Biophysical controls of marsh soil shear strength along an estuarine salinity gradient

Sea-level rise, saltwater intrusion, and wave erosion threaten coastal marshes, but the influence of salinity on marsh erodibility remains poorly understood. We measured the shear strength of marsh soils along a salinity and biodiversity gradient 10 in the York River estuary in Virginia to assess the direct and indirect impacts of salinity on marsh erodibility. We found that soil shear strength was higher in monospecific salt marshes (5 36 kPa) than biodiverse freshwater marshes (4 8 kPa), driven by differences in belowground biomass and rooting structure. However, we also found that shear strength at the marsh edge was controlled by sediment characteristics, rather than vegetation or salinity, suggesting that inherent relationships may be obscured in more dynamic environments. Our results indicate that freshwater marsh soils are weaker than salt marsh soils, and 15 suggest that salinization of freshwater marshes may lead to simultaneous losses in biodiversity and erodibility.


Introduction
Tidal marshes are rapidly evolving ecosystems that sit at the boundary between land and sea, and are influenced by a tight coupling between biological and geomorphic processes (Redfield, 1972). Marshes provide a wide variety of ecosystem services, such as improving water quality, sequestering carbon, and reducing the impacts of storm surge and coastal erosion 20 (Barbier et al., 2011). However, tidal marshes are vulnerable to climate change and its effects-such as sea-level rise, increased storm frequency, and saltwater intrusion (Craft et al., 2009;FitzGerald et al., 2008;Kirwan and Megonigal, 2013).
While many studies have analyzed how sea-level rise and storms influence wave erosion (Howes et al., 2010;Leonardi et al., 2016), less is known about the processes driving the strength of marsh soils and their impact on erosion rates (Jafari et al., 2019). In general, vegetation increases the shear strength and decreases the erodibility of marshes (Ameen et al., 2017;Sasser 35 et al., 2018;Wilson et al., 2012). Nevertheless, it is difficult to attribute large-scale controls on shear strength in marshes due to the heterogeneous distribution of roots, stems, soil types, and shells which all yield variable influences (Jafari et al., 2019).
Previous work has examined shear strength in marsh soils in the context of vegetation, geomorphic setting, and methodology (Ameen et al., 2017;Jafari et al., 2019;Lin et al., 2016;Sasser et al., 2018;Watts et al., 2003;Wilson et al., 2012). However, additional research is required to fully understand the biophysical drivers of marsh soil shear strength. 40 Predicting how erosion rates will change in tidal marshes with increased salinization from sea-level rise requires a more mechanistic understanding of the drivers of shear strength. Salinity likely influences shear strength through ecological factors, such as dominant vegetation communities or distribution of belowground biomass (Ameen et al., 2017;Sasser et al., 2018).
An increase in salinity may increase marsh soil shear strength assuming high salinity marshes favor species with deeper roots 45 (Howes et al., 2010). Alternatively, shear strength may decrease with increasing salinity due to the loss of vegetation biodiversity (Ford et al., 2016). Here, we measure soil shear strength and biophysical parameters along an estuarine salinity gradient and use them to determine how salinity influences marsh soil shear strength.

Study area and approach 50
We measured the erodibility of marshes along a salinity gradient in the York River Estuary, a tributary of the Chesapeake Bay (Virginia, USA). The York River is a microtidal, partially-mixed estuary with a mean tidal range of 0.7 meters at the mouth of the river and 1 meter near the freshwater river sources (Fig. 1a) (Friedrichs, 2009). The York River salinity gradient is created by saltwater from the Atlantic Ocean, and freshwater from the Mattaponi and the Pamunkey rivers (Reay, 2009). Sealevel rise rates are 3-4 times faster than the eustatic levels in this region, which could facilitate faster rates of salinization (Ezer 55 and Corlett, 2012). Various wetland types exist along the York River within different salinity regimes: polyhaline salt marshes that have monocultures of Spartina alterniflora (saltmarsh cordgrass), mesohaline brackish marshes with an extensive array of halophytic grasses, and oligohaline freshwater marshes with dominant plant species Peltandra virginica (arrow arum) and Zizania aquatica (wild rice) (Perry and Atkinson, 2009). Saltwater intrusion is driving an increase in salt-tolerant species at the freshwater marsh sites (Perry and Atkinson, 2009). 60 We chose five marshes along the York River salinity gradient for this study (Fig. 1a). Salinity decreases upriver from 18 ppt at the Goodwin Islands to 0 ppt at the Pamunkey Indian Reservation (Reay, 2009). Within these overall sites, we chose sampling locations along tidal creeks 5-10 m wide, with marsh widths beyond 20 meters and consistent elevations across all study sites (Fig. 1b). We collected samples from both the tidal channel marsh edge and from the marsh interior. Edge sites 65 were located between the tidal channel and any levee (1 m from edge), while interior sites were located at a measured distance of 10-12 m away from the edge site (Fig. 1b). All field work was done in July-August 2018, except for the collection of cores for belowground biomass at the Sweet Hall Marsh and Pamunkey Indian Reservation marsh edge sites (September 2018) and elevation profiles of the Pamunkey Indian Reservation (March 2020).

