- Email: [email protected]
The impact of manipulating surface topography on the hydrologic restoration of a forested coastal wetland
Ecological Engineering 58 (2013) 35–43
Contents lists available at ScienceDirect
Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng
The impact of manipulating surface topography on the hydrologic restoration of a forested coastal wetland夽 Robert D. Jarzemsky a , Michael R. Burchell II b,∗ , Robert O. Evans b a b
U.S. Army Corps of Engineers, Chicago District, 111 N. Canal St. Suite 600, Chicago, IL 60606, United States Biological and Agricultural Engineering Campus, Box 7625, North Carolina State University, Raleigh, NC 27695, United States
a r t i c l e
i n f o
Article history: Received 31 August 2012 Received in revised form 20 April 2013 Accepted 8 June 2013 Available online 6 July 2013 Keywords: Wetland hydrology Restoration Microtopography
a b s t r a c t A wetland, converted to agriculture in the mid-1970s, was restored to re-establish a non-riverine wet hardwood forest community in eastern North Carolina. Three surface techniques were implemented during construction to determine their effect on successfully restoring target wetland hydrology. The surface treatments, replicated within a randomized complete block design, were: plugging field ditches without altering the land surface (PLUG), plugging the field ditches and roughening the surface (ROUGH), and plugging the field ditches and removing the field crown (CR). Hydrologic conditions for the restoration and a nearby reference site were evaluated based on three years of monitoring data. Daily water table depths between the restoration and reference were within 11 cm on average. An initial evaluation found inconsistencies of treatment effect between blocks, and an as-built survey later confirmed surface elevations within Block 3 deviated from the intended design and was excluded from further analysis. Water table and outflow conditions for the remaining treatment plots and the reference were evaluated using several hydrologic criteria. The CR treatment was found to produce the wettest surface conditions and exported the lowest volume of outflow. For the majority of criterion considered, CR also produced significantly wetter conditions than the reference. The PLUG and ROUGH treatments produced similar hydrologic conditions and tracked closely with the median hydrologic conditions in the reference. Based on the results of this study and several others in low lying coastal areas, plugging pre-existing field ditches may be adequate to restore jurisdictional wetland hydrology and match reference hydrologic conditions. However, surface roughening is low cost method to increase surface storage and introduce microtopographic diversity. For many areas, the removal of existing field crown may be cost-prohibitive and produce wetter than desired conditions. Crown removal should be reserved for sites which have borderline historic wetland hydrologic characteristics. © 2013 Elsevier B.V. All rights reserved.
1. Introduction The magnitude of wetland loss and alteration in the U.S. has resulted in measurable negative impacts on habitat and water quality. Therefore, efforts to strategically restore some of these lands back to their natural condition are vital to reclaim damaged or lost ecosystem functions. Approximately 50% of the land in eastern NC originally contained hydric soils, and over 50% of these native wetlands were altered to accommodate other land uses (Cashin et al., 1992). The majority of these wetland alterations occurred to enhance agricultural production, and were achieved
夽 Disclaimer: The views presented in this paper are those of the authors and do not necessarily represent the views of the US Army Corps of Engineers. ∗ Corresponding author. Tel.: +1 919 513 7372; fax: +1 919 515 6772. E-mail address: mike [email protected] (M.R. Burchell II). 0925-8574/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecoleng.2013.06.002
through land clearing, installation of subsurface drainage systems (comprised of open ditches or subsurface drains), and smoothing and crowning of soil surfaces to improve surface drainage (Lilly, 1981). This type of artificial drainage, while necessary for agricultural production in many parts of the country, has also resulted in environmental consequences. Studies have found that drainage from agriculture drainage increases peak runoff rates as much as 300–400% (Skaggs et al., 1980) and can lead to a 10-fold increase in nitrate (NO3 –N) losses (Skaggs et al., 2005). In addition to NO3 –N, agricultural drainage is also linked to increased exports of phosphorus and suspended solids (Skaggs et al., 1980). The goal of wetland restoration/creation is to successfully establish a system that exhibits the same structure and beneficial functions as a targeted wetland community in the most efficient manner possible. Inappropriate designs or inefficient implementation can result in restorations that are too expensive and fall short of achieving target ecosystem services. In a survey of practitioners,
36
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
Holman and Childres (1995) estimated that 40–50% of wetland restoration problems were simply due to poor site selection. Achieving appropriate hydroperiod is a major component in successful wetland restoration. A wetter site is not always better, since many natural wetlands alternate seasonally between dry and wet conditions. This is a major reason that at least one and preferably multiple reference wetlands should be identified prior to restoration to determine the target hydroperiod. To restore wetland hydrology on lands drained for agricultural practices, ditches that were installed must be plugged to reduce drainage from the site. However, Tweedy and Evans (2001) showed that simply plugging ditches may produce conditions that are drier than the targeted wetland community. This is especially true on sites that, prior to artificial drainage, may have only marginally satisfied the U.S. Army Corps of Engineers (USACE) jurisdictional hydrologic criterion (i.e. the water table within 30 cm of the surface continuously for more than 5% of growing season (USACE, 1987)). Removing the field crown and adding microtopography (defined here as slight ridges and depressions along the soil surface) in addition to ditch plugging may improve the chances of meeting wetland hydrology requirements, but it does increase costs. Currently, the U.S. Department of Agriculture – Natural Resources Conservation Service recommends implementing varying levels of microtopography to promote not only hydrologic but habitat diversity (NRCS, 2003). Several studies support this restoration technique because it can increase stormwater storage, prolong surface moisture during dry periods, provide more diverse habitat and vegetation, improve microbially mediated nutrient cycling and removal, and promote more overall ecosystem heterogeneity common in natural wetlands (Smith et al., 2012; Ahn and Dee, 2011; Courtwright and Findley, 2011; Simmons et al., 2007, 2011; Wolf et al., 2011; Alsfeld et al., 2009; Moser et al., 2009). Despite these existing studies, there is a lack of data that strongly supports a particular treatment practice for maximizing appropriate wetland hydrologic characteristics across entire restoration sites. This paper describes how various types of soil surface alterations affected wetland hydrology at a 100 ha forested wetland restoration in eastern North Carolina. In an effort to build on earlier results from research conducted by Wright et al. (2006), field monitoring at the restoration site was intensified to ultimately develop improved recommendations and guidance for future coastal wetland restorations. Goals of this study were to determine (1) differences in
