ARTICLE

Diverse Characteristics of Wetlands Restoredunder the Wetlands Reserve Program in the SoutheasternUnited States

Diane De Steven & Joel M. Gramling

Received: 13 February 2012 /Accepted: 27 March 2012 /Published online: 23 May 2012

Abstract The Wetlands Reserve Program (WRP) restoresconverted or degraded wetlands on private working lands;however, the nature and outcomes of such efforts are un-documented in the Southeastern U.S. Identification of wet-land types is needed to assess the programs conservationbenefits, because ecological functions differ with hydrogeo-morphic (HGM) type. We reviewed >100 WRP projectsacross the Southeast PiedmontCoastal Plain to characterizetheir wetland types and to evaluate whether restorationpractices favored original or modified functions. The proj-ects encompassed four HGM types and diverse pre-restoration conditions. Nearly half were converted wetlandsretired from active agriculture; the remainder were eitherdrained vegetated wetlands or forested bottomlands degrad-ed by timber harvest. Hydrology-repair practices varied bywetland type and prior condition, with differing functionalimplications. Depressions and flats typically were restored,whereas low-order riparian sites and prior-agriculture flood-plains were often modified to enhance water retention.Timber-harvested floodplains were restored by removingbarriers to water flow and biotic connectivity. Vegetationrestoration was generally passive, but tree planting wasfrequent on prior-agriculture sites. Field surveys suggestedthat most projects had positive indicators of wetland hydrol-ogy, vegetation, and faunal use. The variety of Southeastern

WRP wetlands has implications for ecosystem services atlocal and landscape scales.

Keywords Coastal Plain region,USA .ConservationEffectsAssessment Project (CEAP) . Ecosystem services . HGM .

Wetland restoration .Wetlands Reserve Program

Introduction

Wetland degradation and conversion result in the loss ofsignificant ecosystem services that include floodwater storage,water-quality improvement, carbon sequestration, and wildlifehabitat. Agriculture and other working-land uses are welldocumented causes of historical wetland damage and loss(e.g., Dahl 2000). In the United States, the ConservationTitle of the Farm Bill provides financial incentives to privateagricultural landowners for implementing various conserva-tion practices that can reduce the adverse environmentalimpacts of agriculture and enhance the provision of naturalhabitat (Mausbach and Dedrick 2004; Johnson and Monke2010). Certain practices are aimed specifically at recoveringor improving the ecological functions of wetlands on workinglands (Brinson and Eckles 2011).

Interest in accountability for the multi-billion dollar fund-ing of Farm Bill programs led the U.S. Department ofAgricultures Natural Resources Conservation Service(NRCS) to establish the Conservation Effects AssessmentProject (CEAP), a national-scale effort to measure the envi-ronmental benefits of federal conservation programs andpractices (Mausbach and Dedrick 2004; Duriancik et al.2008). Under the CEAPWetlands component, region-based studies are being conducted to assess and quantifythe ecosystem services derived from wetland conservationpractices (Eckles 2011). The core practice is Wetland

D. De Steven (*)U.S. Forest Service, Southern Research Station,Center for Bottomland Hardwoods Research,Stoneville, MS 38776, USAe-mail: ddesteven@fs.fed.us

J. M. GramlingDepartment of Biology, The Citadel,Charleston, SC 29409, USA

Wetlands (2012) 32:593604DOI 10.1007/s13157-012-0303-y

# US Government 2012

Restoration, which has been implemented mainly under theWetlands Reserve and Conservation Reserve Programs.These two programs have been applied extensively inagriculture-dominated regions such as the Northern GreatPlains and the Mississippi Alluvial Valley, where the natureand outcomes of the restoration practice have been fairly wellstudied (reviews in Gleason et al. 2011; Faulkner et al. 2011).However, other U.S. regions have received much less atten-tion (Eckles 2011). In particular, a review for the SoutheastPiedmontCoastal Plain (De Steven and Lowrance 2011)found that information on restorations under Farm Bill pro-grams was scarce to non-existent, with little indication of whatkinds of wetlands were targeted or what specific practiceswere used to rehabilitate them.

Two features of the Southeast have important implica-tions for assessing the regional benefits provided by wetlandpractices and programs (De Steven and Lowrance 2011).First, agriculture is a relatively smaller land use compared toforested land (~20 % vs. >60 %; USDA 2006); this poten-tially limits the landscape extent of program activity com-pared to agricultural regions. Second, the Southeast hasabundant and diverse wetland hydrogeomorphic (HGM)classes, despite historical wetland losses to both agriculturaland forestry activities (Hefner and Brown 1985; Hefner etal. 1994). Wetland diversity complicates assessment becauseHGM types differ in hydrodynamics and thus in their prin-cipal ecosystem services (Brinson and Reinhardt 1998;Smith et al. 2008). Knowledge of the wetland types beingrestored and their interactions with specific practices wouldenhance the ability to assess the services gained from con-servation programs.

