RESEARCH ARTICLE | ECOLOGY OPEN ACCESS Extreme drought impacts have been underestimated in grasslands and shrublands globally Melinda D. Smitha,b,1,2 , Kate D. Wilkinsc,1, Martin C. Holdreged , Peter Wilfahrte, Scott L. Collinsf , Alan K. Knappa,b, Osvaldo E. Salag , Jeffrey S. Dukesh , Richard P. Phillipsi, Laura Yahdjianj , Laureano A. Gherardik, Timothy Ohlerta, Claus Beierl , Lauchlan H. Fraserm , Anke Jentschn , Michael E. Loiko , Fernando T. Maestrep,q , Sally A. Powerr, Qiang Yus, Andrew J. Feltont , Seth M. Munsonu , Yiqi Luov , Hamed Abdoliw , Mehdi Abediw , Concepción L. Aladosx , Juan Albertiy , Moshe Alonz , Hui Anaa, Brian Anackerbb, Maggie Andersone, Harald Augecc,dd, Seton Bachleee,ff, Khadijeh Bahalkehw , Michael Bahngg , Amgaa Batbaatarhh,ii , Taryn Bauerlev , Karen H. Beardd , Kai Behnjj , Ilka Beilkk , Lucio Biancarij , Irmgard Blindowll, Viviana Florencia Bondarukj , Elizabeth T. Borere , Edward W. Borkii , Carlos Martin Bruschettiy , Kerry M. Byrnemm , James F. Cahill Jr.hh, Dianela A. Calvonn , Michele Carbognanioo , Augusto Cardoniy, Cameron N. Carlyleii , Miguel Castillo-­Garciax , Scott X. Changpp, Jeff Chieppar, Marcus V. Cianciarusoqq , Ofer Cohenz, Amanda L. Cordeirorr, Daniela F. Cusackrr , Sven Dahlkell, Pedro Daleoy, Carla M. D'Antonioss , Lee H. Dietterichrr,tt , Tim S. Dohertyuu, Maren Dubbertvv, Anne Ebelingww, Nico Eisenhauerdd,xx , Felícia M. Fischerxx,zz, T'ai G. W. Forteoo , Tobias Gebaueraaa,3 , Beatriz Gozaloq , Aaron C. Greenvilleuu, Karlo G. Guidoni-­Martinsqq , Heather J. Hannuschbbb, Siri Vatsø Haugumccc , Yann Hautierddd , Mariet Heftingddd, Hugh A. L. Henryeee, Daniela Hossdd,xx,yy , Johannes Ingrischgg, Oscar Iribarney, Forest Isbelle , Yari Johnsonfff, Samuel Jordang, Eugene F. Kellyggg, Kaitlin Kimmelhhh, Juergen Kreylingkk , György Kröel-­Dulayiii, Alicia Kröpfljjj , Angelika Kübertkkk, Andrew Kulmatiskid , Eric G. Lamblll , Klaus Steenberg Larsenl , Julie Larsonmmm , Jason Lawsonnnn, Cintia V. Ledernn , Anja Linstädterooo , Jielin Liuppp, Shirong Liuqqq , Alexandra G. Lodgebbb, Grisel Longorrr , Alejandro Loydisss , Junwei Luanttt , Frederick Curtis Lubbeuuu , Craig Macfarlanevvv, Kathleen Mackie-­Haaswww, Andrey V. Malyshevkk, Adrián Maturano-­Ruizq , Thomas Merchantxxx, Daniel B. Metcalfeyyy, Akira S. Morizzz,aaaa, Edwin Mudongobbbb, Gregory S. Newmancccc , Uffe N. Nielsenr , Dale Nimmodddd, Yujie Niun , Paola Nobreqq , Rory C. O'Connormmm, Romà Ogayaeeee,ffff , Gastón R. Oñatibiaj , Ildikó Orbániii,ooo, Brooke Osbornegggg, Rafael Otfinowskihhhh , Meelis Pärteliiii , Josep Penuelaseeee,ffff , Pablo L. Perijjjj , Guadalupe Peternn , Alessandro Petragliaoo , Catherine Picon-­Cochardkkkk, Valério D. Pillaryy , Juan Manuel Piñeiro-­Guerraj,llll , Laura W. Ploughemmmm,4, Robert M. Plowesnnn , Cristy Portales-­Reyesnnnn, Suzanne M. Probervvv, Yolanda Pueyox , Sasha C. Reedoooo , Euan G. Ritchiepppp , Dana Aylén Rodríguezsss , William E. Rogersbbb , Christiane Roscherdd,qqqq, Ana M. Sánchezrrrr , Bráulio A. Santosssss , María Cecilia Scarfósss, Eric W. Seabloome , Baoku Shitttt , Lara Souzacccc,uuuu , Andreas Stampfliwww,vvvv,wwww , Rachel J. Standishvvvv,xxxx , Marcelo Sternbergz , Wei Suntttt, Marie Sünnemanndd,xx, Michelle Tedderyyyy, Pål Thorvaldsenzzzz , Dashuan Tianaaaaa, Katja Tielbörgerbbbbb, Alejandro Valdecantosp,q , Liesbeth van den Brinkbbbbb, Vigdis Vandvikccc , Mathew R. Vankoughnettccccc, Liv Guri Velleddddd , Changhui Wangeeeee, Yi Wangttt, Glenda M. Wardleuu , Christiane Wernerkkk , Cunzheng Weifffff , Georg Wiehlvvv, Jennifer L. Williamsggggg , Amelia A. Wolfhhhhh , Michaela Zeiterwww,vvvv,wwww , Fawei Zhangiiiii, Juntao Zhuaaaaa, Ning Zongaaaaa, and Xiaoan Zuojjjjj Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. Edited by Susan Harrison, University of California, Davis, CA; received June 12, 2023; accepted October 6, 2023 Climate change is increasing the frequency and severity of short-­term (~1 y) drought events—the most common duration of drought—globally. Yet the impact of this intensi- fication of drought on ecosystem functioning remains poorly resolved. This is due in part to the widely disparate approaches ecologists have employed to study drought, variation in the severity and duration of drought studied, and differences among ecosystems in vegetation, edaphic and climatic attributes that can mediate drought impacts. To over- come these problems and better identify the factors that modulate drought responses, we used a coordinated distributed experiment to quantify the impact of short-­term drought on grassland and shrubland ecosystems. With a standardized approach, we imposed ~a single year of drought at 100 sites on six continents. Here we show that loss of a foundational ecosystem function—aboveground net primary production (ANPP)—was 60% greater at sites that experienced statistically extreme drought (1-­in-­100-­y event) The authors declare no competing interest. vs. those sites where drought was nominal (historically more common) in magnitude This article is a PNAS Direct Submission. (35% vs. 21%, respectively). This reduction in a key carbon cycle process with a single Copyright © 2024 the Author(s). Published by PNAS. year of extreme drought greatly exceeds previously reported losses for grasslands and This open access article is distributed under Creative Commons Attribution-­NonCommercial-­NoDerivatives shrublands. Our global experiment also revealed high variability in drought response but License 4.0 (CC BY-­NC-­ND). that relative reductions in ANPP were greater in drier ecosystems and those with fewer Although PNAS asks authors to adhere to United Nations plant species. Overall, our results demonstrate with unprecedented rigor that the global naming conventions for maps (https://www.un.org/ geospatial/mapsgeo), our policy is to publish maps as impacts of projected increases in drought severity have been significantly underestimated provided by the authors. and that drier and less diverse sites are likely to be most vulnerable to extreme drought. 1 M.D.S. and K.D.W. contributed equally to this work. 2 To whom correspondence may be addressed. Email: climate extreme | Drought-­Net | International Drought Experiment | productivity melinda.smith@colostate.edu. 3 Present address: Geo-­konzept Society of Environmental Most terrestrial ecosystems are impacted to some degree by drought, defined meteorolog- Planning mbH, Adelschlag D-­85111, Germany. ically as an anomalous period of low precipitation relative to normal (1). While droughts 4 Present address: National Park Service, Flagstaff, AZ 86001. vary widely with respect to severity, duration, and spatial extent, multi-­year drought events This article contains supporting information online at that incur catastrophic ecological, economic, and societal impacts tend to capture the lion’s https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.​ share of the attention by scientists and the public (e.g., the 1930’s US Dust Bowl, ref. 2; 2309881120/-­/DCSupplemental. the 2000 to 2003 US Southwest drought, ref. 3; the 2012 to 2016 California Drought, Published January 8, 2024. PNAS 2024 Vol. 121 No. 4 e2309881120 https://doi.org/10.1073/pnas.2309881120 1 of 10 ref. 4; the 2001 to 2009 Millennium Drought in Australia, ref. 5; the 2015 to 2017 drought in Cape Town, South Africa, ref. 6). Yet, globally most droughts are short-­term, lasting ~1 Significance y in duration (7). Because short-­term droughts are so numerous, they can cause substantial loss of ecosystem functioning at local, regional, and global scales (8, 9). As Earth’s climate Drought has well-­documented continues to change, short-­term droughts that are statistically extreme in intensity (e.g., societal and economic rare with respect to the long-­term climate record, ref. 10) will become more common (11, consequences. Climate change is 12), with 1-­in-­100-­y droughts potentially happening every 2 to 5 y (7). Indeed, evidence expected to intensify drought to of such drought intensification already exists for some regions (13). Unfortunately, because even more extreme levels, but of the historic rarity of extreme drought, we have limited, and primarily anecdotal, estimates because such droughts have been of the magnitude of their ecological consequences. historically rare, their impact on Knowledge of how short-­term extreme drought may alter ecosystem functioning is ecosystem functioning is not well particularly important for grasslands and shrublands. These ecosystems cover more than known. We experimentally 40% of the ice-­free terrestrial land surface (14, 15) and are found in every region of the globe (15). Grasslands and shrublands are characterized by high variability and frequent imposed the most frequent type of deficits in precipitation (16), and thus, are expected to be the most vulnerable to climate intensified drought—one that is ~1 change (17). Moreover, grasslands and shrublands store more than 30% of the global y in duration—at 100 grassland stock of carbon (15) and contribute significantly to variability in global terrestrial carbon and shrubland sites distributed sinks (18) and atmospheric CO2 concentrations (19). Thus, grassland and shrubland across six continents. We found ecosystems can be expected to cause greater variation in global carbon cycling with inten- that loss of aboveground plant sifying droughts in the future. growth, a key measure of Fortunately, many drought experiments have been conducted in grasslands and shrub- ecosystem function, was 60% lands, relative to other ecosystems (e.g., forests, ref. 1), and a consensus has emerged based greater when short-­term drought on recent meta-­analyses of these studies. These meta-­analyses show the expected—that was extreme (≤1-­in-­100-­y historical drought has negative impacts on multiple aspects of ecosystem functioning, particularly occurrence). This drought-­induced those functions related to C cycling (e.g., productivity)—but also that considerable var- iation in terrestrial ecosystem responses is observed among studies (20, 21). Much of this loss in function greatly exceeds variation could be caused by differences in the magnitude and duration of the droughts previously reported losses for (or alteration in precipitation) imposed among the experiments included in these grasslands and shrublands, meta-­analyses (20, 21). Although most droughts imposed are not statistically extreme, suggesting that the global impacts Wang et al. (21) showed that magnitude and duration were important factors for deter- of projected increases in drought mining variation in ecosystem responses to experimental alterations in precipitation. They severity have been substantially found a linear decrease in ecosystem functioning with greater reductions in precipitation; underestimated. Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. but, over time, productivity became less responsive to altered precipitation (21). One might conclude from this analysis that the effects of droughts, when imposed at statistically (i.e., historically) extreme levels, would result in even further declines in function with the greatest effects manifested in the short term. However, such extreme reductions in precipitation are uncommon in experiments (10); instead, precipitation reductions are for the most part within the range of nominal variability of a particular ecosystem. Consequently, we lack the critical understanding of how grassland and shrubland ecosys- tems will respond to a future where historically extreme droughts will become the norm rather than the exception. Here we report results from the first-­of-­its-­kind coordinated distributed experiment— the International Drought Experiment or IDE—designed to impose a statistically extreme, short-­term (~1 y, Materials and Methods) drought across grassland and shrubland sites globally, using a common methodology (22). At the time of analysis, IDE consisted of 44 sites that experimentally imposed a historically extreme, 1-­in-­100-­y drought treatment for at least a full growing season. The IDE network also provided an additional 56 sites imposing a less severe drought treatment, one that was not extreme by our definition but rather within the range of historic variability (hereafter referred to as nominal drought; Fig. 1 and Materials and Methods). These 100 sites were arrayed across six continents and spanned broad climatic (Fig. 1 and SI Appendix, Table S2) and edaphic gradients (23). At all sites, we measured annual aboveground net primary production (ANPP, Materials and Methods), a foundational component of the global carbon cycle, as a metric of drought-­induced loss of ecosystem functioning in these grasslands and shrublands. The results from this globally distributed experiment allowed us to 1) quantify the effects of short-­term extreme drought on ANPP and determine if this effect differed between grassland and shrubland ecosystems globally, 2) compare the effects of extreme drought to less severe, nominal (or non-­extreme) drought on ANPP, and 3) broadly assess factors potentially contributing to variation in ecosystems’ responses to both extreme and nominal drought. We expected to observe a significant loss in ANPP with extreme drought, and that this loss in ANPP would be greater in grasslands vs. shrublands, con- sistent with past studies (24–26). Furthermore, we expected that extreme drought would suppress ANPP substantially more than nominal drought. We also expected that the extreme drought effects would differ from those derived in previous meta-­analyses of 2 of 10 https://doi.org/10.1073/pnas.2309881120 pnas.org Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. Fig. 1. Geographic extent and climate space encompassed by the IDE. The 100 grasslands (green circles) and shrublands (brown triangles) included in the analysis spanned six continents (A) and broad gradients of mean annual temperature and MAP (B). Closed symbols denote sites (n = 44) that experienced statistically extreme 1-­in-­100-­y drought (i.e., below average annual precipitation during the experiment year). Open symbols denote IDE sites (n = 56) that experienced nominal (not statistically extreme) drought (i.e., average or above-­average annual precipitation during the experiment year). Photos: Shown are drought shelters at representative sites on each continent. Drought shelters were designed to exclude a fixed proportion of each rainfall event from the plots below. The proportion excluded was selected to impose a 1-­in-­100-­y drought for each site during years with average annual precipitation (based on long-­term precipitation records, see Materials and Methods for details; see SI Appendix, Table S2 for site codes). experimental drought results (20, 21, 27), given that past This suggests that ANPP of grasslands and shrublands responds meta-­analyses are subject to publication bias (28) and that their similarly to drought unless droughts are extreme, in which case effect sizes were based on studies that varied widely in the type, these ecosystems are more likely to diverge in their average duration, and magnitude of the drought imposed (29–31). response to even a single year of drought. The 95% CIs for mean reductions in ANPP due to nominal Results and Discussion (non-­extreme) droughts, and for grasslands and shrublands sepa- rately, overlapped with the range of mean effects reported in recent Consistent with most previous research, drought experimentally meta-­analyses of drought experiments (Fig. 2A). This equivalence imposed over ~1 y (<2 y), whether nominal or extreme in mag- between the mean effects of nominal droughts in the IDE network nitude, reduced ANPP relative to ambient conditions (Fig. 1 and and past meta-­analyses occurred despite the wide variety of exper- SI Appendix, Table S3). For those sites that experienced extreme imental protocols for imposing drought treatments included in the drought, ANPP was reduced on average by ~35% overall; by meta-­analyses. In contrast, the effect of extreme drought on ANPP ~38% and 21% for grasslands and shrublands, respectively was well outside the range of these past reported effects, with the (Fig. 2A). Thus, across the 44 sites that experienced extreme reduction in ANPP more than 1.5-­fold greater. Thus, our results drought, grasslands incurred greater losses in ANPP than shrub- suggest that past studies have underestimated the ecosystem effects lands, consistent with previous studies (24–26), though this dif- of statistically extreme droughts—the droughts of the future. ference was not statistically significant (SI Appendix, Table S4). While it is reassuring that there is similarity among the full suite For those sites that experienced nominal drought, ANPP was of sites in our study (n = 100), the subset of sites subjected to suppressed by ~21%, much less (>half ) than in sites experiencing nominal drought (n = 56), and the mean effect sizes of meta-­analyses, extreme drought, and there was a smaller difference between grass- there are several reasons to view the IDE estimates as being more lands and shrublands in these nominal drought responses (Fig. 2A). robust. First, by including results from all sites (including those PNAS 2024 Vol. 121 No. 4 e2309881120 https://doi.org/10.1073/pnas.2309881120 3 of 10 Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. Fig. 2. Response of ANPP to a standardized 1-­y drought for 100 IDE sites. (A) Mean drought response for sites that experienced extreme drought or nominal conditions for: all sites, grasslands, and shrublands (SI Appendix, Tables S3 and S4). The gray bar indicates the range of ANPP loss from −19 to −12.6% found in Song et al. and Wang et al. (20, 21), respectively. Drought response is calculated as: ln(average ANPPDROUGHT/average ANPPCONTROL); 0 (black dashed line) represents no effect of drought, and negative numbers indicate less ANPP in drought vs. control plots. (B) Mean drought response for each site ordered from negative (Top) to positive (Bottom). Site codes and corresponding site information are listed in SI Appendix, Table S2. Shown are 95% CIs for mean site-­level drought responses. *Indicates site with CI that was omitted for clarity because it exceeds the x axis scale. with no evidence of a drought effect, Fig. 2B), we eliminated the responses was also observed with nominal drought, with 39 of 56 long-­standing issue of publication bias affecting meta-­analysis effect sites experiencing a loss in ANPP, but 17 sites displaying high sizes (i.e., bias towards significant results, (28, 32)). In addition, drought resistance. Thus, individual IDE sites still differed in their although the statistical power was relatively low for detecting responsiveness to both extreme and nominal drought despite the drought effects at individual sites—a concern of many global use of common protocols. This begs the question: What factors are change experiments (32)—the large number of IDE sites, almost contributing to the large variation in drought response among sites? twice as many as included in Song et al. (20) and Wang et al. (21), provided broader and in some cases denser geographic coverage. Determinants of Variation in Ecosystem Response to Drought. Furthermore, the standardized experimental design and sampling Although there are myriad factors that may contribute to site-­level protocols we used ensured drought treatments were imposed, and variation in the ANPP responses observed, we focused on seven responses assessed comparably, across all sites. Thus, variation in key abiotic and biotic variables that were reliably available for our dataset should be attributable to ecological differences among >75% of the IDE sites. These included mean annual precipitation sites and not methodological differences inherent in meta-­analyses. (MAP), previous year’s precipitation (relativized by MAP), historic By reducing methodological differences, we anticipated that variability in precipitation (expressed as the interannual coefficient site-­to-­site variability would be reduced in this coordinated, distrib- of variation of MAP), aridity index (AI), soil texture, plant uted experiment (22, 23), at least relative to previous studies. species richness, and the dominant plant growth-­form (expressed Contrary to that expectation, a surprising amount of variation was as proportion of graminoids) of the ecosystem (Materials and still observed in ANPP responses to both extreme and nominal Methods and SI Appendix, Table S5). drought across sites (Fig. 2B). While 79 sites experienced the Past empirical studies have indicated that the factors