hydrologic response of various soil–surface manipulation techniques, (2) how much surface manipulation was required to produce hydrologic conditions within a range of reference hydrologic conditions, and (3) what degree of surface manipulation would be considered optimum for producing acceptable hydrologic conditions at minimal restoration costs. It was hypothesized that roughening the surface and creating microtopographic features similar to those found in natural wetlands would be ideal for achieving hydrologic conditions found in a reference wetland. It was further hypothesized that simply plugging the ditches would produce the driest conditions because the intact field crown and smooth soil surface would result in more rapid surface drainage, while field crown removal would produce the wettest conditions due to the decreased surface flow gradient. Several hydrologic indicators and metrics were used to provide an extensive evaluation of the hydrology at the restoration site because of annual and seasonal climactic variations. A secondary goal was to demonstrate how these metrics might be applied to hydrologic assessments for future wetland restorations. 2. Materials and methods Field-scale studies were conducted at a coastal wetland restoration site located in Carteret County, North Carolina (Fig. 1). The 2400 ha site was artificially drained for agricultural row crop production in the 1970s before being acquired by the NC Coastal Federation for the purpose of restoration in 2002. Ditches were typically spaced at 100 m intervals and dug to 1 m depth, while fields were crowned by about 20 cm. Drainage water from the site discharged south into the North River estuary, an area important to the local shellfishing industry. The overall goals of the project were to restore hydrologic function, increase wetland habitat, and improve downstream water quality. The first of multiple phases of the project included hydrologic restoration of 100 ha of cropland to a non-riverine wet hardwood wetland ecosystem. This ecosystem is described as hardwood forest located on broad inter-stream flats with poorly drained mineral soils, and are common in the lower coastal plain of NC (Schafale and Weakley, 1990). Completed in 2003, Phase I included placement of earthen plugs in field ditches and installation flashboard risers to control water table levels and outflow. Wetland features such as 0.1 ha open water areas and simulated tree falls (about
Fig. 1. Vicinity map of the restoration site in Carteret County, NC.
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
12 per ha) were added across the site to increase topographic diversity. Following earthmoving, the site was planted with over 89,000 native wetland trees such as oaks (Quercus spp.), water tupelo (Nyssa aquatic), black gum (Nyssa sylvatica), bald cypress (Taxodium disticum), atlantic white cedar (Chamaecyparis thyoides), green ash (Fraxinus pennsylvanica), and longleaf pines (Pinus palustris). In addition to plantings, early successional herbaceous and woody vegetation quickly colonized the site. The soils were poorly to very poorly drained and were predominantly Deloss fine sandy loam with pockets of Wasda muck and Leon sand (Goodwin, 1984). Three levels of hydrologic restoration and surface microtopography were implemented on 6.5 ha test plots prior to planting. Each restoration technique was replicated three times in a randomized block design. The construction techniques included: 1. PLUG – hydrologic restoration using a series of earthen ditch plugs and water control structures. The land surface that existed under agricultural production (smoothed and crowned) remained intact, which minimized surface storage and maximized surface runoff toward the ditches. 2. ROUGH – in addition to practices described in PLUG, the field surface was roughed to increase microtopography. Tillage with a disk harrow was applied evenly across 25% of the treatment, creating surface mounds and hollows of ±15 cm, to increase surface storage. 3. CR – in addition to the practices described in PLUG, the field crown was removed and topsoil replaced. Surface storage was not intentionally enhanced, but removal of the crown was intended to lower the surface runoff gradient. Crown removal also resulted in a surface elevation that was closer in proximity to the groundwater table. Each 6.5 ha plot was separated by low earthen berms, constructed at the center point between plugged agricultural ditches to establish a barrier to surface flow between adjacent plots. Outflow
37
from the ditches was controlled by flashboard riser water control structures. Flashboards were installed to average ground surface elevation in each plot, and included a 30◦ V-notch that allowed excess soil–water to slowly drain down to 30 cm below average ground surface. Water table monitoring wells were installed in a single transect across the treatments (Fig. 2). Continuous water table recorders (Infinities USA Inc., Daytona Beach, FL) were installed in 10.2 cm diameter PVC monitoring wells at the mid-point between the plugged ditch and the berm, a location that corresponded to the average elevation of a typical crowned agricultural field. Water table depth was measured and logged hourly by the continuous water table recorders, and data were offloaded monthly. Two additional 5.1 cm diameter PVC wells were installed on either side of the continuous recorders, midway between the plugged ditch on one side and the berm on the other. Water table elevations in these wells, which were installed to provide a more spatial water table profile across the site, were measured manually on a monthly basis and correlated to the automatic measurements. A detailed survey using a Topcon Electronic Total Station was conducted in a swath across that site that included the transects to determine the as-built topography of treatments and the elevations of all the water table monitoring wells. Outflow from each plot was measured using potentiometers connected to a pulley-float system to measure stage over the 30◦ V-notch weir. Measurements were stored in a datalogger (SGT Engineering, Champaign, IL), and appropriate weir equations were applied to the stage readings to calculate drainage flowrate and volume as described in Jarzemsky (2009). Precipitation was recorded in two locations using a tipping bucket rain gauge and HOBO event datalogger (Onset Computer Corp., Bourne, MA). A nearby non-riverine wet hardwood wetland was also instrumented and served as a reference for comparison to the restoration treatments. The reference had a concave topography along an east-west axis (i.e. lowest elevation in the center) that resulted
Fig. 2. Monitoring schematic for Blocks 1 and 2 of the restoration site. PLUG, ROUGH, and CR represent wetland surface treatments. Wells designated NR were continuous water table monitoring, while locations designated by a and b represented the manual water table wells.
38
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
Fig. 3. Hydrologic monitoring schematic for the reference wetland site, which drained to a downstream tidal marsh. All wells had continuous water table measurements. Note: Wells followed by asterisk did not meet minimum USACE jurisdictional criteria and were excluded from analysis. Ref-Wet wells included REF 2, 5, and 8; Ref-Med wells included REF 13, 14, and 16; Ref-Dry wells included REF 6, 9, and 11.
in hydrologic variations across the site. To capture the range of wetness and extent of wetland conditions, the reference was instrumented with 13 continuous water table monitoring wells along three east-west transects (Fig. 3). Water table depths were monitored and recorded using the same instrumentation described earlier for the restoration areas. A detailed topographic survey was also conducted along each of the transects in the reference wetland to relate water table measurements to depth below ground surface. Outflow from the reference wetland was not monitored due to backwater conditions from tidal influences of the marsh located several hundred meters downstream. Water table data for 2006–2008 from the restoration treatments and the reference wetland were evaluated using four hydrologic criteria:
2. Average water table depth (full calendar year). 3. Sum of excess water above 30 cm, SEW30 (full calendar year). 4. Number of days of surface inundation (full calendar year).