Wetland restoration on Southeastern program lands isaccomplished mainly through the Wetlands ReserveProgram (WRP), established nationwide in 1995. TheWRP provides funds for restoring degraded wetlands withan eligible history of agricultural use, and also offers one-time, per-area payments for enrolling the wetland tracts intopermanent or 30-year conservation easements. Non-easement, 10-year agreements fund restoration costs only.Landowner participation is voluntary and based on mutuallyagreed goals. The programs traditional emphasis has beento maximize benefits for migratory waterfowl and otherwildlife, but other ecosystem services may be addressedalso. As part of the CEAPWetlands effort, we analyzedthe characteristics of wetland restorations on SoutheasternWRP lands and assessed coarse-scale evidence for success.Analogous to work by Gwin et al. (1999) on compensatory-mitigation wetlands, we also evaluated how practices usedfor restoration might affect wetland functions from a hydro-geomorphic perspective. Gwin et al. found that mitigationprojects often created atypical HGM types by modifyingsites to enhance water retention. A similar outcome is pos-sible under the WRP, which views restoration in broad terms

and allows establishing managed wetlands for wildlife-habitat objectives. Here we present the study findings inrelation to these questions: 1) what wetland HGM typesare being restored, and by what methods?, 2) do the princi-pal restoration practices result in original or modified HGMfunctions?, and 3) irrespective of HGM effects, does resto-ration successfully establish functional wetland conditions?Based on a survey of more than 100 projects completed overa 12-year period, this summary provides the first systematicdescription of WRP wetland restorations in the SoutheasternUnited States.

Methods

Under the CEAPWetlands framework, the Southeast re-gion encompasses the Piedmont and Coastal Plain physio-graphic province spanning from South Carolina toMississippi, but excluding the Mississippi Alluvial Valley(MAV) and the south Florida OkeechobeeEverglades sys-tem as separate, distinctive eco-regions (Brinson and Eckles2011; De Steven and Lowrance 2011). Our study examinedrestoration projects across three Southeast states with the ma-jority of reported WRP activity: South Carolina (SC), Georgia(GA), and the Coastal Plain portion of Mississippi (MS). Forpurposes of the study, a project is a defined wetland orwetland tract with coordinated planning that may involve mul-tiple landowner contracts.

A sample of projects was chosen by first obtaining allNRCS database records of completed WRP restorations thatwere reported for the three states during a fixed time period(20002008). It can take 24 years to finish a restorationafter a site is enrolled, so the reporting time period capturedan enrollment period spanning the first 910 years of thenationwide WRP. By basing the selection on a wide span ofreporting years, we obtained a broad cross-section of proj-ects without pre-determined selection bias for any wetlandattributes analyzed. After the addition of some projectswhose records were missing in the NRCS database, the finalsample consisted of 109 WRP projects enrolled during19962004 and completed over 12 years (19982009)(Table 1). This represented over half of SC contracts andnearly all GA and MS contracts for the enrollment period,providing a broadly representative sample of completedprojects up to a recent date. The projects occurred in allphysiographic sub-regions (Piedmont, Hilly Coastal Plain,Coastal Flats, Mississippi Loess Uplands). Nearly all WRPtracts (95 %) were in perpetual or 30-year conservationeasements, reflecting the enrollment options that land-owners chose most often.

WRP contract files in NRCS State and Field Offices wereexamined to assemble relevant data for the projects. Allprojects were on private lands, and all information was

594 Wetlands (2012) 32:593604

compiled under confidentiality provisions required by theFarm Bill. Each project was characterized for its originalwetland HGM type, habitat condition before restoration,area of the easement/tract, relative areas of hydric andnon-hydric soils, and the principal methods (practices) usedfor restoration. Wetland types were identified to HGM classor subclass based on topographic maps (USGS 30x60 and7.5). Prior habitat condition (in active agriculture vs. otherland cover) was determined from aerial photos and other filenotes. Easement size and restoration practices were obtainedfrom written project plans; practices included ditch plug-ging, dike construction, installing water-control structures,and vegetation planting. Areas of hydric and non-hydricsoils were estimated from overlays of WRP tract boundariesonto NRCS Web Soil Survey maps (at http://websoilsurvey.nrcs.usda.gov/app/). In addition, each project wasscored for whether the functional outcome of the mainhydrology-repair practice would be to restore or modifythe wetlands inherent hydrodynamic pattern and functions;i.e., whether the wetland retained its original HGM characteror was altered to behave as a different HGM type (cf. Gwinet al. 1999). Appropriate to class variables, three-way con-tingency analyses (log-linear models with likelihood-ratiochi-square tests; Sokal and Rohlf 1981) were used to test iffrequency of the various practices differed in relation towetland type and prior habitat condition, including a testfor type-by-condition interaction. Interaction tests arereported only if significant. For functional outcomes (re-stored, modified, or no change), chi-square analysis wasused to test if outcome frequencies differed between wetlandtypes.

We assessed the field status of project wetlands with acoarse-scale approach. Restoration under the WRP is notcompensatory or explicitly reference-based; the emphasis ison achieving habitat and functional success rather thanemulating specific wetland communities. Pre-restorationfield data suitable for measuring progress are also non-existent. Therefore, as an indication of project success, weused qualitative methods to determine if a sample of re-stored sites met ecological criteria for functional wetlandconditions. Approximately half of the projects (n053) werevisited and scored for field indicators of wetland hydrology,vegetation, and faunal presence. Nearly all GA and MS

projects were included (n013 and 11, respectively), omit-ting a few redundant or newly-completed sites. The selectedSC projects were a random sample (n029) stratified byHGM type and prior habitat condition, excluding someextremely large sites that were not feasible for survey.Time since restoration averaged 6 years (range 211 years)for all 53 projects.

