Catastrophe risk modeling and the road to net zero

Discover five ways the insurance industry can help tackle the climate crisis.

In their unique role in delivering risk information to society, insurers can indicate how insurance costs may be expected to rise over the next couple of decades. They will need to anticipate what lies ahead so they can play an active role and lobby for actions that will help sustain insurability. (Credit: BillionPhotos.com/Stock.adobe.com)

Goal number one for the upcoming COP26 conference in Glasgow, Scotland, is to secure global net-zero carbon emissions by mid-century and keep 1.5 degrees Celsius temperature rise within reach. Everyone will need to work together to accelerate the actions required to tackle the climate crisis, with collaboration between governments, businesses and civil society. How does the insurance and wider risk management industry play its part in the race to net zero, and how can catastrophe risk modeling, which has helped the industry better understand risks, achieve this?

Going back to 1989 when RMS was founded, catastrophe modeling has continued to help provide core capability to enable the functioning of the global property insurance and reinsurance sectors. The understanding that the historical record of hurricanes, floods or earthquakes is too short to reliably base the quantification of risk forms the basis of a catastrophe model.

From a physical understanding of event generation, we create a synthetic history of perhaps a hundred thousand versions of next year. Then, we can enter information into the model on the “exposure” — the buildings and infrastructure in the path of the hazard. Using vulnerability functions to link hazard to damage and loss, we can output the technical price for insuring that exposure. These fundamental modeling principles all apply in modeling physical climate change risk.

Catastrophe models will be critically important in the insurance industry’s partnering and support for delivering net-zero carbon emissions. There are five areas in which we can identify where risk assessment and risk modeling will be key.

1. Surveying future risk in order to sustain insurability

Insurers, brokers, and reinsurers will need to survey how the risks written today are expected to change over the next few decades. While today’s contracts may only last a year, insurers will be expected to continue writing coverage even where the underlying risk is rising on account of climate change. We can also anticipate inflationary pressures and expansions in exposure.

In their unique role in delivering risk information to society, insurers can indicate how insurance costs may be expected to rise over the next couple of decades. They will need to anticipate what lies ahead so they can play an active role and lobby for actions that will help sustain insurability. For example, coastal exposures may require flood defenses and properties in the forest will require action to clear vegetation.

Some changes may arrive faster than originally projected. When a particular catastrophe or set of catastrophes occurs, insurers should have plans in place for how to respond.

2. Proving the power of climate adaptation

The principal response to rising risks associated with climate change will come through “adaptation,” or the measures are taken to bring risk down to tolerable levels.

Examples include:

In order to identify which actions will be most beneficial, we need to assess the cost of the intervention compared with the associated reduction in risk. This reduction in risk may be calculated in financial terms but also the benefits to lives or livelihoods. Risk modeling must measure impacts both with and without the adaptation so alternative interventions can be explored. We will also need to include how risk is expected to change through time, as the benefits brought by adaptation expand into the future.

A simple example is a proposed flood wall. A probabilistic flood hazard and loss model is developed for several time-steps into the future. Flood loss is modeled with and without the flood wall. The reductions in loss brought by the wall are then displayed into the future, from which it is possible to answer: How do the costs of the adaptation compare with the benefits over the next few decades?

3. Expanding the agenda of modeling

Catastrophe loss modeling currently focuses on insured risks. Standard climate peril hazard modeling for insurance includes:

Yet, there are a variety of other perils that attract much less investment in modeling, either because they reflect consequences that are too frequent, too slow in onset, or too “compound” to be insurable.

For example, “sunny-day flooding,” such as king tides flooding low-lying coastal communities, can be too frequent to be insurable. Flood days may also be highly predictable when they are principally driven by extreme astronomical tides – as in 2015, the peak year for extreme high-astronomical tides.

Another uninsured peril is drought, which is considered too slow in onset and long-lasting — and with the moral hazard that people only buy the cover when they can see a drought emerging. With so many industrial processes reliant on water, from agriculture, energy production, textile manufacture and automotive production, drought can have major implications to supply chains.

Drought has systemic impacts as it may leave a legacy of dead trees, which provide fuel for future wildfires. Ground subsidence often follows drought. Water in reservoirs and aquifers can be severely depleted or exhausted. Droughts are often linked with heat waves, which may cause loss of life and interruptions in outside work. These linkages are a key component of the systemic risks of climate change.

In terms of other unquantified perils: Repeated intense rainfalls can overwhelm urban drainage and unusual winter heat may close ski resorts. There are also compound consequences around food security and migration, which will become the subject of modeling.

4. Enabling the insurance of sustainable sources of power generation

New sustainable energy generation technologies bring new risks and will require expanded models and new differentiated vulnerability metrics for measuring potential damage and disruption. For example, in Europe, there are many wind turbines located in the shallow waters of the western and southern North Sea. The windstorm catastrophe model needs to extend coverage over all the principal regions, onshore and offshore, where turbines have been installed.

Solar panels may be installed in desert areas of high sunshine and low population, where relevant catastrophe models for the principal perils of hail and severe wind may not already exist. Tidal energy and wave energy generation will have to withstand the waves from winter storms, requiring an additional parameter to be output from current extratropical cyclone models.

Geothermal energy sources are largely volcanic, with the risk that the facility is terminated by a nearby eruption. Away from volcanoes (in “hot, dry rock” projects), the challenge for energy generation can be project cancellation if an earthquake is triggered above a critical threshold (which ended a geothermal project in Basle in 2007).

For hydroelectricity, beyond damage to the dam or generation facility from a major earthquake, the dominant threat to production may come from a long-lasting drought.

All of these generation sources have associated risks. These risks could all be insured, in some cases, after developing deeper technical insight and hazard modeling.

5. Supporting the insurance of carbon sequestration, storage

Risk modeling will be needed to support the insurance of carbon credits so that sequestered greenhouse gases employed in carbon offsets do not escape back into the atmosphere. While sequestration needs to function over a long timescale — beyond that of an extended insurance coverage — insurance could offer cover for up to five-year periods (subject to renewal).

Each variety of sequestration will require its own risk assessment but with little correlation with other risks in an insurer’s portfolio. Carbon dioxide capture and storage facilities are likely to reuse exhausted natural gas fields. The way that the facility held the gas for millions of years will be used as evidence the geological trap can store the carbon dioxide long term. Each facility will need a risk assessment prior to attracting insurance. What is the chance, for example, an earthquake of sufficient size and impact could occur in the vicinity?

Other forms of sequestration attracting carbon credits include forestry. However, the credit could be lost if the forest were felled by a winter storm, killed by drought or consumed in a fire. Insurance would be taken to offset these eventualities, based on an assessment of the various risks. Windstorm damage to forests is already an insurance coverage in Sweden.

Insurance has an active role to play in supporting the quest for net zero, whether by insureds, by government agencies or by insurers themselves. In this endeavor, risk modeling can play a critical role in sustaining insurability, facilitating the evaluation of adaptation options, supplementing the peril models, enabling appropriate coverage for sustainable generation, and developing the capacity to model and price carbon offset coverages.

Robert Muir-Wood. (Credit: RMS)

Robert Muir-Wood is chief research officer at Risk Management Services. He works to enhance approaches to natural catastrophe modeling, identify models for new areas of risk, and explore expanded applications for catastrophe modeling. Robert has more than 25 years of experience developing probabilistic catastrophe models.

This article originally appeared on the RMS blog and is reprinted here with permission.

Related:

Recommended Stories