Using sub-daily precipitation for grid-based hydrological modelling across Great Britain: hourly model performance and flood impacts under climate change

Gridded sub-daily precipitation data are increasingly available to drive national or regional hydrological models, and sub-daily river flows can be required for high flow analyses in some areas. Here, observation-based hourly 1 km precipitation data are applied with a 1 km hydrological model to simulate hourly mean river flows across Great Britain (GB). On average, performance across a large set of catchments for hourly flows is less than that for daily flows, with less difference for low than high flows and little difference in overall bias, but performance is still reasonable for most catchments. Hourly precipitation data from a convection-permitting climate model are then used to simulate hourly river flows for baseline (1980–2000) and future (2060–2080) periods, to investigate potential differences in peak flow changes derived from annual maxima (AM) of daily and hourly mean flows. On average, future changes in peak flows derived from hourly AM are higher than those from daily AM, with greater differences for higher return period peak flows and for smaller catchments in the north/west. Analysis of AM occurrence dates shows that most daily and hourly AM pairs across GB are from the same event, but some pairs for some locations have large date differences, indicating separate events. Both daily and hourly AM date distributions are generally bimodal, with more peaks in autumn and winter, but with a strong future reduction in autumn peaks and increase in winter peaks. These analyses help to inform where and when use of hourly flows may be required.