One-day site visits were conducted during JulyAugust2010. These are optimal months for determining plant com-position and relative duration of growing-season wetness,although the region experienced below-normal rainfall andsummer drought in that year (NOAA 2010). Each site wasassessed with spot-surveys at 14 wetland locations per site(number scaled to wetland size); the number of locations persite was limited by time and area constraints, but it includeddifferent wetland cover types if these were evident fromaerial photographs. Upland areas were not surveyed. Weassessed wetland conditions by using an adaptation of thescoring method for routine wetland determinations (ACOEEnvironmental Laboratory 1987), omitting soil indicatorsbecause WRP site eligibility is based on presence of hydricsoils. At each location, we traversed the general area andnoted presence/absence of 10 primary and 5 secondary fieldindicators of existing or recent inundation/saturation (fromACOE 2008). We also recorded all dominant plant speciesin four strata (tree, sapling/shrub, herb, woody vine), wherea dominant was any species comprising ~20 % or more oftotal stratum cover by visual estimation. Species wereclassed as to wetland indicator category (Reed 1997), nativestatus (USDA 2011), and (if non-native) potential invasive-ness based on regional exotic-species rankings (at www.invasive.org/south/ and www.invasiveplantatlas.org).Quantitative sampling for fauna was not possible, but at eachlocation we noted any presence of aquatic or wetland-dependent animals in 7 broad taxon groups (wading birds,waterfowl, fish, aquatic insects, amphibians, reptiles,mammals).

Data were compiled to site level and summarized intoecological metrics for the hydrologic and vegetation con-ditions at each project wetland. Metrics included the numberand frequency of hydrology indicators, species richness ofthe dominant plants, and relative percentages of native,hydrophytic (FAC or wetter), and wetland (OBL, FACW)

Table 1 Features of 109 South-eastern WRP restoration projectsevaluated, by state and overall

aCoastal Plain only (excludes theMississippi Alluvial Valleysub-region of the state)

Feature South Carolina Georgia Mississippia All

Number of restoration projects 78 17 14 109

Number of WRP contracts 87 19 12 118

Total area enrolled (ha) 10,690 2,914 964 14,568

WRP enrollment year (range) 19962004 19962004 19972002 19962004

Project completion year (range) 19982009 19982009 20002007 19982009

Projects field-surveyed (number) 29 13 11 53

Wetlands (2012) 32:593604 595

species. A Prevalence Index (mean wetland-indicator scoreweighted by species frequency; ACOE 2008) and an indexof floristic quality (Floristic Assessment Quotient forWetlands, FAQWet4 of Ervin et al. 2006) were also calcu-lated for each site. Percent hydrophytic species equates tothe Dominance Test, where values >50 % indicate hydro-phytic vegetation (ACOE 2008). The Prevalence Index isscaled from 1 (OBL) to 5 (UPL); values

agriculture headwaters and floodplains (88 % and 45 % ofprojects, respectively). On the headwater sites, either creekflows were impounded directly or diked depressionalponds were built on adjacent creek banks. On the floodplainsites, the impoundments were often developed from existingflood-prevention dikes. In contrast, frequent practices onvegetated (forested) floodplains were breaching roads/dikesand installing rock-fill crossings or stream-crossing struc-tures (Fig. 2) to increase hydrologic and biotic connectivityacross the floodplain and to the river. Small, managed

impoundments could be established on these tracts, al-though some were pre-existing from earlier land use.

Excavating small areas of macrotopography (swales,potholes) was a secondary hydrologic practice for enhanc-ing water-depth variety in a wetland, principally on prior-agricultural flats (Fig. 2). In rare cases where a primaryrepair method proved infeasible, macrotopography was usedto provide some water storage in lieu of full restoration.Finally, 11 % of projects had no hydrology practices appliedat all; typically this occurred when it was found that past site

MS GA SCState

0

10

20

30

40

nu

mbe

ro

fpro

jects mainstem floodplain

riparian headwaterwet flatdepression

depression flat headwater floodplainWetland Type

0

10

20

30

40

num

ber

ofp

rojec

ts

natural vegetationagriculture (crop/pasture)

(a)

(b)

Fig. 1 Wetland HGM typesand pre-restoration habitat con-ditions for 109 SoutheasternWRP projects. a Frequency ofwetland types by state. b Fre-quency of prior habitat condi-tions by wetland type

Wetlands (2012) 32:593604 597

drainage was no longer effective, or when the tract (usuallya forested floodplain) enhanced the habitat value of anadjacent existing WRP easement without needing hydrolog-ic repair.

WRP projects are broadly regarded as wetland restora-tion, but a hydrology practice may restore original hydro-dynamic (HGM) function, modify it to create different HGMfunction, or result in no change if no hydrology repairoccurred. For example, blocking drainage would restorethe inherent hydrodynamics of ditched depressions and flats,whereas constructing impoundments would alter riverinewetlands that normally are driven by surface flows andoverbank flooding. From this perspective, the applied prac-tices resulted in differing functional outcomes (Fig. 3). Mosthydrologic regimes of depressions and flats were essentiallyrestored (76 %), with or without added water-control. Incontrast, 83 % of headwater sites were modified (enhanced)to create managed impoundments within a low-order river-ine setting. Functional outcomes for floodplains differedbased on how prior condition influenced restoration practi-ces (Fig. 3). Most forested floodplains were restored bybreach/crossing practices (50 %) or protected without hy-drologic repair (31 %), whereas most prior-agriculturefloodplains (82 %) were modified for managed impound-ments and limited connectivity to adjacent rivers.

