<?xml-model href='http://www.tei-c.org/release/xml/tei/custom/schema/relaxng/tei_all.rng' schematypens='http://relaxng.org/ns/structure/1.0'?><TEI xmlns="http://www.tei-c.org/ns/1.0">
	<teiHeader>
		<fileDesc>
			<titleStmt><title level='a'>Processes and mechanisms of coastal woody‐plant mortality</title></titleStmt>
			<publicationStmt>
				<publisher></publisher>
				<date>10/01/2022</date>
			</publicationStmt>
			<sourceDesc>
				<bibl> 
					<idno type="par_id">10424432</idno>
					<idno type="doi">10.1111/gcb.16297</idno>
					<title level='j'>Global Change Biology</title>
<idno>1354-1013</idno>
<biblScope unit="volume">28</biblScope>
<biblScope unit="issue">20</biblScope>					

					<author>Nate G. McDowell</author><author>Marilyn Ball</author><author>Ben Bond‐Lamberty</author><author>Matthew L. Kirwan</author><author>Ken W. Krauss</author><author>J. Patrick Megonigal</author><author>Maurizio Mencuccini</author><author>Nicholas D. Ward</author><author>Michael N. Weintraub</author><author>Vanessa Bailey</author>
				</bibl>
			</sourceDesc>
		</fileDesc>
		<profileDesc>
			<abstract><ab><![CDATA[Observations of woody plant mortality in coastal ecosystems are globally widespread, but the overarching processes and underlying mechanisms are poorly understood. This knowledge deficiency, combined with rapidly changing water levels, storm surges, atmospheric CO 2 , and vapor pressure deficit, creates large predictive uncertainty regarding how coastal ecosystems will respond to global change. Here, we synthesize the literature on the mechanisms that underlie coastal woody-plant mortality, with the goal of producing a testable hypothesis framework. The key emergent mechanisms underlying mortality include hypoxic, osmotic, and ionic-driven reductions in whole-plant hydraulic conductance and photosynthesis that ultimately drive the coupled processes of hydraulic failure and carbon starvation. The relative importance of these processes in driving mortality, their order of progression, and their degree of coupling depends on the characteristics of the anomalous water exposure, on topographic effects, and on taxa-specific variation in traits and trait acclimation. Greater inundation exposure could accelerate mortality globally; however, the interaction of changing inundation exposure with elevated CO 2 , drought, and rising vapor pressure deficit could influence mortality likelihood. Models of coastal forests that incorporate the frequency and duration of inundation, the role of climatic drivers, and the processes of hydraulic failure and carbon starvation can yield improved estimates of inundation-induced woody-plant mortality.]]></ab></abstract>
		</profileDesc>
	</teiHeader>
	<text><body xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink">
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="1">| INTRODUC TI ON</head><p>Woody-plant dominated coastal ecosystems are showing increasing evidence of large-scale mortality <ref type="bibr">(Sippo et al., 2018)</ref>. Observations of 'ghost forest' formation from inundation are becoming globally widespread for both halophytic and glycophytic woody plants (Figure <ref type="figure">1</ref>; Table <ref type="table">1</ref>; <ref type="bibr">Jimenez et al., 1985;</ref><ref type="bibr">Kirwan &amp; Gedan, 2019;</ref><ref type="bibr">Lovelock et al., 2017;</ref><ref type="bibr">Penfound &amp; Hathaway, 1938;</ref><ref type="bibr">Shreve et al., 1910;</ref><ref type="bibr">Sippo et al., 2018;</ref><ref type="bibr">Wang et al., 2019)</ref>. Accelerating relative sea-level rise and increasing frequency and severity of storm surges are anticipated to continue <ref type="bibr">(Jevrejeva et al., 2016;</ref><ref type="bibr">Vermeer &amp; Rahmstorf, 2009)</ref>, as are increasing relative freshwater level rise in lakes and associated coastal shorelines and wetlands <ref type="bibr">(Woolway et al., 2020)</ref>. These changes are increasing the extent of coastal forest mortality <ref type="bibr">(Schieder &amp; Kirwan, 2019)</ref>.</p><p>The changes in coastal ecosystem water exposure occur on a backdrop of global rises in atmospheric CO 2 , temperature, vapor pressure deficit (VPD), and drought frequency and duration <ref type="bibr">(Cook et al., 2020;</ref><ref type="bibr">Friedlingstein et al., 2019;</ref><ref type="bibr">Trenberth et al., 2014;</ref><ref type="bibr">Ukkola et al., 2020)</ref>, all of which lead to nonlinear interactions that challenge our understanding. An additionally large barrier to our understanding of ghost forest formation is the lack of tests of the physiological mechanisms underlying mortality, specifically conducted on dying, mature plants in coastal systems. These challenges leave our society with large predictive uncertainty regarding when and where ghosts forests will form, and their subsequent consequences on ecosystem resources <ref type="bibr">(Ward et al., 2020)</ref>. Improving our understanding of the drivers and processes of coastal woody plant mortality, in relation to a changing climate, is needed if we are to predict forest loss <ref type="bibr">(Kirwan &amp; Gedan, 2019)</ref>.</p><p>The physiological mechanisms of hypoxia and salinity impacts on plants have been well-studied <ref type="bibr">(Colmer &amp; Flowers, 2008;</ref><ref type="bibr">Duberstein et al., 2020;</ref><ref type="bibr">Flowers &amp; Colmer, 2015;</ref><ref type="bibr">Flowers et al., 1977;</ref><ref type="bibr">Greenway &amp; Munns, 1980;</ref><ref type="bibr">Krauss et al., 2009)</ref>. These mechanisms include hypoxia-and salinity-induced reductions in belowground hydraulic conductance, stomatal closure, reduced photosynthesis, and crown loss among others. However, these mechanisms have not yet been placed within a process framework that integrates the physiological, geomorphic, and climatic drivers and mechanisms into a predictive capacity (wherein mechanisms are components of processes; Table <ref type="table">1</ref>; <ref type="bibr">Merriam-Webster, 2021)</ref>. Existing predictive frameworks of coastal plant loss have advanced representation of the geomorphic drivers, and in some cases, elevation-growth relationships <ref type="bibr">(Carr et al., 2020;</ref><ref type="bibr">Doyle et al., 2010;</ref><ref type="bibr">Enwright et al., 2016;</ref><ref type="bibr">Fagherazzi et al., 2019;</ref><ref type="bibr">Holmquist et al., 2021;</ref><ref type="bibr">Keddy &amp; Campbell, 2019;</ref><ref type="bibr">Kirwan et al., 2016;</ref><ref type="bibr">Theuerkauf &amp; Braun, 2021)</ref>. Such approaches are critical to prediction because ghost forest formation depends on topographic position <ref type="bibr">(Schieder &amp; Kirwan, 2019)</ref>. Integration of these frameworks with mechanistic modeling of the physiological responses to hypoxia and salinity (e.g. <ref type="bibr">Yoshikai et al., 2021)</ref> could enable representation of the compounding plant responses to rising CO 2 , temperature, VPD, and droughts <ref type="bibr">(McDowell et al., 2020)</ref>. Integrating these disparate sets of knowledge can additionally provide a road map for future research into the formation of ghost forests under a changing environment. Here, we bring together the mechanisms of coastal plant mortality within the framework from which the hydraulic failure and carbon starvation hypotheses originated <ref type="bibr">(McDowell et al., 2008</ref><ref type="bibr">(McDowell et al., , 2022))</ref>.</p><p>Hydraulic failure, or the severe and irreversible loss of hydraulic conductance from xylem embolism formation that leads to dehydration of downstream tissues, appears to be a ubiquitous process in drought mortality <ref type="bibr">(Adams et al., 2017)</ref>: a mild degree of hydraulic failure was observed in the only study to date of the hydraulics and gas exchange patterns during coastal-tree mortality <ref type="bibr">(Zhang, Wang, et al., 2021)</ref>. Hydraulic failure represents either a causal or symptomatic step in the mortality process, even in extreme experimentally induced shade or low CO 2 conditions when mortality is driven by the carbon starvation process <ref type="bibr">(Sevanto et al., 2014;</ref><ref type="bibr">Weber et al., 2018)</ref>. The process of carbon starvation, or the decline in carbohydrate supply required to meet metabolic, hydraulic, and defensive needs, promotes cell rupture and mortality <ref type="bibr">(McDowell et al., 2011</ref><ref type="bibr">(McDowell et al., , 2022))</ref>. Severe carbon starvation was observed in the only study of the physiology of ghost forest formation in mature trees <ref type="bibr">(Zhang, McDowell, et al., 2021)</ref>.