Southeastern Naturalist 541 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 22001188 SOUTHEASTERN NATURALIST 1V7o(4l.) :1574,1 N–5o5. 34 Defining Success Criteria for Spartina alterniflora Restoration Projects in Southwestern Louisiana Joshua M. Soileau1, Eddie K. Lyons1,*, Byungkyun Chung1, Justin Hoffman2, and Frederick LeMieux1 Abstract - Evaluation of success in salt-marsh restoration is often driven by subjective measures. By defining success criteria (SC), project effectiveness can be measured quantitatively. Herein, we quantify SC for an existing Spartina alterniflora (Smooth Cordgrass) restoration program. We selected stem density as an indicator of primary production and canopy structure—2 important salt-marsh functions. We established SC through use of stem densities from the literature and prior restoration projects and calculated a coefficient of variation (CV) for reference groups. By using the CV, we determined the strength of a particular reference group for SC formulation. We calculated success criteria for reference groups with CV ≤ 0.25. Three reference groups met the CV requirement. Only 1 project met the SC of the strongest reference group (CV = 0.04). Four of 5 projects met the SC of the weakest reference group (CV = 0.21). Reference sites with low variation in the target indicator should be selected for SC formulation. Introduction Success is one of the terms used most often in wetland-restoration science; it is often used in an imprecise and inconsistent manner because it depends on the situation and the people defining it (Kentula 2000, Whigham 1999, Zedler 1996, Zedler and Callaway 2000). In most cases, the success being referenced is the meeting of established objectives for a permit or contract. Although it is important that mitigation projects fulfill their contract goals, some would argue that this does not necessarily qualify as success from a conservation standpoint. Zedler (1996) termed such a situation—when a project satisfies the permit requirements but is not successful in restoring wetland function—as compliance. In fact, when it comes to compliance, less than 70% of those projects evaluated met all permit requirements (Cole and Shaffer 2002, Minkin and Ladd 2003, Sudol and Ambrose 2002). Another study stated that on average, 73% of permit conditions are met in mitigation projects (Ambrose et al. 2006). Many authors have advocated that success in wetland restoration can only be accomplished if ecological functions such as nutrient cycling, wildlife habitat, and sediment retention are restored (Brown 1996, Cairns 2000, Kentula 2000, Neckles et al. 2002, Quammen 1986, SER 2004). Kentula (2000) termed this functional success. Defining success in such a way means that the system must be able to sustain itself and be biologically viable (West et al. 2000), but it could take many years for 1School of Agricultural Sciences, McNeese State University, Box 92220, Lake Charles, LA 70609. 2Department of Biology/Health Science, McNeese State University, PO Box 92000, Lake Charles LA, 70609. *Corresponding author- elyons@mcneese.edu. Manuscript Editor: Foster Levy Southeastern Naturalist J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 542 that to happen. Zedler (1996) suggested that project managers should seek to report on the progress of a project instead of using the traditional labels—successful and unsuccessful labels—which should only be used when a project has either met or failed to meet all of its objectives. If “success” as opposed to “compliance” is to be evaluated, goals must be established to evaluate whether natural ecological functions have been restored. Defining attainable goals for a project that adequately evaluate ecological functions may not be as simple as meeting the requirements of a permit, or defining a target survival-percentage. Even more daunting is the task of choosing a methodology that adequately monitors characteristics of a project to determine if the goals are met. The major hurdle to determining success for most restoration activities is financial because most projects do not have the budget to conduct monitoring on the ecological function of a site (Ruiz-Jaen and Aide 2005). The Gulf Coast Soil and Water Conservation District (SWCD; Lake Charles, LA) participates in a vegetation program for which the typical objectives of a project are to restore wildlife habitat, provide a seed base, and reduce erosion. However, these functions are never assessed for success or failure because it is outside of the program’s budget to do so. Rather, plant growth and health data (i.e., stem