Coastal forests are increasingly vulnerable to climate change and sea-level rise, with flooding and salinity driving transitions to marsh-dominated ecosystems. Using the coastal version of FATES-Hydro, we conducted 30-year simulations at two coastal forest sites-a broadleaf swamp white oak stand at Lake Erie and a conifer loblolly pine stand at Chesapeake Bay-under historical climate and elevated CO2 (+100 ppm) and temperature (+1.5 degrees C) scenarios. Elevated CO2 increased net primary productivity at both sites, while warming alone intensified hydraulic stress and accelerated mortality, particularly in the conifer stand. Simulations show that elevated temperatures intensify vapor pressure deficit and hydraulic stress on trees already experiencing salinity- and submersion-driven water stress, increasing tree mortality beyond what would be expected in a non-water-limited environment. Marsh expansion partially compensated for tree loss at the Lake Erie site but reduced ecosystem productivity in the conifer forest at Chesapeake Bay. Our results highlight how differences in stand structure, phenology, and local hydrology modulate ecosystem trajectories under climate change, emphasizing the importance of demographic and community-level processes for predicting the fate of coastal forests.

Stomatal closure is a pervasive response among trees exposed to flooded soil. We tested whether this response is caused by reduced hydraulic conductance in the soil-to-leaf hydraulic continuum (ktotal), and particularly by reduced root hydraulic conductance (kroot), which has been widely hypothesized. We tracked stomatal conductance at the leaf level (gs) and canopy scale (Gs) along with physiological conditions in two temperate tree species, Magnolia grandiflora and Quercus virginiana, that were subjected to flood and control conditions in a greenhouse experiment. Flooding reduced gs, Gs, kroot and ktotal. Path analysis showed strong support for direct effects of ktotal on gs and for flood duration on ktotal, but not kroot on ktotal. A process-based model that accounted for the ktotal reduction predicted the timeseries of Gs in flood and control treatment trees reasonably well (predicted versus observed Gs R2 = 0.80 and 0.51 for M. grandiflora and Q. virginiana, respectively). However, accounting only for kroot reduction in flooded trees was insufficient for predicting observed Gs reduction. Together, these results suggest that hydraulic constraints were not limited to roots and highlight the need to account for flooding effects on ktotal when projecting forest ecosystem function using process-based models.

Coastal forests are increasingly experiencing mortality due to inundation by fresh- and seawater, leading to their replacement by marshes. These shifts alter vegetation composition, biogeochemical cycling, carbon storage, and hydrology. Using a hydraulically enabled ecosystem demography model (FATES-Hydro), we conducted numerical experiments to investigate the mechanisms behind inundation-driven forest loss and the ecosystem-scale consequences of forest-to-marsh transitions. We compared mortality processes and their effects across broadleaf and conifer trees at two coastal sites – Lake Erie (freshwater) and Chesapeake Bay (saline). Our simulations show that hydraulic failure, driven by root loss under prolonged flooding, is the primary mortality mechanism across both tree types and sites. Forest replacement by marsh reduced ecosystem-scale leaf area index (LAI), gross primary production (GPP), transpiration, and deep soil water uptake in conifer forests, while broadleaf forests experienced smaller changes due to lower initial LAI and greater marsh compensation. Marsh invasion occurred following canopy thinning driven by tree mortality. These findings suggest that, under similar root loss, hydraulic failure dominates coastal tree mortality regardless of species or water type, with denser forests experiencing stronger ecosystem impacts. Our study identifies key mortality mechanisms and offers testable hypotheses for future empirical studies on coastal vegetation change.

Extreme droughts are a major determinant of ecosystem disturbance that impacts plant communities and feeds back into climate change through changes in plant functioning. However, the complex relationships between aboveground and belowground plant hydraulic traits and their role in governing plant responses to drought are not fully understood. In this study, we use a model, the Functionally Assembled Terrestrial Ecosystem Simulator in a configuration that includes plant hydraulics (FATES-Hydro), to investigate ecosystem responses to the 2012–2015 California drought in comparison with observations at a site in the southern Sierra Nevada that experienced widespread tree mortality during this drought. We conduct a sensitivity analysis to explore how different plant water sourcing and hydraulic strategies lead to differential responses during normal and drought conditions. The analysis shows the following. Deep roots that sustain productivity through the dry season are needed for the model to capture observed seasonal cycles of evapotranspiration (ET) and gross primary productivity (GPP) in normal years, and deep-rooted strategies are nonetheless subject to large reductions in ET and GPP when the deep soil reservoir is depleted during extreme droughts, in agreement with observations. Risky stomatal strategies lead to greater productivity during normal years as compared to safer stomatal control, but they also lead to a high risk of xylem embolism during the 2012–2015 drought. For a given stand density, stomatal and xylem traits have a stronger impact on plant water status than on ecosystem-level fluxes. Our study highlights the significance of resolving plant water sourcing strategies to represent drought impacts on plants and consequent feedbacks in models.