</p><p>The hydraulic framework is appropriate for generating hypotheses (Table <ref type="table">2</ref>) regarding ghost forest formation because the mechanisms underlying hypoxia-and salinity-induced mortality are consistent with those included in the original framework: hydraulic and photosynthetic limitations. Even the mechanism of salt-toxicityinduced reductions in photosynthesis, which was not originally considered, fits squarely within the framework because of its impact on plant carbohydrate supply <ref type="bibr">(Munns &amp; Termaat, 1986)</ref>. This proposed framework is consistent with the two-phase model in which belowground osmotic changes reduce water uptake immediately upon inundation, which increases the risk of hydraulic failure and declining carbon balance, followed by a second, slower phase in which ionic impacts damage foliage, which feeds back to exacerbate carbon starvation <ref type="bibr">(Munns, 1993)</ref>. Hydraulic failure and carbon starvation are now simulated by ecosystem and Earth system models (e.g. <ref type="bibr">Fisher et al., 2010;</ref><ref type="bibr">Jiang et al., 2016;</ref><ref type="bibr">Kennedy et al., 2019;</ref><ref type="bibr">Koven et al., 2020;</ref><ref type="bibr">McDowell, Fisher, et al., 2013)</ref>; thus understanding how the individual mechanisms lead to hydraulic failure and/or carbon starvation provides a direct link to predictive modeling at multiple scales.</p><p>Our objective is to review the state-of-the-knowledge regarding coastal woody-plant mortality in both fresh and saline environments, with the goal of enabling better informed prediction. To achieve this objective, we reviewed the literature regarding ghost TA B L E 1 Glossary.</p><p>Anoxia: the complete absence of O 2 .</p><p>Carbon starvation: mortality resulting from an inability to provide sufficient carbohydrate supply to metabolic processes necessary for survival.</p><p>Cytorrhysis: irreparable damage to cell walls after cellular collapse from the loss of internal positive pressure.</p><p>Die-back: the partial loss of canopy or root biomass, without wholeplant mortality.</p><p>Hydraulic failure: the accumulation of emboli within the sapwood past a threshold after which water transport is irrecoverable leading to dehydration and cytrorrhysis of downstream tissues.</p><p>Hypoxia: reductions in O 2 below levels optimal for root physiology.</p><p>Mechanism: a system of parts working together within a process; a piece of the machinery.</p><p>Mortality: the irreversible cessation of metabolism and the associated inability to regenerate.</p><p>Process: a series of mechanisms that leads to an endpoint.</p><p>Salinity: the amount of salt dissolved in water, which can range from near zero in freshwater to &gt;35 parts per thousand in seawater.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>TA B L E 2 Hypotheses.</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Hypothesis Description 1</head><p>Mortality is increasing globally due to rising sea levels and freshwater levels.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>2</head><p>Mortality results from the interaction of topographic, hydrologic, and physiological factors.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>3</head><p>Glycophytes will experience more rapid and widespread mortality than halophytes.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>4</head><p>Hydraulic failure and carbon starvation are common processes, with taxa-specific adaptations leading to a range of mortality thresholds.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>5</head><p>Hypoxia and salinity drive similar mechanisms of mortality with the exception of toxicity associated with elevated salinity.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>6</head><p>Hydraulic failure is driven by hypoxia-and salinity-induced reductions in belowground water uptake.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>7</head><p>Carbon starvation is driven by reduced stomatal and mesophyll conductance, lost turgor, photosynthetic ion toxicity, phloem transfer reductions, crown loss, and elevated respiration.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>8</head><p>Reduced growth predisposes plants to mortality through physical and physiological feedbacks.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>9</head><p>Rising CO 2 can alleviate or exacerbate mortality associated with rising water levels.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>10</head><p>Rising droughts and VPD will exacerbate mortality due to rising water levels.</p><p>Note: Numerous hypotheses arise from application of the hydraulic framework to predict mortality from sea-level and freshwater-level rise. Here we list some of the most critical hypotheses that should be tested to allow improved understanding and prediction of future coastal woody-plant mortality. These hypotheses address the interaction between topography and physiology, the physiological mechanisms underlying hydraulic failure and carbon starvation, and the potential interactions of changing CO 2 and climate with the physiological responses to rising inundation.</p><p>forest formation from ecological, biophysical, and physiological domains to generate hypotheses regarding the underlying mechanisms of mortality for both halophytic and glycophytic woody plants. We assemble evidence from theory, observations, experiments, and models to generate a hypothesis framework. This framework is consistent with prior theories on the physiological impacts of hypoxia and salinity, while recognizing the underlying mechanisms as interdependent. This framework should also be pertinent to woody-plants of all sizes, from regenerating seedlings to large, mature individuals, and for both halophytic and glycophytic plants. We do not focus on anthropogenic disturbances (e.g. shoreline hardening, levee construction, or deforestation) in this review; however, we cite that literature when appropriate to understand mechanisms. We do, however, briefly address anthropogenic disturbances due to their globally widespread importance and significant impacts on mortality. Likewise, while our focus is on plants located along salt-or fresh-water shorelines, we do draw upon literature from upland, inland river, and wetland ecosystems where appropriate. To the degree that hypoxia drives mortality mechanisms, our framework is pertinent to both saline and freshwater inundation-driven mortality.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2">| PAT TERN S OF WOODY-PL ANT MORTALIT Y IN COA S TAL ECOSYS TEMS</head><p>The term ghost forest was coined originally to describe forest loss along the Gulf of Mexico <ref type="bibr">(Penfound &amp; Hathaway, 1938)</ref>, with documented occurrences along the North American Atlantic coast dating back to 1791 <ref type="bibr">(Spears, 1890;</ref><ref type="bibr">Van Doren, 1928)</ref>. Ghost forests have been observed across a large range of latitudes and ecosystem types, including widespread observations in both halophytic and glycophytic species (Figure <ref type="figure">2</ref>). Despite the variety of drivers of coastal plant mortality, the signal of climate-and sea level rise-driven mortality is becoming apparent <ref type="bibr">(Sippo et al., 2018)</ref>.</p><p>Long-term observations are scarce, but the available site-specific records of sufficient length show increasing rates of mortality over time <ref type="bibr">(Kirwan et al., 2016;</ref><ref type="bibr">Krauss, Demopoulos, et al., 2018;</ref><ref type="bibr">Krauss, Noe, et al., 2018;</ref><ref type="bibr">Lewis et al., 2016;</ref><ref type="bibr">Schieder &amp; Kirwan, 2019)</ref>. The map in Figure <ref type="figure">2</ref> does not represent the global distribution of coastal mortality events in their entirety due to insufficient systematic geographical sampling. However, potential hotspots do emerge such as along the eastern coast of North America, where sea level rise combined with land subsidence, a weakening of the Gulf Stream, and relatively flat topography have promoted widespread plant mortality <ref type="bibr">(Kirwan &amp; Gedan, 2019;</ref><ref type="bibr">Sallenger et al., 2012;</ref><ref type="bibr">Schieder et al., 2018;</ref><ref type="bibr">Smith, 2013;</ref><ref type="bibr">Smith et al., 2017</ref><ref type="bibr">Smith et al., , 2021;;</ref><ref type="bibr">Smith &amp; Kirwan, 2021;</ref><ref type="bibr">Ury et al., 2021)</ref>. The weight of this regional and global evidence leads to the hypothesis that coastal plant mortality is increasing.</p><p>Coastal woody-plant mortality occurs in both mangrove and non-mangrove systems (Figure <ref type="figure">2</ref>). Cyclones, tsunamis, and hurricanes regularly cause damage and mortality in both types of systems <ref type="bibr">(Lugo, 2008;</ref><ref type="bibr">Zeng et al., 2009)</ref>. Likewise, mangrove mortality can occur through freeze events <ref type="bibr">(Cavanaugh et al., 2019)</ref>.</p><p>Observations of mangrove mortality have increased in recent decades, with some individual mortality events exceeding 1000 ha along coastlines <ref type="bibr">(Duke et al., 2017)</ref>. These mortality events have corresponded with periods of severe heat, leading to the hypothesis that rising temperature exacerbates impacts associated with elevated hypoxia and salinity <ref type="bibr">(Allen et al., 2021;</ref><ref type="bibr">Lovelock et al., 2017)</ref>. This is consistent with observations and predictions from upland forests, which suggest that rising temperature, and in particular rising vapor pressure deficit (VPD), is driving increasing tree mortality <ref type="bibr">(McDowell et al., 2020;</ref><ref type="bibr">Williams et al., 2013)</ref>.