key: cord-0262198-k8uicjy1
authors: Carlson, C. J.; Colwell, R.; Hossain, M. S.; Rahman, M. M.; Robock, A.; Ryan, S. J.; Alam, M. S.; Trisos, C. H.
title: Solar geoengineering could redistribute malaria risk in developing countries
date: 2020-10-25
journal: nan
DOI: 10.1101/2020.10.21.20217257
sha: 994339012affb0155e575e53f65013a202cd3137
doc_id: 262198
cord_uid: k8uicjy1

Solar geoengineering is often framed as a stopgap measure to decrease the magnitude, impacts, and injustice of climate change. However, the costs or benefits of geoengineering for human health are largely unknown. We project how geoengineering could impact malaria risk by comparing transmission suitability and populations-at-risk today against moderate and high emissions scenarios (RCP 4.5 and 8.5) with and without geoengineering over the next half-century. We show that if geoengineering deployment cools the tropics, it could help protect high elevation populations in eastern Africa from the encroachment of malaria, but could increase transmission in lowland sub-Saharan Africa and southern Asia. Compared to extreme warming, we also find that by 2070, geoengineering would nullify a projected reduction of nearly one billion people at risk of malaria. Our results indicate that geoengineering strategies designed to offset warming are not guaranteed to unilaterally improve health outcomes, and could produce regional trade-offs among Global South countries that are often excluded from geoengineering conversations.

considered. 48 Among these proposed radical schemes is solar geoengineering (also called solar radiation 49 modification, or SRM), whereby the deliberate injection of aerosols into the stratosphere could 50 reflect a small amount of sunlight back to space, cooling Earth. SRM has never been tested 51 outdoors, but large volcanic eruptions are evidence that increasing stratospheric aerosols would 52 cool the planet (2,3). While responses of temperature and precipitation to SRM have been studied 53 in detail, very little is known about possible consequences for humans or ecosystems (4,5). 54 Although there is high confidence in the adverse impacts of global warming-and therefore, 55 motivation to avoid unmitigated warming-the climate of a world with SRM would still differ in 56 notable ways from the current or preindustrial climates. As a result, there is no a priori reason to 57 think that geoengineering would necessarily improve health outcomes uniformly across regions or 58 health burdens. One study has examined the potential impact of SRM on skin cancer and pollution-59 related illness (6), while another has explored health impacts of urban heat stress (7). However, 60 very little is known about possible impacts on infectious diseases, which account for a much higher 61 proportion of global mortality (especially in low-and middle-income countries). This constitutes 62 a major research gap (8), especially given increased attention on SRM and the critical need to 63 consider risks to health in policymaking on climate change.

Of all the possible infectious diseases to prioritize for health impact assessments, many of the best 65 candidates are vector-borne diseases, given their massive global burden and their well-66 demonstrated (and easily forecasted) climate linkage. Pathogens transmitted by arthropod vectors 67 like mosquitoes or ticks are particularly sensitive to temperature, which determines both their rate 68 of replication in hosts, and the activity and metabolism of their ectothermic (cold-blooded) vectors. 69 Together, these produce a pattern where their transmission responds unimodally to temperature, 70 in a roughly Gaussian response curve (9). Thanks to recent advances in experimental and modeling 71 approaches, scientists can rapidly evaluate these response functions, and confidently identify 72 thermal optima (Topt) and outer limits of transmission (Tmin and Tmax). Using these parameters, Thanks to these approaches, an emerging body of evidence shows a high confidence link between 80 global climate change and a potential resurgence of vector-borne diseases. By 2070, climate 81 change is expected to increase the global population at risk of the Aedes mosquito-borne dengue 82 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 25, 2020. countries agreeing on an optimal amount of geoengineering. We selected the G3 and GLENS 134 scenarios because they simulate the more "realistic" creation of a stratospheric aerosol layer, 135 deploy geoengineering within emissions scenarios widely used in climate impact assessments, and 136 differ in their injection strategies in tropical regions where the burden of malaria is highest. Other 137 scenarios were not used due to potential layers of additional impacts on malaria that might require 138 dedicated attention (e.g., the potential for termination shocks in the G4 scenario, which might 139 interact with how population immunity changes on the timescale of decades, with complex impacts 140 on disease transmission). 141 We use these scenarios to predict the transmission boundaries and intensity of falciparum and 142 vivax malaria in their respective regions, and project their shifting transmission and seasonality 143 over the next half-century (2020 to 2070). We limit our analyses to Africa, Asia, and Latin 144 America, where malaria is endemic today, as a subset of the total global shift in favorable 145 temperatures to higher latitudes that has been well-documented for tropical vector-borne diseases.

