2015/16 El Niño increased water demand and pushed plants from a Mesic tropical montane grassland beyond their hydraulic safety limits
Data files
May 01, 2023 version files 139.14 KB
-
Data_Austral_Ecology_Matos_et_al_2023.xlsx
-
README.md
Abstract
In 2015/16, a strong El Niño event caused anomalously high temperatures and reduced precipitation resulting in Pantropical drought‐induced diebacks and wildfires. Although many studies have documented the El Niño impacts on tropical forests, little we know about its effects on tropical grasslands. Here, we investigated plant drought responses during and after the 2015/16 El Niño event (Jun 2016 to Aug 2017) in 12 species with contrasting drought strategies (tolerance, avoidance and escape) in a Brazilian tropical montane grassland. We tested if (1) the El Niño event induced meteorological drought anomalies, (2) the atmospheric and/or soil drought led to plant water stress and (3) plants showed signs of drought recovery. In contrast to other tropical areas, we found that the 2015/16 El Niño event did not strongly affect precipitation in our study site. However, it increased air temperature and vapour pressure deficit, thus pushing all grassland species, even the most drought‐tolerant ones, beyond their hydraulic safety margins during the dry season. Most species showed signs of drought recovery, returning to positive hydraulic margins in the wet season after the El Niño. However, the finding that all evaluated species, regardless of their drought‐response strategy, are already operating close to their hydraulic safe thresholds for stomatal closure and turgor loss suggests that this cool–humid tropical montane grassland is especially vulnerable to meteorological extremes exacerbated by the additive effects of El Niño and climate change.
Methods
We calculated monthly meteorological anomalies of air temperature (Tair), VPD and precipitation to evaluate whether the EN15/16 event corresponded to drought anomalies in the study area. First, we obtained the 3‐metre air (‘t2m’) and dew point temperature (‘d2m’) meteorological fields from the ERA5‐Land reanalysis product, the current benchmark dataset for analysis of recent climate (https://cds.climate.copernicus.eu/; Copernicus Climate Change Service). Hourly monthly mean of ERA5‐Land data over Itatiaia covering the 1981–2020 period was extracted using Google Earth Engine (https://earthengine.google.com/). After extraction, we selected only the 1 PM time period for our analysis because plants are typically subjected to the driest and hottest conditions shortly after midday (Bhaskar & Ackerly, 2006; Martínez‐Vilalta et al., 2021), and that was also the time when our plant physiological measurements were conducted (see below). Next, we used the fields ‘t2m’ and ‘d2m’ to calculate saturation vapour pressure and the actual vapour pressure, which were then used to calculate VPD using the plantecophys::DewtoVPD R‐package (Duursma, 2015). Finally, we calculated 3‐month centred Tair and VPD monthly mean anomalies in units of standard deviation (σ) for the reference period of 1981–2010.
We followed a similar approach to calculate precipitation monthly anomalies, but because precipitation estimates from reanalysis products are known to possess significant biases over the tropics (Bador et al., 2020; Trenberth et al., 2001), instead of ERA5, we calculated precipitation anomalies using multiple gauge‐based and gauge‐calibrated satellite data products (CHIRPS, GPM‐IMERG, GPCP and GPCC from the FROGS ensemble; Roca et al., 2019) gridded to a common 1∘ × 1∘ daily‐resolution format. We also investigated whether previous strong El Niño events, such as the 1982/83 and 1997/98 El Niños, produced drought anomalies. To do so, we correlate a centred moving 12‐month average of precipitation, Tair and VPD anomalies with the El Niño 3.4 index – EN3.4 (https://esrl.noaa.gov/psd/gcos_wgsp/Timeseries/Data/Niño34.long.anom.data), which indicates average sea surface temperature anomalies. This allowed us to place the severity of the EN15/16 event into the historical context of the study area. If El Niño produces meteorological drought conditions in the study area, we expect EN3.4 anomalies to strongly correlate with positive Tair and VPD anomalies, and with negative precipitation anomalies.
AQ2In addition to the satellite data products, we also monitored meteorological conditions on the ground. Air temperature and humidity were obtained from the Itatiaia weather station located ca.15 km away from the study area (−22.370, −44.700 and >2400 m asl). These data were then used to calculate VPD and the occurrence of fog events (following Rosado et al. 2010). Daily precipitation was determined with a pluviometer located ca. 300 m away from the study area (−22.374 to 44.702 and >2400 m asl), whereas soil temperature and volumetric water content were recorded locally at 20 cm soil depth using two soil moisture sensors (model ECH2O 10HS, Decagon Devices, Pullman, WA, USA USA). Those local meteorological conditions were obtained from Mar 2015 to Aug 2017. Thus, covering the whole EN15/16 event (Jun 2015 to Aug 2016) plus the following year, which was under moderate La Niña influence. As reliable long‐term climatic records are non‐existent for the study area, we could not calculate drought anomalies for those on‐the‐ground meteorological data.
