Forest and woodland replacement patterns following drought-related mortality Enric Batlloria,b,1, Francisco Lloreta,c,1, Tuomas Aakalad, William R. L. Anderegge, Ermias Aynekuluf, Devin P. Bendixseng,2, Abdallah Bentouatih, Christof Bigleri, C. John Burkj, J. Julio Camarerok, Michele Colangelok,l, Jonathan D. Coopm, Roderick Fenshamn, M. Lisa Floydo, Lucía Galianoa, Joseph L. Ganeyp, Patrick Gonzalezq,r, Anna L. Jacobsens, Jeffrey Michael Kanet, Thomas Kitzbergeru,v, Juan C. Linaresw, Suzanne B. Marchettix, George Matusicky, Michael Michaelianz, Rafael M. Navarro-Cerrilloaa, Robert Brandon Pratts, Miranda D. Redmondbb, Andreas Riglingcc,dd, Francesco Ripullonel, Gabriel Sangüesa-Barredaee, Yamila Sasalv, Sandra Saura-Masa, Maria Laura Suarezv, Thomas T. Veblenff, Albert Vilà-Cabreragg, Caroline Vinckehh,ii, and Ben Zeemanjj a Centre de Recerca Ecològica i Aplicacions Forestals (CREAF), 08193 Cerdanyola del Vallès, Barcelona, Spain; bUnitat de Botànica, Departament de Biologia Evolutiva, Ecologia i Ciències Ambientals, Universitat de Barcelona, 08028 Barcelona, Spain; cUnitat d’Ecologia, Departament de Biologia Animal, Biologia Vegetal i Ecologia, Universitat Autònoma Barcelona, 08193 Barcelona, Spain; dDepartment of Forest Sciences, University of Helsinki, FI-00014 Helsinki, Finland; e School of Biological Sciences, University of Utah, Salt Lake City, UT 84112; fWorld Agroforestry Centre (ICRAF), Nairobi, 00100, Kenya; gDepartment of Natural Resource Ecology and Management, Oklahoma State University, Stillwater, OK 74074; hInstitut des Sciences Vétérinaires et des Sciences Agronomiques, Université El Hadj Lakhdar, 05000 Batna, Algérie; iForest Ecology, Department of Environmental Systems Science, ETH Zürich, CH-8092 Zürich, Switzerland; j Department of Biological Sciences, Smith College, Northampton, MA 01066; kInstituto Pirenaico de Ecología (IPE-CSIC), Consejo Superior de Investigaciones Científicas, 50059 Zaragoza, Spain; lSchool of Agricultural, Forest, Food and Environmental Sciences, University of Basilicata, 85100 Potenza, Italy; mSchool of Environment and Sustainability, Western Colorado University, Gunnison, CO 81231; nQueensland Herbarium (DES), Toowong, QLD 4066, Australia; oNatural History Institute, Prescott, AZ 86301; pRocky Mountain Research Station, United States Department of Agriculture (USDA) Forest Service, Flagstaff, AZ 86001; q Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA 94720; rInstitute for Parks, People, and Biodiversity, University of California, Berkeley, CA 94720; sDepartment of Biology, California State University, Bakersfield, CA 93311; tDepartment of Forestry and Wildland Resources, Humboldt State University, Arcata, CA 95521; uDepartamento de Ecología, Universidad Nacional del Comahue (UN Comahue), 8400 Bariloche, Río Negro, Argentina; vLaboratorio Ecotono, Instituto de Investigaciones en Biodiversidad y Medioambiente (INIBIOMA), Consejo Nacional de Ciencia y Tecnología (CONICET)-UN Comahue, 8400 Bariloche, Río Negro, Argentina; wDepartamento Sistemas Físicos, Químicos y Naturales, Universidad Pablo de Olavide, 41013 Sevilla, Spain; xGunnison Service Center, USDA Forest Service, Gunnison, CO 81230; ySchool of Veterinary and Life Sciences, Murdoch University, Murdoch, WA 6150, Australia; zNatural Resources Canada, Canadian Forest Service, Edmonton, AB T6H 3S5, Canada; aaDepartment of Forest Engineering, Laboratory of Dendrochronology, Silviculture and Global Change (DendrodatLab-ERSAF), University of Córdoba, 14071 Córdoba, Spain; bbDepartment of Forest and Rangeland Stewardship, Colorado State University, Fort Collins, CO 80523; ccSwiss Federal Institute for Forest, Snow and Landscape Research (WSL), CH-8903 Birmensdorf, Switzerland; ddInstitute of Terrestrial Ecology, Department of Environmental Sciences, Swiss Federal Institute Technology, ETH Zürich, CH-8092 Zürich, Switzerland; eeDepartamento de Ciencias Agroforestales, Escuela de Ingeniería de Soria (EiFAB), Instituto Universitario de Investigación Gestión Forestal Sostenible (iuFOR), Universidad de Valladolid, 42004 Soria, Spain; ffDepartment of Geography, University of Colorado Boulder, Boulder, CO 80301; ggBiological and Environmental Sciences, Faculty of Natural Sciences, University of Stirling, Stirling FK9 4LA, Scotland, United Kingdom; hhFaculty of Bioscience Engineering, Université Catholique de Louvain,1348 Louvain-la-Neuve, Belgium; iiEarth and Life Institute, Université Catholique de Louvain,1348 Louvain-la-Neuve, Belgium; and jjDepartment of Ecology, Environment and Evolution, La Trobe University, Bundoora, VIC 3086, Australia Edited by Nils Chr. Stenseth, University of Oslo, Oslo, Norway, and approved September 29, 2020 (received for review February 6, 2020) Forest vulnerability to drought is expected to increase under anthro- pogenic climate change, and drought-induced mortality and commu- nity dynamics following drought have major ecological and societal impacts. Here, we show that tree mortality concomitant with drought C limate-induced forest mortality is an emerging global phe- nomenon (1) with major consequences for the functioning of these key ecosystems (2). Reported increases in tree mortality has led to short-term (mean 5 y, range 1 to 23 y after mortality) point toward accelerating global forest vulnerability under hotter vegetation-type conversion in multiple biomes across the world temperatures and longer, more intense droughts associated with (131 sites). Self-replacement of the dominant tree species was only increased climatic variability (2–5). Abrupt drought-induced prevalent in 21% of the examined cases and forests and woodlands forest mortality can trigger substantial changes in the composi- shifted to nonwoody vegetation in 10% of them. The ultimate tem- tion and structure of ecosystems, altering carbon storage and poral persistence of such changes remains unknown but, given the cycling (6), plant productivity (7), ecosystem–atmosphere ex- key role of biological legacies in long-term ecological succession, this changes (8), hydrological cycles (9), and ecosystem resilience to emerging picture of postdrought ecological trajectories highlights the potential for major ecosystem reorganization in the coming decades. Author contributions: E.B. and F.L. designed research; E.B., F.L., T.A., W.R.L.A., E.A., D.P.B., Community changes were less pronounced under wetter postmortal- A.B., C.B., C.J.B., J.J.C., M.C., J.D.C., R.F., M.L.F., L.G., J.L.G., P.G., A.L.J., J.M.K., T.K., J.C.L., ity conditions. Replacement was also influenced by management in- S.B.M., G.M., M.M., R.M.N.-C., R.B.P., M.D.R., A.R., F.R., G.S.-B., Y.S., S.S.-M., M.L.S., T.T.V., tensity, and postdrought shrub dominance was higher when A.V.-C., C.V., and B.Z. performed research; E.B. analyzed data; E.B. wrote the paper with the contribution of F.L., W.R.L.A., C.B., J.D.C., P.G., T.K., J.C.L., M.D.R., T.T.V., and A.V.-C.; pathogens acted as codrivers of tree mortality. Early change in com- and E.B., F.L., T.A., W.R.L.A., E.A., D.P.B., A.B., C.B., C.J.B., J.J.C., M.C., J.D.C., R.F., M.L.F., munity composition indicates that forests dominated by mesic species L.G., J.L.G., P.G., A.L.J., J.M.K., T.K., J.C.L., S.B.M., G.M., M.M., R.M.N.-C., R.B.P., M.D.R., generally shifted toward more xeric communities, with replacing tree A.R., F.R., G.S.-B., Y.S., S.S.-M., M.L.S., T.T.V., A.V.-C., C.V., and B.Z. contributed to result and shrub species exhibiting drier bioclimatic optima and distribution interpretation. ranges. However, shifts toward more mesic communities also oc- The authors declare no competing interest. curred and multiple pathways of forest replacement were observed This article is a PNAS Direct Submission. for some species. Drought characteristics, species-specific environmen- This open access article is distributed under Creative Commons Attribution-NonCommercial- tal preferences, plant traits, and ecosystem legacies govern post- NoDerivatives License 4.0 (CC BY-NC-ND). drought species turnover and subsequent ecological trajectories, 1 To whom correspondence may be addressed. Email: enric.batllori@gmail.com or with potential far-reaching implications for forest biodiversity and francisco.lloret@uab.cat. 2 ecosystem services. Present address: Department of Zoology, Stockholm University, 106 91 Stockholm, Sweden. drought-induced mortality | forest dynamics | forest resilience | global tree This article contains supporting information online at https://www.pnas.org/lookup/suppl/ | mortality climate change doi:10.1073/pnas.2002314117/-/DCSupplemental. First published November 2, 2020. 