</p><p>VPD represents atmospheric evaporative demand and increases non-linearly with temperature, causing significant impacts on plant survival through the increase in water loss relative to supply <ref type="bibr">(Grossiord et al., 2020)</ref>.</p><p>The global distribution of reported coastal mortality events resulting in ghost forests. Regions of mangrove forests are highlighted in green. Map updated with new observations, primarily from non-mangrove systems, from <ref type="bibr">Sippo et al. (2018)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3">| WOODY-PL ANT MORTALIT Y IN FRE S HWATER SYS TEMS</head><p>Most literature on coastal woody-plant mortality is from seashore systems, with far less information available for freshwater coastal systems, despite the critical importance of freshwater lakes and waterways for human uses <ref type="bibr">(Sterner et al., 2020)</ref>. Mortality of freshwater-coastal systems has been observed in relation to shifting water levels in floodplains <ref type="bibr">(Assahira et al., 2017;</ref><ref type="bibr">de Resende et al., 2019;</ref><ref type="bibr">Kramer et al., 2008;</ref><ref type="bibr">Levanic et al., 2011)</ref>, and along lake shorelines <ref type="bibr">(Smith, Fiorino, et al., 2021;</ref><ref type="bibr">Smith &amp; Kirwan, 2021;</ref><ref type="bibr">Theuerkauf &amp; Braun, 2021)</ref>. Freshwater flooding typically occurs in response to anthropogenic activities (e.g. impoundments), anomalously severe precipitation events, and through climatically forced changing water levels <ref type="bibr">(Gronewold et al., 2013;</ref><ref type="bibr">Kozlowski, 1984;</ref><ref type="bibr">Woolway et al., 2020)</ref>. Lake levels can fluctuate at an order of magnitude faster rate than sea levels <ref type="bibr">(Cazenave et al., 2014;</ref><ref type="bibr">Chen et al., 2017)</ref>, forcing rapid vegetation changes along associated shorelines and wetlands <ref type="bibr">(Smith, Fiorino, et al., 2021;</ref><ref type="bibr">Smith &amp; Kirwan, 2021;</ref><ref type="bibr">Theuerkauf &amp; Braun, 2021;</ref><ref type="bibr">Toner &amp; Keddy, 1997)</ref>.</p><p>Freshwater and seawater systems share mortality drivers such as physical disturbance and hypoxia, and therefore share some mechanisms underlying mortality from anomalous exposure to varying water levels. Freshwater-induced mortality is a strong function of the duration of inundation, with longer periods of inundation resulting in greater forest loss <ref type="bibr">(Smith, Fiorino, et al., 2021;</ref><ref type="bibr">Smith &amp; Kirwan, 2021)</ref>. The primary driver of mortality during freshwater inundation is hypoxic conditions that prevent oxidative phosphorylation in the roots and thus create energy deficits that can result in reduced root conductance and eventual root mortality <ref type="bibr">(Colmer &amp; Flowers, 2008;</ref><ref type="bibr">Pezeshki, 2001)</ref>. The subsequent reduced water uptake by roots results in stomatal closure, increased embolism, and crown loss (Figure <ref type="figure">3</ref>). Thus, the impacts of hypoxia can lead to either or both hydraulic failure and carbon starvation, similar to mortality associated with exposure to saline water. To enable better understanding and prediction of freshwater plant mortality from inundation requires testing the subset of mechanistic responses from Figure <ref type="figure">3</ref> that are pertinent to hypoxia-driven mortality. These include all the identified parameters in the framework figures except for osmotic impacts on conductance and ion toxicity to photosynthetic capacity. Ideally, studies could kill plants under varying</p><p>The interacting factors, drivers, mechanisms, and processes that cause coastal forest mortality under increasing seawater exposure. In this schematic, predisposing factors and drivers that occur before inundation (outer spiral) set the stage for subsequent mortality during and after inundation event(s). Pre-disposing factors include site characteristics such as freshwater access, with drivers such as rising VPD, and mechanisms such as salt toxicity, all contributing to the processes of hydraulic failure and carbon starvation. The order of presentation of factors represents a hypothesis regarding the chronology of mechanisms and impacts. The osmotic phase <ref type="bibr">(Munns &amp; Termaat, 1986)</ref> and associated increasing risk of hydraulic failure begins as soon as saline and hypoxic conditions start upon inundation, with the ionic phase requiring more time to manifest as salts accumulate in the crown. Carbon starvation starts initially upon stomatal closure due to the associated reduction in photosynthesis. We note that increasing embolism and stomatal closure typically coincide (Rodriguez-Dominguez &amp; Brodribb, 2020) but are shown separately for clarity.</p><p>duration of inundation with and without salinity to tease apart the relative roles of hypoxia from salt induced death, and their associated mechanisms.</p><p>We note that freshwater systems located upstream of saltwater, such as river estuaries, may experience novel salinity exposure due to rising sea level, storm surges, and human hydrologic alteration, leading to plants tolerant of hypoxia suddenly being exposed to salt <ref type="bibr">(Keim et al., 2019;</ref><ref type="bibr">Stagg et al., 2016)</ref>. How these differences in trait adaptations along the freshwater-saltwater continuum confer tolerance or vulnerability to changing salinity conditions is poorly understood <ref type="bibr">(Colmer &amp; Flowers, 2008)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="4">| PRED IS P OS ING FAC TOR S UNDERLYING COA S TAL WOODY-PL ANT MORTALIT Y</head><p>Ghost forest formation can result from the interaction of predisposing factors, external drivers, and the subsequent responses that lead to hydraulic failure, carbon starvation, and ultimately mortality (Figure <ref type="figure">3</ref>).</p><p>Predisposing factors are those that existed before the anomalous inundation event(s) that can promote either survival or mortality during and after the event(s). Site edaphic factors and external drivers provide the base from which subsequent impacts from rising water and storm surges occur. Topographic position, specifically the elevation above sea-or fresh-water levels, plays a large role because it sets the potential frequency and duration of inundation <ref type="bibr">(Fagherazzi et al., 2019;</ref><ref type="bibr">Schieder &amp; Kirwan, 2019;</ref><ref type="bibr">Smith, Fiorino, et al., 2021;</ref><ref type="bibr">Smith &amp; Kirwan, 2021;</ref><ref type="bibr">Taillie et al., 2019)</ref>. In saline systems, the role of subsurface hydrology on freshwater access during and after periods of inundation plays a large role in survival <ref type="bibr">(Gardner et al., 2002;</ref><ref type="bibr">Hayes et al., 2019;</ref><ref type="bibr">Semeniuk, 1983;</ref><ref type="bibr">Sternberg &amp; Swart, 1987;</ref><ref type="bibr">Williams et al., 1999</ref><ref type="bibr">Williams et al., , 2003))</ref>: mortality likelihood may be reduced when freshwater lenses exist within the belowground volume (soil or sediment) that is explored by plant roots, or when plants are located near rivers <ref type="bibr">(Ewe et al., 2007;</ref><ref type="bibr">Guimond &amp; Michael, 2020;</ref><ref type="bibr">Hsueh et al., 2016;</ref><ref type="bibr">Messerschmidt et al., 2021;</ref><ref type="bibr">Zhai et al., 2018)</ref>. Increasing exposure to kinetic energy from elevated water levels and increasing storms can physically damage root systems <ref type="bibr">(Kazemi et al., 2021)</ref> and thus reduce a plant's ability to forage for water and exchange O 2 during the subsequent increases in hypoxia and salinity. This impact can feed back upon itself because loss of root or whole-plant biomass to mortality will increase the kinetic energy received by the remaining surviving plants <ref type="bibr">(Massel et al., 1999)</ref>. Similarly, decreases in sedimentation rates, subsidence, and organic matter decomposition associated with plant mortality can feed back to produce lower soil surface elevations and hence greater exposure to inundation <ref type="bibr">(Blasco et al., 1996;</ref><ref type="bibr">Cahoon, 2006)</ref>.</p><p>Storms, which could be increasing under climate warming <ref type="bibr">(McDowell et al., 2020)</ref>, can also kill plants directly through stem snapping and uprooting <ref type="bibr">(Canham et al., 2010;</ref><ref type="bibr">Krauss &amp; Osland, 2020;</ref><ref type="bibr">Lugo, 2008;</ref><ref type="bibr">Zeng et al., 2009)</ref>. In surviving plants that lose a fraction of their crown, the likelihood of mortality increases significantly due in part to subsequent reductions in whole-plant photosynthesis relative to respiring biomass, and likely subsequent carbon starvation <ref type="bibr">(Gaylord et al., 2013;</ref><ref type="bibr">Poyatos et al., 2013;</ref><ref type="bibr">Zhang, McDowell, et al., 2021)</ref>.