4.1 Spatiotemporal Variation in Aboveground Vegetation Dynamics
Spatial and temporal aboveground vegetation changes in deltaic wetlands occur as a result of complex feedback mechanisms and environmental gradients, with significant implications for coastal wetlands persistence and sustainability in the face of rising sea level (DeLaune and Pezeshki 1994b; Stagg et al. 2016a; Elsey-Quirk et al. 2019). We found that the saline marshes in both active and inactive delta basins had the highest AGB compared with all other wetland types, and the increase in AGB and ANPP rates occurred along a salinity gradient (Fig. 2; Table 3). Moreover, seasonality had a significant effect on vegetation biomass in both basins (Table 2). The seasonality effects are indicative not only of the obvious phenological changes that wetland plants undergo during their growth cycles (Odum 1988), but also of the effects that river discharge and flood cycles have on plant growth and distribution in these deltaic systems (Whigham and Simpson 1992). Previous research has shown how seasonal patterns observed in river-dominated deltaic wetlands, such as spring floods and high-water levels in the early growing season, can delay germination and re-sprouting of marsh plants (Johnson et al. 1985), change the contribution of annual vs perennial species to the overall biomass (Doumlele 1981), and even change plants’ physiological tolerances and ability to compete with other species for resources (Pasternack et al. 2000).
The highest seasonal change in AGB was observed in saline sites in both active (Atchafalaya) and inactive (Terrebonne) delta basins, with total estimates ranging from 343 ± 101 g m− 2 in early growing season (spring) to 1214 ± 210 g m− 2 during peak biomass season (summer; Fig. 2). In freshwater sites, the most striking seasonal change in AGB occurred in the WLD intermediate intertidal (lower elevation) zone, which was completely depleted of standing biomass in the early growing season and reached AGB values of up to 850 g m− 2 during peak biomass. Necromass also changed with seasonality, with lower estimates during peak biomass season across most of the sites (Appendix S2: Fig. S1). Piazza et al. (2011) reported similar seasonal trends in vegetation across salinity gradients in LA marshes affected by hurricanes Katrina and Rita. Specifically, AGB estimates across the inactive delta (Terrebonne) basin were highest in saline sites during the first year after the hurricane (2006), and in brackish sites during the second year of sampling (Piazza et al. 2011). Studies conducted in freshwater marshes of Western Terrebonne seem to indicate relatively consistent estimates of AGB over the last 30 years, with fluctuations after disturbance events (e.g., hurricanes and droughts) (McKee et al. 2006; Rivera-Monroy et al. 2019). It is likely these marshes benefit from their proximity to Atchafalaya River that penetrates eastward deep into the basin through bayous and canals, providing pulses of nutrient-rich sediments into the system (Sasser et al. 1995; Lane et al. 2010).
Overall, our estimates of ANPP were similar to values reported in the literature for fresh, brackish, and saline wetlands in coastal LA (Appendix S2: Table S1). In particular, ANPP rates estimated using the Smalley method in brackish and saline sites dominated by C4 plants (S. alterniflora and S. patens) (Table 3), are consistent with studies of AGB productivity conducted in the same geographic region (Appendix S2: Table S1 and references therein). Decades of species-level studies have demonstrated the importance of local and regional environmental drivers on AGB primary production (Schubauer and Hopkinson 1984; Pennings et al. 2005; Morris et al. 2013; Token et al. 2022). For example, salinity and increased flooding have been shown to negatively impact ANPP in all wetland types (Howes et al. 1986; Howard and Mendelssohn 1999; Day et al. 2009; Pierfelice et al. 2017), regardless of plant’s tolerance to stressful conditions. However, studies that apply recent scientific advances in predictive modeling and remotely sensed data suggest a more nuanced relationship among zonation, abiotic and biotic stress, and productivity (Schoolmaster and Stagg 2018; Dronova and Taddeo 2022). For instance, in a landscape level study of aboveground biomass production in Terrebonne and Barataria basins, Stagg et al. (2016b) found that while salinity had a negative effect on production within wetland type, along the landscape gradient, brackish marshes were more productive than any other wetland types (fresh and saline). Contrary to our hypothesis that AGB allocation would be lower in inactive Terrebonne basin, saline marshes (dominated by C4 plants) in both basins had the highest biomass production compared to all other wetland types (Table 3).