1. Longest continuous period of water table within 30 cm of the surface (growing season).
AAD =
Criterion 1 was similar to USACE jurisdictional wetland hydrologic criterion (USACE, 1987). Criteria 2 through 4 were used to enhance this simple evaluation. Criterion 2 included two additional metrics, average daily deviation (ADD) and average absolute deviation (AAD), to further evaluate the temporal variation between the restored and reference wetland.
n
ADD =
i
(WTD1 − WTD2 )
n i=1
n |(WTD1 − WTD2 )| n
(1)
(2)
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
where ADD is the average daily deviation of the water table; AAD is the average absolute deviation of the water table; WTD1 is the daily water table depth at site 1 (Restoration); WTD2 is the daily water table depth at site 2 (Reference wetland); n is the number of days of interest. SEW30 is a tool most often used for the hydrologic evaluation of artificially drained sites to determine stress on crops due to high water tables (Evans et al., 1991), but has also been used to evaluate hydrology in wetland sites (Skaggs et al., 1994). The SEW30 is a robust evaluation because it describes both the duration and position of the water table within 30 cm of the soil surface in the wetland restoration and the reference wetland. The 30 cm threshold depth was selected to correspond with the USACE jurisdictional wetland criterion. SEW30 was calculated with the following formula: SEW30 =
n
(30 − Xi )
(3)
i=1
where SEW30 is the sum of excess water above 30 cm (cm days); Xi is the daily average water table depth below the soil surface (cm); n is the number of days of interest (2006–2008 calendar years). A value was computed each day the average water table was within the threshold depth of 30 cm of the soil surface. That daily water table depth was subtracted from 30 cm and multiplied by 1 day, resulting in units of cm days. This hydrologic analysis primarily focused on the entire year since the goal of restoration should be to achieve near-reference conditions across all seasons. Previous studies (Tweedy and Evans, 2001; Wright et al., 2006) also considered hydrologic criteria which temporally encompassed both the full year and the growing season. The hydrologic criteria were applied to six different data sets. Three represented the restoration treatments (PLUG, ROUGH, and CR). Water table conditions at each reference wetland well location
39
were evaluated to determine the areal extent that met minimum jurisdictional wetland hydrology as defined by USACE. Four wells that were located in a higher landscape position (indicated as REF 1, 3, 4, and 7 in Fig. 3) did not meeting the minimum criteria during the three year study period and were excluded from analysis. The remaining nine water table monitoring locations were split into three datasets comprised of three wells each that represented the range of wetness found across the reference wetland area (Fig. 3): Ref-Wet (reference wells that exhibited the wettest conditions), Ref-Med (reference wells that exhibited median conditions), and Ref-Dry (reference wells that exhibited the driest conditions). Additionally, outflow volumes were used to evaluate the ability of each restoration treatment to store water. Reducing outflow volumes was critical for diminishing fluxes of freshwater and pollutant loads to downstream water bodies. In this study, lower outflow indicated an enhancement to the overall hydrologic regulation services provided by a restoration treatment. Data were analyzed using SAS statistical software (SAS Institute Inc., 1985). An ANOVA means test with the Tukey adjustment was applied to the data. The Tukey method was applied to control the Type-I experiment-wise error rate since multiple pair-wise comparisons were made. Comparisons were made at a 95% confidence level. 3. Results and discussion 3.1. Precipitation Annual rainfall totals were highly variable. The site experienced 120 cm in 2006, 154 cm in 2007, and 150 cm in 2008. Average onsite precipitation during the study (141 cm) was slightly lower than the long-term average for nearby Morehead City, North Carolina (148 cm). Wetland hydrologic response to wet, near-normal, and extreme dry periods were captured during this study, which was
Fig. 4. Comparison of the as-built topography of Block 1 (above) and Block 3 (below). In Block 3, note the substantial surface water in the PLUG treatment and higher ground elevation in the CR treatment.
40
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
ideal for evaluating hydrologic response, and supported the reasoning that multiple years of wetland monitoring should be employed when determining success criteria or best design/implementation practices. 3.2. Determination of suitable locations in the restoration for water table analysis During an earlier study that was conducted during the first two years following construction (Wright, 2005), it was observed that the Block 3 PLUG treatment was much wetter than the Block 3 CR treatment, which may have masked the treatment effects of the experiment. Subsequent analysis by Wright et al. (2006) focused exclusively on Blocks 1 and 2, and suggested that landscape or elevation variations may explain the Block 3 hydrologic observations. However, this hypothesis was not verified in the study. An as-built topographic survey performed at the beginning of this study along the monitoring transects in each block verified local elevation was the significant factor that produced the unexpected hydrologic conditions in Block 3. The PLUG treatment contained lower areas while the CR treatment contained higher areas that were not identified during construction (Fig. 4). This produced much wetter conditions in PLUG and drier conditions in CR treatments in Block 3 than intended. The survey also verified that elevations in Blocks 1 and 2 were comparable. As such, analysis of restoration treatments effects in this study focused solely on Blocks 1 and 2 (wells NR1–NR12). While this weakened the statistical power to evaluate treatments effects, eliminating Block 3 (wells NR13–NR18) from analysis produced the most meaningful results. 3.3. Groundwater hydrology observations in the restoration and the reference wetland The overall pattern of water table response in the reference and restoration sites was typical of many eastern NC wetlands. Water table elevations were generally highest during the winter or following tropical events, and lowest during the summer months. Fig. 5 shows the mean water table profiles of the restoration and reference to demonstrate the mean hydroperiod for each location. From an observational standpoint, the average water table response within the restoration matched the reference closely during the study. Average water table conditions in the restoration
easily met minimum USACE jurisdictional wetland criteria – the water table was within 30 cm of the surface for a continuous period of more than 5% of the growing season (12 days in Carteret County, NC). From a USACE regulatory point of view, the restoration design clearly produced minimum wetland hydrology. Over the course of the study (2006–2008), the restoration and reference sites had an ADD of −3.8 cm, demonstrating that the average daily water table depth within the reference site was only 3.8 cm lower than in the restoration sites. The two sites also had an AAD of 10.8 cm meaning the average daily water table depth between the two sites only varied by that amount in either direction (higher or lower). Considering that the reference and restoration experienced an annual water table depth range of over 100 cm, the hydroperiod at both sites were very similar.
