Hydro-GIS Ltd has two decades experience in flood catastrophe modelling. Director of Hydrology, Dr Harvey Rodda, was at Risk Management Solutions form 1999-2001 and led the development of their first UK river flood catastrophe model. He then went on to develop river flood catastrophe models for Benfield covering the Czech Republic and Slovakia. Hydro-GIS Ltd was set-up in 2004 working for Aon to develop river flood catastrophe models covering Austria, Hungary, The Czech Republic, Slovakia and Poland. Over the next 5 years the model development area for Aon was expanded to cover Norway, Sweden, Romania, Bulgaria, Slovenia, Croatia, Bosnia-Herzegovina, Serbia and Montenegro. With the oceanographic expertise of the late Dr Rose Wood Hydro-GIS Ltd also developed a coastal flood cat model for Norway and contributed to a study on the German North Sea and Baltic Sea coasts.
Hydro-GIS Ltd has since provided consultancy services for flood cat model development and review for a number of clients in the re-insurance sector including AIR, JLT Re, Willis, Deloitte, ABS and CEATI. The work has included studies in Europe, North America, South America and Asia. Dr Harvey Rodda also includes a section on flood catastrophe model development as part of the MSc in geophysical hazards and certificate in natural hazards for insurers course at University College London.
The expertise which Hydro-GIS Ltd has gained through flood catastrophe modelling is ideally suited to undertaking reviews and audits of flood catastrophe models for the solvency 2 requirements. Commercial cat modelling software often has extensive documentation on how to apply the software but very little information on how the hazard component of the model has been developed. All too often a modelling company will claim how their model is better than others purely on the use of a higher resolution DTM, when in reality the routines used for generating flood events can be very poor.
Particularly important model assumptions are often not explained in the documentation. For example, the first RMS UK flood cat model had a significant step increase in losses as event return periods increased. Without a knowledge of the model development this was difficult to understand. The change could however be directly attributed to how flood defence failure was simulated in the model with a simple routine removing all defences as the predicted river flows for effected rivers exceeded the 100-year return period, which gave a vast increase in flood extents.