</p><p>Climate and CO 2 also play a role in predisposing plants to mortality (Figure <ref type="figure">3</ref>). Drought (periods of anomalously low precipitation)</p><p>prior to inundation can weaken trees, for example, by producing legacy embolism within the xylem and/or depleted carbohydrate stores, amplifying the likelihood of hydraulic failure and carbon starvation. Drought during or following inundation can promote mortality through the concentration of salt within pore water via high rates of evaporation with little recharge from precipitation <ref type="bibr">(Ardon et al., 2013;</ref><ref type="bibr">Desantis et al., 2007)</ref>. Rising temperature and VPD further exacerbate water stress through increased evaporation from both foliage and soil/sediment surfaces <ref type="bibr">(Grossiord et al., 2020)</ref>. Outbreaks of biotic agents such as insects or fungi can also weaken trees and thus predispose them to subsequent mortality <ref type="bibr">(Gaylord et al., 2013;</ref><ref type="bibr">Werner et al., 2006)</ref>. Elevated CO 2 can either promote or mitigate the likelihood of mortality depending on the balance of increased canopy growth, which requires greater freshwater availability to avoid mortality <ref type="bibr">(Duan et al., 2014)</ref>, versus water savings through stomatal closure, which may promote survival by increasing freshwater availability during drought <ref type="bibr">(McDowell et al., 2022)</ref>. Together, each of these external drivers can predispose woody-plants to survival or mortality under anomalous inundation (Figure <ref type="figure">3</ref>).</p><p>Mortality during and after anomalous inundation can be lessened by vegetation traits that confer tolerance to hypoxia and salinity. Tolerance of hypoxia and hypoxia-related toxins such as ferrous iron and hydrogen sulfide occurs across a variety of taxa and is achieved in part through aerenchyma and lenticels, higher soluble sugar contents, the ability to shift respiratory pathways, and the creation of sufficient antioxidant defenses against free-radical generation during anaerobic conditions <ref type="bibr">(Sairam et al., 2008)</ref>. Halophytes, and some glycophytic woody plants adapted to wet soil conditions, can have aerial roots that increase oxygen supply belowground to mitigate hypoxia impacts <ref type="bibr">(Ball, 1988)</ref>.</p><p>Regarding salt tolerance, halophytes have adaptations that allow them to survive in saline environments <ref type="bibr">(Acosta-Motos et al., 2017;</ref><ref type="bibr">Zhao et al., 2020)</ref>, including some species that have obligate requirements for moderate levels of pore water salinity <ref type="bibr">(Nguyen et al., 2015)</ref>. Halophytes have three general strategies for salttolerance: osmotic adjustment enabling maintenance of turgor and favorable water relations, salt exclusion restricting salt entry into roots during water absorption, and ion compartmentation enabling safe cellular accumulation of ions away from sensitive metabolic sites <ref type="bibr">(Ball, 1988)</ref>. Glycophytes are more typical of upland ecosystems and have fewer adaptations to tolerate hypoxia and salinity <ref type="bibr">(Greenway &amp; Munns, 1980)</ref>. For example, the presence of an exodermis within the roots is a critical adaptation because it reduces the loss of root oxygen to hypoxic soils and reduces the flux of salts into the roots <ref type="bibr">(Enstone et al., 2002;</ref><ref type="bibr">Hose et al., 2001)</ref>. The root exodermis is absent in all gymnosperms and common in angiosperms tolerant of drought, waterlogging and salinity <ref type="bibr">(Artur &amp; Kajala, 2021)</ref>.</p><p>This suggests that increasing water elevations from water level rise, storm surges, and high precipitation events in both saline and freshwater systems should have a disproportionate impact on glycophytic species without the adaptations to hypoxia and salinity, for example gymnosperms and many angiosperms. However, at some threshold, all plants become intolerant of elevated salt concentrations within the cytoplasm, with the species-specific thresholds dependent on their adaptations <ref type="bibr">(Yadav et al., 2011)</ref>. This suggests convergent processes of mortality among taxa, with the mechanisms expressed at different thresholds of hypoxia and salinity, consistent with simulation results <ref type="bibr">(Perri et al., 2019)</ref>. Thus, we propose that hydraulic failure and/or carbon starvation could potentially manifest during inundation-induced mortality across coastal taxa via shared mechanisms, with edaphic, predisposing, and vegetation traits producing effects at varying levels of hypoxia and salinity.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5">| MECHANIS MS UNDERLYING MORTALIT Y FROM HYP OXIA AND SALINIT Y</head><p>Understanding and predicting mortality under future increases in water inundation, atmospheric CO 2 , and VPD requires a framework that provides unambiguous definitions, generates testable hypotheses, and identifies uncertainties. Such a framework should also identify pools and fluxes of critical resources and their potential interdependencies and lethal thresholds, and should be relevant across different taxa and environmental conditions. Here, we propose that the hydraulic framework, from which the hydraulic failure and carbon starvation hypothesis originated, satisfies these criteria, enabling prediction of coastal woody-plant mortality when integrated with existing knowledge regarding inundation impacts. For seashore plants specifically, we integrate the hydraulic framework with the proposed two-phase osmotic and ionic impacts framework, in which water limitations drive rapid, widespread reductions in photosynthesis, which compound with the slower but equally damaging impacts of ion toxicity on photosynthesis and leaf mortality <ref type="bibr">(Munns, 1993)</ref>. We propose that hydraulic failure and carbon starvation drive the final step in mortality, widespread cellular rupture (cytorrhisis; <ref type="bibr">Beckett, 1997;</ref><ref type="bibr">Ding et al., 2014)</ref>, through the progressive impacts of reduced O 2 and pore water osmotic potential and subsequent constraints on root water uptake. Declining water acquisition promotes an increase in the risk of hydraulic failure while photosynthetic limitations due to reduced stomatal conductance, salt toxicity, and crown loss simultaneously promote carbon starvation (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>).</p><p>Notably, these are all whole-plant responses because, even if they start at the organ level, they eventually manifest throughout the </p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.1">| Hydraulic failure</head><p>Mortality of seawater-exposed woody plants begins belowground with prolonged exposure to reduced O 2 and in seawater shorelines, increasingly saline pore water (Figure <ref type="figure">3</ref>). Both hypoxia and elevated salinity reduce the water uptake capacity of roots <ref type="bibr">(Drew, 1997;</ref><ref type="bibr">Duberstein et al., 2020;</ref><ref type="bibr">Muller et al., 2009;</ref><ref type="bibr">Nicol&#225;s et al., 2005;</ref><ref type="bibr">Krauss &amp; Duberstein, 2010;</ref><ref type="bibr">Krauss et al., 2014;</ref><ref type="bibr">Pedersen et al., 2021)</ref>, forcing a large decline in transpiration that subsequently triggers the mechanisms underlying hydraulic failure and carbon starvation (Figure <ref type="figure">4</ref>). Hypoxia impacts the roots' capacity to acquire water due to reduced hydraulic conductance and ultimately through root death from O 2 starvation of respiration and toxic impacts on root metabolism <ref type="bibr">(Dat et al., 2006;</ref><ref type="bibr">Malik et al., 2002;</ref><ref type="bibr">Pedersen et al., 2021;</ref><ref type="bibr">Sairam et al., 2008;</ref><ref type="bibr">Youssef &amp; Saenger, 1998)</ref>. This loss of belowground water uptake decreases water supply relative to the whole-plant demand and hence places increased tension within the hydraulic system, elevating the risk of runaway cavitation within the xylem, leading to loss of water transport to downstream tissues, that is, hydraulic failure <ref type="bibr">(Tyree &amp; Sperry, 1989)</ref>.</p><p>Even relatively small increases in salinity, when they occur in ecosystems adapted to freshwater dominance, can lead to rapid ghost forest formation <ref type="bibr">(Pezeshki et al., 1987;</ref><ref type="bibr">Salinas et al., 1986)</ref>.