Tropical deforestation accelerates local warming and loss of safe outdoor working hours

•Recent tropical deforestation increased heat exposure relative to forested areas•Work in deforested areas was associated with disproportionate lost safe work time•Deforestation >0.5 h/day of for 4.9 million people•Additional global climate change will worsen Recent research has highlighted the effects warming on low- and middle-income countries tropics. Additionally, remains an issue concern mitigation. We link these bodies by examining how exacerbates local impacts leads unsafe thermal environments outdoor workers. show that across tropics is increases humid large enough exceed established thresholds worker health. These findings suggest hastening arrival impacts. Furthermore, this shows ecosystem services provided forests are important resilience vulnerable populations. Climate many parts tropics, negatively impacting productivity Although it known warming, extent which additional affects people unknown. In modeling study, we combine health guidelines satellite, reanalysis, population data investigate recent (2003–2018) working conditions low-latitude countries, future magnify areas. find from 15 years losses 2.8 also large-scale forest loss particularly populations locations such as Brazilian states Mato Grosso Pará. Future Tropical serve carbon sinks provide critical communities,1Brondizio E.S. Settele J. Díaz S. Ngo H.T. Global Assessment Report Biodiversity Ecosystem Services Intergovernmental Science-Policy Platform Services. IPBES Secretariat, 2019Google Scholar,2Cook-Patton S.C. Leavitt S.M. Gibbs D. Harris N.L. Lister K. Anderson-Teixeira K.J. Briggs R.D. Chazdon R.L. Crowther T.W. Ellis P.W. Mapping accumulation potential natural regrowth.Nature. 2020; 585: 545-550Crossref PubMed Scopus (89) Google Scholar but driven expansion agriculture logging accelerated decades.3Curtis P.G. Slay C.M. Tyukavina A. Hansen M.C. Classifying drivers loss.Science. 2018; 361: 1108-1111Crossref (603) Trees help regulate environment.4Ellison Morris C.E. Locatelli B. Sheil Cohen Murdiyarso Gutierrez V. Van Noordwijk M. Creed I.F. Pokorny Trees, water: cool insights a hot world.Glob. Environ. Change. 2017; 43: 51-61Crossref (388) Scholar,5Scott Monks S.A. Spracklen D.V. Arnold S.R. Forster P.M. Rap Äijälä Artaxo P. Carslaw K.S. Chipperfield M.P. Impact short-lived forcers projected due deforestation.Nat. Commun. 9: 1-9Crossref (48) The removes cooling temperatures,4Ellison Scholar,6Mildrexler D.J. Zhao Running S.W. A comparison between station air temperatures MODIS land surface reveals role forests.J. Geophys. Res. Biogeosci. 2011; 116https://doi.org/10.1029/2010JG001486Crossref (162) Scholar, 7McAlpine C.A. Johnson Salazar Syktus Wilson Meijaard E. Seabrook L. Dargusch Nordin H. Forest Borneo’s climate.Environ. Lett. 13: 044009Crossref (38) 8Wolff N.H. Masuda Y.J. Wells J.A. Game E.T. Impacts temperature human well-being perceptions.Glob. 52: 181-189Crossref (40) 9Cohn A.S. Bhattarai N. Campolo Crompton O. Dralle Duncan Thompson Brazil maximum within 50 km.Environ. 2019; 14: 084047Crossref (24) 10Masuda Y. Garg T. Anggraeni I. Wolff Ebi K.L. E.E. et al.Heat decreases cognitive performance rural workers: experimental study.Environ. 15: 124015https://doi.org/10.1088/1748-9326/abb96cCrossref (9) 11Masuda Castro Aggraeni Krenz Spector How healthy, affected increasing tropics? Implications adaptation policies.Glob. 56: 29-40Crossref (27) mediated changes shade, evapotranspiration, reflectivity.12Prevedello Winck G.R. Weber M.M. Nichols Sinervo forestation globe.PLoS One. e0213368Crossref (35) often occurs contiguous patches cleared create or pasture land. magnitude strongly their area, annually averaged up 10°C warmer (>100 km2) patches,13Zeppetello L.V. Parsons Naylor R. Battisti Large scale drives extreme warming.Environ. 