Unlike freshwater C3 plants, C4 plants have better photosynthetic and water use efficiency (Pezeshki and Delaune 1993; DeLaune and Pezeshki 1994a), which can stimulate productivity (Stagg et al. 2016b). In our study, porewater salinity (Fig. 6) had a strong positive effect on AGB, explaining 85% of the variability in AGB across both basins. Specifically, saline sites in both active and inactive delta basins had the highest salinity (range: 13 to 14.5 ppt) (Table 1) and the highest AGB during peak biomass season (Fig. 2; Fig. 6). Indeed, this is consistent with earlier studies that showed that S. alterniflora dominated marshes reach their highest biomass at porewater salinity levels between 10–15 ppt (Mooring et al. 1971; Linthurst and Blum 1981). In addition, salinity has been observed to mobilize NH4+ from soils (Ardon et al. 2013), which could result in a fertilization effect and higher rates of productivity (Herbert et al. 2015). Further, plant communities in ADB and IDB saline sites might benefit from the constant pulses of flood waters and sediment deposition into the system. For instance, Fontenot (2022) found that surface accretion and inorganic and organic matter sedimentation rates increased along the salinity gradient in the IDB and that the saline site had the highest (2.94 cm yr− 1) surface accretion (Table 1; (Fontenot 2022; Twilley et al. 2022)). This finding, coupled with low marsh elevation (0.04 m) at this site compared to brackish and freshwater sites (Table 1; (Twilley et al. 2022)), might indicate that the IDB saline site is benefiting from higher sedimentation rates enhanced by frequent flooding events and increased input of allochthonous nutrients into the system. Indeed, recent studies have shown that re-suspended sediment input into the distal areas of the inactive Terrebonne delta from bay bottoms or the nearshore Gulf of Mexico contribute to accretion and the resilience of marshes to SLR (Twilley et al. 2016; Cortese and Fagherazzi 2022; Day et al. 2023). This sediment dynamics might be beneficial to marsh vegetation, leading to higher biomass allocation in saline sites (Table 1). Overall, this landscape trend along the salinity gradient, might indicate that aboveground production is maintained with shifts to more salt-tolerant wetland types, as well as resources (e.g., nutrients) that are available at the level of marsh type (Merino et al. 2008; Meert and Hester 2009).
In a model describing plant biomass as a function of resource competition and salinity stress, Schoolmaster and Stagg (2018) showed that, if the supply of the limiting resource in the system and the nutrient use efficiency of plants increases with salinity, then the negative relationship between production and salinity can be reversed. Our results show that in the active delta basin, soil total N was a significant predictor of higher biomass along a salinity gradient (Fig. 7A), while in the inactive basin, both soil total N and P had a significant positive effect on AGB allocation (Fig. 7B; D). The rate of N supply has been shown to influence plant productivity and the composition of plant communities in coastal LA marshes (DeLaune et al. 2016). However, high riverine nutrient pulses are not always beneficial to plants, especially in freshwater marshes where high nutrient availability may reduce plant root growth, decrease soil strength, and make marshes more susceptible to disturbances (Darby and Turner 2008; Howes et al. 2010; Morris et al. 2013). Also, under conditions of high N enrichment, plants tend to assimilate more N by increasing foliar N and decreasing C:N (Shaver and Melillo 1984). This is often the case in younger deltaic systems such as WLD where plant growth is initially N-limited (Henry and Twilley 2014). As soils develop, plant communities become more diverse and established, which in turn adds to the soil organic matter pool (Craft 1997) and shifts plant growth to P limitation over time (Chapin et al. 1986; Shaffer et al. 1992). Our results suggest that in the inactive (Terrebonne) basin (a P-limited system; N:P range: 35–45) (Fig. 8A, B), both soil total N and P strongly controlled biomass allocation (Fig. 7B, D), while in the active (Atchafalaya) basin (N-limited system), only soil total N had a significant effect on biomass along a salinity gradient (Fig. 7A, C).
