Climate Risk Projections with Pattern Scaling

The financial sector is becoming increasingly interested in understanding how it is exposed to the risks due to climate change. At Climate X our multi-disciplinary team of hazard and climate scientists work to generate useful projections of risk for a variety of users.

To assess future changes in weather-related hazards we use publicly available climate model outputs from projects such as CMIP and CORDEX. However, these experiments are often not designed with decision-makers and risk assessment at the forefront. Most global climate models are still run at relatively low resolution, whereas decision makers are interested in very local changes (down to asset level). Projects that are run at high resolution, such as HighResMIP and CORDEX, often do not include all the scenarios that decision-makers are interested in and have limited ensemble members.

This talk will explore how the use of pattern scaling can address these limitations. Pattern scaling extracts the signal from local changes in atmospheric variables to global mean temperatures (GMT). It can therefore be used to explore emissions scenarios for which there are limited (or no) GCM runs. This allows us to generate custom scenarios such as global warming levels or a client’s individual projections with only a trend in GMT. Additionally, by extracting temperature uncertainty in the climate sensitivity, local hazard responses and internal variability can be separated.