084012Crossref (26) can extend kilometers beyond sites.9Cohn Deforestation-induced may be detrimental communities they dependent have comparatively low adaptive capacity adjust environmental change.14McKinnon M., Buckle E., Gueye K., Toroitich I., Ionesco D., Mach al. 2016. https://www.ilo.org/global/topics/green-jobs/publications/WCMS_476194/lang--en/index.htm.Google Scholar,15Coffel E.D. Horton R.M. de Sherbinin Temperature humidity based projections rapid rise stress during 21st century.Environ. 014001Crossref (129) Increased higher threaten communities, if regularly safety.16American Conference Governmental Industrial Hygienists (ACGIH)2015 TLVs BEIs: documentation threshold limit values chemical substances physical agents & biological indices. American Hygienists, Cincinnati, OH2015Google Scholar,17Mora C. Dousset Caldwell I.R. Powell F.E. Geronimo R.C. Bielecki C.R. Counsell C.W.W. Dietrich B.S. Johnston Louis risk deadly heat.Nat. Clim. 7: 501-506Crossref (508) Contributors include ambient high humidity, internal generated heavy work, nonbreathable clothing.18Spector J.T. Sheffield P.E. Re-evaluating occupational changing climate.Ann. Occup. Hyg. 2014; 58: 936-942PubMed Working increase core body temperatures, leading strain potentially fatal stroke, even among young otherwise healthy individuals. well outdoors reduces strain,19Flouris A.D. Dinas P.C. Ioannou L.G. Nybo Havenith G. Kenny G.P. Kjellstrom Workers' under strain: systematic review meta-analysis.Lancet Planet. Health. 2: e521-e531Abstract Full Text PDF (108) Scholar,20Masuda Warming communities.Nat. 2021; 12: 1-8Crossref (6) deforestation-driven This noteworthy because already daytime much tropics.21Kjellstrom Freyberg Lemke Otto Estimating conjunction change.Int. Biometeorol. 62: 291-306Crossref (78) 22Kjellstrom Holmer Workplace stress, productivity–an challenge change.Glob. Health Action. 2009; 2047Crossref (356) 23Kjellstrom South-East Asia: ongoing time trends 1980–2011 estimates 2050.Ind. 2013; 51: 56-67Crossref (71) 24Watts Amann Arnell Ayeb-Karlsson Belesova Boykoff Byass Cai W. Campbell-Lendrum Capstick 2019 report Lancet Countdown change: ensuring child born today not defined climate.The Lancet. 394: 1836-1878Abstract (482) 25Romanello McGushin Di Napoli Drummond Hughes Jamart Kennard Lampard Rodriguez 2021 code red future.Lancet. 398: 1619-1662Abstract 26Parsons Shindell Tigchelaar Zhang labor decreased world.Nat. (In press)Crossref (3) Scholar,27Parsons Kroeger underestimated workers.Environmental Research Letters. press)https://doi.org/10.1088/1748-9326/ac3daeCrossref (2) Studies focused quantifying at regional scales,13Zeppetello Scholar,28Alkama Cescatti Biophysical cover.Science. 2016; 351: 600-604Crossref (331) relationship established.4Ellison Scholar,7McAlpine Scholar,29Crompton Corrêa S.E. influenced fragmentation spatial Maritime Southeast Asia.Environ. 16: 114018https://doi.org/10.1088/1748-9326/ac2fdcCrossref (1) Scholar,30Alves Oliveira B.F. Bottino M.J. Nobre Deforestation Amazon.Commun. Earth However, only few small-scale studies evaluated nearby communities,8Wolff Scholar,31Suter M.K. Miller K.A. Sheppard Association deforested, compared forested, study environment.Environ. (11) Scholar,32Wolff Zeppetello L.R.V. L.A. D.S. effect all-cause mortality Berau, Indonesia: modelling study.Lancet https://doi.org/10.1016/S2542-5196(21)00279-5Abstract (4) limiting our understanding whether globally similar settings countries. quantify geographic deforestation-associated 94 forests. estimate hours (the amount day performed safely thresholds) been combining satellite-derived reanalysis-based calculate Heat Index (HI) both locations. last alone (2003–2018), exposure, were almost 5 people, including workers, recently experienced least half hour per day. estimated greater experiencing extensive deforestation. areas, more hours. Our further emphasize valuable forests, increasingly workers who limited adapt warming. Of ∼23.07 km2 encompassing subtropical biome (green regions Figure 1A before year 2000; Dinerstein al.),33Dinerstein Olson Joshi Vynne Burgess N.D. Wikramanayake Hahn Palminteri Hedao Noss An ecoregion-based approach protecting terrestrial realm.Bioscience. 