29720–29729 | PNAS | November 24, 2020 | vol. 117 | no. 47 www.pnas.org/cgi/doi/10.1073/pnas.2002314117 resistance or potential vegetation-type conversion based on the Significance relative cover and identity of neighboring and understory trees, shrubs, and well-established saplings occupying the space released Forests are experiencing growing risks of drought-induced by dead trees. We then examine the influence of drought charac- mortality in a warming world. Yet, ecosystem dynamics fol- teristics and site climate, dominant species, management, and lowing drought mortality remain unknown, representing a concurrent biotic disturbances on the observed replacement pat- major limitation to our understanding of the ecological con- terns and thus ecosystem resistance. Finally, we use the bioclimatic sequences of climate change. We provide an emerging picture niches of predrought dominant and postdrought replacing species of postdrought ecological trajectories based on field indicators in relation to precipitation regime and aridity to determine shifts of forest dynamics. Replacement patterns following mortality toward more xeric or more mesic communities and to ascertain indicate limited short-term persistence of predrought domi- which factors (bioclimatic optimum, driest edge, range, succes- nant tree species, highlighting the potential for major ecosys- sional index) are associated with the observed shifts. Despite the tem reorganization in the coming decades. The great variability limited temporal perspective of the available information, our ap- of the observed dynamics within and among species reinforces proach provides an emerging picture of current postdrought eco- the primary influence of drought characteristics and ecosystem logical trajectories across a broad range of global forest biomes. legacies, modulated by land use, management, and past dis- turbances, on ongoing drought-related species turnover and Forest and Woodland Replacement Patterns their potential implications for future forest biodiversity and Following tree mortality concomitant with drought, persistence ecosystem services. of dominant forest and woodland species is limited across our sites. Although predrought dominant tree species are present in subsequent disturbances (10, 11). Global trends in the potential the community after mortality episodes in 63% of the sites, self- ecological consequences of increased tree mortality, however, replacement of the dominant species is only prevalent in 33% of remain largely unknown. these cases (21% out of the total; Fig. 2A). Replacement by Research on recent drought-induced vegetation change (since other woody species (trees and shrubs) is the dominant process the 1970s) has mostly focused on plant mortality processes and in 69% of the sites. Of those, shrubs are the main replacing not on community recovery, which still represents a research species in 42% of the cases (29% of total cases) following ECOLOGY frontier in our understanding of forest mortality (2, 12, 13). mortality, pointing to important postdrought alterations of eco- Performance and sensitivity of the postdrought community to system structure and function (11). Further, in 10% of the sites contemporary and future climatic conditions are predicted to (Fig. 2A), no replacement by woody vegetation prevailed, indi- have a lasting effect on ecosystem dynamics (14), which in turn cating at least a transient loss of forest and woodland cover are central to projecting climate-induced vegetation change and promoted by drought-related mortality (up to ∼20 y following drought; Dataset S1). Such replacement patterns are consistent Earth system feedbacks (15). Thus, characterizing ecological regardless of recovery time and forest successional state (SI communities following forest mortality is crucial to understand- Appendix, Fig. S3 and Dataset S1). Ecologists increasingly rec- ing subsequent consequences for ecosystems. ognize that long-term outcomes of ecological succession can be Postdisturbance recruitment is a key ecological process, as it disproportionately shaped by early patterns of species occupancy sets the stage for long-term ecosystem recovery and dynamics. and ecosystem legacies (20–22), particularly under directional Initial recruitment stages are, however, highly sensitive to envi- climate change (10). Drought represents, therefore, a likely ronmental constraints, typically exhibiting high mortality rates driving factor of species turnover and structural change with (16) that could be exacerbated under climate change (17). In strong influences on ecosystem resistance and thus short-term addition, the climatic requirements and tolerance of juveniles resilience (2, 19). Whether such changes correspond to tran- often differ from those of adults (18). Alternatively, the assess- sient communities within successional trajectories or to stable ment of intermediate stages, such as saplings or subcanopy trees state shifts will only be unraveled through long-term monitoring, representing dominant and well-established plants occupying the and is a key future research direction. space and using resources released by dead trees, can provide Short-term forest and woodland replacement patterns are not better short-term indicators of forest dynamics and ecosystem consistently related to the bioclimatic characteristics (drier ver- resistance (19) following mortality events. Forest ecosystems sus wetter) of the affected sites or species (SI Appendix, Fig. S4). comprise long-lived species and, in the absence of further dis- This is consistent with the equivalent relationships between turbance, we can expect that the influence of the persisting mortality and drought intensity reported among biomes and species [i.e., disturbance legacies (10)] and their demographic between major plant taxonomic levels (4) (angiosperms versus responses will remain for decades (13, 16). gymnosperms). However, replacement patterns are significantly Here, we assess short-term replacement patterns following tree modulated by drought characteristics (Fig. 3 and SI Appendix, mortality associated with drought in 131 forest and woodland sites Fig. S5). Specifically, wetter conditions after drought-related encompassing a wide range of vegetation types across multiple mortality are associated with smaller compositional and struc- biomes, from dry tropical to temperate and boreal systems but tural differences between pre- and postdrought community. excluding species-rich tropical forests (Fig. 1 and Dataset S1). In all Conversely, the less favorable postmortality drought conditions sites studied, tree mortality was linked to drought evidence re- are relative to the drought event, the greater the community gardless of among-site differences in drought intensity, duration, change irrespective of the intensity of the event. This suggests seasonality, and pre- and postdrought conditions relative to the stronger and more sustained shifts under directional precipita- reported mortality events (i.e., drought characteristics; SI Appendix, tion changes (e.g., longer-term, increasing drought trends). Fig. S1 and Datasets S1 and S2). Postmortality surveys (mean 5 y, Premortality drought conditions are not significantly related to range 1 to 23 y after mortality; SI Appendix, Fig. S2) were per- the replacement patterns reported here. Regardless, the influ- formed to assess dominant woody species composition, allowing ence of postmortality drought conditions on forest and woodland characterization of the degree of self-replacement (i.e., replace- replacement patterns highlights the potential of drought-induced ment of a dominant tree species with itself), replacement by other ecosystem reorganizations under ongoing trends of increased woody species including trees and shrubs, or lack of replacement by aridity (1, 