Scenario Analysis for Physical Climate Risk: Foundations

The TCFD Status Report published early June 2019 reiterates the need for corporations and financial institutions to perform scenario analysis in a context of uncertainty over climate risk. It notes that while about 56% of companies use scenario analysis, only 33% perform scenario analysis for physical risk. Even fewer firms (43% of those using scenario analysis) disclose their assumptions and findings. The report contains useful case studies, but most focus on transition risk.

Yet a growing number of corporations and financial institutions recognize the need to integrate physical risk into scenario analysis and to develop resilience strategies that address imminent challenges from climate impacts. For example, the most recent IPCC report illustrating the impact of 1.5˚C increase in global temperatures on mean temperatures, extreme temperatures, extreme precipitation and sea levels shows that there will be significant implications for economies even with a 1.5˚C increase in global temperatures. This is still a best case scenario compared to impacts of 2˚C or 2.5˚C warming.

Scenario analysis for physical risk is fundamentally different from transition risk in its challenges and assumptions. This blog series provides our current reflections on how corporations and financial institutions can integrate physical climate risk into scenario analysis. This first blog presents the Foundations, focusing on important characteristics of climate science that affect how climate data can be used to inform scenario analysis for economic and financial risk. The next blog focuses on Equity Markets, with concrete examples of how available data can inform financial stakeholders ready to start putting scenario analysis into action. A forthcoming post will discuss scenario analysis at the asset level for real asset investments and corporate facilities.

Part 1: Foundations

The physical impacts of climate change encompass a range of direct and indirect hazards caused or exacerbated by the concentration of greenhouse gases in the atmosphere. Previous publications such as Advancing TFCD Guidance for Physical Risks and Opportunities, for which Four Twenty Seven was a lead author, provide background on these hazards as they pertain to corporate value chains and economic activities. Further information is also available in Cicero’s excellent report, Shades of Climate Risk. Categorizing climate risk for investors.

The Science: Uncertainties and Relevant Time Frames

Rapid developments in atmospheric and climate science over the past 30 years enable us to understand how these physical hazards will evolve over time due to climate change. Sophisticated global climate models project expected changes in key physical phenomena affected by greenhouse gas (GHG) concentration: heat, humidity, precipitation, ocean temperature, ocean acidification, etc. Like any other models, climate models have limitations in their accuracy and ability to correctly simulate complex and interrelated phenomena. However, it is worth noting that since 1973 models have been consistently successful in projecting within the range of warming that we have experienced in the past twenty years. More details on climate data and uncertainties from global climate models can be found in our report, Using Climate Data.

The Bad News: Impacts Are Locked In

Global climate models project different possible outcomes using scenarios called Representative Concentration Pathways (RCPs). RCP scenarios capture differing GHG emissions trajectories based on a representation of plausible global policy outcomes, without specifying the details of the underlying policies that could generate this outcome. These scenarios show that GHG emissions generated over the coming decades will influence the severity of impacts in the long-term, but also that we are already committed to some impacts through 2100 and beyond.

This is particularly noticeable over the “short term.”  When looking at the next 10 to 20 years, projections for temperature and other physical hazards do not present significant differences under different emissions scenarios (Fig 1). This is due to the massive inertia of the Earth’s systems, and the life expectancy of the stock of greenhouse gases already in the atmosphere. To put it simply, significantly reducing GHG emissions is akin to applying the brakes on a rapidly moving truck. It won’t stop instantaneously. Even if we were to stop emitting GHG altogether, climate change would persist. In the words of the Intergovernmental Panel On Climate Change (IPCC), climate change “represents a substantial multi-century commitment created by the past, present, and future emissions of CO2.”

Figure 1. Temperature increases under different GHG emissions scenarios in the near term. Source: IPCC, as published by Climate Lab Book.

This is by no mean an invitation to give up on reducing GHG emissions. Quite the opposite. Emission reductions are critical to curbing long term impacts and avoiding irreversible effects to our environment (Fig. 2). But for organizations looking at climate data and scenario analysis for risk management and strategy, with a focus on the coming decade(s), this is a critical fact to understand.

Figure 12.5 | Time series of global annual mean surface air temperature anomalies (relative to 1986–2005) from CMIP5 concentration-driven experiments. Projections are shown for each RCP for the multi-model mean (solid lines) and the 5 to 95% range (±1.64 standard deviation) across the distribution of individual models (shading). Discontinuities at 2100 are due to different numbers of models performing the extension runs beyond the 21st century and have no physical meaning. Only one ensemble member is used from each model and numbers in the figure indicate the number of different models contributing to the different time periods. No ranges are given for the RCP6.0 projections beyond 2100 as only two models are available.Source IPCC AR5: Collins, M., R. Knutti, J. Arblaster, J.-L. Dufresne, T. Fichefet, P. Friedlingstein, X. Gao, W.J. Gutowski, T. Johns, G. Krinner, M. Shongwe, C. Tebaldi, A.J. Weaver and M. Wehner, 2013: Long-term Climate Change: Projections, Commitments and Irreversibility. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 1029–1136, doi:10.1017/CBO9781107415324.024.
Figure 2. Temperature increases under different GHG emissions scenarios through 2300 from IPCC AR 5: Figure 12.5 | Time series of global annual mean surface air temperature anomalies (relative to 1986–2005) from CMIP5 concentration-driven experiments. Projections are shown for each RCP for the multi-model mean (solid lines) and the 5 to 95% range (±1.64 standard deviation) across the distribution of individual models (shading). Discontinuities at 2100 are due to different numbers of models performing the extension runs beyond the 21st century and have no physical meaning. Only one ensemble member is used from each model and numbers in the figure indicate the number of different models contributing to the different time periods. No ranges are given for the RCP6.0 projections beyond 2100 as only two models are available.
Aside from RCP-driven scenarios, there is, of course, a broad range of possible increases in temperature (and other climate hazards) even when looking at the 2030-2040 time frame. These plausible differences are not so much policy-driven as science-driven, demonstrating the different possible responses from the Earth’s systems to the existing stock of GHG.

