Leaf Physiological Responses and Early Senescence Are Linked to Reflectance Spectra in Salt-Sensitive Coastal Tree Species

Abstract

Salt-sensitive trees in coastal wetlands are dying as forests transition to marsh and open water at a rapid pace. Forested wetlands are experiencing repeated saltwater exposure due to the frequency and severity of climatic events, sea-level rise, and human infrastructure expansion. Understanding the diverse responses of trees to saltwater exposure can help identify taxa that may provide early warning signals of salinity stress in forests at broader scales. To isolate the impacts of saltwater exposure on trees, we performed an experiment to investigate the leaf-level physiology of six tree species when exposed to oligohaline and mesohaline treatments. We found that species exposed to 3–6 parts per thousand (ppt) salinity had idiosyncratic responses of plant performance that were species-specific. Saltwater exposure impacted leaf photochemistry and caused early senescence in Acer rubrum, the most salt-sensitive species tested, but did not cause any impacts on plant water use in treatments with <6 ppt. Interestingly, leaf spectral reflectance was correlated with the operating efficiency of photosystem II (PSII) photochemistry in A. rubrum leaves before leaf physiological processes were impacted by salinity treatments. Our results suggest that the timing and frequency of saltwater intrusion events are likely to be more detrimental to wetland tree performance than salinity concentrations.

1. Introduction

Increased salinization in surface water, soil, and groundwater is leading to increased plant death in forests and agricultural lands in coastal regions of the world [1]. In coastal freshwater wetlands of North America, there have been dramatic increases in dying forests, or ghost forests, which results in altered ecosystem stability and function [2,3,4]. Salinization is a primary cause of plant physiological stress [5,6] and an increasing threat to forest health [7,8,9] that can lead to forests transitioning to shrub–scrub, brackish marsh, or open water systems [2] (Figure 1). Salinization has major implications for the fates of biodiverse wetland plant communities [10,11,12] that maintain ecosystem resilience, especially during periods of ecosystem reorganization following stochastic events like hurricanes and droughts [13,14]. However, the ability of coastal forests to recover from stochastic events decreases as the frequency and severity of disturbance and environmental stressors increase.

Salinization affects the metabolism of many plant species by disrupting normal physiological function [5,15]. For continued growth and reproduction when exposed to abiotic stressors, plants must adjust physiological processes such as plant water status, photosynthetic activity, and carbon allocation [16]. In long-lived plants, like trees, physiological responses to stressors are often temporary but can significantly affect overall plant performance, hindering growth and the ability to compete for resources in stressful environments for years [17]. Our ability to detect a plant’s stress response is dependent upon the type of environmental stressor, or the combination of stressors, and responses are not always consistent across species, time, and/or space. Plant responses to salt stress are well studied, and changes in plant water status and photosynthetic function are often reliable indicators of changes in function due to stress [5,18]. However, salt sensitivity of some well-studied woody plants (e.g., Taxodium distichum (L.) Rich., Pinus taeda L., and Morella cerifera L.) have been found to be variable across species, geographies, and life-history phases [12,17,18,19,20,21,22,23].

Salinity, drought, and flooding are primary environmental stressors in low-lying coastal ecosystems experiencing climate change (e.g., sea-level rise, saltwater intrusion, and frequent storm events) and anthropogenic impacts on hydrology and nutrient balance [1,24]. Overall, the physiological responses of plants to these stressors are similar in many ways. Salt stress disrupts plant water use, alters leaf pigmentation, and inhibits photosystem II, all of which inhibit photosynthesis [5,25]. When combined with flooding, the negative effects of salinity exposure in terrestrial plants are exacerbated [17,18,20,26]. Many trees exposed to oligohaline and mesohaline conditions (~0.5–8 parts per thousand (ppt)) in the porewater or surface water experience osmotic stress and salt toxicity (i.e., chloride, sodium, sulfate) [12,20,27]. If salt exposure persists, plants undergo photoinhibition and photooxidation [28] similar to the physiological responses during drought [29]. Dissolved salts reduce the osmotic potential of water and can induce water stress in plants, resulting in changes in stomatal conductance, which cause leaf wilting and/or stem embolism [20,30]. In forested wetlands with high tree density, water stress also reduces transpiration [7] and increases greenhouse gas emissions [31]. To prevent irreversible damage to photosynthetic machinery, plants can avoid salt stress by closing stomata to reduce water use or through ionic uptake and/or excretion of salts through photosynthetic leaf tissue [17]. Plant responses to salt exposure are dependent on prior exposure and species-specific levels of plasticity [32]. However, there are limits (thresholds) to a plant’s ability to adjust to stress, which in turn has broader implications for transitions occurring in wetland plant communities [33,34,35].

Salt-stressed plants use light differently. To avoid photoinhibition and photooxidation, chlorophyll molecules in the leaf normally compensate for excess light energy in three different ways: (1) by converting it to photochemical energy via photosynthesis, (2) by releasing it as heat in non-photochemical quenching (NPQ) by photosystem II, or (3) by emitting it as fluorescence [36,37]. Shifts in all three light-use strategies to indicate stress responses in plants have been detected using leaf-level measurements of chlorophyll fluorescence [38]. In combination with analyses of leaf gas exchange and water potential, we can gain detailed insight into how plants metabolize and use water differently under stress. In general, it can be difficult to diagnose what stressors are affecting a plant or an ecosystem, but fortunately, there are non-invasive methods for detecting plant stress at the leaf-level, local, or regional scales [8]. Some of these methods can even target specific stress responses related to leaf photosynthesis, leaf pigmentation, biochemistry, and water content. When combined with controlled mesocosm experiments in the laboratory or greenhouse, we can target specific stressors more easily while controlling the severity of a single stressor. In combination, chlorophyll fluorescence and spectral reflectance indexes have been shown to be reliable indicators of salt stress in coastal tree and shrub species M. cerifera and Baccharis halimifolia L. [39,40].

The goal of this study was to assess the sensitivity of coastal tree species to seawater using multiple different potential indicators of stress. We investigated whole plant performance, productivity, water use, leaf physiology, and leaf fluorescence [41,42,43] in dominant coastal tree species that we would expect to have a range of sensitivities to soil salinity [10] when exposed to pulses of seawater. Salinity treatments were modeled after prior studies by Poulter et al. [20] and Powell et al. [29] and verified by species-specific threshold change points calculated by Anderson et al. [10]. We sought to investigate the relationship between leaf-level physiological changes and leaf spectral reflectance.

The two primary questions that we sought to answer were the following: (1) Which common taxa found in freshwater forested wetlands of the southeastern U.S. are the most and least sensitive to elevated saltwater exposure? And (2) what physiological responses are reliable indicators of plant performance when exposed to moderate pulses of seawater? To address these questions, we conducted a greenhouse experiment to assess the responses of the following contrasting but co-occurring hardwood (Acer rubrum L., Nyssa sylvatica Marshall, Quercus nigra L.) and conifer (Juniperus virginiana L., Pinus taeda L., Taxodium distichum (L.) Rich.) tree species when exposed to a gradient of seawater concentrations.

