Author(s): Marina Cantalejo; Manuel Cobos; Agustin Millares; Juan Antonio Perez-Luque; Rosa Maria Mateos; Asuncion Baquerizo

Linked Author(s): Marina Cantalejo, Manuel Cobos, Agustín Millares

Keywords: High-mountains basins; Climate projections; Regionalization; Uncertainty; Water resources

Abstract: Semi-arid high mountainous environments have a very sensitive response to changes in hydro-meteorological forcings, with a direct impact on water resources. The variability of meteorological drivers and associated hydrological processes, make these areas of special interest as observatories of climate change. To design management strategies for upcoming decades, it is critical to estimate the impact of future scenarios and to assess its uncertainty. However, the available climate projections from global circulation models (GCM), even regionalized (RCM), present important limitations especially in areas with a complex orography. Indeed, the spatial resolution is not sufficient to capture the local behavior of variables such as precipitation and temperature, that have significant impacts on the flow regime. The bias adjustment of time series, usually done with the well-known empirical quantile mapping, does not consider the high seasonal variability. In contrast, a better frequency adaptation of dry/wet days with non-stationary approaches of bias correction have been previously reported in basins with complex topography. In this work, a non-stationary quantile mapping technique have been used in order to better adapt the results of different regionalized global circulation models to a semi-arid high mountain basin. The method was based on the fit to data of theoretical non-stationary piecewise marginal distributions that allowed to better characterize the behavior of the upper tails along time and its seasonal variability. It therefore permits the treatment of intense precipitation events out of range of observed data. Corrected projections of precipitation and temperature were then used to estimate the hydrological response of the Guadalfeo river basin, a semi-arid high mountainous basin conditioned by snow dynamics. Results from the physical-based distributed hydrological modelling show that the non-stationary bias adjustment better gives more precise estimations than other bias adjustment techniques, improving the inter-annual behavior of modelled flow response, which is clue for hydrological impact assessment and, hence, for future resilience strategies and water resources management

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221640

Year: 2022