</p><p>High salt concentrations can linger or even increase following the flooding event depending on subsequent freshwater input, VPD (through its impact on evaporation), soil hydraulic conductivity, and cation exchange capacity <ref type="bibr">(Ardon et al., 2013;</ref><ref type="bibr">Fagherazzi et al., 2019)</ref>. Rising salinity of soil pore water induces a cascade of changes in plant water relations driven foremost by the elevated osmotic potential of saline water, which subsequently reduces the potential gradient for water flux through the rhizosphere (Figure <ref type="figure">4</ref>; <ref type="bibr">Boursiac et al., 2005)</ref>. Standard seawater has an osmotic potential of -2.4 MPa <ref type="bibr">(Harvey, 1966)</ref>, which is lower than root water potential for most coastal taxa, and thus exposure to saline pore water can reduce or eliminate water flow into roots. Additionally, root hydraulic conductivity declines with increasing salinity, exacerbating the decline in root water impact beyond that of the osmotic impact <ref type="bibr">(L&#243;pez-Berenguer et al., 2006;</ref><ref type="bibr">Nedjimi, 2014)</ref>. This is called the osmotic phase of plant response to salinity (Figure <ref type="figure">3</ref>; <ref type="bibr">Munns &amp; Tester, 2008)</ref>. These mechanisms lead to reduced water uptake that promotes reductions in leaf water potential and subsequently declines in stomatal conductance, which minimizes transpiration and thus reduces the risk of hydraulic failure <ref type="bibr">(Clough &amp; Simm, 1989;</ref><ref type="bibr">Orsini et al., 2012;</ref><ref type="bibr">Perri et al., 2019;</ref><ref type="bibr">Sperry et al., 2016)</ref>. Stomatal closure also aids in avoiding turgor loss <ref type="bibr">(M&#233;ndez-Alonzo et al., 2016)</ref>.</p><p>However, under these conditions hydraulic failure becomes increasingly likely as water loss continues through the cuticle and bark even after complete stomatal closure <ref type="bibr">(Cochard et al., 2021;</ref><ref type="bibr">Wolfe, 2020)</ref>, resulting in a progressively more negative water balance (Figure <ref type="figure">4</ref>).</p><p>Stomatal closure also has implications for carbon starvation through reduced photosynthesis via reduced CO 2 diffusion through stomata (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>).</p><p>Hydraulic failure from salinity or hypoxia-driven root death can lead to partial or complete crown mortality, and to mortality of entire individuals if it occurs widely throughout the entire plant, including the meristems that are the source of regenerating tissue. Hydraulic failure in dying coastal plants has not been examined in depth, but recent evidence from conifer trees exposed to novel tidal inputs of saline water demonstrated that mortality was associated with up to 60% loss of xylem conductance <ref type="bibr">(Zhang, Wang, et al., 2021)</ref>. Increased resistance to hydraulic failure caused by exposure to saline and hypoxic conditions, for example acclimation may occur in seedlings <ref type="bibr">(Pandolfi et al., 2012</ref><ref type="bibr">(Pandolfi et al., , 2016;;</ref><ref type="bibr">Pezeshki et al., 1987;</ref><ref type="bibr">Stiller, 2009)</ref>, but this has not been investigated in mature plants to our knowledge. Reef &amp; Lovelock, 2015; Figure <ref type="figure">4</ref>). Additionally, reductions in turgor associated with higher salinity cause reduced cell expansion during growth, leading to fewer and/or smaller conduits for water to move through the xylem <ref type="bibr">(Munns, 1993)</ref>. These smaller elements subsequently reduce hydraulic conductance, and without reductions of distal foliar biomass, the rate of water supply-to-demand decreases significantly, increasing the tension within the xylem. Thus, sustained turgor loss anywhere in the root-to-leaf hydraulic system may cascade towards an accumulation of xylem emboli and lead to hydraulic failure.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.2">| Carbon starvation</head><p>Carbon starvation is the process by which limitations of carbon supply relative to demand cause failure to maintain metabolic and hydraulic functions above thresholds that prevent widespread cellular cytorrhysis <ref type="bibr">(McDowell et al., 2022)</ref>. Inundation increases the likelihood of starvation through reducing the ratio of carbon uptake to demand that manifest at leaf-and whole-plant scales over minutes to years (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>). The first impacts of increased hypoxia and salinity occur by the aforementioned reduction in belowground hydraulic conductance (Figure <ref type="figure">4</ref>). Reductions in hydraulic conductance force declines in CO 2 diffusion into foliage, and hence photosynthesis, due to stomatal closure (Figure <ref type="figure">5a</ref>; <ref type="bibr">Ball &amp; Farquhar, 1984;</ref><ref type="bibr">Brugnoli &amp; Lauteri, 1991;</ref><ref type="bibr">Li et al., 2021;</ref><ref type="bibr">Mustroph &amp; Albrecht, 2003;</ref><ref type="bibr">Zhang, Wang, et al., 2021)</ref>. This reduction in whole-plant photosynthesis is the most immediate mechanism driving carbon starvation <ref type="bibr">(Wang et al., 2019)</ref>. A second mechanism of reduced photosynthesis is declining mesophyll conductance to CO 2 induced by elevated salinity <ref type="bibr">(Flexas et al., 2004)</ref>. A third mechanism of reduced photosynthesis is turgor loss resulting from salinity stress, because photosynthetic rates decline with declining turgor and reach zero photosynthesis when turgor is lost (Figure <ref type="figure">4</ref>; <ref type="bibr">Brodribb et al., 2003;</ref><ref type="bibr">M&#233;ndez-Alonzo et al., 2016)</ref>.</p><p>A fourth mechanism driving a reduction in whole-plant photosynthesis under seawater exposure occurs via salt-induced damage to photosynthetic biochemistry (Figures 3-5a; <ref type="bibr">Ball &amp; Farquhar, 1984;</ref><ref type="bibr">Delatorre-Herrera et al., 2021;</ref><ref type="bibr">Li et al., 2021;</ref><ref type="bibr">Yadav et al., 2011)</ref>. Salt toxicity takes longer to manifest than the more immediate hypoxic and osmotic limitations upon water uptake <ref type="bibr">(Munns, 1993)</ref> due to the time lag in salt transport to shoots via the transpiration stream and the time required for salt to accumulate to toxic levels. Accumulation of salts in the foliar cytoplasm, and inhibition of nitrogen uptake, both directly impact photosynthetic capacity by reducing electron transport and carboxylation capacities <ref type="bibr">(Ball et al., 1987;</ref><ref type="bibr">Ge et al., 2014;</ref><ref type="bibr">James et al., 2006;</ref><ref type="bibr">Munns, 2005;</ref><ref type="bibr">Su&#225;rez &amp; Medina, 2006)</ref>. Salt accumulation in foliage can thus lead to dramatic reductions in photosynthetic capacity, particularly in glycophytes that have limited ability to exclude salts at the root surface (Figure <ref type="figure">5a</ref>; <ref type="bibr">Li et al., 2021)</ref>.</p><p>A fifth mechanism that reduces carbon uptake at the wholeplant scale is crown dieback, or foliage loss (Figure <ref type="figure">4</ref>  Beyond reduced photosynthesis, the process of carbon starvation can be exacerbated if there is an insufficient transport rate of carbon substrates to required metabolic processes <ref type="bibr">(Yu, 1999)</ref>.</p><p>Increased risk of carbon starvation can result from transport constraints through source and sink strength impacts upon phloem loading and unloading, and increasing viscosity (Figure <ref type="figure">4</ref>; <ref type="bibr">Mencuccini et al., 2015;</ref><ref type="bibr">Saglio, 1985)</ref>. The impact of salinity upon phloem function is poorly understood in seashore plants, but there is reason to suspect that phloem flow can be reduced under highly saline conditions <ref type="bibr">(Perri et al., 2019)</ref>. Lowered turgor due to increasing salt concentrations within the foliage apoplast (e.g., <ref type="bibr">Ottow et al., 2005;</ref><ref type="bibr">Zeng et al., 2009)</ref> can reduce phloem water transport, especially when associated with reduced carbohydrate availability to drive water flow, as occurs with inundation <ref type="bibr">(Chen &amp; Polle, 2009)</ref>. Reduced xylem water potentials due to constrained water uptake will reduce phloem water content, concentrating phloem sucrose and elevating the viscosity and hence further reducing flow <ref type="bibr">(H&#246;ltt&#228; et al., 2009;</ref><ref type="bibr">Jensen et al., 2013)</ref>. High osmotic pressure in the roots, if not accompanied by effective salt extrusion processes, can reduce phloem sink strength, potentially eliminating flow of carbohydrates <ref type="bibr">(Abdolzadeh et al., 2008;</ref><ref type="bibr">Matsushita et al., 1992;</ref><ref type="bibr">Yang et al., 2019)</ref>. Notably, osmoregulation could counteract reduced phloem transport if it generates sufficiently elevated osmolyte concentrations to maintain flow <ref type="bibr">(Perri et al., 2019;</ref><ref type="bibr">Plaut et al., 2012)</ref>. Thus, salinity and hypoxia F I G U R E 5 Declining tree carbon uptake promotes carbon starvation. Sitka spruce (Picea Sitchensis) trees exposed to seawater inundation exhibit strong declines in (a) photosynthetic capacity (the slope of photosynthesis versus internal CO 2 concentration), (b) the fraction of their crowns that are foliated, and (c) starch concentrations in the foliage, branches, stems and roots. Combined, the decline in both photosynthetic capacity and in crown leaf area result in a significant decline in starch concentrations, which is the primary storage pool of carbon. As trees approach death the starch concentrations reach zero, which is rarely seen except for in studies that force carbon starvation <ref type="bibr">(Quirk et al., 2013;</ref><ref type="bibr">Sevanto et al., 2014;</ref><ref type="bibr">Weber et al., 2018)</ref>. Data used to create these figures are from <ref type="bibr">Li et al. (2021)</ref> and <ref type="bibr">Zhang, McDowell, et al. (2021)</ref>.