2. Materials and Methods

2.1. Experimental Approach: Tree Species and Salinity Treatments

We conducted a six-month mesocosm experiment in the greenhouse from May to October 2018 using 118 tree seedlings to investigate plant performance and leaf physiology of common tree species from the coastal plain of the South Atlantic United States when exposed to seawater. Species responses to pulses of seawater were assessed based on the magnitude of change in performance including plant water status, leaf functional traits, leaf gas exchange and fluorescence, and leaf spectral reflectance. Here, we sought to gain a deeper quantitative understanding of tree sensitivity to low, relevant concentrations of soil salinity found in forested freshwater wetlands in eastern North Carolina experiencing episodic pulses of saltwater through surface and groundwater.

This study focused on six species that are common to fresh and brackish wetlands in the coastal plain of North Carolina and throughout the Southeastern U.S.: Acer rubrum (ACRU), Juniperus virginiana (JUVI), Nyssa sylvatica (NYSY), Pinus taeda (PITA), Quercus nigra (QUNI), and Taxodium distichum (TADI). We chose these species to represent a range of natural history strategies (i.e., growth strategy, successional status, leaf longevity, and structure), societal benefits (i.e., economic work/lumber value, wildlife preference, conservation concern), and species occurrence along a natural salinity gradient at field locations throughout the Albemarle-Pamlico Peninsula in eastern North Carolina. In this region, Anderson et al. [10] found soil sodium (the second most dominant ion in seawater after chloride (Cl⁻)) ranged from 0.1–9.3 ppt, and a salinity threshold for sensitive plant taxa at ~0.3 ppt where a significant portion of species was lost. Five of the six tree species included in this study begin to decrease in their dominance at 0.1–0.3 ppt of sodium (Na⁺). In addition, we selected several tree taxa based on how frequently they were studied for comparison (e.g., T. distichum, P. taeda) [44,45,46] or paucity of knowledge regarding their salinity tolerance (e.g., A. rubrum, J. virginiana, and Q. nigra) [12,46] and the strategies they employ to cope with salinity stress.

All trees were planted as one-year-old bare root seedlings except N. sylvatica and several Q. nigra individuals, which were germinated from seed in October 2017 (Ernst Conservation Seed Inc., Meadville, PA, USA). All other seedlings were germinated from seed in spring of 2017 from various geographic sources (Mellow Marsh, Siler City, NC, USA, and U.S Forest Service, Raleigh, NC, USA). J. virginiana and N. sylvatica seedlings experienced higher rates of mortality from overwintering in cold chambers prior to propagation; therefore, fewer replicates were included in comparison to the other species (18 and 12 total individuals, respectively). Q. nigra also experienced increased mortality, so several replicates were of younger germinates; however, the same age tree from a different source was tracked throughout the experiment. If a basin was missing a replicate of one of the six species, an additional seed-germinated Q. nigra seedling replaced the missing tree. There were not enough individual N. sylvatica trees to include four intermediate salinity treatments (0.5, 1, 5, and 6 ppt).

In February 2018, all seedlings were planted in standard sterile peat-based media (Fafard^® P4 Mix, Sun Gro Horticulture, Agawam, MA, USA) and acclimated in the greenhouse for three months (700 µmol m² s photosynthetically photon flux density (PPFD) for 16 h per day, 27.8 °C day, and 24.4 °C night temperature and 35% relative humidity). Then, all trees were moved to a larger greenhouse once ambient conditions met those of the prior greenhouse before salinity dosing. Trees were exposed to a salinity gradient with eight treatment concentrations, ranging from controls (no salt added) to 6 ppt simulating fresh (oligohaline) to brackish (mesohaline) conditions found in forested wetland in eastern NC [34,47]. Salinity treatments were dosed with the amount of dissolved Instant Ocean^® (Spectrum Brands, Inc., Blacksburg, VA, USA) according to manufacturer’s guidelines in ppt. Instant Ocean^® mimics concentrations of salts and micronutrients in seawater, and here, it was used for salinity treatments to be consistent across replicates and decrease the potential for introducing pollutants not relevant to this study. Actual salinity of the stock solution diluted for each treatment level was analyzed to reach target concentrations after the first dose. The two primary treatments (control and 3 ppt) were set up in a randomized complete block design, each with 8 replicates. Individual, non-replicated trees were grown in intermediate (0.5, 1, and 2 ppt) and high (4, 5, and 6 ppt) salinity saltwater concentrations for regression analyses only and not used for inference of statistical results (Figure 2).

Salt treatments were administered using a simulated groundwater approach—submerging all tree pots into basins inoculated with stock water collected from Timberlake Observatory Wetland Restoration (TOWeR) site in Alligator, NC (35.898095, −76.167048) [48]. Applying the seawater treatments from belowground allowed us to simulate a realistic delay in salt exposure to living belowground tissue via groundwater pathways. All six tree species were grown in each salinity treatment-specific basin (i.e., block). Before treatments were introduced, water chemistry (pH, DOC, ortho-P, NH₄, anions, and cations) was measured on the stock water used as the solvent for each treatment, and reverse osmosis (RO) water was tested for salts prior to use for water replacement in the basins throughout the experiment.

To isolate the influence of salinity on tree physiology, we applied two pulses of seawater and maintained constant root inundation throughout the entire experiment. The first saltwater pulse was applied on 1 May 2018, after trees had established roots and have begun to leaf out (~12 weeks). A second pulse was dosed on 15 August 2018, once saltwater concentrations in the groundwater of salt-treated basins from the first dose were diluted and no longer different across treatments. Water depth in each basin was monitored biweekly to maintain three moisture horizons below the soil surface: dry/unflooded (0–8 cm), intermittent saturated (8–16 cm), and permanently saturated (16–24 cm) (Figure 2). We chose moisture horizons to reflect the annual average water-table depth in research plots at the TOWeR site where we have monitored groundwater levels continuously throughout 2017 and 2018. Our aim was that soils in the upper horizon (0–8 cm) remained permanently dry, the lower horizon (16–24 cm) remained permanently saturated, and the intermittent zone was allowed to dry before rewetting. Any water loss via transpiration and/or evaporation was replaced by adding RO water directly to the treatment basin up to the top of the intermittent horizon (8 cm from the soil surface).

Ambient greenhouse conditions (27.1 +/− 7.23 °C, 80.0 +/− 9.82% relative humidity, 152 +/− 291 PAR) were monitored continuously using in-house sensors (Duke Phytotron) and two temperature/light sensors (Onset Computer Corp. Hobo Pendant^®, Bourne, MA, USA) placed in the greenhouse. Shortly after dosing salt treatments (15 May 2018), 250 g of sterilized leaf mulch was added to the soil surface of each pot to reduce soil surface drying. Starting on June 12, 200 mL of fertilizer (half-strength Hoagland’s solution) was added to the soil surface of each pot every two weeks throughout the entire experiment to reduce nutrient limitation in the soil. Soil temperature and soil moisture were monitored weekly at 12 cm depth from the soil surface (Campbell Scientific HydroSense II with CS655 sensor, Logan, UT, USA).

Conductivity, temperature, and pH of the groundwater in each basin were monitored biweekly using a multi-parameter water sensor (YSI Professional Plus (Yellow Springs, OH, USA) or Eureka Manta2 (Austin, TX, USA)) including before and after water was added to a basin to ensure consistent water chemistry after additions. Monthly water samples were collected from each of the 22 basins, filtered (glass-fiber filters 0.7 μm), and frozen for additional water chemistry analysis. Analyses of anions (Cl⁻, SO₄⁻, NO₃⁻) and cations (Na⁺, Ca⁺, Mg⁺, K⁺) were run on a Dionex ICS 2000 (Dionex Corporation, Sunnyvale, CA, USA), nutrients (NH₄⁺, PO₄⁻) on a Lachat QuickChem 8500 (Lachat Instruments, Loveland, CO, USA), and total nitrogen (TN) and dissolved organic carbon (DOC) on a AS-18 analytical column and Shimadzu TOC-VCPH Analyzer with TNM-1 module and ASI-V autosampler.