3.4. Evaluation of restoration treatment effects using hydrologic criteria Criterion 1, the longest continuous period when the water table was within 30 cm of the surface during the growing season, was used as an initial comparison since it was analogous to the minimum jurisdictional criterion used for wetland hydrology regulation and permitting in the United States (USACE, 1987). As shown in Fig. 6, all treatments produced hydrologic conditions that met the 5% threshold for jurisdictional hydrology. In the restored wetland, the water table in the CR treatment was consistently within 30 cm of the surface for the longest continuous period, while the continuous duration in the PLUG treatment was the shortest. Results of this comparison seemed to support our hypothesis that the crown removal process produced wetter conditions. The reference wetland exhibited a much wider range of hydrologic conditions; the wettest locations (Ref-Wet) experienced a much longer duration of high water table, while the driest location (Ref-Dry) only exceeded the 5% threshold in 2007 when rainfall exceeded the long-term average. The remaining criteria were all applied across the entire threeyear study period (Table 1). Comparisons using both Criterion 2 (average water table depth) and Criterion 3 (SEW30 ) yielded results similar to those observed using Criterion 1. For both of these criteria, no significant difference was found among observed water table conditions of PLUG, ROUGH, and Ref-Med (P > 0.05) which suggested that, based on Criterion 2 and 3, the ROUGH and PLUG treatments produced conditions most similar to the
Fig. 5. Mean daily water table profiles for the restoration site (12 continuous monitoring locations) and in the reference site (9 continuous monitoring locations). Mean soil surface is at 0 cm, and the 30 cm depth represents the threshold depth for USACE jurisdictional wetland hydrology.
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
41
Fig. 6. Criterion 1 – longest continuous period of water table within 30 cm of the surface during the growing season. Note: Horizontal dotted line indicates the 5% growing season threshold (12 days) for this location.
Table 1 Criteria 2 through 4 evaluated for the entire study period (2006–2008) for the full calendar year. Values followed by same value vertically are not significantly different (˛ = 0.05). Dataset
Criterion 2 (cm)
Criterion 3 (cm days)
PLUG ROUGH CR Ref-Dry Ref-Med Ref-Wet
42 (b) 42 (b) 34 (c) 51 (a) 44 (b) 35 (c)
7944 (c) 8865 (c) 13,777 (a) 2873 (d) 8750 (c) 10,562 (b)
Criterion 4 (days) 13 (b) 15 (b) 107 (a) 0 (b) 3 (b) 5 (b)
Note: Criterion 2, average water table depth for 2006–2008; Criterion 3, total SEW30 (sum of excess water above 30 cm of the soil surface) for 2006–2008; Criterion 4, total days of surface inundation for 2006–2008.
median hydrologic conditions observed within the reference wetland. Additionally these criteria suggested that the CR treatment and Ref-Wet were both significantly wetter than other areas of the restoration and reference wetlands, while Ref-Dry was significantly drier. Criterion 3, SEW30 , was considered to be a more robust metric for evaluating wetland hydroperiod since it captured both the
duration and degree for which the water table was higher than the threshold depth of 30 cm. The evaluation using this criterion generally ranked the treatment and reference areas in the same order of wetness as with Criterion 2. However, the SEW30 evaluation suggested that CR was significantly wetter than Ref-Wet, which was not demonstrated by the previous criteria. The final criterion (Criterion 4) used for the hydrologic evaluation was the number of days of surface inundation (water table at or above ground surface) during the study period. Unlike the previous criteria, the CR treatment was not only significantly wetter than other restoration and reference wetland areas, but wetter by almost an order of magnitude. The CR treatment experienced 105 days of inundation while the next wettest area, ROUGH only experienced 15 days of inundation. Only Ref-Dry failed to experience inundated surface conditions. Given the extreme difference in surface inundation for CR compared to all of the reference wetland areas, it clearly did not effectively mimic reference conditions for this criterion. However, this criterion did demonstrate that the CR treatment was more effective at surface water storage than predicted. For all of the hydrologic criteria, Ref-Dry was found to exhibit the driest conditions, which was not surprising since its corresponding
Fig. 7. Cumulative outflow from the restoration treatment plots during 2006–2008. Values with different letters are significantly different (˛ = 0.05).
42
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
wells were located in higher elevation areas. While neither PLUG nor ROUGH experienced water table conditions significantly different than Ref-Med, the ROUGH treatment appeared to be a closer match to median reference conditions in terms of actual numerical difference. 3.5. Effect of restoration treatments on surface outflow Wetlands such as these whose main inputs are precipitation will have surface outflow volumes that are inversely proportional to surface storage. Minimizing outflow was particularly important for this restoration, because this site discharged to an estuary important to the local shellfishing industry. During the study, only 25–32% of rainfall recorded at the site was exported from the restored wetlands as outflow. In comparison, long-term hydrologic simulations performed by Wright et al. (2006) showed that under the previous agricultural conditions, the site exported around 36% of annual rainfall volumes on average. Given that the restoration was still young, outflow from the site should gradually decrease as it reaches maturity. The simulations by Wright et al. (2006) showed that outflow volumes from this site when mature would be reduced by almost 30% compared to previous agricultural conditions. It was hypothesized that the PLUG treatment would have the greatest outflow, since the intact field crown and smoother surface conditions would provide less surface storage. The ROUGH and CR treatments were expected to both export significantly less outflow than the PLUG treatment. Given that the CR treatment produced the wettest soil–water conditions, it was not surprising that it exported significantly less total outflow (104 cm) than the ROUGH (136 cm) and the PLUG treatments (129 cm). An unexpected result was the outflow from the ROUGH treatment was significantly greater than CR or PLUG (Fig. 7). A post-construction topographic survey of site showed that some surface water conveyance paths may have unintentionally been created while constructing the topographic highs and lows within the ROUGH treatments. Since the field crown was not removed, surface flow through these conveyances may have also been enhanced due to the higher lateral gradient. 4. Conclusions This research evaluated hydrologic conditions at a restored wetland in coastal North Carolina that was constructed by plugging drainage ditches, controlling outflow with water control structures, and using three surface construction techniques that created varying degrees of microtopography and surface storage. The results of field monitoring and analysis showed that the CR treatment areas produced significantly wetter soil conditions and exported a significantly lower volume of water than the ROUGH and PLUG treatments areas. However, CR appeared to produce a hydroperiod that was wetter than the reference wetland for 3 out of 4 hydrologic evaluation criteria. As expected, hydrologic conditions within the ROUGH treatment areas closely matched median hydrologic conditions observed within the reference. For the majority of evaluation criteria, there was not a significant difference between hydrologic conditions observed in the ROUGH and PLUG treatments – though ROUGH did appear to be slightly wetter. Surprisingly, the ROUGH treatment exported the greatest volume of water. The roughening process, which was intended to create surface storage and reduce runoff, may have unintentionally created interconnected surface flow conveyances that actually increased surface drainage. While this research did not find clear evidence to support the application of any particular surface treatment, mostly because this was already a sufficiently wet site, surface roughening still presents
many wetland restoration benefits for a nominal expense. Plugging pre-existing field ditches may be adequate to restore jurisdictional wetland hydrology and match reference hydrologic conditions. However, this minimalistic approach does introduce a greater risk of restoration failure at sites that possessed borderline wetland characteristics prior to agricultural conversion. For most sites, surface roughening should be included in the wetland restoration design. Besides improved habitat diversity and enhanced biogeochemical conditions found in natural wetlands, the increased surface storage should also help ensure that a restoration site meets jurisdictional criteria for minimum wetland hydrology even in years with below normal rainfall. Care should be taken during construction to avoid the creation of unwanted surface water conveyances as was observed during this study. Crown removal will add significant costs to a wetland restoration project since it requires topsoil to be removed, stockpiled, and then reapplied to the field following crown excavation. The expense may be justified for sites that historically exhibited extremely marginal wetland hydrology, but proper site selection can help reduce the need for this extensive earthwork.
