Exploring the likelihood and mechanism of aclimate-change-induced dieback of theAmazon rainforestYadvinder Malhia,1, Luiz E. O. C. Araga˜oa, David Galbraithb, Chris Huntingfordc, Rosie Fisherd, Przemyslaw Zelazowskia,Stephen Sitche, Carol McSweeneya, and Patrick MeirbaEnvironmental Change Institute, School of Geography and the Environment, University of Oxford, South Parks Road, Oxford OX1 3QY, United Kingdom;bSchool of GeoSciences, University of Edinburgh, Edinburgh EH8 9XP, United Kingdom;cCentre for Ecology and Hydrology, Wallingford OX10 8BB, UnitedKingdom;dDepartment of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN, United Kingdom; andeMet Office Hadley Centre, JointCentre for Hydro-Meteorological Research, Wallingford OX10 8BB, United KingdomEdited by Hans Joachim Schellnhuber, Potsdam Institute for Climate Impact Research, Potsdam, Germany, and approved December 15, 2008 (received forreview June 10, 2008)We examine the evidence for the possibility that 21st-centuryclimate change may cause a large-scale ‘‘dieback’’ or degradationof Amazonian rainforest. We employ a new framework for eval-uating the rainfall regime of tropical forests and from this deduceprecipitation-based boundaries for current forest viability. Wethen examine climate simulations by 19 global climate models(GCMs) in this context and find that most tend to underestimatecurrent rainfall. GCMs also vary greatly in their projections offuture climate change in Amazonia. We attempt to take intoaccount the differences between GCM-simulated and observedrainfall regimes in the 20th century. Our analysis suggests thatdry-season water stress is likely to increase in E. Amazonia over the21st century, but the region tends toward a climate more appro-priate to seasonal forest than to savanna. These seasonal forestsmay be resilient to seasonal drought but are likely to face inten-sified water stress caused by higher temperatures and to bevulnerable to fires, which are at present naturally rare in much ofAmazonia. The spread of fire ignition associated with advancingdeforestation, logging, and fragmentation may act as nucleationpoints that trigger the transition of these seasonal forests intofire-dominated, low biomass forests. Conversely, deliberate limi-tation of deforestation and fire may be an effective intervention tomaintain Amazonian forest resilience in the face of imposed21st-century climate change. Such intervention may be enough tonavigate E. Amazonia away from a possible ‘‘tipping point,’’beyond which extensive rainforest would become unsustainable.carbon dioxide 兩 drought 兩 fire 兩 tropical forests 兩 adaptationThe response of c omponents of the Earth system to increasinglevels of anthropogen ic g reenhouse-gas forcing is unlikely tobe c ontinuous and gradual; instead there may be ‘‘tippingelements’’ in the system (1). Among the most iconic of these isthe Amazon rainforest, with some projections suggesting thepossibilit y of substantial and rapid ‘‘dieback’’ (2–4). The Ama-zon forest biome is biologically the richest region on Earth,hosting ⬇25% of global biodiversity, and is a major contributorto the biogeochemical functioning of the Earth system (3). Itslarge-scale deg radation would leave an enduring legacy on thefunction ing and diversity of the biosphere. We review theevidence for such a tipping element in Amazonia and examineclimate model projections in the context of rainforest viabilityc onsidering direct human pressures on the forest system.There is clear and ongoing change in the physical environmentof Amazon ia, whether through increasing atmospheric carbondioxide concentrations, associated imposed climate change, ormore direct intervention because of the spread of settlement,deforestation, forest timber extraction, or any related fire initi-ation. Such perturbations are certain to persist on policy-relevanttimescales. The challenge is to identify and characterize systemnonlinearities, thresholds, and feedbacks and determine whichc omponents of this system are open to manipulation and man-agement in a manner beneficial to the long-term sust ainability ofthe Amazonian social ec ological system.In this paper, we ask 3 questions. First, what are the climaticthresholds that favor the current presence of a forest as opposedto savanna? Second, what do climate models say about the likelydirection of expected climate change in Amazonia and anyassociated likelihood of large regions of Amazonia crossingthresholds of forest viability during the 21st century? Third, towhat extent does direct human pressure (deforestation, f rag-ment ation, and fires) influence the transition?Current Climate and Vegetation in AmazoniaThe lowland forests of Amazonia have a mean annual temper-ature of 26 °C, with very little spatial variability, and a meanannual precipitation of ⬇2400 mm, ranging from ⬎3000 mm inNorth West Amazon ia to ⬍1500 mm at the forest–savannatransition zones (5).Two relevant features of the rainfall regime are (i) theintensit y and duration of the dry season and (ii) the overall watersupply. To describe the accumulated water stress that occursacross a dry season, we employ the maximum climatologicalwater deficit (MCWD) (6).MCWD is defined as the most negative value of climatologicalwater deficit (CWD), att ained over a year, where the monthlychange in water deficit is precipitation (P) (mm/month) ⫺evapotranspiration (E) (mm/month). For month n,CWDn⫽ CWDn⫺1⫹ Pn⫺ En; Max共CWDn兲 ⫽ 0;CWD0⫽ CWD12; MCWD ⫽ Min共CWD1... CWD12兲At the wettest time of the year, we assume the soil is saturated (i.e.,set CWD ⫽ 0) and start the 12-month cycle of calculation from thiswet phase. For any multiyear period, we apply this calculation to themean annual cycle of precipitation (rather than calculating for eachyear and then taking the mean MCWD). We do not attempt tomodel E but fix it at 3.33 mm/day, ⬇100 mm/month. Hence, theCWD is only an approximate indicator of actual soil water deficitAuthor contributions: Y.M. designed research; Y.M., L.E.O.C.A., R.F., P.Z., and S.S. per-formed research; Y.M., L.E.O.C.A., D.G., R.F., P.Z., and C.M. analyzed data; and Y.M., C.H.,S.S., and P.M. wrote the paper.The authors declare no conflict of interest.This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: [email protected] article contains supporting information online at