Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Hydraulic diversity of forests regulates ecosystem resilience during drought


Plants influence the atmosphere through fluxes of carbon, water and energy1, and can intensify drought through land–atmosphere feedback effects2,3,4. The diversity of plant functional traits in forests, especially physiological traits related to water (hydraulic) transport, may have a critical role in land–atmosphere feedback, particularly during drought. Here we combine 352 site-years of eddy covariance measurements from 40 forest sites, remote-sensing observations of plant water content and plant functional-trait data to test whether the diversity in plant traits affects the response of the ecosystem to drought. We find evidence that higher hydraulic diversity buffers variation in ecosystem flux during dry periods across temperate and boreal forests. Hydraulic traits were the predominant significant predictors of cross-site patterns in drought response. By contrast, standard leaf and wood traits, such as specific leaf area and wood density, had little explanatory power. Our results demonstrate that diversity in the hydraulic traits of trees mediates ecosystem resilience to drought and is likely to have an important role in future ecosystem–atmosphere feedback effects in a changing climate.

This is a preview of subscription content, access via your institution

Relevant articles

Open Access articles citing this article.

Access options

Rent or buy this article

Prices vary by article type



Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Variation in hydraulic traits mediates ecosystem flux response to drought.
Fig. 2: Ecosystem sensitivity to drought as a function of community variation in hydraulic safety margin.
Fig. 3: Forest ecosystem response to drought estimated from remote-sensing-derived vegetation water content variation is influenced by species richness.

Data availability

Eddy flux data are available at; community trait data are available at; detailed trait data are available in Extended Data Figs. 110 and at, and from a previous publication18.


  1. Bonan, G. B. Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science 320, 1444–1449 (2008).

    ADS  CAS  PubMed  Google Scholar 

  2. Seneviratne, S. I. et al. Impact of soil moisture–climate feedbacks on CMIP5 projections: first results from the GLACE-CMIP5 experiment. Geophys. Res. Lett. 40, 5212–5217 (2013).

    ADS  Google Scholar 

  3. Berg, A. et al. Land–atmosphere feedbacks amplify aridity increase over land under global warming. Nat. Clim. Change 6, 869–874 (2016).

    ADS  Google Scholar 

  4. Dirmeyer, P. A. Vegetation stress as a feedback mechanism in midlatitude drought. J. Clim. 7, 1463–1483 (1994).

    ADS  Google Scholar 

  5. Sulman, B. N. et al. High atmospheric demand for water can limit forest carbon uptake and transpiration as severely as dry soil. Geophys. Res. Lett. 43, 9686–9695 (2016).

    ADS  CAS  Google Scholar 

  6. Novick, K. A. et al. The increasing importance of atmospheric demand for ecosystem water and carbon fluxes. Nat. Clim. Change 6, 1023–1027 (2016).

    ADS  CAS  Google Scholar 

  7. Khanna, J., Medvigy, D., Fueglistaler, S. & Walko, R. Regional dry-season climate changes due to three decades of Amazonian deforestation. Nat. Clim. Change 7, 200–204 (2017).

    ADS  Google Scholar 

  8. Teuling, A. J. et al. Contrasting response of European forest and grassland energy exchange to heatwaves. Nat. Geosci. 3, 722–727 (2010).

    ADS  CAS  Google Scholar 

  9. IPCC. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change. (Cambridge Univ. Press, Cambridge, 2012).

  10. Tilman, D. Wedin, D. & Knops, J. Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379, 718–720 (1996).

    ADS  CAS  Google Scholar 

  11. Doak, D. F. et al. The statistical inevitability of stability–diversity relationships in community ecology. Am. Nat. 151, 264–276 (1998).

    CAS  PubMed  Google Scholar 

  12. Tilman, D., Isbell, F. & Cowles, J. M. Biodiversity and ecosystem functioning. Annu. Rev. Ecol. Evol. Syst. 45, 471–493 (2014).

    Google Scholar 

  13. Naeem, S. & Wright, J. P. Disentangling biodiversity effects on ecosystem functioning: deriving solutions to a seemingly insurmountable problem. Ecol. Lett. 6, 567–579 (2003).

    Google Scholar 

  14. Mori, A. S., Furukawa, T. & Sasaki, T. Response diversity determines the resilience of ecosystems to environmental change. Biol. Rev. Camb. Philos. Soc. 88, 349–364 (2013).

