October 2016 v Volume 7(10) v Article e014871 v www.esajournals.org
Disturbance in Georgia salt marshes: variation across space and time
Shanze Li1,2 and Steven C. Pennings2,†
1State Key Joint Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875 China
2Department of Biology and Biochemistry, University of Houston, Houston, Texas 77204 USA
Citation: Li, S., and S. C. Pennings. 2016. Disturbance in Georgia salt marshes: variation across space and time. Ecosphere 7(10):e01487. 10.1002/ecs2.1487
Abstract. We documented the frequency and effect on live biomass of five different types of disturbance over 14 years in creekbank and mid- marsh zones of eight salt marshes dominated by Spartina alterniflora in Georgia, USA. Wrack (floating debris) and creekbank slumping were the most common disturbances at the creekbank, and snails were the most common disturbance agent in the mid- marsh. Disturbance frequency varied among sites due to differences in plot elevation and landscape position. Wrack disturbance at the creekbank was positively correlated with plot elevation, and both initial slumping and terminal slumping of creekbank plots were negatively correlated with plot elevation. Wrack disturbance at the creekbank and snail disturbance in the mid- marsh were also most common at barrier island vs. interior marshes. Distur-bance varied up to 14- fold among years. Wrack disturbance at the creekbank was negatively correlated with river discharge and sea level, and initial slumping of creekbank plots was also negatively correlated with sea level. The different disturbance types varied in their effects on end- of- year standing plant bio-mass. At the creekbank, wrack disturbance reduced biomass in affected plots by ~46%, but slumping did not affect biomass until the plot was totally lost. In the mid- marsh, slumping and wrack were not import-ant disturbances, but snail disturbance reduced biomass in affected plots by ~70%. In addition, abiotic conditions (river discharge, maximum monthly temperature, sea level, and precipitation) strongly affected year- to- year variation in biomass. Across the entire landscape, fewer than a quarter of the plots on average were disturbed, and disturbance reduced overall standing biomass by ~18% in the creekbank zone and ~3% in the mid- marsh zone. Our results indicate that wrack has fairly strong effects on end- of- year bio-mass at the creekbank. Overall, however, variation in abiotic conditions among years had stronger effects on end- of- year standing biomass in both marsh zones than did disturbance.
Natural disturbances play important roles in many coastal ecological systems, including kelp forests, coral reefs, mangroves, oyster reefs, and sea grass beds (Short and Wyllie- Echeverria 1996, Kimbro and Grosholz 2006, Wilson et al. 2006, Hoegh- Guldberg et al. 2007, Reed et al. 2011, Guo et al. 2013, Pickett and White 1985). Disturbance can strongly affect primary production (Reed
et al. 2011), species composition (Shumway and Bertness 1994, Platt et al. 2015), diversity (Strong 1977, Connell 1978), and biogeochemical cycling (Bernhard et al. 2015). For logistical reasons, many studies have investigated the effects of disturbance at small spatial and short temporal scales (Dayton 1971, Bertness and Ellison 1987, Fischer et al. 2000), and we have a weaker under-standing of how disturbance varies at the land-scape level or over longer periods of time.
October 2016 v Volume 7(10) v Article e014872 v www.esajournals.org
LI AND PENNINGS
In coastal salt marshes, a variety of distur-bances affect plant and animal communities and ecosystem function (Valiela and Rietsma 1995, Silliman et al. 2005, McFarlin et al. 2015, Reidenbaugh and Banta 2015). Common dis-turbances in salt marshes around the world include deposition of floating wrack (Bertness and Ellison 1987, Tolley and Christian 1999), ice damage (Pennings and Bertness 2001), creek-bank slumping (Pethick 1974), fire (Smith and Kadlec 1985, Smith et al. 2013), and herbivory or excavation by wildlife (Ellison 1987, Pennings et al. 2009). By killing vegetation, providing nutrients, and altering abiotic conditions, these disturbances affect species composition and suc-cession (Bertness and Ellison 1987, Bertness et al. 1992), and primary production (Pennings and Richards 1998). Many of these studies empha-size higher marsh elevations that have higher plant species richness. In addition, with a few exceptions (Valiela and Rietsma 1995, Brewer et al. 1998, Fischer et al. 2000), these studies were conducted at single sites over one or two years. Finally, few of these studies compared distur-bance with other factors that might affect plant productivity. Thus, despite considerable interest in disturbance in coastal salt marshes, we lack an understanding of how frequent disturbance is across all marsh elevations, how it varies among sites and over time, and what types of distur-bance and abiotic variables (elevations, physical locations, and climate) are most important in affecting plant productivity. As a consequence, we also do not know whether the interest in dis-turbance in coastal salt marshes reflects its eco-logical importance in this habitat compared to other systems.