</p><p>can exacerbate localized carbon starvation through limitations to phloem flow, with the threshold for carbon starvation varying with adaptations to maintain phloem function. While these hypotheses are relatively well supported with evidence from studies of surviving plants in upland systems <ref type="bibr">(Chen &amp; Polle, 2009;</ref><ref type="bibr">Wu, 2018)</ref>, they should be tested on dying plants in coastal systems in order to draw robust conclusions <ref type="bibr">(McDowell, Ryan, et al., 2013)</ref>.</p><p>Concurrently with decreasing carbon supply due to hypoxia-and salinity-exposure, the metabolic demand for carbon could increase as plants shift resource allocation to promote survival.</p><p>Carbohydrates are needed in part for regulation of xylem sap osmotic potential, which contributes to the soil-to-root water potential gradient <ref type="bibr">(Acosta-Motos et al., 2017)</ref>. Declining resistance to xylem embolism has been associated with low carbohydrate concentrations in multiple experimental and observational studies, suggestive that reduced carbon balance can promote a hydraulic-carbon feedback <ref type="bibr">(Tomasella et al., 2020)</ref>. Respiratory consumption of carbohydrates increases for many reasons under hypoxia and salt stress <ref type="bibr">(Jacoby et al., 2011</ref><ref type="bibr">(Jacoby et al., , 2013))</ref>. Carbohydrates are consumed for use in both cellular osmoregulation and for free-radical scavenging, the requirements of which increase under salt-stress <ref type="bibr">(Acosta-Motos et al., 2017;</ref><ref type="bibr">Jouve et al., 2004;</ref><ref type="bibr">Munns &amp; Gilliham, 2015)</ref>. The demands to maintain protein and membrane stability and cross-membrane transport of salts and amino acids increase under hypoxia and salt stress <ref type="bibr">(Hoekstra et al., 2001;</ref><ref type="bibr">Koster &amp; Leopold, 1988;</ref><ref type="bibr">Matros et al., 2015;</ref><ref type="bibr">Ramel et al., 2009;</ref><ref type="bibr">Rolland et al., 2006;</ref><ref type="bibr">Sapes et al., 2021;</ref><ref type="bibr">Van den Ende &amp; Valluru, 2009;</ref><ref type="bibr">Yu, 1999)</ref>, thus creating feedback loops that promote subsequent starvation. Finally, increased respiration occurs through shifts in metabolic pathways used to maintain metabolism under hypoxia <ref type="bibr">(Drew, 1997)</ref>  Figure <ref type="figure">5a</ref>,b), can increase the likelihood of carbohydrate levels falling below thresholds for survival (e.g. Figure <ref type="figure">5c</ref>).</p><p>A final driver of mortality related to plant carbon relations under hypoxia and salinity stress is reduced growth <ref type="bibr">(Ball, 1988)</ref>. Reduced root and mycorrhizal growth limit the ability of plants to forage for freshwater <ref type="bibr">(Evelin et al., 2009;</ref><ref type="bibr">Hsueh et al., 2016;</ref><ref type="bibr">Sternberg &amp; Swart, 1987;</ref><ref type="bibr">Zhai et al., 2018)</ref>. Because roots play a critical role in sediment accretion <ref type="bibr">(Kirwan &amp; Megonigal, 2013;</ref><ref type="bibr">Kumara et al., 2010)</ref>, reduced root growth could cause decreased rates of sedimentation, leading to greater subsequent submergence <ref type="bibr">(Cahoon, 2003;</ref><ref type="bibr">Krauss, Demopoulos, et al., 2018;</ref><ref type="bibr">Krauss, Noe, et al., 2018)</ref>. Reduced carbon allocation to suberin in the roots, which excludes salt, can promote greater salt uptake <ref type="bibr">(Reef &amp; Lovelock, 2015)</ref>. Xylem growth dominates recovery of hydraulic conductance after embolism at mid-to long-term scales <ref type="bibr">(Brodribb et al., 2010)</ref>, hence lower xylem growth means less recovery of hydraulic conductance and greater risk of both hydraulic failure and carbon starvation. Finally, reduced growth of foliage limits whole-plant photosynthetic potential, thereby promoting subsequent starvation <ref type="bibr">(Poyatos et al., 2013)</ref>.</p><p>Together, these growth reductions drive hydraulic, carbohydrate, and sedimentation impacts that promote hydraulic failure and carbon starvation.</p><p>The bulk of evidence points to initial hydraulic limitations of hypoxia and salinity driving a cascade of impacts that can lead to hydraulic failure and carbon starvation, particularly when coupled with ion-toxicity F I G U R E 6 Hypothesized responses of mortality to rising hypoxia and salinity and the influence of rising CO 2 , VPD, and droughts. (a) As the cumulative duration of exposure to elevated hypoxia and salinity increases, hydraulic failure should be the first process to manifest because of the immediate declines in belowground hydraulic conductance that subsequently promote increased embolism, crown dieback, and eventual mortality. Carbon starvation will increase in likelihood at a slower rate because of the buffering capacity of stored carbohydrates but can become a dominant driver of mortality due to the large declines in wholeplant photosynthetic potential. (b) Glycophytes are more sensitive to hypoxia and salinity than halophytes, however, their responses to elevated CO 2 and VPD are similar. Elevated CO 2 can mitigate mortality likelihood at low levels of hypoxia and salinity, but these benefits are lost with rising hypoxia and salinity. The relative range of exposure shown in panel (b) allows for a range of exposure at across taxa, topography, soil depths, and climate among other variables that could drive variation in any convergence. Elevated VPD, in contrast, will increase mortality likelihood throughout the range of hypoxia and salinity, regardless of vegetation-type.</p><p>to photosynthesis and reduced crown surface area (Figures <ref type="figure">3</ref><ref type="figure">4</ref><ref type="figure">5</ref>). The manifestation of these mechanisms may vary under rising CO 2 and changing climate, thus we next consider the hypothesis framework under these increasingly important drivers of plant physiology.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="6">| THE ROLE OF RIS ING CO 2 , DROUG HT, AND VPD ON COA S TAL WOODY-PL ANT MORTALIT Y</head><p>Atmospheric CO 2 , droughts, and VPD are all rising rapidly and are expected to continue increasing for the foreseeable future (IPCC, <ref type="formula">2021</ref>), but the interactions of these drivers with hypoxia and salinity in causing mortality are largely unknown. We hypothesize that while halophytes and glycophytes have different sensitivities to hypoxia and salinity, they have similar responses to rising CO 2 , VPD, and drought (Figure <ref type="figure">6</ref>). This convergence hypothesis results from the consistent evidence suggesting that halophytes and glycophytes may both die via the same mechanisms, just at different points on the hypoxia and salinity curves (Figure <ref type="figure">6</ref>; reviewed above). Below, we review the specific mechanisms underlying potential shifts in mortality due to rising CO 2 , VPD, and drought.</p><p>In mesic, young, upland ecosystems, CO 2 fertilization can decrease stomatal conductance <ref type="bibr">(Ainsworth &amp; Rogers, 2007;</ref><ref type="bibr">Klein &amp; Ramon, 2019)</ref>, and can increase total biomass <ref type="bibr">(Walker et al., 2021)</ref>, fine root biomass <ref type="bibr">(Anderson-Teixeira et al., 2013)</ref>, rooting depths <ref type="bibr">(Iversen, 2010)</ref>, plant leaf area <ref type="bibr">(Norby &amp; Zak, 2011;</ref><ref type="bibr">Walker et al., 2021)</ref>, and leaf biomass to total biomass ratio <ref type="bibr">(Lovelock et al., 2016)</ref>. Specific to seashore systems, CO 2 fertilization can reduce oxidative stress under elevated salinity <ref type="bibr">(Geissler et al., 2009)</ref>.</p><p>CO 2 fertilization can reduce the likelihood of mortality (Figure <ref type="figure">6</ref>) if carbohydrate storage is enhanced by CO 2 <ref type="bibr">(Li et al., 2018)</ref>. Water savings from reduced stomatal conductance could maintain more freshwater availability to plants <ref type="bibr">(Keenan et al., 2013;</ref><ref type="bibr">Niklaus et al., 1998;</ref><ref type="bibr">Walker et al., 2021)</ref>.