Modelling the Rainfall-Runoff Relationships in a Large Olive Orchard Catchment in Southern Spain

Water balance models on the monthly scale are commonly used for planning purposes due to the relative simplicity of their parameterization and because monthly data are more readily available than daily data. The aim of this study was to evaluate the performance of three hydrological models, suitable as hydrological planning tools in rural engineering projects: one multiple linear regression (MLR) and two water balance models (one with daily and the other with a monthly time step, named DWBR and SIMPA, respectively). Runoff in both models are based on the Curve Number approach. The evaluation was conducted on a large olive orchard catchment of 308 km(2) using a daily rainfall-runoff dataset of 9 years. SIMPA and DWBR performed better than the MLR model. The SIMPA results were heavily dependent on the parameter soil water storage capacity, as determined from Monte Carlo analysis, although they showed the best adjustments (with a mean Nash-Sutcliffe Efficiency = 0.78 and 0.66 for calibration and validation, respectively). In addition, inconsistent parameterization could be obtained in both SIMPA and DWBR when the aquifer recharge coefficient was included in the set of parameters to be calibrated. The advantage of DWBR against SIMPA is that the daily temporal scale is more physically meaningful than the monthly scale. Extreme runoff values were responsible for most simulated-measured runoff deviations for the three models. Despite the good performance and conceptual advantages of SIMPA and DWBR, they should not be applied without previous calibration.