全球智库信息集成服务系统

Climate shows a natural variability that influences the dynamics of river discharges. In particular, intense precipitations would cause floods, while prolonged dry periods are associated to droughts phenomena. In the Mediterranean area, climate change is expected to increase the frequency of these phenomena due to variations in the precipitation partitioning in both space and time. To evaluate the impacts of these changes on the Po river daily discharges, we have developed a modelling chain that includes both climate and hydrological models. The performances of the chain are currently under testing through different simulations over the period 1971-2000. These simulations are driven by precipitation and temperature from a high resolution observed climate dataset and from the regional climate model COSMO-CLM, driven by perfect boundary conditions given by ERA40 Reanalysis and by suboptimal boundary conditions, using the global climate model CMCC-CM. The aim of these simulations is to investigate the uncertainties introduced by the components of the modelling chain and their effects on simulated discharges: the first simulation is used as reference simulation, the second one aims to evaluate how the uncertainties, introduced by the RCM, COSMO-CLM, propagate to the simulated discharges; and the last one is designed to evaluate the joint effects of the GCM, CMCC-CM, and the RCM on the simulation outputs. The results of such analysis will be used to qualify the XXI century climate projections and to correctly interpret climate change impacts on hydrological cycle in the future. The simulations performances are evaluated by comparing the precipitation and discharge seasonality and through five indices based on the flow-duration curve, that is representative of the probability distribution function of the river flow. To improve the simulation results a quantile-quantile correction is applied to simulated discharges using 1972-1990 data as calibration period and validating the results on 1991-2000. The quantile-quantile corrected simulations better resemble discharge seasonality and flow-duration curve. Results show how probabilistic bias correction helps in reducing the overall uncertainty.