count, vigor, spread, etc.) are gathered at predetermined time intervals after planting, and reports are submitted annually. Success is then determined based on survival rate of plants, or the quality of the project (i.e., excellent, fair, poor), a subjective measure that may change from year-to-year or with the personnel monitoring the projects. Our study sought to utilize the data gathered on these projects to quantitatively determine if functional success has been attained. To more accurately determine the success of a project, biologists must assess indicators of ecological functions. Two functions often associated with salt-marsh ecosystems are canopy structure and primary production (Gleason et al. 1979, Kaswadji et al. 1990, Short et al. 2000). These 2 functions relate to the habitat, refuge, and food available to wildlife and fisheries (Short et al. 2000). Short et al. (2000) identified stem density as an indicator for both of these functions. Stem density is a good candidate indicator for determining success because of its ease of use and low cost (Craft et al. 2003), and restoration programs, such as the one administered by the Gulf Coast SWCD, often utilize it. However, stem density is rarely utilized when determining project success. The objectives of our study were to: (1) define success criteria for Spartina alterniflora L. (Smooth Cordgrass) restoration projects by using data already gathered within the current monitoring program, and (2) apply this criterion to recently completed (planting, and initial and 1-y monitoring completed) projects to determine if functional success has been achieved. Study Area We conducted our study during salt-marsh revegetation projects in southwestern Louisiana. These marshes are dominated by Smooth Cordgrass. Salinities ranged from 7 ppt to >20 ppt (J.M. Soileau et al., unpubl. data). Rainfall in the area averages 134.62 cm/y (NOAA 2012). The major soil unit in the area is Clovelly muck. Southeastern Naturalist 543 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 Salt marshes in this part of Louisiana occur on both organic and mineral soils (Midkiff 1995) and were severely impacted by hurricanes in 2005 and 2008, with large areas lost due to scouring (Barras 2006, 2009). The revegetation projects consisted of either bare-root plugs or commercially available trade gallons (i.e., plants in 1-gallon pots, each pot typically with 5–12 aerial stems that are 0.46–0.61 m tall) of S. alterniflora ‘Vermilion’, a native Louisiana strain of Smooth Cordgrass. Project locations in 2011 were mud flats along the banks of the Mermentau River in southeastern Cameron Parish, LA; sandy banks along the Calcasieu Ship Channel in southern Calcasieu Parish, LA; and levees along the banks of Kelso Bayou in Cameron Parish (Fig. 1). Objectives for the projects were to (1) provide a seed base, (2) create wildlife habitat, and (3) reduce erosion. Methods In this study, we adapted the methods employed by Short et al. (2000) for success-criteria (SC) calculation to accommodate the limited capabilities of the monitoring program. Modifications are discussed here. Restoration sites Projects were planted in 2011 throughout Calcasieu and Cameron parishes, LA, while J.M. Soileau was employed by Gulf Coast SWCD (Fig. 2). We planted Figure 1. Salt-marsh restoration sites in southwestern Louisiana in 2011. Southeastern Naturalist J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 544 Smooth Cordgrass plugs at the Kelso Bayou site; the 4 remaining sites were planted with trade gallons. We conducted monitoring 1 y after planting. We gathered quantitative and qualitative information on ~10% of the total project, divided into segments. For each segment, we recorded vigor, spread, and average stems per square meter (Table 1). Reference sites In order to calculate SC for reference sites and success ratios (SR) for restoration sites, reference sites must be established and candidate indicators selected. The purpose of this research is to define success criteria for projects with a limited budget; thus, natural reference sites were not defined for each project, unlike in Short et al. (2000). Rather, we gathered data on stem density from natural and restoration Smooth Cordgrass marshes from the literature (Table 2; Craft et al. 2003, Currin et Table 2. Stem densities from natural and restored Smooth Cordgra ss marshes from literature. Author/Project State Type Stem density stems/(m2) Notes Minello and Zimmerman 2003 TX Natural 124 Matagorda Bay TX Transplant 175 Matagorda Bay TX Natural 129 Chocolate Bay TX Transplant 184 Chocolate Bay TX