Acknowledgements The authors wish to thank the NC Department of Environment and Natural Resources – Ecosystem Enhancement Program for supporting this research. The North Carolina Coastal Federation was instrumental in supporting our efforts at their North River Farms restoration site. Special thanks to the NC State University Bio&Ag Engineering staff, graduate, and undergraduate students that supported well installation, site maintenance, and field data collection during the project.
References Ahn, C., Dee, S., 2011. Early development of plant community in a created mitigation wetland as affected by introduced hydrologic design elements. Ecol. Eng. 37 (2011), 1324–1333. Alsfeld, A.J., Bowman, J.L., Jacobs, A.D., 2009. Effects of woody debris microtopography, and organic matter amendments on the biotic community of constructed depressional wetlands. Biol. Conserv. 142, 247–255. Cashin, G.E., Dorney, J.R., Richardson, C.J., 1992. Wetland alteration trends on the North Carolina coastal plain. Wetlands 12 (2), 63–71. Courtwright, J., Findley, S.E.G., 2011. Effects of microtopography on hydrology physiochemistry and vegetation in a tidal swamp of the Hudson River. Wetlands 31, 239–249. Evans, R.O., Skaggs, R.W., Sneed, R.E., 1991. Stress day index models to predict corn and soybean relative yield under high water table conditions. Trans. ASAE 34 (5), 1997–2005. Goodwin, R.A., 1984. Soil Survey of Carteret County. United States Department of Agriculture, Soil Conservation Service, North Carolina. Holman, R.E., Childres, W.S., 1995. Wetland restoration and creation: Development of a handbook covering six coastal wetland types. Report No. 289. North Carolina State University, Water Resources Research Institute, Raleigh, NC. Jarzemsky, R.D., 2009. Hydrologic Evaluation of a Restored Wetland in Eastern North Carolina. Biological and Agricultural Engineering, North Carolina State University, Raleigh, NC (M.S. Thesis). Lilly, J.P.,1981. The Blackland soils of North Carolina: Their characteristics and management for agriculture. Technical Bulletin No. 270. Raleigh, NC, North Carolina State University, North Carolina Agricultural Research Service. Moser, K.F., Ahn, C., Noe, G.B., 2009. The influence of microtopography on soil nutrients in created mitigation wetlands. Restorat. Ecol. 17 (5), 641–651. Natural Resource Conservation Service (NRCS), 2003. Wetland Restoration, Enhancement, and Management. United States Department of Agriculture (USDA), Washington, DC. SAS Institute Inc., 1985. SAS/STAT Guide for Personal Computers, Version 6 Edition. SAS Institute, Cary, NC. Schafale, M.P., Weakley, A.S., 1990. Classification of the Natural Communities of North Carolina North Carolina Natural Heritage Program. Department of Environment and Natural Resources, NC. Simmons, M.E., Wu, X.B., Whisenant, S.G., 2011. Plant and soil responses to created microtopography and soil treatments in bottomland hardwood forest restoration. Rest Ecol. 19 (1), 136–146.
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43 Simmons, M.E., Wu, X.B., Whisenant, S.G., 2007. Bottomland hardwood forest species response to flooding regimes along an urbanization gradient. Ecol. Eng. 29, 223–231. Skaggs, R.W., Amatya, D., Evans, R.O., Parsons, J.E., 1994. Characterization and evaluation of proposed hydrologic criteria for wetlands. J. Soil Water Conserv. 49 (5), 501–510. Skaggs, R.W., Gilliam, J.W., Sheets, T.J., Barnes, J.S., 1980. Effect of agricultural land development on drainage waters in the North Carolina Tidewater Region. Water Resources Research Institute of the University of North Carolina Rep. No. 159. Skaggs, R.W., Youssef, M.A., Chescheir, G.M., Gilliam, J.W., 2005. Effects of drainage intensity on nitrogen losses from drained lands. Trans. ASAE 48 (6), 2169– 2177. Smith, M.L., Meiman, P.J., Brummer, J.E., 2012. Characteristics of hummocked and non-hummocked Colorado riparian areas and wetlands. Wetlands Ecol. Manage. 20, 409–418.
43
Tweedy, K.L., Evans, R.O., 2001. Hydrologic characterization of two prior converted wetland restoration sites in Eastern North Carolina. Trans. ASAE 44 (5), 1135–1142. U.S. Army Corps of Engineers, 1987. Corps of Engineers Wetlands Delineation Manual. Technical Report Y-87-1. US Army Corps of Engineers Waterways Experiment Station – Environmental Laboratory, Vicksburg, MS. Wolf, K.L., Ahn, C., Noe, G.B., 2011. Microtopography enhances nitrogen cycling and removal in created mitigation wetlands. Ecol. Eng. 37 (9), 1398–1406. Wright, J.D., 2005. The evaluation and modeling of the effects of surface treatments on the hydrology of a restored wetland in the Coastal Plain of North Carolina. Biological and Agricultural Engineering, North Carolina State University, Raleigh, NC (M.S. Thesis). Wright, J.D., Burchell, M.R., Evans, R.O., Shelby, J.D., 2006. Hydrologic effect of three restoration techniques on a prior converted wetland in eastern NC. In: Hydrology and Management of Forested Wetlands: Proceedings of the International Conference, New Bern, NC, April 8–12.