Soil conductivity was measured in each individual pot at 5 cm and 10 cm depths using a soil probe (Hanna HI98331 Soil Test™, Woonsocket, RI, USA) pre-calibrated to salinity tested in the field. Soil conductivity was intensively measured every other day during the first 2 weeks and once a week thereafter to identify when plant roots in the upper top 10 cm were fully exposed to the salinity treatments. During the final harvest, soil conductivity was measured three times at each of the three soil depths using the same probe. We also collected bulk soil subsamples composited from all three horizons to measure soil properties and water-extractable chemistry, including % soil moisture, organic matter content, pH (Mettler Toledo DL18 titrator pH probe, Columbus, OH, USA), and anions (Cl⁻, SO₄⁻, PO₄⁻, NO₃⁻) and cations (NH₄⁺, Na⁺, Ca⁺, Mg⁺, K⁺) on a Dionex ICS 2000, and Lachat QuickChem 8500.

2.2. Functional Traits, Leaf Loss, and Biomass

We quantified performance by evaluating aboveground biomass by plant compartment, growth rate, total leaf area, leaf litter fall, multiple leaf functional traits, and belowground biomass by depth and flooding frequency. Functional traits included specific leaf area (SLA), leaf dry matter content (LDMC), and leaf carbon and nitrogen concentrations. Leaf litter was collected whenever fallen leaves were observed or weekly from 19 July, when substantial leaf litter began to accumulate, until 17 October. Leaf litter was dried at 65 °C for a minimum of 72 h and weighted. Mass of leaf litter loss was assessed over time and summed to calculate proportion of leaves lost (relative to the whole-plant leaf mass) and rate of leaf loss (mg leaf dry matter per day). During final harvest, we measured the dead P. taeda leaves (void of chlorophyll) separate from healthy/green leaves since leaves on the terminal branch generally did not drop.

We destructively harvested above- and belowground biomass at the conclusion of the experiment (17–19 October 2018) to assess total productivity for each tree. All replicate trees of each species were processed in a single day. Seedling height and diameter were measured monthly to assess relative growth over the course of the experiment. Height was a measure of each individual tree from the soil surface to the highest photosynthetically active materials and diameter at soil surface/root collar [49]. Biomass was compartmentalized into leaf (excluding petiole), stem, and root mass to calculate total biomass, root-to-shoot ratio and root mass fraction, and total root mass at each of the three soil horizons (i.e., dry, intermittent saturated, and permanently saturated) (Figure 2). All biomass was dried for a minimum of 72 h at 65 °C before weighing. Relative growth rate (RGR) for each tree was calculated as the following:

(H_post − H_pre)/H_pre,

where the difference in tree height of post-treatment (H_post) and pre-treatment (H_pre) represent biomass allocation over the 6 months of the experiment [20]. Stem diameter (D_post and D_pre) was calculated in the same way.

2.3. Osmotic and Leaf Water Potential

Prior to harvesting all plant biomass at the end of the experiment, we measured leaf water potential (Ψ_L in units of mPa) in three fresh leaves of each treatment per species (18 total leaves) at predawn (Ψ_pd) from 05:20–07:10 across two consecutive days using a pressure chamber (PMS Instruments, Albany, OR, USA). We collected an additional leaf from each tree to measure midday leaf water potential (Ψ_md) from 12:00–14:00 to test the difference in plant water status between Ψ_md. and Ψ_pd of each species in response to the gradient of saltwater exposures. Leaves were selected from the upper portion of the tree between the 3rd and 6th leaf position down from the terminal end of the apical meristem [50] to capture mature new growth. Single leaves of J. virginiana and P. taeda were too small to be measured in the pressure chamber, so we measured Ψ_L on terminal branch segments that had not become woody and were comparable in size to leaf samples from other species. Leaves were later dried, weighted, and tallied with total leaf biomass of each tree.

Additionally, we measured osmotic potential (Ψ_π in mPa) of groundwater after dosing saltwater in the groundwater of each treatment basin. We used groundwater concentrations of NaCl (mg/L) two hours after the second dosing on August 15 to calculate Ψ_π using the Van’t Hoff equation:

Ψ_π = −iCRT,

where i is 2 (for NaCl), C is concentration of NaCl in molarity, R is gas constant (8.32 × 10⁻³ MPa L mol⁻¹), and T is temperature in Kelvin. Treatment-specific Ψ_π was compared to Ψ_pd, which is a proxy for soil water potential [51] to assess species-specific responses to the expected osmotic potential of the solute concentrations. From this, we estimated the position along the salinity gradient that each species reached equilibrium (Ψ_π − Ψ_pd = 0). Equilibrium is the point where water would stop flowing from the soil into the roots, in this case, driven by increasing Ψ_π, or soil salinity high on the gradient. It is beyond this point where we would expect salinity stress to have major implications on plant function, behaving much like drought during low water supply to the roots.

2.4. Leaf Gas Exchange, Chlorophyll Fluorescence, and Spectral Reflectance

Once a month, leaf photosynthesis (A_net μmol CO₂ m⁻² s⁻¹), stomatal conductance to H₂O vapor (g_wv mol H₂O m⁻² s⁻¹), and chlorophyll fluorescence were measured simultaneously using an open gas exchange system (Li-Cor 6400; Li-Cor Biosciences, Lincoln, NE, USA) mounted with a pulse-amplitude modulated fluorometer chamber (Li-Cor 6400-40). One mature sun leaf (5–7 leaves down from the terminal end) was selected per tree (a different leaf each month) with parameters of light intensity set to 700 µmol m⁻² s⁻¹, block temperature 25 °C, and sample CO₂ at 400 µmol m⁻² s⁻¹ following the setup procedure used by [52]. Once the sample chamber reached equilibrium, five evenly spaced samples were logged for each leaf and averaged for analysis. Sample relative humidity in the chamber was maintained at ambient concentrations (mean = 61%, s.d. = 10.4%). If desiccant scrubber was not able to regulate chamber humidity to expected levels, sampling was discontinued, and desiccant was changed before next use.

We measured carbon-13 (¹³C) isotopes in leaves of four of the six tree species (A. rubrum, J. virginiana, P. taeda, and T. distichum) as a measure of intrinsic water-use efficiency (δ¹³C) to assess if leaves exposed to increased salinity of 3 ppt were discriminating against the lighter carbon isotope, ¹²C, due to stomatal closure with increase salt stress [53]. Of the five leaves used to measure leaf functional traits, three were ground and analyzed for carbon isotope ¹³C on a Carlo Erba NA 1500 Elemental Analyzer (Thermo Scientific, Waltham, MA, USA), fitted with a Costech zero-blank autosampler (Duke Environmental Stable Isotope Laboratory, Durham, NC, USA). The methodology we used for sample collection and processing of all traits listed above follows Pérez-Harguindeguy et al. [49].