Our results are consistent with studies that found that as salinity increased, nutrient tissue concentrations declined as production increased, indicating greater nutrient use efficiency with increased stress (Linthurst and Seneca 1981; Chapin et al. 1986; Aerts 1996). This interesting pattern, especially in the active delta fresh site seems to indicate that these freshwater species adopted a strategy of storing nutrient capital, without immediately reinvesting N and P in the production of new tissues (i.e., growth). This strategy, also known in plant ecology as “luxury consumption”, has been observed in previous studies (Twilley et al. 1985; Chapin et al. 1990; Aerts and Chapin 1999) and is attributed to plant’s ability to store nutrients for the following year (Chapin et al. 1986). Moreover, in conditions of increased nutrient availability, leaf nutrient content at the community level increases due to species replacement and various phenotypic responses (Chapin et al. 1990; Aerts and Chapin 1999), coupled with decreasing foliar C:N (Shaver and Melillo 1984). This C:N trend is evident in ADB freshwater plants throughout both seasons (Appendix S2: Fig. S4C) (Castañeda-Moya and Solohin 2022). In contrast, in the IDB fresh site lower foliar nutrients and high biomass seem to indicate an investment of resources (nutrients) into production of new tissue (i.e., growth). These results support our hypothesis that nutrient uptake in plants is higher in the mineral-rich soils of Atchafalaya basin (especially in the fresh site). Nevertheless, high uptake of nutrients by plants does not necessarily result in higher biomass production, suggesting that plant communities respond to nutrient availability based on their intra- and interspecific adaptations to local environmental conditions.
4.2 Changes in Plant Community Composition along Environmental Gradients
Numerous species-level studies in herbaceous wetlands have shown the importance of allogenic and autogenic factors on plant community development and productivity (Shaffer et al. 1992; White 1993; Holm and Sasser 2001; Carle 2013; Hupp et al. 2019). For example, plants found in younger river-dominated deltaic wetlands such as WLD, are exposed to various allogenic factors such as sedimentation, river currents, and erosion (Carle and Sasser 2015; Carle et al. 2015). These processes are usually indicative of geomorphological changes occurring within the young delta, leading to establishment of an elevation gradient, that is considered one of the most important predictors of plant species distribution within a growing deltaic wetland system (Johnson et al. 1985; Holm and Sasser 2001). Our study showed that vegetation in the ADB fresh site displayed a distinct zonation along the elevation gradient, from higher elevation (marsh edge) to lower elevation in the intermediate intertidal zone (Appendix S2: Fig. S2B). This result is supported by previous studies in WLD that have documented shifts in plant community composition and distribution over time (Carle et al. 2015; Bevington et al. 2022), specifically species like P. nodosus and emergent Nelumbo lutea spreading over much of the lower elevation areas of the delta previously dominated by Sagittaria spp., grasses, and forbs (like Polyganum spp.) dominating intermediate elevations, and C. esculenta and woody S. nigra growing along higher elevation on channel levees (Carle 2013; Jensen et al. 2021). Furthermore, a recent study of plant biomass in WLD, based on an in-situ mesocosm experiment, found that hydrogeomorphic zones explained species biomass variability, and reported a specifically strong positive correlation of C. esculenta AGB with percent inundated time, increasing from higher elevation to the lower elevation (intermediate intertidal) zone (Rovai et al. 2022). Our study also found variability in seasonal succession of freshwater species in the inactive Terrebonne basin (Appendix S2: Fig. S2). Unlike brackish and saltmarsh plant communities, species composition and diversity in freshwater marshes can change dramatically throughout the course of one growing season (Odum 1988; Holm and Sasser 2001). This is especially the case in river-dominated deltaic wetlands, where intra-annual vegetation changes have been well-documented (Stern et al. 1986) and are usually caused by variations in water level and river’s flood cycles throughout the year. Our assessment of the plant communities in the active and inactive delta freshwater sites provides a clear example of such changes: shifts in species composition between early and peak growing seasons, and even changes between marsh edge vs interior (Appendix S2: Fig. S2A, B; (White 1993)). Seasonal changes in brackish and saline plant communities in both basins were not as evident as in the freshwater sites. In fact, even though AGB in our saline sites increased during peak biomass season, species richness of these marshes declined (Fig. S3). For instance, the ADB saline site was dominated by monotypic stands of Spartina spp., while at the IDB saline site, S. alterniflora replaced other salt-tolerant plants (e.g., J. roemerianus), and became the only dominant species during peak biomass season (Appendix S2: Fig. S2A, B). Overall, seasonal and site-specific changes in plant community composition assessed in our study suggest that the distribution of species depends both on their physiological tolerances to stress (e.g., salinity, prolonged inundation) and their ability to adapt to changes in resources (e.g., nutrients) over time.