Contents lists available at ScienceDirect
Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng
The impact of manipulating surface topography on the hydrologic restoration of a forested coastal wetland夽 Robert D. Jarzemsky a , Michael R. Burchell II b,∗ , Robert O. Evans b a b
U.S. Army Corps of Engineers, Chicago District, 111 N. Canal St. Suite 600, Chicago, IL 60606, United States Biological and Agricultural Engineering Campus, Box 7625, North Carolina State University, Raleigh, NC 27695, United States
a r t i c l e
i n f o
Article history: Received 31 August 2012 Received in revised form 20 April 2013 Accepted 8 June 2013 Available online 6 July 2013 Keywords: Wetland hydrology Restoration Microtopography
a b s t r a c t A wetland, converted to agriculture in the mid-1970s, was restored to re-establish a non-riverine wet hardwood forest community in eastern North Carolina. Three surface techniques were implemented during construction to determine their effect on successfully restoring target wetland hydrology. The surface treatments, replicated within a randomized complete block design, were: plugging field ditches without altering the land surface (PLUG), plugging the field ditches and roughening the surface (ROUGH), and plugging the field ditches and removing the field crown (CR). Hydrologic conditions for the restoration and a nearby reference site were evaluated based on three years of monitoring data. Daily water table depths between the restoration and reference were within 11 cm on average. An initial evaluation found inconsistencies of treatment effect between blocks, and an as-built survey later confirmed surface elevations within Block 3 deviated from the intended design and was excluded from further analysis. Water table and outflow conditions for the remaining treatment plots and the reference were evaluated using several hydrologic criteria. The CR treatment was found to produce the wettest surface conditions and exported the lowest volume of outflow. For the majority of criterion considered, CR also produced significantly wetter conditions than the reference. The PLUG and ROUGH treatments produced similar hydrologic conditions and tracked closely with the median hydrologic conditions in the reference. Based on the results of this study and several others in low lying coastal areas, plugging pre-existing field ditches may be adequate to restore jurisdictional wetland hydrology and match reference hydrologic conditions. However, surface roughening is low cost method to increase surface storage and introduce microtopographic diversity. For many areas, the removal of existing field crown may be cost-prohibitive and produce wetter than desired conditions. Crown removal should be reserved for sites which have borderline historic wetland hydrologic characteristics. © 2013 Elsevier B.V. All rights reserved.
1. Introduction The magnitude of wetland loss and alteration in the U.S. has resulted in measurable negative impacts on habitat and water quality. Therefore, efforts to strategically restore some of these lands back to their natural condition are vital to reclaim damaged or lost ecosystem functions. Approximately 50% of the land in eastern NC originally contained hydric soils, and over 50% of these native wetlands were altered to accommodate other land uses (Cashin et al., 1992). The majority of these wetland alterations occurred to enhance agricultural production, and were achieved
夽 Disclaimer: The views presented in this paper are those of the authors and do not necessarily represent the views of the US Army Corps of Engineers. ∗ Corresponding author. Tel.: +1 919 513 7372; fax: +1 919 515 6772. E-mail address: mike [email protected] (M.R. Burchell II). 0925-8574/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecoleng.2013.06.002
through land clearing, installation of subsurface drainage systems (comprised of open ditches or subsurface drains), and smoothing and crowning of soil surfaces to improve surface drainage (Lilly, 1981). This type of artificial drainage, while necessary for agricultural production in many parts of the country, has also resulted in environmental consequences. Studies have found that drainage from agriculture drainage increases peak runoff rates as much as 300–400% (Skaggs et al., 1980) and can lead to a 10-fold increase in nitrate (NO3 –N) losses (Skaggs et al., 2005). In addition to NO3 –N, agricultural drainage is also linked to increased exports of phosphorus and suspended solids (Skaggs et al., 1980). The goal of wetland restoration/creation is to successfully establish a system that exhibits the same structure and beneficial functions as a targeted wetland community in the most efficient manner possible. Inappropriate designs or inefficient implementation can result in restorations that are too expensive and fall short of achieving target ecosystem services. In a survey of practitioners,
36
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
Holman and Childres (1995) estimated that 40–50% of wetland restoration problems were simply due to poor site selection. Achieving appropriate hydroperiod is a major component in successful wetland restoration. A wetter site is not always better, since many natural wetlands alternate seasonally between dry and wet conditions. This is a major reason that at least one and preferably multiple reference wetlands should be identified prior to restoration to determine the target hydroperiod. To restore wetland hydrology on lands drained for agricultural practices, ditches that were installed must be plugged to reduce drainage from the site. However, Tweedy and Evans (2001) showed that simply plugging ditches may produce conditions that are drier than the targeted wetland community. This is especially true on sites that, prior to artificial drainage, may have only marginally satisfied the U.S. Army Corps of Engineers (USACE) jurisdictional hydrologic criterion (i.e. the water table within 30 cm of the surface continuously for more than 5% of growing season (USACE, 1987)). Removing the field crown and adding microtopography (defined here as slight ridges and depressions along the soil surface) in addition to ditch plugging may improve the chances of meeting wetland hydrology requirements, but it does increase costs. Currently, the U.S. Department of Agriculture – Natural Resources Conservation Service recommends implementing varying levels of microtopography to promote not only hydrologic but habitat diversity (NRCS, 2003). Several studies support this restoration technique because it can increase stormwater storage, prolong surface moisture during dry periods, provide more diverse habitat and vegetation, improve microbially mediated nutrient cycling and removal, and promote more overall ecosystem heterogeneity common in natural wetlands (Smith et al., 2012; Ahn and Dee, 2011; Courtwright and Findley, 2011; Simmons et al., 2007, 2011; Wolf et al., 2011; Alsfeld et al., 2009; Moser et al., 2009). Despite these existing studies, there is a lack of data that strongly supports a particular treatment practice for maximizing appropriate wetland hydrologic characteristics across entire restoration sites. This paper describes how various types of soil surface alterations affected wetland hydrology at a 100 ha forested wetland restoration in eastern North Carolina. In an effort to build on earlier results from research conducted by Wright et al. (2006), field monitoring at the restoration site was intensified to ultimately develop improved recommendations and guidance for future coastal wetland restorations. Goals of this study were to determine (1) differences in