Fluorescence measurements were taken on light-adapted leaves across all species and treatments simultaneously with gas exchange. Three randomly selected leaves from three individual trees were selected per species for fluorescence of dark-adapted leaves for a minimum of 30 min using dark-adapting leaf clips (LI-COR Biosciences, Lincoln, NE, USA). We used light-adapted leaf fluorescence measurements to calculate operating efficiency of PSII photochemistry (ΦPSII) in light-adapted leaves, which measures a proportion of absorbed light used in PSII photochemistry [36,37,38] and used as an indicator of salinity stress [27,43]. ΦPSII is calculated as

ΦPSII = F’m − F’/F’m,

where F’m is the maximum fluorescence following a light pulse in a light-adapted leaf, and F’ is the steady-state fluorescence of that leaf after a waiting period when fluorescence can reach equilibrium again.

We measured leaf reflectance in the visible and near-infrared spectrum (400–1100 nm) monthly using a spectroradiometer (Apogee PS-100, Apogee Instruments Inc., Logan, UT, USA) across species and all eight salinity treatments. After calibrating the spectroradiometer to a white and black standard (Apogee AS-004), a fixed light source was held at 45° from the leaf/leaves surface with a reflectance probe (Apogee AS-003, Apogee Instruments Inc.). We measured the leaf reflectance of five leaves per individual tree by logging ten measurements per leaf every three seconds. Later, we generated a reflectance curve by averaging the five leaves per tree and then across salinity treatment. We extracted wave lengths from the reflectance data in R [54] to calculate normalized difference vegetation index (NDVI) [55] and photochemical reflectance index (PRI) [56]:

NDVI = (800 − 680)/(800 + 680), and

PRI = (570 − 531)/(570 + 531),

where each number is a wavelength in nanometers (nm). We also assessed the entire wavelength profile to visually inspect the sensitivity of each curve to changes in soil salinity. We merged reflectance data profiles generated in the Apogee software using the Specalyzer web application and in apogeereader package [57] in R [54].

2.5. Statistical Analysis

We analyzed all soils and water chemistry parameters as response variables in generalized linear mixed-effect models (glmer) with the lme4 package [58] in R [54] to determine the differences in salinity treatments across all species over time and their interactions. We included basin (each containing all tree species of like salinity treatment) number in the model as a random effect to reduce the influence of basin position in the greenhouse. Generalized linear models (glm) and two-way analysis of variance (ANOVA) with Tukey’s test for mean separation were used to test the significance of the effects of salinity treatment and species on soil nutrients (NO₃⁻, NH₄⁺, and PO₄⁻). Since much of these data were continuous, non-negative, and right-skewed, we used gamma or Gaussian family distributions and compared models with log and identity links using Akaike’s information criteria (AIC). We also used lsmeans as a post-hoc test of the significance of contrasts between each salinity treatment (p ≤ 0.05). All statistics were performed in R [54].

Additionally, we used generalized linear models with an analysis of variance (ANOVA) to test the variation in mean effects of functional traits within and across all salinity treatments for each of the six tree species included in the experiment. Two-way ANOVA (using Bonferroni correction) was used to test for differences (p < 0.05) in leaf traits (e.g., SLA, LDMC) across species and salinity treatments as measured at final harvest. Leaf litter data were analyzed using glmer to test for differences across species and salinity treatments and their interactions over time. Again, we used AIC in all our analyses to compare models using gamma family, log and identity links, and lsmeans to conduct a post-hoc test with Tukey’s adjustment to test significance of contrasts between controls and each salinity treatment. Effect sizes between mean treatments (control and 3 ppt) were calculated using Hedge’s g using cohen.d in effsize package, and effect sizes were categorized as small (|g| < 0.5), medium (|g| = 0.5–0.8), or large (|g| > 0.8) effect sizes. All data visualizations were generated using the ggplot2 [59] package in R Studio [60].

3. Results

3.1. Groundwater and Soil Chemistry

We achieved significant separation of salinity across each replicated treatment (controls and 3 ppt) shown by measures of specific conductance and ion concentrations in groundwater (Figure 3 and

Table A1. Chemistry of water drawn from TOWeR restoration wetland and seawater stock concentrations for dosing initial groundwater addition to all 22 treatment basins. Salt concentrations for mean and standard deviation of groundwater in controls and salinity treatments (0.5–6 ppt) after the first pulse dosing on 1 May 2018. All concentrations are presented in mg/L.

	Cl−	SO4−	NO3−N	Na+	K+	Mg+	Ca+	NH4−N	PO4−P
TOWeR	7.78	0.31	0.02	4.22	0.87	0.68	1.82	0.05	<0.01
	0.005	0.005	<0.001	0.001	0.010	0.002	0.010	<0.001	-
Control	23.0	3.7	2.31	2.7	85.2	19.6	49.7	0.26	0.67
	3.45	2.68	1.69	1.42	35.1	11.2	23.8	0.12	0.34
Salinity Stock (0.5–6 ppt)	31,349	5546	14.78	18,308	635	2924	849	-	-
0.5 ppt	493	144	3.79	134	0.69	44.2	105	-	-
	-	-	-	-	-	-	-
1 ppt	1184	248	1.79	335	0.82	69.4	39.7	-	-
	-	-	-	-	-	-	-
2 ppt	2051	396	2.04	599	1.27	105	44.7	-	-
	-	-	-	-	-	-	-
3 ppt	2972	622	2.44	861	2.29	143	73.1	-	-
	572	95.5	0.76	169	0.31	21.5	7.92
4 ppt	3108	615	3.54	889	2.10	173	79.1	-	-
	-	-	-	-	-	-	-
5 ppt	4363	863	4.46	1267	2.88	219	92.4	-	-
	-	-	-	-	-	-	-
6 ppt	5604	1126	4.41	1617	3.75	267	115	-	-
	-	-	-	-	-	-	-

). Aside from periodic overlap between low (0.5–2 ppt) and high (4–6 ppt) non-replicated treatments (e.g., early June and mid-September), the two peaks (~9000–10,000 μs/cm) in conductivity were the days when the salt stock was dosed, with a gradual decline from May 1–August 15 in groundwater salinity following each dosing event (Figure 3). Overall, groundwater temperature and pH (22.4–29.9 °C and 6.0–7.5, respectively) were relatively constant across treatments throughout the course of this experiment with some increase and decrease spikes in all treatments likely due to instrument calibration. The groundwater in the control basins at the end of this experiment (10 October 2018) had PO₄⁺ concentrations of 0.75 +/− 0.09 mg/L in the control basins and 0.70 +/− 0.18 mg/L in the 3 ppt basins. Ammonium (NH₄⁺) concentrations were <0.01 mg/L in both the control and 3 ppt treatment basins.

Soil conductivity increased faster at a 10 cm depth than at a 5 cm depth (Figure 3). Soils reached maximum salinity after ~21 days at 10 cm, while it took approximately three times as many days (after the second salinity pulse) for soils at 5 cm to reach comparable levels (Figure 3). Soil salinity was consistently higher at 10 cm than at 5 cm. Regardless of sampling depth, salinity increased continuously throughout the entire experiment, not showing any signs of two pulse additions dosed to the groundwater (Figure 3). Soil temperature was largely consistent across all treatments but varied week to week as the basin positions within the greenhouse were randomized weekly. Soil moisture was also largely consistent across treatments except that the 0.5 ppt and 6 ppt treatments tended to have a higher soil moisture than all other treatments on 1 May and 30 July, respectively.

