Poor, low-emitting regions bear brunt of economic burden of extreme heat

High-income countries that emit more greenhouse gases have lost relatively little, say scientists.

A herd of sheep on the cracked earth of Al Massira dam at a village some 140 km south of Morocco’s economic capital, Casablanca, on August 8. The country is facing its worst drought in at least four decades.  | Photo Credit: FADEL SENNA/AFP

The loss to the global economy between 1992 and 2013 because of climate change was between $5 trillion and $29.3 trillion, according to a recent estimate by two climate scientists, Christopher W. Callahan and Justin S. Mankin, from Dartmouth College, Hanover, New Hampshire, US. Their analysis shows that much of this loss was borne by low-income countries of tropical regions, which are not the primary drivers of human-induced global warming (Fig. 1). The analysis is based on a sample that covers 66 per cent of the world’s population. The work was published in the October 28 issue of Science Advances.

Ironically, while these countries had a 6.7 per cent average reduction in national income, the richer nations, who are historically the greatest emitters of greenhouse gases and primarily responsible for anthropogenic global warming, experienced only a 1.5 per cent reduction. The analysis assumes significance in the context of the inclusion of “compensation for loss and damage” in the agenda of the ongoing COP27 at Sharm El Sheikh in Egypt and the contentious negotiations that are expected to ensue in the coming days on the issue.

Fig. 1:  Relationship between each country’s 2010 GDP per capita percentile and the regional-average effect of anthropogenic changes to extreme heat in that country. The colours denote each country’s fossil fuel CO2 (FF-CO2) emissions anomaly (difference between its log cumulative FF-CO2 emissions and global mean log cumulative FF-CO2 emissions). The colour bar units are the log of gigatons of carbon (GtC). The black line statistical fit to data with the 95 per cent confidence intervals shaded.  | Photo Credit: Science Advances, October 28, 2022

Increased extreme heat, according to the authors, is one of the clearest impacts of global warming. The model that the researchers used combined extreme heat metrics measuring the temperature of the hottest five days (Tx5d) each year from 1992 to 2013 and an ensemble of climate models and subnational economic data to quantify the effect of extreme heat on economic growth globally. “Days that are very very hot are one of the most tangible ways that we feel climate change,” Callahan told the journal Nature. “We know that they destroy crops, they reduce labour productivity, they cause more workplace injuries,” he added.

Noting that the warmest regions of the earth, which also happen to be the poorest, were the first to experience changes in extreme temperatures as a result of global warming, the authors say that risks due to extreme heat are particularly acute in countries of these regions. Added to this is the fact that warmer years also tend to be drier and so it is a combined effect of low income and drought that influences the impact of extreme heat. “Because of their warmth,” the authors point out, “tropical regions are at risk to cross physiological temperature thresholds for human morbidity and mortality. Moreover, lower incomes make tropical economies less able to adapt to increases in extreme heat. Even modest increases in mean temperatures can cause large increases in extremes. So increased heat extremes due to warming will stress adaptive capacities in the low-income regions that have contributed least to climate change.”

An empirical gap remains

As the paper notes, despite the centrality of extreme heat in the overall impact of climate change, there has been practically no study quantifying the economic loss to nations globally because of it. Earlier studies in this direction, as the authors point out, were often sector or region specific. A theoretical and empirical gap still remains, says the paper, between the non-linearities identified at the local and sectoral level and the global assessment required to evolve appropriate climate change measures.

Also, attempts at bridging this gap have looked through the lens of changes in average temperature and temperature variability. However, as the authors point out, the physical processes driving average temperature and extreme temperatures are fundamentally different. While it is true that regions that are warmer tend to have both greater annual temperatures and extreme hot days, the analysis found that in the regions considered in the study anomalies in annual average temperatures explained only less than 13 per cent of the variation in the Tx5d values.

“Extreme heat events,” say the authors, “are driven by atmospheric blocking events and land-atmosphere feedbacks, such as soil drying, which can amplify the anti-cyclonic circulation patterns required for multiday heat accumulation. These processes take place on characteristic daily-to-weekly time scales and have length scales associated with the synoptic or finer. While related, these processes are not the same as those that determine climatological quantities such as annual mean temperature.”