5). woody vegetation. We use a joint compositional and structural Replacement patterns are often influenced by major species- community resemblance index to quantify forest and woodland specific traits such as resprouting capacity or light requirements Batllori et al. PNAS | November 24, 2020 | vol. 117 | no. 47 | 29721 Fig. 1. Location of the 131 field sites (Dataset S1) for which this research assessed tree species replacement patterns after mortality concomitant with drought. The analysis considers forest and woodland sites across Earth’s forested biomes, excluding species-rich tropical biomes (gray areas in the map). Global forest cover is based on Global Forest Watch (http://globalforestwatch.org). Biome classification (61): BorF, boreal forests/taiga; Des, deserts and xeric shrublands; MedF, Mediterranean forests, woodlands, and scrub; MnG, montane grasslands; TeBF, temperate broadleaf and mixed forests; TeCF, temperate conifer forests; TeG, temperate grasslands, savannas, and shrublands; TrG, tropical and subtropical grasslands, savannas, and shrublands. Photos exemplifying the four replacement processes considered are as follows. (A) Self-replacement; E. marginata, Northern Jarrah Forest, Australia (G.M., 2014). (B) Replacement by another tree species; Cedrus atlantica, Middle Atlas, Morocco (E.B., 2017). (C) Replacement by shrub species; Abies pinsapo, Sierra de las Nieves, Spain (E.B., 2017). (D) No replacement by woody vegetation; P. edulis, New Mexico, USA (F.L., 2012). (Fig. 2B and SI Appendix, Fig. S6). Resprouting tree species (e.g., successional trajectories toward communities better adapted to Populus spp., Eucalyptus spp., Quercus spp.) tend to show higher the conditions of the affected sites. Prescribed burning, thinning, levels of self-replacement than obligate seeding species (e.g., and, to a lesser degree, grazing, deplete plant density and bio- Pinus spp., Abies spp.), although mortality rates concomitant mass and can enhance ecosystem water availability (29). These with drought are similar among them (Fig. 2B and SI Appendix, mechanisms, together with the unintended selection by these Fig. S6). Additionally, community-level resprouting capacity in- practices of species resistant to disturbance, could explain higher creases as a result of species-specific replacement patterns (SI self-replacement patterns associated with these management Appendix, Fig. S6). Other functional attributes, such as light re- types. Nevertheless, high management intensity, irrespective of quirements, will also mediate forest recovery patterns. Shade- management type, is associated with higher replacement by tolerant plants present in the understory increase in abundance shrubs compared with low or moderate management intensities following canopy mortality (SI Appendix, Fig. S6). Therefore, it is (Fig. 4 and SI Appendix, Figs. S7 and S8). Whether and how unlikely that either resprouting or seeding becomes the prevalent postdrought communities differ from recovery patterns associ- regenerative strategy of the dominant postmortality species un- ated with management practices inducing tree mortality (e.g., der the stochastic spatiotemporal patterns of extreme droughts timber harvest) is uncertain but may be a function of biome, (23). Moreover, the interaction of drought with other distur- climate, and management type and goals, among others. For bances, such as fire, may also lead to different outcomes in the instance, in forests dominated by shade-tolerant species such as prevalence of regenerative traits (24, 25). Picea abies or Pinus edulis, advanced regeneration in the un- Persistence versus turnover of dominant tree species is mod- derstory would lead to self-replacement following any distur- ulated by the influence of past and concurrent codrivers such as bance affecting overstory trees (e.g., selective thinning, drought). management type, management intensity, and biotic distur- However, even in these systems, climate trends (e.g., warming) bances (26, 27) (Fig. 4). Although highly variable, replacement may preclude self-replacement where increased moisture stress by shrubs following drought tends to be highest in communities results in recruitment failure and high juvenile mortality (30). As dominated by naturalized species originated by historical plant- another example, microenvironmental conditions generated by ing (Fig. 4 and SI Appendix, Fig. S7), likely reflecting the absence tree mortality associated with drought can be clearly different of other tree species in these areas and a stronger disequilibrium from those generated after harvesting. In Nothofagus dombeyi between predrought dominant species and the climatic regime in forests, dead trees remain standing for many years following these sites (28). In such cases, replacement by shrubs may initiate drought, generating less direct radiation but high levels of diffuse 29722 | www.pnas.org/cgi/doi/10.1073/pnas.2002314117 Batllori et al. Fig. 2. Postdrought replacement patterns by vegetation replacement type (NR, no replacement by woody vegetation; Self, self-replacement; Shrub, re- placement by shrublands; Tree, replacement by other tree species) and by tree genus. In A, each bar depicts the possible combinations of replacement by the different types (e.g., Self + Tree corresponds to sites in which self-replacement and replacement by other tree species are observed), whereas the proportion of each replacement type across all sites is depicted by the size of the gray dots. The overall proportion of sites showing a given replacement type is shown (Right). Colors depict major replacing categories in which trees (green), shrubs (violet), or lack of replacement by woody vegetation (brown) dominate. In B, outer-level colored bars show the dominant (predrought) genus and the most important replacing woody genera (NR, lack of replacement by woody vegetation; Shrub, replacement by shrub species; Tree_other, replacement by other scarcely represented tree genera: Acer, Arbutus, Austrocedrus, Betula, Carya, Dasyphyllum, Fagus, Ilex, Lomatia, Sorbus, Ulmus, Weinmannia). Inner links are directional, joinning predrought dominant genera (flat ends) and postdrought replacing genera (arrow ends). The inner links depict replacement proportions, so link width is proportional to the number of cases showing any ECOLOGY given replacement pattern. radiation, favoring replacement by tree species with intermediate level drought tolerance that results from shifts in tree species shade tolerance, rather than shade-tolerant or shade-intolerant composition due to incremental increases in climatic water def- species (31). icit [e.g., United States (33, 34)]. Extreme drought events may Biotic disturbances contemporaneous with the drought event accelerate, in some cases, the effect of long-term climate trends show the strongest effect on replacement patterns (Fig. 4). When on the species replacement process, which may have initiated fungi and other pathogens act as codrivers of tree mortality, the prior to the onset of drought (35). Yet, pulses of forest mortality highest levels of shrub replacement and the lowest levels of self- concomitant with drought show highly variable population re- replacement are observed. In contrast, bark beetles do not seem sponses. In 24% of the study sites, replacement involved species to systematically relate to low levels of self-replacement. Thus, with less arid bioclimatic optima (Fig. 5B). Finally, in 21% of the together with climate, replacement patterns may reflect distinct cases, no changes in bioclimatic characteristics occurred after ecosystem legacies depending on the type of biotic agent in- forest mortality, corresponding both to self-replacement and to volved. For example, insects may have tree size requirements replacement by species with equivalent bioclimatic characteris- which, unlike pathogens and fungi, may leave smaller conspecific tics following drought (Fig. 5E). trees intact in the understory (32). Additionally, fungal patho- Although the occurrence of mortality events coincident with gens could be primary factors of tree mortality and vegetation drought is observed irrespective of site-level and species-specific