hydrologic response of various soil–surface manipulation techniques, (2) how much surface manipulation was required to produce hydrologic conditions within a range of reference hydrologic conditions, and (3) what degree of surface manipulation would be considered optimum for producing acceptable hydrologic conditions at minimal restoration costs. It was hypothesized that roughening the surface and creating microtopographic features similar to those found in natural wetlands would be ideal for achieving hydrologic conditions found in a reference wetland. It was further hypothesized that simply plugging the ditches would produce the driest conditions because the intact field crown and smooth soil surface would result in more rapid surface drainage, while field crown removal would produce the wettest conditions due to the decreased surface flow gradient. Several hydrologic indicators and metrics were used to provide an extensive evaluation of the hydrology at the restoration site because of annual and seasonal climactic variations. A secondary goal was to demonstrate how these metrics might be applied to hydrologic assessments for future wetland restorations. 2. Materials and methods Field-scale studies were conducted at a coastal wetland restoration site located in Carteret County, North Carolina (Fig. 1). The 2400 ha site was artificially drained for agricultural row crop production in the 1970s before being acquired by the NC Coastal Federation for the purpose of restoration in 2002. Ditches were typically spaced at 100 m intervals and dug to 1 m depth, while fields were crowned by about 20 cm. Drainage water from the site discharged south into the North River estuary, an area important to the local shellfishing industry. The overall goals of the project were to restore hydrologic function, increase wetland habitat, and improve downstream water quality. The first of multiple phases of the project included hydrologic restoration of 100 ha of cropland to a non-riverine wet hardwood wetland ecosystem. This ecosystem is described as hardwood forest located on broad inter-stream flats with poorly drained mineral soils, and are common in the lower coastal plain of NC (Schafale and Weakley, 1990). Completed in 2003, Phase I included placement of earthen plugs in field ditches and installation flashboard risers to control water table levels and outflow. Wetland features such as 0.1 ha open water areas and simulated tree falls (about
Fig. 1. Vicinity map of the restoration site in Carteret County, NC.
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
12 per ha) were added across the site to increase topographic diversity. Following earthmoving, the site was planted with over 89,000 native wetland trees such as oaks (Quercus spp.), water tupelo (Nyssa aquatic), black gum (Nyssa sylvatica), bald cypress (Taxodium disticum), atlantic white cedar (Chamaecyparis thyoides), green ash (Fraxinus pennsylvanica), and longleaf pines (Pinus palustris). In addition to plantings, early successional herbaceous and woody vegetation quickly colonized the site. The soils were poorly to very poorly drained and were predominantly Deloss fine sandy loam with pockets of Wasda muck and Leon sand (Goodwin, 1984). Three levels of hydrologic restoration and surface microtopography were implemented on 6.5 ha test plots prior to planting. Each restoration technique was replicated three times in a randomized block design. The construction techniques included: 1. PLUG – hydrologic restoration using a series of earthen ditch plugs and water control structures. The land surface that existed under agricultural production (smoothed and crowned) remained intact, which minimized surface storage and maximized surface runoff toward the ditches. 2. ROUGH – in addition to practices described in PLUG, the field surface was roughed to increase microtopography. Tillage with a disk harrow was applied evenly across 25% of the treatment, creating surface mounds and hollows of ±15 cm, to increase surface storage. 3. CR – in addition to the practices described in PLUG, the field crown was removed and topsoil replaced. Surface storage was not intentionally enhanced, but removal of the crown was intended to lower the surface runoff gradient. Crown removal also resulted in a surface elevation that was closer in proximity to the groundwater table. Each 6.5 ha plot was separated by low earthen berms, constructed at the center point between plugged agricultural ditches to establish a barrier to surface flow between adjacent plots. Outflow
37
from the ditches was controlled by flashboard riser water control structures. Flashboards were installed to average ground surface elevation in each plot, and included a 30◦ V-notch that allowed excess soil–water to slowly drain down to 30 cm below average ground surface. Water table monitoring wells were installed in a single transect across the treatments (Fig. 2). Continuous water table recorders (Infinities USA Inc., Daytona Beach, FL) were installed in 10.2 cm diameter PVC monitoring wells at the mid-point between the plugged ditch and the berm, a location that corresponded to the average elevation of a typical crowned agricultural field. Water table depth was measured and logged hourly by the continuous water table recorders, and data were offloaded monthly. Two additional 5.1 cm diameter PVC wells were installed on either side of the continuous recorders, midway between the plugged ditch on one side and the berm on the other. Water table elevations in these wells, which were installed to provide a more spatial water table profile across the site, were measured manually on a monthly basis and correlated to the automatic measurements. A detailed survey using a Topcon Electronic Total Station was conducted in a swath across that site that included the transects to determine the as-built topography of treatments and the elevations of all the water table monitoring wells. Outflow from each plot was measured using potentiometers connected to a pulley-float system to measure stage over the 30◦ V-notch weir. Measurements were stored in a datalogger (SGT Engineering, Champaign, IL), and appropriate weir equations were applied to the stage readings to calculate drainage flowrate and volume as described in Jarzemsky (2009). Precipitation was recorded in two locations using a tipping bucket rain gauge and HOBO event datalogger (Onset Computer Corp., Bourne, MA). A nearby non-riverine wet hardwood wetland was also instrumented and served as a reference for comparison to the restoration treatments. The reference had a concave topography along an east-west axis (i.e. lowest elevation in the center) that resulted
Fig. 2. Monitoring schematic for Blocks 1 and 2 of the restoration site. PLUG, ROUGH, and CR represent wetland surface treatments. Wells designated NR were continuous water table monitoring, while locations designated by a and b represented the manual water table wells.