Vegetation Restoration and Upland Habitat

Practices to restore wetland vegetation varied with priorhabitat condition and by state. Tree planting was frequenton prior-agriculture tracts (3582 % of projects per wetlandtype), but not on prior-vegetated sites (Fig. 2). Active refor-estation was typical on MS projects (93 % of sites), giventhat all were prior-agriculture riverine lowlands. On suchtracts, a mix of 58 bottomland species (e.g., oaks, Quercusspp.; green ash, Fraxinus pennsylvanica; baldcypress,Taxodium distichum) was planted across areas outside ofthe managed moist-soil units. The approach favoredheavy-seeded oaks on the presumption that light-seededspecies would colonize naturally (see King and Keeland1999). Passive revegetation was favored in SC and GA,

with supplemental use of planted trees (47 % of GA projectsand 12 % of SC projects). Planting was done mainly indepressions and flats and was based on microsite, typicallywith 12 wetland species (cypress, Taxodium ascendens orT. distichum; swamp tupelo, Nyssa biflora) placed centrallyand 23 bottomland oak species on higher wetland margins.On a few depressions and flats, vegetation restoration con-sisted solely of allowing natural recovery after removal oflivestock grazing. Likewise, timber-harvested floodplainswere not actively reforested, but instead were left to regen-erate naturally.

Upland areas, when present, were restored to naturalvegetation and/or managed with various wildlife habitatpractices. In SC and GA, practices on non-floodplainuplands included establishing native early-successiongrassland or longleaf pine (Pinus palustris) forest, bothof which might be managed with prescribed fire; thesepractices were planned on nearly 60 % of the non-floodplain tracts with available uplands. Planting agri-cultural grains (e.g., corn, sorghum) in small wildlife-foodplots was planned on almost 50 % of all projects, including onfloodplains.

Field Condition of Project Wetlands

Of the 53 project sites visited in late summer 2010, most hadmultiple positive metrics of hydrology function and wetlandvegetation, irrespective of wetland type (Table 3). No met-rics were correlated with time since restoration (Pearsoncorrelations, all P n.s.). With respect to hydrology, 38 sites(72 %) were inundated or saturated despite summer-droughtconditions, and 7 other sites without visible water (13 %)had at least 3 other primary indicators of recent inundation.The number of hydrology indicators averaged 45 per site(Table 3), but six sites that were dry when visited (3 depres-sions, 3 flats) had either no or few other indicators.Hydrology indicators did not differ with prior habitat con-dition (all P n.s.). Nearly all sites with installed water-control structures had water present when sampled, but theapparent extent of water-level management was variable. Of28 sites with control structures or impoundments, an

Table 2 Easement/tract size andwetland area of 109 Southeast-ern WRP projects, where wet-land area is estimated as the % ofeasement area with hydric soils

Wetland HGM type n Easement area (ha) Wetland area (ha)

Mean Range Mean Range

Depression 29 45 2249 27 2168

Flat 20 151 5779 114 4748

Riparian headwater 23 34 4167 24 3135

Mainstem floodplain (all) 37 256 41093 209 4886

Timbered floodplains 26 321 121093 263 10886

Prior-agriculture floodplains 11 101 4270 83 4212

598 Wetlands (2012) 32:593604

estimated 57 % were being actively managed, 18 % werenot managed, and the rest (25 %) could not be determined.

Almost 380 plant taxa were recorded as dominants acrossall sites; the list included aquatic plants (5 %), grasses andsedges (31 %), forbs (29 %), and woody species (35 %).Overall, the dominant vegetation averaged 88 % hydrophyt-ic species, 63 % wetland species, and 95 % native species,with little difference among wetland types (Table 3).Prevalence Index and FAQWet4 values averaged 2.1 and

10.7, respectively, indicating native wetland vegetation. Anapparent difference in species richness (Table 3) was likelyan artifact of differing numbers of survey locations per site(mean of ~2 in depressions/flats versus ~3 in headwaters/floodplains). No site had less than ~60 % hydrophyticspecies as dominants; however, 68 sites (mainly depres-sions or flats) had 2.5, and most of these were siteswith few to no hydrology indicators.

