Performance of three infiltration models under surge irrigation

Abstract

Three models were developed to predict infiltration under a surge-flow irrigation regime. One model (2P) utilized the Kostiakov-Lewis infiltration equation. The other two models (2PF and 3P) utilized the modified Kostiakov-Lewis infiltration equation. All three models combined continuous-flow furrow irrigation volume-balance hydraulic models presented by Reddell and Latortue (1986, 1988) with a numerical variation of the Cycle Ratio-Time Model (CRTM) for surge irrigation presented by Blair and Smerdon (1987). The volume–balance component of the models estimated the soil's continuous-flow furrow irrigation infiltration function using first surge-cycle advance data. The numerical CRTM component used the estimated infiltration function in conjunction with the surge-flow irrigation infiltration opportunity-time to predict infiltration. An assessment of overall model accuracy was based upon an infiltrated volume root mean square error (RMSE) statistic. Overall, the 2PF model returned the lowest RMSE values and the 3P model was the next most accurate. The 2P model was the least accurate. While the 2PF model was the most accurate, this model has limited applicability since it requires inflow and outflow measurements throughout an irrigation event to evaluate the soil's basic infiltration constant. Despite its good overall performance, anomalous results from the 3P model suggest that further testing is needed. Although the 2P model was the least accurate of the three models evaluated here, and although it consistently underpredicted infiltration, it appears to be the most stable and effective of the three models evaluated.