To address these issues, we examined distur-bance to permanent plots over 14 years at eight salt marshes in coastal Georgia, USA, that were dominated by the grass Spartina alterniflora. We tested the hypotheses that (1) the type and fre-quency of disturbance would vary among sites, marsh zones, and years; (2) the frequency of disturbance would vary as a function of abiotic conditions including climate and sea level; (3) different types of disturbance would have differ-ent effects on standing biomass in affected plots; and (4) disturbance would be similar in impor-tance to local climate and abiotic drivers in affect-ing variation in plant productivity.
Methods
Our study area was within the domain of the Georgia Coastal Ecosystems Long- Term Eco-logical Research (GCE- LTER) program (http:// gce-lter.marsci.uga.edu/). The GCE- LTER pro-gram includes 10 permanent sites (GCE1–10) that inc lude fresh, brackish, and salt marshes that are dominated by the plant species Zizaniopsis milia-cea, Spartina cynosuroides and Juncus roemerianus, and Spartina alterniflora, respectively. Tides are semidiurnal and mesotidal, with a range of 2–3 m. We focused on the eight sites where the creekbank or marsh platform was dominated by S. alterniflora (creekbank zones at GCE sites 1, 2–6, and 9–10; mid- marsh zones at GCE sites 2–6 and 9–10, Fig. 1; Appendix S1: Fig. S1). At each site, we located eight plots at the creekbank and eight in the mid- marsh. S. alterniflora height var-ies with elevation, and ecologists often distin-guish “tall Spartina” (typically at the creekbank, >100 cm tall), “medium Spartina” (50–99 cm tall), and “short Spartina” (<50 cm tall) (Schalles et al. 2013). Using these criteria, mid- marsh plots at sites 2–4 would probably be categorized as short Spartina, mid- marsh plots at sites 5, 6, 9, and 10 would be categorized as medium Spartina, and
Fig. 1. Map of the study site. GCE- LTER perma-nent monitoring sites that were included in this study are marked with filled circles. GCE 3 and 6 were coded as barrier island sites; 2, 5, 9, and 10 as intermediate; and 1 and 4 as mainland sites.
October 2016 v Volume 7(10) v Article e014873 v www.esajournals.org
LI AND PENNINGS
creekbank plots at all the sites would be catego-rized as tall Spartina. Elevations of plots varied among sites, but with considerable overlap between sites due to variation among plots within each site (Appendix S1: Table S5). We elected not to include the various different types of high marsh vegetation present at these sites in our sampling design so as to maximize our abil-ity to compare disturbance to a single dominant vegetation type across as many sites as possible. We measured shoot heights and flowering status of S. alterniflora every October from 2000 to 2014 and used allometric relationships (Appendix S1: Table S1) to estimate standing biomass (Więski and Pennings 2013). Since 2000 was the first year of the project, all plots were initially undisturbed, so disturbances were recorded from 2001 to 2014.
Each year when standing biomass was esti-mated in the plots, we also recorded distur-bances due to wrack (floating debris, usually consisting of the stems of dead plants), creek-bank slumping (the collapse of the creekbank toward the bottom of the creek), or aggregations of snails in each plot. The decision to mark a plot as disturbed was based on clear evidence of the purported disturbance agent. For exam-ple, we recorded plots as disturbed by wrack if we found abundant wrack in the plot or broken plant stems indicative of plants having been crushed by wrack. We recorded plots as slump-ing (initial slump) based on crevices in the marsh or displaced plot stakes indicating that the plot was sliding into the creek. We also recorded plots as slumping (terminal slump) if the plot stakes had slipped beyond the lower distribu-tion limit of intertidal vegetation (in these cases, we scored standing biomass as zero and estab-lished a replacement plot at higher elevations in the same general location). Finally, we recorded plots as disturbed by snails if we observed high densities of snails in the plot (>~250/m2) coupled with visual observation of snail grazing damage to plants (Silliman and Zieman 2001, Schalles et al. 2013). Thus, we did not score plots as dis-turbed if the plots contained small amounts of wrack or normal densities of snails without evi-dence of damage to vegetation, and we might have missed disturbances that happened ear-lier in the growing season if plants had mostly recovered and the evidence of the disturbance had disappeared.