0.0

0.2

0.4

0.6

0.8

1.0

pro

porti

onofp

rojec

ts Natural VegetationAgriculture

Prior Habitat Condition

0.0

0.2

0.4

0.6

0.8

1.0

prop

ort i

ono

fpro

jects

depression flat headwater floodplainWetland Type

0.0

0.2

0.4

0.6

0.8

1.0

propo

rtion

ofp

rojec

t s

depression flat headwater floodplainWetland Type

gnissorC/hcaerBkaerBeliT/gulPhctiD

tnemdnuopmIlortnoCretaW

gnitnalPeerTyhpargopotorcaM

Phgm**, Ppc n.s. Phgm**, Ppc**

Phgm*, Ppc* Phgm*, Ppc*, Pint*

Phgm n.s., Ppc**Phgm*, Ppc**

Fig. 2 Proportion of WRP projects in a given wetland type and priorhabitat condition that used each of 6 restoration practices. P-valuesfrom chi-square tests for effect of wetland HGM type (Phgm, df03)

and prior condition (Ppc, df01) are noted by ** P0.01; * P0.05;n.s., not significant. Only one P-value for type-by-condition interaction(Pint, df03) was significant

Wetlands (2012) 32:593604 599

Vegetation metrics did not differ with prior habitat con-dition, except for a few related to species nativity. One ormore non-native species was recorded in 74 % of agricul-tural sites vs. 46 % of prior-vegetated sites (chi-square, df01,P

78 % of prior-agriculture sites were re-established as open ormixed habitats, whereas 69 % of prior-vegetated sites wereshrub or forested wetland (chi-square P 75 % of sites, withthe number of observed faunal groups averaging 23 per site.Wading birds, waterfowl, and/or amphibians occurred at nearly60 % of the GA and MS sites.

Discussion

Farm Bill program wetlands have been understudied in partbecause of privacy and access constraints; however, the CEAPWetlands assessment (Eckles 2011) will enhance opportunitiesfor wider examination. The most-studied WRP efforts to datehave been in major agricultural regions (e.g., Great Plains,Mississippi Alluvial Valley), where there may be one dominantwetland type and where converted wetlands on highly alteredcroplands are targeted for restoration (King et al. 2006;Faulkner et al. 2011; Gleason et al. 2011). Our broad assess-ment of projects in a mixed land-use region revealed severalcontrasts and unexpected program versatility. SoutheasternWRP wetlands are characterized by multiple HGM types andby varied pre-restoration conditions that include cropped orgrazed sites, drained wetlands with semi-natural vegetation,and timber-harvested bottomlands with no contemporary crop-ping history. Restoration approaches were partly adjusted towetland type and initial condition, but with differing functionaloutcomes. Rapid field surveys indicated generally successfulestablishment of functional wetlands with a native flora andhabitat favorable to wetland-dependent wildlife.

Wetlands in the Southeastern WRP

The characteristics of the enrolled wetlands reflected inher-ent program flexibility. The WRP is generally thought to

target prior-converted or farmed wetlands on croplands;however, program guidelines allow for eligibility of hydro-logically degraded wetlands in various working-land set-tings, including rangelands and forest production lands(NRCS 2009). Individual states can establish priority rank-ings for habitats of concern in relation to the prevailingpatterns of land use and landowner participation. InCoastal Plain Mississippi, low- and high-order riverine wet-lands were most often converted to cropland and thus werepriortized for restoration (although high competitive de-mand for the WRP in the MAV portion of the state tendsto limit the number of Coastal Plain enrollments). The othertwo states also targeted recovery of certain distinctive wet-lands uncommon in Mississippi (e.g., isolated depressions,Carolina bays), and South Carolina developed an option in2002 to address degraded bottomland habitats. Collectively,these approaches yielded a region-wide diversity of restoredwetland types, initial conditions, and ecosystem servicesthat may be recovered. There may be a similar variety ofWRP wetlands in other mixed land-use regions of the U.S.,but the extent is unknown because HGM type generally hasnot been used for tracking program wetlands.

Functional Effects of WRP Restorations

The WRP can broadly address recovery of wetland functions,but its traditional focus on wildlife benefits influences resto-ration methods. Hydrology enhancement (water-level con-trol, partial impoundments, macrotopography) is allowed onup to 30 % of a restored tract (NRCS 2009), though in realitythe spatial extent may be greater. WRP projects also mayincorporate other targeted wildlife-management techniques(vegetation manipulation, food plots, etc.; Strader andStinson 2005), thus constrasting with mitigation or ecologicalrestorations that focus on maximizing equivalency to undis-turbed wetlands (e.g., Kentula 2000; Zedler 2000; Reiss et al.2009). The traditional wildlife-oriented practices may some-times result in poor ecological fit between managementgoals and natural ecosystem processes (Euliss et al. 2008).

On many of the Southeastern projects, hydrology repairinvolved installing some form of water-level control, reflect-ing an explicit or implicit landowner desire for waterfowl

Table 4 Principal vegetationhabitats of 53 WRP project sites.Values are the number of proj-ects representing a given vege-tation class. Habitat-classfrequency differed among wet-land types (chi-square P

habitat. An HGM perspective illustrated that the principalrepair methods would recover inherent hydrologic functionsin some cases, but substantially modify them in others (seeSmith et al. 2008). Blocking ditches/drains in depressionsand flats tends to re-establish natural rainfall-driven hydro-logic regimes and functions, with added water control alter-ing hydroperiod timing but not the basic functional type.Similarly, the breach/crossing practices applied on forestedfloodplains promote natural flooding dynamics and hydro-logic connectivity. Conversely, on low-order riparian sites,installing small dams or impoundments produces an atypicaldepression-in-riverine geomorphic setting. Gwin et al.(1999) had observed a similar pattern of depressions createdwithin riparian wetlands for mitigation purposesan ex-change of one functional type for another. While enhancingwater retention and allowing for periodic de-watering, thesesemi-enclosed riparian impoundments can reduce hydrolog-ic and biogeochemical functions related to surface flows andflood pulsing (Smith et al. 2008). Impoundments on prior-agriculture floodplains can have similar effects but reflectpractical trade-offs. Such tracts may be adjacent to otherprivate farm properties, which constrains the ability to allownatural river flooding. Instead, managed moist-soil units andexcavated swales add hydrologic variety to agriculturallyleveled sites, while retaining the flood-control dikes thatprovide landowner access to the mix of open and reforestedhabitats.

