Leaf acclimation to soil flooding and light availability underlies photosynthetic capacity of Lindera melissifolia, an endangered shrub of bottomland forests in the Mississippi Alluvial Valley, USA

Two prominent environmental factors in alluvial floodplains, soil flooding and light availability, interacted to influence leaf characteristics and photosynthetic capacity in Lindera melissifolia. Post-flood acclimation of photosynthetic capacity appears to involve transition from the physiological factors predominant during soil flooding to the structural factors predominant in the absence of flooding.

Abstract

Lindera melissifolia is an endangered shrub indigenous to the broadleaf forest of the Mississippi Alluvial Valley (MAV). In this region, extant colonies of the species are found in periodically ponded habitats where a diversity of broadleaf trees can form well-developed overstory and sub-canopies—these habitat characteristics suggest that soil flooding and light availability are primary drivers of L. melissifolia ecophysiology. To understand how these two factors affect its photosynthetic capacity, we quantified leaf characteristics and photosynthetic response of plants grown in a large-scaled, field setting of three distinct soil flooding levels (no flood, 0 day; short-term flood, 45 days; and extended flood, 90 days) each containing three distinct light availability levels (high light, 30% shade cloth; intermediate light, 63% shade cloth; and low light, 95% shade cloth). Lindera melissifolia leaves showed marked plasticity to interacting effects of flooding and light with lamina mass per unit area (L_m/a) varying 78% and total nitrogen content per unit area (N_a) varying 63% from the maximum. Photosynthetic capacity (A_1800-a) ranged 123% increasing linearly with N_a from low to high light. Extended flooding decreased the slope of this relationship 99% through a reduction in N availability and metabolic depression of A_1800-a relative to N_a. However, neither soil flooding nor light imposed an additive limitation on photosynthetic capacity when the other factor was at its most stressful level, and the A_1800-a–N_a relationship for plants that experienced short-term flooding suggested post-flood acclimation in photosynthetic capacity was approaching the maximal level under respective light environments. Our findings provide evidence for wide plasticity and acclimation potential of L. melissifolia photosynthetic capacity, which supports active habitat management, such as manipulation of stand structure for improved understory light environments, to benefit long-term conservation of the species in the MAV.

Introduction

The Mississippi Alluvial Valley (MAV), a 10-million-ha geographic region surrounding the lower Mississippi River in the south-central United States, holds global significance for ecologically sustainable management and conservation of natural resources (Stanturf et al., 2000). The fertile alluvial soils and temperate climate of the region support species-rich, deciduous broadleaf forests that are foundational to terrestrial and aquatic systems extending well beyond the margins of the MAV (Gardiner, 2014; Miranda et al., 2021). The region, ideal for row crop production, has historically supported a robust, globally prominent agricultural economy (USDA Economic Research Service, 2015; Olson and Speidel, 2021). Successes in developing and protecting the MAV for agricultural production, however, have been at the expense of forest conservation such that indigenous bottomland forest extent has declined 70% (Gardiner, 2014).

Agricultural development and the subsequent forest lost in the MAV have produced consequences spanning several environmental and ecological fronts. Degradation of soil quality by cultivation and the resulting impacts of agricultural runoff to water quality have been found in local floodplain catchments (Bhattacharya et al., 2016) and downstream of the region into the Gulf of Mexico (Rabotyagov et al., 2014). Groundwater extraction for crop irrigation has surpassed aquifer recharge rates thereby lowering water tables and altering seasonal dynamics of the hydrologic cycle (Dakhlalla et al., 2016). Terrestrial carbon pools have been exhausted as contemporary land use practices continue to suppress carbon storage functions in the MAV (Hanberry et al., 2015). Habitat available to native flora and fauna has been reduced and critically degraded—impacts most obvious for fauna requiring expansive tracts of particular forest environments (Buehler et al., 2008) or habitats created by functionally connected river-floodplain systems (Dembkowski and Miranda, 2011). The extensive deforestation has predictably led to fragmentation of remaining MAV forests (Rudis, 1995). This fragmentation has increased vulnerability to external threats and exacerbated degradation of remnant forests through reduced diversity of plant communities across the landscape, increased influx of invasive plant species, elevated susceptibility to forest decline and dieback syndromes, and persistent application of misguided and ineffective forest management activities (Rudis, 1995; Gardiner, 2014).

Deforestation, alteration of floodplain hydrologic regimes and forest degradation have been implicated in the imperilment of L. melissifolia Lauraceae (pondberry), a shrub indigenous to the MAV (US Fish and Wildlife Service, 1993; Devall, 2013). In this region, L. melissifolia forms clonal colonies in the understory of broadleaf forests, typically where floodplain geomorphology and subsurface soil characteristics lead to microsites of ponded water during winter and spring months (Devall et al., 2001; Hawkins et al., 2010). Botanists historically have noted L. melissifolia as a rare plant (Steyermark, 1949), so directly linking human activities to the current population status of this species is tenuous. However, it has been reasoned that MAV land use has limited the availability of suitable habitats (Carter et al., 1990), contributed to colony isolation in fragmented forests (Devall, 2013), and heightened vulnerability of existing colonies to external threats, including the non-native fungal pathogen Raffaelea lauricola (Fraedrich et al., 2011). Accordingly, the US Fish and Wildlife Service listed L. melissifolia as federally endangered to highlight the need to thwart risks of habitat loss and degradation that could further jeopardize the persistence of this species (US Fish and Wildlife Service, 1993; Devall, 2013).

Life history accounts of L. melissifolia in the MAV routinely have highlighted a clear association between colony occurrence and recognizable commonalities among site factors of habitats (Morgan, 1983; Devall, 2013). Environmental conditions associated with these common site factors have been interpreted to represent life cycle requirements of the species. In the MAV, for example, L. melissifolia colonies typically occur where surface depressions are underlain with soil that exhibits poor internal drainage (Priest and Wright, 1991; Devall, 2013). During winter and early spring periods of high precipitation and low evapotranspiration, rainfall accumulates to form a perched water table that maintains inundation or ponding within surface depressions (Priest and Wright, 1991). Some workers have viewed this ability of L. melissifolia to colonize and proliferate into colonies of several hundred stems on these seasonally flooded sites as evidence that soil flooding is a biological requisite of the species (US Fish and Wildlife Service, 1993). Others held that this correlation with soil flooding may not be a biological requirement of the L. melissifolia life cycle but it likely allows the species to thrive in areas where other species cannot because they lack comparable flood tolerance (Wright, 1990; Devall, 2013; Lockhart et al., 2013).