Measurements of shear strength, vegetation, and soil properties 70
We measured shear strength and a variety of biophysical characteristics at each location in this study. We used a Humboldt Shear Vane to determine the shear strength of marsh soils. Although the effectiveness of the shear vane in wetland soils is undetermined (Jafari et al., 2019), it remains the most widely used method to quantify wetland soil shear strength in coastal science (Ameen et al., 2017;Howes et al., 2010;Valentine and Mariotti, 2019). We used a 1 m long shear vane, with a 10 cmlong head for relatively weak marsh soils, and a 5 cm-long head to measure relatively strong marsh soils. The shear vane was 75 turned until the marsh soil was broken, and unitless values on the shear vane were converted to kPa using the manufacturer's conversion formulas (10 cm head: kPa=shear vane value*10*0.0625; 5 cm head: kPa=shear vane value*10*0.5). Each shear strength profile was one meter long with measurements taken at 10 cm intervals. Ten replicate profiles were taken per marsh location at each study site for a total of 100 total soil shear strength profiles along the York River.

80
We measured elevation profiles for all five sites using real-time kinematic (RTK) GPS across a transect from the marsh edge to the interior (Fig. 1b). For each site, five ground control points were taken on the marsh edge, 1 meter in from the edge, and in the marsh interior to obtain more accurate elevations at our specific sampling locations. All sites were located at similar elevations except for the Pamunkey Indian Reservation, which was lower in elevation than the other sites. This discrepancy is likely due to the other elevation profiles being measured during the growing season while the Pamunkey Indian Reservation 85 profile was measured in early Spring. There is no vegetation cover in freshwater marshes along the York River during colder months (Perry and Atkinson, 2009), and the marsh platform may lose elevation due to seasonal erosion or subsidence.
Aboveground and belowground biomass were measured destructively. We collected standing vegetation from three replicate above-ground stem clip plots (25 cm x 25 cm) per marsh location at each study site. Samples were counted for the total number 90 of stems, dried to a constant weight, and separated by species to calculate species richness. Three replicate belowground biomass soil cores (15 cm diameter, 50-70 cm depth) were collected at each location within sites (except for the edge site at the Pamunkey Indian Reservation where only 2 cores were taken), and used to measure belowground root and rhizome biomass. We sectioned cores into 10 cm increments and washed over a 1 mm screen sieve. Live belowground biomass was separated based on color, turgidity, and buoyancy in water. Live belowground biomass was dried to a constant weight, and 95 used to generate belowground biomass profiles for each study site.
Water content, bulk density, and organic matter content were determined from two Russian peat cores (2 cm diameter, 1 m depth) collected at each sampling location per study site. Cores were sectioned into 1 cm segments for the top 50 cm, every 3 cm for the bottom 50 cm. Samples were dried to a constant weight, homogenized, and combusted for 6 h at 550°C to burn off 100 organic material.

