A model for stochastic simulation of daily streamflow

Assessment of fluvial flood risk is typically based on a single 1:N year extreme flow event, which cannot assure the long-term sustainability of flood defense assets. To achieve long-term sustainability, the estimation of daily streamflow time series is of paramount importance. Traditional, indirect approaches, combining stochastic simulation of rainfall with hydrological rainfall-runoff models are limited by uncertainties in model calibration and computational expense. For stochastic modelling of daily streamflow, this paper presents a simple, direct, approach that combines a Hidden Markov Model with an extreme value distribution. Model has been validated across three hydrologically distinct catchments in the UK. Results show that the model produces excellent performance (relative mean absolute differences of < 2%), appropriately captures extreme events, and is generically applicable across a range of hydrological regimes. Thus, proposed model can be readily applied to a range of catchment types for various flood risk studies.