above may expected losses in ANPP, 21 sites were insensitive to the 1-­y extreme underpin variations in ANPP responses among sites to drought. drought, i.e., control and treatment ANPP means differed by <1% For example, evidence suggests that drier and more arid sites (low or were slightly higher in treatment plots, suggestive of high resist- MAP and AI) tend to be more sensitive to drought than wetter ance to short-­term drought (Fig. 2B). Variation in ecosystem or less arid sites (33–36). But, in addition to MAP (or AI), historic 4 of 10 https://doi.org/10.1073/pnas.2309881120 pnas.org variability in precipitation also may contribute to variation in (low values indicate lower plant water availability or more arid drought responses (37, 38). In this case, sites that experience sites) being related to magnitude of drought responses (SI Appendix, higher year-­to-­year variation in MAP are expected to be less sen- Table S6). Drier sites (lower MAP or greater aridity) experienced sitive to extreme drought, a product of these ecosystems being greater losses in productivity than wetter grassland and shrub- adapted to large interannual variations in precipitation. Research lands. This finding matches studies demonstrating that produc- also suggests that legacy effects of the previous year’s precipitation tion losses at more arid sites are greater in response to drought may play an important role in determining plant productivity, (32, 35–37) and supports the Huxman-­Smith model (35) of such that responses lag behind the increases or decreases in pre- greater sensitivity of ANPP to interannual variation in precipita- cipitation from the previous year (39–41). Finally, the inverse soil tion (and dry years) in more arid sites. Finally, as demonstrated texture hypothesis (42) proposes that plants growing in coarse-­ previously (44), there was moderate evidence for more species-­rich textured (sandy) soils should experience less water stress than sites being more resistant to a loss in productivity than less rich plants growing in fine-­textured soils in relatively arid ecosystems, sites (P = 0.04). with the opposite pattern for ecosystems with higher precipitation. In contrast, we found no evidence that previous year’s precipita- This interaction between soil texture and MAP is expected to be tion (as relativized by MAP), CV in MAP, percent sand (a key amplified with drought, but this prediction has rarely been tested component of soil texture), or the proportion graminoids explained (43). With IDE spanning a broad range of edaphic conditions the variation in drought responses observed across all sites (Fig. 3 (23), we provide one of the first tests of the inverse soil texture and SI Appendix, Table S6). We also found no evidence for an inter- hypothesis on drought responses. active effect between MAP and percent sand on drought response In addition to the abiotic factors listed above, plant species (SI Appendix, Table S7), and thus no support for the inverse soil richness has been shown to influence the magnitude of ecosystem texture hypothesis affecting differential drought sensitivity. response to drought, with more species-­rich communities being more resistant to drought than less species-­rich communities (44). The Importance of Drought Severity. As indicated above, drought There is also abundant evidence that growth forms differ in their magnitude or severity (i.e., % reduction in precipitation relative sensitivity to drought, with grasses and grass-­like plants (i.e., to the control) was an important predictor of ecosystem response graminoids) typically more sensitive to water deficits than woody to drought in the recent meta-­analysis by Wang and et al. (21). plants (45, 46) or forbs (47). Given that the IDE sites represent Because the passive approach to imposing drought employed with a range of plant species richness and graminoid abundance IDE (Fig. 1; Materials and Methods) relies on ambient precipitation (SI Appendix, Table S5), we evaluated the relationship between levels, the actual amount of precipitation that was reduced with the average plot-­level species richness and abundance of graminoids drought treatment at each site (with respect to the long-­term record, and the magnitude of the drought response observed. i.e., MAP) varied with the amount of annual precipitation received Three of these potential sources of variation in drought response during the year of the experiment. We used this variation in drought had statistical support in the IDE dataset. We found weak evidence severity to determine if differences in drought responses could be Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. (P = 0.08) for MAP and moderate evidence (P = 0.02) for aridity explained simply by the magnitude of drought imposed (calculated Fig. 3. Effect of abiotic and biotic factor on drought response. Relationships between 1-­y drought responses across IDE sites and (A) MAP, (B) previous year’s precipitation (relativized by MAP), (C) interannual percent coefficient of variation (CV) of MAP, (D) natural log of the AI, (E) percent sand, (F) average proportion of graminoids, and (G) average richness of plant species. Information on abiotic and biotic characteristics for each site can be found in SI Appendix, Table S5. Model results are summarized in SI Appendix, Table S6. Drought response is calculated as: ln(average ANPPDROUGHT/average ANPPCONTROL); 0 (black dashed line) represents no effect of drought, and negative numbers indicate less ANPP in drought vs. control plots. Lines are shown only for significant relationships. Shaded area represents the 95% CI. *Previous year’s precipitation included the precipitation in the 365 to 730 d preceding the biomass collection date and was relativized by MAP. PNAS 2024 Vol. 121 No. 4 e2309881120 https://doi.org/10.1073/pnas.2309881120 5 of 10 could be dismissed, suggesting that other factors are contributing to high resistance of these sites to a single year of extreme drought. With the above analyses, we were able to evaluate how climate, soil texture, vegetation structure, and drought severity broadly influence drought response. However, given that drought severity was the primary determinant of variation in drought response, other factors that we were unable to include in our analyses are likely contributing to observed high variability in ecosystem response to extreme drought. Moving forward, a key challenge is to determine what these other factors are and identify those eco- system attributes (e.g., soil fertility, root:shoot ratios, plant com- munity composition, plant-­microbe interactions, etc.) that may strongly influence resistance to a single year of extreme drought. Measurement of these factors at the site level and inclusion of these factors in future analyses will be crucial for predicting and mitigating the impacts of extreme drought as climate changes. Despite the uncertainty of what may be determining variation Fig. 4. Relationship between drought severity and drought response. A linear in drought response in grasslands and shrublands globally, our mixed effects model found strong evidence for a negative effect of increasing analysis suggests that overall, ANPP declines as a linear function drought severity (becoming more negative) on drought response for 1 y of of increasing drought severity (Fig. 4). In other words, there was drought across all sites (intercept = −0.11; slope = 0.53; P = 0.009; adjusted R2 = 0.06). Model results for the effects of drought severity on drought response no evidence for catastrophic or nonlinear losses in ANPP when for extreme and nominal sites are summarized in the Main Text. Shaded single-­year droughts become statistically extreme. However, the area represents the 95% CI. Drought response is calculated as: ln(average results from this globally distributed drought experiment do indi- ANPPDROUGHT/average ANPPCONTROL). For drought responses, 0 represents no cate that losses in ANPP are greater than previously expected when effect of drought, negative numbers indicate less ANPP in drought vs. control plots. Drought severity is calculated as: (PrecipDROUGHT-­MAP)/MAP; MAP = mean drought is historically extreme. With climate change, droughts annual precipitation. Because ambient precipitation during the experiment are not only expected to become more extreme, but also more year determines the severity of the imposed drought, positive drought severity frequent and longer in duration. It remains unknown what effects can occur during anomalously wet years when plots beneath drought shelters also experience above average precipitation. The open symbols denote those these aspects of intensified drought may have when overlaid with IDE sites (n = 56) where ambient precipitation was above average, and thus greater losses in ANPP with increased drought severity. the imposed drought was not statistically extreme (1-­in-­100 y). Closed symbols In conclusion, given that many ecosystems, particularly grass- denote those sites (n = 44) with average or below average annual precipitation during the experiment year. All of these IDE sites experienced statistically lands and shrublands, experience substantial interannual variability extreme drought. The filled orange square denotes the mean drought in precipitation (16, 49), it is not surprising that short-­term pre- Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. response for sites experiencing extreme drought whereas the open orange cipitation reductions that are not statistically extreme would result square is the mean for sites that experienced less severe drought. Note that in only “moderate” (~20%) losses in productivity. Even such mod- there was no relationship between drought severity and drought response when only those sites that experienced extreme drought are considered. The erate responses are likely to have important implications for the red dashed lines provide visual guides for 50% and 75% reductions in ANPP. global carbon cycle and the wildlife, livestock, and human popu- lations that rely on plant production. Of greater concern, however, is that grassland sites and grassland and shrubland sites combined as drought severity = (PrecipitationDROUGHT-­MAP)/MAP). As experienced a magnified loss of function (more than 1.8-­ and expected, we found strong evidence (P = 0.009) that increasing 1.5-­fold greater reduction, respectively) when drought was statis- (more negative) drought severity led to larger reductions in ANPP tically extreme for ~1 y. Clearly, with climate change increasing when examined across all 100 sites (Fig. 4). Further, the amount of drought intensity and frequency (50), and given that effects can variability explained doubled when drought severity, MAP, previous linger long after drought ends, even more substantial impacts on year’s precipitation, CV in MAP, percent sand, and proportion of the global carbon cycle can be expected. Indeed, reductions in graminoids were included together in the model (SI Appendix, ANPP exceeding 35% are not often observed in moderate droughts Table S8); though, drought severity remained the only significant of longer duration (24, 51), and a recent study suggests that exper- factor in the model, underscoring the primary importance of this imental droughts may underestimate the magnitude of ANPP loss metric in determining the magnitude of the drought response. by more than half when compared to naturally-­occurring droughts One concern with passively reducing ambient precipitation is (52). Thus, results from our distributed experimental approach that in particularly dry years, differences in ecosystem responses reveal that extreme droughts are likely to substantially slow C between drought and control treatments tend to be minimized sequestration in grasslands and shrublands, surpassing predictions (48). In other words, if ambient precipitation is well below aver- from past meta-­analyses (20, 53) and experiments (9, 52). Finally, age, ANPP would be expected to be low even in control treatments the underlying cause of the striking range in ecosystem responses and further reductions in precipitation with the drought treatment to short-­term extreme drought, from highly resistant to highly may not cause any additional appreciable reductions in ANPP. vulnerable, remains unresolved. Results from our globally distrib- Thus, the difference between the drought and control treatments uted and standardized drought experiments demonstrated little to would be small, resulting in an effect size close to zero. This would no evidence for key factors typically thought to drive ecosystem give the appearance of the site being highly resistant to drought. variability in response to drought: CV of MAP, previous year’s We examined whether this was a possibility for sites with a large precipitation, soil texture, and proportion of graminoids. Yet, we drought severity index (45% reduction in precipitation) and yet found strong support for MAP/aridity and plant species richness a drought response close to zero. We found that this phenomenon being at least partially predictive of ANPP response to extreme may indeed be responsible for the high resistance observed for drought. If traditionally invoked variables do not explain most of these few sites (n = 4), but for most sites that exhibited resistance the cross-­site variation in responses to intensified droughts, we must to drought (n = 17), this potential experimental phenomenon rethink our measurements and experiments to allow us to identify 6 of 10 https://doi.org/10.1073/pnas.2309881120 pnas.org other underexplored factors. Understanding the determinants of investment of time than other measures of ecosystem functioning—a crucial fea- differences in both short-­ and longer-­term drought vulnerability ture of response variables in CDEs and other successful experimental networks, will provide critical insight into both the mitigation potential and such as the Nutrient Network (22, 61). ANPP was estimated either destructively adaptative capacity of ecosystems in a future where today’s extremes and/or non-­destructively using methods appropriate for the particular ecosystem become the norm rather than the exception. as cited in Fahey and Knapp (62), with herbaceous-­dominated sites encouraged to follow the Nutrient Network’s protocols (https://nutnet.org/). We relied on investi- gators to use their expertise in determining the most appropriate methods—either Materials and Methods destructive or non-­destructive—to estimate ANPP for their study system. Sites then separated ANPP estimates into live and dead before further classifying live The IDE. IDE was initiated in 2013 as part of the Drought-­Net Research Coordination biomass by growth form (graminoid, grass, forb, woody, etc.) and submitting Network, funded by the US NSF. Drought-­Net is a global network of researchers com- all estimates in grams of dry biomass per m2. Standing dead biomass could be mitted to understanding how terrestrial ecosystems respond to extreme drought. For separated into current and previous year’s growth where appropriate. all network investigators, we provided and continue to provide standard experimen- tal protocols on the DroughtNet website (droughtnet.weebly.com). Sites must follow Drought Response Metric. For each site, we calculated relative drought these protocols for their data to be included in our analyses. At the time of analysis, response as the ratio of average ANPP in the drought plots compared to average 141 sites had joined our network, but data from only 100 of the sites (Fig. 1A and ANPP in the control plots, as a metric of ecosystem response to imposed drought. SI Appendix, Table S2) had been submitted and/or met our criteria for inclusion in Specifically, we adapted equations from Smith et al. (63) and Kreyling et al. (64) this analysis (SI Appendix, Table S9). These 100 sites were well distributed across to define relative drought response as: ln(ANPPDROUGHT/ANPPCONTROL). gradients of MAT and MAP (Fig. 1B) and represented two ecosystem types: grass- Drought Severity Metric. We calculated the actual severity of the drought that lands and shrublands. Most (>90%) of the sites were dominated by perennials. was imposed during the year of precipitation manipulation using meteorological Furthermore, like most grasslands and shrublands globally, all the sites had some data that was either collected 1) on-­site (site-­submitted) or 2) from a nearby history of management (SI Appendix, Table S10), but only 13 sites that we know of weather station (mean distance = 10.2 km). The nearby weather station data were actively mowed (n = 6), burned (n = 5), or grazed (n = 2). was obtained either from the Global Historical Climatology Network (GHCN, 65) The target for the IDE drought treatment was a statistically extreme, 1-­in-­100-­y or the Climate Hazards Group InfraRed Precipitation with Stations (CHIRPS, 66). drought imposed year-­round. To achieve the target level of extremeness, each When data from GHCN, CHIRPS, or local weather stations were not available, site’s treatment magnitude (reduction in precipitation) was based on the past 100 y we used site-­submitted annual precipitation estimates. We first calculated the of climate data from the site or 100 y of interpolated data from the Terrestrial amount of precipitation each site received in the 365 d preceding collection Precipitation Analysis tool (54). Because precipitation history and variability are of ANPP. We then used this precipitation calculation, the site-­reported percent unique to each site, this approach allowed us to target the common level of reduction in precipitation imposed with the drought treatment, and days the statistical extremeness by allowing the proportional reduction in precipitation to drought shelters were in place prior to the collection of ANPP to estimate the vary across sites (55 and SI Appendix, Fig. S1). This contrasts with the alternative total amount of precipitation reduced at each site. To qualify as 1 y of drought, a approach of imposing a fixed reduction in precipitation (e.g., 50%), which can site’s shelters needed to be in place for at least one full growing season (within result in very different levels of extremeness across sites (55). 1 wk of 120 treatment days prior to harvest), but less than 2 y (within 1 wk of The target level of extreme drought was imposed at each site using infra- Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. 650 d). For the drought severity metric, we compared the estimated precipitation structure that is commonly used in short-­statured ecosystems (56). The infra- received in drought plots to MAP (mm) as reported by each site, and calculated structure consisted of two or more open-­sided shelters, each a minimum of the deviation from this number: (Precipdrought – MAP)/MAP. 2 m × 2 m, with roofs that were partially covered with transparent strips of plastic (either V-­shaped or corrugated). The percentage roof coverage was dictated by Abiotic and Biotic Factors. We included in our analysis seven abiotic and the target level of precipitation reduction (Fig. 1). This shelter design has been biotic factors that have been hypothesized to be important in influencing var- shown to have minimal effects on microclimate (57–59), while matching key iability in drought response among sites (SI Appendix, Table S5). These factors characteristics (e.g., number of consecutive dry days, size of events, number of included: MAP (mm), previous year’s precipitation (mm), historical variability events) of naturally occurring extreme dry years across a range of ecosystems in MAP (interannual CV), AI, soil texture, proportion of graminoids, and site (55). For those sites (n = 9) with both control plots (no infrastructure) and plots richness. As indicated above, we used MAP as reported by each site. We used with an infrastructure control (i.e., structures that mimic the shading of shelters either site-­submitted, GHCN, or CHIRPS data to calculate how much precipita- but allow rain to pass through), there was no evidence that ANPP was affected by tion each site received in the 365 to 730 d preceding the ANPP harvest and then the shelter infrastructure (mean difference = 5.25%; 95% CI = −6.39%, 18.33%; relativized this value by site MAP. We refer to this as previous year’s precipitation t-­value = 0.89; df = 33; P = 0.38). This suggests an absence of significant non-­ in our analyses. Historical variability of precipitation was estimated using the target effects of the drought shelters, which has also been