Another commonality among L. melissifolia colonies in the MAV is their occurrence in the understory of mature, structurally complex forests (Hawkins et al., 2009). Because these forests characteristically have a closed overstory canopy and a developed midstory that heavily shade the understory, Tucker (1984) considered L. melissifolia a shade dependent species that flourishes best in environments of “considerable shade”. Tucker’s (1984) interpretation conflicts with physiological models of growth and reproductive performance for shade-tolerant woody plants and our broader field observations and experimentation (Lockhart et al., 2013; Lockhart et al., 2017; Lockhart et al., 2018; Leininger et al., 2021). Oliver et al. (2005) explained that stand development patterns in MAV bottomland forests naturally advance toward structural complexity and canopy closure. Thus, occurrence of L. melissifolia in heavily shaded environments is not biologically obligatory; rather, it reflects morphological and physiological plasticity to the light environment that confers acclimation and shade tolerance (Lockhart et al., 2013; Lockhart et al., 2017).

The aforementioned habitat characteristics suggest that soil flooding and light availability interact as primary drivers of L. melissifolia ecophysiology and persistence in MAV bottomland environments. The purpose of this work was to assess through controlled experimentation the interaction of soil flooding and light availability as regulators of photosynthetic acclimation of L. melissifolia, particularly acclimation of photosynthetic capacity. We quantified leaf characteristics and photosynthetic response of L. melissifolia shrubs that were subjected to imposed gradients of soil flooding and light availability. Quantifying acclimation of the L. melissifolia photosynthetic mechanism to these environmental factors revealed information directly relevant to management of forest stand structure to benefit growth and reproduction of this plant. Thus, we interpret the implications of findings from our research to inform development of improved management approaches that target progressive conservation of this imperiled species.

Materials and Methods

Experiment establishment

This research was established at a field site within the MAV on the Theodore Roosevelt National Wildlife Refuge in Sharkey County, Mississippi, USA (32° 58’N, 90° 44’W). The site has soils developed from alluvium (Sharkey series: very-fine, smectitic, thermic, chromic EPIAQUERTS), and a mild temperate climate (humid, subtropical: mean annual rainfall of 1394 mm and mean annual temperature of 17.8°C) (NOAA, 2010). In 2005, we established an experimental population of L. melissifolia in a complex of 12, 0.4-ha earthen impoundments known as the Flooding Research Facility (FRF) (Lockhart et al., 2006). The FRF was designed to permit random assignment of independent flooding and draining regimes to individual, replicate units on native soil. Additionally, three, 187-m² shade houses—wooden-frames covered with neutral density shade cloth—constructed in each impoundment permitted random assignment and independent control of light availability (Lockhart et al., 2006).

L. melissifolia planting stock was generated with micropropagation techniques (Hawkins et al., 2007) from 20 genotypes collected from MAV populations growing within 2.5 to 100 km of the FRF. Single-stemmed stecklings produced by rooting shoot cuttings from micropropagated plants (Hawkins et al., 2007) were raised for 5 months in 0.98-L containers in a climate-controlled greenhouse (Lockhart et al., 2013). In April 2005, 3456 stecklings were outplanted into native soil within FRF shade houses. Each shade house received 96 stecklings that were randomly assigned a planting position on a 6 × 16 rectangular grid with positions spaced 1.2-m apart (Lockhart et al., 2013). Stecklings averaged 22 cm tall at outplanting (Lockhart et al., 2013). Competition from other plants was eliminated from within shade houses for the duration of the study by hoeing and directed application of glyphosate herbicide.

The FRF was configured according to a split-plot design with three levels of the whole plot factor (soil flooding regime), and three levels of the split-plot factor (light availability regime) (Table 1). Impoundments were randomly assigned a soil flooding level (no flood, short-term flood or extended flood) and shade houses in an impoundment were randomly assigned a light availability level (high light, intermediate light or low light) resulting in nine soil flooding and light availability treatment combinations, which were replicated four times. Outplanted stecklings were established and acclimated to their respective light environments for one growing season before we initiated soil flooding treatment in advance of leaf-out on 1 March 2006.

Table 1.

Soil flooding and light availability treatment levels assigned to experimental units planted in L. melissifolia

Factor	Treatment level
Soil floodinga	No flood = 0 days of soil flooding
	Short-term flood = 45 days of soil flooding
	Extended flood = 90 days of soil flooding
Light availabilityb	High light = 30% shade cloth
	Intermediate light = 63% shade cloth
	Low light = 95% shade cloth

^{a a}Impoundments assigned short-term or extended flood were inundated to 12 cm above the soil surface for the duration of the assigned regime.

^{b b}Quantum sensors maintained in shade houses indicated the high light regime received an annual (2006) photosynthetic photon flux density (PFD) that was 71.5% of ambient sunlight; the intermediate light regime received an annual PFD that was 32.6% of ambient sunlight; and the low light regime received an annual PFD that was 2.2% of ambient sunlight (Lockhart et al., 2013).

Plant sampling

Measurements to quantify the photosynthetic light response (A/PFD) of established L. melissifolia plants were conducted in May 2006 (Figure 1.). This timing allowed for sampling of plants established in soil that did not receive flooding (no flooding), plants established in soil recently drained of floodwater (short-term flooding), and plants established in soil that was presently flooded (extended flooding). Sample plants were chosen from one randomly selected replicate of each whole plot level (soil flooding regime). In each selected whole plot, six plants were randomly selected from each split-plot (light availability regime). Thus, we selected six plants from each of the nine treatment combinations for a total of 54 sample plants.

Figure 1.

Partial view of the Flooding Research Facility capturing a worker preparing to measure photosynthetic light response of L. melissifolia growing in an experimental unit assigned extended flood and high light and showing the arrangement of shade houses within impoundments in the left background (Photo credit USFS files, Brian Lockhart).

The evening before each selected plant was measured it was covered with opaque cloth to limit light incident on the crown for the duration of the measurement; and, in plots absent of floodwater, soil within the dripline of the crown was irrigated to field capacity. These methods ensured sample leaves were hydrated and photosynthetically inactive prior to initiation of A/PFD measurements. We chose sample leaves that were mature, without visible damage, and positioned in the upper crown of the sample plant. CO₂ exchange of sample leaves was measured with a Cirus-2 infrared gas analyzer fitted with an automatic leaf cuvette and programmable light attachment (PP Systems, Amesbury, MA, USA). For each sample leaf, net photosynthesis (A_net) was recorded when CO₂ exchange stabilized at five fixed levels of PFD (0, 1800, 800, 200 and 50 μmol photons m⁻² s⁻¹). Reference gas was maintained at 375 ppm of CO₂ and temperature within the leaf cuvette averaged 30°C during sampling.

Each sample leaf was harvested for laboratory processing immediately upon completion of respective CO₂ exchange measurements. In the laboratory, lamina area (L_a) was measured with a LI-3100 Area Meter (LI-COR Biosciences, Inc., Lincoln, NE, USA), then leaves were dried to constant mass at 70°C before dry mass was measured with an analytical balance. Dried sample leaves were pulverized with a ball mill to pass a 100-mesh sieve (150 μm nominal sieve opening), and total nitrogen (N) concentration of each sample was quantified with a PE 2400 Series II CHNS/O Analyzer (PerkinElmer, Waltham, MA, USA).

