Parameterisation of hydrological model for application in ungauged basins

From a classical point of view, hydrological models can only provide warnings on key points of the river drainage system, previously identified as risk points and with available hydrological records. However the new ideas about flood warning and active control strategies require tools able to take into account the particular spatial and temporal characteristics of a given storm, put in interaction them with the geomorphological and hydrographic characteristics of the area being affected and produce real-time forecasts of the flood risk whatever it could be the location potentially affected at a given moment. Unfortunately, most of locations sensitive to flash flood occurrences (campings, industrial areas, road bridges, etc.) are located in ungauged basins and have no available flow measurements in their proximity. Therefore, the problem of how to issue sensible flow forecasts on ungauged basins should be addressed. This report summarizes the research activities made for GRAHI-UPC staff inside of FLOODSite ECProject related with the estimation of parameters of a distributed hydrological model based on the relationship of physical, geomorphological and hydraulics characteristics of the catchments with their parameters, i.e., based on the link between distributed catchment characteristics and hydrological model parameters. The estimation parameters ethodology was developed using the distributed conceptual rainfallrunoff model DiCHiTop (Corral, 2004) on the 1000 km2 Besòs river basin. Using information from 7 level gauges inside the Besòs catchment, parameter calibration was made for all these points. Then distributed characteristics and calibrated parameters of each subcatchment were statistically analyzed to evaluate their relationships. It was focused on: 1) analysis of topography and derived drainage features (for example local slopes) in relation to model routing parameters and 2) analysis of land use and vegetation cover, as well as topography via the topographic index, as descriptors of runoff generation process. From these analyses, a proposition of parameter distribution based on distributed catchment information was formulated. The proposed methodology for parameter regionalisation was extended and validated on a test case area on the Catalonia region, the Anoia catchment (904 km2). This basin is located inside the area covered by the warning centre of the Water Agency of Catalonia (ACA).Floodsit