Valuing the Mortality Risks of Climate Change
An interview with CEGA affiliated faculty Tamma Carleton
How will future temperature increases caused by climate change impact global human health? — Examining the first globally comprehensive and empirically grounded estimates.
In honor of Earth Day last week, we are highlighting research led by CEGA affiliated faculty Tamma Carleton (UC Santa Barbara), estimating the mortality risk due to future temperature increases caused by climate change. By accounting for an often-overlooked factor — the costs and benefits of climate adaptation efforts — Carleton, CEGA affiliated faculty Sol Hsiang (UC Berkeley), and co-authors are transforming our understanding of how climate change will impact real lives. Their research is one way the CEGA network is helping transform how decision-makers address climate change and other environmental risks. CEGA’s Data Science for Development portfolio is advancing work like this by leveraging approaches that use non-traditional data — such as satellite and remotely sensed data — to answer urgent questions about environmental change, economic development, and more.
CEGA: At CEGA’s Measuring Development conference a few weeks ago, you shared your research on the global mortality consequences of climate change and the economic impacts of these consequences. Could you tell us your motivations behind this study and what you were looking to model?
Tamma: Communities are already experiencing a changing climate as record-breaking warmth becomes a familiar trend around the globe. In this study, we aim to quantify the risks these extreme temperatures pose to future global human health under climate change. In particular, we study changing mortality rates under a changing global climate. What role might efforts to adapt play in reducing those risks? And, at what cost?
To answer these questions, we assemble a large dataset of subnational mortality records covering ~40% of the global population. We pair these data with climate observations in a flexible statistical model to quantify how death rates across the globe have been historically affected by temperature. We additionally estimate how adaptations to different climates can influence the relationship between mortality and daily temperatures, and we quantify the effectiveness of rising incomes for reducing temperature-related mortality.
Finally, we use these data-driven results to project the future impact of climate change on mortality rates, as well as the costs and benefits of adaptation measures populations are likely to undertake to protect themselves from warming. We show that while global average mortality rates are likely to rise substantially under a high emissions scenario, the mortality burden of climate change is disproportionately borne by today’s poorest locations.
C: How do you hope your research could contribute to the Biden Administration’s climate policy in the future? What added resources does it suggest might need to be put toward climate adaptation and mitigation going forward?
T: There are at least two ways that I hope these findings can be useful in forming climate policy.
First, the Biden Administration is working to update the United States Government’s Social Cost of Carbon (SCC). The SCC represents the total dollar value of all future damages caused by the release of one additional ton of CO2. The SCC is central to U.S. climate policy, as it is used to determine whether the benefits of any emissions mitigation policy outweigh the costs of lowering emissions.
While the SCC is critical in determining whether climate policies are enacted and how stringent those policies are, its value is incredibly outdated. Our new estimates of the mortality component of the SCC — that is, the value of the mortality risk caused by an additional ton of CO2 emissions — are ten times higher than the corresponding estimates behind the current U.S. SCC. We hope that our findings contribute to an updated SCC that better reflects current scientific and economic evidence.
Second, our findings describe hyperlocal health damages from climate change. This high spatial resolution uncovers enormous disparities in the impacts of climate change on mortality across the globe, making clear that adaptation efforts will need to be targeted to today’s poorest populations and those already living in hot climates. While these inequities are most stark when looking across the globe, impacts remain highly unequal even within the United States, pointing toward both domestic and foreign policies that put distributional impacts center stage.
C: What should other researchers take away from this study and what further research is needed to better model the economic impacts of climate change worldwide?
T: A key finding for future research is the importance of modeling the benefits and costs of adaptation when conducting climate change impact analysis. We find that failing to account for adaptation overstates the future impacts of climate change by a factor of 2.6. However, populations differentially invest in adaptation, with poor regions bearing most of the mortality-related burden of climate change as rising mortality risk, and wealthy region experiencing net declines in mortality rates alongside increases in adaptation spending.
Second, our study uncovers one important link between climate change and human health, focusing on the mortality implications of extreme temperatures. However, multiple other dimensions of health are likely to be influenced by rising greenhouse gas emissions, such as vector-borne and water-borne diseases, childhood undernutrition, and mental health. Data-driven, high-resolution, globally-comprehensive estimates of these climate change impacts are needed to inform mitigation and adaptation planning, both within the United States and abroad.
C: What else are you working on that might help researchers and policymakers better address climate change in low-resource contexts?
T: Evidence-based policy and the research behind it are often constrained in low-resource contexts by limited data availability. For example, our global mortality estimates infer mortality sensitivity to temperature for all of Africa based on data from other locations, because no population-wide, subnational, panel datasets exist tracking death rates for any country across the continent. Additionally, weather station data are sparse and irregularly provided.
In a separate project, I’m collaborating with computer scientists and other economists from Berkeley to leverage satellite imagery and machine learning to begin to fill some of these data gaps. Importantly, we’re focused on building a generalizable and accessible approach to learning about environmental and socioeconomic conditions from space. That means our method can be used to predict many different outcomes from satellite imagery using a laptop computer and a simple linear regression. We hope that this tool increases access to the powerful technology of remote sensing and can contribute to building a robust evidence base on climate change across the developing world.