Data analysis

Variables of morphology and N composition resulting from sample leaf processing included L_a, lamina mass (L_m), lamina mass per unit area (L_m/a), total N content (N_a: area basis) and total N concentration (N_m: mass basis). We modeled A/PFD and estimated key physiological variables for each sample leaf with Photosyn Assistant software (Dundee Scientific Ltd, Dundee, Scotland, UK), which fits photosynthetic response curves according to methods described by Prioul and Chartier (1977). Variables quantified through A/PFD modeling included dark respiration (R_d), light compensation point (I_c), apparent quantum yield (ϕ), maximum photosynthesis at light saturation (A_max) and saturating light intensity (I_s). Photosynthesis variables were scaled relative to L_a and L_m where appropriate.

Effects of soil flooding, light availability, and their interaction on response variables were determined with Analysis of Variance (ANOVA) for a completely random factorial experiment arranged in a split-plot design. The analysis was structured to handle each sample leaf as an independent sample. Thus, sample size for each of the nine soil flooding and light availability treatment combinations totaled six observations. The ANOVA model held 53 total degrees of freedom (df); 10 df were partitioned in the whole plot error term and 30 df were partitioned in the split-plot error term. We separated means within significant source effects using Least Significant Difference. Simple linear regression was used to model relationships between A measured at a PFD of 1800 μmol photons m⁻² s⁻¹ (A_1800-a: L_a basis; A_1800-m: L_m basis) and L_m/a, N_a and N_m. All analyses were conducted with SAS software version 9.4 (SAS Institute, Inc., Cary, NC USA); an α of 0.05 determined significance of all tests.

Results

Leaf characteristics

Bud break and leaf development initiated about two weeks earlier in low light than in high light. Soil flooding did not appear to impact this process. Full leaf-out across all treatment combinations was completed more than a month in advance of photosynthesis measurements. Soil flooding was not a significant source of variation in L. melissifolia L_a or L_m, but light availability accounted for differences observed for each of these variables (Tables 2 and 3). L_a was inversely related to light availability such that leaves produced by plants in the high light environment were 46% as large as leaves of the low light environment (Tables 2 and 3). Conversely, L_m measured on plants from either high or intermediate light was twice that of plants receiving low light (Tables 2, 3).

Table 2.

Source effect significance for leaf trait and photosynthesis variables measured on L. melissifolia established along gradients of soil flooding and light availability. F-values are noted with * if P < 0.05, ** if P < 0.01 and *** if P < 0.001

Variable	Source
	Soil flooding (df = 2)	Light availability (df = 2)	Soil flooding x light availability (df = 4)
La	0.9	20.1***	1.3
Lm	1.9	11.5***	2.0
Lm/a	30.3***	365.9***	7.3***
Na	11.9***	112.7***	4.8**
Nm	93.7***	164.9***	1.3
Rd	1.2	43.5***	0.3
ϕ	1.0	0.3	1.5
Ic	0.8	24.5***	0.5
Is	11.0**	25.5***	6.0**
Amax-a	70.6***	97.4***	25.5***
Amax-m	56.7***	33.8***	15.3***

Table 3.

Leaf traits (mean ± standard error)^a of L. melissifolia established along gradients of soil flooding and light availability

Variable	Soil flooding regime
	0 days	45 days	90 days	Light regime mean
La (cm2)
High light	18.6 ± 2.1	32.3 ± 4.9	17.9 ± 1.9	22.9 ± 2.4c
Inter. light	39.6 ± 3.3	35.9 ± 7.8	29.5 ± 4.1	34.9 ± 3.1b
Low light	49.8 ± 7.1	47.7 ± 7.9	50.4 ± 4.5	49.3 ± 3.6a
Flooding regime mean	36.0 ± 4.0A	38.6 ± 4.1A	32.6 ± 3.8A
Lm (g)
High light	0.10 ± 0.02	0.22 ± 0.03	0.15 ± 0.02	0.16 ± 0.02a
Inter. light	0.16 ± 0.02	0.17 ± 0.04	0.19 ± 0.03	0.17 ± 0.02a
Low light	0.07 ± 0.01	0.07 ± 0.01	0.09 ± 0.01	0.08 ± 0.01b
Flooding regime mean	0.11 ± 0.01A	0.15 ± 0.02A	0.14 ± 0.01A
Lm/a (g m−2)
High light	52.8 ± 4.9C,a	67.4 ± 2.1B,a	81.4 ± 2.9A,a	67.2 ± 3.4
Inter. light	39.4 ± 3.2B,b	45.2 ± 2.1B,b	62.1 ± 2.3A,b	48.9 ± 2.7
Low light	14.2 ± 1.0A,c	14.8 ± 0.9A,c	19.3 ± 0.9A,c	16.1 ± 0.8
Flooding regime mean	35.5 ± 4.3	42.5 ± 5.3	54.3 ± 6.4
Na (g m−2)
High light	1.80 ± 0.09A,a	1.86 ± 0.05A,a	1.32 ± 0.06B,a	1.67 ± 0.07
Inter. light	1.38 ± 0.10A,b	1.27 ± 0.11A,b	1.01 ± 0.07B,b	1.22 ± 0.06
Low light	0.74 ± 0.07A,c	0.68 ± 0.06A,c	0.75 ± 0.03A,c	0.73 ± 0.03
Flooding regime mean	1.31 ± 0.12	1.27 ± 0.12	1.03 ± 0.06
Nm (mg g−1)
High light	34.9 ± 2.4	27.8 ± 1.3	16.2 ± 0.4	26.3 ± 2.1b
Inter. light	35.1 ± 0.9	27.9 ± 1.8	16.4 ± 1.1	26.5 ± 2.0b
Low light	52.3 ± 1.5	46.3 ± 2.0	39.2 ± 0.9	45.9 ± 1.6a
Flooding regime mean	40.8 ± 2.2A	34.0 ± 2.3B	23.9 ± 2.6C

^{a a}Uppercase letters indicate difference among soil flooding regime means in a row, while lowercase letters indicate difference among light regime means in a column (α = 0.05).

Soil flooding and light availability interacted to affect L_m/a of L. melissifolia leaves (Table 2). L_m/a tended to increase with flood duration for plants grown under high or intermediate light, but not for plants grown under low light (Table 3). Soil flooding and light availability interacted also to influence N_a (Table 2). Extended flooding reduced N_a at least 27% in plants raised under high or intermediate light, while it did not affect N_a in plants raised under low light (Table 3). N_m of L. melissifolia was lowered 17% by short-term flooding, and a further 30% by extended flooding (Tables 2, 3). Plants grown under low light produced leaves with the greatest N_m irrespective of flooding regime (Table 3).