While temperatures may become more favorable for malaria transmission at higher latitudes, This approach cannot necessarily predict total incidence, because major factors such as population 165 density, malaria control, or elimination progress are not included. However, mapping R0(T) can be 166 a first-order proxy of transmission suitability, and, by comparing between scenarios, can indicate 167 where the intensity of transmission and the potential resulting burden of malaria would be higher 168 or lower in different pathways. We make two such comparisons: first, we compare possible futures 169 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint 6 for 2070, with and without SRM deployment; second, we compare a future with SRM deployment 170 against the present-day climate. 171 We found that across all scenarios, the highest intensity transmission in 2070 will remain where 172 malaria is hyperendemic today, particularly sub-Saharan Africa and the Indian subcontinent 173 (Figure 1) . However, we identified several major differences in transmission suitability, which 174 were more pronounced when geoengineering is used to offset warming from RCP 8.5, the high 175 emissions scenario (Figure 2 ). Compared to climate change without solar geoengineering, we To approximate these relative burdens, we calculated population at risk from stable and unstable 209 transmission in these areas, using future population projections based on the shared socioeconomic 210 pathways paired with our climate scenarios (SSP2-RCP 4.5 and SSP5-RCP 8.5; see Methods).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint

In climate change scenarios without geoengineering, much of the world faces increased malaria 212 risk over at least the next decade; in the longer term, regional redistributions of both falciparum 213 and vivax malaria risk are expected throughout the tropics, on the order of hundreds of millions of 214 people. For both RCP 4.5 and RCP 8.5, substantial increases in populations at risk from malaria 215 are projected in east Africa and to a lesser degree central Africa, particularly in high elevation 216 regions where colder temperatures have previously limited malaria transmission. In west Africa, 217 malaria risk also increases, but is dependent on the amount of warming; in RCP 8.5, population 218 exposed to stable transmission risk peaks mid-century and then declines, as much of the region 219 becomes too warm. South and southeast Asia show a similar trend, where warming temperatures 220 in RCP 8.5 lead to massive declines in population at risk (~200 million people in each region), 221 including some populations shifting from stable into unstable risk. In the rest of the tropics, 222 warming temperatures lead to little change or mild declines in total risk, again with shifts from 223 stable to unstable transmission risk. 224 We find that solar geoengineering to stabilize planetary temperatures at 2020 levels despite 225 moderate greenhouse gas emissions (the G3 scenario) is projected to initially mitigate population 226 at risk of malaria very slightly, but this effect is short-lived and uncertain, and the total population 227 at risk converges on climate change without geoengineering (RCP 4.5) by mid-century (Figure 3) . 228 This largely holds across regions (Extended Data Figures 1 and 2) ; the most pronounced (Figure 3) . We also find strong regional patterns for these projected impacts. For Finally, we observed that uncertainty about both baseline climate change impacts and 249 geoengineering impacts showed a tremendous degree of regional variation, which was largely 250 consistent between mid-and high-emissions scenarios. In Latin America and central Africa, 251 changes were highly consistent across climate model runs, and differences between scenarios were 252 minimal. In east and west Africa, we found that differences between climate scenarios were 253 pronounced, but again largely consistent across runs by the end of the century. However, we found 254 that southern Asia-a hotspot of projected changes, and the global hotspot of vivax malaria 255 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint 8 endemicity-showed a tremendous degree of variability, so much so that the differences between 256 the RCP 4.5 and G3 scenario are proportionally much smaller, and essentially impossible to 257 evaluate given model uncertainty (Extended Data Figure 3) . This likely reflects a combination 258 of climate model uncertainty and highly aggregated populations, where any one pixel's suitability 259 for malaria may have a marked impact on the total population at risk. warming is projected to increase risk, our results show that case by case assessment is urgently 281 needed, and regional differences in outcomes must be disaggregated. Without specific research, 282 assumptions that solar geoengineering's health impacts would be intrinsically fair and effective 283 are unsupported, even compared to the most extreme scenarios for climate change. Emissions 284 reduction for climate change mitigation is widely agreed to produce major net benefits across the 285 health sector; climate geoengineering strategies will likely not be as easily weighed, and may be is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint willpower. After COVID-19, the burden of malaria -and many other infectious diseases that have 299 been the subject of large-scale elimination programs -hangs in the balance. Even if COVID-19 is 300 eventually contained, most countries' health systems will remain weakened, and some will face a 301 resurgence of diseases that were well-controlled or even nearing elimination before interventions 302 were disrupted (39-41). In a world that eliminates COVID-19, "builds back better," and recoups 303 these losses, malaria transmission might not be an important issue for climate policy. In a world 304 that fails to do so, with lasting damage from the pandemic, malaria might still be one of the biggest 305 climate related priorities for developing countries, and therefore, one of the greatest potential 306 downsides (or negative repercussions) of solar geoengineering.

Our study underscores the need for involvement and leadership of developing countries throughout is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint

Climate projections 337 We used two sets of scenarios with different greenhouse gas concentrations and geoengineering The total value of this formula is then scaled between zero and one. Andean). We focused on falciparum risk for East Africa given both the more severe presentation, 441 and the difficulty of comparing the relative burden of the two given diagnostic challenges.

To estimate future populations at risk, we paired representative concentration pathways (RCPs) 444 with shared socioeconomic pathways (SSPs) using the conventional scenario matrix, which To project populations-at-risk from malaria, we rasterized region boundaries (using a fractional 455 approach to proportional area, rounded to the nearest 1% of a grid cell), and then summed the 456 populations in every grid cell that fell within thermal bounds and precipitation cutoffs.

