Intensification of the global hydrological cycle, ranging from larger individual precipitation events to more extreme multi-year droughts, has the potential to cause widespread alterations in ecosystem structure and function. With evidence that the incidence of extreme precipitation years (defined statistically from historical precipitation records) is increasing, there is a clear need to identify ecosystems that are most vulnerable to these changes and understand why some ecosystems are more sensitive to extremes than others. To date, opportunistic studies of naturally occurring extreme precipitation years, combined with results from a relatively small number of experiments, have provided limited mechanistic understanding of differences in ecosystem sensitivity suggesting that new approaches are needed. Coordinated distributed experiments (CDEs) arrayed across multiple ecosystem types and focused on water can enhance our understanding of differential ecosystem sensitivity to precipitation extremes, but there are many design challenges to overcome (e.g., cost, comparability, standardization). Here we evaluate contemporary experimental approaches for manipulating precipitation under field conditions to inform the design of "Drought-Net", a relatively low cost CDE that simulates extreme precipitation years. A common method for imposing both dry and wet years is to alter each ambient precipitation event. We endorse this approach for imposing extreme precipitation years because it simultaneously alters other precipitation characteristics (i.e., event size) consistent with natural precipitation patterns. However, we do not advocate applying identical treatment levels at all sites -a common approach to standardization in CDEs. This is because precipitation variability varies >5-fold globally resulting in a wide range of ecosystemspecific thresholds for defining extreme precipitation years. For CDEs focused on Accepted ArticleThis article is protected by copyright. All rights reserved. precipitation extremes, treatments should be based on each site's past climatic characteristics. This approach, though not often used by ecologists, allows ecological responses to be directly compared across disparate ecosystems and climates, facilitating process-level understanding of ecosystem sensitivity to precipitation extremes.

Plant traits can provide unique insights into plant performance at the community scale. Functional composition, defined by both functional diversity and community‐weighted trait means (CWMs), can affect the stability of above‐ground net primary production (ANPP) in response to climate extremes. Further complexity arises, however, when functional composition itself responds to environmental change. The duration of climate extremes, such as drought, is expected to increase with rising global temperatures; thus, understanding the impacts of long‐term drought on functional composition and the corresponding effect that has on ecosystem function could improve predictions of ecosystem sensitivity to climate change. We experimentally reduced growing season precipitation by 66% across six temperate grasslands for 4 years and measured changes in three indices of functional diversity (functional dispersion, richness and evenness), community‐weighted trait means and phylogenetic diversity (PD). Specific leaf area (SLA), leaf nitrogen content (LNC) and (at most sites) leaf turgor loss point (πTLP) were measured for species cumulatively representing ~90% plant cover at each site. Long‐term drought led to increased community functional dispersion in three sites, with negligible effects on the remaining sites. Species re‐ordering following the mortality/senescence of dominant species was the main driver of increased functional dispersion. The response of functional diversity was not consistently matched by changes in phylogenetic diversity. Community‐level drought strategies (assessed as CWMs) largely shifted from drought tolerance to drought avoidance and/or escape strategies, as evidenced by higher community‐weighted πTLP, SLA and LNC. Lastly, ecosystem drought sensitivity (i.e. relative reduction in ANPP in drought plots) was positively correlated with community‐weighted SLA and negatively correlated with functional diversity. Synthesis. Increased functional diversity following long‐term drought may stabilize ecosystem functioning in response to future drought. However, shifts in community‐scale drought strategies may increase ecosystem drought sensitivity, depending on the nature and timing of drought. Thus, our results highlight the importance of considering both functional diversity and abundance‐weighted traits means of plant communities as their collective effect may either stabilize or enhance ecosystem sensitivity to drought.

Plant traits can be used to predict ecosystem responses to environmental change using a response–effect trait framework. To do this, appropriate traits must be identified that explain a species' influence on ecosystem function (“effect traits”) and the response of those species to environmental change (“response traits”). Response traits are often identified and measured along gradients in plant resources, such as water availability; however, precipitation explains very little variation in most plant traits globally. Given the strong relationship between plant traits and ecosystem functions, such as net primary productivity (NPP), and between NPP and precipitation, the lack of correlation between precipitation and plant traits is surprising. We address this issue through a systematic review of >500 published studies that describe plant trait responses to altered water availability. The overarching goal of this review was to identify potential causes for the weak relationship between commonly measured plant traits and water availability so that we may identify more appropriate “response traits.” We attribute weak trait–precipitation relationships to an improper selection of traits (e.g., nonhydraulic traits) and a lack of trait‐based approaches that adjust for trait variation within communities (only 4% of studies measure community‐weighted traits). We then highlight the mechanistic value of hydraulic traits as more appropriate “response traits” with regard to precipitation, which should be included in future community‐scale trait surveys. Trait‐based ecology has the potential to improve predictions of ecosystem responses to predicted changes in precipitation; however, this predictive power depends heavily on the identification of reliable response and effect traits. To this end, trait surveys could be improved by a selection of traits that reflect physiological functions directly related to water availability with traits weighted by species relative abundance. A http://onlinelibrary.wiley.com/doi/10.1111/1365-2435.13135/suppinfo is available for this article.