</p><p>Not all CO 2 -induced changes are beneficial to survival <ref type="bibr">(McDowell et al., 2022)</ref>. The increased demand for water due to elevated growth of the plant canopy could nullify any stomatal water savings by increasing whole plant water demand during stressful periods <ref type="bibr">(Duan et al., 2014)</ref>. This concept is referred to as 'structural overshoot', where plants cannot provide sufficient water to supply the larger crown during periods of resource limitation <ref type="bibr">(Jump et al., 2017;</ref><ref type="bibr">Ward et al., 2013</ref><ref type="bibr">Ward et al., , 2015;;</ref><ref type="bibr">Tor-ngern et al., 2015)</ref>. This whole-plant response may explain why nutrient enrichment promotes mangrove mortality <ref type="bibr">(Lovelock et al., 2009)</ref>, given that elevated nutrient availability typically increases the ratio of above-to below-ground biomass and thus predisposes trees to mortality under drought due to elevated water demand relative to supply <ref type="bibr">(Ewers et al., 2001)</ref>. Consistently, while CO 2 benefits plant growth under relatively low salinity, it does not benefit plants under elevated salinity levels (Figure <ref type="figure">6</ref>; Alongi, 2015; <ref type="bibr">Ball et al., 1987)</ref>. This is likely due to limited stomatal conductance to CO 2 via osmotic constraints on water flow and due to exceedance of the threshold for photosynthetic damage due to salt toxicity, both of which can nullify the benefits of elevated CO 2 <ref type="bibr">(Reef et al., 2016)</ref>.</p><p>This leads to the hypothesis that while CO 2 fertilization may or may not reduce mortality likelihood at low levels of hypoxia and salinity, the likely response to increasingly low O 2 and high salinity conditions should be a particularly steep response of mortality likelihood due to structural overshoot (Figure <ref type="figure">6</ref>). Thus, chronic increases in inundation events should overwhelm CO 2 benefits in locations where water level rise is accelerating.</p><p>Meteorological droughts, when precipitation is anomalously low, are anticipated to become more frequent <ref type="bibr">(Ukkola et al., 2020)</ref> and can drive widespread mortality in coastal systems (Figure <ref type="figure">6</ref>; <ref type="bibr">et al., 2007;</ref><ref type="bibr">Lovelock et al., 2017;</ref><ref type="bibr">Sippo et al., 2018)</ref>. Drought prior to extreme inundation events can predispose plants to mortality through increased xylem embolism and reduced carbohydrate storage and growth, thus exacerbating the likelihood of hydraulic failure and carbon starvation. Similarly, drought during or after inundation should promote these mortality processes through exacerbating the negative impacts of hypoxia and salinity on embolism likelihood, and subsequent reductions in photosynthesis and increased likelihood of mortality (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>). Drought after seawater inundation will increase soil or sediment pore water salinity <ref type="bibr">(Lovelock et al., 2017)</ref>, thus exacerbating osmotic impacts. Drought combined with inundation events are thus compounding drivers of hydraulic impacts that should exacerbate the risk of hydraulic failure and carbon starvation (Figure <ref type="figure">6</ref>).</p><p>VPD is rising in concert with increasing atmospheric temperature <ref type="bibr">(Grossiord et al., 2020)</ref>, with the increase in soil and leaf evaporation increasing mortality likelihood under anomalous water inundation (Figure <ref type="figure">6</ref>). Prior to stomatal closure, VPD-induced increases in transpiration elevate the risk of hydraulic failure as water demand becomes larger than supply, particularly when hydraulic supply is reduced due to hypoxia and salinity. After stomatal closure, elevated VPD continues to drive water loss through the foliar cuticle and bark, which becomes the dominant source of water loss after stomatal closure <ref type="bibr">(Duursma et al., 2019)</ref>. This residual, non-stomatal water loss exhibits a strongly non-linear, positive response to rising temperature, and heat waves are thus of particular concern for accelerating water loss and the likelihood of hydraulic failure <ref type="bibr">(Duursma et al., 2019)</ref>.</p><p>In seashore systems exposed to significant tidal fluctuations, increased evaporation due to rising VPD can cause elevated soil pore water salinity through concentration of salts during low-tides <ref type="bibr">(Cormier et al., 2013;</ref><ref type="bibr">Desantis et al., 2007)</ref>. This process of increasing salinity could have played a large role in the recent, widespread mangrove mortality in northern Australia, in which a drought/high VPD period corresponded with a period of particularly low sea level that resulted in extremely high pore water salinity concurrent with mortality <ref type="bibr">(Lovelock et al., 2017)</ref>. Thus, we hypothesize that the mortality likelihood is higher across the possible range of hypoxia and salinity under elevated VPD (Figure <ref type="figure">6</ref>). Together, rising CO 2 , VPD, and drought are likely to push plants towards accelerated mortality under rising hypoxia and salinity, though some degree of stress alleviation through elevated CO 2 might occur at low degrees of hypoxia and salinity (Figure <ref type="figure">6</ref>).</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="7">| ANTHROP OG ENIC DIS TURBAN CE S</head><p>Human-induced disturbances along coastal shorelines are globally widespread <ref type="bibr">(Goldberg et al., 2020;</ref><ref type="bibr">Kirwan &amp; Gedan, 2019)</ref>. Such disturbances can have significant, large-scale impacts on coastal woody-plants via the same mechanistic responses as to increases hypoxia and salinity (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>). These disturbances are often in the form of deltaic system alteration such as river dredging, and construction of dams, levees, navigation canals, road construction, and seawalls <ref type="bibr">(Kirwan &amp; Gedan, 2019;</ref><ref type="bibr">Lewis et al., 2016;</ref><ref type="bibr">Salinas et al., 1986)</ref>. In many regions globally, conversion of land to aquaculture or agriculture is a dominant disturbance <ref type="bibr">(Goldberg et al., 2020)</ref> as is tree cutting for timber <ref type="bibr">(Roma&#241;ach et al., 2018)</ref>.</p><p>Such landscape alterations can result in large-scale mortality due to shifts in hydrologic and sediment fluxes <ref type="bibr">(Jones et al., 2017;</ref><ref type="bibr">Syvitski et al., 2009)</ref>. Drying or ponding of freshwater dominated ecosystems due to hydrologic shifts can promote mortality through severe reductions in soil moisture availability or increased hypoxia, respectively. If such shifts result in increased intrusion of saline water, then osmotic and ionic impacts increase.</p><p>Loss of woody-plants due to human disturbance can result in reduced rates of sediment accumulation and compaction <ref type="bibr">(Lang'at et al., 2014)</ref>, while planting increases surface elevation through increased sedimentation. These changes in elevation make large differences in plant exposure to hypoxia, salinity, and kinetic energy from waves. Fragmentation and reduced stem densities likewise increase mortality from storm surges <ref type="bibr">(Kumara et al., 2010;</ref><ref type="bibr">Lagomasino et al., 2021)</ref>.</p><p>Protection of coastal ecosystems dominated by woody-plants is a large challenge <ref type="bibr">(Lee et al., 2019;</ref><ref type="bibr">Osland et al., 2018)</ref>. Conservation is complex due to the wide range of land stewardship of shorelines globally <ref type="bibr">(Titus et al., 2009)</ref> and because of the required coordination between national policies and local communities <ref type="bibr">(Roma&#241;ach et al., 2018)</ref>. Avoiding the construction of shoreline structures through coastal land conservation is critical, while restoration of degraded systems can require investment into structures that move the local hydrology back towards a pre-disturbance state <ref type="bibr">(Temmerman &amp; Kirwan, 2015)</ref>. Economic factors play a significant role influencing restoration due to construction costs <ref type="bibr">(Hinkel et al., 2014)</ref>. Balancing flood mitigation with restoration is another critical factor determining the ability to restore damaged ecosystems <ref type="bibr">(Calil et al., 2015)</ref>.</p><p>Global conservation agreements have resulted in protection of large swaths of coastal forests, as have public outreach efforts to educate society regarding the ecosystem services lost through coastal plant mortality <ref type="bibr">(Roma&#241;ach et al., 2018)</ref>. Ultimately, balancing the economic benefits of coastal development against the ecosystem services provided by intact ecosystems can drive policy-and decision-making regarding future human-induced disturbances <ref type="bibr">(Barbier et al., 2011;</ref><ref type="bibr">Friess et al., 2016;</ref><ref type="bibr">Hochard et al., 2019)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="8">| FUTURE DIREC TIONS</head><p>There have been significant advances in our understanding of how hypoxia and salinity impact plant physiology <ref type="bibr">(Ball &amp; Farquhar, 1984;</ref><ref type="bibr">Kozlowski, 1997;</ref><ref type="bibr">Longstreth &amp; Nobel, 1979;</ref><ref type="bibr">Munns, 1993;</ref><ref type="bibr">Munns &amp; Gilliham, 2015)</ref>. The various deleterious impacts of hypoxia and salinity on root function, water uptake and demand, photosynthetic biochemistry, carbon transport and metabolism, and on turgor are relatively well understood at cellular and molecular levels <ref type="bibr">(Gupta &amp; Huang, 2014;</ref><ref type="bibr">Hasegawa et al., 2000;</ref><ref type="bibr">Munns &amp; Tester, 2008)</ref>.