Multi-GCM ensembles performance for climate projection on a GIS platform

Climate impact studies especially in the field of hydrology often depend on climate change projections at fine spatial resolution. General circulation models (GCMs), which are the tools for estimating future climate scenarios, run on a very coarse scale, so the output from GCMs need to be downscaled to obtain a finer spatial resolution. This paper aims to present GIS platform as a downscaling environment through a suggested algorithm, which applies statistical downscaling models to multidimensional GCM-Ensembles simulations. Climate change projections for the Shannon River catchment in Ireland were developed for several climate variables from multi-GCM ensembles for three future time intervals forcing by different Representative Concentration Pathways (RCP): all these processes are implemented in a GIS platform through designed and developed GIS-based algorithm. This algorithm is used as a downscaling tool in GIS environment, which is unprecedented in literature. Statistical downscaling methods were used in the projection process after a particular verification and performance evaluation using several techniques such as Taylor diagram for each GCM-ensembles within independent sub-periods. The established statistical relationships were used to predict the response of the future climate from simulated climate model changes of the coarse scale variables. Significant changes in temperature, precipitation, wind speed, solar radiation and relative humidity were projected at a very fine spatial scale. It was concluded that the main source of uncertainty was related to the GCMs simulation and selection. In addition, it was obvious to conclude that GIS platform is an efficient tool for spatial downscaling using raster data forms.