We selected 12 co‐occurring C3 plant species to investigate the potential effects of the EN15/16 event on the tropical montane grassland vegetation. Those species differ in their habit (grasses, shrubs and forbs) and dominances (dominant, rare and subordinate), and are classified into three contrasting eco‐physiological strategies (Matos et al., 2021): S‐tolerant/avoiders (Achyrocline satureioides (Lam.) DC., Baccharis uncinella DC., Chionolaena capitata (Baker) Freire, Chusquea pinifolia (Nees) Nees, Gamochaeta purpurea (L.) Cabrera and Pleroma hospita (Schrank et Mart. ex DC.) Triana), CR‐escape/avoiders (Hypochaeris lutea (Vell.) Britton, Leptostelma maximum D. Don and Mikania glaziovii Baker) and CS‐escape/tolerant (Cortaderia modesta (Döll.) Hack, Eryngium glaziovianum Urb. and Machaerina ensifolia (Boeckeler) T. Koyama). Those strategies unify Grime's CSR (competitor: species that maximize resource acquisition in productive niches; stress‐tolerant: species that maintain metabolic performance in stressful niches; and ruderal: species that favours rapid propagation or regeneration in frequently disturbed niches) triangle ecological scheme (Grime, 1977; Pierce et al., 2017) and Levitt's physiological classification (avoiders, escapers and tolerant; Levitt, 1972) to describe plant responses to drought (Matos et al., 2021). Specifically, S‐tolerant/avoiders include species with high stress tolerance in the CSR triangle and with a combination of both drought tolerance and avoidance traits. CR‐escape/avoiders include competitor and ruderal species that can either avoid or escape drought. CS‐escape/tolerant includes species with intermediate trait values for tolerance and escape and with an intermediate position along the competitor–stress‐tolerant axis in the CSR triangle (Matos et al., 2021).
For each species, we conducted monthly measurements, from Jun 2016 to Aug 2017, of midday (Ψmd measured from 1 to 2 PM) and predawn leaf water potentials (Ψpd measured from 4 to 6 AM) on 2–3 mature and sun‐exposed leaves of 2–3 individuals per species. Water potential measurements were conducted on similarly clear and sunny days using a pressure chamber (model 3005; Soil Moisture Equipment, Santa Barbara, CA, USA). Then, the minimum midday leaf water potential (Ψmin) was determined for each species as the minimum midday leaf water potential (Ψmd) recorded during the whole study period. Therefore, Ψmin represents the maximum xylem tension or the minimum water potential that a species is subject to under natural conditions, with more negative values suggesting higher drought tolerance (Bhaskar & Ackerly, 2006; Martínez‐Vilalta et al., 2021).
On the same days of water potential measurements, we also obtained values of midday stomatal conductance (gsmd, mmol m−2 s−1). gsmd was measured with a steady‐state leaf porometer (model SC1, Decagon Devices, Pullman, WA, USA) on the same individuals used for Ψmd measurements. Then, both Ψmd and gsmd were used for the calculation of the leaf water potential thresholds inducing 50% (Ψgs50, MPa) and 95% (Ψgs95, MPa) of stomatal closure. As the stomatal sensitivity to leaf water potential varies across species (Klein, 2014), we regressed the seasonal time series of gsmd against Ψmd using linear, exponential, sigmoidal and logistic functions. Next, we selected the function with the lowest Akaike information criterion (following Scoffoni et al., 2012) to calculate Ψgs50 and Ψgs95 (see Appendix S1). Those two traits reflect the stringency of the plant's stomatal control (Klein, 2014), with less negative Ψgs50 and Ψgs95 values indicating a tighter stomatal control and more negative values suggesting a looser stomatal regulation.
We also sampled five mature leaves per species during the 2016 dry season to construct pressure–volume (PV) curves using the bench‐drying method (Turner, 1988). PV curves were constructed by plotting the inverse of leaf water potential (−1/Ψleaf) against the leaf relative water content. Then, the leaf water potential at turgor loss point (Ψtlp in MPa) was determined as the point of transition between the linear and non‐linear portions of the curve. Ψtlp indicates the leaf water potential at which the average leaf cell turgor pressure is lost (i.e. wilting point), with more negative Ψtlp values reflecting higher drought tolerance (Bartlett et al., 2014; Zhu et al., 2018). As some species are known to adjust their Ψtlp under water stress to improve drought tolerance (Bartlett et al., 2014), our PV curves constructed during the peak of the dry season (Jul–Aug 2016) probably report the most negative Ψtlp value that the evaluated species manifest under natural conditions.
We calculated monthly HSMs from Jun 2016 to Aug 2017 to identify which (if any) of the 12 studied species experienced physiological water stress during the EN15/16 event and whether such species exhibited signs of drought recovery after El Niño. HSMs were calculated for turgor loss (HSMtlp = Ψmin – Ψtlp, where Ψmin is the minimum Ψmd at each month), for partial (HSMgs50 = Ψmin – Ψgs50) and almost complete stomatal closure (HSMgs95 = Ψmin – Ψgs95). As plants often experience hotter and drier conditions shortly after midday, those HSMs calculated based on Ψmd likely represent the narrower HSMs for each species throughout the seasons. To evaluate whether plants were able to reopen their stomata and/or regain turgor after this midday depression, we also calculated HSMs using the predawn leaf water potentials, that is, using Ψmin as the minimum Ψpd for each month. Positive HSMs values indicate that plants are operating at a safer range of water potentials, while negative HSMs indicate that plants have crossed their hydraulic safety thresholds and are likely under water stress. Note that, rather than lethal thresholds of hydraulic failure and plant death, those HSMs calculated based on stomatal closure and turgor loss are primarily useful to describe the short‐term plants' responses to drought.
Usage notes
Excel
R