According to them, climate change studies have shown that anthropogenic warming causes increased warming of the hottest days of the year more than it increases annual mean temperatures. This scientific reasoning has informed the authors’ rationale for their approach: assessing the effects of the hottest few days of the year to fully quantify the costs of global warming.

The authors use the metric of the temperature of the hottest five-day period (Tx5d) in each year, which, according to them, captures the damaging multiday periods of extreme heat “while avoiding the arbitrary [temperature] thresholds used in other metrics”. This time period of extreme heat, they argue, is consistent with the synoptic time scale of heat waves, which are generally driven by large-scale high-pressure systems that evolve on daily-to-weekly time scales associated with continental-scale atmospheric circulation.

Since they combine data from global climatological models, their empirical model includes the effects of both Tx5d and annual average temperature, and their interaction allows the effects of extremes to vary with average temperature and temperature variability, according to the authors. This, they say, allows the model to account for the heterogeneity in the effect of extreme heat and for the consequent differing responses of different regions on the basis of their annual average temperatures. Therefore, the researchers have inferred the effects of extreme heat for all regions on the basis of average temperatures data. The sample that was included for the estimate spans regions with average temperatures exceeding 30 °C and included tropical countries such as Brazil, Indonesia, and India.

The results of the analysis have shown that there is significant independence in the way both average and extreme temperatures affect a given region. The research found that while increases in average temperatures have a weakly positive effect in cold regions, they have increasingly harmful effects in warmer regions. This additional effect arises because of the interaction between warmer average temperatures and increased extreme heat intensity during the warmest part of the year.

For regions where economic data are not available, which are mostly poor countries in warm regions, assessing the economic impact of extreme heat conditions is based on an assumed latitudinal structure of responses to such conditions: tropical regions lose income significantly when there is an increase in extreme temperatures, mid-latitude regions in places in the US and southern Europe are weakly impacted transition zones, and high-latitude regions gain economically as they are too cold for optimal growth. This extrapolation procedure, the authors admit, is a key limitation of their analysis. “Gathering additional economic data in the regions most prone to climate impacts given their geography and income is an important focus for better attribution of climate impacts and therefore management of future climate risks,” they write.

The authors’ approach

To summarise the authors’ approach, estimating the economic impact of anthropogenic extreme heat requires three things: (i) the effect of extreme heat on economic growth; (ii) anthropogenic changes in extreme heat; and (iii) continuous GDP per capita data. The empirical model the authors used provided the first. To estimate the second, they used historical and natural climate model experiments from the sixth phase of the Coupled Model Intercomparison Project to calculate “counterfactual” Tx5d (Fig. 2). For the third, to ensure the availability of continuous per capita GDP data, they used a statistical model to infer regional per capita GDP time series over 1992-2013 for regions where they were not available.

Fig. 2:   Anthropogenic changes in extreme heat. (A) Ensemble mean change in each region’s average Tx5d (hottest five-day period) value between the observed and counterfactual climates estimated using CMIP6 (sixth phase of the Coupled Model Intercomparison Project) climate models. (B) Ensemble average change in the probability of each region’s counterfactual 90th percentile Tx5d value between the observed and counterfactual climates. Increases in both quantities imply that the values are higher in the observed climates than in the counterfactual climate.  | Photo Credit: Science Advances, October 28, 2022

The frequency and intensity of extreme heat events due to anthropogenic warming increased globally during 1992-2013, but the spatial pattern is heterogeneous, increasing most strongly in the tropics (Fig. 2). On average, regional Tx5d values have increased by 0.77 °C more than they would have without warming, with increases of more than 1°C in much of the tropics but less than 0.5 °C in the US and Europe (Fig. 2A). The probability of extreme Tx5d values (the 90th percentile in each region calculated from the counterfactual time series) has also substantially increased, with probabilities across regions rising by 13 percentage points on average, and even more intensely across South America, Africa, and West Asia (Fig. 2B). In contrast, the probabilities of 90th percentile Tx5d values have risen less than 5 percentage points, or even decreased, in much of the mid latitudes.