Photosynthetic light response

R_d and I_c of L. melissifolia leaves were unchanged by soil flooding, but light availability produced a response in both variables (Tables 2, 4). Leaves that developed in the high light environment showed R_d rates 38% greater than those in intermediate light, while leaves that developed in the intermediate light environment showed R_d rates 150% greater than those in low light (Table 4). I_c exhibited a response to light availability like that observed for R_d. Neither soil flooding nor light availability influenced ϕ of L. melissifolia (Tables 2, 4). Interaction between treatment effects accounted for significant variation in I_s, A_max-a, and A_max-m (Table 2). I_s was generally lowest for plants experiencing the extended flood, or for plants grown in low light (Table 4). Lindera melissifolia leaves showed the greatest rates of A_max-a when plants were raised under high light in the absence of soil flooding (Table 4). The extended flood reduced A_max-a by more than 63% for plants in environments of high or intermediate light, but did not influence A_max-a of plants in low light (Table 4). Consistent with the response observed for A_max-a, soil flooding inhibited A_max-m when L. melissifolia was grown under high or intermediate light, and this effect was most pronounced under extended flooding (Tables 2, 4). Soil flooding did not impact A_max-m of plants raised under low light (Table 4).

Table 4.

Photosynthesis variables (mean ± standard error)^a of L. melissifolia established along gradients of soil flooding and light environment

Variable	Soil flooding regime
	0 days	45 days	90 days	Light regime mean
Rd (μmol CO2 m−2 s−1)
High light	−1.02 ± 0.08	−0.85 ± 0.13	−0.95 ± 0.05	−0.94 ± 0.05a
Inter. light	−0.76 ± 0.11	−0.63 ± 0.14	−0.66 ± 0.04	−0.68 ± 0.06b
Low light	−0.27 ± 0.07	−0.23 ± 0.07	−0.32 ± 0.04	−0.27 ± 0.03c
Flooding regime mean	−0.68 ± 0.09A	−0.57 ± 0.09A	−0.64 ± 0.07A
ϕ (μmol CO2 (μmol photons)−1)
High light	0.043± 0.004	0.033 ± 0.003	0.037 ± 0.004	0.038 ± 0.002a
Inter. light	0.037 ± 0.002	0.048 ± 0.008	0.029 ± 0.002	0.038 ± 0.003a
Low light	0.025 ± 0.012	0.042 ± 0.012	0.034 ± 0.001	0.034 ± 0.006a
Flooding regime mean	0.035 ± 0.004A	0.041 ± 0.005A	0.033 ± 0.002A
Ic (μmol photons m−2 s−1)
High light	24.8 ± 3.3	28.5 ± 6.2	27.4 ± 2.9	26.9 ± 2.4a
Inter. light	20.3 ± 2.9	16.4 ± 5.7	23.7 ± 2.6	20.2 ± 2.3b
Low light	7.4 ± 0.9	5.7 ± 0.9	9.2 ± 0.8	7.4 ± 0.6c
Flooding regime mean	17.5 ± 2.3A	16.8 ± 3.5A	20.1 ± 2.3A
Is (μmol photons m−2 s−1)
High light	379.0 ± 20.2A,a	447.8 ± 60.7A,a	192.2 ± 23.2B,a	339.7 ± 33.8
Inter. light	358.7 ± 29.3A,a	275.0 ± 61.2AB,b	179.0 ± 12.2B,a	270.9 ± 27.9
Low light	147.6 ± 14.8A,b	129.2 ± 19.9A,c	164.8 ± 9.3A,a	147.2 ± 9.0
Flooding regime mean	295.1 ± 28.1	284.0 ± 42.0	178.7 ± 9.1
Amax-a (μmol CO2 m−2 s−1)
High light	15.2 ± 0.9A,a	12.9 ± 0.4B,a	5.6 ± 0.3C,a	11.2 ± 1.0
Inter. light	12.4 ± 0.7A,b	10.9 ± 1.2A,b	4.4 ± 0.3B,a	9.2 ± 0.9
Low light	4.4 ± 0.3A,c	4.1 ± 0.2A,c	5.3 ± 0.3A,a	4.6 ± 0.2
Flooding regime mean	10.6 ± 1.2	9.3 ± 0.9	5.1 ± 0.2
Amax-m (nmol CO2 g−1 s−1)
High light	299.3 ± 30.2A,a	194.5 ± 9.2B,b	69.0 ± 3.4C,b	187.6 ± 24.9
Inter. light	322.7 ± 21.1A,a	241.0 ± 24.8B,a	70.5 ± 3.9C,b	211.4 ± 27.5
Low light	312.5 ± 17.9A,a	278.7 ± 12.4A,a	274.3 ± 11.4A,a	288.5 ± 8.7
Flooding regime mean	311.5 ± 13.0	238.0 ± 12.4	137.9 ± 23.7

^{a a}Uppercase letters indicate difference among soil flooding regime means in a row, while lowercase letters indicate difference among light regime means in a column (α = 0.05).

Photosynthetic capacity

N_a of L. melissifolia leaves increased with increasing L_m/a (Fig. 2a). Extended flooding depressed the slope in this relationship because leaves with relatively high values for L_m/a, i.e. those in high and intermediate light, showed reduced N_a under this flood regime. In contrast to N_a, N_m decreased with increasing L_m/a (Fig. 2b). Slopes of curves fit to this relationship were similar for each of the three flood regimes, illustrating a decrease in N_m of about 12% for a 10 g m⁻² increase in L_m/a.

Figure 2.

Response of leaf nitrogen to lamina mass per unit area in L. melissifolia. Data are plotted by soil flood and light availability treatment combinations with regression lines fit by soil flooding regime. (a) Relationship between N_a and L_m/a (no flood: r² = 0.88, P < 0.0001; short-term flood: r² = 0.90, P < 0.0001; extended flood: r² = 0.78, P < 0.0001). (b) Relationship between N_m and L_m/a (no flood: r² = 0.74, P < 0.0001; short-term flood: r² = 0.69, P < 0.0001; extended flood: r² = 0.86, P < 0.0001).

A 10 g m⁻² rise in L_m/a elevated A_1800-a about 19% for plants raised absent of soil flooding, and about 17% for plants that experienced the short-term flood (Fig. 3). A_1800-a was not responsive to L_m/a if plants were subjected to the extended flood. A_1800-a showed strong response to N_a, except in plants subjected to the extended flood (Fig. 4a). Plants receiving no flooding or short-term flooding exhibited an increase in A_1800-a of at least 47% per 0.5 g m⁻² rise in N_a. A_1800-m responded positively to N_m if plants experienced short-term soil flooding, and the steepest response (49% rise in A_1800-m for a 10 mg g⁻¹ increase in N_m) was marked by plants that experienced extended flooding (Fig. 4b).