Natural 103 Stadman Island TX Transplant 193 Stadman Island Craft et al. 2003 NC Natural ~250 1-y-old marsh NC Transplant ~500 1-y-old marsh NC Natural ~200 3-y-old marsh NC Transplant ~250 3-y-old marsh NC Natural ~300 28-y-old marsh NC Transplant ~300 28-y-old marsh Currin et al. 2008 NC Natural 149 NC Restored 70 NC Natural 130 NC Restored 150 NC Natural 222 NC Restored 161 Mullens 2007 LA Restored 42 LA Restored 75 LA Natural 72 Table 1. Stem densities at 1-y monitoring for Smooth Cordgrass projects installed in 2011 by the Gulf Coast SWCD, Calcasieu and Cameron parishes, LA. Stem density Project (stems/m2) Notes East Moss Lake 3 Heavily grazed by cattle Kelso Bayou 360 Mermentau River North 93 Some of the project area had been grazed by cattle West Turner's Bay 108 Mud Lake 89 Some of the project area had been grazed by cattle Southeastern Naturalist 545 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 Table 3. Past projects installed by the Gulf Coast Soil and Water Conservation District, Calcasieu and Cameron parishes, LA. Stem density was measured 1 y after plant ing. Stem density Project State Type (stems/m2) Notes Mermentau River (2000) LA Restored 217 Choupique Bayou (2001) LA Restored 146 DU Terraces 1 (2002) LA Restored 449 Hackberry, trade gallon containers DU Terraces 2 (2002) LA Restored 229 Bare-root plugs Moss Lake (2003) LA Restored 54 Moss Lake (2003) Rockefeller Refuge (2006) LA Restored 158 Rockefeller Refuge (2007) LA Restored 112 Headquarters Canal al. 2008, Minello and Zimmerman 2003, Mullens 2007). In addition to the densities gathered from the literature, we also selected as candidate reference-sites 7 areas that had previously been restored and monitored by the Gulf Coast SWCD prior to J.M. Soileau’s employment by that organization (Table 3; Gulf Coast SWCD, unpubl. data). We selected these areas based on proximity to the sites being analyzed, as well as for their similar environmental characteristics (i.e., soils, salinity). Stem densities from these projects were taken 1 y after planting. We divided candidate reference-sites into reference groups based on geographic location and type (restoration or natural; Table 4). We created the groups to allow computation of mean and standard deviation (SD) because no data were available for the individual sites. Although natural reference sites near the restoration projects are ideal, 1 study indicated that Smooth Cordgrass is adapted on a regional scale and that performance among several populations over a given distance may not differ (Travis and Grace 2010). Selecting literature-reported stem densities from numerous Smooth Cordgrass populations along the Atlantic and Gulf Coasts provides a larger set of reference sites from which to calculate SC. Selecting candidate indicators Another difference between this study and Short et al. (2000) is the lack of options for candidate indicators. The monitoring program already in place was limited to measurements of plant growth such as survival, spread, and stem density. However, these parameters are sufficient because stem density is an indicator of functions such as primary production, canopy structure, organic-matter accumulation, and sediment trapping (Gleason et al. 1979, Kaswadji et al. 1990, Short et al. 2000). According to Short et al. (2000), these functions, represent 2 of the 3 major goals (wildlife habitat, providing a seed base, and reducing erosion) of the restoration projects: canopy structure is representative of wildlife habitat; and organic-matter accumulation and sediment trapping are representative of erosion reduction. We used stem densities to calculate SC for reference sites and SR for restoration sites, and compared them to the reference site SC individually to determine if a given project was successful based on a given SC. Southeastern Naturalist J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 546 Descriptive statistics We calculated mean, SD, and coefficient of variation (CV) for 8 groups (Table 5). The groupings were based on geographic location (i.e., Texas, North Carolina, Louisiana, Gulf Coast) and type of reference site (i.e., whether the site was a restored or natural marsh). We used the CV to rank the reference sites for SC calculation. A lower CV indicates that there is less dispersion in the variable (i.e., stem density at reference sites), signifying that a particular group is suitable for creating an SC. Low variability across reference sites means that success of the project can be more accurately determined; high variability in the reference sites makes it difficult to determine if a