WRP Project Success

Whether wetlands were restored or modified, the field sur-veys indicated that most projects (8587 %) are providingfunctional hydrology and wetland habitat. Diverse vegeta-tion communities were dominated by native hydrophyticspecies that are common in natural wetlands. Some poten-tially invasive species occurred locally, but non-nativesappeared less frequent overall compared to other regionswhere exotics may comprise >20 % of the restored flora(e.g., Aronson and Galatowitsch 2008; Gleason et al. 2008).Consistent with their younger successional age, prior-agricultural sites had more open vegetation communities,sometimes with remnant old-field species. Few other fieldindicators differed based on prior habitat status, suggestingthat restoration improved wetland conditions in the moredegraded agricultural sites relative to already vegetatedsites. Wetland and aquatic fauna were seen at many projectsites, although frequency of habitat use or species-specifichabitat quality could not be quantified.

Achieving close resemblance to reference wetlands is notan explicit success goal for WRP projects, partly becausethey are not compensatory for permitted wetland losses. Atminimum, the project sites had a plant composition indica-tive of native wetlands, but more complete floristic data

would be needed to evaluate similarity to possible referenceplant communities. The task is complex for Southeasternwetlands, where the historic communities are often un-known and where species composition is highly variable.Riverine wetlands are naturally forested, but depressionsand flats can support a range of vegetation types dependingupon temporal variability in hydroperiod, frequency of fire,and other factors (De Steven and Lowrance 2011). Wewould expect the greatest floristic divergence from naturalcommunities in the managed wetlands, which are maintainedin early-succession stages with a relatively narrow suite ofplant species. Wetlands that were vegetated before restorationwould likely be more similar to natural communities.

A small percentage of WRP projects (1115 %) appearedto be less successful hydrologically, possibly reflecting in-compatibility between project goals and site hydrogeomor-phic limitations. Most of the drier sites were depressions andflats, some of which can have inherently temporary orsaturation-driven hydroperiods that cannot be enhanced eas-ily (Rheinhardt et al. 2002; De Steven et al. 2010). Theremaining drier sites were first-order stream features thatmay naturally flood only temporarily or intermittently,resulting in unsuccessful enhancement with ditch plugs ordikes. Failure to recognize such site limitations can lead tounsatisfactory outcomes relative to expectations, particular-ly when the goal is a floodable waterfowl pond. Even wherea practice can enhance water retention, inappropriate appli-cation may result in poor wetland quality (Euliss et al.2008). For example, we noted that on two agricultural flood-plains where deep macrotopography had been excavated,the resulting ponds were stagnant with turbid water, anoxicsediments, and depauperate biota. Attention to HGM set-tings could improve project success by identifying site-compatible practices and realistic expectations for hydrologyrestoration. Projects that re-establish or mimic natural process-es are most likely to be self-sustaining in the long run (Smithet al. 2008). For example, inherently drier wetland sites thatmay not be favorable for waterfowl could still provide impor-tant habitat for other fauna (e.g., amphibians) that rely ontemporary or highly dynamic hydroperiods (e.g., Snodgrasset al. 2000).

Implications for Ecosystem Services and Assessment

The variety of program wetlands has implications for eco-system services at local and landscape scales. MostSoutheast Coastal Plain states have relatively few WRPenrollments (2050 per state) compared to high numbersin some agriculture-dominant regions (>400500 per state),although South Carolina now has >200 contracts (NRCSdata; cf. De Steven and Lowrance 2011). Because landownerparticipation is opportunistic, most Southeastern project types

602 Wetlands (2012) 32:593604

(depression, flat, low-order riparian, farmed floodplains) arespatially scattered and would provide ecosystem servicesmainly at a local scale. These projects have varied easementsizes that reflect both wetland geomorphology and theamounts of land that landowners offer for enrollment; however,native wetland plant communities may be supported irrespec-tive of wetland type and size. Smaller wetland easements mightlack sufficient area to form core habitat for some wildlifespecies, but they could still provide corridor or stepping-stoneareas to larger wetland habitats within the regional forest-mosaic landscape. Local wetland functions (hydrologic, bio-geochemical, biologic) may vary with the extent of restoredupland buffer habitat, which differed considerably among ease-ments. Targeted studies that quantify the ecosystem services ofthese local projects could assist future project planning byidentifying ways to optimize multiple environmental benefits(Brinson and Eckles 2011).

