SAMPLING STRATEGIES FOR MEASUREMENT OF SOIL HYDRAULIC-PROPERTIES TO PREDICT RICE YIELD USING SIMULATION-MODELS

Measurements of soil hydraulic conductivity and water retention functions that are needed to run soil water balance and/or crop growth simulation models are expensive and time consuming. The number of such measurements can be reduced considerably if ''classical'' soil horizons that are to be sampled can be merged into broader, hydraulic functional horizons. A hydraulic-functional horizon comprises one or more soil horizons that are similar in terms of soil hydraulic functions. A detailed soil survey was carried out in a 50 ha dryland rice area to test the concept of hydraulic-functional horizons and to obtain soil hydraulic input data for a crop growth simulation model. The feasibility of identifying hydraulic-functional horizons from visual appraisal of texture and structure during the field survey was demonstrated. Soil water retention and conductivity data determined for the functional horizons and weather data from a nearby weather station were used to simulate potential and rainfed rice yield for 25 years. A variability analysis showed that the effect of variation in soil hydraulic properties on simulated rice yield was relatively small compared to the effect of variation in year to year weather conditions. It was concluded that sampling strategies for soil hydraulic functions should be based on a first rough comparison of the impact of weather variability and soil heterogeneity on yield variability using simulation and sensitivity analyses.