Relative sea level rise (SLR) increasingly impacts coastal ecosystems through the formation of ghost forests. To predict the future of coastal ecosystems under SLR and changing climate, it is important to understand the physiological mechanisms underlying coastal tree mortality and to integrate this knowledge into dynamic vegetation models. We incorporate the physiological effect of salinity and hypoxia in a dynamic vegetation model in the Earth system land model, and used the model to investigate the mechanisms of mortality of conifer forests on the west and east coast sites of USA, where trees experience different form of sea water exposure. Simulations suggest similar physiological mechanisms can result in different mortality patterns. At the east coast site that experienced severe increases in seawater exposure, trees loose photosynthetic capacity and roots rapidly, and both storage carbon and hydraulic conductance decrease significantly within a year. Over time, further consumption of storage carbon that leads to carbon starvation dominates mortality. At the west coast site that gradually exposed to seawater through SLR, hydraulic failure dominates mortality because root loss impacts on conductance are greater than the degree of storage carbon depletion. Measurements and modeling focused on understanding the physiological mechanisms of mortality is critical to reducing predictive uncertainty.

As temperatures decrease in autumn, vegetation of temperate and boreal ecosystems increases its tolerance to freezing. This process, known as hardening, results in a set of physiological changes at the molecular level that initiate modifications of cell membrane composition and the synthesis of anti-freeze proteins. Together with the freezing of extracellular water, anti-freeze proteins reduce plant water potentials and xylem conductivity. To represent the responses of vegetation to climate change, land surface schemes increasingly employ ‘hydrodynamic’ models that represent the explicit fluxes of water from soil and through plants. The functioning of such schemes under frozen soil conditions, however, is poorly understood. Nonetheless, hydraulic processes are of major importance in the dynamics of these systems, which can suffer from e.g. winter ‘frost drought’ events. In this study, we implement a scheme that represents hardening into CLM5.0-FATES-Hydro. FATES-Hydro is a plant hydrodynamics module in FATES, a cohort model of vegetation physiology, growth and dynamics hosted in CLM5.0. We find that, in frozen systems, it is necessary to introduce reductions in plant water loss associated with hardening to prevent winter desiccation. This work makes it possible to use CLM5.0-FATES-Hydro to model realistic impacts from frost droughts on vegetation growth and photosynthesis, leading to more reliable projections of how northern ecosystems respond to climate change.

Abstract. Plant community composition influences carbon, water and energy fluxes at regional to global scales. Composition is a dynamic property of ecosystems, arising from complex feedbacks among the environment, disturbance, and plant physiology. Vegetation demographic models (VDMs) allow investigation of the effects of changing climate and disturbance regimes on vegetation composition and fluxes. Such investigation requires that the models can accurately resolve these feedbacks to simulate realistic composition. Vegetation in VDMs is composed of plant functional types (PFTs), which are specified according to plant traits. Defining PFTs is challenging due to large variability in trait observations within and between plant types and a lack of understanding of model sensitivity to these traits. Here we present an approach for developing PFT parameterizations that are connected to the underlying ecological processes determining forest composition in the mixed-conifer forest of the Sierra Nevada Mountains of California, USA. We constrain multiple relative trait values between PFTs, as opposed to randomly sampling within the range of observations. An ensemble of PFT parameterizations are then filtered based on emergent forest properties meeting observation-based ecological criteria under alternate disturbance scenarios. A small ensemble of alternate PFT parameterizations is identified that produces plausible forest composition, and demonstrates variability in response to disturbance frequency and regional environmental variation. Retaining multiple PFT parameterizations allows us to quantify the uncertainty in forest responses due to variability in trait observations. Vegetation composition is a key emergent outcome from VDMs and our methodology provides a foundation for robust PFT parameterization across ecosystems.

Live fuel moisture content (LFMC) plays a critical role in wildfire dynamics, but little is known about responses of LFMC to multivariate climate change, e.g., warming temperature, CO2 fertilization, and altered precipitation patterns, leading to a limited prediction ability of future wildfire risks. Here, we use a hydrodynamic demographic vegetation model to estimate LFMC dynamics of chaparral shrubs, a dominant vegetation type in fire-prone southern California. We parameterize the model based on observed shrub allometry and hydraulic traits and evaluate the model's accuracy through comparisons between observed and simulated LFMC of three plant functional types (PFTs) under current climate conditions. Moreover, we estimate the number of days per year of LFMC below 79 % (which is a critical threshold for wildfire danger rating of southern California chaparral shrubs) from 1960 to 2099 for each PFT and compare the number of days below the threshold for medium and high greenhouse gas emission scenarios (RCP4.5 and 8.5). We find that climate change could lead to more days per year (5.2 %–14.8 % increase) with LFMC below 79 % between the historical (1960–1999) and future (2080–2099) periods, implying an increase in wildfire danger for chaparral shrubs in southern California. Under the high greenhouse gas emission scenario during the dry season, we find that the future LFMC reductions mainly result from a warming temperature, which leads to 9.1 %–18.6 % reduction in LFMC. Lower precipitation in the spring leads to a 6.3 %–8.1 % reduction in LFMC. The combined impacts of warming and precipitation change on fire season length are equal to the additive impacts of warming and precipitation change individually. Our results show that the CO2 fertilization will mitigate fire risk by causing a 3.5 %–4.8 % increase in LFMC. Our results suggest that multivariate climate change could cause a significant net reduction in LFMC and thus exacerbate future wildfire danger in chaparral shrub systems.

