Runoff estimation in southern Brazil based on Smith's modified model and the Curve Number method

Abstract

► We model surface runoff for different soil classes and management system. ► Runoff can be estimated based on rainfall intensity , land slope and soil porosity . ► The Smith's modified model estimates runoff for no-tillage management system. ► Curve Number overestimates the cumulative runoff for no-tillage management system. The objective of this work was to measure and model the runoff for different soils classes at different rainfall intensities (30, 60 and 120 mm h −1 ) in Southern Brazil. A portable rainfall simulator with multiple nozzles was used to simulate these rainfall intensities. For each soil, the initial time and runoff rate, rainfall characteristics (total, duration and intensities), surface slope, crop residue amount and cover percentage, soil densities (bulk and particle), soil porosity (bulk, macro and micro), textural fractions (clay, silt and sand), and the initial and saturated soil water content were measured. The runoff measured was compared to Smith's modified and Curve Number (USDA-SCS) models. The cumulative runoff losses were 67, 45 and 27% of the total rainfall, for a Rhodic Paleudalf, Typic Quartzipsamment and Rhodic Hapludox, respectively. An inverse relationship was observed between initial runoff and the runoff rate, independently of the soil surface and rainfall conditions. Increasing rainfall intensity decreased the time to runoff and increased runoff rate. The Smith's modified model overestimated the cumulative runoff by about 4%. The Smith's modified model presented a better estimate for both higher and lower rainfall intensities (120 and 30 mm h −1 ). The SCS Curve Number model overestimated the cumulative runoff by about 34%. This large overestimate is probably due to that the model did not take into account the soil tillage system used in the field by farmers, particularly for irrigated conditions. The combination of high porosity, low bulk density and presence of crop residue on soil surface decreased runoff losses, independently of the soil texture class. Smith's modified model better estimated the surface runoff for soil with a high soil water content, and it was considered satisfactory for Southern Brazil runoff estimations. The SCS Curve Number model overestimated the cumulative runoff and its use needs adjustments particularly for no-tillage management system.