Figure 3 shows the unequal economic effects of extreme heat caused by anthropogenic changes. In tropical countries such as Brazil, Venezuela, and Mali, the per capita GDP was lower by more than 5per cent a year than it would have been otherwise. In high-latitude nations such as Canada and Finland, anthropogenic extreme heat changes lowered the per capita GDP only by about 1 per cent a year. The cumulative loss in the average Brazilian region during 1992-2013 was $39 billion (2010-equivalent dollars), which is more than half of its 2010 GDP, and $6.5 billion in the average Indonesian region, which is over 44 per cent of its 2010 GDP. Many high-income countries have lost little in relative terms but tens of billions in absolute terms due to their large economies.

Fig. 3: Unequal economic effects of anthropogenic changes to extreme heat intensity. (A) Average annual change in regional per capita GDP due to anthropogenic changes in Tx5d intensity during 1992–2013. (B) Cumulative 1992–2013 change in regional GDP in 2010 U.S. dollars due to anthropogenic changes in Tx5d intensity.  | Photo Credit: Science Advances, October 28, 2022

Since low-income countries have higher baseline temperatures and lower temperature variability, these regions both experience the signal of extreme heat from warming first and suffer most when extreme heat increases. “However,” write the authors, “the inequality of climate change extends to its causes, not just its effects. Rich countries that experience limited damage are also large emitters of fossil fuel carbon dioxide (FF-CO₂), making them primarily responsible for increases in global temperatures and associated heat extremes. Given the strong relationship between cumulative CO₂ emissions and changes in local temperature extremes, high-emitting nations can be considered directly responsible for a large fraction of warming-induced heat extremes and, by extension, the income losses suffered by individual regions.”

In sum, therefore, this first-ever quantitative analysis on the global economic burden of climate change has highlighted the following: (i) Increased extreme heat intensity has caused economic losses greatly in relatively warm tropical regions and weakly in relatively cool mid-latitude regions; (ii) human-driven global warming has increased the frequency and intensity of these heat extremes; and (iii) the effects of climate change on extreme heat have amplified underlying inequality, disproportionately harming low-income, low-emitting regions, with the historically major emitters being primarily responsible for billions of dollars of losses in the tropics. The authors add that their estimates are conservative because, for reasons of simplicity and clarity, they focussed only on the peak intensity of extreme heat, and multiple periods of extreme heat can have compounding and non-linear effects, which would enhance the effect of extreme heat.

Behaviourial adaptations

Even though many adaptations have been undertaken to deal with hot conditions even without climate change, the authors have emphasised that people are poorly adapted to deal with extreme heat. While in high-income regions, this includes air conditioning of indoor spaces and a move towards to service-dominated economies, in low-income regions it is more behavioural, such as resting in the shade, drinking more water, and shifting to non-outdoor labour. “However,” the authors point out, “there are physiological thresholds for extreme heat exposure in people and agriculture, which challenge the efficacy of behavioural adaptations.”

Their work, in effect, shows that current adaptations have not been successful in combating impacts of effects of extreme heat. The findings, they hope, will inform the way in which region-specific strategies to combat the effects of climate change are obtained, and they have stressed the need for more such adaptation investments alongside climate mitigation. “The fact that we were able to pinpoint this effect of five hottest days of the year on the whole year, as economic effects, implies that these few days have really outsized effects,” Callahan told Nature. “So investments targeted at mitigating the effects of heat extremes in the hottest parts of the year could deliver major economic returns,” he added.

The study also has brought out clearly the responsibility of rich countries to pay their share. The journal also quoted Erich Fischer of the Swiss Federal Institute of Technology Zurich as saying: “Given the unequal burden and the share of historical emissions… the global north needs to support the global south in terms of coping with these adverse effects.” One hopes that words to this effect ring loud and clear among the nations at COP27 as it begins to discuss “compensation for loss and damage”. 

The Crux

• Christopher W. Callahan and Justin S. Mankin, two climate scientists of Dartmouth College, carried out an analysis that shows that the loss to the global economy between 1992 and 2013 because of climate change was $5 trillion and $29.3 trillion. Their work was published in Science Advances.
• The authors say that extreme heat is one of the clearest impacts of global warming, but despite that there are few studies quantifying the economic loss it is responsible for globally.
• Their analysis shows that much of this loss was borne by low-income countries of tropical regions, which are not the primary drivers of human-induced global warming.
• The study clearly brought out the responsibility of rich countries to pay their share to help poor countries adapt to global warming. The work is timely as compensation for loss and damage is on the agenda of COP27 at Sharm El Sheikh.