资源环境科技发展态势分析平台

A robust hydrological impact assessment is indispensable for mitigation and adaptation planning. This study presents an integrated modelling methodology for evaluating climate change impacts on water availability, sediment yield and extreme events at the catchment scale. We propose the use of the spatial-temporal Neyman-Scott Rectangular Pulses (STNSRP) modelRainSim V3 and the rainfall conditioned daily weather generatorICAAM-WG, as well as the physically based spatially distributed hydrological modelSHETRAN. The change factor approach was applied for obtaining unbiased rainfall and temperature statistics. The ICAAM-WG was developed based on the modified Climate Research Unit daily Weather Generator (CRU-WG). The methodology is proposed to generate synthetic series of hourly precipitation, daily temperature and potential evapotranspiration, hourly runoff and hourly sediment discharge. We demonstrated a possible application in a 705-km(2) Mediterranean climate basin in southern Portugal. The case study showed the evaluation of future climate change impacts on annual and monthly water balance components and sediment yield, annual and seasonal flow duration curves, empirical extreme value distributions and the theoretical fits. It did not consider the possible uncertainty due to the limit of computational resources. The methodology can be well justified as follows: (a) the use of synthetic hourly instead of daily precipitation enables SHETRAN to be more capable of reproducing reliable storm runoff processes and the consequent sediment transport processes; (b) the use of SHETRAN makes possible the impact assessment to be accessible for any model grid square within the study basin; (c) the use of a statistical-stochastic downscaling method facilitates the generation of the synthetic series with unlimited length. It makes possible robust hydrological impact assessments if uncertainties related to the global climate model, regional climate model, greenhouse gas emission scenario, downscaling method, hydrological model and observational data are considered.