    PubMed  Google Scholar 

  15. Scheiter, S., Langan, L. & Higgins, S. I. Next-generation dynamic global vegetation models: learning from community ecology. New Phytol. 198, 957–969 (2013).

    PubMed  Google Scholar 

  16. Matheny, A. M., Mirfenderesgi, G. & Bohrer, G. Trait-based representation of hydrological functional properties of plants in weather and ecosystem models. Plant Divers. 39, 1–12 (2017).

    PubMed  Google Scholar 

  17. Anderegg, W. R. L. et al. Woody plants optimise stomatal behaviour relative to hydraulic risk. Ecol. Lett. 21, 968–977 (2018).

    PubMed  Google Scholar 

  18. Choat, B. et al. Global convergence in the vulnerability of forests to drought. Nature 491, 752–755 (2012).

    ADS  CAS  PubMed  Google Scholar 

  19. Sperry, J. S. et al. Pragmatic hydraulic theory predicts stomatal responses to climatic water deficits. New Phytol. 212, 577–589 (2016).

    CAS  PubMed  Google Scholar 

  20. Brodribb, T. J. Xylem hydraulic physiology: the functional backbone of terrestrial plant productivity. Plant Sci. 177, 245–251 (2009).

    CAS  Google Scholar 

  21. Anderegg, W. R. Spatial and temporal variation in plant hydraulic traits and their relevance for climate change impacts on vegetation. New Phytol. 205, 1008–1014 (2015).

    PubMed  Google Scholar 

  22. Konings, A. G. & Gentine, P. Global variations in ecosystem-scale isohydricity. Glob. Change Biol. 23, 891–905 (2017).

    ADS  Google Scholar 

  23. Martin-StPaul, N., Delzon, S. & Cochard, H. Plant resistance to drought depends on timely stomatal closure. Ecol. Lett. 20, 1437–1447 (2017).

    PubMed  Google Scholar 

  24. Suding, K. N. et al. Scaling environmental change through the community-level: a trait-based response-and-effect framework for plants. Glob. Change Biol. 14, 1125–1140 (2008).

    ADS  Google Scholar 

  25. Zhang, T., Niinemets, Ü., Sheffield, J. & Lichstein, J. W. Shifts in tree functional composition amplify the response of forest biomass to climate. Nature 556, 99–102 (2018).

    ADS  CAS  PubMed  Google Scholar 

  26. Wolf, A., Anderegg, W. R. L. & Pacala, S. W. Optimal stomatal behavior with competition for water and risk of hydraulic impairment. Proc. Natl Acad. Sci. USA 113, E7222–E7230 (2016).

    CAS  PubMed  Google Scholar 

  27. Bonan, G. B., Williams, M., Fisher, R. A. & Oleson, K. W. Modeling stomatal conductance in the earth system: linking leaf water-use efficiency and water transport along the soil–plant–atmosphere continuum. Geosci. Model Dev. 7, 2193–2222 (2014).

    ADS  Google Scholar 

  28. Trugman, A., Medvigy, D., Mankin, J. & Anderegg, W. R. L. Soil moisture stress as a major driver of carbon cycle uncertainty. Geophys. Res. Lett. 45, 6495–6503 (2018).

    ADS  Google Scholar 

  29. Nardini, A., Battistuzzo, M. & Savi, T. Shoot desiccation and hydraulic failure in temperate woody angiosperms during an extreme summer drought. New Phytol. 200, 322–329 (2013).

    CAS  PubMed  Google Scholar 

  30. Anderegg, W. R. L. et al. Meta-analysis reveals that hydraulic traits explain cross-species patterns of drought-induced tree mortality across the globe. Proc. Natl Acad. Sci. USA 113, 5024–5029 (2016).

    ADS  CAS  PubMed  Google Scholar 

  31. Pastorello, G. Z. et al. The FLUXNET2015 dataset: The longest record of global carbon, water, and energy fluxes is updated. Eos 98, (2017).

  32. Clogg, C. C., Petkova, E. & Haritou, A. Statistical methods for comparing regression coefficients between models. Am. J. Sociol. 100, 1261–1293 (1995).