Abstract The leaf economic traits such as leaf area, maximum carbon assimilation rate, and venation are all correlated and related to water availability. Furthermore, leaves are often broad and large in humid areas and narrower in arid/semiarid and hot and cold areas. We use optimization theory to explain these patterns. We have created a constrained optimization leaf model linking leaf shape to vein structure that is integrated into coupled transpiration and carbon assimilation processes. The model maximizes net leaf carbon gain (NPPleaf) over the loss of xylem water potential. Modeled relations between leaf traits are consistent with empirically observed patterns. As the results of the leaf shape–venation relation, our model further predicts that a broadleaf has overall higher NPPleaf compared to a narrowleaf. In addition, a broadleaf has a lower stomatal resistance compared to a narrowleaf under the same level of constraint. With the same leaf area, a broadleaf will have, on average, larger conduits and lower total leaf xylem resistance and thus be more efficient in water transportation but less resistant to cavitation. By linking venation structure to leaf shape and using water potential as the constraint, our model provides a physical explanation for the general pattern of the covariance of leaf traits through the safety–efficiency trade‐off of leaf hydraulic design.

Western U.S. conifer forests harbor diverse ecological strategies that enable species to persist across a wide range of hydroclimate conditions, along with wildfire and eruptive insect outbreaks. Assessing climate influences on future forest composition and carbon sequestration requires vegetation process models that have sufficient ecological resolution to simulate this range of ecological variability. Here we present progress towards incorporating multiple shade and drought tolerance strategies in a vegetation demographic model parameterized for Western U.S. forests. We used the Functionally Assembled Terrestrial Ecosystem Simulator (FATES) to simulate a mixed conifer forest dominated by ponderosa pine and incense cedar in the Sierra Nevada Mountains of California. FATES resolves plant growth and respiration at the level of cohorts, defined by size and plant functional type. Incense cedar is shade and drought tolerant, while ponderosa pine is shade intolerant and the canopy dominant. We synthesized literature values of plant traits that correspond to important physiological and allometric parameters in FATES and conducted a sensitivity analysis within the observed parameter ranges with respect to carbon and water fluxes. Model output was benchmarked against carbon flux, water flux, and leaf area index measurements from the Critical Zone Observatory/AmeriFlux CZ2 site during 2010-2012. Specific leaf area, Vcmax, rooting distribution, and allometric equations had the most influence on simulated carbon and water fluxes. Final simulated average annual gross primary production (GPP) over 2010-2012 (1156 +- 79.2 gC/m2/yr) was 3.8% lower than observed GPP (1202 +-138.2 gC/m2/yr). Simulated evapotranspiration (ET, 373 +- 25 mm/yr) was 62% lower than measured ET (993 +-158 mm/yr). Simulated leaf area index (LAI, 1.2) was within the range of measured LAI (0.5-1.5). Preliminary analysis indicates underestimation of ET is likely due to an overestimation of soil water drainage. Our final parameter set allows pine and cedar coexistence to emerge from a bare ground initialization, and additional sensitivity testing of parameters important for coexistence are in progress. Clearly, observationally constrained parameters are critical for simulating ecosystem dynamics in Western U.S. forests.

The diffusive and advective erosion‐created landscapes have similar structure (hillslopes and channels) across different scales regardless of variations in drivers and controls. The relative magnitude of diffusive erosion to advective erosion (D/K ratio) in a landscape development model controls hillslope length, shape, and drainage density, which regulate soil moisture variation, one of the critical resources of plants, through the contributing area (A) and local slope (S) represented by a topographic index (TI). Here we explore the theoretical relation between geomorphic processes, TI, and the abundance and distribution of plants. We derived an analytical model that expresses the TI with D, K, and A. This gives us the relation between soil moisture variation and geomorphic processes. Plant tolerance curves are used to link plant performance to soil moisture. Using the hypothetical tolerance curves of three plants, we show that the abundance and distribution of xeric, mesic, and hydric plants on the landscape are regulated by the D/K ratio. Where diffusive erosion is the major erosion process (large D/K ratio), mesic plants have higher abundance relative to xeric and hydric plants and the landscape has longer and convex‐upward hillslope and low channel density. Increasing the dominance of advective erosion increases relative abundance of xeric and hydric plants dominance, and the landscape has short and concave hillslope and high channel density.