A significant amount of tidal marsh restoration has occurred over the past two decades. However, restored marshes often fail to recover biological structure and ecosystem functions comparable to reference marshes. We implemented a 13-site inventory to evaluate the recovery of zooplankton and meroplankton abundance and community composition along the Mississippi and Alabama Gulf coasts. Understanding the recovery of zooplankton and meroplankton communities in restored marshes is critical, as many planktonic invertebrate species contribute to nutrient cycling and food web dynamics. We found that zooplankton and meroplankton communities in restored tidal marshes were comparable in total abundance, taxonomic richness, and taxonomic composition to communities observed in reference tidal marshes — with composition being driven mainly by surface water salinity. But zooplankton and meroplankton communities in restored marshes did have lower evenness and diversity than comparable reference marshes. These results suggest that zooplankton and meroplankton communities in restored marshes along the Mississippi and Alabama Gulf coast tend to recover after 7–34 years and support robust populations of prey items for larger, ecologically and economically-important species (e.g., fishes). 

Site Descriptions and Marsh Restoration

We used a series of 13 tidal marshes in coastal wetlands along the Alabama and Mississippi, USA coast to evaluate the diversity and composition of planktonic invertebrate communities in restored tidal marshes of the northern Gulf of Mexico (GOM). Ten of these tidal marshes had brackish surface waters ranging in salinity from 2–13 ppt. The additional three marshes had fresh surface waters with a salinity of 0.1 ppt. All tidal marsh sites were exposed to meteorologically influenced, diurnal micro tides with a maximum amplitude of 0.8 m.

Our 13 tidal marshes were mainly nested within three watersheds along the Mississippi and Alabama coastlines. Three marshes, one natural (Fowl River Natural) and two restored (Fowl River CON-1 and CON-2), were clustered along the West Fowl River (30°21'59.5"N, 88°09'33.2"W) on the western side of Mobile Bay. One natural (Grand Bay Natural) and one restored (Grand Bay Restored) marsh were located approximately 0.5 km from the mouth of Bayou Heron inside the Grand Bay National Estuarine Research Reserve (30°24'43.6"N, 88°24'19.1"W). Finally, one natural (Weeks Bay Natural) and two restored marshes (Weeks Bay Restored and Magnolia Springs) were in Weeks Bay, located on the eastern side of Mobile Bay (30°23'49.2"N 87°49'42.3"W).

Five restored marshes were located outside of these three main waterways: Perdido Beach, Helen Wood Park, Deer Island 1, Deer Island 2, and Greenwood Island. Perdido Beach is a small brackish marsh along Perdido Bay, east of Mobile Bay (30°20'28.5"N, 87°29'56.6"W). Helen Wood Park is a brackish marsh along the western coast of Mobile Bay near the mouth of Dog River (30°34'11.5"N, 88°05'06.9"W). Deer Island 1, Deer Island 2, and Greenwood Island are open-coast marshes located west of Mobile Bay along the Mississippi coastline (Deer Island 1 and 2: 30°22'17.8"N, 88°50'07.5"W; Greenwood Island: 30°20'00.1"N, 88°31'00.7"W).

The ten restored tidal marshes included in our study were developed using three distinct strategies. The two restored tidal marshes along the West Fowl River (Fowl River CON-1 and CON-2) were created in the 1980’s as mitigation efforts for a coal and grain facility by harvesting pine savanna habitat, excavating topsoil down to the water table to allow for hydrological connectivity with the West Fowl River, and planting the area with native marsh plants. Five of the restored tidal marshes were living shorelines (Grand Bay Restored, Helen Wood Park, Weeks Bay Restored, Magnolia Springs, Perdido Beach) that range in age from 7–19 years old and are green infrastructure planted with native marsh plants. Finally, three of our restored tidal marshes were created 7–19 years ago by the Mississippi Department of Marine Resources through the beneficial use of dredge materials (Deer Island 1, Deer Island 2, and Greenwood Island). Specifically, these beneficial use marshes were built by spreading dredge material to establish a marsh platform and allowing vegetation to naturally populate the site.

Most (n = 9) of our tidal marshes had plant communities dominated by native rushes in the genus Juncus. Eight of these marshes were composed mainly of J. roemerianus (a common species in oligohaline and brackish marshes of the northern GOM), and one was composed of J. effusus (a common species in oligohaline marshes of the GOM). The remaining four marshes (Helen Wood Park, Deer Island 1, Deer Island 2, and Greenwood Island) had plant communities dominated by Spartina alterniflora and S. patens, grasses common in brackish and saline marshes along the Atlantic and Gulf coasts. In our study, tidal marshes dominated by Juncus species were mainly oligohaline, while marshes dominated by Spartina species are mainly brackish. At all tidal marshes, dominant plant communities covered 30-95% of the marsh surface. 

Invertebrate Surveys

At all 13 tidal marshes, we surveyed subtidal invertebrate communities using a combination of passive collectors and vertical plankton tows. We used passive collectors to target benthically-recruiting meroplankton communities (e.g., swimming crabs). Vertical plankton tows were used to target the total zooplankton community, which are commonly monitored in restored tidal marshes to understand the availability of prey resources for higher trophic levels.  Both survey techniques were deployed in ~0.5 m of water at every marsh between July and August 2021. We sampled in the late summer since this is when the abundance of juvenile fishes, known to consume zooplankton communities in tidal marshes, peaks throughout Mobile Bay. 