Soils from each pot were mixed, subsampled, and analyzed for % soil moisture, organic matter content, pH, and anions (Cl⁻, SO₄⁻, PO₄⁻, NO₃⁻) and cations (NH₄⁺, Na⁺, Ca⁺, Mg⁺, K⁺), shown in

Table A2. Soil chemistry from composited soil subsamples from the final harvest (6 months/October). All values are means and standard deviation within each treatment across all tree species within that treatment. The number of trees sampled (n) is listed under each salinity treatment header.

	Salinity Treatments (Mean + st. dev.)
Soil Parameters	Control (n = 43)	0.5 ppt (n = 5)	1 ppt (n = 5)	2 ppt (n = 5)	3 ppt (n = 45)	4 ppt (n = 5)	5 ppt (n = 5)	6 ppt (n = 5)
pH	6.07 + 0.2	6.09 + 0.2	6.07 + 0.2	6.02 + 0.1	6.01 + 0.1	6.06 +0.2	6.02 + 0.1	6.06 + 0.1
% soil moisture	22.4 + 3.1	22.6 + 4.2	21.9 + 2.6	20.6 + 1.9	22.2 + 2.8	22.1 + 3.4	21.3 + 3.6	23.1 + 5.3
AFDM g/g dry soil	1.05 + 0.3	1.08 + 0.3	0.94 + 0.2	0.95 + 0.2	1.02 + 0.2	0.98 + 0.2	0.98 + 0.3	1.05 + 0.3
Sp. Cond. µs/cm	0.10 + 0.03	0.19 + 0.1	0.32 + 0.1	0.52 + 0.3	0.74 + 0.3	1.04 + 0.4	1.01 + 0.4	1.36 + 0.4
Chloride (Cl−) µg/g	40.9 + 52	178 + 96	360 + 145	588 + 254	896 + 249	1295 + 733	1016 + 89	1710 + 504
Sulfate (SO4−) µg/g	167 + 115	73.9 + 20	233 + 205	172 + 63	215 + 96	230 +191	138 + 71	316 + 116
Sodium (Na−) µg/g	55.5 + 14	126 + 50	229 + 94	336 + 130	502 + 121	703 + 342	609 + 65	956 + 260
Calcium (Ca+) µg/g	49.7 + 24	42.5 + 8.3	161 + 55	143 + 40	151 + 39	161 + 96	110 + 37	192 + 54
Magnesium (Mg+) µg/g	19.6 + 11	16.1 + 4.1	53.1 + 21	56.4 + 15	68.6 + 20	82.2 + 58	53.9 + 16	101 + 35
Potassium (K+) µg/g	85.2 + 35	83.3 + 28	123 + 57	125 + 21	136 + 37	150 + 81	111 + 41	158 + 26
Ammonium (NH4+) mg/g	0.26 + 0.1	0.24 + 0.1	0.21 + 0.1	0.33 + 0.1	0.292 + 0.2	0.46 + 0.2	0.327 + 0.2	0.48 + 0.1
Nitrate (NO3−) µg/g	2.31 + 1.7	1.85 + 0.6	1.9 + 0.3	3.73 + 3.9	2.57 + 1.4	2.11 + 0.7	2.31 + 1.8	2.37 + 1.3
Phosphate (PO4−) mg/g	0.67 + 0.3	0.61 + 0.2	0.68 + 0.4	0.82 + 0.5	0.705 + 0.3	0.66 + 0.4	0.60 + 0.1	0.81 + 0.4

. Interestingly, maximum soil salinity was greater (p < 0.01) in conifers (max = 0.66–0.79 μg Na⁺ g⁻¹) compared to hardwoods (max = 0.42–0.59 μg Na⁺ g⁻¹) (

Table 1. Soil salinity broken down by species. Chloride (Cl⁻) and sodium (Na⁺) concentrations are in bold and highlighted grey to show the elevated concentrations in the conifers compared to the hardwood species. Additional soil salinity statistics are reported in Table A2.

		Cl− mg/L		SO4− mg/L		Na+ mg/L		K+ mg/L		Ca+ mg/L		Mg+ mg/L
	Species	Min	Max	Min	Max	Min	Max	Min	Max	Min	Max	Min	Max
Hardwoods	Nyssa sylvatica	35.8	822	43.8	318	49.5	453	80.1	157	43.3	191	16.3	82.3
	Acer rubrum	18.6	1276	22.7	628.4	36.7	691	23.5	201	26.5	183	7.9	90.0
	Quercus nigra	21.6	1363	50.8	492.2	38.6	712	47.6	194	29.5	251	14.2	109
Conifers	Juniperus virginiana	21.3	2228	72.3	309	36.6	1196	57.6	169	35.5	232	128	121
	Taxodium distichum	22.1	2344	20.7	523	36.2	1294	28.7	183	20.2	255	6.9	156
	Pinus taeda	21.9	2676	62.5	559	35.3	1340	44.6	287.1	25.9	323	9.4	187

). All environmental variables unrelated to salinity treatment (% soil moisture, organic matter content, pH) were not different across treatments. The only unexpected difference across treatments was that the 5 ppt treatment had lower mean salt concentrations compared to the 4 ppt treatment. The 0.5 ppt treatment also had lower concentrations of SO₄⁻, Ca⁺, Mg⁺, and K⁺ compared to the control. These results are also reflected in the soil conductivity measured at various times throughout the experiment (Figure 3).

We analyzed water-extractable soil nutrient concentrations (PO₄⁻, NO₃⁻, and NH₄⁺ g per g soil) across all species binned (control, low, mid, and high) salinity treatments. Soil NH₄⁺ increased in soils with 4–6 ppt salinity (p < 0.001), and soil NO₃⁻ increased when exposed to 3 ppt salinity but not at higher concentrations (p = 0.007) (Figure A1). Interestingly, we found an interaction between species and treatment on NH₄⁺ and NO₃⁻ (p = <0.001). That is, NH₄⁺ and NO₃⁻ concentrations were dependent on what species was growing in the soil. NH₄⁺ increased with 3–6 ppt salinity in Q. nigra and T. distchum (p = <0.006) and with 4–6 ppt salinity in A. rubrum and P. taeda pots (p < 0.007). Soil NO₃⁻ increased in the presence of T. distchum in trees exposed to 3 ppt salinity (p = 0.0001). We found soil NH₄⁺ and NO₃⁻ concentrations in soils exposed to high salinity (4–6 ppt) decreased in J. virginiana soils (p = 0.04 and 0.001, respectively). N. sylvatica soil NH₄⁺ and NO₃⁻ did not have a consistent pattern related to increased salinity; however, increased replication for high salinity concentrations would likely result in a similar outcome to all other increasing species since the effect was marginally significant (p = 0.08).

3.2. Functional Traits, Leaf Loss, and Biomass

Total whole-plant leaf area (LA) decreased in A. rubrum (p = 0.02) and tended to increase in P. taeda (p = 0.06) trees exposed to salinity 4–6 ppt (Figure A2). However, when assessing leaf area by absolute soil Na⁺ concentrations, we found a negative relationship between species and soil Na⁺ in A. rubrum and P. taeda (R² = 0.40 and 0.17, and p = 0.04 and 0.05, respectively). A. rubrum also experienced a significant amount of leaf loss when exposed to 3 ppt (p = 0.037) and 4–6 ppt (p = 0.041) salinity treatments (Figure A3). Leaf loss in A. rubrum was also dependent on the interaction of time and salinity treatment. Acer rubrum trees in high salinity treatments (4–6 ppt) did not lose a large proportion of leaf mass until after the second saltwater pulse (15 August 2018).