4.3 Foliar and Soil Isotopic Signatures of Herbaceous Wetlands
Stable isotopes (δ13C, δ15N) have been used to identify the general sources of organic material in wetlands (Chmura and Aharon 1985), to assess sources of nutrient and to understand the effects of environmental conditions on plant metabolism and growth (Dawson et al. 2002; Gouveia et al. 2019). Our study showed that foliar δ13C varied among sites, as well as between species, with higher (more enriched) δ13C values in saline and brackish species compared to freshwater species in both basins (Fig. 3A). These trends are consistent with previously reported estimates, which confirm that C4 plants (dominating saline sites) trap a greater proportion of 13CO2 relative to C3 plants (present in fresh and brackish sites) as a result of plant adaptations to environmental stressors such as elevated salinity (Voss et al. 2017). In general, stable C isotopic composition of soils is expected to be similar to that of the source plant material (Choi et al. 2001). Interestingly, our results showed a discrepancy in the expected trends between soil and foliar δ13C (Fig. 3A; Table 1). Specifically, soil δ13C estimates in IDB brackish and saline sites were on average 1.4 times lower than foliar δ13C estimates at the same sites (Fig. 3A; Table 1). Also, soil δ13C values in IDB saline site showed a decreasing trend with depth (Fig. S5). Since saline marshes in proximity to our IDB saline site have been accreting at a rate of ~ 0.8 cm yr− 1 (Baustian et al. 2021), we estimate that the shift in soil δ13C isotopic signatures at this site started in the early 1960s, which is indicative of shifts in vegetation from a brackish C3-plant dominated community to the current saline (C4-dominated) marsh. This shift in dominance of C4 vegetation is likely an indicator of changes in salinity regime as a result of landward expansion of the coastal wetland due to accelerated regional SLR in the past century. This is supported by recent studies that estimate that Terrebonne basin has experienced a significant landward migration (17 km) over the last 78 years, resulting in an increase (25%) of saltmarsh vegetation in this region (Twilley et al. 2016; Twilley et al. 2019). Moreover, the observed differences in δ13C of vegetation and soil in IDB might be indicative of greater accumulation of lignin in the substrate of Terrebonne soils (Benner et al. 1987; Craft et al. 1988), due to high rates of productivity of these marshes throughout time (Chmura et al. 1987; Choi et al. 2001; Finlay and Kendall 2007).
