A physically based analytical model to predict infiltration under surge irrigation.

Abstract

A significant advantage attributed to surge flow irrigation is that for the same volume of water applied the stream will advance farther along the furrow than with continuous flow. This potentially will reduce runoff and deep percolation which will improve uniformity and application efficiency where this advance phenomenon holds. The mechanism for improvement in advance time has generally been ascribed to surface sealing and surface layer consolidation. However, these phenomena do not satisfactorily explain improved advance times in sandy soils. Widely used infiltration equations which require the determination of empirical coefficients are unsatisfactory as predictors of infiltration conditions of intermittent wetting. The Green-Ampt model and a simple redistribution model are combined into an analytical model to predict infiltration under surge irrigation. The model results are compared to infiltration tests on soil columns of three soils of different soil textures. Also the model and the experimental results from the soil columns are compared to predictions made by two numerical solutions of the Richard's equation. One of the numerical models includes the effect of hysteresis by the use of Mualem's model to predict the variation of moisture content with potential, the other numerical model neglects the effect of hysteresis. A comparison of the analytical and the numerical models shows good agreement in their predictions for the soils and surge cycles tested. A comparison of predictions made by all three models shows good correlation to the experimental results. Although the number of tests done on the analytical model were limited it appears to be nearly as good a predictor of infiltration as the numerical models. The greatest strength of the analytical model is that while the numerical models took many hours to do a single run, the analytical model took only a few minutes. Both model and experimental results indicate that there was no reduction in infiltration rates or volumes infiltrated with intermittent as compared to continuous wetting. Thus the reduction in hydraulic gradient is not a factor in the reduced infiltration observed by others.