnic Institute


Techniques were developed to simulate the soil-plant environment as a semi-dynamic hydrologic system. Exponential expressions were found to give reasonable estimates of soil evaporation prior to corn plant emergence. A family of non-linear curves were used to estimate evapotranspiration after corn plant emergence. These curves assumed that pan evaporation was the integrated result of atmospheric demands and was therefore defined as the potential condition. Precipitation excess was estimated from an exponential expression that was developed by Holtan45 from the 'total water storage capacity in the Ap or A soil horizon. Corn plant heights were simulated with a 2-part non-dimensional exponential relationship. Rainfall interception by the corn plant canopy was assumed to be a function of rainfall, plant height, and density of cover with the maximum allowable quantity at full tassel limited to 0.03 inch per storm per 15,000 plants. Vertical water movement through the soil was assumed to be related to its saturated hydraulic conductivity.

The soil profile was subdivided into seven zones with the first zone terminating at the bottom of the Ap or A horizon and the seventh zone terminating at maximum root depth. Intermediate zones· were arbitrarily spaced at 6 and 12 inch intervals. The potential water holding capacity of each zone was further categorized as free water and plant available water. All recharge of the system took place in zone one. Subsequent recharge of lower depths or zones were allowed when free water existed in the zone immediately above.

All aspects of the model were programmed in the Fortran language for an IBM 360 Computer. Each logical segment of the model was sub-routined to allow minimum effort for changes and/or modification.

A comparison of simulated with observed data indicates very good agreement. With few exceptions, the simulated results were well within sampling variation.