Foliar δ15N of the freshwater species in ADB were significantly more enriched compared to that of freshwater species in IDB (Fig. 3B). These differences likely reflect the ability of freshwater species to fix N and adapt to N-limited soil conditions in the Atchafalaya Basin. However, elevated foliar δ15N levels in ADB plants might also reflect the increased rates of N mineralization happening in the freshwater site (McKee et al. 2002; BryantMason et al. 2013) as a result of the continuous nitrate loading in the Mississippi River waters into the coastal zone over the years, with current nitrate concentrations five times higher compared to the pre-industrial time (Goolsby et al. 2000). Previous studies have shown that changes in δ15N in wetland plants might be due to variation in source δ15N (e.g., anthropogenic NO− inputs) and/or to physiological fractionation by plants growing under different stress conditions (Evans 2001; BryantMason et al. 2013; Tanu et al. 2020). For instance, Henry and Twilley (2014) reported that under nitrate-enriched conditions, the dominant N2 transformation in WLD is removing reactive N via denitrification, rather than increasing N via fixation, making WLD a N-limited system. The nitrate-enriched conditions in the active Atchafalaya basin coincide with increased rates of N mineralization and nitrification (Henry and Twilley 2014). This, in turn, results in δ15N enrichment of soil pool as observed in WLD (Appendix S2: Fig. S), as the lighter δ14N isotopes are preferentially lost through leaching and denitrification (Dawson et al. 2002). Plants accessing this soil N pool can then become relatively δ15N-enriched over time as observed in our study (Evans 2001; Dawson et al. 2002). In contrast, foliar δ15N of IDB freshwater plants were significantly lower than those of ADB plants (Fig. 3B), reflecting the lack of N-pulsing events in the sediment-deprived basin. Overall, these results support our hypothesis that changes in plant and soil isotopic signatures follow the spatiotemporal patterns in environmental stressors affecting coastal plant communities.
4.4 Landscape-level Carbon Stocks
The landscape distribution of AGB C stocks (AVIRIS-NG analysis) during both seasons is consistent with previous estimates reported in coastal Louisiana (Thorhaug et al. 2019). For instance, plant C stocks across Terrebonne and St. Mary’s Parishes (LA) ranged between 3.4 and 3.8 MgC ha− 1 in palustrine, and estuarine emergent marshes, respectively (Byrd et al. 2018). Moreover, Thomas et al. (2019) estimated marsh and forested wetland AGB within Atchafalaya and Terrebonne basins using high-resolution Sentinel-2 optical imagery and reported an average AGB of ~ 382 g m− 2 and ~ 798 g m− 2 during early growing and peak biomass seasons, respectively. Specifically, the authors documented a high seasonal increase in biomass in IDB saline marshes (stations CRMS0307 and CRMS0434), in close proximity to our intensive saline study site in Terrebonne basin, ranging between 915 and 1014 g m− 2 during peak biomass season (Thomas et al. 2019). Our landscape-level maps show seasonal trends in C stocks dynamics in both basins, with C densities ranging between 1.2 and 3.2 MgC ha− 1 during early and peak biomass season, respectively. Furthermore, even though AGB C stocks estimates did not vary significantly across basins, we were able to detect regions of high C stocks (> 5 MgC ha− 1) in southeastern Terrebonne marshes (Fig. 5B). This is consistent with previously reported regional AGB and C estimates (Thomas et al. 2019; Jensen et al. 2023), as well as with our field observations in this region of Terrebonne basin.
The experimental approach of our study is unique in that it integrated field biomass data and state-of-the-art airborne-derived reflectance spectra captured by NASA’s AVIRIS-NG, to compile landscape level vegetation cover and C stocks maps for the active and inactive basins in coastal Louisiana. Our findings offer an insight to our understanding of species- and landscape-level changes in plants’ adaptive strategies to environmental stressors and availability of resources. Basins with contrasting sediment delivery and hydrological regimes, such as Atchafalaya and Terrebonne, can both support high biomass and C stocks, however, it is not clear whether this is indicative of overall ecosystem health, given current and future anthropogenic disturbances and climate-induced changes in this coastal region. This underscores the need for further research to quantify additional spatiotemporal changes across these coastal basins, such as patterns in belowground biomass and distribution. Elucidating these biomass and C partitioning patterns between above- and belowground components of deltaic wetlands vegetation will improve our understanding of factors that lead to changes in plant adaptive strategies and the feedback mechanisms that control landscape vegetation gradients and coastal resilience in the region. This is particularly relevant in the context of ongoing and future river diversions to restore Louisiana’s coastal wetlands.