Validating the KAGIS black-box GIS-based model in a Mediterranean karst aquifer: Case of study of Mela aquifer (SE Spain)

摘要

Karst Aquifer GIS-based model (KAGIS model) is developed and applied to Mela aquifer, a small karst aquifer located in a Mediterranean region (SE Spain). This model considers different variables, such as precipitation, land use, surface slope and lithology, and their geographical heterogeneity to calculate both, the run-off coefficients and the fraction of precipitation which contributes to fill the soil water reservoir existing above the aquifer. Evapotranspiration uptakes deplete water, exclusively, from this soil water reservoir and aquifer recharge occurs when water in the soil reservoir exceeds the soil field capacity. The proposed model also obtains variations of the effective porosity in a vertical profile, an intrinsic consequence of the karstification processes. A new proposal from the Nash-Sutcliffe efficiency index, adapted to arid environments, is presented and employed to evaluate the model's ability to predict the water table oscillations. The uncertainty in the model parameters is determined by the Generalized Likelihood Uncertainty Estimation method. Afterwards, when KAGIS is calibrated, wavelet analysis is applied to the resulting data in order to evaluate the variability in the aquifer behaviour. Wavelet analysis reveals that the rapid hydrogeological response, typical of a wide variety of karst systems, is the prevailing feature of Mela aquifer. This study proves that KAGIS is a useful tool to quantify recharge and discharge rates of karst aquifers and can be effectively applied to develop a proper management of water resources in Mediterranean areas.