Taxodium distichum trees exposed to 4–5 ppt did not experience a significant increase in leaf loss; however, the individuals exposed to 6 ppt salinity experienced a significant amount of leaf loss toward the end of the experiment. In addition, P. taeda experienced mixed responses in leaf loss and increased (p = 0.009) and decreased (p = 0.007) in 3 ppt and 4–6 ppt salinity exposure, respectively. However, it is important to note that P. taeda only lost ~2%–8% of total leaf mass. We observed a significant amount of leaf discoloration and death in salt-exposed P. taeda trees, but many of these leaves were not lost as litter and therefore not collected as litter loss. Dead leaves represented 6.6%–49.5% of total leaves in salinity treatments from trees exposed to 3 ppt and 6 ppt salinity treatments.

Untreated evergreen trees (J. virginiana and P. taeda) had different specific leaf area (SLA) compared to the deciduous species, and A. rubrum and Q. nigra had similar SLA, while N. sylvatica and T. distichum had different SLA (Figure A4). When comparing seawater-treated trees to controls, A. rubrum and T. distichum SLA increased in mid- and high-salinity treatments (p = ≤0.01). N. sylvatica leaves also increased with 4–6 ppt salinity compared to low (3 ppt) salinity exposure (p = 0.049; p = 0.054 compared to controls). J. virginiana was the only species that decreased in SLA (i.e., leaves became smaller and/or thicker) when trees were exposed to ≥3 ppt salinity compared to 0.5–2 ppt exposed trees (p = 0.046). A. rubrum, J. virginiana, Q. nigra, and T. distichum all had unique leaf dry matter content (LDMC mg g⁻¹ (mg dry leaf mass per g water-saturated fresh leaf mass), p = < 0.0001–0.046,), while N. sylvatica and P. taeda had comparable LDMC. A. rubrum and T. distichum trees exposed to 3–6 ppt and 4–6 ppt salinity increased their LDMC (p = 0.0002–0.0054, and p = 0.0004 respectively). Conversely, the LDMC in Q. nigra decreased with salinity for the 4–6 ppt treatment (p = 0.048).

Leaf percent carbon and nitrogen content varied by species (A. rubrum, J. virginiana, P. taeda, and T. distichum) and salinity treatment (control and 3 ppt). Percent leaf carbon decreased while nitrogen increased in A. rubrum (p = <0.001 and 0.009), and nitrogen increased in T. distichum when exposed to 3 ppt salinity (p = 0.003). P. taeda had the lowest nitrogen content across all four species, but neither leaf carbon nor nitrogen content changed with increased salt exposure.

Differences between the control and 3 ppt treatments for biomass, growth, and functional trait responses are shown as effect sizes (Figure 4) and actual mean values (

Table A3. Tree species means for all plant performance parameters shown, including leaf mass and area (LM and LA), specific leaf area (SLA), leaf dry matter content (LDMC), leaf percent carbon and nitrogen (%C and %N), leaf carbon–nitrogen ratio (C:N), aboveground biomass (AGB), belowground biomass (BGB), height and diameter growth rates (HGR and DGR), and root-to-shoot ratio/root mass fraction (RSR).

		ACRU		NYSY		QUNI		JUVI		PITA		TADI
	Treatment	Mean	s.d.	Mean	s.d.	Mean	s.d.	Mean	s.d.	Mean	s.d.	Mean	s.d.
AGB g	Control	26.9	6.31	12.9	4.83	19.2	10.2	29.7	14.9	36.5	8.71	54.2	14.7
	3 ppt	26.2	10.3	22.5	7.44	14.2	7.03	32.5	20.7	35.9	10.1	51.3	8.56
BGB g	Control	52.4	14.5	15.6	5.47	34.5	17.7	17.3	10.9	27.2	9.13	67.0	15.4
	3 ppt	38.2	21.0	29.9	9.23	31.3	15.9	14.8	10.2	26.4	5.39	70.5	16.3
LEAF MASS g	Control	9.96	2.70	5.04	1.67	6.50	3.03	17.3	7.41	24.5	6.06	14.6	2.81
	3 ppt	12.4	2.97	7.78	2.43	5.45	2.63	20.7	12.1	23.0	7.29	13.8	3.27
LEAF AREA mm2	Control	1206	258	845	231	777	353	974	251	1598	531	2268	590
	3 ppt	1050	800	1200	403	664	293	1169	553	1319	23	2357	768
HGR cm/day	Control	0.50	0.64	4.27	2.44	1.09	1.74	0.87	1.34	0.80	0.94	0.41	0.71
	3 ppt	0.27	0.56	4.27	3.75	0.61	0.62	0.96	0.93	0.66	0.86	0.41	0.73
DGR mm/day	Control	0.83	0.38	2.56	1.82	1.43	1.32	0.93	0.32	0.46	0.28	0.51	0.23
	3 ppt	0.41	0.27	2.18	0.84	1.50	1.96	1.22	0.39	0.72	0.20	0.52	0.27
ROOT: SHOOT	Control	3.17	1.20	2.05	0.61	2.89	0.70	1.44	0.34	2.15	0.75	1.84	0.69
	3 ppt	2.02	1.06	2.07	0.45	4.01	1.24	1.30	0.60	1.94	0.59	1.98	0.74
SLA mm2/mg	Control	135	15.9	174	20.0	124	21.9	59.8	11.9	70.1	21.2	158	21.3
	3 ppt	172	34.0	175	37.2	128	18.8	61.6	13.1	65.8	16.9	175	27.5
LDMC mg/g	Control	460	36.3	385	30.1	536	43.8	418	47.9	399	34.5	374	25.2
	3 ppt	503	84.7	385	48.8	526	44.2	423	52.7	409	53.3	368	30.9
%C	Control	46.3	0.73	-	-	-	-	48.4	0.94	49.0	1.41	48.7	0.79
	3 ppt	44.1	3.69	-	-	-	-	48.0	0.59	47.8	0.94	49.3	1.30
%N	Control	0.94	0.29	-	-	-	-	0.75	0.06	0.49	0.09	1.07	0.06
	3 ppt	1.20	0.34	-	-	-	-	0.74	0.14	0.46	0.07	1.35	0.24
C:N	Control	53.4	14.1	-	-	-	-	64.9	5.45	104	17.0	45.6	2.12
	3 ppt	41.2	11.6	-	-	-	-	67.6	14.7	106	12.0	37.6	5.69

). Of the twelve performance parameters we measured in this study, A. rubrum showed seven mean decreases (BGB, LM, LA, DGR, RSR, leaf %C, and C:N) and two increases (leaf %N and SLA) in 3 ppt salinity treated plants compared the controls. T. distichum also had a mean increase in leaf %N and a mean decrease in leaf C:N like A. rubrum.