During the 1930s Dust Bowl drought in the central United States, species with the C3 photosynthetic pathway expanded throughout C4-dominated grasslands. This widespread increase in C3 grasses during a decade of low rainfall and high temperatures is inconsistent with well-known traits of C3 vs. C4 pathways. Indeed, water use efficiency is generally lower, and photosynthesis is more sensitive to high temperatures in C3 than C4 species, consistent with the predominant distribution of C3 grasslands in cooler environments and at higher latitudes globally. We experimentally imposed extreme drought for 4 y in mixed C3/C4 grasslands in Kansas and Wyoming and, similar to Dust Bowl observations, also documented three- to fivefold increases in C3/C4 biomass ratios. To explain these paradoxical responses, we first analyzed long-term climate records to show that under nominal conditions in the central United States, C4 grasses dominate where precipitation and air temperature are strongly related (warmest months are wettest months). In contrast, C3 grasses flourish where precipitation inputs are less strongly coupled to warm temperatures. We then show that during extreme drought years, precipitation–temperature relationships weaken, and the proportion of precipitation falling during cooler months increases. This shift in precipitation seasonality provides a mechanism for C3 grasses to respond positively to multiyear drought, resolving the Dust Bowl paradox. Grasslands are globally important biomes and increasingly vulnerable to direct effects of climate extremes. Our findings highlight how extreme drought can indirectly alter precipitation seasonality and shift ecosystem phenology, affecting function in ways not predictable from key traits of C3 and C4 species.

No abstract

Extreme drought decreases aboveground net primary production (ANPP) in most grasslands, but the magnitude of ANPP reductions varies especially in C3‐dominated grasslands. Because the mechanisms underlying such differential ecosystem responses to drought are not well resolved, we experimentally imposed an extreme 4‐yr drought (2015–2018) in two C3 grasslands that differed in aridity. These sites had similar annual precipitation and dominant grass species (Leymus chinensis) but different annual temperatures and thus water availability. Drought treatments differentially affected these two semiarid grasslands, with ANPP of the drier site reduced more than at the wetter site. Structural equation modeling revealed that community‐weighted means for some traits modified relationships between soil moisture and ANPP, often due to intraspecific variation. Specifically, drought reduced community mean plant height at both sites, resulting in a reduction in ANPP beyond that attributable to reduced soil moisture alone. Higher community mean leaf carbon content enhanced the negative effects of drought on ANPP at the drier site, and ANPP–soil‐moisture relationships were influenced by soil C:N ratio at the wetter site. Importantly, neither species richness nor functional dispersion were significantly correlated with ANPP at either site. Overall, as expected, soil moisture was a dominant, direct driver of ANPP response to drought, but differential sensitivity to drought in these two grasslands was also related to soil fertility and plant traits.

Community-scale surveys of plant drought tolerance are essential for understanding semi-arid ecosystems and community responses to climate change. Thus, there is a need for an accurate and rapid methodology for assessing drought tolerance strategies across plant functional types. The osmometer method for predicting leaf osmotic potential at full turgor (πo), a key metric of leaf-level drought tolerance, has resulted in a 50-fold increase in the measurement speed of this trait; however, the applicability of this method has only been tested in woody species and crops. Here, we assess the osmometer method for use in herbaceous grassland species and test whether πo is an appropriate plant trait for understanding drought strategies of herbaceous species as well as species distributions along climate gradients. Our model for predicting leaf turgor loss point (πTLP) from πo (πΤLΡ = 0.80πo-0.845) is nearly identical to the model previously presented for woody species. Additionally, πo was highly correlated with π-^ for graminoid species (πtlp = 0.944πo-0.611; r2 = 0.96), a plant functional group previously flagged for having the potential to cause erroneous measurements when using an osmometer. We report that πo, measured with an osmometer, is well correlated with other traits linked to drought tolerance (namely, leaf dry matter content and leaf vulnerability to hydraulic failure) as well as climate extremes linked to water availability. The validation of the osmometer method in an herb-dominated ecosystem paves the way for rapid community-scale surveys of drought tolerance across plant functional groups, which could improve trait-based predictions of ecosystem responses to climate change.

Experiments are widely used in ecology, particularly for assessing global change impacts on ecosystem function. However, results from experiments often are inconsistent with observations made under natural conditions, suggesting the need for rigorous comparisons of experimental and observational studies. We conducted such a "reality check" for a grassland ecosystem by compiling results from nine independently conducted climate change experiments. Each experiment manipulated growing season precipitation (GSP) and measured responses in aboveground net primary production (ANPP). We compared results from experiments with long-term (33-yr) annual precipitation and ANPP records to ask if collectively (n = 44 experiment-years) experiments yielded estimates of ANPP, rain-use efficiency (RUE, grams per square meter ANPP per mm precipitation), and the relationship between GSP and ANPP comparable to observations. We found that mean ANPP and RUE from experiments did not deviate from observations. Experiments and observational data also yielded similar functional relationships between ANPP and GSP, but only within the range of historically observed GSP. Fewer experiments imposed extreme levels of GSP (outside the observed 33-yr record), but when these were included, they altered the GSP-ANPP relationship. This result underscores the need for more experiments imposing extreme precipitation levels to resolve how forecast changes in climate regimes will affect ecosystem function in the future.