Southeastern Naturalist 463 E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 22001144 SOUTHEASTERN NATURALIST 1V3o(3l.) :1436,3 N–4o7. 43 Small-Mammal Occupancy in Freshwater Marshes of Mandalay National Wildlife Refuge, Louisiana Eric J. Tobin1,2,*, Jenneke M. Visser1,3, James K. Peterson3, and Paul L. Leberg2 Abstract - Small mammals are key consumers in the marsh food web and could serve as indicators of a marsh’s potential to support higher-level predators. We studied how smallmammal occupancy varied among plant communities in coastal Louisiana freshwater marshes. We sampled small mammals at 36 sites on 4 different occasions during the late spring in freshwater marshes of the Mandalay National Wildlife Refuge, LA. Mammalian diversity was low; we captured only Oryzomys palustris (Marsh Rice Rats). Occupancy modeling revealed a positive association between Marsh Rice Rat site occupancy and Sagittaria lancifolia (Bulltongue Arrowhead) biomass. Our data suggest that subtle changes in plant-species composition within a marsh may affect the distribution of the most common small mammal in the ecosystem. Introduction Coastal Louisiana has undergone drastic habitat modification during the last century, including major conversion of wetlands to open water (Barras et al. 2008, Mitsch et al. 2009). In particular, freshwater marshes in Louisiana have experienced major shifts in plant-species composition (Visser et al. 1999). Driving factors behind these changes include water-level increase, salinity alterations, grazing behavior by native and invasive species, lack of particulate deposition, and oil and gas extraction activities (Gosselink et al. 1998, Penland et al. 2001, Visser et al. 2012). In Louisiana’s Terrebonne Basin, these complex and interacting environmental changes have lead to a shift in freshwater-marsh vegetation types, such as the replacement of Panicum hemitomon Schul. (Maidencane) with Eleocharis spp. (spikerush) as the dominant plant (Visser et al. 1999). Wildlife investigations and management activities in coastal wetlands are often concerned with species responses across salinity zones due to a lack of fine-scale habitat-use data (e.g., King and Michot 2002, Miller et al. 2009, Plattner et al. 2010, Ysebaert et al. 2000). For wildlife management purposes, the Louisiana coastal zone is divided into 5 habitats based on salinity: Taxodium sp. (cypress)–Nyssa sp. (tupelo) swamp, and fresh, brackish, intermediate, and saline marshes (Palmisano 1972). Although cypress–tupelo swamps are classified as freshwater wetlands, they are distinct from the other coastal-zone habitats because they are forests. Vegetation-community structure varies widely within each of these marsh habitats; Bulltongue, Maidencane, and Spikerush are all types of freshwater marshes (Visser 1Institute for Coastal Ecology and Engineering, University of Louisiana at Lafayette, Lafayette, LA 70504. 2Department of Biology, University of Louisiana at Lafayette, Lafayette, LA 70504. 3School of Geosciences, University of Louisiana at Lafayette, Lafayette, LA 70504. *Corresponding author - EricTobin09@gmail.com. Manuscript Editor: Andrew Edelman Southeastern Naturalist E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 464 et al. 1999, 2002) and each has a distinct species composition, structure, function, soil, and hydrology (Sasser et al. 1996). Under current wildlife management practices (e.g., Alligator mississippiensis Daudin [American Alligator] harvest regulation), however, they are all placed in the same, homogenous classification of freshwater marsh habitat. Pioneering work has been done to define the differences in vegetation types in Louisiana’s freshwater marshes, but little research has focused on the smallmammal communities of these habitats. To date, only Abernethy et al. (1985) and Martin et al. (1991) included freshwater marshes in assessments of small-mammal diversity and abundance; however, these studies were not designed to examine the influence of freshwater-marsh