3.3. Osmotic and Leaf Water Potential

We found significant effects of species (F = 11.6 and, p = <0.001) and salinity treatment (F = 5.4, p = <0.001) on predawn water potential (Ψ_pd). Leaf Ψ_pd in A. rubrum was less negative than Q. nigra, J. virginiana, and P. taeda (p = <0.03). Both A. rubrum and T. distichum were closest to equilibrium (Ψ_π − Ψ_pd = 0) and gradually declined toward zero as salinity treatment and Ψ_π increased (Figure 5 and Figure A5). Conversely, J. virginiana leaf Ψ_pd was more negative than A. rubrum and T. distichum (p < 0.021) and did not change at any salinity/solute osmotic potential. Overall, with salinity exposure < 6 ppt, Ψ_pd increased, and Ψ_π decreased with increased salinity; however, water transport was maintained in all species except for P. taeda, where equilibrium was reached at Ψ_π −0.35 MPa. This is where water ceases to move from the soil into the roots (Figure 5). Hardwoods did not reach equilibrium, and Ψ_π was never less negative than Ψ_pd.

Midday leaf water potential (Ψ_md) was different across species (F = 18.4, p = < 0.001) but not statistically affected by salinity (F = 0.674, p = 0.724). The difference between midday and predawn potential (Ψ_md − Ψ_pd) was also different across species (F = 13.54, p = < 0.001) but not affected by salinity. Interestingly, P. taeda treated with 4 ppt salinity (Figure 5) did cross zero/equilibrium (Ψ_md = Ψ_pd, i.e., Ψ_md − Ψ_pd = 0 MPa) around Ψ_π = −0.2 MPa Ψ_π where we expect water transport would stop midday due to salinity or drought since Ψ_pd was quite low at around −0.8 MPa. However, diurnal water transport continued in all other treatments.

3.4. Leaf Gas Exchange and Chlorophyll Fluorescence

There was an interaction between the species and grouped salinity treatment for photosynthesis (A_net) rate (p = 0.013); however, when we tested the main effects of species and salinity treatment (control and 3 ppt), species was a significant predictor of photosynthesis over time (p = <0.001), but salinity was not (p = 0.75). We found that each species’ responses differed among salinity treatments. A. rubrum, T. distichum, and N. sylvatica all had decreases in the mean photosynthesis rate in the salinity treatments compared to the controls as the experiment progressed (Figure A6). A. rubrum photosynthesis decreased when exposed to 3 ppt (p = 0.005) and marginally to 4–6 ppt salinity in October (p = 0.083) and July (p = 0.08). N. sylvatica photosynthesis decreased in August when exposed to 3 ppt salinity (p = 0.008), as did T. distichum in September, but in low salinity (0.5–2 ppt) (p = 0.03).

We did not find any decreases in stomatal conductance to H₂O vapor (g_wv) across all species exposed to 0.5–6 ppt salinity exposure (Figure A7). However, we did find an increase in g_w in the T. distichum exposed to 4–6 ppt salinity in September (p = 0.04) and 3 ppt salinity in October (p = 0.016). Additionally, we observed a marginal increase in A. rubrum exposed to 3–6 ppt salinity (p = 0.07–0.08). We compared the linear relationship between δ¹³C (the difference between ¹³C and ¹²C ratios between a sample and a standard) and C_i/C_a (the ratio of internal leaf CO₂ and ambient CO₂ in the air) to detect increased enrichment in the leaves of plants exposed to 3 ppt salinity. We found no significant linear relationships across treatment within all four species tested (A. rubrum, J. virginiana, P. taeda, and T. distichum).

Like photosynthesis, we found a significant interaction between salinity and month on the operating efficiency (in light-adapted leaves) of photosystem II (ΦPSII) (p = <0.001). However, the response of ΦPSII to salinity was not consistent across all species compared to photosynthesis (Figure A8). In July, ΦPSII in A. rubrum decreased when the trees were exposed to 3 ppt salinity and again in August and October in salinity 3–6 ppt; however, these decreases were marginally significant (p = 0.05–0.09). A_net and g_wv are highly correlated with leaf physiological processes in terrestrial plants. We found a moderate yet highly significant correlation between A_net and g_wv across all species, salinity treatments, and sample dates throughout this study (R² = 0.52, p = <0.001) (Figure A9). However, when testing the correlation of A_net and g_wv in the control and 3 ppt treatments of A. rubrum alone, the relationship in controls (R² = 0.81, p = <0.001) had a remarkably consistent yet stronger linear relationship compared to the trajectory across species and treatments. Although the linear trends in A. rubrum exposed to 3 ppt salinity were less strong (R² = 0.59, p = <0.001), and the slope of the linear relationship between A_net and g_wv only increased slightly (m = 67.5) compared to the controls (m = 67.1); however, these relationships were not specifically tested. However, the y-intercept decreased as the salt-exposed plants began to decrease photosynthesis but not in g_wv.

3.5. Spectral Reflectance

The reflectance spectra in A. rubrum match the expected curvature across the entire visible spectrum, visibly changing slope with increased salinity (Figure 6). We found an interaction between species, salinity treatment, and month (June, September, and October 2018) on PRI and NDVI (p = <0.001) (Figure A10). The PRI decreased in 3, 4, and 6 ppt salinity-exposed trees in June (−0.025, 0.004, and −0.011 respectively), September (−0.076, −0.127, and −0.141), and October (−0.080, −0.080, and −0.149) (p = 0.001–0.04). We also found that NDVI decreased in A. rubrum leaves when exposed to 3, 4, and 6 ppt salinity in September (0.69, 0.41, and 0.45, respectively) and October (0.67, 0.51, and 0.38) (p = <0.001) but not in June (0.84, 0.84, and 0.82). When testing the relationship between PRI and ΦPSII, a non-linear exponential relationship emerged after the second salinity dose in September (F = 7.1, p = 0.015) and became more significant and reflected the salinity gradient treatments more closely in October (F = 8.54, p = 0.009) when compared to June. The NDVI showed the same pattern as PRI.

4. Discussion

This study sought to deepen our understanding of tree species-specific sensitivities to low levels of episodic seawater exposure to improve predictions of non-tidal coastal wetland change at broader scales. We have documented the physiological sensitivity of six tree species to oligohaline and mesohaline (0.5–6 ppt) salinity by evaluating the relationships between various measures of plant physiological responses including leaf water potential, gas exchange, chlorophyll fluorescence, and spectral reflectance. We found that A. rubrum was the most sensitive, and J. virginiana the most tolerant species to moderate levels of soil salinity after two exposures of seawater. Juniperus virginiana trees did not display any responses to ≤6 ppt salinity in Ψ_L, photosynthesis, or chlorophyll fluorescence making it a particularly tolerant species to pulses of salt exposure. J. virginiana is known to be drought and salt-tolerant in a variety of landscapes but is not particularly flood-tolerant [12]. Additionally, there is a subspecies Juniperus virginana var. silicicola (Small) E. Murray, commonly found in sand dunes, marshes, and forests of the Southeastern U.S., and is likely to be even more salt tolerant than J. virginiana var. virginiana [61].

Conversely, A. rubrum, was particularly sensitive to salinity. A. rubrum Ψ_md was not significantly impacted by moderate increases in salinity; however, we found no evidence of a reduction in Ψ_pd in treatments other than the non-replicated tree exposed to 6 ppt salinity and, therefore, maintained water uptake from soil into the roots. Yet, we found that the A. rubrum tree’s leaf functional traits and biomass allocated were impacted more than any other species when exposed to 3 ppt salinity (Figure 4). In addition, A. rubrum photosynthesis decreased after two pulses of seawater at 3–6 ppt within four months of exposure. A. rubrum photosynthesis declined in September and October which coincided with leaf wilt and early leaf senescence in 3–6 ppt salinity treatments. In contrast to J. virginiana, A. rubrum is tolerant to flooding but more sensitive to salt and flooding in combination [62]. Therefore, J. virginiana and A. rubrum may be good indicator species of saltwater stress in freshwater forested wetlands, which reflects what is observed in the coastal plain of North Carolina on the Albemarle-Pamlico Peninsula [10,34]. J. virginiana is often the last tree species scattered throughout non-tidal freshwater and brackish marshes, and A. rubrum trees are not abundant in forested wetlands transitioning to marsh or open water, and seedlings and saplings struggle to survive.