changes by themselves but empirical and experimental evidence bioclimatic characteristics (SI Appendix, Fig. S4), our analyses at supports in many cases higher mortality rates when strong the community level reveal that sites with a higher dominance of physiological stress (e.g., drought) limits the vigor of trees and mesic species tend to shift consistently to more xeric communi- predisposes them to parasite attacks (27). Overall, these results ties (Fig. 6A). Such shifts are imparted by replacing species highlight that interacting disturbances can disrupt ecosystem exhibiting distribution limits over drier bioclimatic conditions than resistance and amplify processes that can lead to state changes in the replaced species, likely due to a higher tolerance to water forest ecosystems (26). deficit (Figs. 5C and 6B). This higher tolerance to arid conditions may be behind some of the iconic cases of drought-related forest Vegetation Dynamics and Bioclimate transformation such as P. edulis replacement by Juniperus osteo- Across multiple forested biomes, drought-related mortality sperma, Juniperus monosperma, and Purshia tridentata in North mostly promoted shifts toward more xeric communities, although America, Pinus sylvestris replacement by Quercus pubescens or shifts toward more mesic communities also occurred. This is il- Quercus ilex in Europe, and Eucalyptus marginata replacement by lustrated by the bioclimatic characteristics of the predrought Corymbia calophylla in Australia (Fig. 5 C and D). In contrast, dominant and postdrought replacing species along the environ- shifts toward more mesic communities are imparted by replacing mental variability axis encompassing aridity (Fig. 5 and SI Ap- species exhibiting distributions over a wider range of climatic pendix, Figs. S9–S12). Changes indicative of shifts toward more conditions in relation to precipitation regimes and aridity (Figs. 5B xeric postdrought communities (55% of the assessed cases) re- and 6C and SI Appendix, Fig. S13). This has occurred, for instance, flect increased bioclimatic optima in terms of aridity levels and in Mediterranean landscapes, where Pinus pinea has been reduced optima in relation to precipitation amount and sea- replaced by Q. ilex, and Quercus cerrioides by Buxus sempervirens sonality of the replacing woody species (Fig. 5A). This is con- (SI Appendix, Fig. S13). sistent with the potential for abrupt vegetation transitions toward Replacement by early- versus late-successional species also ap- drought-tolerant species formerly reported by Martínez-Vilalta pears to influence the observed bioclimatic shifts (Fig. 6D). Shifts to and Lloret (13). It is also consistent with increases in population- more xeric communities can be associated with replacing species Batllori et al. PNAS | November 24, 2020 | vol. 117 | no. 47 | 29723 Fig. 3. Effects of drought conditions before (pre), during, and after (post) tree mortality on the reported replacement patterns. The panels show the results of a beta regression model where replacement pattern (included in the model as the community resemblance index) is the dependent variable and drought conditions are the explanatory variables. The CRI is a joint compositional and structural index that quantifies the vegetation-type change in initial tree forest composition. CRI = 0 reflects no change in composition or structure (complete self-replacement by neighboring canopy trees) and CRI = 1 corresponds to the maximum possible change (no woody replacement). (A) The model’s coefficient estimates. (B) Influence of the difference between postmortality and during- mortality drought conditions on CRI; larger values correspond thus to more favorable conditions after mortality. (C) Influence of the interaction between during-mortality drought and the difference between postmortality and during-mortality drought conditions on CRI. Model pseudo-R2 = 0.274. that are late-successional. This is consistent, for instance, with on- pose additional challenges to understanding community dynamics going replacement patterns in some European temperate forests following drought. Remote sensing approaches to develop com- where P. sylvestris, a pioneer conifer that has been favored by sil- posite and multiindicator drought monitoring at broad spatial vicultural practices, is replaced by late-successional species such as scales (38), together with the design and implementation of field- Q. pubescens. However, successional trajectories can also result in based monitoring systems under standardized protocols (12, 13), shifts toward more mesic communities when mortality of the forest are needed to better understand the potential long-term conse- canopy favors shade-tolerant, more mesic species growing in the quences of mortality events related to drought. understory, for example the replacement of Pinus flexilis by Abies The observed variability in short-term replacement patterns lasiocarpa or Picea engelmannii in western North America. There- and associated bioclimatic shifts reported here indicates that the fore, community-level bioclimatic changes resulting from succes- ecological outcomes of extreme climatic events, such as drought, sional trajectories are driven, among other factors, by the interplay depend on the relative strength of multiple ecological processes. of climate variability (e.g., drought), the available species pool and On one hand, ongoing trends of increasing aridity (5) can pro- their bioclimatic characteristics, and management- or disturbance- mote mortality of major forest species (1), hinder their recovery, caused modifications to stand structure and competition processes and favor replacement by drought-resistant tree and shrub spe- (13, 36). cies (i.e., species with broader bioclimatic niches or drier ranges). Mortality linked to climate variability could thus lead to bio- Future Perspectives geographical shifts or permanent range contractions of tree We found limited short-term persistence of dominant tree spe- species with climate niches less suitable to current and future cies after drought-related mortality across major forested and conditions (2, 39) or to geographic shifts in vegetation at the woodland biomes. Our work thus represents an important step in biome level (40). Species functional traits other than those di- documenting the ecological consequences of forest mortality rectly related to drought resistance (e.g., resprouting capacity, associated with climate variability at the multibiome scale (1, 2, plant light requirements, dispersal and colonizing ability) may 13, 37). Empirical understanding of long-term trajectories of have, however, paramount importance in modulating community species persistence versus vegetation shifts following mortality, dynamics and the compositional and functional resilience ca- however, remains very limited. This is particularly true for pacity of the system (4, 10). On the other hand, drought-related species-rich tropical forests, not included in this analysis, where forest mortality may enhance transitions within typical succes- the lack of data and the high biological and functional diversity sional pathways by removing early-successional dominant species 29724 | www.pnas.org/cgi/doi/10.1073/pnas.2002314117 Batllori et al. ECOLOGY Fig. 4. Postdrought replacement patterns in relation to management, management intensity, and biotic disturbances. The CRI is a joint compositional and structural index that quantifies the vegetation-type change in initial tree forest composition (Fig. 3). The symbols (Top Right) in each plot show significant differences among the different classes in each panel: +P < 0.05, ++P < 0.01. and favor replacing communities better suited to the climatic associated with drought and their potential implications for future regime of the affected area. By contrast, drought could also forest biodiversity and ecosystem services. disrupt successional pathways by removing late-successional species, potentially promoting shifts to divergent trajectories Materials and Methods and alternative vegetation states that could be maintained by Mortality Dataset. This work is based on information from 131 forest and emerging climates or subsequent