Trait velocities reveal that mortality has driven widespread coordinated shifts in forest hydraulic trait composition

Abstract

Understanding the driving mechanisms behind existing patterns of vegetation hydraulic traits and community trait diversity is critical for advancing predictions of the terrestrial carbon cycle because hydraulic traits affect both ecosystem and Earth system responses to changing water availability. Here, we leverage an extensive trait database and a long-term continental forest plot network to map changes in community trait distributions and quantify “trait velocities” (the rate of change in community-weighted traits) for different regions and different forest types across the United States from 2000 to the present. We show that diversity in hydraulic traits and photosynthetic characteristics is more related to local water availability than overall species diversity. Finally, we find evidence for coordinated shifts toward communities with more drought-tolerant traits driven by tree mortality, but the magnitude of responses differs depending on forest type. The hydraulic trait distribution maps provide a publicly available platform to fundamentally advance understanding of community trait change in response to climate change and predictive abilities of mechanistic vegetation models.

Global ecosystem-scale plant hydraulic traits retrieved using model-data fusion

“…Consistent with the FIA-based estimate, the retrieved ψ 50,x are overall lower in pixels dominated by evergreen needleleaf forests than in evergreen and deciduous broadleaf forests and mixed forests. However, across pixels, the ecosystem-scale ψ 50,x derived from remote sensing vary significantly more than the estimates from the Trugman et al (2020) dataset. Some fraction of this discrepancy might be due to intra-species variability in ψ 50,x , which is not accounted for in the FIA-based estimate, and due to uncertainty in the kriging-based interpolation used for upscaling from the sparse FIA plots to each 1 • pixel.…”

Low forest productivity associated with increasing drought‐tolerant species is compensated by an increase in drought‐tolerance richness

“…Similar to previous studies using the XFT dataset (e.g. Trugman et al, 2020), we only considered P 50 and P 88 data measured in branches and we discarded measurements from r‐shaped vulnerability curves. For species for which there were more than one trait measurements in the XFT database, we used the mean species value.…”

“…Peters et al (2015) also found that in U.S. forests, especially in western areas, drought‐tolerant tree species are increasing in abundance, relative to drought‐vulnerable species. Trugman et al (2020) used tree hydraulic traits to measure changes in community‐level drought tolerance (DT) and found that U.S. forests have become increasingly drought tolerant in the last 20 years. However, such increased representation of drought‐tolerant species may have consequences for forest functioning.…”

“…Brodribb, 2017). A recent example of this approach applied at large spatial scales was provided by Trugman et al (2020), who used tree hydraulic traits to study changes in forest tolerance to drought in the United States. We argue that such an approach provides a step forward in assessing the relationship between forest DT and its productivity.…”

Precipitation–productivity relationships and the duration of precipitation anomalies: An underappreciated dimension of climate change

“…Intrinsic, demographic responses may also mediate or buffer cumulative impacts over time, leading to diminished rather than amplified NPP responses. For example, there is observational evidence that trait distributions of some forests are shifting toward drought tolerance and that this is driven by tree mortality (Trugman et al, 2020). Long‐term studies show that prolonged drought selects for species that are drought resistant (Fauset et al, 2012; Liu et al, 2020), which supports the existence of stabilizing mechanisms at the community level that dampen ecosystem responses to PPT events (Felton & Smith, 2017; Jentsch et al, 2011; Lloret et al, 2012).…”