These differences have significant implications for businesses and investors. For example, a model of sea level rise developed in 2018 incorporates accelerated rates of melting and recent advancements in modelling ice-cliff dynamics to capture extreme risk of coastal flooding. The model shows the Atlantic rising by 1.2m (3.9ft) by 2060 on the Florida coastline, which would equate to widespread flooding of coastal properties with potential domino effects on real estate prices across the state (Fig 3). The ‘intermediate’ scenario, however, most often used for planning, predicts only a 55cm (1.8ft) rise in water levels. While reducing GHG emissions does reduce the risk of more extreme sea level rise millennia into the future, year after year, scientists find that the Antarctic is warming faster than anybody predicted, and there is increasing concern that the process of ice sheet melt may be too far advanced to be stopped.

Figure 3. Building-level perspective of inundation in downtown Miami under 1m (3.3ft). Red buildings are those most likely to be impacted and blue areas are inundated. Source: NOAA Office for Coastal Management.

Thus, performing scenario analysis where the key variable is GHG emission reduction targets may not be an accurate representation of the range of possible outcomes for the near future. Rather, looking at high and low warming projections across a large set of models to understand the range of potential outcomes (independent of the underlying RCP scenario) is a better way to understand potential risk. In other words, physical risks over the next 10-20 years are largely independent from policy decisions and emission pathways, and a rapid, orderly, effective transition to a low-carbon economy could still come with massive physical impacts as these processes are already under way, fueled by the past 150 years of GHG emissions.

The Worse News: Tipping Points

Another challenge is that climate scientists are not currently able to model certain possible impacts from climate change, commonly known as “tipping points.” Tipping points is a catch-all term for a wide range of phenomena that may accelerate feedbacks due to climate change, though the timing or probability of their manifestation is currently not well understood. The phenomena are known as tipping points because past a certain threshold, they may not be reversible, even with a dramatic reduction in GHG emissions. Tipping points of most concern to the scientific community are presented in this report from the Environmental Defense Fund.

Figure 4. Melting permafrost is a powerful feedback loop exacerbating climate change. Source: UNEP, Woods Hole Research Center.

Some tipping points catalyze “feedback loops” which can worsen and dramatically accelerate climate change beyond human control. Such is the case, for example, with melting ice sheets, which would not only lead to catastrophic sea level rise, but would also further heat up the planet as the poles’ albedo (reflectivity) is reduced after the ice disappears. Thawing permafrost could lead to massive amounts of methane, a particularly powerful GHG, to be released from the frozen tundra into the atmosphere (in addition to many direct impacts for local communities, infrastructure  and ecosystems in the region) (Fig. 4).

 

Tipping points further reinforce uncertainty about severity and timing of these extreme impacts and the limitations of using RCP scenarios to understand the range of outcomes for physical risk.

Another source of uncertainty for physical climate impacts are knock-on effects, or ‘indirect hazards,’ from the primary expression of global warming (rising temperature and humidity), ranging from biodiversity losses and ecosystem collapses, human health impacts, impacts on crop yields, pests and soil, impacts on human society, increased violence, and rates of war and migration, etc. (Fig 5)

Figure 5. The likely risks to human and natural systems under several global warming scenarios, with dark purple representing high risks of severe impacts with limited reversibility and white indicating no attributable impacts. Source: IPCC, 2018

These indirect or second-order hazards are as relevant as first-order impacts to understand the implications of physical climate change on economic outcomes, but they’re not captured by RCP scenarios and many require stand-alone models that cannot easily be integrated into one clean set of scenarios.

Conclusion

Scenario analysis is often approached from the perspective of transition risk, where policy developments and GHG emission targets are the key drivers of risk pathways over the next 10 to 30 years. Physical risk, however, requires a different approach. Impacts over the coming decades are largely locked-in and are only marginally influenced by GHG emission pathways. In contrast, uncertainty looms large regarding how severe these physical hazards will be, and exploring a range of possible outcomes for physical risk, including looking at tail-risks, provides important insights for risk management and financial analysis.  In summary, the current state of scientific knowledge and the nature of the Earth’s atmospheric systems call for the developments of scenarios that are decoupled from transition/policy scenarios and instead focused on key scientific drivers of uncertainty and risks that may be experienced regardless of policy decisions over the short to medium term (2020-2040).

While efforts to develop easy-to-use tools for physical risk analysis are nascent, organizations can still extract important insights from climate data and leverage estimates of risk exposure across portfolios. Our next blog in this series provides examples of how financial institutions can leverage data on physical risk exposure in equities to inform some early scenario analysis in equity markets.

———————————–

Four Twenty Seven’s data products and portfolio analytics support risk reporting and enable investors and businesses to understand their exposure to physical climate risks across asset classes.