Plant hydraulics improves and topography mediates prediction of aspen mortality in southwestern USA

Abstract

SummaryElevated forest mortality has been attributed to climate change-induced droughts, but prediction of spatial mortality patterns remains challenging. We evaluated whether introducing plant hydraulics and topographic convergence-induced soil moisture variation to land surface models (LSM) can help explain spatial patterns of mortality.A scheme predicting plant hydraulic safety loss from soil moisture was developed using field measurements and a plant physiology-hydraulics model, TREES. The scheme was upscaled to Populus tremuloides forests across Colorado, USA, using LSM-modeled and topographymediated soil moisture, respectively. The spatial patterns of hydraulic safety loss were compared against aerial surveyed mortality.Incorporating hydraulic safety loss raised the explanatory power of mortality by 40% compared to LSM-modeled soil moisture. Topographic convergence was mostly influential in suppressing mortality in low and concave areas, explaining an additional 10% of the variations in mortality for those regions.Plant hydraulics integrated water stress along the soil-plant continuum and was more closely tied to plant physiological response to drought. In addition to the well-recognized topo-climate influence due to elevation and aspect, we found evidence that topographic convergence mediates tree mortality in certain parts of the landscape that are low and convergent, likely through influences on plant-available water.

Tree hydraulic traits are coordinated and strongly linked to climate‐of‐origin across a rainfall gradient

“…Plant functional traits are typically well correlated with the physical environment and are important in determining plant performance under environmental stress (Soudzilovskaia et al, ). Drought‐induced hydraulic dysfunction may strongly limit the distribution of species along environmental gradients of water availability (Choat, Sack, & Holbrook, ; Kursar et al, ; Larter et al, ; Pockman & Sperry, ), define critical drought tolerance thresholds (Brodribb & Cochard, ; Urli et al, ), and be used to predict drought‐induced tree mortality (Tai, Mackay, Anderegg, Sperry, & Brooks, ; Xu, Medvigy, Powers, Becknell, & Guan, ). However, determining species overall drought susceptibility will depend on a more thorough understanding of the suite of traits that define plant water use and drought response strategies (Blackman, Aspinwall, Resco de Dios, Smith, & Tissue, ; Brodribb, McAdam, & Carins Murphy, ; Gleason, Blackman, Cook, Laws, & Westoby, ) and their relationship to traits regulating growth and productivity.…”

Accounting for landscape heterogeneity improves spatial predictions of tree vulnerability to drought

“…Tai et al . () also found that including topography to approximate lateral redistribution improved predictions of Aspen, Populus tremuloides , mortality in Colorado. Our approach expands upon Tai et al .…”

“…Our approach expands upon Tai et al . (), by providing an alternative framework that cohesively integrates plant physiological thresholds limiting hydraulic capacity (Fig. b) and landscape processes.…”

“…We first aggregated both modeled results and observations of drought‐impacted area to hydrologically similar but noncontiguous stands of J. ashei , similar to Tai et al . (). Hydrologically similar stands, n = 24, were defined as stands with similar aspects, NE, −45° to 135° vs SW, 135° to 315°; soil depths, < 100 cm vs > 100 cm; soil texture, silty clay, clay loam or other; and topographic divergence vs convergence, TWI below or above the mean, respectively.…”

“…Alternatively, DGVMs include various mechanisms for modeling tree mortality; however, their algorithms are rarely tested with spatially explicit canopy loss observations (McDowell et al ., ), and often do not account for landscape heterogeneity of abiotic factors at fine spatial resolutions, such as soil conditions, slope and aspect (Moorcroft, ), even though these data are readily available for many locations. Many studies have focused on improving model representations of the physiological mechanisms of tree mortality (McDowell et al ., ; Parolari et al ., ; Mackay et al ., ); however, only a few have examined if including landscape heterogeneity would improve predictions of tree mortality (Tague et al ., ; Anderegg et al ., ; Tai et al ., ).…”