Compared to other project types, the forested floodplainprojects offer unique potential to provide ecosystem servicesat a watershed/landscape scale by encompassing large, to-pographically complex areas. Easements typically span thebreadth of a floodplain or braided channel system from theriver to the upland terrace. With restoration methods thatencourage natural floodplain hydrodynamics, there are syner-gistic benefits for habitat, floodwater storage, water quality,and other riverine functions that also affect environmentalquality in downstream estuarine habitats. Total area of flood-plain tracts in South Carolina had reached 19,000 ha by 2010,with nearly all in perpetual easements (SC-NRCS State Officedata). In several river systems, the program was able to as-semble small groups of adjacently owned tracts, ranging sin-gly from 20 to 1,100 ha, into larger patches of river corridorhabitat totaling 2501,600 ha (2.516 km2) or more. The fieldvisits suggested that these degraded floodplain tracts havevigorous tree regeneration, though it will take many years toassess the eventual forest composition and stand quality.Long-term prospects for recovery are favorable, given thatthe harvested forests had successfully regrown from wide-spread historic logging during the previous century (Sharitzand Mitsch 1993; Lockaby et al. 1997).

The study results will assist future CEAPWetlandsefforts to develop regional models for quantifying the eco-system services delivered by wetland practices and pro-grams. Conceptually, the Southeastern WRP sites reflect acomplex condition gradient from highly degraded to mini-mally disturbed, including sites restored from row-cropping,grazing, timber-harvest, or simple drainage. Restorationpractices seek to move degraded sites closer to naturalwetland functioning (NRCS 2006). Knowledge of wetlandtypes is key to understanding the principal services affectedas well as the potential trade-offs resulting from choice ofpractices and goals. Initial habitat condition offers a meansto estimate gains in services, since semi-natural vegetated

sites have likely retained or recovered some wetland func-tions compared to more degraded and younger agriculturalsites. While the scope of the WRP allows for a variety oftargeted goals, incorporating an HGM and site-conditionframework could improve assessments of cumulative bene-fits and performance relative to regional resource concerns(Brinson and Eckles 2011).

Acknowledgments We greatly appreciated the support of CEAPWetlands Science Coordinator Diane Eckles and the generous cooper-ation of NRCS State and Field Office staffs, especially WRP Coordi-nators Glenn Sandifer (SC), Mike Oliver (MS), and Keith Wooster(GA). We also thank the landowners for access to their WRP sites.Steven Hughes assisted in field work and data compilation, withadditional help from Kristie Burr, Cody Clark, Anne Cubeta, and KristiWharton. Editorial and statistical reviews were provided by Kay Kirkman,Rebecca Sharitz, Ray Souter, and TomDell. The study was funded throughtheUSDANRCS, theU.S. Forest Service Southern Research Station, and aUSFS Joint Venture Agreement with The Citadel Graduate College.

References

ACOE [Army Corps of Engineers] Environmental Laboratory (1987)Corps of Engineers wetlands delineation manual. Technical Re-port Y-87-1. U.S. Army Engineer Waterways Experiment Station,Vicksburg, MS

ACOE (2008) Interim regional supplement to the Corps of Engineerswetland delineation manual: Atlantic and Gulf Coastal PlainRegion. ERDC/EL TR-08-30. U.S. Army Engineer Researchand Development Center, Vicksburg, MS

Aronson MFJ, Galatowitsch S (2008) Long-term vegetation develop-ment of restored prairie pothole wetlands. Wetlands 28:883895

Brinson MM, Eckles SD (2011) U.S. Department of Agriculture con-servation program and practice effects on wetland ecosystemservices: a synthesis. Ecology Applied Supply 21(3):S116S127

Brinson MM, Reinhardt RD (1998) Wetland functions and relations tosocietal values. In: Messina MG, Conner WH (eds) Southernforested wetlands: ecology and management. Lewis Publishers,Boca Raton, pp 2948

Dahl TE (2000) Status and trends of wetlands in the conterminousUnited States, 1986 to 1997. U.S. Dept of Interior, Fish andWildlife Service, Washington, DC

De Steven D, Lowrance R (2011) Agricultural conservation practicesand wetland ecosystem services in the wetland-rich PiedmontCoastal Plain region. Ecology Applied Supply 21(3):S3S17

De Steven D, Sharitz RR, Barton CD (2010) Ecological outcomes andevaluation of success in passively restored Southeastern depres-sional wetlands. Wetlands 30:11291140

Duriancik LF, Bucks D, Dobrowolski JP, Drewes T, Eckles SD, JolleyL, Kellogg RL, Lund D, Makuch JR, ONeill MP, Rewa CA,Walbridge MR, Parry R, Weltz MA (2008) The first five years ofthe Conservation Effects Assessment Project. Journal Soil andWater Conservation 63:185A197A

Eckles SD (2011) Linking science, policy, and management to con-serve wetlands in agricultural landscapes: introduction to theSpecial Issue. Ecology Applied Supply 21(3):S1S2

Ervin GN, Herman BD, Bried JT, Holly DC (2006) Evaluating non-native species and wetland indicator status as components ofwetlands floristic assessment. Wetlands 26:11141129

Euliss NH Jr, Smith LM, Wilcox DA, Browne BA (2008) Linkingecosystem processes with wetland management goals: charting acourse for a sustainable future. Wetlands 28:553562

Wetlands (2012) 32:593604 603

Faulkner S, Barrow W Jr, Keeland B, Walls S, Telesco D (2011) Effects ofconservation practices on wetland ecosystem services in theMississippi Alluvial Valley. Ecology Applied Supply 21(3):S31S48

