Modeling Runoff from Semi-Arid Agricultural Lands in Northwest Iran

Transformation of rainfall into runoff over an area is a very complex process which exhibits both temporal and spatial variability; runoff in a defined area can be affected by factors such as topography, vegetation, rainfall characteristics and soil properties. This study was conducted to develop an empirical model using the rainfall characteristics and soil properties for predicting runoff from dry-farming lands in a semi-arid agricultural area in Hashtroud, Northwest Iran. Runoff plots (1.83 m x 22.1 m) in triplicate were installed in thirty-six sloped dry-farming lands in the study area. Runoff under natural rainfalls was measured in each plot during a 2-year period. The results showed that runoff for 41 runoff-producing rainstorm events with duration longer than 30 min was largely associated with a rainfall index obtained by multiplying the positive square root of rainfall depth (h(0.5)) by the logarithm of the maximum 30-minute intensity (LogI(30)) (R-2 (=) 0.81). Runoff significantly varied among the plots (P < 0.001), which was considerably related to the effective soil properties (R-2 = 0.74), i.e., soil permeability (Per) and aggregate stability (AS). A multiple linear regression model was developed between runoff and the rainfall index (h(0.5)logI(30)) and the effective soil properties (AS and Per). Evaluation of the model using 34 runoff-producing rainstorm events that occurred during the next two years resulted in high values of the efficiency coefficient and R2 (0.88 and 0.91, respectively), which revealed that the model developed in this study could be used in predicting runoff from the dry-farming lands in the semi-arid regions.