We qualitatively compared the frequency of different types of disturbance among years, sites, and marsh zones from 2001 to 2014. To deter-mine how disturbance affected the fall standing biomass of S. alterniflora, we compared biomass of plots experiencing the more common types of disturbances at the creekbank (wrack distur-bance, initial slump, terminal slump, both wrack disturbance and initial slump, and no distur-bance) and in the mid- marsh (wrack disturbance, snail disturbance, and no disturbance) using one- way analysis of variance (ANOVA) with all years combined. Data were ln(x + 1)- transformed before analysis to improve homogeneity of variance.
We collected data on six factors that might affect disturbance and plant biomass (site, ele-vation of each plot, temperature, precipitation, sea level, and river discharge) from GCE- LTER data sets or public sources. We used maximum temperature instead of mean temperature as a predictor because maximum temperature was previously shown to be a better predictor of plant biomass (Więski and Pennings 2013). Site loca-tions, elevation of each plot, temperature, precip-itation, and river discharge are accessible on the GCE- LTER Web site (http://gce-lter.marsci.uga.edu/portal/monitoring.htm). Sea level is avail-able on the Web site of National Oceanographic and Atmospheric Administration (8670870 Fort Pulaski, Georgia, http://www.noaa.gov/). Data sources are described more fully in Appendix S1: Tables S2–S5. Predictors that varied during each year (river discharge, sea level, precipitation, and maximum temperature) were averaged over the growing season (April–September) to provide a single value per year. We addressed the con-tributions of abiotic (landscape location, plot elevation) or climate factors (sea level and river discharge) in predicting disturbance frequency using stepwise logistic regression. We started with a full model and sequentially removed terms with P > 0.25 at each step.
We examined the importance of the six envi-ronmental factors in combination with distur-bance type (wrack disturbance, snail disturbance, initial slump, and terminal slump) on plant bio-mass using analyses of covariance (ANCOVAs) for the creekbank and mid- marsh zones sepa-rately. Plots (8 zone−1 year−1 site−1) were used as the unit of replication. In the creekbank zone (n = 1015), we modeled S. alterniflora biomass
October 2016 v Volume 7(10) v Article e014874 v www.esajournals.org
LI AND PENNINGS
as the dependent variable, with wrack distur-bance, initial slump, terminal slump, and site as fixed factors, and plot elevation, river discharge, sea level, precipitation, and maximum monthly temperature as covariates (no creekbank plots were disturbed by snails). In the mid- marsh zone (n = 837), we modeled S. alterniflora biomass as the dependent variable, with snail disturbance and site as fixed factors (mid- marsh plots did not experience slumping disturbance, and wrack disturbance was rare, n = 6), and plot elevation, river discharge, sea level, precipitation, and max-imum temperature as covariates. Plot elevation and sea level were removed from the mid- marsh model because they were not significant (Abdala- Roberts and Mooney 2014). All statistical analyses were performed in SPSS 22 (Allen et al. 2014).
results
Standing biomass of Spartina alterniflora was ~2.6 times higher in the creekbank vs. the mid- marsh zone (Fig. 2A). Standing biomass varied ~threefold among years in the creekbank zone and ~twofold among years in the mid- marsh zone.
Across all years and disturbance types, the fre-quency of disturbance was ~eight times higher in the creekbank zone than in the mid- marsh zone (Fig. 2B). Disturbance varied ~fivefold among years in the creekbank zone and ~11- fold among years in the mid- marsh zone.
The frequency of different types of disturbance varied between zones, among sites, and among years. The most common disturbances in the
Fig. 2. (A) Variation in Spartina alterniflora fall standing biomass, all sites combined. (B–H) Frequency of different disturbances in creekbank and mid- marsh zones.
October 2016 v Volume 7(10) v Article e014875 v www.esajournals.org
LI AND PENNINGS
creekbank were wrack, which affected about 15% of plots, and slumping, which affected about 16% of plots, with initial slumping and terminal slumping equally common (Fig. 2C). The most common disturbance in the mid- marsh was snails, which affected about 4% of the plots over-all (Fig. 2D). The four disturbance types varied in different ways among sites (Fig. 2C, D) and years, with wrack and snail disturbance most important early in the study period, and termi-nal slumping most important late in the study period (Fig. 2E–H).
Wrack disturbance at the creekbank was more common at barrier island sites than at sites fur-ther inland (Fig. 3A) and was also affected pos-itively by plot elevation, negatively by river discharge, and negatively by sea level (Table 1; Appendix S1: Fig. S2A–C). Initial slumping at the creekbank was affected negatively by plot elevation and negatively by sea level (Table 1; Appendix S1: Fig. S2D, E). Terminal slumping at the creekbank was affected negatively by plot elevation (Table 1; Appendix S1: Fig. S2F). Snail disturbance in the mid- marsh was more common at barrier island sites than at sites further inland (Fig. 3B) and was negatively affected by plot ele-vation (Table 1; Appendix S1: Fig. S2G).