Shear strength
Shear strength measurements generally ranged from 0-36 kPa, but differed between locations within a site, and across sites.
There was an observable trend of shear strength increasing with depth at freshwater marsh sites for the upper 30 cm of the soil 105 profile (Fig. 2). For brackish and salt marsh sites, patterns of shear strength with depth were inconsistent. There was a large increase in shear strength below 30 cm at the Goodwin Islands edge location (Fig. 2a), which we attribute to measurements that were within the antecedent lithology (i.e. non-marsh soils). In analyses related to the effect of salinity on soil shear strength (discussed below), we used depth-averaged shear strength values from the upper 30 cm of the soil profile. We selected the upper 30 cm as our window of averaging, because shear strength varied little with depth beyond 30 cm at most sites, it excludes 110 any influence of antecedent parent material, and corresponds to typical vegetation rooting depths. This approach allows for comparisons between sites that are based on the same depth interval at each site.
Depth-averaged shear strength increased significantly with salinity across study sites for interior locations only (R 2 = 0.81; p = 0.038) (Fig. 3b). There was no significant relationship between shear strength and salinity for edge locations across all 115 sites (R 2 = 0.04; p = 0.74) (Fig. 3a).  (Fig. 4). Shear strength values were nearly equal between the edge and interior 120 sites for freshwater marshes (Fig. 4). The most substantial difference between edge and interior shear strength values occurred at the salt marsh sites, with an increase from 5.1 kPa at the edge to 18.5 kPa in the interior.

Biophysical drivers
In the marsh interior, vegetation properties largely explained variability in soil shear strength. Belowground biomass had the most significant influence on shear strength in the marsh interior (p = 1.086e-5) (Fig. 5). Aboveground biomass was also correlated with shear strength in the marsh interior but was marginally significant (R 2 = 0.64; p = 0.105). At the marsh edge, soil properties explained most of the variability in shear strength on edge sites. Organic content was correlated with edge shear strength values but was marginally significant (R 2 = 0.72; p = 0.059). However, other properties that co-varied with organic content (i.e. water content, bulk density) were also important, including the relationship between bulk density and shear strength in salt marshes (Fig. 6). 130

Discussion
The results from this study are consistent with previous work that identifies vegetation and soil properties as important drivers of marsh soil shear strength. For example, soil shear strength is well known to vary with dominant plant species (Howes et al., 2010;Sasser et al., 2018). Our work extends this concept and finds that soil shear strength is positively correlated with belowground biomass in the marsh interior (Fig. 5). This finding aligns with natural and manipulative experiments that show 135 shear strength increases with belowground biomass (Wilson et al., 2012), and that the mortality of belowground roots and rhizomes is related to enhanced erosion (Coleman and Kirwan, 2019;Lin et al., 2016;Silliman et al., 2019;Wilson et al., 2012). Like previous work (Ameen et al., 2017;Wilson et al., 2012), our results demonstrate that soil properties such as bulk density are also important drivers of marsh erodibility (Fig. 6). However, we uniquely show that the relative importance of vegetation and soil properties depends on the location within a marsh (edge vs. interior). 140 Our work indicates that the marsh interior has a higher soil strength than the marsh edge at our salt and brackish marsh sites ( Fig. 4). We ascribe this variability in saline marshes to biological drivers influencing marsh interior soils (Fig. 5), and soil properties determining soil shear strength at the seaward marsh boundary (Fig. 6). The differing influences are due primarily to various processes occurring at different places in the marsh. Sedimentation is low in the marsh interior leading to more 145 compacted stronger soils, and biomass tends to be concentrated closer together with higher stem densities. The tightly packed belowground root network adds cohesion to marsh soils without the active edge processes frequently reworking sediment (Silliman et al., 2019). In contrast, the marsh edge is typically more dynamic, where increased inorganic sediment deposition and resuspension leads to more unconsolidated, mineral-rich soils that can impact soil cohesion and shear strength (Ameen et al., 2017). Enhanced nutrient loading, particularly in wetlands undergoing eutrophication, at the marsh edge weakens soils and 150 may also influence shear strength variability at the seaward boundary (Turner et al., 2020). It is unclear why similar spatial patterns were not observed in the freshwater marsh locations. Perhaps the overall lower belowground biomass and shear strength of the freshwater marshes precludes our ability to detect patterns across the marsh. Nevertheless, our findings indicate that belowground biomass drives soil shear strength variability in the marsh interior (Fig. 5), and soil properties influence marsh edge shear strength (Fig. 6) for brackish and salt marshes. 155 Salinity potentially plays an important role in determining the erodibility of marsh soils, through its combined influence on vegetation type and belowground biomass. Prior work examining the relationship between salinity and marsh soil strength concludes that salt marshes are more resistant to lateral edge erosion than freshwater marshes, ascribing variation in rooting depth to differences in soil shear strength (Howes et al., 2010). While our study also finds that salt marshes are stronger than 160 freshwater marshes in the marsh interior (Fig. 3), we find that salt marshes are generally stronger than freshwater marshes regardless of the depth within the soil and rooting profile (Fig. 2b). We therefore attribute stronger salt marsh soils in our study area simply to the greater belowground biomass of Spartina alterniflora relative to freshwater species such as Peltandra virginica. Interestingly, we found no relationship between salinity and soil shear strength at the marsh edge, where erosion would actually occur (Fig. 3a). Although this finding warrants more attention, we suggest that processes associated with a 165 more dynamic marsh edge (e.g. sediment deposition, erosion, and resuspension) obscure patterns that would otherwise be evident.
Root structure and geometry may also have considerable influence over marsh soil shear strength (Ameen et al., 2017;Howes et al., 2010). In the salt marsh, Spartina-dominated systems, the belowground root network consists of fibrous, tightly 170 interlocking strands that may hold soil more effectively (Fig. 7a). Peltandra virginica is the most abundant plant species in the freshwater sites examined in this study, whose belowground biomass consists of large tubers with aerenchyma tissue and easily broken roots spaced out throughout the marsh (Fig. 7b). This laterally heterogeneous distribution of roots across the surface of the freshwater marsh may lead to overall decreased soil shear strength.