Figure 3.

Response of photosynthesis to lamina mass per unit area in L. melissifolia. Data are plotted by soil flood and light availability treatment combinations with regression lines fit by soil flooding regime. Relationship between A_1800-a and L_m/a (no flood: r² = 0.73, P < 0.0001; short-term flood: r² = 0.72, P < 0.0001; extended flood: r² = 0.08, slope not different from 0, P = 0.26).

Figure 4.

Response of photosynthesis to leaf nitrogen in L. melissifolia. Data are plotted by soil flood and light availability treatment combinations with regression lines fit by soil flooding regime. (a) Relationship between A_1800-a and N_a (no flood: r² = 0.83, P < 0.0001; short-term flood: r² = 0.81, P < 0.0001; extended flood: r² = 0.0001, slope not different from 0, P = 0.96). (b) Relationship between A_1800-m and N_m (no flood: r² = 0.12, slope not different from 0, P = 0.15; short-term flood: r² = 0.52, P = 0.0007; extended flood: r² = 0.97, P < 0.0001).

Discussion

Photosynthetic function has been a focal point of L. melissifolia ecophysiology research since Wright (1990) initiated his field studies. Though few, investigations conducted before the current research provide important records of L. melissifolia photosynthesis across the range of the species (Wright, 1990; Aleric and Kirkman, 2005 ; Lockhart et al., 2017), contributing elemental information on variability of leaf function under a variety of plant and site conditions. But experimental data useful to inform active management to sustain species conservation are lacking, e.g. evaluation of plasticity of the photosynthetic mechanism to important environmental factors, and available research findings appear confounded or conflicting (Aleric and Kirkman, 2005; Lockhart et al., 2017). Our work provides fundamental insight into the L. melissifolia photosynthetic mechanism by showing how photosynthetic capacity acclimates to two interacting environmental factors, soil flooding and light availability, prominent in alluvial floodplain habitats occupied by this species.

Forested floodplains are characteristically dynamic environments with high spatial and temporal variability in abiotic site factors. Many woody plants of temperate floodplains possess plasticity of leaf response that enables acclimation to these dynamic environments. Leaf plasticity is clearly expressed along a light availability gradient by deciduous trees and shrubs of temperate floodplains (McMillen and McClendon, 1983; Niinemets, 1997; Granata et al., 2020), including L. melissifolia (Aleric and Kirkman, 2005; Lockhart et al., 2017). Anatomical, morphological, and physiological acclimation support optimization of photosynthetic capacity under the “chronic” stress of the ambient light environment. Leaf plasticity and acclimation of the photosynthetic mechanism are also evident when edaphic or hydrologic factors impose stress on some woody plants of floodplains (Angelov et al., 1996; Gardiner and Krauss, 2001; Niinemets and Kull, 2003). In cases of flooded or waterlogged soil, which for river floodplains is typically a temporary but “acute” stress that is most deleterious during the growing season, leaf acclimation is primarily physiological (Tatin-Froux et al., 2014). Some well-adapted floodplain species exhibit resilience by maintaining sub-optimal photosynthetic function while soil is waterlogged, then appear to recover to full photosynthetic capacity when aerobic conditions return to the soil environment (Angelov et al., 1996; Anderson and Pezeshki, 2000; Rodríguez et al., 2015). Observations across soil flooding and light availability gradients established in this study reveal that L. melissifolia expresses leaf plasticity through structural and physiological mechanisms that foster photosynthetic acclimation and regulate its photosynthetic capacity.

Plasticity in leaf morphology and N-status

Soil flooding has been linked to a reduction in leaf lamina area and mass of some broadleaf trees and shrubs, including L. melissifolia (Liu and Dickmann, 1992; Lockhart et al., 2017), but observation of these morphological responses for species endemic to temperate floodplains is seldom reported. Reduced lamina size is viewed as indicative of root hypoxia stress and likely results from limitations on cell wall extensibility that inhibit complete lamina expansion (Smit et al., 1989). Observations noted in this report are inconsistent with those previously reported by Lockhart et al. (2017), who found that 90 days of soil flooding reduced L. melissifolia lamina area 35% and reduced lamina mass 22%. Our study shows soil flooding of a similar duration imposed by Lockhart et al. (2017) did not affect L. melissifolia lamina area or mass, and when viewed together with Lockhart et al. (2017) reveals interannual variation in leaf size that may have contributed to the differing responses.

The light environment imposes a more influential role in leaf development of floodplain trees and shrubs than does soil flooding. Like other temperate broadleaves (Koike et al., 2001; Muraoka and Koizumi, 2005; Sack et al., 2006; Granata et al., 2020), L. melissifolia lamina area was inversely related to light availability, while lamina mass showed a positive relationship to the gradient. This plasticity in lamina morphology is associated with acclimation that generally supports efficiencies in light harvesting and CO₂ uptake across the irradiance gradient. Our findings verify other published examinations of L. melissifolia leaf morphology relative to light availability—phenotypic variation in lamina area, particularly, ranges greatly (> 115% from high to low light in the current study) across a gradient of light availability (Lockhart et al., 2012; Lockhart et al., 2017). Thus, the species exhibits wide plasticity for acclimation of gross leaf morphology in simulated understory environments of markedly different light availability.

The observed plasticity in lamina area and mass relative to light availability would be expected to impact other key leaf traits more directly associated with photosynthetic capacity. The ratio of lamina mass to lamina area, i.e. L_m/a, reflects leaf tissue structure and cell composition (John et al., 2017) and provides an informative index of acclimation strongly linked to photosynthetic capacity (Ellsworth and Reich, 1992; Ellsworth and Reich, 1993; Niinemets, 1997; Sack et al., 2006; Niinemets, 2010). Within a species, relatively low L_m/a indicates leaves structured to favor light absorbance in low light environments owing to its association with a thin epidermis, and low palisade and spongy mesophyll volumes (Niinemets, 2010; Villar et al., 2013; Gratani, 2014). Relatively high L_m/a indicates leaves structured to favor CO₂ fixation in high light environments because of its association with a high palisade density and volume and high spongy mesophyll volume (Poorter et al., 2009; Niinemets, 2010; Villar et al., 2013; Gratani, 2014). As such, L_m/a of temperate broadleaf species characteristically decreases along gradients of high to low irradiance (Niinemets, 1997). By contrast, L_m/a reportedly increases during periods of soil flooding (Liu and Dickmann, 1992; Gardiner and Krauss, 2001; Poorter et al., 2009). This increase is not attributed to structural acclimation, but appears to be driven by an accumulation of non-structural carbohydrates caused by interruption of their translocation out of the leaf (Vu and Yelenosky, 1991; Gardiner and Krauss, 2001; Martínez-Alcántara et al., 2012; Sloan et al., 2016). L_m/a of L. melissifolia examined in this study responded to the interacting effects of soil flooding and light availability. L_m/a was greatest under the longest flood duration and highest light availability, but the effect of soil flooding was not evident where light availability was low. This observation appears consistent with the supposition that soil flooding led to an accumulation of non-structural carbohydrates that boosted L_m/a where photosynthesis was not limited by light. Within a flood regime, L_m/a of L. melissifolia leaves ranged more than 270% from low to high light, confirming that development of leaf structure exhibits strong plasticity to light availability (Aleric and Kirkman, 2005; Lockhart et al., 2017), and suggesting this plasticity is a primary mechanism supporting photosynthetic acclimation to the light environment.