vegetation composition on small-mammal communities. Small mammals play important roles in many ecosystems, such as seed dispersal and insect control (Lowery 1974, Ryszkowski 1975). In freshwater-marsh ecosystems, rodents serve as primary, secondary, and tertiary consumers and as prey in both terrestrial and aquatic food webs (Eubanks et al. 2011, Gosselink et al. 1998, Lowery 1974, Negus et al. 1961). In the case of omnivorous rodents, their variable trophic position likely allows for both top down and bottom up influence on other species’ populations. Our objective was to determine if variation in vegetation composition of freshwater marshes in coastal Louisiana resulted in differences in small-mammal site occupancy. Due to the low abundances of small mammals detected in Abernethy et al. (1985) and Martin et al. (1991), and the disturbance caused to floating marsh habitats by a large number of visits to the same site, we examined site occupancy instead of measuring population densities. Occupancy modeling is a fairly new statistical tool that estimates the probability that a species occurs at a site (MacKenzie et al. 2006), and is well suited for application to the noncontiguous Louisiana freshwater marshes that are interspersed with canals, rivers, lakes, and ridges. Occupancy modeling also allows the inclusion of covariates such as vegetation, which is useful for comparing marsh features associated with the occurrence of small mammals. Methods We selected Mandalay National Wildlife Refuge (MNWR) as the study area because of its abundant freshwater marshes and relative ease of access. The refuge is situated roughly 8 km west (29°31'12.44"N, 90°48'03.21"W) of Houma, LA (USFWS2009). The climate is subtropical, with an annual average rainfall of 165 cm. The influx of freshwater from the Atchafalaya River results in classification of MNWR as a freshwater system. Periodic flooding from the Atchafalaya River deposits sediments and increases water level in MNWR swamps and marshes. Additional freshwater comes via run-off from adjacent uplands and local streams. As with the rest of Louisiana’s coastal marshes, climate change and hydrologic alterations are affecting the refuge, leading to land loss and salt-water intrusion, with catastrophic losses predicted for the near future. Cypress swamps, freshwater marshes, upland ridges, Lake Hatch, and oil-field canals are found on the refuge, and the Gulf Intracoastal Waterway crosses the Southeastern Naturalist 465 E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 parcel. Marshes are generally dominated by Sagittaria lancifolia L. (Bulltongue Arrowhead) to the north of the waterway and Maidencane to the south. The refuge is primarily managed for migratory and endangered birds. Faunal species diversity of the refuge has not been described, but is likely similar to the species list for the freshwater marshes and swamps of the Barataria-Terrebonne Estuary—another freshwater wetland system in the region (Condrey et al. 1995). We randomly selected 36 navigable sites for live trapping on MNWR using ARCGIS (version 9.3.1), and we conducted an on-site determination of our ability to traverse the marsh to set and retrieve traps. To allow access by boat, all of the sites we selected for study were near canals. Due to the structural and ecological fragility of floating freshwater marshes, we elected to place traps in a circular layout (Fig. 1) because occupancy modeling relies only on detection and not abundance measurements. This modification of a traditional linear trap-line is similar to the point-count method commonly used when sampling birds for occupancy studies (MacKenzie et al. 2006). We employed standard (233 mm x 75 mm x 90 mm) live traps (model Folding Trap-LFA, Sherman Trap Company, Tallahassee, FL) to non-lethally sample small mammals, and selected quick oats as bait because they afforded the best capture rate with the least attraction of Solenopsis invicta Buren (Red Imported Fire Ants ) in a pilot study. We used Velcro