Gleason RA, Laubhan MK, Euliss NH Jr (eds) (2008) Ecosystemservices derived from wetland conservation practices in theUnited States Prairie Pothole Region with an emphasis on theU.S. Department of Agriculture Conservation Reserve andWetlands Reserve Programs. U.S. Geological Survey ProfessionalPaper 1745, Reston, VA

Gleason RA, Euliss NH Jr, Tangen BA, Laubhan MK, Browne BA(2011) USDA conservation program and practice effects on wet-land ecosystem services in the Prairie Pothole Region. EcologyApplied Supply 21(3):S65S81

Gwin SE, Kentula ME, Shaffer PW (1999) Evaluating the effects ofwetland regulation through hydrogeomorphic classification andlandscape profiles. Wetlands 19:477489

Hefner JM, Brown JD (1985) Wetland trends in the southeasternUnited States. Wetlands 4:111

Hefner JM, Wilen BO, Dahl TE, Frayer WE (1994). Southeastwetlands: status and trends, mid-1970s to mid-1980s. U. S.Department of Interior, Fish and Wildlife Service, Atlanta, GA

Johnson R, Monke J (2010) What is the Farm Bill? Report RS22131.Congressional Research Service, Washington, DC (http://www.nationalaglawcenter.org/assets/crs/RS22131.pdf)

Kentula ME (2000) Perspectives on setting success criteria for wetlandrestoration. Ecological Engineering 15:199209

King SL, Keeland BD (1999) Evaluation of reforestation in the LowerMississippi River Alluvial Valley. Restoration Ecology 7:348359

King SL, Twedt DJ, Wilson RR (2006) The role of the WetlandReserve Program in conservation efforts in the Mississippi RiverAlluvial Valley. Wildlife Society Bulletin 34:914920

Lockaby BG, Stanturf JA, Messina MG (1997) Effects of silviculturalactivity on ecological processes in floodplain forests of the south-ern United States: a review of existing reports. Forest Ecology andManagement 90:93100

Mausbach MJ, Dedrick AR (2004) The length we go: measuringenvironmental benefits of conservation practices. Journal Soiland Water Conservation 59:96A103A

National Research Council (1995) Wetlands: characteristics andboundaries. National Academy Press, Washington, DC

NRCS [Natural Resources Conservation Service] (2006) ConservationEffects Assessment Project: National Assessment-Wetlands com-ponent. USDA NRCS, Beltsville, MD

NRCS (2009) Wetlands Reserve Program interim final rule. FederalRegister 74(10):23172337

NOAA [National Oceanic and Atmospheric Administration] (2010)Climatological data annual summaries. National Climatic DataCenter, Asheville, NC

Reed PB Jr (1997) Revision of the national list of plant species thatoccur in wetlands. U.S Dept of Interior, Fish and Wildlife Service,Washington, DC

Reiss KC, Hernandez E, Brown MT (2009) Evaluation of permit successin wetland mitigation banking: a Florida case study. Wetlands29:907918

Rheinhardt RD, Rheinhardt MC, Brinson MM (2002) A regionalguidebook for applying the hydrogeomorphic approach to assess-ing wetland functions of wetland pine flats on mineral soils in theAtlantic and Gulf Coastal Plains. ERDC/EL TR-02-9. U.S. ArmyEngineer Research and Development Center, Vicksburg MS

Sharitz RR, Mitsch WJ (1993) Southern floodplain forests. In: MartinWH, Boyce SG, Echternacht AC (eds) Biodiversity of the south-eastern United States: lowland terrestrial communities. Wiley,New York, pp 311372

Smith LM, Euliss NH Jr, Wilcox DA, Brinson MM (2008) Applicationof a geomorphic and temporal perspective to wetland manage-ment in North America. Wetlands 28:563577

Snodgrass JW, Komoroski MJ, Bryan AL Jr, Burger J (2000) Relation-ships among isolated wetland size, hydroperiod, and amphibianspecies richness: implications for wetland regulations. ConservationBiology 14:414419

Sokal RR, Rohlf FJ (1981) Biometry, 2nd edn. Freeman, New YorkStrader RW, Stinson PH (2005) Moist-soil management guidelines for

the U.S. Fish and Wildlife Southeast Region. U.S. Department ofInterior, Fish and Wildlife Service, Jackson, MS

USDA (2006) Agricultural Statistics. U.S. Department of Agriculture,National Agricultural Statistics Service, Washington, DC

USDA (2011) The PLANTS database (http://plants.usda.gov). US De-partment of Agriculture, National Plant Data Team, Greensboro, NC

Zedler JB (2000) Progress in wetland restoration ecology. Trends inEcology & Evolution 15:402407

604 Wetlands (2012) 32:593604

Diverse Characteristics of Wetlands Restored under the Wetlands Reserve Program in the Southeastern United StatesAbstractIntroductionMethodsResultsCharacteristics of the Wetland RestorationsHydrology Repair and Functional OutcomesVegetation Restoration and Upland HabitatField Condition of Project Wetlands

DiscussionWetlands in the Southeastern WRPFunctional Effects of WRP RestorationsWRP Project Success

Implications for Ecosystem Services and AssessmentReferences