Coupled with the observed variation in morphology and structure, we found soil flooding and light availability to affect N content and concentration of L. melissifolia leaves. Foliar N tends to drop, sometimes leading to leaf chlorosis, in woody plants experiencing waterlogged soil. The decline in leaf N results from root hypoxia which can inhibit N uptake and alter N translocation to foliage (DeLaune et al., 1998; Martínez-Alcántara et al., 2012). Additionally, the decline in leaf N is sometimes associated with an increase in L_m/a as was the case in this study (Gardiner and Krauss, 2001). This suggests that it may also be linked, in part, to the previously discussed carbohydrate accumulation thought to have increased L_m/a (Martínez-Alcántara et al., 2012; Sloan et al., 2016; Lockhart et al., 2017). Leaf chlorosis was not evident in L. melissifolia studied in this experiment, even for plants that experienced extended flooding (personal observation by Gardiner). But, extended flooding led to a lower N_a in leaves of high and intermediate light environments, and N_m decreased with flood duration regardless of light regime suggesting that soil flooding likely limited N uptake, translocation to leaves, and/or the N:carbohydrate ratio in leaves.

Consistent with other reports (Ellsworth and Reich, 1992; Niinemets, 1997; Aranda et al., 2004; Lockhart et al., 2017), we found a positive relationship between N_a and light availability that bore a similar pattern to our findings for treatment effects on L_m/a and N_m. Variation in N_a relative to the light environment has been linked to plasticity in lamina morphology and anatomy, and to plasticity in N partitioning that supports efficient photosynthesis under the prevailing light environment. Accordingly, a relatively high N_a is typical of leaves acclimated to light-rich environments and is associated with an increase in the fraction of N partitioned to RuBP carboxylase (rubisco) to sustain high carbon fixation. Respectively, a relatively low N_a is typical of leaves acclimated to light-poor environments and is associated with an increase in the fraction of N partitioned to chlorophyll to enhance light harvesting (Evans, 1989; Ellsworth and Reich, 1992; Niinemets, 1997). N_a status of L. melissifolia in this study varied with soil flooding and light availability consistent with other temperate broadleaves, and implies shifts in N partitioning between rubisco and chlorophyll fractions that support photosynthetic acclimation may have occurred with the environmentally driven variation in N_a and L_m/a.

Plasticity in photosynthetic traits

Variation in leaf traits due to soil flooding and light availability was accompanied by substantial variation in L. melissifolia photosynthesis as was expected from results of previous work (Wright, 1990; Aleric and Kirkman, 2005; Lockhart et al., 2017). Analysis of A/PFD allowed us to partition variation in photosynthetic response to effects of soil flooding and light availability, and to identify which photosynthetic traits exhibit plasticity in response to these environmental factors. Unfortunately, we cannot compare our findings to those of Wright (1990) or Aleric and Kirkman (2005) because soil flooding, which was reportedly present at their respective study sites, was not accounted for in their sampling or analyses. Three traits, R_d, ϕ, and I_c, showed no variation attributable to soil flooding. Reports detailing the impact of flooding stress on these traits are few, but reduced ϕ has been noted as a primary impairment to photosynthesis of flood-stressed, mesophytic shrubs and trees (Davies and Flore, 1986; Gardiner and Krauss, 2001). The ability to maintain ϕ during seasonal soil flooding events appears significant in the adaptiveness of L. melissifolia to floodplain habitats and warrants further ecophysiological investigation.

In contrast to soil flooding, light availability influenced R_d and I_c. Both traits ranged more than 70% and were least where leaves developed in low light. These observations are consistent with photosynthetic acclimation to light availability reported for some temperate broadleaf species (Ellsworth and Reich, 1993; Kubiske and Pregitzer, 1996; Oguchi et al., 2005; Granata et al., 2020). Ranges in R_d and I_c are often driven by the extent of adjustment in L_m/a across the irradiance gradient (Mitchell et al., 1999; Lusk and Reich, 2000).

Soil flooding and light availability effects on L. melissifolia were most impactful to photosynthetic traits indicative of leaves functioning at maximal CO₂ uptake and fixation capacity, i.e. I_s, A_max-a, and A_max-m. Calculation of a simple index (Valladares et al., 2002) quantifies plasticity in A_max-a at 0.69 and 0.66 for the soil flooding and light availability effects, respectively. Though flooding obviously constrains photosynthesis in most temperate floodplain broadleaves, including L. melissifolia, (Dreyer, 1994; Anderson and Pezeshki, 2000; Li et al., 2015; Lockhart et al., 2017), the literature remains ambiguous concerning how this stress impacts photosynthetic traits derived from A/PFD response curves (Gardiner and Krauss, 2001). With respect to light availability, plasticity observed for A_max-a is within the range of plasticity reported for other temperate broadleaves (Valladares et al., 2002). Broadleaf species for which light availability alters leaf morphology generally show comparatively low values for I_s and A_max-a under the stress of low light, while values for A_max-m tend to covary with shade tolerance and successional status (Bazzaz and Carlson, 1982; Walters et al., 1993; Niinemets and Valladares, 2004). Much less is known about the combined effects of the two environmental stresses on expression of these photosynthetic traits in temperate trees and shrubs. We found variation in response by L. melissifolia for each of these traits to be attributable to interacting effects of the environmental factors, but the effects were not additive at the most stressful level for each factor, i.e. they did not appear amplified (sensuLenssen et al., 2003). In other words, soil flooding did not co-limit A_max-a if L. melissifolia was raised under low light, nor did light availability co-limit A_max-a if plants received extended flooding.