disturbances. Further, extreme woodland sites (Fig. 1) in which drought is considered to have played a drought may also act as an environmental filter for species that relevant role in tree mortality (SI Appendix, Fig. S1 and Datasets S1 and S2). Therefore, the sites exhibited mortality events exceeding the presumed are particularly sensitive to water deficit (41) irrespective of baseline or background tree mortality rates in the absence of drought successional status. Therefore, drought may not necessarily lead (standing mortality: mean 37%; median 31%; range 5 to 97%). Nevertheless, to altered trajectories in comparison with succession following other agents (e.g., insects, diseases) could have acted as codrivers of the other disturbance types, but drought-associated tree mortality observed forest mortality. Basic pre- and postdrought community data could accelerate the spatial extent and temporal rate of vegeta- (Dataset S3) came from the contributing authors, who performed field- tion turnover, with broad ecological impacts (6–11, 14, 15). Fu- based studies in which mortality of dominant tree species concomitant ture research is needed to unravel how postdrought vegetation with drought and subsequent replacing species was assessed (Dataset S1). patterns differ from those following tree mortality from other Additional data come from authors’ observations within or near the field types of disturbance, such as wildfire, pest infestations, or sites, from the CIPHA database (Climate Impacts on the Productivity and Health of Aspen; Natural Resources Canada) (44), and from previously un- windthrow, particularly under a changing climate. reported field data collected within the project BIOCLIM (CGL2015-67419-R; Disturbance-related changes in the relative abundance of native principal investigator F.L.). Each contributing author provided data on major and newly introduced or invasive species could lead to new species site characteristics (location, past management and disturbance legacies) combinations (i.e., novel ecosystems), representing future, un- and tree mortality and replacement patterns of the dominant and replacing known ecological pathways for specific forest ecosystems (42). woody species (Datasets S1 and S3). The average lag between the last year of Similarly, interactions among disturbance events (e.g., drought, fire, reported tree mortality and assessment of replacement patterns is 5 y (range insect and pathogen outbreaks) may trigger compositional and 1 to 23 y after mortality; mean 11 y; range 1 to 37 y from the initial year of structural changes affecting ecological trajectories (26, 43). A key tree mortality; SI Appendix, Fig. S2). constraint on postdrought dynamics and vegetation-type transitions A comprehensive search through the Thomson Reuters Web of Science was may thus be the availability of climatically suitable species at a given performed on 2 May 2017 with the following keywords (topic search): (vegetation OR forest OR woodland OR shrubland OR biome) AND (shift$ OR site (13) as suggested by the great variability of observed dynamics change$ OR transition$ OR replacement$ OR substitution$ OR succession) among and within species, including multiple pathways of forest AND drought$ AND (mortality OR die-off OR dieoff OR decline OR die-back replacement for some species. Overall, short-term replacement OR dieback) AND (regeneration OR recruitment). The search yielded 351 patterns reinforce the primary influence of drought characteristics studies, from which we selected field-based studies reporting drought- and ecosystem legacies, modulated by land-use changes, manage- related mortality after the 1950s, excluding species-rich tropical forests ment history, and past disturbances, on ongoing species turnover (e.g., ref. 1; see next paragraph). The selection included research from the Batllori et al. PNAS | November 24, 2020 | vol. 117 | no. 47 | 29725 Fig. 5. Community bioclimatic shift as a result of forest mortality associated with drought. (Left) Relative change within the environmental space defined by precipitation regime and aridity. Environmental axes 1 and 2 encompass 82.4% of the variability of individual variables (PCA-derived axes). Each arrow represents the bioclimatic shift for a given forest site computed as the difference between the bioclimatic centroids of the dominant (predrought) and the replacing woody species weighted by the relative abundance of each species at the site. Orange and blue arrows illustrate shifts toward more xeric and more mesic communities, respectively. (Right) From A to E are examples of bioclimatic niches of the predrought dominant (blue) and postdrought replacing (red) species. Solid lines show the abundance and distribution range of each species along environmental axis 1, whereas the dotted vertical lines correspond to the species’ bioclimatic optima (center of mass of the distribution). contributing authors of this study and several additional potential works was obtained from the Global Drought Monitor (http://spei.csic.es/map; (n = 168 sites). However, we received no response in some cases (32 studies) accessed March 2019) and used to characterize drought (i.e., pre-, during, and data were not suitable for this survey in others (5 studies). Eventually, and postmortality drought conditions) at each site in relation to the period data from 131 forest and woodland sites were compiled for this assessment. of observed tree mortality. We selected the SPEI temporal scale that Within our framework, dominant (predrought) species refers to all tree matched the period of tree mortality concomitant with drought as described species with a canopy cover >25% of total canopy cover of the site that were by the individual study authors (SI Appendix, Fig. S1). Premortality drought affected by mortality. Following this definition, we excluded species-rich conditions correspond to mean SPEI values of 1, 5, and 10 y prior to tree tropical forests affected by mortality coincident with drought where many mortality, whereas during-mortality drought conditions correspond to the species are present. The high levels of biological diversity and functional minimum SPEI value (i.e., the most extreme water deficit experienced in composition in these forests would have required a different approach from each site) during the period in which tree mortality was reported. Post- the one presented here. Replacing woody species refers to those species mortality drought conditions correspond to mean SPEI values after mortality already growing in the gaps produced by the death of the dominant tree and until postmortality vegetation surveys were conducted (up to 10 y after species and, therefore, occupying the space released by the mortality event mortality). Negative and positive SPEI values correspond to dry and wet associated with drought. Replacing species include understory trees, shrubs, conditions, respectively. The differences between premortality and during- and well-established saplings (with measurable diameter at breast height) mortality drought conditions and between during-mortality and post- or, alternatively, neighboring trees whose crown occupied the space re- mortality drought conditions were used, together with during-mortality leased by the dead trees. Importantly, young seedlings which we assume to drought conditions, to quantify drought in each site (Dataset S1). have uncertain fate (16) were not considered here. The replacement of the dominant species with the same species (i.e., by itself) is referred to as self- Bioclimatic Niche Characterization. We obtained global presence data for the replacement. Drought within the reporting sites was inferred, non- 122 species constituting the dominant (predrought) tree species and the exclusively, from tree-ring analysis (n = 43), instrumental long-term climate replacing woody species from the Global Biodiversity Information Facility records (n = 61), or a combination of local weather surveys and symptomatic (GBIF; https://www.gbif.org/). In all cases, data from botanical gardens were (observational) tree dieback and mortality (n = 106). The standardized excluded and, when available, coordinate uncertainty was used to filter out precipitation evapotranspiration index [SPEI (45)] at a 1° spatial resolution data with geolocation errors >10 km. Only occurrence data from the 29726 | www.pnas.org/cgi/doi/10.1073/pnas.2002314117 Batllori et al. better characterize the capacity of species to endure different levels of aridity and precipitation variation. For very widely distributed species (e.g., P. sylvestris), the bioclimatic optima may be less representative of a “typical” climate niche. However, this was partially ameliorated by using the 95% distribution kernel to define the species’ bioclimatic optima and ranges (SI Appendix, Fig. S14). In addition, uncertainty in niche characterization is di- rectly related to the number of occurrence records for each species (SI Ap- pendix, Fig. S15). Also, niche estimations presented here correspond to the realized species niche that, among others, reflects all past processes that have influenced current species distributions (e.g., land use). Replacement Pattern Analysis. First, for each site and affected dominant species, we characterized replacement following drought-related mortality by assessing the proportion of mortality-created gaps in which the dominant vegetation occupying the gaps corresponded to the same tree species (self-replacement), other tree species (tree replacement), shrub species (shrub replacement), and lack of replacement by woody vegetation (NR). Species-specific data from all sites were then aggregated to obtain a sum- mary of the replacement patterns by these four replacement typologies. Subsequently, we summarized the replacement patterns at the genus level by computing the overall proportion of sites where each dominant (pre- drought) genus was present and the identity of replacing woody genera in Fig. 6. Contribution of (A) the bioclimatic characteristics of the study sites, each case. We used a circular layout [R package circlize (50)] as an efficient (B) the dry bioclimatic edge of the replacing species (Rep_sp), (C) the range method for visualizing the multidimensional replacement patterns across of the bioclimatic distribution of the replacing woody species, and (D) the the multiple biomes included in this study. successional index of the replacing species versus the dominant (predrought) Second, replacement patterns were quantified by means of a community species on the bioclimatic shift index. ICBS is the difference between the resemblance index (CRI). The CRI corresponds to a joint measure of com- bioclimatic optima of the replacing woody species and the bioclimatic op- munity resemblance in terms of composition and structure, in which these tima of the dominant (predrought) species along environmental axis 1 two ecosystem properties have the same weight. Compositional and struc- ECOLOGY (Fig. 4). Positive ICBS values indicate shifts toward more xeric communities, tural differences were characterized by means of Bray–Curtis dissimilarity, an whereas negative values indicate shifts toward more mesic communities. A index that is appropriate to measure community resemblance in terms of shows a linear fit and B–D depict the component smooth functions of a size structure and species composition (51). Pre- and postdrought composi- generalized additive model fitted using the four variables depicted here; tional differences were computed with the function vegdist from the R model R2 = 0.612, explained deviance = 67.2%; all variables are significant at package vegan (52) on the basis of community data matrices (i.e., matrices P < 0.05. with the relative abundance of predrought dominant and postdrought replacing woody species in each site). Structural differences were computed with the function vegdiststruct from the R package vegclust (53) on the basis 1950-to-2017 period were retained. Subsequently, a 95% density distribu- of five structural typologies or size classes of replacement: neighboring tion kernel was used as the CI to systematically remove outliers according to canopy, (subcanopy) tree replacement, sapling replacement, shrub replace- the bulk of occurrences for each species (SI Appendix, Fig. S14). This method ment, and no woody replacement. We assumed that replacement by was used to remove single-occurrence points far from the major species neighboring canopy trees represents the minimum structural change and no distribution clusters (46) (SI Appendix, Fig. S14). To correct for geographical bias woody replacement represents the maximum change. Tree, sapling, and in sampling efforts within the GBIF data and to remove points that were too shrub replacement represent intermediate structural changes between the close to each other, we resampled occurrence data falling within 10 arc-min two extremes. The CRI is the mean of the compositional and structural dis- (∼18.5 km) distance from each other as described in ref. 46. This enabled us to similarities, taking a value of 0 when there is neither change in composition obtain a balanced dataset with a minimum number of occurrence data to nor in structure (i.e., replacement by neighboring canopy trees of the same characterize the bioclimatic niche of each species. species) and a value of 1 when the change is largest (i.e., no woody re- The environmental space used to define species’ bioclimatic niches in- placement). corporated four variables to characterize precipitation regimes [precipita- Third, we used the CRI to assess the relationship between replacement tion of the driest and warmest quarters, precipitation seasonality, annual patterns and drought conditions before, during, and after mortality by means precipitation; WorldClim 2.0; https://worldclim.org (47)] and the aridity index of beta regression models. To do this, CRI values (y) were previously trans- [AI; Global Potential Evapotranspiration database, Consortium of Interna- formed as (y·(n − 1) + 0.5)/n, where n is the sample size (54) to obtain values tional Agricultural Research Centers for Spatial Information (CGIAR-CSI); in the open standard unit interval (0,1). We also used the CRI to examine the https://cgiarcsi.community (48)]. The spatial resolution of the AI was adjusted relationship between replacement patterns and the bioclimatic characteristics of to the 10 × 10-km resolution of the precipitation data and, for each pixel, the the study sites, as inferred from their geographical location and also in relation aridity level was computed as 1 − AI (46). The four precipitation variables and to the bioclimatic optima of the dominant (predrought) species. We mapped the aridity values were standardized, and the climate data from species-rich PCA loadings to extract the PC1 and PC2 values at the x–y coordinates of the tropical biomes (Fig. 1) were excluded from the datasets, before principal- sites, and subsequently computed the environmental centroid of each site as the component analysis (PCA) was performed to reduce the dimensionality of weighted mean of the bioclimatic optima of the dominant (predrought) species the five climate variables included in this analysis. We retained the first two based on their abundance in each site. Finally, we used the CRI and the pro- PCA axes (PC1 and PC2), which comprise two orthogonal axes of variation of portion of self-replacement and replacement by shrubs to assess the influence of precipitation regimes and aridity explaining 82.4% of the variability within the codrivers such as management type, management intensity, and the occurrence original climate dataset (PC1, 60.2%, mostly related to aridity levels [AI], and of biotic disturbances (Dataset S1) in the overall replacement patterns. Man- PC2, 22.2%, related to precipitation amount and seasonality). Subsequently, agement intensity was characterized by means of a semiquantitative index (low, PCA loadings were used to compute the PC1 and PC2 values for each pixel to moderate, high) based on the author’s expert knowledge of the study sites and map the environmental information to geographic space. on-site evidence (e.g., stumps, trampling). For instance, high thinning intensity Occurrence data and the defined environmental space were then used to would correspond to a forest site with abundant stumps whereas a low thinning characterize the mean features of the bioclimatic niche—or bioclimate—for intensity would imply a much lower density of stumps. However, given the di- each species (SI Appendix, Fig. S14). Features incorporated in our analysis versity of the assessed forests, a one-size-fits-all threshold to define intensity were: 1) the bioclimatic optimum—defined as the weighted centroid in (e.g., stump density) is hard to establish and that is why such estimates rely on environmental joint PC1–PC2 space [center of mass of the distribution kernel authors’ expertise in the study areas. A Kruskal–Wallis rank-sum test was applied of each species; COGravity function, SDMTools package (49)]; 2) the driest to perform a nonparametric one-way ANOVA and determine whether there are bioclimatic edge—defined as the maximum value of the species distribution significant differences in the observed replacement patterns in relation to the along PC1; and 3) the bioclimatic range—defined as the width of the climate management and biotic agent codrivers. Also, we used beta regression models envelope in PC1. The bioclimatic edge and range were incorporated to and multiple comparisons (Tukey honest significant differences) to assess pairwise Batllori et al. PNAS | November 24, 2020 | vol. 117 | no. 47 | 29727 differences on the mean levels of replacement between management and dis- early-successional and positive values when replacing species are late- turbance types. Note that the proportion of self-replacement and replacement by successional. For those species where PLRs were missing, we used the R shrubs was rescaled between 0 and 1 before applying beta regression models and package mice (59) (multivariate imputation chained equations) to create that rescaled variables were transformed to obtain values in the open unit in- multiple imputations (n = 100) for multivariate missing data. For the mul- terval (0,1) as described above for the CRI. Beta regression models were computed tivariate imputation, we used a suite of 13 plant traits, including specific leaf with the R package betareg (55) and pairwise differences with the package area, plant vegetative height, plant height growth, seed dry mass, seedling multcomp (56). These analyses were performed for 102 (out of the 131) forest growth rate, plant resprouting capacity, plant light requirements, wood sites with detailed quantitative information on replacement patterns (Dataset S1). density, leaf dry mass, leaf dry matter content, leaf lifespan, seedbank type, Fourth, a generalized additive model [GAM; R package mgcv (57)] was and seed longevity. Subsequently, we used the mean of the imputed values used to determine which factors, when replacement by other woody vege- at the species level to assign the imputed PLR values. The mean, mode, and tation occurred, were associated with shifts toward more xeric or more mesic median of the imputed values yielded equivalent results. The same impu- woody replacing communities. We used the site-level community bioclimatic tation procedure was used to obtain missing data for plant resprouting shift index (ICBS) as the response variable, which was calculated as the dif- capacity of all species. ference between the bioclimatic optima of the replacing woody species and Data treatment, analyses, and visualization were all performed in R (60). the bioclimatic optima of the predrought forest along PC1, the axis mainly driven by differences in aridity. As a result, positive ICBS values indicate a Data Availability. All data related to this paper are available in the main text, shift toward more xeric species whereas negative values indicate a shift SI Appendix, and Datasets 1–3. toward more mesic species. When multiple replacing-dominant species were present, we used a weighted mean procedure based on species abundance ACKNOWLEDGMENTS. We thank two anonymous reviewers who helped to in each site to obtain a site-level ICBS. Four explanatory variables were used improve the manuscript and also acknowledge M. De Cáceres for his assis- in the GAM: the bioclimatic optimum of the predrought forest, the driest tance with community indices. E.B. thanks the support of Project CGL2017- bioclimatic edge of the replacing woody species, the bioclimatic range of the 87176-P, and E.B. and F.L. thank the support of Project CGL2015-67419-R and replacing woody species, and the successional index of the replacing woody research group 2017 SGR 1001. W.R.L.A. acknowledges funding from the species versus the predrought forest, characterized as the difference be- University of Utah Global Change and Sustainability Center (NSF Grants tween plant light requirements (PLRs) of the replacing community and PLRs 1714972 and 1802880) and the United States Department of Agriculture of the predrought community. Species-specific PLRs and the other plant National Institute of Food and Agriculture, Agricultural and Food Research traits (e.g., resprouting capacity) used in our study were obtained from the Initiative Competitive Program, Ecosystem Services and Agro-Ecosystem Management (Grant 2018-67019-27850). A.V.-C. was funded by the Euro- publicly available, open-access TRY database [https://www.try-db.org (58)]. pean Union’s Horizon 2020 research and innovation programme under PLR was converted to a quantitative index, where 1 indicates full light, 2 Marie Skłodowska-Curie Grant Agreement 656300, and the 50th Anniversary indicates half-shadowy, and 3 indicates shadowy (following notation in Fellowship program of the University of Stirling. G.S.-B. is supported by a TRY). Therefore, the successional index or PLR difference between post- and Juan de la Cierva-formación Grant (FJCI 2016-30121) from the Spanish Min- predrought community yields negative values when replacing species are istry of Science and Innovation. 1. C. D. Allen, D. D. Breshears, N. G. McDowell, On underestimation of global vulnera- 19. D. Hodgson, J. L. McDonald, D. J. Hosken, What do you mean, ‘resilient’? Trends Ecol. bility to tree mortality and forest die-off from hotter drought in the Anthropocene. Evol. 30, 503–506 (2015). Ecosphere 6, 1–55 (2015). 20. S. T. A. Pickett, P. White, The Ecology of Natural Disturbance and Patch Dynamics 2. W. R. L. Anderegg, J. M. Kane, L. D. L. Anderegg, Consequences of widespread tree (Academic Press, 2013). mortality triggered by drought and temperature stress. Nat. Clim. Chang. 3, 30–36 21. M. G. Turner, W. L. Baker, C. J. Peterson, R. K. Peet, Factors influencing succession: (2013). Lessons from large, infrequent natural disturbances. Ecosystems 1, 511–523 (1998). 3. A. P. Williams et al., Temperature as a potent driver of regional forest drought stress 22. R. Seidl, W. Rammer, T. A. Spies, Disturbance legacies increase the resilience of forest and tree mortality. Nat. Clim. Chang. 3, 292–297 (2013). ecosystem structure, composition, and functioning. Ecol. Appl. 24, 2063–2077 (2014). 4. S. Greenwood et al., Tree mortality across biomes is promoted by drought intensity, 23. P. J. Bellingham, A. D. Sparrow, Resprouting as a life history strategy in woody plant lower wood density and higher specific leaf area. Ecol. Lett. 20, 539–553 (2017). communities. Oikos 89, 409–416 (2000). 5. Intergovernmental Panel on Climate Change (IPCC), “Summary for policymakers” in 24. R. B. Pratt et al., Mortality of resprouting chaparral shrubs after a fire and during a Global Warming of 1.5°C. An IPCC Special Report on the Impacts of Global Warming record drought: Physiological mechanisms and demographic consequences. Glob. of 1.5°C above Pre-Industrial Levels and Related Global Greenhouse Gas Emission Change Biol. 20, 893–907 (2014). Pathways, in the Context of Strengthening the Global Response to the Threat of 25. J. Henzler, H. Weise, N. J. Enright, S. Zander, B. Tietjen, A squeeze in the suitable fire Climate Change, Sustainable Development, and Efforts to Eradicate Poverty, interval: Simulating the persistence of fire-killed plants in a Mediterranean-type V. Masson-Delmotte, Ed. et al. (World Meteorological Organization, Geneva, Swit- ecosystem under drier conditions. Ecol. Modell. 389, 41–49 (2018). zerland, 2018), pp. 1–32. 26. J. S. Clark et al., The impacts of increasing drought on forest dynamics, structure, and 6. D. Frank et al., Effects of climate extremes on the terrestrial carbon cycle: Concepts, biodiversity in the United States. Glob. Change Biol. 22, 2329–2352 (2016). processes and potential future impacts. Glob. Change Biol. 21, 2861–2880 (2015). 27. H. Jactel et al., Drought effects on damage by forest insects and pathogens: A meta- 7. L. Huang et al., Drought dominates the interannual variability in global terrestrial net analysis. Glob. Change Biol. 18, 267–276 (2012). primary production by controlling semi-arid ecosystems. Sci. Rep. 6, 1–7 (2016). 28. J.-C. Svenning, B. Sandel, Disequilibrium vegetation dynamics under future climate 8. S. Vicca et al., Remotely-sensed detection of effects of extreme droughts on gross change. Am. J. Bot. 100, 1266–1286 (2013). primary production. Sci. Rep. 6, 28269 (2016). 29. A. Giuggiola et al., Improvement of water and light availability after thinning at a 9. B. P. Wilcox, Transformative ecosystem change and ecohydology: Ushering in a new xeric site: Which matters more? A dual isotope approach. New Phytol. 210, 108–121 era for watershed management. Ecohydrology 3, 126–130 (2010). (2016). 10. J. F. Johnstone et al., Changing disturbance regimes, ecological memory, and forest 30. M. D. Redmond, P. J. Weisberg, N. S. Cobb, M. J. Clifford, Woodland resilience to resilience. Front. Ecol. Environ. 14, 369–378 (2016). regional drought: Dominant controls on tree regeneration following overstorey 11. R. Seidl et al., Forest disturbances under climate change. Nat. Clim. Chang. 7, 395–402 mortality. J. Ecol. 106, 625–639 (2018). (2017). 31. M. L. Suarez, T. Kitzberger, Recruitment patterns following a severe drought: Long- 12. H. Hartmann et al., Research frontiers for improving our understanding of drought- term compositional shifts in Patagonian forests. Can. J. For. Res. 38, 3002–3010 (2008). induced tree and forest mortality. New Phytol. 218, 15–28 (2018). 32. D. Sauvard, “General biology of bark beetles” in Bark and Wood Boring Insects in Living 13. J. Martínez-Vilalta, F. Lloret, Drought-induced vegetation shifts in terrestrial ecosys- Trees in Europe, a Synthesis, F. Lieutier, K. Day, A. Battisti, J.-C. Grégoire, H. Evans, Eds. tems: The key role of regeneration dynamics. Global Planet. Change 144, 94–108 (Springer, 2004), pp. 63–88. (2016). 33. T. Zhang, Ü. Niinemets, J. Sheffield, J. W. Lichstein, Shifts in tree functional compo- 14. K. J. Anderson-Teixeira et al., Altered dynamics of forest recovery under a changing sition amplify the response of forest biomass to climate. Nature 556, 99–102 (2018). climate. Glob. Change Biol. 19, 2001–2021 (2013). 34. A. T. Trugman, L. D. L. Anderegg, J. D. Shaw, W. R. L. Anderegg, Trait velocities reveal 15. H. Hartmann, H. D. Adams, W. R. L. Anderegg, S. Jansen, M. J. B. Zeppel, Research that mortality has driven widespread coordinated shifts in forest hydraulic trait frontiers in drought-induced tree mortality: Crossing scales and disciplines. New composition. Proc. Natl. Acad. Sci. U.S.A. 117, 8532–8538 (2020). Phytol. 205, 965–969 (2015). 35. B. J. Zeeman, I. D. Lunt, J. W. Morgan, Can severe drought reverse woody plant en- 16. C. D. Canham, L. Murphy, The demography of tree species response to climate: croachment in a temperate Australian woodland? J. Veg. Sci. 25, 928–936 (2014). Sapling and canopy tree survival. Ecosphere 8, e01701 (2017). 36. A. Rigling et al., Driving factors of a vegetation shift from Scots pine to pubescent oak 17. S. Z. Dobrowski et al., Forest structure and species traits mediate projected recruit- in dry Alpine forests. Glob. Change Biol. 19, 229–240 (2013). ment declines in western US tree species. Glob. Ecol. Biogeogr. 24, 917–927 (2015). 37. R. C. Cobb et al., Ecosystem dynamics and management after forest die-off: A global 18. D. M. Bell, J. B. Bradford, W. K. Lauenroth, Early indicators of change: Divergent synthesis with conceptual state-and-transition models. Ecosphere 8, e02034 (2017). climate envelopes between tree life stages imply range shifts in the western United 38. A. AghaKouchak et al., Remote sensing of drought: Progress, challenges and op- States. Glob. Ecol. Biogeogr. 23, 168–180 (2014). portunities. Rev. Geophys. 53, 1–29 (2015). 29728 | www.pnas.org/cgi/doi/10.1073/pnas.2002314117 Batllori et al. 39. F. Lloret, T. Kitzberger, Historical and event-based bioclimatic suitability predicts re- 49. J. VanDerWal, L. Falconi, S. Januckowski, L. Shoo, C. Storlie, SDMTools: Species Dis- gional forest vulnerability to compound effects of severe drought and bark beetle tribution Modelling Tools: Tools for Processing Data Associated with Species Distri- infestation. Glob. Change Biol. 24, 1952–1964 (2018). bution Modelling Exercises (R Package Version 1.1-221, 2014). https://CRAN.R-project. 40. P. Gonzalez, R. P. Neilson, J. M. Lenihan, R. J. Drapek, Global patterns in the vulnerability org/package=SDMTools. Accessed 20 October 2020. of ecosystems to vegetation shifts due to climate change. Glob. Ecol. Biogeogr. 19, 50. Z. Gu, L. Gu, R. Eils, M. Schlesner, B. Brors, circlize implements and enhances circular 755–768 (2010). visualization in R. Bioinformatics 30, 2811–2812 (2014). 41. M. Á. Pérez Navarro et al., Climatic suitability derived from species distribution 51. M. De Cáceres, P. Legendre, F. He, Dissimilarity measurements and the size structure models captures community responses to an extreme drought episode. Ecosystems 22, of ecological communities. Methods Ecol. Evol. 4, 1167–1177 (2013). 77–90 (2019). 52. J. Oksanen et al, vegan: Community Ecology Package (R Package Version 2.5-5, 2019). 42. R. J. Hobbs et al., Novel ecosystems: Theoretical and management aspects of the new https://CRAN.R-project.org/package=vegan. Accessed 20 October 2020. ecological world order. Glob. Ecol. Biogeogr. 15, 1–7 (2006). 53. M. de Cáceres, X. Font, F. Oliva, The management of vegetation classifications with 43. E. Batllori et al., Compound fire-drought regimes promote ecosystem transitions in fuzzy clustering. J. Veg. Sci. 21, 1138–1151 (2010). Mediterranean ecosystems. J. Ecol. 107, 1187–1198 (2019). 54. M. Smithson, J. Verkuilen, A better lemon squeezer? Maximum-likelihood regression 44. M. Michaelian, E. H. Hogg, R. J. Hall, E. Arsenault, Massive mortality of aspen fol- with beta-distributed dependent variables. Psychol. Methods 11, 54–71 (2006). lowing severe drought along the southern edge of the Canadian boreal forest. Glob. 55. F. Cribari-Neto, A. Zeileis, Beta regression in R. J. Stat. Softw. 34, 1–24 (2010). Change Biol. 17, 2084–2094 (2011). 56. T. Hothorn, F. Bretz, P. Westfall, Simultaneous inference in general parametric 45. S. Vicente-Serrano, S. Beguería, J. I. López-Moreno, A multiscalar drought index sensitive to global warming: The standardized precipitation evapotranspiration in- models. Biom. J. 50, 346–363 (2008). dex. J. Clim. 23, 1696–1718 (2010). 57. S. N. Wood, Fast stable restricted maximum likelihood and marginal likelihood esti- 46. M. Berdugo et al., Aridity preferences alter the relative importance of abiotic and mation of semiparametric generalized linear models. J. R. Stat. Soc. 73, 3–36 (2011). biotic drivers on plant species abundance in global drylands. J. Ecol. 107, 190–202 58. J. Kattge et al., TRY—A global database of plant traits. Glob. Change Biol. 17, (2019). 2905–2935 (2011). 47. R. J. Hijmans, S. E. Cameron, J. L. Parra, P. G. Jones, A. Jarvis, Very high resolution 59. S. van Buuren, K. Groothuis-Oudshoorn, mice: Multivariate imputation by chained interpolated climate surfaces for global land areas. Int. J. Climatol. 25, 1965–1978 equations in R. J. Stat. Softw. 45, 1–67 (2011). (2005). 60. R Core Team, R: A Language and Environment for Statistical Computing, (R Foun- 48. R. J. Zomer, A. Trabucco, D. A. Bossio, L. V. Verchot, Climate change mitigation: A dation for Statistical Computing, Vienna, 2018). spatial analysis of global land suitability for clean development mechanism affores- 61. D. M. Olson et al., Terrestrial ecoregions of the world: A new map of life on Earth. tation and reforestation. Agric. Ecosyst. Environ. 126, 67–80 (2008). Bioscience 51, 933–938 (2001). ECOLOGY Batllori et al. PNAS | November 24, 2020 | vol. 117 | no. 47 | 29729