Spartina alterniflora biomass was reduced by ~46% in creekbank plots experiencing wrack disturbance compared to undisturbed plots (Fig. 4A). Initial slumping had no effect on bio-mass, but terminal slumping resulted in total biomass loss. In the mid- marsh zone, wrack dis-turbance did not affect standing biomass, but snail disturbance reduced biomass by ~70% in affected plots (Fig. 4B).
We multiplied the frequency of disturbance by its effect on standing biomass in affected plots to calculate the total effect of disturbance on S. alterniflora biomass at the landscape level (Table 2). Depending on the site, wrack
disturbance reduced biomass in the creekbank zone by 3–15%, loss of creekbank plots due to terminal slump reduced biomass by 2–16%, and snails reduced biomass in the mid- marsh zone by 1–7%. On average, across all sites, years, and types of disturbance, disturbance reduced bio-mass by ~18% in the creekbank zone and ~3% in the mid- marsh zone.
Both disturbance and abiotic drivers helped explain variation in S. alterniflora biomass. Creekbank biomass was negatively affected by wrack disturbance, terminal slumping, maximum
Fig. 3. (A) Variation of wrack disturbance frequency among different landscape locations at the creekbank zone and (B) variation of snail disturbance frequency among different landscape locations at the mid- marsh zone.
Table 1. Logistic regression relationships between disturbances and abiotic drivers.
Note: Values are regression coefficients and P-values; only significant predictors are shown.
October 2016 v Volume 7(10) v Article e014876 v www.esajournals.org
LI AND PENNINGS
monthly temperature, and plot elevation and positively affected by river discharge, sea level, and precipitation (Table 3). Plant biomass in the mid- marsh zone was negatively correlated with snail disturbance and maximum monthly tem-perature and positively correlated with local pre-cipitation and river discharge (Table 3).
dIscussIon
We found that type, frequency, and importance of disturbance varied strongly and predictably
among marsh zones, sites, and years. Overall, plant biomass in these salt marshes was more strongly affected by abiotic drivers than by dis-turbance, in contrast to some other ecological sys-tems where productivity is strongly dependent on disturbance (Reed et al. 2011).
Disturbance was more common at the creek-bank than in the mid- marsh. Many previous studies of salt marsh disturbance have focused on the importance of wrack disturbance in the high marsh, where floating wrack can be deposited by spring tides, remain in place for weeks, and affect vegetation composition (Bertness et al. 1992). Both Bertness and Ellison (1987), working in Rhode Island, and Valiela and Rietsma (1995), working in Massachusetts, found wrack disturbance to be greatest in the high marsh. Both studies, moreover, concluded that wrack disturbance in the low marsh did not strongly affect plant community compo-sition because wrack was relatively rare, wrack patches were regularly moved by high tides, and plant species richness was low. Although uninter-esting from a community ecology point of view, creekbank disturbance by wrack may neverthe-less strongly affect plant biomass. We found much higher rates of wrack disturbance at the creekbank than in the mid- marsh. Creekbank plots often had modest rather than heavy mats of wrack at the time that we sampled the plots, suggesting that the heavy mat of wrack that initially disturbed the plot had since been largely transported else-where, consistent with the view of past workers that wrack in the low marsh is relatively transient. Nevertheless, the tall stems of S. alterniflora along the creekbank are vulnerable to breaking when crushed by mats of wrack, even if the wrack mat persists in a given location for only a single low tide. Thus, even if wrack does not persist long at any single location along the creekbank and does not affect species composition, it may nevertheless have strong effects on standing biomass and pro-ductivity. Moreover, if wrack mats at the creek-bank move around frequently, the cumulative disturbance to the creekbank zone may be much greater than that suggested by a snapshot survey of wrack abundance at any one time.
Other disturbance types also varied with marsh zone. Because the mid- marsh is almost flat, slumping did not occur in the mid- marsh zone. In contrast, snail disturbance was only observed in the mid- marsh, because predators
Fig. 4. Biomass of Spartina alterniflora plots experi-encing different disturbances in creekbank and mid- marsh zones (all years and sites combined). Data are means ± 1 SE (n = 149, 51, 74, 14, and 727 for wrack disturbance, initial slump, terminal slump, wrack + initial slump, and no disturbance plots, respectively, in the creekbank zone, and n = 6, 29, and 802 for wrack disturbance, snail disturbance, and no disturbance plots, respectively, in the mid- marsh). Bars not shar-ing a letter are significantly different from one another (P < 0.05).
