Flood Forecasting Model Selection

A comparison of gauge and radar precipitation data for simulating an extreme hydrological event in the Severn Uplands, UK

“…The greater the distance from the radar source, the greater is the error, which can result from the overshooting of precipitation by the radar beam at long ranges (Tilford et al , 2003). The nearest ground‐based radar dish, at Clee Hill, is located to the south‐east outside the Severn Uplands catchment boundary.…”

“…Gauge adjustment has been performed operationally in the United Kingdom for around two decades (Gjertsen et al , 2003). However, Tilford et al (2003) stated that, in recent years, radar measurements are not adjusted in real time but are in fact only corrected on a weekly basis to remove systematic bias due to hardware calibration errors and variation in radar sensitivity. In addition, only a few point observations are used for ground reference.…”

“…These signals can observe moderate and high precipitation rates fairly accurately, but generally produce errors when measuring light rain and drizzle (Golding, 2000). During heavy rainfall, especially short duration, intense rainfall events, such as thunderstorms, weather radars can be more useful than point estimates from gauge locations (Tilford et al , 2003).…”

“…Progress in the operational use of weather radar as an input to flood forecasting models has been slow within the United Kingdom (Moore et al , 2004) and the link between radar systems and hydrological forecasting models has not been fully exploited (Tilford et al , 2003). In part, this is due to uncertainty associated with recent changes in the quality of radar data and how this feeds through to the performance of flood forecasting (Moore et al , 2004).…”

“…There is caution in the automated use of weather radar data as an input to hydrological models for real‐time flood forecasting due to the lack of consistent data quality in time and space of weather radar data (Vehvilainen et al , 2004). Tilford et al (2003) identified a lack of knowledge in the coupling of meteorological and hydrological models, and suggested that the reliability of rainfall‐runoff modelling needs to be increased, particularly under extreme conditions. Collier et al (2002) stated that ‘the most extreme hydrometeorological events that are likely to be experienced in the United Kingdom have received only limited study from the point of view of underlying consistency and predictability’ and that research advances are required to determine how precipitation forecasts can be presented effectively to hydrologists so that they can be used effectively in combination with hydrological models to provide indications of future flows and river levels.…”