demonstrated in other average coefficient of variation for the 30 y of precipitation from each site using experiments (57–59). GHCN station data. The AI (an estimate of plant moisture availability), was cal- We also chose the shelter infrastructure to impose drought because it is highly culated as the MAP divided by potential evapotranspiration (67); Trabucco et al. cost-­effective and can be consistently deployed across a range of short-­statured (67) uses estimates from World Clim v2 (68) to calculate these values. Lower AI ecosystems, making it amenable for use in a coordinated distributed experiment values indicate lower plant water availability. For the interannual coefficient of (22). However, because we manipulate precipitation passively, the target level of variation in MAP, we pulled data from the Multi-­Source Weighted-­Ensemble drought extremeness may or may not be achieved in any given year, depending Precipitation tool (69). on ambient precipitation amount (48). For example, an above-­average year of As a proxy for soil texture, we used a weighted mean for percent sand in the top precipitation will result in drought treatments that are less severe, while a below-­ 100 cm of soil for each site (n = 96) available in the ISRIC World Soil Information average precipitation year will result in an even more extreme drought than the (70), which yielded information for 96 sites. We used this global dataset because target level. In total, 44 of the 100 IDE sites received average or below-­average site-­level data was only available for 27 of the sites. However, for those 27 sites precipitation in the first year of the treatment, and thus imposed the target sta- we conducted a Pearson correlation test between site-­derived values and the tistically extreme, 1-­in-­100-­y drought (SI Appendix, Fig. S2). The remaining 56 global dataset to confirm whether the global dataset measures were accurate, sites received above-­average precipitation, and thus imposed a non-­extreme which we found to be the case (r = 0.67, P < 0.001). To calculate the proportion (“nominal”) drought, which was within the range of historic variability. of graminoids, we used 77 sites that submitted ANPP data for each functional To be included in this analysis, sites needed to collect peak live aboveground group. For each plot at a site, we divided the ANPP of graminoids by the total biomass as an estimate of annual ANPP (60). While we recognize that there are ANPP (all functional groups added together). We then averaged these proportions numerous other ecosystem processes that can be impacted by drought, ANPP across plots for each site. For richness, we used plant species composition data was selected because it can be comparably estimated and readily standardized submitted by 68 of the 100 sites at the time of analysis. We only used data from across sites. Moreover, ANPP is a low-­cost measurement that requires much less control plots and averaged richness at the plot level for each site. PNAS 2024 Vol. 121 No. 4 e2309881120 https://doi.org/10.1073/pnas.2309881120 7 of 10 Statistical Analyses. For all statistical analyses, we used the language of evi- Forest Service Rocky Mountain Research Station and the USDA Agricultural Research dence (71) to describe our results as an alternative to typical statistical significance Service, and the findings and conclusions are those of the authors and should not testing (i.e., using a significance cut-­off of P-­value ≤ 0.05). With this approach, be construed to represent any official USDA determination of policy. Any use of we ascribe P-­values as following Muff et al. (71): >0.1 = little or no evidence, trade, firm, or product names is for descriptive purposes only and does not imply between 0.05 and 0.10 = weak evidence, between 0.01 and 0.05 = moderate endorsement by the U.S. Government. Additional acknowledgements include: evidence, and ≤0.01 = strong evidence. NSF Long Term Research in Environmental Biology (LTREB DEB 1754106 and To test whether ANPP in drought plots differed from that in control plots, we 2326482) to Arizona State University and Long Term Ecological Research (LTER) used six one-­sided t tests in R (version 4.0.2, 72), identifying whether responses program to New Mexico State University (DEB-­2025166); Sevilleta and Central were significantly less than 0. We chose to use one-­sided t tests given that out Arizona-­Phoenix Long-­Term Ecological Research Programs (NSF DEB-­1655499 ad a priori hypothesis was for drought to result in a reduction in ANPP (rather than DEB-­1832016); The Institute for the Study of Ecological and Evolutionary Climate either a decrease or increase, which would be relevant for an expectation of Impacts supported sites at University of California-­Santa Cruz; USDA Agriculture increased ANPP with drought). However, to examine whether results differed and Food Research Initiative (AFRI) Physiology of Agricultural Plants Program, based on one-­sided vs. two-­sided t tests, we conducted two-­sided t tests for the Grant #2017-­67013-­26191; US Department of Energy (DOE) Environmental six comparisons and found that this did not affect the results, except in the case System Science Program (DE-­SC0021980); Federal Ministry of Education and of the shrubland nominal drought test (one-­sided P-­value = 0.07, two-­sided Research (grant 031B1067C); European Research Council [ERC Grant agreement P-­value = 0.15). We tested sites with nominal and extreme droughts separately, 647038 (BIODESERT)] and Generalitat Valenciana (CIDEGENT/2018/041); USGS and tested ecosystem types together and separately (SI Appendix, Table S3). To Ecosystems Mission Area; CONICET and Universidad de Buenos Aires; Tarbiat test for differences in grassland and shrubland responses to drought across all Modares University; Department of Biology at Kansas State University and NSF sites, we used linear regression (SI Appendix, Table S4). LTER program to Kansas State University (DEB-­144048); Austrian Science Fund, To examine whether the seven abiotic and biotic factors described above Austrian Academy of Sciences and Austrian Research Promotion Agency; Utah explained any variation in the drought response observed, we used linear mixed Agricultural Experimental Station; Grants from the US NSF Long-­Term Ecological effects models for all 100 sites, or separately for sites that experienced extreme or Research Program (LTER) including DEB-­1234162 and DEB-­1831944. Further nominal drought. We built separate models that tested the following explanatory support was provided by the Cedar Creek Ecosystem Science Reserve and the variables: ecosystem type (grassland or shrubland), MAP (mm), previous-­year’s University of Minnesota; US Bureau of Land Management (Grant No. L16AS00178); precipitation (mm, relativized by MAP), average coefficient of variation (%) for NSERC Discovery Grants to J.F.C., E.G.L., and J.L.W.; a joint strategic grant from MAP, AI (scaled using the natural log), average percent sand, average proportion the Alberta Livestock and Meat Agency (now Alberta Ministry of Agriculture and of graminoids, and average plot-­level richness. We also explored the inverse Forestry) and Emissions Reduction Alberta; CNPq/FAPEG -­PELD-­PNE (Site 13); DOE soil texture hypothesis using a linear mixed effects model (site set as a random Office of Science Early Career Award DE-­SC0015898; German Research Foundation effect) to test how the interaction of average percent sand and MAP affected (DFG), Grant number DU1688/1-­1; iDiv and sDiv, the Synthesis Centre of iDiv (FZT drought response. 118, 202548816) and the Gottfried Wilhelm Leibniz Prize (Ei 862/29-­1), both We used a Pearson correlation to test collinearity among drought severity (pro- granted by the DFG; F.M.F. and D.H. received a PhD scholarship by Coordenação portion), length of drought (days), and a categorical variable for whether sites had de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Ministry of Education, ambient precipitation equal to or above (group 1) or below (group 2) MAP. We set Federal Government of Brazil; Australian Research Council (DP210102593); Texas drought severity as our fixed effect as it integrates the length of drought (r = −0.59, A&M Savanna Long-­Term Research and Education Initiative, Sid-­Kyle Foundation, Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. P < 0.001), deviation in ambient precipitation from MAP (r = 0.60, P < 0.001), and Sonora Research Station personnel; Staatsbosbeheer (Dutch State Forestry and a site’s drought shelter design into a single variable. To test the nature of the Service) for giving permission to use the protected nature area at Rhijnauwen; relationship between drought response and various factors (drought severity, MAP, Colorado Agricultural Experiment Station; National Research, Development and previous year’s precipitation, aridity, coefficient of variation, and plant community Innovation Fund (Fund) of Hungary (112576, 129068); City of Boulder Open Space richness) we compared both linear and non-­linear models (asymptotic regression & Mountain Parks Funded Research Program, Garden Club of America, and USDA and general additive model with a spline function set to 3) using AIC. We did not National Institute of Food and Agriculture Predoctoral Fellowship (Project Accession find evidence for a non-­linear response (SI Appendix, Table S11) and proceeded Number, 1019166); German Federal Government (BMBF) through the SPACES with building linear models using the lmer function (package stats) in R. We also initiative (“Limpopo Living Landscapes” project—grant 01LL1304D; “SALLnet” pro- built a multiple linear regression to see whether drought severity combined with ject—grant 01LL1802C); Post-­doctoral fellowship of CAPES–Brasil, Programa CsF; MAP, previous year’s precipitation, proportion graminoids, CV, and average percent PGI UNS 24/ZB81; Swiss NSF, grants 149862, 185110; Environment Research sand could further explain variation in drought response. and Technology Development Fund (JPMEERF15S11420) of the Environmental Data, Materials, and Software Availability. Derived data are provided in Restoration and Conservation Agency of Japan; The Teshio Research Forest of SI Appendix, Table S5. All code and derived data have been deposited