to attach traps to 41-cm2 platforms. Figure 1. Diagram of the general trap arrangement at small-mammal sampling sites in freshwater marshes of the Mandalay National Wildlife Refuge, LA. Southeastern Naturalist E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 466 The platforms consisted of plywood and foam layers that we glued together, similar to the designs used by Desa et al. (2012), Hotaling et al. (2010) and Wolfe (1985), but designed independently. We cleaned the traps weekly to prevent disease and rancid odors and allowed them to air dry for at least 36 h before being used again. At each site, we evenly spaced 10 traps along a 16-m-diameter circle around a center point that was a minimum of 20 m into the marsh from the canal. We marked the plot center with a PVC pipe and recorded the location using a GPS unit. At each site, we installed traps in the afternoon and retrieved them early the following morning to exploit the nocturnal nature of small mammals and prevent the overheating of trapped animals. We deployed a total of 360 traps (10 per site) in each of 4 sampling periods, for a total of 1440 trap-nights during mid-May–mid-June 2011. All sites were trapped for one night within each of the sampling periods, with a minimum of 4 and a maximum of 5 days between samples. The number of sites sampled and sampling occasions was a compromise between the need for statistical power and the resources (personnel, traps, boat, and truck) available. Due to the fragile nature of floating marsh, we limited the number of visits so that we measured small mammal and plant interaction rather than researcher-induced habitat changes. Floating marsh root mats eventually give way to open water if traversed too often. We placed sites a minimum of 200 m apart unless there was a large (>8 m cross section) canal between them that we assumed would be a barrier to animal movement. We considered movement across large canals to be as unlikely as movement of >200 m because of high levels of boat traffic and the presence of American Alligators in the canals. We made our site placements so as to provide the maximum number of sites in the limited study area with little likelihood of movement of small mammals between sites. The number of sites sampled every day was determined by the average time needed to check them in the morning before the animals became too stressed by the heat. We sampled during the early summer, and daily high temperatures averaged 32 °C, daily low temperatures averaged 22 °C, and cumulative precipitation was 0.58 cm (http://www.nws.noaa.gov/climate/). Water levels remained relatively constant; however, we observed acute increases after periods of heavy precipitation associated with thunderstorms. We emptied all traps placed on a given day between sunrise (approximately 06:30) and 10:00 the following day. We strived to sample 10 sites per day, but sampled fewer sites on days with inclement weather. The timing of site visits was structured to allow the sites to recover from researcher disturbance between samples and to satisfy the assumptions of occupancy modeling: the area does not change between sample periods, animals are not harmed by sampling, the population remains closed, there is no movement between sites, sites do not become colonized or extinct, each site has the same occupancy and capture probability, and each site is homogeneous (MacKenzie et al. 2006). We photographed captured animals and determined species, body metrics, gender, and age characteristics. Before release, we applied an ear tag to identify previously captured individuals to help us determine if movement occurred Southeastern Naturalist 467 E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 between sites. Recaptured animals were examined for any substantial change from the previous sampling, such as bodily injury or change in reproductive status. We harvested vegetation 14–18 June at five 0.25-m2 plots, located at every other trap station at each site (a total of 180 stations), placed samples in black plastic bags, and took them back to the laboratory, where they were refrigerated. We sorted the collected aboveground biomass by species, stored samples in brown paper bags, and dried them at 70 °C for a minimum of 48 hours. Species dry weight was determined by weighing the bag with the sample to the nearest 0.1 g; the weight of the bag was subtracted from this measurement to obtain each species’ dry weight. We multiplied all weights by 4 to transform the sample weight to the dry weight per m². We obtained total aboveground biomass for each plot by averaging the weights of the species harvested from that plot (Evers et al. 1998). The program PRESENCE (version 6.1) (available from http://www.mbr-pwrc. usgs.gov/software/presence.html) was used to conduct occupancy modeling of small mammals across sites (MacKenzie et al. 2006). In the occupancy model, two parameters were estimated from covariates and occupancy data. The first parameter estimated is psi (Ψ), the probability of a site being occupied. The second parameter estimated was probability of detection at a site (p). We used a single-species, single-season model that held probability of detection constant across sites and sampling sessions. Assuming a constant detection probability was reasonable because trap placement, time of deployment, and baiting were standardized, while weather conditions varied randomly across sites and the sampling season. The equation for determining Ψ for a single-species, single-season model with covariates is: logit(Ψi) = β0 + β1xi1 + β2xi2 + ... + βuxiu, where β0 represents intercept, βu represents covariates associated with site i, and xi are the values for the covariates at site i. PRESENCE evaluated null, full, and covariate models using the biomass of each of the 5 most common plant taxa as the covariates, both singularly and with additive effects, for a total of 31 models. We did not include climatic covariates because they tended to either be relatively invariant, such as temperature, or too spatially heterogeneous to monitor at each sample site, such as rainfall. Model averaging was used to estimate parameter and unconditional standard errors (Burnham and Anderson 2002). We used using Program R (Version 2.12.1) (R Foundation for Statistical Computing) to perform spatial autocorrelation on biomass of each of the 5 most common taxa to determine the spatial independence of sites. Spatial autocorrelation uses distance and site characteristics to determine if sites in a given system fall out into spatial clusters (Lichstein et al. 2002). We also evaluated whether capture of individual mammals was due to spatial proximity of sites, using a Mantel test (XLSTAT 2010, New York, NY). We compared a matrix of similarity of sites based on capture status to a matrix of geographic distance between sites; statistical significance of the matrix association was calculated by running 10,000 permutations. Results The only small-mammal species we captured was Oryzomys palustris Harlan (Marsh Rice Rat). Marsh Rice Rats were captured at 21 of the 36 sampled sites, for Southeastern Naturalist E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 468 a total of 34 individuals captured during the study. We recaptured only 3 individuals: 2 at the same site of their initial capture and one that had traversed a floating marsh and one small canal to move 1.5 km between sites over a 5-day period. The individual that had moved was a juvenile male, and it is likely that this movement is indicative of the lack of male philopatry rather than overall species behavior. The captured Marsh Rice Rats included 7 females, 25 males, and 2 individuals that escaped before being sexed. We classified 3 juveniles, 6 sub-adults going through molts, and 25 adults. We collected 33 plant species across the refuge. We identified 5 dominant plant taxa that made up the majority (89%) of the live biomass and occurred at the majority of the sites (Fig. 2): Alternanthera philoxeroides Grisebach (Alligatorweed), Bulltongue Arrowhead, spikerush, Leersia oryzoides L. Swartz (Rice Cutgrass), and Typha spp. (cattail). Alligatorweed was the most common plant and was found at every site. The most dominant species were cattail and Bulltongue Arrowhead, at 42% and 28% of total biomass, respectively. Alligatorweed, spikerush and Rice Cutgrass made up 10%, 4%, and 4% of total biomass, respectively. We found no significant correlations (Pearson’s r ≤ 0.1018) between the biomass of the 5 most common plant taxa, so we used the biomass of each of these 5 taxa in subsequent analyses. The results of the spatial autocorrelation analysis showed that all sites were spatially independent (Fig. 3); strong positive autocorrelation only occurred within Figure 2. Dry live biomass of the 5 most common plant species at each site sampled in the Mandalay National Wildlife Refuge. The 36 sites are listed left to right by decreasing number of Marsh Rice Rats captured, then by decreasing Bulltongue Arrowhead biomass, and last by increasing cattail biomass. Bars represent the average from 5 plots. Southeastern Naturalist 469 E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 the first 200 m for the biomass of most plant species among sites. The results of the Mantel test indicated no association between the distance among sites and Marsh Rice Rat captures (r = -0.01, P = 0.787). We ran occupancy models with null, full, singular, and additive effects of the 5 most abundant plant taxa. After removing models in which SE could not be estimated for 1 or more parameters, the highest ranked model, based on the lowest Akaike information criterion corrected for small sample sizes (AICc), retained Bulltongue Arrowhead as the only covariate (Table 1). A goodness-of-fit test of the saturated model (χ² = 27.9, P ≥ 0.02) suggested that the data were overdispersed. However, the variance-inflation factor was relatively small (ĉ = 1.93) compared to an expectation of 1. Burnham and Anderson (2002) suggested that there is no need to adjust AICc values for overdispersed data if ĉ is between 1 and 4, so we present unadjusted estimates of AICc. Results of model averaging returned a βS. lancifolia of 15.0 (SE calculated from unconditional variances = 8.0), which indicated a positive association between Marsh Rice Rat occupancy and Bulltongue Arrowhead biomass. The lowest level of Bulltongue Arrowhead biomass observed in this study offered a 100% probability of occupancy. There was little support for other covariates containing information on Marsh Rice Rat occupancy because model-averaged parameter estimates for all other vegetation covariates were near zero with large Figure 3. Spatial autocorrelation of the 5 most common plant species (Alligatorweed, Rice Cutgrass, spikerush, Bulltongue Arrowhead, and cattail) on sampled sites in Mandalay National Wildlife Refuge, LA (n = 180). Spatial lag (in meters) is plotted against covariance for each site. Southeastern Naturalist E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 470 Table 1. Results of occupancy modeling for the presence of Marsh Rice Rats in freshwater marshes of coastal Louisiana. Only the 5 most common vegetation species were used in the analysis. Included on the table are the 15 most informative models ordered by AICc. AIC is corrected for small sample sizes (AICc), Model likelihood is AIC weight of the given model relative to the top model, and -2LL is 2 times the negative log-likelihood. Model covariates AICc ΔAICc AIC weight Model likelihood # of parameters -2LL Sagittaria lancifolia 149.69 0.00 0.2837 1.0000 2 145.60 S. lancifolia + Leersia oryzoides 151.04 1.35 0.1511 0.5326 3 144.86 S. lancifolia + Eleocharis spp. 151.62 1.93 0.1131 0.3985 3 145.44 S. lancifolia + Alternanthera philoxeroides 152.83 3.14 0.0617 0.2176 3 146.65 S. lancifolia + L. oryzoides + Eleocharis spp. 153.09 3.40 0.0578 0.2039 4 144.78 Eleocharis spp. 153.31 3.62 0.0464 0.1637 2 149.22 Eleocharis spp. + A. philoxeroides 154.09 4.40 0.0329 0.1159 3 147.91 S. lancifolia + L. oryzoides + Eleocharis spp. + Typha spp. 154.57 4.88 0.0298 0.1049 5 144.11 S. lancifolia + Eleocharis spp. + A. philoxeroides 154.58 4.89 0.0275 0.0968 4 146.27 S. lancifolia + L. oryzoides + Eleocharis spp. + A. philoxeroides 154.97 5.28 0.0244 0.0859 5 144.51 A. philoxeroides 154.63 4.94 0.0240 0.0846 2 150.54 Total biomass 154.84 5.15 0.0216 0.0762 2 150.75 L. oryzoides + Eleocharis spp. 155.07 5.38 0.0201 0.0710 3 148.89 Eleocharis spp. + Typha spp. 155.39 5.70 0.0172 0.0605 3 149.21 Null model 155.48 5.79 0.0157 0.0553 1 151.39 Southeastern Naturalist 471 E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 standard errors. A Mantel test indicated no significant association between the biomass of Bulltongue Arrowhead and spatial proximity of sites (r = 0.03, P = 0.456). The estimated overall Ψ for all sites using the null model was 0.8032 (SE = 0.1557) with probability of detection at 0.2767 (SE = 0.0639). Therefore, we estimate that about 80% of sampled sites on MNWR were occupied by Marsh Rice Rats, although the species was captured at only 21 (58%) of the sampled sites. Discussion Our results suggest that variation in plant composition in freshwater marshes affects Marsh Rice Rat occupancy over a small spatial scale. Marsh Rice Rat ecology and population dynamics in salt marshes and terrestrial edge habitats (barrier islands, levees, etc.) have been extensively described (Eubanks et al. 2011, Kruchek 2004, Negus et al. 1961, Smith 1980, Smith and Vrieze 1979, Wolfe 1985), but little is known about this marsh-obligate species in coastal freshwater-marsh habitat. Different wetland plant communities reflect differences in the environment, such as soil structure and composition, flooding regimes, and water chemistry (Sasser et al. 1996). These changes in the environment can directly or indirectly affect Marsh Rice Rat occupancy through characteristics of the plants. The relationship between Bulltongue Arrowhead and Marsh Rice Rat occupancy found in our study may be related to several characteristics of plant-community structure and Marsh Rice Rat ecology. Marsh Rice Rats are omnivorous (Gosselink et al. 1998), and their diet includes crabs, insects, green vegetation, seeds and fruits of marsh plants, snails, fishes, and carrion (Davis and Schmidly 1997, Lowery 1974, Wolfe 1985). Populations fluctuate seasonally and are largely driven by resource availability (Bloch and Rose 2005, Lowery 1974). Marsh Rice Rats’ resource-based home ranges are extremely variable in size, ephemeral, and overlap with conspecifics (Negus et al. 1961). Bulltongue Arrowhead is perennial and grows in thick clumps, with broad, rigid leaves that allow Marsh Rice Rats to move under cover but above the water of the marsh, helping them to avoid aerial predators, including herons and raptors. Another feature of Bulltongue Arrowhead is that it forms floating mats (Sasser et al. 1996) that rise and fall with the water levels, providing reliable living space and sites for Marsh Rice Rats to rest and rear young. Marsh Rice Rats utilize all parts of plants as food (Davis and Schmidly 1997, Lowery 1974, Wolfe 1985) and, because Bulltongue Arrowhead has relatively high biomass, this species could potentially be an important food source. Marsh Rice Rats have shown an affinity for broadleaved plant cover in other marsh environments and the species would likely be extirpated if marshes shift to woody vegetation, despite the cover shrubs provide, as the species has not been found to inhabit sites dominated by shrubs (Eubanks et al. 2011, Hofmann et al. 1990, Negus et al. 1961). Bulltongue Arrowhead has been shown to have a higher salt tolerance than related species and is a good candidate for use in restoration of freshwater marsh that has suffered saltwater intrusion and conversion to open water (Martin and Shaffer 2005). A full understanding of how Marsh Rice Rats and Bulltongue Arrowhead interact may be important to the success of recovery efforts and the longevity of Southeastern Naturalist E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 472 Louisiana’s coastal freshwater marshes. More research is needed on behavior, diet, plant use, and life history of small mammals in these systems, as well as the influence of plant–animal interactions. Acknowledgments We would like to thank the staff of the Mandalay National Wildlife Refuge and the University of Louisiana at Lafayette Institutional Animal Care and Use Committee (IACUC No. 2010-8717-007) for their help in making this project possible. Dr. Scott Duke-Sylvester’s help with the data analysis for this project is greatly appreciated. We extend our gratitude to Andrew Edelman and the two anonymous reviewers for their patience and comments, which have greatly improved this manuscript. We would also like to thank Edward Delaney IV for his assistance in editing figures, and Jennifer Rasch and Clara Weidman for their technical support. Support for E. Tobin was provided by the Louisiana Environmental Health Association. Literature Cited Abernethy, R.K., G.W. Peterson, and J.G. Gosselink. 1985. A small-mammal trapping study of the floating freshwater marshes surrounding Lake Boeuf, Louisiana. Northeast Gulf Science 7:177–180. Barras, J.A., J.C. Bernier, and R.A. Morton. 2008. Land-area change in coastal Louisiana: A multidecadal perspective (from 1956 to 2006). Scientific investigations map 3019. US Geological Survey Reston, VA. Bloch, C.P., and R.K. Rose. 2005. Population dynamics of Oryzomys palustris and Microtus pennsylvanicus in Virginia tidal marshes. Northeastern Naturalist 12:295–306. Burnham, K.P., and D.R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach, 2nd Edition. Springer-Verlag, New York, NY. 488 pp. Condrey, R., P. Kemp, J. Visser, J. Gosselink, D. Lindstedt, E. Melancon, Jr., G. Peterson, and B. Thompson. 1995. Characterization of the current status, trends, and probable causes of change in living resources in the Barataria-Terrebonne estuarine system. BTNEP Publ. No. 21, Barataria-Terrebonne National Estuarine Program, Thibodaux, LA. 434 pp. Davis, W.B., and D.J. Schmidly. 1997. The mammals of Texas. Texas Tech University. Available online at http://www.nsrl.ttu.edu/tmot1/Default.htm. Accessed 12 September 2010. Desa, M.A., C.L. Zweig, H.F. Percival, W.M. Kitchens, and J.W. Kabohm. 2012. Comparison of small-mammal sampling techniques in tidal salt marshes of the central Gulf Coast of Florida. Southeastern Naturalist 11:89–100. Eubanks, B.W., E.C. Hellgren, J.R. Nawrot, and R.D. Bluett. 2011. Habitat associations of the Marsh Rice Rat (Oryzomys palustris) in freshwater wetlands of southern Illinois. Journal of Mammalogy 92:552–560. Evers, D.E., C.E. Sasser, J.G. Gosselink, D.A. Fuller, and J.M. Visser. 1998. The impact of vertebrate herbivores on wetland vegetation in Atchafalaya Bay, Louisiana. Estuaries 21:1–13. Gosselink, J.G., J.M. Coleman, and R.E Steward, Jr. 1998. Coastal Louisiana. Pp. 385–436, In M.J. Mac, P.A. Opler, C.E. Puckett Haacker, and P.D. Doran (Eds.). Status and Trends of the Nation’s Biological Resources, Vol. 1. US Department of the Interior, US Geological Survey, Fort Collins, CO. 434 pp. Southeastern Naturalist 473 E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 Hofmann, J.E., J.E. Gardner, and M.J. Morris. 1990. Distribution, abundance, and habitat of the Marsh Rice Rat (Oryzomys palustris) in southern Illinois. Transactions Illinois State Academy of Science 83:162–180. Hotaling, A., H.F. Percival, W.M. Kitchesn, and J.W. Kasbohm. 2010. The persistence of endangered Florida Salt Marsh Voles in salt marshes of the central Florida Gulf Coast. Southeastern Naturalist 9:795–802. King, D.T., and T.C. Michot. 2002. Distribution, abundance, and habitat use of American White Pelicans in the delta region of Mississippi and along the western Gulf of Mexico coast. Waterbirds 25:410–416. Kruchek, B.L. 2004. Use of tidal marsh and upland habitats by the Marsh Rice Rat (Oryzomys palustris). Journal of Mammalogy 85: 569–575. Lichstein, J.W., T.R. Simons, S.A. Shriner, and K.E. Franzreb. 2002. Spatial autocorrelation and autoregressive models in ecology. Ecological Monographs 72:445–463. Lowery, G.H., Jr. 1974. The Mammals of Louisiana and its Adjacent Waters. Louisiana State University Press, Baton Rouge, LA. 565 pp. MacKenzie, D.I., J.D. Nichols, J.A. Royle, K.H. Pollock, L.L. Bailey, and J.E. Hines. 2006. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Elsevier Publishing, Oxford, UK. 323 pp. Martin, R.P., R.B. Hamilton, P.M. McKenzie, R.H. Chabreck, and D.H. Dell. 1991. Habitat use by small mammals in coastal marshes of southwestern Louisiana. Estuaries 14:107–110. Martin, S.B, and G.P. Shaffer. 2005. Sagittaria biomass partitioning relative to salinity, hydrologic regime, and substrate type: Implications for plant distribution patterns in coastal Louisiana, United States. Journal of Coastal Research 21:167–174. Miller, M.R., E.G. Burns, B.E. Wickland, and J.M. Eadie. 2009. Diet and body mass of wintering ducks in adjacent brackish and freshwater habitats. Waterbirds 32:374–387. Mitsch, W.J., J.G. Gosselink, C.J. Anderson, and L. Zhang. 2009. Wetland Ecosystems. John Wiley and Sons, Hoboken, NJ. 304 pp. Negus, N.C., E. Gould, and R.K. Chipman. 1961. Ecology of the rice rat Oryzomys palustris (Harlan) on Breton Island, Gulf of Mexico, with a critique of the social stress theory. Tulane Studies in Zoology 8:95–123. Palmisano, A.W. 1972. Habitat preference of waterfowl and fur animals in the northern Gulf Coast marshes. Pp. 163–190, In R.H. Chabreck (Ed.). 2nd Coastal Marsh and Estuary Management Symposium. Louisiana State University, Baton Rogue, LA. 316 pp. Penland, S., L. Wayne, L.D. Britsch, S.J. Williams. A.D. Beal, and V. Carridas-Butterworth. 2001. Process classification of coastal land loss between 1932 and 1990 in the Mississippi River deltaic plain, southeastern Louisiana. US Geological Survey Open File Report 00-418. Washington, DC. Plattner, D.M., M.W. Eichholz, and T. Yerkes. 2010. Food resources for wintering and spring-staging Black Ducks. Journal of Wildlife Management 74:1554–1558. Ryszkowski, L. 1975. The ecosystem role of small mammals. Ecological Bulletins, Biocontrol of Rodents 19:139–145 SAS Institute Inc. 2009. SAS Enterprise Guide 4.1 for SAS OnDemand. Cary, NC. Sasser, C.E., J.G. Gosselink, E.M. Swenson, C.M. Swarzenski, and N.C. Leibowitz. 1996. Vegetation, substrate and hydrology in floating marshes in the Mississippi River delta plain wetlands, USA. Vegetatio 122:129–142. Smith, A.T. 1980. Lack of interspecific interactions of Everglades rodents on two spatial scales. Acta Theriologica 25:61–70. Southeastern Naturalist E.J. Tobin, J.M. Visser, J.K. Peterson, and P.L. Leberg 2014 Vol. 13, No. 3 474 Smith, A.T., and J.M. Vrieze. 1979. Population structure of Everglades rodents: Responses to a patchy environment. Journal of Mammalogy 60:778–794. US Fish and Wildlife Service (USFWS). 2009. Draft comprehensive conservation plan and environmental assessment: Mandalay National Wildlife Refuge: Terrebone Parish, Louisiana. Washington, DC. 151 pp. Visser, J.M., C.E. Sasser, R.H. Chabreck, and R.G. Linscombe. 1999. Long-term vegetation change in Louisiana tidal marshes. Wetlands 19:168–175. Visser, J.M., C.E. Sasser, R.H. Chabreck, and R.G. Linscombe. 2002. The impact of a severe drought on the vegetation of a subtropical estuary. Estuaries 25:1185–1196. Visser, J. M., J.W. Day, Jr., L.L. Battaglia, G.P. Shaffer, and M.W. Hester. 2012. Mississippi River Delta Wetlands. Pp. 6–74, In D. Batzer and A. Baldwin (Eds.). Wetland Habitats of North America: Ecology and Conservation Concerns. University of California Press, Berkeley, CA. 408 pp. Ysebaert, T., P.L. Meininger, P. Meire, K. Devos, C.M. Berrevoets, R.C.W Strucker, and E. Kuijken. 2000. Waterbird communities along the estuarine salinity gradient of the Schelde estuary, NW-Europe. Biodiversity and Conservation 9:1275–1296. Wolfe, J.L. 1985. Population ecology of the rice rat (Oryzomys palustris) in a coastal marsh. The Journal of Zoology, London 205:235–244