October 2016 v Volume 7(10) v Article e014877 v www.esajournals.org
LI AND PENNINGS
typically limit snails at the creekbank to densi-ties below those at which snails harm vegetation (Silliman and Bertness 2002).
Disturbance also varied among sites and across the landscape. We know little about how dis-turbance varies among salt marsh sites, because most previous studies focused only on one or a few sites. Brewer et al. (1998) found variation in disturbance among three salt marshes in Rhode Island, but did not study enough sites to gener-alize their results. Fischer et al. (2000) studied spatial variation in wrack disturbance in salt marshes at Sapelo Island and found that it was highly dependent on small- scale aspects of creek-bank morphology. Our plots were located along relatively straight sections of creeks and did not include sufficient variation in creekbank mor-phology to assess its importance. However, we found considerable variation among sites in all four disturbance types, with wrack disturbance
varying ~12- fold, initial slumping varying ~eight-fold, terminal slumping varying ~18- fold, and snail disturbance varying ~sixfold among sites.
Variation among sites is probably largely due to differences in plot elevation and to landscape position. Wrack disturbance at the creekbank was positively correlated with plot elevation, probably because wrack was more likely to be stranded in plots that were higher in the inter-tidal frame. Both initial slumping and termi-nal slumping at the creekbank were negatively correlated with plot elevation, perhaps simply reflecting the fact that lower elevation plots are the closest to the creeks. Finally, snail disturbance in the mid- marsh was negatively correlated with plot elevation; however, this correlation had a very poor fit and was highly influenced by a rela-tively small number of plots at three or four sites (Fig. S2G), and we therefore do not place a high level of confidence in this result.
Table 2. Effects of wrack, initial slump, terminal slump, and snail disturbances on standing biomass of Spartina alterniflora in creekbank and mid- marsh zones, respectively.
Notes: Data are percentage reductions in biomass at each zone at each site by the various disturbances. Only results for sta-tistically significant disturbances are shown. ND = no data because this zone at site 1 is not dominated by S. alterniflora.
Table 3. ANCOVA table estimating effects of different disturbance and climate drivers on plant biomass in the creekbank and mid- marsh plots (n = 1015 and 837, respectively).
Creekbank (adjusted R2 = 0.47) Mid- marsh (adjusted R2 = 0.24)Source F P Sign Source F P Sign
Note: Variables that were not significant were removed from the models.
October 2016 v Volume 7(10) v Article e014878 v www.esajournals.org
LI AND PENNINGS
Wrack and snail disturbance also varied pre-dictably across the landscape. Wrack distur-bance was most common on barrier island sites, likely because the net transport of wrack by river flow is offshore. Snail disturbance was also more common on barrier island sites than main-land marshes, likely because of increased larval recruitment at sites closer to the ocean (B. R. Silliman, personal communication). The overall frequency of significant Littoraria irrorata snail disturbance that we observed is consistent with the study of Schalles et al. (2013), who estimated that 4.5% of the marsh (all vegetation types on the creek- shed on the west side of Sapelo Island) had snail densities >250/m2, but is lower than one might expect from studies of the impact that L. irrorata has on marsh vegetation in South Atlantic salt marshes, which often report aver-age snail densities of >300/m2 (Silliman and Bertness 2002, Silliman and Bortolus 2003). It is likely that this discrepancy is largely explained by these studies working with L. irrorata where it is most abundant in the short Spartina zone at the upper reaches of the intertidal range dom-inated by S. alterniflora. In contrast, the major-ity of the mid- marsh plots in this study were located in medium Spartina, where snails are considerably less abundant (77 vs. 237/m2 in short Spartina, Schalles et al. 2013).
Finally, disturbance varied up to 14- fold among years, and different types of disturbance varied out of synchrony with each other. Most previous studies of disturbance in salt marshes have sam-pled over only one or a few years and therefore are likely to poorly estimate the true frequency of disturbance. Wrack disturbance at the creekbank was negatively correlated with river discharge, perhaps because wrack is transplanted offshore during periods of high river discharge. Wrack disturbance at the creekbank was also negatively correlated with sea level, likely for the same reason that it was positively correlated with plot elevation. Both of these variables address how deeply the marsh is flooded, with flood-ing greater in years of high sea level or at plots located at low elevation. In both cases, wrack is more likely to be transported over the creekbank plots and into the remainder of the marsh or to be exported from the marsh and out to sea.
