CICERO - Center for International Climate Research

Catastrophe modelling and the use of loss data from the insurance sector

Climate risk events - Climate risk seminars, events and conferences

Published 06.02.2020

February 25th, 2020. To be able to quantify and reduce climate risks, methods and tools are developed for assessment andmapping of risk. 

A plethora of tools are used for various types of hazards, exposure andvulnerabilities. There is, however, an imbalance in favour of hazard assessments, in comparison withassessments of exposure and vulnerability. The development of modelling tools for e.g. flood hazardshas been rapid, with good support from large sets of meteorological and hydrological data frommonitoring programs all over the world. Data on exposure and vulnerability, on the other hand, arescarce, scattered, and seldom available for academic research.

Catastrophe models (also called cat models) are one type of tools, used primarily within theinsurance sector. One aim of these models is to estimate the need for reinsurance among insurancecompanies. Large disaster events lead to huge total claims from policyholders, and it is crucial thateach company has the appropriate level of reinsurance. The strength of this type of models is thatthe insurance sector has good knowledge about the exposure and economic value of insuredproperty. In many high-income countries the level of insurance of homes and property is high, andthe value of insured objects and properties can be used as a proxy for the real economic value. Oneshould, however, remember that all kind of values are not insured, such as values related toecosystems, cultural heritage, and certain aspects of human health. The use of data from theinsurance sector is also limited by secrecy regulations to prevent unfair competition and to protectcustomer’s privacy.

The presentation will give an overview of the research on disaster risk management and modelling atthe Centre for Climate and Safety, Karlstad University ( One of the projects thatwill be presented is SPLASH (, where the objectives have been todevelop a Swedish catmodel for cloudbursts, including the secure use of insurance loss data.