Acclimation of photosynthetic capacity to soil flooding and light availability

Development of photosynthetic capacity in temperate broadleaves is inextricably linked to leaf N status because leaf N is predominantly held in proteins supporting light harvesting, electron transport, and carbon fixation functions (Evans, 1989; Reich et al., 1995). Accordingly, photosynthetic capacity responds to variation in leaf N content and allocation brought about by leaf acclimation, particularly to the light environment, that affects leaf anatomy and structure, leaf N concentration, or N partitioning among light harvesting and carbon fixation functions (Niinemets, 1997; Aranda et al., 2004; Kitaoka and Koike, 2005; Delagrange, 2011). In this study, photosynthetic capacity (A_1800-a) of L. melissifolia showed significant acclimation to environmental gradients by ranging about 121% across soil flooding regimes (extended flood to no flood) and about 123% across light environments (low light to high light). This suggests the species is capable of acclimation across broader environmental gradients, particularly in respect to light availability, than what might be concluded from most observations of its natural occurrence in the MAV (Hawkins et al., 2009; Devall, 2013). Obviously, natural occurrence of L. melissifolia is limited by several factors in addition to those affecting its ecophysiology but the potential for photosynthetic acclimation bares relevance to tolerance of soil flooding and the light environment. An understanding of the mechanisms that support this photosynthetic acclimation in L. melissifolia can be gained from relationships between photosynthetic capacity and indices of physiological plasticity, principally leaf N, and structural plasticity, principally L_m/a, established in the literature (Ellsworth and Reich, 1992; Reich et al., 1995; Niinemets and Valladares, 2004; Kitaoka and Koike, 2005).

L. melissifolia developed greatest photosynthetic capacity (A_1800-a) in the absence of soil flooding and in the environment of highest light. A_1800-a decreased as light availability decreased; response along this gradient was linear to variation in N_a. Acclimation to open versus shaded environments observed for some temperate broadleaves, including Quercus petraea and Acer saccharum, suggests variation in N_a that drives the linear A_1800-a–N_a relationship arises exclusively from structural acclimation of the leaf, i.e. change in N_a is proportional to change in L_m/a (Gardiner et al., 2009; Delagrange, 2011; except see Ellsworth and Reich, 1992). For those species, the preeminence of leaf structure in determining N_a was confirmed by observation of a corresponding relationship between photosynthetic capacity and L_m/a, or stable N_m across the light gradient or range of L_m/a. Yet, many species, including L. melissifolia in this study, show change in N_m accompanying change in L_m/a in response to the light environment. As with this species, the relationship between N_m and L_m/a most often is negative because structural change to the lamina that increases L_m/a tends to move leaf chemical composition to a lower N fraction (Niinemets, 1997; Poorter et al., 2009). The significance of this pattern is that the amounts of N partitioned to light harvesting versus carbon fixation functions varies with leaf N availability (Evans, 1989; Niinemets, 1997) and can impart a physiological impact on the A_1800-a–N_a relationship (Ellsworth and Reich, 1992; Delagrange, 2011).

We expected to find an element of metabolic plasticity affecting the A_1800-a–N_a relationship that would substantiate a physiological component of acclimation of L. melissifolia photosynthetic capacity across the studied range of light environments after recording variation in N_m associated with L_m/a. This should be revealed through variation in A_max-m and the A_1800-m–N_m relationship. Reich et al. (1998) and Kaelke et al. (2001) illustrated the dependency of A_m upon N_m for a range of deciduous broadleaf tree species but noted the strength of this relationship varied between species and especially for intra-specific comparisons of photosynthetic acclimation across light environments. For example, Gardiner et al. (2009) demonstrated the importance of N availability to acclimation of photosynthetic capacity of Fagus sylvatica in contrasting understory environments, but Quercus robur photosynthetic acclimation showed much less reliance on N_m. In this study, A_max-m of L. melissifolia did not differ by light level (Table 3), and A_1800-m was unrelated to N_m for plants grown in the absence of soil flooding. While N_m differed between light environments, and this was likely accompanied with shifts in N partitioning among proteins of the photosynthetic process (Niinemets, 1997; Delagrange, 2011), photosynthetic capacity lacked responsiveness to N availability. This finding suggests acclimation of photosynthetic capacity to the prevailing light environment is primarily accomplished through adjustment of leaf structure and the resulting impact on N_a rather than adjustment of metabolic efficiencies. This is consistent with analysis by Reich et al. (1998), who found the A_1800-m–N_m relationship often lacks prominence for species that express strong variation in leaf structure relative to light availability.

Soil flooding too had great impact on photosynthetic capacity of L. melissifolia but the effect arose physiologically and was expressed only if plants were raised under environments of intermediate or high light. Extended flooding flattened slope of the A_1800-a–N_a relationship 99% which appeared attributable to two sources: depression of A_1800-a for a given level of N_a and a reduction in the range of N_a. The depression of A_1800-a relative to N_a indicates metabolic impairment of photosynthetic capacity—confirmed by the substantial decline in A_1800-m relative to N_m. Because of the differential effect among light environments and the lack of an effect on ϕ, we suspect soil flooding reduced photosynthetic capacity of L. melissifolia primarily through impairment of carbon fixation functions (Gardiner and Krauss, 2001). Evidence of reduced N partitioning to rubisco or suppressed rubisco activity has been reported for other flood-stressed temperate broadleaves that exhibited decline in photosynthetic capacity (Davies and Flore, 1986; Gardiner and Hodges, 1996; Gardiner and Krauss, 2001; Kitao et al., 2022). Vu and Yelenosky (1991), who worked with Citrus spp., reported direct evidence that suppressed carboxylation contributes to the metabolic decline by measuring decreased rubisco activation that ranged between 35% and 60% for two cultivars during 24 to 54 days of soil flooding. Though we cannot confirm soil flooding reduced rubisco activity in L. melissifolia our observation of depressed A_1800-a for a given level of N_a indicates contributions of this factor to the metabolic reduction in photosynthetic capacity.

Soil flooding is suspected to also impact photosynthetic capacity by lowering the amount of N partitioned to carboxylation functions (Liao and Lin, 1996; Herrera et al., 2008; De Pedro et al., 2020; Kitao et al., 2022). We recorded a reduction in the range of N_a relative to L_m/a that accompanied a depressed A_1800-a with extended flooding in high and intermediate light environments. This pattern diminished the dependency of A_1800-a upon L_m/a as observed for the light effect and shifted importance of N availability to the relationship, i.e. high A_1800-m was maintained through extended flooding in the low light environment where plants retained relatively high N_m. Given that L. melissifolia in high and intermediate light environments was unable to maintain N_m and consequently N_a under extended flooding, N partitioning may have receded away from carboxylation functions. Evans (1989) illustrated that the proportion of total leaf N partitioned to thylakoid proteins remains relatively stable across the range of N_a, while the proportion partitioned to Calvin Cycle proteins increases with increasing N_a.