given site was successfully restored because of the differences in the reference sites. A CV of less than 0.2 is ideal because the Table 4. Reference groups established from literature and past-r estoration sites. Stem density Reference group Author/project (stems/m2) Notes Texas natural Minello and Zimmerman 2003 124 Matagorda Bay 129 Chocolate Bay 103 Stadman Island Texas restored Minello and Zimmerman 2003 175 Matagorda Bay 184 Chocolate Bay 193 Stadman Island Gulf Coast natural Minello and Zimmerman 2003 124 Matagorda Bay 129 Chocolate Bay 103 Stadman Island Mullens 2007 72 Gulf Coast restored Minello and Zimmerman 2003 175 Matagorda Bay 184 Chocolate Bay 193 Stadman Island Mullens 2007 42 75 Mermentau River (2000) 217 Choupique Bayou (2001) 146 DU Terraces 1 (2002) 449 Hackberry, trade 1-gal containers DU Terraces 2 (2002) 229 Bare-root plugs Moss Lake (2003) 54 Moss Lake (2003) Rockefeller Refuge (2006) 158 Rockefeller Refuge (2007) 112 Headquarters Canal North Carolina natural Craft et al. 2003 ~250 1-y-old marsh ~200 1-y-old marsh ~300 3-y-old marsh Currin et al. 2008 149 130 222 North Carolina restored Craft et al. 2003 ~500 1-y-old marsh ~250 3-y-old marsh ~300 28-y-old marsh Currin et al. 2008 70 150 161 Southeastern Naturalist 547 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 variation within the reference sites is low, something that is difficult to surpass regularly in the biological sciences (Short et al. 2000). For this study, we selected reference sites with a CV ≤ 0.25 for SC calculation. We chose this value because it indicates that the variation is equal to a quarter or less of the mean, indicating lower variation within the sampled areas. Short et al. (2000) noted that a CV of 0.2 is a “somewhat” arbitrary value, but used it because it is difficult to surpass with regularity in biological field-sampling. A smaller SD, which is used in the calculation of the CV, indicates a smaller degree of variation within a dataset. Calculating success criteria The formula for calculating SC is (Short et al. 2000): SC = 100 x (mean of reference site - 1 SD) / (mean of reference site) The SD is used to calculate SC for 3 reasons: (1) it holds the CV within the distribution of the data but keeps it above the lowest 16.7% (i.e., 1 SD) of the distribution, (2) it is the statistical standard of variability and is independent of sample size, and (3) the relationship between the SC and CV (a small SD leads to a low CV and a high SC; Short et al. 2000). Assessment time-frame In the current monitoring program, all required monitoring was complete after 1 y; the first monitoring occurred 30–60 d after planting and the second event was 1 y after planting. A third monitoring may be conducted no earlier than 5 y after planting. These projects have yet to be monitored at 5 y; thus, we made our assessments 1 y after monitoring was completed. In biological terms, this would be considered inadequate; many studies state that restored marshes do not exhibit the characteristics of natural marsh for 3–5 y (Confer and Niering 1992, Currin et al. 2008, Edwards and Proffitt 2003, Short et al. 2000). Table 5. Mean, standard deviation, and coefficient of variation (CV) for 3 different mean calculations for stem density. Success Criteria not calculated for reference groups with CV ≥ 0.25. Reference Mean SD CV SC TX restored marshes 184.00 7.35 0.04 96.01 TX natural marshes 118.67 11.26 0.09 90.51 Gulf Coast natural marshes 107.00 22.44 0.21 79.03 SWLA restored marshes (excluding DU Terraces 1 and 172.40 44.15 0.26 - Moss Lake) NC natural marshes 208.50 57.81 0.28 - Gulf Coast restored marshes (excluding DU Terraces 1 and 153.10 57.51 0.38 - Moss Lake) LA restored marshes (excluding Du Terraces 1 and Moss Lake) 139.86 64.17 0.46 - NC restored marshes 238.50 138.16 0.58 - SWLA restored marshes (all) 195.00 117.53 0.60 - Gulf Coast restored marshes (all) 169.50 102.97 0.61 - LA restoration marshes (all) 164.67 118.43 0.72 - Southeastern Naturalist J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 548 Calculating success ratio The success ratio (SR) is the measure of how successful each restoration site is when compared to reference sites (Short et al. 2000). The SR is calculated as follows (Short et al. 2000): SR = 100 x (mean of restoration site) / (mean of selected reference site) We compared the SR for each restoration site to the SC for the reference sites. A given restoration site was successful if it met or exceeded the SC of a given reference site (Short et al. 2000). If the SR for a restoration site exceeds the SC, this can result in unequal development