Initial slumping of creekbanks was negatively correlated with sea level, perhaps because periods
of low sea level lead to increased pore water drainage, and dewatered soils have a lower shear strength (Gallage and Uchimura 2006). Terminal slumping, however, was not correlated with sea level. It is likely that terminal slumping, which indicates that the slumping plot has reached ele-vations below which S. alterniflora can survive, is a function of multiple additional variables, such as creekbank steepness and wave exposure, which were not measured in this study.
One caveat of our study is that we only scored disturbances that were obvious in October when we sampled the plots. Thus, we would have missed disturbances that occurred early in the growing season if the vegetation was largely recovered, and we would have ignored small disturbances that had only weak effects on the vegetation. As our interest, however, was how disturbance affected end- of- year standing bio-mass, neither of these would have had a large effect on our results.
How important are disturbances vs. abiotic factors in controlling end- of- year biomass of salt marsh plants? Several previous studies have examined the importance of abiotic factors in mediating the productivity of salt marsh plants (Morris et al. 1990, Więski and Pennings 2013). In addition, several studies have looked at the effect of disturbances on plant productivity at the plot scale (Tolley and Christian 1999, Reidenbaugh and Banta 2015). For example, Silliman and Zieman (2001) found that snail grazing reduced Spartina biomass by 51–66% (similar to our esti-mate of 70%). None of these previous studies of disturbance in salt marshes, however, took a landscape- scale perspective, making it difficult to scale their results up. And none of these past studies integrated the importance of abiotic fac-tors and disturbance.
We estimated the effect of disturbance at the landscape scale by multiplying the frequency of disturbance at each site with its overall effect on standing biomass. This calculation suggested that disturbance was of moderate importance in mediating standing biomass at the land-scape scale. On average, fewer than a quarter of the plots were disturbed, and only some disturbances (wrack and terminal slumping at the creekbank and snails in the mid- marsh) strongly reduced standing biomass. The total effect of disturbance was to reduce standing
October 2016 v Volume 7(10) v Article e014879 v www.esajournals.org
LI AND PENNINGS
biomass by ~18% at the creekbank and ~3% in the mid- marsh. Although this is an important effect, abiotic factors such as maximum tem-perature, river discharge, and sea level produce much greater variation in end- of- year standing biomass from year to year (Morris et al. 1990, Więski and Pennings 2013). High temperatures in salt marshes may reduce productivity if they are beyond the thermal optimum for photosyn-thesis of Spartina (Giurgevich and Dunn 1982, Więski and Pennings 2013). River discharge and sea level affect productivity because they mediate salinity of the pore water surrounding the plant roots (Morris et al. 1990, Więski and Pennings 2013). In comparison, disturbance is not common enough across the landscape to strongly affect standing biomass. Recovery of S. alterniflora from disturbance depends on dis-turbance size, abiotic conditions, and top- down pressure (Angelini and Silliman 2012, Silliman et al. 2013), all of which likely vary among sites and years. Thus, not only disturbance but also recovery is likely to be spatially and temporally variable.
We are aware of few studies in other systems that have explicitly contrasted the effects of abi-otic factors and disturbance on primary produc-tion. One notable exception is a study of giant kelp in California, where disturbance (storm waves) overwhelmed bottom- up (nutrient availability) and top- down (grazing pressure) factors in determining net primary production (Reed et al. 2011). Our study compared effects of abiotic factors and disturbance on end- of- year biomass, but did not explicitly consider the role of consumers other than snails. A vari-ety of herbivorous insects and crabs also affect S. alterniflora biomass in southeastern Atlantic salt marshes (Pennings et al. 2012) and so the next step in understanding S. alterniflora pro-ductivity would be to integrate the effects of these consumers.
In summary, our results indicate that distur-bance in coastal salt marshes dominated by S. alterniflora varies among marsh zones, plots, sites, and years. Previous studies that have focused on one or another type of disturbance, often over a limited range of marsh elevations, sites, or years, have been important in under-standing the mechanisms by which disturbance affects marsh structure and function. Our work
has begun to place these previous studies in a broader context and has shown that while dis-turbance can be locally important in affecting marsh productivity, abiotic drivers have a stron-ger effect on standing plant biomass across the entire landscape.
AcknowledgMents
This research was funded by the National Science Foundation (OCE99- 82133, OCE06- 20959, OCE12- 37140). Shanze Li was able to participate thanks to the National Key Basic Research Program of China (2013CB430406) and the China Scholarship Council. We thank everyone who helped sample the permanent plots over the years and thank Merryl Alber, Brian Silliman, and Kazimerz Więski for helpful comments on the manuscript. This work is a contribution of the Georgia Coastal Ecosystems LTER program and con-tribution number 1051 of the University of Georgia Marine Institute.
lIterAture cIted
Abdala-Roberts, L., and K. A. Mooney. 2014. Ecologi-cal and evolutionary consequences of plant geno-type diversity in a tri- trophic system. Ecology 95: 2879–2893.
Allen, P., K. Bennett, and B. Heritage. 2014. SPSS Sta-tistics version 22: a practical guide. Cengage Learn-ing, South Melbourne, Victoria, Australia.
Angelini, C., and B. R. Silliman. 2012. Patch size- dependent community recovery after massive dis-turbance. Ecology 93:101–110.
Bernhard, A. E., C. Dwyer, A. Idrizi, G. Bender, and R. Zwick. 2015. Long- term impacts of disturbance on nitrogen- cycling bacteria in a New England salt marsh. Frontiers in Microbiology 6:46. http://dx.doi.org/10.3389/fmicb.2015.00046
Bertness, M. D., and A. M. Ellison. 1987. Determinants of pattern in a New England salt marsh plant com-munity. Ecological Monographs 57:129–147.
Bertness, M. D., L. Gough, and S. W. Shumway. 1992. Salt tolerances and the distribution of fugitive salt marsh plants. Ecology 73:1842–1851.
Brewer, J. S., J. M. Levine, and M. D. Bertness. 1998. Interactive effects of elevation and burial with wrack on plant community structure in some Rhode Island salt marshes. Journal of Ecology 86: 125–136.
Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302–1310.
Dayton, P. K. 1971. Competition, disturbance, and community organization: the provision and
October 2016 v Volume 7(10) v Article e0148710 v www.esajournals.org
LI AND PENNINGS
sub sequent utilization of space in a rocky intertidal community. Ecological Monographs 41:351–389.
Ellison, A. M. 1987. Effects of competition, distur-bance, and herbivory on Salicornia europaea. Ecol-ogy 68:576–586.
Fischer, J. M., T. Reed-Andersen, J. L. Klug, and A. G. Chalmers. 2000. Spatial pattern of localized distur-bance along a southeastern salt marsh tidal creek. Estuaries 23:565–571.
Gallage, C. P. K., and T. Uchimura. 2006. Effects of wetting and drying on the unsaturated shear strength of a silty sand under low suction. Unsat-urated Soils 2006:1247–1258. Proceedings of the Fourth International Conference on Unsaturated Soils, held in Carefree, Arizona, April 2–6, 2006. http://dx.doi.org/10.1061/40802(189)102
Giurgevich, J. R., and E. L. Dunn. 1982. Seasonal pat-terns of daily net photosynthesis, transpiration and net primary productivity of Juncus roemerianus and Spartina alterniflora in a Georgia salt marsh. Oeco-logia 52:404–410.
Guo, H., Y. Zhang, Z. Lan, and S. C. Pennings. 2013. Biotic interactions mediate the expansion of black mangrove (Avicennia germinans) into salt marsh-es under climate change. Global Change Biology 19:2765–2774.
Hoegh-Guldberg, O., et al. 2007. Coral reefs under rap-id climate change and ocean acidification. Science 318:1737–1742.
Kimbro, D. L., and E. D. Grosholz. 2006. Disturbance influences oyster community richness and even-ness, but not diversity. Ecology 87:2378–2388.
McFarlin, C. R., T. D. Bishop, M. W. Hester, and M. Alber. 2015. Context- dependent effects of the loss of Spartina alterniflora on salt marsh inverte-brate communities. Estuarine, Coastal and Shelf Science 163:218–230.
Morris, J. T., B. Kjerfve, and J. M. Dean. 1990. Dependence of estuarine productivity on anomalies in mean sea level. Limnology and Oceanography 35:926–930.
Pennings, S. C., and M. D. Bertness. 2001. Salt marsh communities. Pages 289–316 in M. D. Bertness, S. D. Gaines, and M. E. Hay, editors. Marine Com-munity Ecology, Sinauer Associates, Sunderland, Massachusetts, USA.
Pennings, S. C., C.-K. Ho, C. S. Salgado, K. Wieski, N. Davé, A. E. Kunza, and E. L. Wason. 2009. Lati-tudinal variation in herbivore pressure in Atlantic Coast salt marshes. Ecology 90:183–195.