    Google Scholar 

  33. Zanne, A. E. et al. Global wood density database. Dryad (2009).

  34. Chave, J. et al. Towards a worldwide wood economics spectrum. Ecol. Lett. 12, 351–366 (2009).

    PubMed  Google Scholar 

  35. Maire, V. et al. Global effects of soil and climate on leaf photosynthetic traits and rates. Glob. Ecol. Biogeogr. 24, 706–717 (2015).

    Google Scholar 

  36. Gleason, S. M. et al. Weak tradeoff between xylem safety and xylem-specific hydraulic efficiency across the world’s woody plant species. New Phytol. 209, 123–136 (2016).

    CAS  PubMed  Google Scholar 

  37. Anderegg, L. D. L., Anderegg, W. R. L., Abatzoglou, J., Hausladen, A. M. & Berry, J. A. Drought characteristics’ role in widespread aspen forest mortality across Colorado, USA. Glob. Change Biol. 19, 1526–1537 (2013).

    ADS  Google Scholar 

  38. Burnham, K. P. & Anderson, D. R. Multimodel inference understanding AIC and BIC in model selection. Sociol. Methods Res. 33, 261–304 (2004).

    MathSciNet  Google Scholar 

  39. Breiman, L. Random forests. Mach. Learn. 45, 5–32 (2001).

    MATH  Google Scholar 

  40. Jackson, T. J. & Schmugge, T. J. Vegetation effects on the microwave emission of soils. Remote Sens. Environ. 36, 203–212 (1991).

    ADS  Google Scholar 

  41. Momen, M. et al. Interacting effects of leaf water potential and biomass on vegetation optical depth. J. Geophys. Res. Biogeosci. 122, 3031–3046 (2017).

    Google Scholar 

  42. Du, J., Kimball, J. S. & Jones, L. A. Passive microwave remote sensing of soil moisture based on dynamic vegetation scattering properties for AMSR-E. IEEE Trans. Geosci. Remote Sens. 54, 597–608 (2016).

    ADS  Google Scholar 

  43. Du, J. et al. A global satellite environmental data record derived from AMSR-E and AMSR2 microwave earth observations. Earth Syst. Sci. Data 9, 791–808 (2017).

    ADS  Google Scholar 

  44. Jones, L. A. et al. Satellite microwave remote sensing of daily land surface air temperature minima and maxima from AMSR-E. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 3, 111–123 (2010).

    ADS  Google Scholar 

  45. Huffman, G. J. et al. Global precipitation at one-degree daily resolution from multisatellite observations. J. Hydrometeorol. 2, 36–50 (2001).

    ADS  Google Scholar 

  46. Konings, A. G., Williams, A. P. & Gentine, P. Sensitivity of grassland productivity to aridity controlled by stomatal and xylem regulation. Nat. Geosci. 10, 284–288 (2017).

    ADS  CAS  Google Scholar 

  47. Jenkins, C. N., Van Houtan, K. S., Pimm, S. L. & Sexton, J. O. US protected lands mismatch biodiversity priorities. Proc. Natl Acad. Sci. USA 112, 5081–5086 (2015).

    ADS  CAS  PubMed  Google Scholar 

  48. Ellis, E. C., Antill, E. C. & Kreft, H. All is not loss: plant biodiversity in the Anthropocene. PLoS ONE 7, e30535 (2012).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  49. Olson, D. M. et al. Terrestrial ecoregions of the world: a new map of life on earth: a new global map of terrestrial ecoregions provides an innovative tool for conserving biodiversity. Bioscience 51, 933–938 (2001).

    Google Scholar 

  50. Sanderson, E. W. et al. The human footprint and the last of the wild. Bioscience 52, 891–904 (2002).

    Google Scholar 

  51. Dormann, C. F. et al. Methods to account for spatial autocorrelation in the analysis of species distributional data: a review. Ecography 30, 609–628 (2007).

    Google Scholar 

  52. Simard, M., Pinto, N., Fisher, J. B. & Baccini, A. Mapping forest canopy height globally with spaceborne lidar. J. Geophys. Res. 116, G04021 (2011).

    ADS  Google Scholar 

  53. R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, Vienna, 2012).

  54. Ripley, B. et al. MASS: Support Functions and Datasets for Venables and Ripley’s MASS. R package version 7.3-50 (2013).