in Dryad Hokkaido University provided in situ support; German Academic Exchange (73). All other data are included in the manuscript and/or SI Appendix. Programme (DAAD)–SPACES scholarship for short term visit to Germany 2015 to 2017; Australian Research Council (DP150104199; DP190101968); Alexander von ACKNOWLEDGMENTS. We are grateful to our IDE collaborators who estab- Humboldt Foundation (AvH; grant 33000351); The University of Winnipeg, In-­kind lished and maintained the IDE experiments, collected field data, and shared support provided by Manitoba Beef and Forage Initiatives Inc.; Catalan Government their data with the IDE community (SI Appendix, Table S1). We thank A. Tatarko grants SGR 2921-­1333, the Spanish Government grant PID2022-­140808NB-­I00, for assisting with data management. We also thank the landowners that made and the Fundación Ramón Areces grant CIVP20A6621; PI-­IUNRN 40-­C-­873; the the IDE possible. Research support was provided by the following: US National French government IDEX-­ISITE initiative 16-­IDEX-­0001 (CAP 20-­25); Brazilian Science Foundation (NSF) Research Coordination Network grant to M.D.S., National Research Council (CNPq grant 307689/2014-­0) and Fundação de Amparo O.E.S., and R.P.P. (DEB-­1354732); US Department of Agriculture’s National à Pesquisa do Estado do Rio Grande do Sul (FAPERGS grant 17/2551-­0001106-­6); Institute of Food and Agriculture (USDA-­NIFA) Postdoctoral Fellowship grant Post-­doctoral fellowship of Inter-­American Institute for Global Change Research to K.D.W. (2020-­67034-­31898); USDA-­NIFA Conference Grant to M.D.S. (IAI)CRN3005, which is supported by the US NSF (Grant GEO-­1128040); Great (2020-­67019-­31757); US Geological Survey (USGS) John Wesley Powell Center Western Woodlands Supersite, part of Australia's Terrestrial Ecosystem Research for Analysis and Synthesis grant to M.D.S., S.L.C., and S.M.M.; USGS grant to M.D.S. Network, and thanks to the Department of Biodiversity, Conservation and Attractions (G21AC10266-­00); Global Drought Synthesis Group grant to M.D.S., K.D.W., Western Australia for hosting the site and assistance with construction of shelters; P.W., O.E.S., L.A.G. funded by the NSF Long-­term Ecological Research Network US Department of Agriculture–National Institute of Food and Agriculture award Office and the National Center for Ecological Analysis and Synthesis, University 2019-­68012-­29819; National Council of Scientific and Technological Development of California-­Santa Barbara; National Key Research and Development Program (CNPq, grant number 310340/2016-­0); Research Funds for ICBR (1632021023); of China (2022YFE0128000; 2021YFD2200405); and National Natural Science Israel Ministry of Science and Technology; National Natural Science Foundation of Foundation of China (32061123005, 32071627, 31930078, 31971461). Funding China (31870456); National Research Foundation, Grant No: 116262; DFG Priority for specific experimental sites within this synthesis paper came from the USDA Program SPP-­1803 “EarthShape: Earth Surface Shaping by Biota” (TI 338/14-­1 8 of 10 https://doi.org/10.1073/pnas.2309881120 pnas.org and -­2), and the German Ministry of Education and Research (BMBF); Generalitat Universitario Regional Zona Atlántica, Viedma 85009, Argentina; kkkEcosystem Physiology, Faculty of Environment and Natural Resources, Albert-­Ludwig-­University of Freiburg, Valenciana, Project R2D–RESPONSES TO DESERTIFICATION (CIPROM/2021/001); Freiburg 79110, Germany; lllDepartment of Plant Sciences, University of Saskatchewan, NORWEGIAN RESEARCH COUNCIL MILJØFORSK project 255090 (LandPress: Saskatoon, SK S7N5A8, Canada; mmmRange and Meadow Forage Management Research, Eastern Oregon Agricultural Research Center, US Department of Agriculture (USDA)-­ Land use management to ensure ecosystem service delivery under new soci- Agricultural Research Service, Burns, OR 97720; nnnBrackenridge Field Laboratory, etal and environmental pressures in heathlands); Australian Research Council University of Texas, Austin, TX 78747; oooDepartment of Biodiversity Research and (DP150104199; DP190101968, DP210102593); Centre for Integrative Ecology, Systematic Botany, University of Potsdam, Potsdam 14469, Germany; pppPrataculture Research Institute, Heilongjiang Academy of Agricultural Sciences, Haerbin 150086, China; Deakin University; The Hermon Slade Foundation, Australia; Estonian Research qqq Key Laboratory of Forest Ecology and Environment of National Forestry and Grassland Council (PRG609); Research Station Bad Lauchstädt of the Helmholtz-­Centre for Administration, Ecology and Nature Conservation Institute, Chinese Academy of Forestry, Beijing 100091, China; rrrPrograma de Posgrado en Desarrollo y Medio Ambiente– Environmental Research–UFZ, Germany; National Research Foundation Grant Universidade Federal da Paraíba, Cidade Universitária, Castelo Branco, João Pessoa, PB CSRU180504326326; and European Commission (GYPWORLD, H2020-­MSCA-­ 58051-­900, Brazil; sssCentro de Recursos Naturales Renovables de la Zona Semiárida– RISE-­777803) and Spanish Government (Querpin PID2021-­126927NB-­I00). CONICET, Departamento de Biología, Bioquímica y Farmacia, Universidad Nacional del Sur, Bahía Blanca 8000FTN, Argentina; tttInstitute of Resources and Environment, International Centre for Bamboo and Rattan, Key Laboratory of National Forestry and Grassland Administration and Beijing for Bamboo and Rattan Science and Technology, Beijing Author affiliations: aDepartment of Biology, Colorado State University, Fort Collins, CO 100102, China; uuuInstitute of Botany, Czech Academy of Sciences, Třeboň 379 01, Czech 80523; bGraduate Degree Program in Ecology, Colorado State University, Fort Collins, CO Republic; vvvCommonwealth Scientific and Industrial Research Organization (CSIRO) 80523; cDenver Zoo, Denver, CO 80205; dDepartment of Wildland Resource and the Environment, Wembley, WA 6913, Australia; wwwSchool of Agricultural, Forest and Food Ecology Center, Utah State University, Logan, UT 84322; eDepartment of Ecology, Evolution, Sciences, Bern University of Applied Sciences, Zollikofen 3052, Switzerland; xxxDepartment and Behavior, University of Minnesota, Saint Paul, MN 55108; fDepartment of Biology, of Ecology and Evolutionary Biology, Institute for Arctic and Alpine Research, University of University of New Mexico, Albuquerque, NM 87131; gSchool of Life Sciences, Global Colorado, Boulder, CO 80309; yyyDepartment of Ecology and Environmental Science, Umeå Drylands Center, Arizona State University, Tempe, AZ 85281; hDepartment of Global University, Umeå S-­901 87, Sweden; zzzResearch Center for Advanced Science and Ecology, Carnegie Institution for Science, Stanford, CA 94305; iDepartment of Biology, Technology, University of Tokyo, Meguro, Tokyo 153-­8904, Japan; aaaaGraduate School of Indiana University, Bloomington, IN 47405; jInstituto de Investigaciones Fisiológicas y Environment and Information Sciences, Yokohama National University, Yokohama 240-­ Ecológicas Vinculadas a la Agricultura (IFEVA), National Scientific and Technical Research 8501, Japan; bbbbConservancy-­Communities Living Among Wildlife Sustainably (CLAWS) Council (CONICET), Faculty of Agronomy, University of Buenos Aires, Buenos Aires Botswana, Seronga 00000, Botswana; ccccSchool of Biological Sciences, University of C1417DSE, Argentina; kDepartment of Environmental Science, Policy, and Management, Oklahoma, Norman, OK 73019; ddddGulbali Institute, Charles Sturt University, Albury, NSW University of California, Berkeley, CA 94720; lDepartment of Geosciences and Natural 2640, Australia; eeeeGlobal Ecology Unit Center for Ecological Research and Forestry Resource Management, University of Copenhagen, Frederiksberg C 1958, Denmark; Applications (CREAF)-­National Research Council (CSIC)-­Universitat Autonoma de Barcelona m Department of Natural Resource Science, Thompson Rivers University, Kamloops, BC V2C (UAB), National Research Council (CSIC), Bellaterra, Catalonia 08194, Spain; ffffCenter for 0C8, Canada; nDepartment of Disturbance Ecology and Vegetation Dynamics, Bayreuth Ecological Research and Forestry Applications (CREAF), Cerdanyola del Vallès, Barcelona, Center of Ecology and Environmental Research, University of Bayreuth, Bayreuth 95447, Catalonia 08193, Spain; ggggDepartment of Environment and Society, Utah State University, Germany; oDepartment of Environmental Studies, University of California, Santa Cruz, CA Moab, UT 84532; hhhhDepartment of Biology, The University of Winnipeg, Winnipeg, MB R3B 95064; pDepartamento de Ecologia, Universidad de Alicante, 03690 Alicante, Spain; 2E9, Canada; iiiiInstitute of Ecology and Earth Sciences, University of Tartu, Tartu EE50409, q Instituto Multidisciplinar para el Estudio del Medio “Ramón Margalef”, Universidad de Estonia; jjjjInstituto Nacional de Tecnología Agropecuaria–Universidad Nacional d ela Alicante, 03690 Alicante, Spain; rHawkesbury Institute for the Environment, Western Patagonia Austral–CONICET, Río Gallegos, Caleta Olivia Z9011, Argentina; kkkkUniversité Sydney University, Penrith, NSW 2751, Australia; sSchool of Grassland Science, Beijing Clermont Auvergne, National Research Institute for Agriculture, Food and the Environment, Forestry University, Beijing 100083, China; tDepartment of Land Resources and VetAgro Sup, Research Unit for Grassland Ecosystems, Clermont-­Ferrand 63000, France; llll Environmental Sciences, Montana State University, Bozeman, MT 59717; uU.S. Geological Laboratório de Ecologia Aplicada e Conservação, Departamento de Sistemática e Survey, Southwest Biological Science Center, Flagstaff, AZ 86001; vSoil and Crop Sciences Ecologia, Universidade Federal da Paraíba, Cidade Universitária, Castelo Branco, João Section, School of Integrative Plant Science, Cornell University, Ithaca, NY 14853; Pessoa, PB 58051-­900, Brazil; mmmmDepartment of Biological Sciences, Purdue University, w Department of Range Management, Faculty of Natural Resources and Marine Sciences, West Lafayette, IN 47907; nnnnDepartment of Biology, Saint Louis University, St. Louis, MO Tarbiat Modares University, Noor 46417-­76489, Iran; xDepartamento de Biodiversidad y 63103; ooooU.S. Geological Survey, Southwest Biological Science Center, Moab, UT 84532; pppp Restauración, Instituto Pirenaico