4.1. Salinity Treatments

Our efforts to initiate salinity treatments representative of the targeted saltwater exposure levels, based on the groundwater and soil chemistry, were successful. Even though trees were only dosed with two pulses of saltwater stock, soil conductivity continued to increase throughout the entire experiment, as water lost (via transpiration or evaporation) was replaced with fresh reverse osmosis water (Figure 3).

The goal of our study was to assess the prolonged impacts of salinity pulses. Prior experiments [18] have dosed plants with chronic additions of targeted salt concentrations throughout entire experiments that can quickly accumulate on the soil surface, therefore resulting in reduced growth and physiological activity within 35 days, causing higher mortality rates. This sort of approach causes an accumulation of seawater-derived salts in the soil. In a study conducted by Poulter et al. [20], salt addition of 4 and 8 ppt of P. taeda and Pinus serotina Michx. sapling resulted in complete mortality over 2.5 months. And similarly, Naumann et al. [27] exposed M. cerifera to chronic seawater treatments at 5, 10, and 15 ppt. Salinity exposure at these levels resulted in a rapid (~3–9 days) decline in plant growth and performance at 5 ppt, the complete die-off of leaves, and eventually mortality >10 ppt. Our study highlights the importance of dosing strategy (both concentrations and treatment pathway) when conducting environmentally controlled salinity stress experiments in mesocosms. Additionally, treatments need to be tailored to the species being studied since the maximum soil salinity was greater in conifers compared to hardwoods (Table 1). This result could indicate reduced plant water use in hardwoods, which would dilute available salts in the soil. Alternatively, conifers are more successful at excluding salts from being taken up through the roots [7].

Salt stress in the context of coastal freshwater wetlands is largely an episodic event resulting from storms, droughts, or human disturbance [1]. Although the experimental design is ecologically relevant, we assumed that the susceptibility of earlier growth stages of trees to salinity stress is higher and does not necessarily directly relate to mature trees in situ. However, using seedlings to quantify tree salinity sensitivity is more critical than mature individuals because if young fail to survive salt-stressed conditions, we would not expect recruitment of juvenile trees in forests, resulting in long-term changes in coastal forested wetlands [11].

4.2. Leaf Water Potential, Gas Exchange, Fluorescence, and Senescence

Typically, salinity causes a drop in photosynthesis in salt-sensitive plant species, reducing osmotic potential and stomatal conductance [20]. In our study, P. taeda was the only species where Ψ_pd increased (nearing equilibrium with Ψ_π or Ψ_md) when exposed to 5–6 ppt salinity. Therefore, water transport was maintained in A. rubrum, J. virginiana, N. sylvatica, Q. nigra, and T. distichum when exposed to salinity <6 ppt. However, the A. rubrum tree treated with 6 ppt salinity had already lost all its leaves by the time we measured Ψ_L. We cannot be certain when A. rubrum crossed equilibrium, but our data suggest that A. rubrum is sensitive to salinity >4 ppt since multiple trees exposed to 3 ppt had significant leaf loss. Pinus taeda Ψ_L tended to reach equilibrium (Ψ_π = Ψ_pd) when water transport of all other species was maintained (Figure 5). Interestingly, P. taeda also reduced the whole-plant leaf area with increased salinity, so we may expect P. taeda growth to suffer in years following saltwater pulse events due to lack of water transport and loss of photosynthetic capacity [18].

Although juvenile trees tend to be more sensitive to salinity and less variable in their tolerance to salinity, their sensitivity is a characteristic ideal for studying plant responses to low concentrations (≤6 ppt) of soil salinity [17,63]. We expected more negative impacts of increased salinity treatments on the photosynthesis rate (A_net) in the six species we tested in this study. Yet, we found A. rubrum to have the most instances of decreased photosynthesis, which coincided with the second salinity pulse. It is not surprising that A. rubrum is the most sensitive to salinity [62]; however, few studies have investigated how A. rubrum, a species known for its wide distribution, responds to multiple pulses of low concentrations of seawater that accumulate in coastal soils.

Trees exposed to low to moderate levels of salinity stress (<6 ppt) leave stomata open when access to freshwater is abundant to allow the plant to continue photosynthesizing under stress; however, increased salinity causes leaf wilt and necrosis, which leads to premature senescence of leaves [64]. Any reduction in photosynthetic rates was a result of nonstomatal effects, in this case, early leaf abscission, which may explain high leaf N content in A. rubrum and T. distichum [17]. However, stomatal conductance increases temporarily in T. distichum leaves in September (4–6 ppt) and October (3 ppt).

4.3. Links between Photosystem II and Leaf Reflectance

We found periodic reductions in photosynthesis, ΦPSII, NDVI, and PRI in A. rubrum treated with 3–6 ppt salinity over time. Reflectance spectra in A. rubrum showed the gradual breakdown of pigmentation in the leaves of plants exposed to increased salinity, most significantly in 3–6 ppt salinity treatments (Figure 6). Further investigation of the relationship of NDVI, PRI, and additional remotely sensed indexes (between 400–531, and 570–680 nm) to ΦPSII may show a breakdown in leaf pigmentation. We anticipate the loss of leaf pigmentation was not completely consistent with the loss of functioning in ΦPSII and photosynthesis; however, in a study on M. cerifera, Naumann et al. [40] exposed trees to continuous salinity (0, 5, 10, and 15 g/L) with and without flooding and found a positive linear relationship between PRI and ΦPSII. This suggests that PRI may have detected a decline in carotenoid pigments prior to any decrease in photosynthesis, indicating the beginning of early senescence.

5. Conclusions

This study highlights the diverse responses of wetland tree species to the timing and frequency of saltwater intrusion events that impact plant physiological performance when exposed to pulses of oligohaline and mesohaline salinity concentrations. We found that J. virginana was the most tolerant to salinity pulses ≤6 ppt. In contrast, A. rubrum was the most sensitive to salinity pulses ≤3 ppt, and changes in leaf pigmentation occurred prior to declines in photosynthesis rates late in the growing season. We did not find any adverse effects of increased salinity (≤6 ppt) on water transport in A. rubrum; however, we did find evidence of early leaf senescence. Our initial investigation of leaf reflectance shows promise for linking experimental approaches, used to isolate environmental stressors, to remotely sensed techniques for assessing wetland transition at broader scales. As vast forested wetlands are transitioning to marsh or open water in the coastal plain of the Southeastern U.S., remote sensing is quickly becoming the future of saltwater intrusion research, as high-resolution geospatial data becomes more readily available and accessible. Additional studies on leaf reflectance spectra in coastal tree species exposed to salt stress are needed. Improving our approaches to detect plant physiological change in coastal forested wetlands where sea-level rise and saltwater intrusion are causing wetland ecosystem change over short time frames would support mitigation strategies to climate-driven impacts in coastal forests.