175
Previous work in U.K. marshes finds that freshwater marshes with a more diverse array of vegetation species have stronger marsh soils due to greater belowground biomass (Ford et al., 2016). However, our study finds that marsh soil shear strength increases with a decrease in plant biodiversity. Like other estuaries (Brock et al., 2005;Engels and Jensen, 2010;Grenier La Peyre et al., 2001;Odum, 1998), biodiversity decreases from freshwater sites to salt marsh sites in our study area (species richness = 6 for Pamunkey Indian Reservation, species richness = 1 at Goodwin Islands and Catlett Islands). Salt marshes in 180 the York River are dominated by S. alterniflora and S. patens, which are highly productive species that create a dense network of belowground biomass (Perry and Atkinson, 2009;Silliman et al., 2019). We suggest that the overwhelming influence of these highly productive salt marsh species explains the high shear strength of less biodiverse marshes, as these species are absent in our freshwater York River estuary sites along the York River and those studied in the U.K (dominant species: Puccinellia maritima and Juncus gerardii/maritimus). While there was a positive correlation between shear strength and 185 belowground biomass in both the York River and U.K. marshes, the relationship between biodiversity and belowground biomass differs. We find that a decrease in biodiversity leads to an increase in belowground biomass, while the U.K. work determines the opposite trend (Ford et al., 2016). While biodiversity may not be driving soil shear strength variability in this study, it may be a more critical factor in the typically low productivity, high salt marsh platforms present across the U.K. (Ford et al., 2016).
Sea-level rise and saltwater intrusion are impacting wetlands in the York River estuary (Perry and Atkinson, 2009), and globally (Herbert et al., 2015;Neubauer, 2013;Noe and Zedler, 2000). While accelerated rates of sea-level rise could enhance wave erosion (Mariotti and Fagherazzi, 2010), and threaten the survival of marshes (Kirwan and Megonigal, 2013), sea-level rise also leads to changes in vegetation type and productivity (Donnelly and Bertness, 2001;Kirwan et al., 2009;Morris et al., 195 2002). Although these changes will have a variety of ecological and geomorphic consequences, our work suggests that saltwater intrusion may be accompanied by stronger salt marsh soils that are less easily eroded.

Data Availability
Data has been archived to the Long Term Ecological Research repository, and can be found at 200 doi:10.6073/pasta/26c848ab288cc14a2edb106f5800cfc8.   biomass measurement with its associated shear strength value at concurrent depths in the soil profile. Only the relationship between belowground biomass and shear strength in the interior was significant (R 2 = 0.58; p = 1.09e-5).