A London-based non-profit firm is aiming to provide a "new method of catastrophe modelling" this year, according to Lloyd’s.
Oasis Modelling Framework LP, which is “supported” by insurers including Lloyd's, will let insurance firms "undertake full uncertainty calculations within a given model combination and to test a variety of models to illustrate the sensitivity of results from each model," according to an article posted last week to Lloyd's website.
"Its ambition aim is to deliver a new method of catastrophe modelling to the industry in 2013," wrote Kassy Dignam, digital content executive for Lloyd's.
Founded in 2011, Oasis LMF plans to use IBM Corp.'s Netazza data warehouse appliance and computer clustering methods to form a "cloud" computing architecture. It plans to create an "open community" to include hardware and software firms and to create what it calls "a marketplace for models, data and services that will lower the barrier to entry, so that more parties can develop modules and users can obtain relevant information about catastrophic risk.
"By providing the simulation kernel and financial module and with the development of cat model standards, we will allow a broad range of academic and specialist hazard organizations to provide their intellectual capital direct to the desktop of the insurance industry," Oasis LMF states on its website.
One aim is to let insurance firms "demonstrate to regulators a detailed understanding of the models and their inherent uncertainties."
Dignam noted in her article that Oasis LMF associate members include the British Met Office. In addition to running the public weather service in the U.K. , Met Office operates the the Hadley Centre, a climate research unit funded by the British government.
Other associate members of Oasis LMF cited by Dignam include University College London, Karen Clark & Co, JBA Risk Management and Perils AG.
"The European Union funded Climate-KIC programme is backing the initiative and has sponsored numerous model developers," she wrote.
Oasis LMF aims to help users tests a variety of models and to create both alternative models for existing perils and new models for what it calls "unmodelled perils."
It also aims to improve computer run times to reduce the need to simplify assumptions.