67: 534-545Crossref (554) 13.9 covered >75% tree cover 2000 (gray 1A, procedures; al.).34Hansen Potapov P.V. Moore Hancher Turubanova Thau Stehman S.V. Goetz S.J. Loveland T.R. High-resolution maps 21st-century change.Science. 342: 850-853Crossref (5508) Between 2003 2018, 8.7% (1.22 degraded (red 1A), 6.9% Americas (∼0.56 km2), 6.5% Africa (∼0.14 13.6% Asia (∼0.52 km2). Throughout subtropics, exceeds interannual variability near equator,13Zeppetello afternoon satellite analyzed here (experimental procedures). Specifically, Americas, Africa, maintained 2018 mean variations ∼0.2°C, ∼0.4°C, ∼-0.01°C, respectively. consistent model representations anthropogenic change; example, observations global-mean, near-surface less than 0.4°C over period.35Lenssen N.J. Schmidt G.A. J.E. Menne Persin Ruedy Zyss Improvements GISTEMP uncertainty model.J. Atmos. 124: 6307-6326Crossref (249) By contrast, 1.6°C, 1.2°C, 0.55°C, Asia, respectively (Figure 1B; see Prevedello al.);12Prevedello same period. apparent tails distributions showing largest increases.13Zeppetello For ∼78% 6°C despite accounting 9% original area. gain loss,12Prevedello here, focus timing regrowth Here, “lost hours” deemed includes conducted construction sectors21Kjellstrom Scholar,25Romanello given HI values. reflect rather individually verified that, average, 0.33, 0.13, 0.16 h day, respectively, 0.05, 0.03, 0.03 1C). experience some degree weather therefore variation hours25Romanello regardless 1C), striking decades occurred deforestation.11Masuda proportion 30 min 4% lost, whereas 31% 2A). Using Gridded Population World version 4 (GPWv4) data,36CIESIN: Center International Science Information Network World, Version (GPWv4): Administrative Unit Points Estimates. NASA Socioeconomic Data Applications (SEDAC), 2016Google live 2B). sensitivity tests LandScan data,37Rose A.N. McKee J.J. Sims K.M. Bright E.A. Reith A.E. Urban M.L. 2019. ed. Oak Ridge National Laboratory, 2020Google assume different distribution populations, main results remain unchanged analysis period, living 0.5 1.6% 8.2% losses. similarly Asia. 8.8% time, 1.1% 15.4% 4.7% Loss relatively common where fractions places. Similar virtually nonexistent (Figures 1C two 93.8% had Human highly concentrated certain Therefore, examine total number country administrative division 1—the highest subnational unit each country.36CIESIN: size circles 3 denotes division; color percent 2018. Darker colors highlight “hotspots,” divisions >30% portions population. “hotspots” Brazil, Belize, Cambodia, Vietnam, Malaysia, Myanmar, Nigeria, Cameroon. Comparing predicted Rural-Urban Project (GRUMP) assessment urban settlements densely populated latter surprising observed come partially—and cases—predominantly expense forests.38Liu X. Huang Xu Li Ciais Lin Gong Ziegler Chen High-spatiotemporal-resolution mapping 1985 2015.Nat. Sustain. : 1-7Google During 2003–2018, ∼4.9 losing ∼91,000 2 (Table S1). general statistics (i.e., agriculture, forestry, fisheries sectors),39Ruggles Kugler T.A. Fitch Riper D.C. Terra Populus: integrated environment.in: Proceedings. IEEE, 2015: 222-231Google deforestation; available 41 7.9%, 4.3%, 10.8%, 2003–2018 loss, causing all three combined S2). 1.24%, 0.16%, 2.79%,