How extreme can natural events get? What do they look like in a geographical
sense, how do they interact with each other, and what’s the likelihood of them
happening in immediate succession? That’s what I’m working on. As a case study,
I’m building a model to explore extreme rainfall in Peninsular Malaysia (West
Malaysia). This information is valuable for businesses like reinsurance companies, as
well as for governments.
How do extreme events look like that we haven’t seen before? We built a model to explore it
I simulated 10,000 years of rainfall in West Malaysia, using a statistical model.. Currently, there is no extreme rainfall data set for this region. The model will also be applicable to other countries and other natural hazards, such as river flooding in order to better understand the risk of future natural disasters.
In my research I introduce three different weather generators. One of them is purely parametric and quite conventional. In the other two, I introduce novel approaches. One of them reproduces low-frequency, seasonal and daily fluctuations using a non-parametric method.
My research will enable insurance
companies to improve their pricing models. With this new model, we can predict the risks of certain
disasters in a way they haven’t seen before. For instance, how large can
flooding or rainfall events get? With our model you can also see the
correlation between different natural hazards happening at the same time in
different countries, for example extreme rainfall and flooding. Is there a
correlation between them? This model will help you explore this.
I’ve always been interested in living abroad
and exploring new horizons. After a study abroad and experience in Sweden, I
finished my two degrees in International Economics and Geography. Thinking it
would be time for a break, I worked in Tokyo, learned Japanese and met lots of
interesting people from all over the globe. After being accepted for a masters
at UCL in London, I moved back to Europe, and am now finishing my PhD at
Reinsurance companies need a model to predict future extreme events. With this information, they can spread their risks, work more accurately and avoid bankruptcy. We use the information available from the OASIS Loss Modelling Framework. This is a non-for-profit organization aiming to deliver an open source catastrophe modelling service for (re)insurance companies, financial institutions and public bodies.
Governments also benefit from this model. The US already has good models for hurricane risks, for instance in Texas or Florida. But in other countries, awareness is still low. Our models can give people in developing countries the opportunity to find out whether they live in a hazard zone. They might have return level (i.e. flood) maps already, but these are are often conventional and ‘simplistic’. Using our model, they can combine them with new information.By understanding the potential risks of extreme events, governments can also take the corrective or preventative measures.
There are other people working on the OASIS project as well. One of them
is focussing on hurricanes. It would be
great to incorporate all of this information; this would be very valuable in
Natural hazards, Malaysia, flooding, weather generators, ARMA, Markov chain models.