Geoscience Reference
In-Depth Information
The insurance sector would benefi t from complementing existing risk analysis
techniques with forecasts of future risk profi les. A number of emerging trends sug-
gest that there is a growing need to utilise the power of mathematical modelling and
computational simulations to deliver prospective risk analyses. For example, it is
possible to utilise data generated by atmospheric models to improve the quantifi ca-
tion of risk for infrastructure arising from windstorms (Anastasiades and McSharry
2013
). This ability to harness data from models is particularly important for wind
farm risk assessment where insuffi cient wind speed data is typically available at the
sites of the wind turbines.
Assessing the risk of extreme events, managing this risk and correctly quantify-
ing its market price is at the heart of the insurance industry. The general assumption
underlying the majority of today's transactions in the insurance industry is that the
future will resemble the past. If this is the case, it seems reasonable to price risk
based on historical observations.
However, if evidence of different regimes exists, such as low or high levels of
hurricane activity, for example, pricing must be adjusted accordingly. Alternatively,
if we have reason to believe that there is a slow trend causing the level of risk to rise
or fall systematically over time, such as climate change, then prices should be modi-
fi ed to refl ect this.
Model uncertainty arises from poor quality data and insuffi cient availability.
Initiatives led by government or the market will be required to make appropriate
datasets more accessible. Uncertainty in risk assessment typically leads to higher
costs for insurance, making it less affordable to many of those who most need it.
Access to appropriate models for understanding and assessing risk is therefore cru-
cial not only for insurers but for the society as a whole.
Box 17.1 Types of Models
There are two types of models that could be employed for quantifying and
forecasting future risks: statistical and physical.
Statistical models
are constructed by analysing historical data and identifying
signifi cant mathematical relationships. These fl exible models can describe
trends that are emerging over time but are limited to historical events that have
been observed.
Physical models
such as Numerical Weather Prediction (NWP) models and
Global Circulation Models (GCMs), attempt to explicitly describe our under-
standing of the dynamics of the atmosphere and ocean. An investigation of
trends and the potential of exploring new regimes is possible with these physi-
cal models. Physical models have been used extensively by the International
Panel on Climate Change (IPCC) to assess the impacts of increasing green
house gases.
,
Search WWH ::
Custom Search