Temperature cutoffs were applied at the daily level and summarized for each pixel, and classified 458 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint Extended Data Fig. 4 . Populations at risk from vivax malaria in RCP 8.5 versus GLENS. 520 Transmission risk is split below, into stable risk (left) and unstable risk (right). Populations at risk 521 are reported in millions of people. 522 523 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint

Paris Agreement climate proposals need a boost to keep warming well below 526 2 C

Land surface cooling 528 induced by sulfate geoengineering constrained by major volcanic eruptions

Estimating global agricultural 531 effects of geoengineering using volcanic eruptions

Potentially dangerous consequences for biodiversity of solar 533 geoengineering implementation and termination

Quantifying the impact of 537 sulfate geoengineering on mortality from air quality and UV-B exposure

Impact of Solar Geoengineering on Human Health from Urban Heat 540 Stress

Climate engineering needs a clean bill of health

Thermal biology of mosquito-borne disease

Mapping the distribution of malaria: current approaches and future 546 directions. Analyzing and modeling spatial and temporal dynamics of infectious diseases

Global expansion and 549 redistribution of Aedes-borne virus transmission risk with climate change

The current and future global distribution and population at risk of 552 dengue

Warming temperatures could expose more than 1.3 billion new people to 554 Zika virus risk by 2050

The effect of climate 556 change on Yellow Fever disease burden in Africa

The global distribution and burden of dengue

The effect of malaria control on Plasmodium falciparum in Africa between 559

Malaria eradication within a generation: ambitious, achievable, and 561 necessary

Identifying and combating the impacts of COVID-19 on malaria

Malaria and Parasitic Neglected Tropical Diseases: Potential Syndemics 565 with COVID-19?

Shifting transmission risk for malaria in Africa 568 with climate change: a framework for planning and intervention

Climate change unlikely to increase malaria 571 burden in West Africa

Mapping physiological suitability limits for malaria in Africa under climate 573 change. Vector-Borne and Zoonotic Diseases

Climate 575 change could shift disease burden from malaria to arboviruses in Africa. The Lancet 576 Planetary Health

Assessing the role of 578 climate change in malaria transmission in Africa. Malaria Research and Treatment

Projecting malaria hazard from climate change in eastern Africa using large 581 ensembles to estimate uncertainty

Plasmodium vivax malaria viewed through the lens of an eradicated 583 European strain

Malaria in Brazil, Colombia, Peru and Venezuela: current challenges in 585 malaria control and elimination

Effects of political instability in Venezuela on malaria resurgence 587 at Ecuador-Peru border

Optimal temperature for malaria transmission is dramatically lower 589 than previously predicted

Comparing surface and stratospheric impacts of geoengineering with 591 different SO2 injection strategies

The geoengineering model intercomparison project (GeoMIP)

CESM1 (WACCM) stratospheric aerosol geoengineering large ensemble 596 project

International funding for malaria 598 control in relation to populations at risk of stable Plasmodium falciparum transmission

Climate suitability for stable malaria transmission in Zimbabwe under 601 different climate change scenarios

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted

Reaching 1.5 and 2.0° C global surface temperature targets using 605 stratospheric aerosol geoengineering

Incorporating hydrology into climate suitability models changes projections 607 of malaria transmission in Africa

A choice between two futures for pandemic recovery

39. Jesudason, T. New WHO road map for NTD recovery post COVID-19. The Lancet 611 Infectious Diseases

COVID-19 and neglected tropical 613 diseases in Africa: impacts, interactions, consequences

Strategies supporting the prevention and control of neglected tropical 615 diseases during and beyond the COVID-19 pandemic

Engaging the Global South on climate 618 engineering research

Climate emergencies do not justify engineering the climate

Temperature impacts the 622 transmission of malaria parasites by Anopheles gambiae and Anopheles stephensi 623 mosquitoes

Modelling the global constraints of temperature on transmission of 625

A global map of dominant malaria vectors

Malaria in selected non-Amazonian countries of Latin America

Assessing the global burden of ischemic heart disease: part 1: methods 631 for a systematic review of the global epidemiology of ischemic heart disease in 1990 and 632

A new scenario framework for climate change research: scenario 634 matrix architecture

Spatially explicit global population scenarios consistent with the 636

the author/funder, who has granted medRxiv a license to display the preprint in perpetuity

into two strata of risk: unstable or epidemic risk (over 30 days and under 180 days, or 1-6 months 459 of the year), and stable or endemic risk (over 6 months of the year). These classifications are 460 adapted from prior work that has stratified based on monthly temperature cutoffs, but have been 461 adapted to using daily-level climate data. The split can be thought of as indicative of a first-order 462 stratification of burden: most of the burden of malaria, especially mortality, is clustered in places 463 where malaria is endemic or hyperendemic. However, epidemics can be particularly severe in 464 places where malaria is rare, and population immunity is therefore lower (e.g., high-elevation 465 communities in east Africa, or near-elimination communities in Latin America).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)The copyright holder for this preprint this version posted October 25, 2020. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)The copyright holder for this preprint this version posted October 25, 2020. ; https://doi.org/10.1101/2020.10.21.20217257 doi: medRxiv preprint