How do plant trait strategies underpin ecosystem stability in temperate and alpine grasslands?

One trait, two opposite stability outcomes

The paper asks whether the link between plant functional traits and ecosystem stability is the same everywhere. It is not. Across 160 plots in 20 sites on the Mongolian Plateau and the Tibetan Plateau, acquisitive strategies went with greater temporal stability and resistance in temperate grasslands but with lower temporal stability and resistance in alpine grasslands. In the alpine system, slow-growing perennial species with conservative traits were the ones associated with stability.

The split matters because it contradicts the common expectation that only conservative, stress-tolerant plants confer resistance and temporal stability. The authors frame the result as distinct routes to the same endpoint: different starting conditions, different stabilizing strategy.

Three stability measures are used here and should not be conflated. Temporal stability is how little a system varies over time; resistance is its capacity to withstand a perturbation; resilience is its ability to return to its prior state afterwards. The trait contrast appears in the first two, not in the third.

Linking field-measured traits to satellite stability

Traits were measured in situ during July and August 2018 for 343 species, covering 15 morphological and chemical traits. These include size-related measures such as plant height and leaf area, chemistry such as leaf and root nitrogen and phosphorus, and traits tied to the economics of resource capture and retention: specific leaf area (SLA), and leaf, stem and root dry matter content (LDMC, SDMC, RDMC). High SLA and high leaf nitrogen sit on the acquisitive side; high dry matter content sits on the conservative side.

Stability was not measured in the field. It was derived from satellite vegetation indices combined with 23 years (2000–2022) of Standardized Precipitation–Evapotranspiration Index data, which identifies dry, normal and wet growing-season conditions using a ±0.67 threshold. Resistance and resilience were computed from how vegetation responded around those hydroclimatic extremes.

The two grassland types differ in climate but not in everything: growing-season temperature was significantly higher in temperate grasslands, while the types did not differ in precipitation, soil properties or species richness. The proportion of annual plants did differ sharply, at 27% versus 9% of species richness and 32% versus 7% of aboveground biomass in temperate versus alpine grasslands.

Annuals carry the acquisitive signal in temperate grasslands

The mechanism the authors propose runs through life history. Temperate grasslands, with longer and warmer growing seasons, hold more annual plants, and the data show annuals tend to have faster trait syndromes such as large SLA. Because annuals complete their life cycles quickly, they can escape prolonged stress rather than endure it, and the paper also reports that temperate annuals tended to have smaller stomata, a trait linked to faster stomatal responses to stress.

This is the paper's interpretation of why acquisitive traits could stabilize a system, not a directly manipulated test. The authors describe it as stress escape and stress avoidance, and note that annuals contributed substantially to community-level acquisitive economics, particularly in temperate grasslands.

Alpine grasslands run the other way. With cold, short growing seasons and few annuals, the community's economics are dominated by slow-growing perennials, and there conservative traits were the ones associated with higher temporal stability and resistance.

Resilience moves with acquisitive traits in both biomes

The biome dependence is not uniform across all three stability dimensions. Resilience showed consistent positive responses to acquisitive traits in both grassland types, unlike temporal stability and resistance, which flipped sign between biomes.

That divergence is a substantive result rather than a technicality: resistance and resilience describe different things, and the paper treats them as separate response variables rather than interchangeable proxies. A community that holds its ground under a dry spell and one that bounces back afterwards need not be the same community.

The authors also caution that their supplementary site-level analyses treat resistance and resilience as distinct components, and they present the resilience pattern as something that may hold more broadly but should be tested across a wider range of ecosystems.

The pattern survives controls for diversity, climate and soil

A natural objection is that the trait signal is just a stand-in for diversity, climate or soil. The paper reports that the biome-dependent patterns persisted after accounting for plant diversity, climate and soil properties. In the authors' comparison, the trait effects were comparable in magnitude to abiotic variables and consistently exceeded the influence of plant diversity, which is often emphasized as a key stability driver.

The site-level structural equation models used LDMC as a parsimonious proxy for the conservative–acquisitive axis, growing-season temperature as the focal climatic driver, and soil pH and species composition as covariates. Biome dependence was tested as a trait-by-temperature interaction rather than by fitting separate models per grassland type, because grassland type aligned closely with the temperature gradient and including both would have been largely redundant at the available site-level sample size.

Soil pH behaved in an ecosystem-dependent way too: higher pH went with lower stability in temperate grasslands but greater stability in alpine grasslands. The authors offer a nutrient-availability explanation, but explicitly call for long-term pH manipulation or nutrient addition experiments to clarify causality.

What the observational design cannot settle

Traits were sampled in a single year, while the stability metrics span multiple decades. That mismatch means the analysis cannot assess whether shifts in trait expression over time contribute to stability; repeated trait measurements would be needed to separate temporal and intraspecific trait variation.

The study is observational, so the direction of influence is not established. The authors note that stability may also shape which traits persist and dominate, potentially creating a feedback loop that their design cannot resolve.

The two grassland types also sit in different geographic regions, so the contrast may partly reflect spatially structured context such as latitude and regional productivity. Models controlling for latitude and aboveground biomass as a productivity proxy gave qualitatively similar trait–stability relationships, which the authors use to argue the conclusions are not driven solely by broad geographic gradients. Historical or low-level grazing could not be ruled out because spatially explicit, time-resolved grazing data were unavailable for all sites.

Match the stabilizing strategy to the dominant constraint

The paper's practical priority follows from its own contrast rather than from a general rule. Where stress is persistent and resource supply chronically constrained, as in the alpine system, conservative traits were associated with higher resistance and temporal stability. Where resources are less chronically limiting, as in the temperate system, acquisitive strategies were associated with the same two outcomes.

So the decision-relevant question is not which trait syndrome is better but which constraint dominates at a given site, and which stability dimension is being managed for. If the goal is resilience after disturbance, acquisitive traits were positively associated with it in both grassland types, which cuts across the biome split seen for resistance and temporal stability.

The authors extend this to management: strategies should be tailored to ecosystems with distinct trait–stability patterns rather than transferred wholesale. They also restrict the interpretation of conservative and acquisitive syndromes to grasslands along environmental gradients, noting that the same traits may reflect different ecological strategies in other systems.

Sources used in this answer