Passive Collectors: We deployed three cylindrical passive hog’s hair collectors at each tidal marsh site. Specifically, the collectors consisted of a 25 cm x 20 cm piece of hog’s hair (Rheem Dust & Pollen High Performance Indoor Air Filter) zip tied around a 76 cm long PVC pipe with a 3.8 cm diameter. Passive hog’s hair collectors have frequently and successfully been used to monitor meroplankton communities in tidal habitats. We deployed the passive collectors in the shallow subtidal adjacent to the tidal marsh platform by inserting the sampler’s PVC into the soil until only the portion covered by the hog’s hair was above the soil. This standardized the portion of the water column that interacted with our passive samplers across tidal marsh sites— the bottom 25 cm of the water column. We chose this orientation to maximize the colonization of our passive collectors by both intertidal and subtidal meroplankton. While our sample size of three collectors per marsh is small, past studies using passive hog’s hair collectors have suggested that three replicates is sufficient to characterize recruiting meroplankton communities.

Passive collectors were deployed in July 2021 and were collected two weeks later in August 2021. Upon collection, we placed each individual passive collector in a 1L glass mason jar filled with 70% ethanol. Collectors remained preserved in 70% ethanol until they were processed by rinsing the hog’s hair filter with 70% ethanol for 5 minutes, with the rinse ethanol and storage ethanol from the mason jar collected in a 33 cm x 23 cm glass pan with a 1 cm x 1 cm grid on the bottom. We let the ethanol in the glass pan settle for 5 minutes before we scanned the pan for any macroinvertebrates (e.g., snails, juvenile crabs, shrimp). When macroinvertebrates were found, they were identified to the lowest taxonomic group possible using a standard key, counted, and removed from the sample at this time. We then surveyed the remaining ethanol for microinvertebrates by randomly sampling at least three, 1 cm² grids in the glass pan for identification. We sampled more than three grid samples if the third subsample included two or more taxa that had not previously been observed on the collector. All microinvertebrates were identified under a dissecting microscope to the lowest taxonomic group possible using a standard key. After identification, proportions of each taxonomic group in the subsamples were extrapolated to total density of each taxon per m².

Vertical Tows: To compare the meroplankton composition on the passive collectors to that available in the zooplankton, we conducted three vertical tows at each site in August 2021. Each vertical tow was conducted in habitat adjacent to the passive collectors [depth = 0.3-0.8 m; volume = 2.4 ± 0.2 L (mean ± 1SE)] using a plankton net with a 30 cm diameter opening and 80 µm Nitex Nylon mesh. Tow volume differed slightly between marshes because marshes varied in their local hydrology. However, zooplankton abundance and diversity were unaffected by tow volume. Samples were preserved in 70% ethanol until processed in the laboratory. We chose to use vertical tows, rather than horizontal tows, because we wanted to sample the same section of the water column that was in contact with our passive collectors to facilitate comparisons of these distinct datasets.

We processed all vertical tows by emptying the full sample into a 33 cm x 23 cm glass pan with a 1 cm x 1 cm grid on the bottom. We let the sample in the glass pan settle for 5 minutes before we randomly sampled at least three, 1 cm ² grids in the glass pan for species composition and density. We sampled more than three grid samples if the third subsample included two or more taxa that had not previously been observed in the sample. We identified all organisms to the lowest taxonomic group possible using a dissecting microscope and a standard key for the northern GOM. After identification, proportions of each taxonomic group in the subsamples were extrapolated to total density of each taxon per m³.

Data Processing

Recovery trajectories: We calculated the total abundance, species richness, species evenness, and Shannon Index of zooplankton and meroplankton communities observed on each passive collector (n = 3) and in each vertical tow (n = 3) collected at every tidal marsh site. Species richness was calculated using the Margalef index, species evenness was calculated using Pielou’s evenness index, and Shannon Index was calculated using a log base ‘e’. We had to exclude one of the 10 restored sites (Helen Wood Park) from our passive collector dataset because two of the collectors deployed at this site were damaged.

To evaluate the recovery of invertebrate communities in tidal marshes along the northern GOM, we calculated the recovery trajectories of planktonic invertebrate total abundance, species richness, species evenness, and Shannon Index. Prior to the analysis, we paired all restored tidal marshes with the most comparable reference tidal marsh sampled. Pairings were based on the relative surface water salinity. We chose to group reference and restored tidal marshes based on surface water salinity because prior studies indicated that salinity influences invertebrate community structure. We obtained response ratios for all diversity indices at each restored tidal marsh site using the equation: ln((Xrest+1)/(Xref+1).

Xrest is the mean value observed in the restored marsh and Xref is the mean value observed in the reference marsh. The value “1” was added to the numerator and denominator to avoid zeros. Positive response ratios thus indicate that restored marsh zooplankton and meroplankton communities outperform reference marsh communities, while negative response ratios indicate that restored marsh zooplankton and meroplankton communities lag reference marsh communities. We chose to calculate the response ratio, rather than directly compare diversity indices between restored and reference tidal marshes, because salinity can affect the diversity of marsh planktonic invertebrate communities — making direct comparisons between fresh and brackish tidal marshes problematic. 

All data are provided in CSV formats and should be usable in basic programs, like Excel.