  55. Hijmans, R. J. & van Etten, J. raster: Geographic Data Analysis and Modeling. R package version 2.6-7 (2014).

  56. South, A. rworldmap: Mapping Global Data. R package version 1.3-6 (2011).

  57. Bates, D. et al. lme4: Linear Mixed-Effects Models using ‘Eigen’ and S4. R package version 1.1-18-1 (2014).

  58. Wood, S. N. mgcv: GAMs and generalized ridge regression for R. R News 1 (2), 20–25 (2001).

    Google Scholar 

Download references


We thank M. Beninati, R. O’Dell and J. Gallafent for assistance with trait compilation. W.R.L.A. acknowledges funding from the University of Utah Global Change and Sustainability Center, NSF Grant 1714972 and 1802880 and the USDA National Institute of Food and Agriculture, Agricultural and Food Research Initiative Competitive Programme, Ecosystem Services and Agro-ecosystem Management, grant no. 2018-67019-27850. A.T.T. acknowledges funding from USDA National Institute of Food and Agriculture Postdoctoral Research Fellowship grant no. 2017-07164. This work used eddy covariance data acquired and shared by the FLUXNET community, including these networks: AmeriFlux, AfriFlux, AsiaFlux, CarboAfrica, CarboEuropeIP, CarboItaly, CarboMont, ChinaFlux, Fluxnet-Canada, GreenGrass, ICOS, KoFlux, LBA, NECC, OzFlux-TERN, TCOS-Siberia and USCCC. The ERA-Interim reanalysis data are provided by ECMWF and processed by LSCE. The FLUXNET eddy covariance data processing and harmonization was carried out by the European Fluxes Database Cluster, AmeriFlux Management Project and Fluxdata project of FLUXNET, with the support of CDIAC and ICOS Ecosystem Thematic Center and the OzFlux, ChinaFlux and AsiaFlux offices.

Reviewer information

Nature thanks D. Baldocchi, S. Delzon, S. Jansen and the other anonymous reviewer(s) for their contribution to the peer review of this work.

Author information

Authors and Affiliations



W.R.L.A. designed the study with all authors providing input. A.G.K., K.Y., R.G. and N.Z. contributed data and assisted with data collection. W.R.L.A., A.G.K. and A.T.T. analysed the data. W.R.L.A. wrote the paper with all authors providing input.

Corresponding author

Correspondence to William R. L. Anderegg.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended Data Fig. 1 Map of the included eddy covariance flux sites overlaid on species richness.

Species richness is shown by different colours (data from Size of the circle is representative of the sample size of included days ranging from 10 (smallest circles) to 1,057 (largest circles) samples.

Extended Data Fig. 2 Variable importance analysis for traits at flux sites.

a, b, Variable importance (total decrease in node impurities) results from the machine-learning algorithm, random forests, for each variable for drought sensitivity (a) and drought coupling (b) metrics. Traits include SLA, (Amax), P50 and HSM. The suffix ‘m’ indicates the community-weighted mean; the suffix ‘SD’ indicates the community-weighted standard deviation. See Supplementary Table 3 for sample sizes.

Extended Data Fig. 3 Increased hydraulic variation buffers ecosystem drought responses.

a, c, Drought coupling is expressed as the percentage of explained variation (R2). b, d, Drought sensitivity is shown as the summed absolute values of standardized coefficients for drought variables that are regressed against latent energy (LE) exchange. Regression: LE ≈ VPD + SM + VPD × SM. a, b, Panels are identical to Fig. 2 but with site identifications shown. Hydraulic variation is expressed as the community-weighted standard deviation in the hydraulic safety margin of species. c, d, The hydraulic safety margin was calculated from the 50% loss of hydraulic conductivity in gymnosperms and 88% loss of hydraulic conductivity in angiosperms. Colours indicate biomes of deciduous broadleaf (green) and needleleaf (red) forests. The size of the dot indicates the number of days included for each flux site. The solid black line is the best fit of the ordinary least-squares linear regression (c, P = 0.008; d, P = 0.01) and dashed lines are the 95% confidence interval of the regression fit.

Extended Data Fig. 4 Satellite comparisons to flux towers.

ac, Relationship between daily total latent energy (LE) exchange measured via eddy covariance and midday canopy water content from remote-sensing of VOD for global forest sites (a; n = 4,525 grid cells), broadleaf forest sites (b; n = 1,915 grid cells) and evergreen forest sites (c; n = 2,610 grid cells). Red lines indicate best fits for ordinary least-squares regressions. Note that the canopy water content at each pixel integrates a spatial area that is two orders of magnitude greater than the eddy covariance sites.

