ABSTRACT
This study aimed to compare models for predicting soil water infiltration rate and erosive rates using a rainfall simulator in different systems of common bean (Phaseolus vulgaris L.) cultivation. The evaluated mathematical models were: Kostiakov, Kostiakov-Lewis, Green-Ampt and Horton. Infiltration tests were carried out considering six treatments: bean cultivated on contour with rock barriers spaced at 0.5 m between crop rows (BC1); bean cultivated on contour with rock barriers spaced at 1.0 m between crop rows (BC2); bean cultivated downslope (BDS); bean cultivated on contour with mulch (BCM); bare soil (BS) and soil under natural cover (NC). Four replicates were considered, totaling 24 field tests. Kostiakov-Lewis's equation showed the lowest values of standard error. Soil water infiltration rate was equal to 53.3 mm h-1 in the natural vegetation treatment and to 9.49 mm h-1 in the downslope treatment. Surface roughness and the time of beginning of surface runoff were significantly higher for the conditions with mulch cover.
Key words:
semiarid; bean; soil management