A significant aspect of the physiological impairment discussed above is the apparent post-flood acclimation of L. melissifolia towards maximal photosynthetic capacity of the light environment. Other authors studying temperate broadleaf species emphasized that the ability to recover physiological function after floodwaters recede confers a level of fitness in floodplain environments (Wagner and Dreyer, 1997; Anderson and Pezeshki, 2000; Glenz et al., 2006; Li et al., 2015; Sample et al., 2023). Measurements from plants assigned the short-term flood were collected 30–40 days after floodwater was drawn down and all relationships between A_1800-a, or A_1800-m, and L_m/a, N_a, and N_m held slopes intermediate to those for extended and no flooding. The slope of the A_1800-a–N_a relationship for plants that experienced short-term flooding was 79% of the slope held by plants that had experienced no flooding indicating that metabolic constraints to photosynthetic capacity during soil flooding were waning. Post-flood acclimation appeared to be driven by increased N_m, and subsequently N_a, relative to L_m/a which shifted dependence of A_1800a towards N_a as observed in the absence of flooding. However, the A_1800-m–N_m relationship remained significant, marking persistence of the flooding effect that depressed N_m and A_1800-m in high and intermediate light environments. Thus, post-flood acclimation of photosynthetic capacity appears to involve transition from the physiological factors predominant during soil flooding to the structural factors predominant in the absence of flooding.

Conservation Implications

Long-term conservation of L. melissifolia in the MAV could benefit from active management that improves size and vigor of existing colonies and promotes regeneration of additional colonies on suitable, unoccupied sites within forests the species currently inhabits. This study shows how two prominent environmental factors in the MAV floodplain interact to govern acclimation of photosynthetic capacity in L. melissifolia and contributes important information towards developing silvicultural prescriptions that foster improved colony vigor, growth, and reproduction. Key findings from our work that hold conservation implications include: 1) photosynthetic capacity of L. melissifolia exhibited wide plasticity and acclimation to environmental factors; 2) the photosynthetic mechanism showed resistance amid co-occurring stresses; 3) physiological plasticity enabled acclimation of the photosynthetic mechanism after floodwater removal; and 4) the light environment was primal to development of L. melissifolia photosynthetic capacity.

Our research revealed that photosynthetic capacity of L. melissifolia acclimated to prevailing light and soil hydrologic regimes through structural and physiological mechanisms of considerable plasticity. The range of plasticity observed for this understory shrub was comparable to that expressed by other temperate broadleaf species of more common occurrence and wider range (Valladares et al., 2002; Takahashi et al., 2005; Villar et al., 2013). We confirm L. melissifolia can adjust leaf structure and physiology to maintain a positive carbon balance under multiple stresses indicative of a wide range of alluvial floodplain habitats. This confirmation substantiates previous findings for growth and reproduction of L. melissifolia (Lockhart et al., 2018; Leininger et al., 2021) and dismisses concerns of narrow habitat suitability for this species relative to development of the photosynthetic mechanism. The plasticity and acclimation potential associated with the photosynthetic mechanism imparts an element of flexibility in design and implementation of silvicultural practices aimed at habitat improvement for the species. Lindera melissifolia is inherently capable of photosynthetic acclimation across the range of understory microsites that could be created by silvicultural practices of different intensities, and application of practices need not be restricted to sites of specific hydrologic regimes.

Following the first conservation implication, neither soil flooding or light imposed an additive limitation on L. melissifolia photosynthetic capacity when the other factor was at its most stressful level. This finding connotes a strong tolerance to soil flooding under the chronic stress of low light and implies a level of resistance inherent to the L. melissifolia photosynthetic mechanism. Niinemets (2010) stated that survival under a chronic stress like shade often depends on the capacity of a species to resist a second stress while structurally and physiologically acclimated to the first. Indeed, the ability to survive soil flooding in a shaded understory environment is not common among woody broadleaf shrubs in forests of the MAV (Hawkins et al., 2009). The resistance exhibited by the L. melissifolia photosynthetic mechanism is likely fundamental to stress tolerance and persistence of this plant in habitats of marginal suitability or declining quality. This characteristic affords conservationists the ability to prioritize effort on sites holding colonies in most need of active management.

Photosynthetic capacity of L. melissifolia is dynamic and not fixed by soil flooding regime when leaves develop under sufficient light at the onset of the growing season. The observed impairment to photosynthetic capacity by soil flooding was primarily physiological, and plants responded readily to removal of floodwater from soil through shifts in leaf N that supported increased photosynthesis. Thus, the effect of soil flooding appeared temporary as photosynthetic capacity of plants acclimated to fit their light environment after floodwater receded. The ability to respond photosynthetically to an improved soil environment is an attribute beneficial to other floodplain species (Pezeshki, 2001) and would also be important to active management for sustaining L. melissifolia in MAV floodplain habitats. The fact that leaf area develops while soil is flooded and photosynthetic capacity acclimates quickly after floodwater recedes would allow L. melissifolia to compete for site resources made available through active management. We speculate that the resistant and dynamic nature of its photosynthetic mechanism imparts a physiological component to the competitive adaptiveness of L. melissifolia in its floodplain habitats.

Our observations substantiate primacy of the light environment on development of photosynthetic capacity in L. melissifolia during and after soil flooding. The prevailing influence of light availability over photosynthetic capacity is consistent among temperate broadleaf species, including understory broadleaves, which require an ability to acclimate to understory light environments that range from heavily shaded to sunlit. The range of acclimation observed in L. melissifolia photosynthetic capacity relative to light establishes compatibility of the species with silvicultural manipulation of stand structure for improved habitat where overstory and midstory canopies create heavy shade. The importance of this characteristic cannot be understated for conservation of this species because forests of the MAV where L. melissifolia is endemic naturally develop multi-storied, dense canopies (Oliver et al., 2005; Hawkins et al., 2009). In forests of this condition, understory light availability is easy to quantify and monitor and light deficiencies can be alleviated through manipulation of forest stand structure. In application, though, we recognize that stand disturbance would release competing vegetation and alter competitor interactions with L. melissifolia, especially on sites or in years of short-duration surface inundation (Wright, 1990; Hawkins et al., 2016). This underscores the need for applied research to support development of comprehensive silvicultural prescriptions that account for the environmental and ecological complexities inherent to natural L. melissifolia colonies. Still, active management of stand structure to improve understory light environments offers an approach, compatible with its ecophysiology, to bolster colony vigor, growth, and reproduction for sustained conservation of L. melissifolia.

Author contributions

ESG, TDL, KFC, MSD, PBH, ADW, and NMS conceived the idea and contributed to funding acquisition, which was led by TDL. TDL provided project administration. ESG and TDL co-supervised the work. ESG developed the methods, led the investigation, conducted the analysis, and archived data. ESG drafted the text, tables and figures. ESG, TDL, KFC, MSD, PBH, ADW and NMS evaluated results, reviewed findings, edited drafts and contributed to the final manuscript.

Conflicts of interest

None declared.

Funding

This work was supported by the US Army Corps of Engineers, Vicksburg District (grant/award no: SRS 01-IA-11330127-527) and by the US Forest Service (operating funds of the Southern Research Station).

Data Availability Statement

Data pertaining to this research are on file with the US Forest Service, Center for Bottomland Hardwoods Research and are available on request from the corresponding author.