Extended Data Fig. 5 Hydraulic trait variation compared to species richness.

Comparison of dominant tree species richness from gridded data of US forests against the hydraulic diversity—the standard deviation in HSM—at six eddy covariance sites in the United States that have adequate trait data. The red line indicates the best fit of the ordinary least-squares linear regression (n = 6 sites; R2 = 0.65; P = 0.04).

Extended Data Fig. 6 Higher species diversity is associated with more buffered drought responses in US forests.

a, b, Drought coupling (a; n = 163 grid cells) and drought sensitivity (b; n = 163 grid cells) from variation in remotely sensed canopy vegetation water content compared to tree species richness in the United States. Black line is the best fit of the ordinary least-squares linear regressions.

Extended Data Fig. 7 Higher species diversity is associated with more-buffered drought responses in forests globally.

a, Drought sensitivity as the slope (β) in an ordinary least-squares linear regression of an index of the variation in aboveground plant water content at midday compared to night (regression: VODmidday = β × VODnight + ε). b, Native plant species richness (percentage of maximum; data from ch, Ordinary least-squares linear regressions between these two variables for six major biomes. c, Tropical and subtropical moist broadleaf forests (n = 1,380 grid cells). d, Tropical and subtropical dry broadleaf forests (n = 241 grid cells). e, Temperate broadleaf and mixed forests (n = 1,289 grid cells). f, Temperate coniferous forests (n = 318 grid cells). g, Boreal forests (n = 1,784 grid cells). h, Mediterranean-type forests, woodlands and shrub (n = 319 grid cells). Each point represents an individual grid cell from the map and colours that are more red indicate a higher density of points. Red lines show the best fit of ordinary least-squares linear regression lines.

Extended Data Fig. 8 Analyses of the importance of variables using satellite data.

The importance of the variables (total decrease in node impurities) results obtained using the machine learning algorithm, random forests, for each variable of the drought coupling metric (n = 6,698 grid cells). ‘CanHeight’, lidar-derived canopy height.

Extended Data Fig. 9 Ecosystem sensitivity to drought saturates with species richness (the percentage of maximum) across forests globally.

a, Drought coupling is expressed as the explained variation (R2) of midday aboveground plant water content in forest ecosystems regressed against drought variables using ordinary least-squares linear regression. b, Drought sensitivity is expressed as the regression coefficient of midday aboveground plant water content regressed against a metric of soil water stress using ordinary least-squares linear regression. n = 6,698 grid cells. The black line shows the best fit generalized additive model and dashed lines are the 99% confidence interval.

Extended Data Fig. 10 Performance of the multivariate drought regression model.

Nine randomly selected sites of observed latent energy (LE) fluxes versus predicted fluxes from the multiple regression model based on VPD, soil moisture and their interaction. Red lines are the ordinary least-squares best-fit regression line. Sites are as follows (abbreviations can be found in Supplementary Table 1). a, US-UMB (R2 = 0.12, P < 0.0001). b, US-WCr (R2 = 0.13, P < 0.0001). c, DK-Sor (R2 = 0.11, P < 0.0001). d, IT-Ca1 (R2 = 0.44, P < 0.0001). e, IT-PT1 (R2 = 0.06, P = 0.002). f, IT-Ro1 (R2 = 0.10, P < 0.0001). g, JP-SMF (R2 = 0.40, P < 0.0001). h, NL-Loo (R2 = 0.25, P < 0.0001). i, US-NWR (R2 = 0.15, P < 0.0001). k, Performance of the drought multivariate ordinary least-squares linear regression model (LE = f(VPD, SM, VPD × SM) across eddy covariance sites shown in a histogram of site-level model P values from those regressions.

Supplementary information

Supplementary Information

This file contains five supplementary tables that contain details about the flux sites used in the study, a brief supplementary discussion, and supplementary notes and references for the flux sites.

Reporting Summary

Rights and permissions

Reprints and Permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Anderegg, W.R.L., Konings, A.G., Trugman, A.T. et al. Hydraulic diversity of forests regulates ecosystem resilience during drought. Nature 561, 538–541 (2018).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing