My research concerns the development of stochastic space-time rainfall models to support the analysis of hydrological systems under present day and future climatic scenarios.
For rainfall modelling applications the Rainsim model (Burton et al., 2008) can simulate time series for a single raingauge or spatial-temporal fields of rainfall at either daily or hourly time steps for European climates. This model may be used to downscale General Circulation Model scenarios to scales of relevance to hydrological systems of up to 10000km2. This allows
comparative hydrological studies of rainfall impacts for present day and for future climates. This capability has led to a range of future climate assessments including modelling studies of flash-flood risk, urban flooding, droughts and rainfall-triggered landsliding. This rainfall model is being actively extended, refined and applied in support of a range of recent and ongoing research projects.
The Rainsim model in concert with a weather generator can produce consistent rainfall and weather time series for present day and for future climate scenarios. This potential has been realized in the EARWIG system (Kilsby et al., 2007) which provides a user friendly interface to allow rapid preparation of climatic scenarios for UK catchments. This process is considerably eased through the utilization of databases of river catchments,
meteorological observations and UKCIP02 future climate projections.
The latest UK climate projections (UKCIP09) also incorporate a version of the Rainsim model with a weather generator (Jones et al., 2009). Here the EARWIG approach is extended and updated by using a range of more recent climate model projections, by representing future uncertainty as an ensemble of equal-likelihood possibilities, by providing a wider range of future scenario time slices and by increasing the availability of these projections through the use of a web interface.
Modelling the temporal development of climate-sensitive hydrological systems may require transient-climate simulations of rainfall. A new downscaling methodology has therefore been developed (Burton et al., 2010) so that the Rainsim model can generate climatically non-stationary simulations. For a application to the Brevilles spring (Seine, France) transient single-site rainfall model parameterizations were fitted and used to simulate 100 80-year transient-climate timeseries corresponding to the projections of 13 RCM's.
Ewen J, Bathurst JC, Parkin G, O'Connell E, Birkinshaw SJ, Adams R, Hiley RA, Kilsby CG, Burton A. SHETRAN physically-based distributed river basin modelling system.
In: Singh VP; Frevert DK, ed. Mathematical Models of Small Watershed Hydrology and Applications. Highlands Reach, Colorado: Water Resources Publications, 2002, pp. 43-68.
Kilsby CG, Burton A. Developing a regionalized rainfall model for Europe.
In: Proceedings of the British Hydrological Society 7th National Hydrology Symposium. 2000, University of Newcastle upon Tyne: British Hydrological Society.
Burton A, Bathurst JC. Modelling shallow landslide erosion sediment yield at the basin scale.
In: Proceedings International Association Hydraulic Research International Workshop on Floods and inundations related to large earth movements. 1994, Padua, Italy: University of Padua.