de Ecología, Consejo Superior de Investigaciones Centre for Integrative Ecology, School of Life and Environmental Sciences, Deakin Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. Científicas (CSIC), Zaragoza 50059, Spain; yLaboratorio de Ecología, Instituto de University, Burwood, VIC 3125, Australia; qqqqDepartment of Physiological Diversity, Investigaciones Marinas y Costeras (IIMyC), Universidad Nacional de Mar del Plata Helmholtz-­Centre for Environmental Research–UFZ, Leipzig 04318, Germany; rrrr (UNMdP)-­Consejo Nacional de Investigación Ciencia y Técnica (CONICET), CC 1260 Correo Department of Biology and Geology, Rey Juan Carlos University, Madrid 28032, Spain; ssss Central, Mar del Plata B7600WAG, Argentina; zSchool of Plant Sciences and Food Security, Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba, Cidade Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel; aaSchool of Ecology and Universitária, Castelo Branco, João Pessoa, PB 58051-­900, Brazil; ttttInstitute of Grassland Environment, Ningxia University, Yinchuan 750021, China; bbCity of Boulder Open Space Science, Key Laboratory of Vegetation Ecology of the Ministry of Education, Jilin Songnen and Mountain Parks, Boulder, CO 80301; ccDepartment of Community Ecology, Helmholtz-­ Grassland Ecosystem National Observation and Research Station, Northeast Normal Centre for Environmental Research–UFZ, Halle 06120, Germany; ddGerman Centre for University, Changchun 130024, China; uuuuOklahoma Biological Survey, University of Integrative Biodiversity Research (iDiv) Halle-­Jena-­Leipzig, Leipzig 04103, Germany; Oklahoma, Norman, OK 73019; vvvvInstitute of Plant Sciences, University of Bern, Bern ee Division of Biology, Kansas State University, Manhattan, KS 66506; ffLI-­COR Biosciences, 3013, Switzerland; wwwwOeschger Center for Climate Change Research, University of Bern, 4647 Superior Street, Lincoln, NE 68505; ggDepartment of Ecology, University of Innsbruck, Bern 3012, Switzerland; xxxxEnvironmental and Conservation Sciences, Murdoch University, Innsbruck 6020, Austria; hhDepartment of Biological Sciences, University of Alberta, Murdoch, WA 6150, Australia; yyyySchool of Life Sciences, University of Kwazulu-­Natal, Edmonton, AB T6G 2E9, Canada; iiDepartment of Agricultural, Food and Nutritional Science, Pietermaritzburg 3201, South Africa; zzzzNorwegian Institute of Bioeconomy Research, University of Alberta, Edmonton, AB T6G 2P5, Canada; jjInstitute of Crop Science and Department of Landscape and Biodiversity, Tjøtta 8860, Norway; aaaaaKey Laboratory of Resource Conservation, Department of Plant Nutrition, University of Bonn, Bonn 53115, Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Germany; kkInstitute of Botany and Landscape Ecology, Department of Experimental Plant Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China; bbbbb Ecology, University of Greifswald, Greifswald D-­17498, Germany; llBiological Station of Plant Ecology Group, Department of Biology, University of Tübingen, Tübingen 72076, Hiddensee, Department of Biology, University of Greifswald, Kloster D-­18565, Germany; Germany; cccccNova Scotia Community College, Annapolis Valley Campus, Applied Research, mm Department of Environmental Science and Management, California State Polytechnic Middleton, NS B0S 1P0, Canada; dddddMøreforsking, Aalesund 6021, Norway; eeeeeCollege of University, Humboldt, Arcata, CA 95521; nnUniversidad Nacional de Río Negro, Centro de Grassland Science, Shanxi Agricultural University, Jinzhong 030801, China; fffffState Key Estudios Ambientales desde la NorPatagonia (CEANPa), Sede Atlántica–CONICET, Viedma Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy 8500, Argentina; ooDepartment of Chemistry, Life Sciences and Environmental of Sciences, Beijing 100093, China; gggggDepartment of Geography and Biodiversity Sustainability, University of Parma, Parma I-­43124, Italy; ppDepartment of Renewable Research Centre, University of British Columbia, Vancouver, BC V6T 1Z4, Canada; hhhhh Resources, University of Alberta, Edmonton, AB T6G 2E3, Canada; qqDepartment of Ecology, Department of Integrative Biology, University of Texas, Austin, TX 78712; iiiiiKey Universidade Federal de Goiás, Goiânia, GO 74690-­900, Brazil; rrDepartment of Ecosystem Laboratory of Adaptation and Evolution of Plateau Biota, Northwest Institute of Plateau Science and Sustainability, Colorado State University, Fort Collins, CO 80523; ssDepartment Biology, Chinese Academy of Sciences, Xining, Qinghai 810008, China; and jjjjjUrat Desert-­ of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, CA 93106; grassland Research Station, Northwest Institute of Eco-­Environment and Resources, tt US Army Engineer Research and Development Center, Environmental Laboratory, Chinese Academy of Science, Lanzhou 730000, China Vicksburg, MS 39180; uuSchool of Life and Environmental Sciences, The University of Sydney, Camperdown, NSW 2006, Australia; vvIsotope Biogeochemistry and GasFluxes, Author contributions: M.D.S., K.D.W., S.L.C., A.K.K., O.E.S., J.S.D., R.P.P., and L.Y. designed Leibniz-­Zentrum fürAgrarlandschaftsforschung (ZALF), Müncheberg 15374, Germany; research; M.D.S., K.D.W., J.S.D., L.H.F., A.J., M.E.L., F.T.M., S.A.P., Q.Y., A.J.F., S.M.M., Y.L., ww Institute of Ecology and Evolution, Friedrich Schiller University Jena, Jena 07743, H. Abdoli, M. Abedi, C.L.A., J.A., M. Alon, H. An, B.A., M. Anderson, H. Auge, S.B., K. Bahalkeh, Germany; xxInstitute of Biology, Leipzig University, Leipzig 04103, Germany; yyDepartment M.B., A.B., T.B., K.H.B., K. Behn, I. Beil, L.B., I. Blindow, V.F.B., E.T.B., E.W.B., C.M.B., K.M.B., of Ecology, Universidade Federal do Rio Grande do Sul, Porto Alegre 91501-­970, Brazil; J.F.C., D.A.C., M.C., A.C., C.N.C., M.C.-­G., S.X.C., J.C., M.V.C., O.C., A.L.C., D.F.C., S.D., P.D., zz Centro de Investigaciones sobre Desertificación, Consejo Superior de Investigaciones C.M.D., L.H.D., T.S.D., M.D., A.E., N.E., F.M.F., T.G.W.F., T.G., B.G., A.C.G., K.G.G.-­M., H.J.H., Científicas (CSIC)-­Universitat Valencia (UV) -­ Generalitat Valenciana (GV), Valencia 46113, S.V.H., Y.H., M.H., H.A.L.H., D.H., J.I., O.I., F.I., Y.J., S.J., E.F.K., K.K., J.K., G.K.-­D., A. Kröpfl, Spain; aaaGeobotany, Faculty of Biology, University of Freiburg, Freiburg D-­79104, Germany; A. Kübert, A. Kulmatiski, E.G.L., K.S.L., J. Larson, J. Lawson, C.V.L., A. Linstädter, J. Liu, bbb Department of Ecology and Conservation Biology, Texas A&M University, College S.L., A.G.L., G.L., A. Loydi, J. Luan, F.C.L., C.M., K.M.-­H., A.V.M., A.M.-­R., T.M., D.B.M., Station, TX 77843; cccDepartment of Biological Sciences, University of Bergen, Bergen 5007, A.S.M., E.M., G.S.N., U.N.N., D.N., Y.N., P.N., R.C.O., R. Ogaya, G.R.O., I.O., B.O., Norway; dddEcology and Biodiversity Group, Department of Biology, Utrecht University, R. Otfinowski, M.P., J.P., P.L.P., G.P., A.P., C.P.-­C., V.D.P., J.M.P.-­G., L.W.P., R.M.P., C.P.-­R., Utrecht, 3584 CH, Netherlands; eeeDepartment of Biology, University of Western Ontario, S.M.P., Y.P., S.C.R., E.G.R., D.A.R., W.E.R., C.R., A. Sánchez, B.A.S., M.C.S., E.W.S., B.S., L.S., London, ON N6A 5B7, Canada; fffU.S. Army Corps of Engineers, Sacramento, CA 95814; A. Stampfli, R.J.S., M. Sternberg, W.S., M. Sünnemann, M.T., P.T., D.T., K.T., A.V., L.v.d.B., ggg Department of Soil and Crop Sciences, Colorado State University, Fort Collins, CO 80523; V.V., M.R.V., L.G.V., C. Wang, Y.W., G.M.W., C. Werner, C. Wei, G.W., J.L.W., A.A.W., M.Z., hhh Global Water Security Center, The University of Alabama, Tuscaloosa, AL 35487; iiiCentre F.Z., J.Z., N.Z., and X.Z. performed research; K.D.W., M.C.H., P.W., L.A.G., and T.O. for Ecological Research, Institute of Ecology and Botany, Vácrátót 2163, Hungary; analyzed data; and M.D.S., K.D.W., M.C.H., P.W., S.L.C., A.K.K., O.E.S., J.S.D., R.P.P., L.Y., jjj Departamento de Gestión Agropecuaria, Universidad Nacional del Comahue, Centro and C.B. wrote the paper. PNAS 2024 Vol. 121 No. 4 e2309881120 https://doi.org/10.1073/pnas.2309881120 9 of 10 1. I. J. Slette et al., How ecologists define drought, and why we should do better. Global Change Biol. 38. A. K. Knapp et al., Consequences of more extreme precipitation regimes for terrestrial ecosystems. 25, 3193–3200 (2019). BioScience 58, 811–821 (2008). 2. A. K. Knapp et al., Resolving the Dust Bowl paradox of grassland responses to extreme drought. 39. R. A. Sherry et al., Lagged effects of experimental warming and doubled precipitation on annual Proc. Natl. Acad. Sci. U.S.A. 117, 22249–22255 (2020). and seasonal aboveground biomass production in a tallgrass prairie. Global Change Biol. 14, 3. D. D. Breshears et al., Rangeland responses to predicted increases in drought extremity. Rangelands 2923–2936 (2008). 38, 191–196 (2016). 40. J. Dudney et al., Lagging behind: Have we overlooked previous-­year rainfall effects in annual 4. P. A. Ullrich et al., California’s drought of the future: A midcentury recreation of the exceptional grasslands? J. Ecol. 105, 484–495 (2017). conditions of 2012–2017. Earth’s Future 6, 1568–1587 (2018). 41. L. Yahdjian, O. E. Sala, Vegetation structure contrains primary production response to water 5. A. I. J. M. van Dijk et al., The Millennium Drought in southeast Australia (2001–2009): Natural availability in the Patagonian Steppe. Ecology 87, 952–962 (2006). and human causes and implications for water resources, ecosystems, economy, and society. Water 42. I. Noy-­Meir, Desert ecosystems: Environment and producers. Annu. Rev. Ecol. Syst. 4, 25–51 (1973). Resour. Res. 49, 1040–1057 (2013). 43. O. E. Sala, W. J. Parton, W. K. Lauenroth, L. A. Joyce, Primary production of the central grassland 6. A. Maxmen, Cape Town scientists prepare for ‘Day Zero’. Nature 554, 2 (2018). region of the United States. Ecology 69, 40–45 (1988). 7. G. Naumann et al., Global changes in drought conditions under different levels of warming. 44. F. Isbell et al., Biodiversity increases the resistance of ecosystem productivity to climate extremes. Geophys. Res. Lett. 45, 3285–3296 (2018). Nature 526, 574–577 (2015). 8. P. Ciais et al., Europe-­wide reduction in primary productivity caused by the heat and drought in 45. A. Kulmatiski, K. H. Beard, Woody plant encroachment facilitated by increased precipitation 2003. Nature 437, 529–533 (2005). intensity. Nat. Clim. Change 3, 833–837 (2013). 