In this fourth edition of the Finity Climate Risk Blog, our Climate Risk Team focus on climate change financing issues – its getting warmer but how do the economics play out.
We look at how climate change is impacting tropical cyclones and what extra costs may follow. We follow a stress testing framework developed at Finity to assess the cost and we’ll share our results in this edition. The framework is intended to provide transparency in order to assist the user to understand the nature of the risk and the key drivers.
Starting with the science
Our starting point is a broad engagement with the science. Using our partnership with Newcastle University’s Centre for Water, Climate and Land (CWCL), we asked their cyclone genesis expert, Dr Andrew Magee to scour the latest research for us in relation to the impacts of climate change on cyclones.
His findings are shown in the diagram below and although there are some wide ranges for the potential magnitude of change, the expected changes to the physical process are generally consistent across a range of scientific papers.
Interestingly, not all the impacts will make things worse. For example, the expected increases to vertical wind shear (instability) in the tropics will make it harder for cyclones to form and reduce their frequency. When they do occur however, they are expected to have stronger winds and increased rainfall.
The broad engagement with the literature leads to insights that would not have been gleaned from relying on outputs from any single or ensemble Global Climate Model (GCM). The southerly migration of cyclones is a good illustration of this as the migration appears to be happening at a significantly faster rate than the GCMs predict.
A new world
Once we have identified the physical processes that are expected to change, we can recalibrate our view of the world to reflect the changed environment. For this we used Finity’s address level cyclone risk model, fincyclone.
The recalibration was initially completed for each impact in isolation and sensitivities were used to understand how risk differs for varying but plausible levels of change. This can be useful to start quantifying the risk even before we have settled on how much the extreme weather may vary by.
To illustrate, the graphic below shows how storm surge risk in Cairns will increase (green is low risk, red is high risk) compared to the base level as a result of a 0.5m, 1m and 2m sea level rise.
There is a significant non-linear step increase in risk if sea level rise was to reach 2m as compared to 1m or 0.5m. Repeating this process across all of Australia, reveals that we expect storm surge losses to be a more than ten times higher than they are today if sea level rises reach this 2m level, for the current stock of housing.
The next step is to combine the individual impacts into plausible scenarios to understand the compound effects. A ’middle of the road’ type scenario that we have tested combines the following individual impacts:
Running this scenario through our recalibrated ‘fincyclone’ model results in some significant increases in risk with a tripling of cost in Brisbane/Rockhampton and doubling of cost Australia-Wide. The impact on far North Queensland is expected to be lower (in percentage terms) as the expanding tropics and hence polewards shift isn’t expected to impact areas already in tropical regions.
It is also interesting to note that the combined impact of the scenario is higher than the product of the individual components as these effects compound on top of each other.
With the increasing pressure from both regulators and investors for companies to be managing and disclosing their climate risk, it is important for companies to start quantifying their climate exposure. APRA has explicitly stated that they are expecting companies to be moving beyond the awareness phase and into the action phase of their climate risk management plans. The framework above can be replicated for all natural perils and is a good starting point for physical climate risk quantification.
1 Abbs 2012; Walsh 2015; Yoshida et al. 2017, 2 Abbs 2012, 3 Kossin et al. 2014, 4 Knutson et al. 2010: Patricola and Wehner 2018; Parker et al. 2018, 5 Takayabu et al. 2015, 6 Kossin 2018
Read our previous blog post: Australian natural perils cost $13.5bn