of ecological functions in the site, such as plant response and predation (Short et al. 2000). Results A total of 3 reference groups produced a CV ≤ 0.25 (Table 5). In descending order by CV, these were: Gulf Coast natural marshes (GCNat), Texas natural marshes (TXNat), and Texas restored marshes (TXRes). Only 2 groups met the criterion of CV ≤ 0.2 established by Short et al. (2000): the Texas natural marshes (CV = 0.09) and the Texas restored marshes (CV = 0.04). We report mean, SD and CV for all reference groups (Table 4). The GCNat group had the lowest mean (n = 107), while the TXRes group had the lowest SD (SD = 7.35). Applying success criteria The TXRes group had the lowest CV (CV = 0.04) of all reference groups (Table 5); thus, it was the highest-ranked mean, and the best for SC formulation (Short et al. 2000). Figure 2 shows SC and SR values for reference groups and restoration sites. When compared to the TXRes group (SC = 96.01%), only 1 project, Kelso Bayou (SR = 195.65%), met or exceeded the SC. The TXNat group (SC = 90.51%) was the second best for SC calculation (CV = 0.09), and 2 restoration projects met or exceeded the SC—Kelso Bayou (SR = 303.36%) and West Turner’s Bay (SR = 91.01%). The GCNat group (SC = 79.03%) ranked third for SC calculation (CV = 0.21), and 4 of the 5 restoration projects met the SC—Kelso Bayou (SR = 336.45%), West Turner’s Bay (SR = 100.93%), Mermentau River North (SR = 86.92%), and Mud Lake (SR = 83.18%). All Louisiana-based reference groups had a CV > 0.25, and we did not calculate SC for any group with a CV > 0.25. Discussion Although the Gulf coast marshes used in this study differed from one another in a number of aspects (Table 2), Smooth Cordgrass along the Gulf coast typically exhibits the same growth form (Seliskar et al. 2002). Based on the low variation, the TXRes and TXNat groups were reliable reference sites for the study (Table 2). Only the Kelso Bayou restoration project was successful when compared to the TXRes group, which had the strongest CV (CV = 0.04). Using the TXNat group (CV = 0.09), Kelso Bayou and West Turner’s Bay were successful. Since stem Southeastern Naturalist 549 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 density can be used to determine the restoration of certain functions, such as wildlife habitat and sediment retention, it can be assumed through these results that these functions have been restored at the Kelso Bayou and West Turner’s Bay sites. Personal observations of wildlife use at both sites and the narrowing of levee breaches at Kelso Bayou confirm this conclusion (Fig. 3). Fruiting culms Figure 2. Success criteria (SC) and success ratios (SR) for stem density from Smooth Cordgrass restoration sites in 2001 in southwestern Louisiana. SC for each reference group represented by dark gray horizontal line. Figure 3. Smooth Cordgrass at Kelso Bayou. Picture taken at 1-y monitoring in 2012. Southeastern Naturalist J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 550 were present at West Turner’s Bay,[indicating that the goal of providing a seed base had been met (Fig. 4). Of the 3 projects that failed to meet the SC for both reference groups, Mud Lake and Mermentau River North each had monitoring segments with stem densities above the SC (Fig. 5). Portions of these projects were planted on harder clays and soils with high shell-content and transplants did not perform as well here, which Figure 4. Smooth Cordgrass fruiting culms at West Turner’s Bay in 2012. Figure 5. Smooth Cordgrass at Mud Lake project in 2012. A dense stand has established on previous mud-flat area. Southeastern Naturalist 551 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 resulted in lower stem densities. In the areas that did have high stem densities, project objectives were being met. Although these projects fell short of meeting the SC of the 2 Texas groups, both met the SC established by the GCNat group (CV = 0.21). East Moss Lake was the only project that failed to meet any of the SC established in this study. In this project, it appeared that grazing by Bos taurus L. (Cattle) or Myocastor coypus (Molina) (Nutria) had reduced survival to less than 10%. The use of past restoration projects as reference sites can be a viable choice if the sites have achieved the desired ecological functionality. In the current study, the DU Terraces 1 and Moss Lake restoration projects were obvious outliers (stem densities = 449 stems/m2 and 54 stems/m2, respectively). By removing these 2 projects from the SWLA restored marshes group (CV = 0.60), the CV was reduced by 56% (CV = 0.26). Although this CV was still outside the designated selection criterion of CV ≤ 0.25, this group would serve as a good reference for SC calculation under less-selective criteria. However, the projects used in this reference group were selected due to proximity to the restoration sites established in 2011. Using past projects that have been monitored at 5 y, when restoration trajectories have typically leveled off and are closer to conditions seen in natural marshes, would be a better alternative. Projects tend to exceed natural marsh production within the first few years of establishment (Short et al. 2000). The Kelso Bayou restoration project is an example of this phenomenon: the SR for the site exceeded all 3 SC with which it was compared. Future use of past restoration sites should be subjected to a more strenuous evaluation process. The effectiveness of this methodology of setting SC is largely dependent on the reference sites used. Examination of the reference groups indicated that using multiple reference sites from different geographic regions, even along the Gulf coast, or using restoration sites which have a wide range of densities, can cause a site to be incorrectly deemed successful. The GCNat group led to the conclusion that 4 of the 5 restoration projects were successful, and was also the only group for which SC was calculated that had a CV > 0.2, the criterion in Short et al. (2000). We chose to depart from the 0.2 CV applied in Short et al. (2000) to illustrate how different methods of SC calculation can lead to different outcomes of success determinations. More-rigorous criteria for SC calculation would reduce the likelihood of inaccurately classifying the success of a project, as was seen with the TXRes and TXNat reference groups. However, reference groups or sites with SC < 0.1 may create criteria that are too strict, even when data and observations show healthy, functioning marsh, as was evident in the West Turner’s Bay project. This finding further exemplifies the need for proper reference site selection . We echo the recommendations of others (Kentula 2000, Short et al. 2000) that natural sites near the proposed restoration project are the best ones for referencing project outcomes. The use of information from other geographic localities can lead to improper designations of success or failure. Assessment time-frame is also an important consideration. These decisions of success or failure are based on 1 y of growth. A decision on success or failure of a project may best be made 5 y Southeastern Naturalist J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 552 after restoration; this is the point when restored marshes begin to resemble natural marshes (Confer and Niering 1992, Currin et al. 2008, Edwards and Proffitt 2003, Short et al. 2000). However, some candidate indicators, such as stem density, may take as much as 15 y to resemble those present in natural marshes (Craft et al. 2003). For this reason, managers must continually monitor restoration projects to accurately determine the trajectory of restoration because some projects may take longer to reach equivalency with their natural counterparts. Acknowledgments We thank the Gulf Coast Soil and Water Conservation District for providing technical support and equipment for the study. Funding was provided by the Coastal Protection and Restoration Authority of Louisiana. Literature Cited Ambrose, R.F., J.C. Callaway, and S.F. Lee. 2006. An evaluation of compensatory mitigation projects permitted under Clean Water Act Section 401 by the California State Water Quality Control Board, 1991–2002. Los Angeles Regional Water Quality Control Board, Los Angeles, CA. Barras, J.A. 2006. Land-area change in coastal Louisiana after the 2005 hurricanes: A series of three maps: US Geological Survey Open-File Report 2006-1274. Available online at http://pubs.usgs.gov/of/2006/1274/. Accessed 12 December 2012. Barras, J.A. 2009. Land-area change and overview of major hurricane-impacts in coastal Louisiana, 2004–2008: US Geological Survey Scientific Investigations Map 3080. US Geological Survey, Reston, VA. Brown, S.C. 1999. Vegetation similarity and avifaunal food value of restored and natural marshes in northern New York. Restoration Ecology 7:56–68. Cairns, J., Jr. 2000. Setting ecological-restoration goals for technical feasibility and scientific validity. Ecological Engineering 15:17–18. Cole C.A., and D. Shaffer. 2002. Section 404 wetland mitigation and permit success-criteria in Pennsylvania, USA, 1986–1999. Environmental Management 30:50 8–518. Confer, S.R., and W.A. Niering. 1992. Comparison of created and natural freshwater emergent wetlands in Connecticut (USA). Wetlands Ecology and Management 2:143–156. Craft, C., P. Megonigal, S. Broome, J. Stevenson, R. Freese, J. Cornell, L. Zheng, and J. Sacco. 2003. The pace of ecosystem development of constructed Spartina alterniflora marshes. Ecological Applications 13:1417–1432. Currin, C.A., P.C. Delano, and L.M. Valdes-Weaver. 2008. Utilization of a citizen monitoring- protocol to assess the structure and function of natural and stabilized fringing saltmarshes in North Carolina. Wetlands Ecology and Management 16:97–118. Edwards, K.R., and C.E. Proffitt. 2003. Comparison of wetland structural characteristics between created and natural salt-marshes in southwest Louisiana, USA. Wetlands 23:344–356. Gleason, M.L., D.A. Elmer, N.C. Pien, and J.S. Fisher. 1979. Effects of stem density upon sediment retention by Salt-marsh Cordgrass, Spartina alterniflora Loisel. Estuarine Research Federation 2:271–273. Kaswadji, R.S., J.G. Gosselink, and R.E. Turner. 1990. Estimation of primary production using five different methods in a Spartina alterniflora salt-marsh. Wetlands Ecology and Management 1:57–64. Southeastern Naturalist 553 J.M. Soileau, E.K. Lyons, B. Chung, J. Hoffman, and F. LeMieux 2018 Vol. 17, No. 4 Kentula, M.E. 2000. Perspectives on setting success criteria for wetland restoration. Ecological Engineering 15:199–209. Midkiff, C.T., A.J. Roy, and R. Nolde. 1995. Soil survey of Cameron Parish, Louisiana. Soil Conservation Service, Alexandria, LA. Minello, T.J., and R.J. Zimmerman. 2003. Utilization of natural and transplanted Texas salt–marshes by fish and decapod crustaceans. Marine Ecology 90: 273–285. Minkin, P., and R. Ladd. 2003. Success of corps-required mitigation in New England. USACE New England District, Concord, MA Mullens, A.W. 2007. Strategies for establishing Spartina alterniflora on newly constructed marsh terraces in coastal Louisiana. M.Sc. Thesis, Louisiana State University, Baton, Rouge, LA. National Oceanic and Atmospheric Administration (NOAA). 2012. National Weather Service Weather Forecast Office. Yearly climate summaries, Lake Charles, Louisiana. Available online at http://www.noaa.gov/. Accessed 21 December 2012. National Research Council (NRC). 1992. Restoration of Aquatic Ecosystems. National Academy Press, Washington, DC. 552 pp. Neckles, H.A., M. Dionne, D.M. Burdick, C.T. Roman, R. Buchsbaum, and E. Hutchins. 2002. A monitoring protocol to assess tidal restoration of salt-marshes on local and regional scales. Restoration Ecology 10:556–563. Quammen, M.L. 1986. Summary of the conference and information needs for mitigation in wetlands. Pp. 151–158, In R. Strickland (Ed.). Wetland Functions, Rehabilitation, and Creation in the Pacific Northwest: The State of Our Understanding. Washington State Department of Ecology, Olympia, WA. Ruiz-Jaen, M.C., and T.M. Aide. 2005. Restoration success: How is it being measured? Restoration Ecology 13:569–577. Seliskar, D.M., J.L. Gallagher, D.M. Burdick, and L.A. Mutz. 2002. The regulation of ecosystem function by ecotypic variation in the dominant plant: A Spartina alterniflora salt-marsh case study. Journal of Ecology 90:1–11. Short, F.T., D.M. Burdick, C.A. Short, R.C. Davis, and P.A. Morgan. 2000. Developing success criteria for restored Eelgrass, salt-marsh, and mud-flat habitats. Ecological Engineering 15:239–252. Society of Ecological Restoration International Science and Policy Working Group (SER). 2004. The SER International Primer on Ecological Restoration. Available online at http// www.ser.org. Accessed 12 October 2012. Sudol, M.F., and R.F. Ambrose. 2002. The US Clean Water Act and habitat replacement: Evaluation of mitigation sites in Orange County, California, USA. Environmental Management 30:727–734. Travis, S.E., and J.B. Grace. 2010. Predicting performance for ecological restoration: A case study using Spartina alterniflora. Ecological Applications 20:192–204. West, T.L., L.M. Clough, and W.G. Ambrose. 2000. Assessment of function in an oligotrophic environment: Lessons learned by comparing created and natural habitats. Ecological Engineering 15:303–321. Whigham, D.F. 1999. Ecological issues related to wetland preservation, restoration, creation, and assessment. The Science of the Total Environment 240:31–40. Zedler, J.B. 1996. Tidal-wetland restoration: A scientific perspective and Southern California focus. Publication Number T-038, California Sea Grant College System, La Jolla, CA. Zedler, J.B., and J.C. Callaway. 2000. Evaluating the progress of engineered tidal wetlands. Ecological Engineering 15:211–225.