</p><p>However, the application of this mechanistic knowledge to understanding coastal woody-plant mortality from hypoxia and salinity exposure is relatively unexplored.</p><p>We suggest a testable framework for understanding and modeling inundation-and salinity-induced mortality (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>).</p><p>We hypothesize that hypoxia-and salinity-driven shifts in water uptake increase the risk of hydraulic failure and carbon starvation through multiple mechanisms (Table <ref type="table">2</ref>). The immediate and sustained reductions in hydraulic conductance due to belowground osmotic constraints and potentially root death should simultaneously increase likelihoods of hydraulic failure and carbon starvation, with interdependent feedbacks between them (Figure <ref type="figure">4</ref>). Photosynthetic declines due to declining stomatal conductance, ion toxicity to photosynthetic biochemistry, reduced phloem transport, and loss of crown surface area, combined with increasing respiratory demand, all should drive large reductions in carbohydrate supply (Figure <ref type="figure">5</ref>).</p><p>We additionally hypothesize that rising CO 2 can either benefit or hurt plants during inundation events, whereas rising VPD will predispose plants to mortality above that expected from inundation alone (Table <ref type="table">2</ref> and<ref type="table">Figure 6</ref>). These and other hypotheses embodied by the framework should be testable in both field and numerical modeling experiments.</p><p>This mechanistic framework allows identification of both challenges and solutions to understanding and modeling coastal woodyplant mortality. Critically, there remain too few observations of the global distribution of ghost forests to allow strong conclusions about geographic patterns or trends of ghost forest formation over time (Figure <ref type="figure">2</ref>). There is also a paucity of measurements on mature coastal plants needed for testing the hypotheses proposed within  <ref type="figure">4</ref>). Identifying when and where belowground conductance approaches zero due to osmotic constraints is critical because this sets the remainder of the mortality mechanisms into motion <ref type="bibr">(K&#246;rner, 2019)</ref>. Likewise, identifying dehydration and carbon supply thresholds and how long plants can survive on water stores, while losing water to residual conductance from foliage and bark under rising VPD, are large and critical challenges <ref type="bibr">(McDowell et al., 2022)</ref>.</p><p>A new generation of observations should ideally test the many potential mechanisms underlying hydraulic failure and carbon starvation (Table <ref type="table">2</ref>), such as the individual and combined impacts of hypoxia and salinity on root hydraulics, xylem embolism, crown leaf area, and shifts in carbon demand such as for osmoregulation and cellular maintenance (Figure <ref type="figure">4</ref>). To understand the progression and thresholds of hydraulic failure and carbon starvation, and their underlying mechanisms, will require observations of these parameters in dying woody-plants in the field; it could also be investigated using 'point-of-no-return' studies that measure the key mechanisms during and after relief of hypoxia and salinity, including stress to the point of mortality (e.g. <ref type="bibr">Hammond et al., 2019;</ref><ref type="bibr">Sapes et al., 2021)</ref>.</p><p>Partitioning the interdependency of hydraulic failure and carbon starvation will require studies that manipulate the likelihood of one or the other process, for example CO 2 and light starvation to deplete carbohydrate stores, while simultaneously pushing plants to mortality through anomalous inundation <ref type="bibr">(Hartmann et al., 2013;</ref><ref type="bibr">Quirk et al., 2013;</ref><ref type="bibr">Sevanto et al., 2014)</ref>. Finally, inundation with and without elevated salinity should be conducted to tease apart the relative roles of hypoxia and salt on mortality mechanisms.</p><p>Increasing our understanding of coastal woody-plant mortality requires not only advanced measurements but also model development and testing against measurements <ref type="bibr">(Collier et al., 2018;</ref><ref type="bibr">Dietze et al., 2018;</ref><ref type="bibr">Medlyn et al., 2015)</ref>. Models of coastal woodyplant mortality are in their infancy, but significant potential exists for rapid improvement in their mechanistic realism. Models of sealevel driven marsh evolution include feedbacks between flooding frequency and duration, plant growth, and sediment accretion <ref type="bibr">(Fagherazzi et al., 2012;</ref><ref type="bibr">Kirwan et al., 2010;</ref><ref type="bibr">Schile et al., 2014)</ref>.</p><p>However, attempts to couple marsh models with retreating coastal forests still rely on an assumption of passive retreat, where forest mortality depends on the static inundation of topography, with no tree physiology or groundwater dynamics <ref type="bibr">(Carr et al., 2020;</ref><ref type="bibr">Kirwan et al., 2016)</ref>. In contrast to the coastal models, models of upland woody plant mortality include both hydraulic failure and carbon starvation, allowing representation of the feedbacks between drivers, mechanisms, and processes underlying mortality (e.g. <ref type="bibr">Christoffersen et al., 2016;</ref><ref type="bibr">Fisher et al., 2010;</ref><ref type="bibr">Koven et al., 2020;</ref><ref type="bibr">McDowell, Fisher, et al., 2013)</ref>.</p><p>Both coastal and upland models lack many of the mechanisms associated with shoreline woody-plant mortality, such as osmotic constraints on transpiration or salt toxicity to photosynthesis.</p><p>Therefore, regardless of the modeling approach, new development is required to increase the predictive certainty in our model outcomes.</p><p>We propose model development for coastal woody-plant mortality should take advantage of the existing model structures that consider topography while adding the critical drivers and physiological mechanisms of mortality (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>). The responses to both hypoxia and salinity should be incorporated to allow for simulations of both fresh and saline systems. In addition, such models should include plant responses to climate, such as to rising CO 2 and VPD, to capture the multiple impacts of both changing water inundation and changing climate (Figure <ref type="figure">6</ref>).</p><p>Given the rapid rate of change not only in water levels but also in atmospheric CO 2 concentrations, drought, and VPD (IPCC 2021), it is imperative that both measurement and modeling studies of inundation-induced mortality be conducted within the context of changing climate (Figure <ref type="figure">6</ref>; <ref type="bibr">Lovelock et al., 2016;</ref><ref type="bibr">Reef et al., 2015)</ref>.</p><p>In upland systems, rising CO 2 can either increase or decrease the likelihood of mortality, rising VPD increases the likelihood of mortality, and their combination may balance the likelihood of mortality <ref type="bibr">(McDowell et al., 2022)</ref>. We have very limited knowledge of their interactive physiological impacts when combined with increasing hypoxia and salinity, and the impacts of their combinations on mortality per se have not been studied. Application of both empirical and numerical experiments can yield significant new insights into the mechanisms of mortality through testing the hypotheses outlined here.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="9">| CON CLUS IONS</head><p>Woody plant mortality and ghost forest formation are global phenomena in both fresh and seawater systems (Figure <ref type="figure">2</ref>). Our understanding and predictive capacity of coastal woody-plant mortality are both in their infancies; however, the abundance of information from studies of hypoxia and salinity impacts on plants, along with the increasing research focus on ghost forests globally, can enable leaps forward towards better understanding. We hypothesize that both hydraulic failure and carbon starvation emerge from the short and long-term impacts of hypoxia and salinity on plant water uptake, stomatal closure, photosynthetic biochemistry, phloem transport constraints, and crown loss (Figures <ref type="figure">3</ref> and<ref type="figure">4</ref>). The risk of hydraulic failure precedes that of carbon starvation as determined by the rate of embolism formation relative to the rate of declining carbohydrate content, and their respective thresholds for cytrorrhysis. Feedbacks of reduced growth on mortality appear likely, with decreased growth leading to increased likelihood of future inundation, xylem embolism, and reduced whole-plant photosynthesis (Figure <ref type="figure">4</ref>). The mortality processes can be strongly influenced by climatic changes including global rises in CO 2 and VPD, and such interactions should be considered in future research on ghost forest formation (Figure <ref type="figure">6</ref>).</p><p>Modeling mortality processes and their interactions with hypoxia, </p></div></body>
		</text>
</TEI>