Pennings, S. C., and C. L. Richards. 1998. Effects of wrack burial in salt- stressed habitats: Batis mariti-ma in a southwest Atlantic salt marsh. Ecography 21:630–638.
Pennings, S. C., et al. 2012. South Atlantic tidal wetlands. Pages 45–61 in A. Baldwin and D. Batzer,
editors. Wetland habitats of North America: ecol-ogy and conservation concerns. University of California Press, Berkeley, California, USA.
Pethick, J. 1974. The distribution of salt pans on tidal salt marshes. Journal of Biogeography 1: 57–62.
Pickett, S. T., and P. S. White. 1985. The ecology of natural disturbance and patch dynamics. Academic Press, Orlando, USA.
Platt, W. J., D. Joseph, and D. P. Ellair. 2015. Hurricane wrack generates landscape- level heterogeneity in coastal pine savanna. Ecography 38:63–73.
Reed, D. C., A. Rassweiler, M. H. Carr, K. C. Cavanaugh, D. P. Malone, and D. A. Siegel. 2011. Wave disturbance overwhelms top- down and bottom- up control of primary production in Cali-fornia kelp forests. Ecology 92:2108–2116.
Reidenbaugh, T. G., and W. C. Banta. 2015. Origins and effects of Spartina wrack in a Virginia salt marsh. Gulf and Caribbean Research 6:393–401.
Schalles, J. F., C. M. Hladik, A. A. Lynes, and S. C. Pennings. 2013. Landscape estimates of habitat types, plant biomass, and invertebrate densities in a Georgia salt marsh. Oceanography 26:85–87.
Short, F. T., and S. Wyllie-Echeverria. 1996. Natural and human- induced disturbance of seagrasses. Environmental Conservation 23:17–27.
Shumway, S. W., and M. D. Bertness. 1994. Patch size effects on marsh plant secondary succession mech-anisms. Ecology 75:564–568.
Silliman, B. R., and M. D. Bertness. 2002. A trophic cascade regulates salt marsh primary production. Proceedings of the National Academy of Sciences USA 99:10500–10505.
Silliman, B. R., and A. Bortolus. 2003. Underestimation of Spartina productivity in western Atlantic marsh-es: marsh invertebrates eat more than just detritus. Oikos 101:549–554.
Silliman, B. R., M. W. McCoy, C. Angelini, R. D. Holt, J. N. Griffin, and J. Van de Koppel. 2013. Consumer fronts, global change, and runaway collapse in eco-systems. Annual Review of Ecology, Evolution and Systematics 44:503–538.
Silliman, B. R., J. Van de Koppel, M. D. Bertness, L. E. Stanton, and I. A. Mendelssohn. 2005. Drought, snails, and large- scale die- off of southern US salt marshes. Science 310:1803–1806.
Silliman, B. R., and J. C. Zieman. 2001. Top- down control of Spartina alterniflora production by peri-winkle grazing in a Virginia salt marsh. Ecology 82:2830–2845.
Smith III, T. J., A. M. Foster, G. Tiling-Range, and J. W. Jones. 2013. Dynamics of mangrove–marsh eco-tones in subtropical coastal wetlands: fire, sea- level rise, and water levels. Fire Ecology 9:66–77.
October 2016 v Volume 7(10) v Article e0148711 v www.esajournals.org
LI AND PENNINGS
Smith, L. M., and J. A. Kadlec. 1985. Fire and herbivory in a Great Salt Lake marsh. Ecology 66: 259–265.
Strong Jr., D. R. 1977. Epiphyte loads, tree falls, and perennial forest disruption: a mechanism for main-taining higher tree species richness in the trop-ics without animals. Journal of Biogeography 4: 215–218.
Tolley, P. M., and R. R. Christian. 1999. Effects of inc-reased inundation and wrack deposition on a high salt marsh plant community. Estuaries 22:944–954.
Valiela, I., and C. S. Rietsma. 1995. Disturbance of salt marsh vegetation by wrack mats in Great Sip-pewissett Marsh. Oecologia 102:106–112.
Więski, K., and S. C. Pennings. 2013. Climate drivers of Spartina alterniflora saltmarsh production in Geor-gia, USA. Ecosystems 17:473–484.
Wilson, S. K., N. A. Graham, M. S. Pratchett, G. P. Jones, and N. V. Polunin. 2006. Multiple distur-bances and the global degradation of coral reefs: Are reef fishes at risk or resilient? Global Change Biology 12:2220–2234.
supportIng InforMAtIon
Additional Supporting Information may be found online at: http://onlinelibrary.wiley.com/doi/10.1002/ecs2.1487/supinfo
