Seasonal hydrological prediction in Great Britain – an assessment

This thesis assesses seasonal hydrological prediction in Great Britain. Firstly, the study evaluates river flow prediction using climate model output to drive a rainfall-runoff model in the Dyfi basin, Wales. Results show that climate model precipitation can not skilfully simulate Dyfi discharge. When a downscaling process is employed to generate precipitation time series, river flow forecast skill improves, but historical river flows still provide superior forecasts. Secondly, large-scale climatic control on British precipitation/discharge and European precipitation is investigated by correlation analysis. Results show spatiotemporal hydroclimatological variation, with western regions generally having stronger empirical relationships. River flow has weaker associations because of basin controls and evapotranspiration. The dynamic nature of precipitation/discharge generating mechanisms is not captured by the North Atlantic Oscillation Index. Thirdly, seasonal climate model forecast skill is evaluated. Limited skill exists over land and over all extratropical regions for forecasts beyond month-1; precipitation has lower skill than 2-metre air temperature and mean sea level pressure. Seasonal climate models exhibit higher idealised predictive skill indicating potential for future increases in actual predictive skill. In conclusion, seasonal hydrological prediction using a climate-to-river modelling chain could be improved through consideration of the uncovered spatiotemporal hydroclimatological variability and through seasonal climate modelling improvements.EThOS - Electronic Theses Online ServiceGBUnited Kingdo