9. L. Du et al., Global patterns of extreme drought-­induced loss in land primary production: 46. D. E. Winkler, J. Belnap, D. Hoover, S. C. Reed, M. C. Duniway, Shrub persistence and increased grass Identifying ecological extremes from rain-­use efficiency. Sci. Total Environ. 628–629, 611–620 mortality in response to drought in dryland systems. Global Change Biol. 25, 3121–3135 (2019). (2018). 47. K. A. Mackie, M. Zeiter, J. M. G. Bloor, A. Stampfli, Plant functional groups mediate drought 10. M. D. Smith, The ecological role of climate extremes: Current understanding and future prospects: resistance and recovery in a multisite grassland experiment. J. Ecol. 107, 937–949 (2019). Ecological role of climate extremes. J. Ecol. 99, 651–655 (2011). 48. D. L. Hoover, K. R. Wilcox, K. E. Young, Experimental droughts with rainout shelters: A 11. IPCC, "The physical science basis. Contribution of working group 1 to the fifth assessment report methodological review. Ecosphere 9, e02088 (2018), 10.1002/ecs2.2088. of the intergovernmental panel on climate change" (Cambridge University Press, Cambridge, UK, 49. D. A. Frank, R. S. Inouye, Temporal variation in actual evapotranspiration of terrestrial ecosystems: 2013). Patterns and ecological implications. J. Biogeogr. 21, 401 (1994). 12. B. I. Cook, T. R. Ault, J. E. Smerdon, Unprecedented 21st century drought risk in the American 50. F. Chiang, O. Mazdiyasni, A. AghaKouchak, Evidence of anthropogenic impacts on global drought Southwest and Central Plains. Sci. Adv. 1, e1400082 (2015). frequency, duration, and intensity. Nat. Commun. 12, 2754 (2021). 13. A. P. Williams, B. I. Cook, J. E. Smerdon, Rapid intensification of the emerging southwestern North 51. F. Zhang et al., When does extreme drought elicit extreme ecological responses? J. Ecol. 107, American megadrought in 2020–2021. Nat. Clim. Change 12, 232–234 (2022), 10.1038/s41558-­ 2553–2563 (2019). 022-­01290-­z. 52. G. Kröel-­Dulay et al., Field experiments underestimate aboveground biomass response to drought. 14. F. T. Maestre et al., Biogeography of global drylands. New Phytol. 231, 540–558 (2021). Nat. Ecol. Evol. 6, 540–545 (2022), 10.1038/s41559-­022-­01685-­3. 15. R. P. White, S. Murray, M. Rohweder, Grassland Ecosystems (World Resources Institute, Washington, 53. C. Wang, Y. Sun, H. Y. H. Chen, J. Yang, H. Ruan, Meta-­analysis shows non-­uniform responses of DC, 2000). above-­and belowground productivity to drought. Sci. Total Environ. 782, 146901 (2021). 16. A. K. Knapp, M. D. Smith, Variation among biomes in temporal dynamics of aboveground primary 54. N. P. Lemoine, J. Sheffield, J. S. Dukes, A. K. Knapp, M. D. Smith, Terrestrial Precipitation Analysis production. Science 291, 481–484 (2001). (TPA): A resource for characterizing long-­term precipitation regimes and extremes. Methods Ecol. 17. C. A. E. Strömberg, A. C. Staver, The history and challenge of grassy biomes. Science 377, 592–593 Evol. 7, 1396–1401 (2016). (2022). 55. A. K. Knapp et al., Characterizing differences in precipitation regimes of extreme wet and dry years: 18. A. Ahlström, J. Xia, A. Arneth, Y. Luo, B. Smith, Importance of vegetation dynamics for future Implications for climate change experiments. Global Change Biol. 21, 2624–2633 (2015). terrestrial carbon cycling. Environ. Res. Lett. 10, 054019 (2015). 56. L. Yahdjian, O. E. Sala, A rainout shelter design for intercepting different amounts of rainfall. 19. B. Poulter et al., Contribution of semi-­arid ecosystems to interannual variability of the global carbon Oecologia 133, 95–101 (2002). cycle. Nature 509, 600–603 (2014). 57. N. B. English, J. F. Weltzin, A. Fravolini, L. Thomas, D. G. Williams, The influence of soil texture and 20. J. Song et al., A meta-­analysis of 1,119 manipulative experiments on terrestrial carbon-­cycling vegetation on soil moisture under rainout shelters in a semi-­desert grassland. J. Arid Environ. 63, responses to global change. Nat. Ecol. Evol. 3, 1309–1320 (2019). 324–343 (2005). 21. J. Wang et al., Precipitation manipulation and terrestrial carbon cycling: The roles of treatment 58. C. Signarbieux, U. Feller, Effects of an extended drought period on physiological properties of magnitude, experimental duration and local climate. Global Ecol. Biogeogr. 30, 1909–1921 (2021). grassland species in the field. J. Plant Res. 125, 251–261 (2012). Downloaded from https://www.pnas.org by 83.41.84.181 on January 12, 2024 from IP address 83.41.84.181. 22. L. H. Fraser et al., Coordinated distributed experiments: An emerging tool for testing global 59. M. E. Loik, J. C. Lesage, T. M. Brown, D. O. Hastings, Drought-­Net rainfall shelters did not cause hypotheses in ecology and environmental science. Front. Ecol. Environ. 11, 147–155 (2013). nondrought effects on photosynthesis for California central coast plants. Ecohydrology 12, e2138 23. L. Yahdjian et al., Why coordinated distributed experiments should go global. BioScience 71, (2019), 10.1002/eco.2138. 918–927 (2021). 60. A. K. Knapp, J. M. Briggs, D. L. Childers, O. E. Sala, “Estimating aboveground net primary production 24. J. Gao, L. Zhang, Z. Tang, S. Wu, A synthesis of ecosystem aboveground productivity and its process in grassland-­and herbaceous-­dominated ecosystems” in Principles and Standards for Measuring variables under simulated drought stress. J. Ecol. 107, 2519–2531 (2019). Primary Production, T. J. Fahey, A. K. Knapp, Eds. (Oxford University Press Inc, New York, NY, 2007), 25. L. Zhang et al., Drought events and their effects on vegetation productivity in China. Ecosphere 7, pp. 27–48. e01591 (2016), 10.1002/ecs2.1591. 61. E. T. Borer et al., Finding generality in ecology: A model for globally distributed experiments. 26. S. Báez, S. L. Collins, W. T. Pockman, J. E. Johnson, E. E. Small, Effects of experimental rainfall Methods Ecol. Evol. 5, 65–73 (2014). manipulations on Chihuahuan Desert grassland and shrubland plant communities. Oecologia 172, 62. T. J. Fahey, A. K. Knapp, Eds., Principles and Standards for Measuring Primary Production (Oxford 1117–1127 (2013). University Press, New York, 2007). https://oxford.universitypressscholarship.com/10.1093/ 27. Z. Wu, P. Dijkstra, G. W. Koch, J. Peñuelas, B. A. Hungate, Responses of terrestrial ecosystems to acprof:oso/9780195168662.001.0001/acprof-­9780195168662. temperature and precipitation change: A meta-­analysis of experimental manipulation. Global 63. M. D. Smith, K. R. Wilcox, S. A. Power, D. T. Tissue, A. K. Knapp, Assessing community and Change Biol. 17, 927–942 (2011). ecosystem sensitivity to climate change–Toward a more comparative approach. J. Veg. Sci. 28, 28. J. Koricheva, J. Gurevitch, Uses and misuses of meta-­analysis in plant ecology. J. Ecol. 102, 235–237 (2017). 828–844 (2014). 64. J. Kreyling et al., Species richness effects on grassland recovery from drought depend on 29. K. J. Komatsu et al., Global change effects on plant communities are magnified by time and the community productivity in a multisite experiment. Ecol. Lett. 20, 1405–1413 (2017). number of global change factors imposed. Proc. Natl. Acad. Sci. U.S.A. 116, 17867–17873 (2019). 65. J. H. Lawrimore et al. “Global historical climatology network-­monthly (GHCN-­M), version 3” (NOAA 30. J. Peñuelas et al., Nonintrusive field experiments show different plant responses to warming and National Centers for Environmental Information, Asheville, NC, 2011), 10.7289/V5X34VDR. drought among sites, seasons, and species in a North-­South European gradient. Ecosystems 7, 66. C. C. Funk et al., “A quasi-­global precipitation time series for drought monitoring: U.S. Geological 598–612 (2004), 10.1007/s10021-­004-­0179-­7. Survey Data Series Report Number 832” (U.S. Geological Survey, Reston, VA, 2014), p. 4. 31. A. Batbaatar, C. N. Carlyle, E. W. Bork, S. X. Chang, J. F. Cahill, Multi-­year drought alters plant species 67. R. J. Zomer, J. Xu, A. Trabucco, Version 3 of the Global Aridity Index and Potential Evapotranspiration composition more than productivity across northern temperate grasslands. J. Ecol. 110, 197–209 Database, Scientific Data 9, 409 (2022). Figshare. 10.6084/m9.figshare.7504448.v2. Accessed 18 (2022). April 2022. 32. Y. Yang, H. Hillebrand, M. Lagisz, I. Cleasby, S. Nakagawa, Low statistical power and overestimated 68. S. E. Fick, R. J. Hijmans, WorldClim 2: New 1-­km spatial resolution climate surfaces for global land anthropogenic impacts, exacerbated by publication bias, dominate field studies in global change areas. Int. J. Climatol. 37, 4302–4315 (2017). biology. Global Change Biol. 28, 969–989 (2021). 69. H. E. Beck et al., MSWEP V2 global 3-­hourly 0.1° precipitation: Methodology and quantitative 33. C. Golodets et al., Climate change scenarios of herbaceous production along an aridity gradient: assessment. Bull. Am. Meteorol. Soc. 100, 473–500 (2019). Vulnerability increases with aridity. Oecologia 177, 971–979 (2015). 70. N. H. Batjes, E. Ribeiro, A. J. M. van Oostrum, Standardised soil profile data for the world (WoSIS 34. A. K. Knapp et al., Differential sensitivity to regional-­scale drought in six central US grasslands. Snapshot–September 2019), (2019), 10.17027/isric-­wdcsoils.20190901. Oecologia 177, 949–957 (2015). 71. S. Muff, E. B. Nilsen, R. B. O’Hara, C. R. Nater, Rewriting results sections in the language of evidence. 35. T. E. Huxman et al., Convergence across biomes to a common rain-­use efficiency. Nature 429, Trends Ecol. Evol. 37, 203–210 (2022). 651–654 (2004). 72. R Core Team, R: A Language and Environment for Statistical Computing (R Foundation for Statistical 36. G. E. Maurer, A. J. Hallmark, R. F. Brown, O. E. Sala, S. L. Collins, Sensitivity of primary production to Computing, Vienna, Austria, 2021), https://www.R-­project.org/. precipitation across the United States. Ecol. Lett. 23, 527–536 (2020). 73. M. D. Smith et al., Data for “Extreme drought impacts have been underestimated in grasslands 37. E. Hou et al., Divergent responses of primary production to increasing precipitation variability in and shrublands globally.” Dryad. https://doi.org/10.5061/dryad.3j9kd5lrb. Deposited 25 October global drylands. Global Change Biol. 27, 5225–5237 (2021). 2023. 10 of 10 https://doi.org/10.1073/pnas.2309881120 pnas.org