38
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
Fig. 3. Hydrologic monitoring schematic for the reference wetland site, which drained to a downstream tidal marsh. All wells had continuous water table measurements. Note: Wells followed by asterisk did not meet minimum USACE jurisdictional criteria and were excluded from analysis. Ref-Wet wells included REF 2, 5, and 8; Ref-Med wells included REF 13, 14, and 16; Ref-Dry wells included REF 6, 9, and 11.
in hydrologic variations across the site. To capture the range of wetness and extent of wetland conditions, the reference was instrumented with 13 continuous water table monitoring wells along three east-west transects (Fig. 3). Water table depths were monitored and recorded using the same instrumentation described earlier for the restoration areas. A detailed topographic survey was also conducted along each of the transects in the reference wetland to relate water table measurements to depth below ground surface. Outflow from the reference wetland was not monitored due to backwater conditions from tidal influences of the marsh located several hundred meters downstream. Water table data for 2006–2008 from the restoration treatments and the reference wetland were evaluated using four hydrologic criteria:
2. Average water table depth (full calendar year). 3. Sum of excess water above 30 cm, SEW30 (full calendar year). 4. Number of days of surface inundation (full calendar year).
1. Longest continuous period of water table within 30 cm of the surface (growing season).
AAD =
Criterion 1 was similar to USACE jurisdictional wetland hydrologic criterion (USACE, 1987). Criteria 2 through 4 were used to enhance this simple evaluation. Criterion 2 included two additional metrics, average daily deviation (ADD) and average absolute deviation (AAD), to further evaluate the temporal variation between the restored and reference wetland.
n
ADD =
i
(WTD1 − WTD2 )
n i=1
n |(WTD1 − WTD2 )| n
(1)
(2)
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
where ADD is the average daily deviation of the water table; AAD is the average absolute deviation of the water table; WTD1 is the daily water table depth at site 1 (Restoration); WTD2 is the daily water table depth at site 2 (Reference wetland); n is the number of days of interest. SEW30 is a tool most often used for the hydrologic evaluation of artificially drained sites to determine stress on crops due to high water tables (Evans et al., 1991), but has also been used to evaluate hydrology in wetland sites (Skaggs et al., 1994). The SEW30 is a robust evaluation because it describes both the duration and position of the water table within 30 cm of the soil surface in the wetland restoration and the reference wetland. The 30 cm threshold depth was selected to correspond with the USACE jurisdictional wetland criterion. SEW30 was calculated with the following formula: SEW30 =
n
(30 − Xi )
(3)
i=1
where SEW30 is the sum of excess water above 30 cm (cm days); Xi is the daily average water table depth below the soil surface (cm); n is the number of days of interest (2006–2008 calendar years). A value was computed each day the average water table was within the threshold depth of 30 cm of the soil surface. That daily water table depth was subtracted from 30 cm and multiplied by 1 day, resulting in units of cm days. This hydrologic analysis primarily focused on the entire year since the goal of restoration should be to achieve near-reference conditions across all seasons. Previous studies (Tweedy and Evans, 2001; Wright et al., 2006) also considered hydrologic criteria which temporally encompassed both the full year and the growing season. The hydrologic criteria were applied to six different data sets. Three represented the restoration treatments (PLUG, ROUGH, and CR). Water table conditions at each reference wetland well location
39
were evaluated to determine the areal extent that met minimum jurisdictional wetland hydrology as defined by USACE. Four wells that were located in a higher landscape position (indicated as REF 1, 3, 4, and 7 in Fig. 3) did not meeting the minimum criteria during the three year study period and were excluded from analysis. The remaining nine water table monitoring locations were split into three datasets comprised of three wells each that represented the range of wetness found across the reference wetland area (Fig. 3): Ref-Wet (reference wells that exhibited the wettest conditions), Ref-Med (reference wells that exhibited median conditions), and Ref-Dry (reference wells that exhibited the driest conditions). Additionally, outflow volumes were used to evaluate the ability of each restoration treatment to store water. Reducing outflow volumes was critical for diminishing fluxes of freshwater and pollutant loads to downstream water bodies. In this study, lower outflow indicated an enhancement to the overall hydrologic regulation services provided by a restoration treatment. Data were analyzed using SAS statistical software (SAS Institute Inc., 1985). An ANOVA means test with the Tukey adjustment was applied to the data. The Tukey method was applied to control the Type-I experiment-wise error rate since multiple pair-wise comparisons were made. Comparisons were made at a 95% confidence level. 3. Results and discussion 3.1. Precipitation Annual rainfall totals were highly variable. The site experienced 120 cm in 2006, 154 cm in 2007, and 150 cm in 2008. Average onsite precipitation during the study (141 cm) was slightly lower than the long-term average for nearby Morehead City, North Carolina (148 cm). Wetland hydrologic response to wet, near-normal, and extreme dry periods were captured during this study, which was
Fig. 4. Comparison of the as-built topography of Block 1 (above) and Block 3 (below). In Block 3, note the substantial surface water in the PLUG treatment and higher ground elevation in the CR treatment.
40
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
ideal for evaluating hydrologic response, and supported the reasoning that multiple years of wetland monitoring should be employed when determining success criteria or best design/implementation practices. 3.2. Determination of suitable locations in the restoration for water table analysis During an earlier study that was conducted during the first two years following construction (Wright, 2005), it was observed that the Block 3 PLUG treatment was much wetter than the Block 3 CR treatment, which may have masked the treatment effects of the experiment. Subsequent analysis by Wright et al. (2006) focused exclusively on Blocks 1 and 2, and suggested that landscape or elevation variations may explain the Block 3 hydrologic observations. However, this hypothesis was not verified in the study. An as-built topographic survey performed at the beginning of this study along the monitoring transects in each block verified local elevation was the significant factor that produced the unexpected hydrologic conditions in Block 3. The PLUG treatment contained lower areas while the CR treatment contained higher areas that were not identified during construction (Fig. 4). This produced much wetter conditions in PLUG and drier conditions in CR treatments in Block 3 than intended. The survey also verified that elevations in Blocks 1 and 2 were comparable. As such, analysis of restoration treatments effects in this study focused solely on Blocks 1 and 2 (wells NR1–NR12). While this weakened the statistical power to evaluate treatments effects, eliminating Block 3 (wells NR13–NR18) from analysis produced the most meaningful results. 3.3. Groundwater hydrology observations in the restoration and the reference wetland The overall pattern of water table response in the reference and restoration sites was typical of many eastern NC wetlands. Water table elevations were generally highest during the winter or following tropical events, and lowest during the summer months. Fig. 5 shows the mean water table profiles of the restoration and reference to demonstrate the mean hydroperiod for each location. From an observational standpoint, the average water table response within the restoration matched the reference closely during the study. Average water table conditions in the restoration
easily met minimum USACE jurisdictional wetland criteria – the water table was within 30 cm of the surface for a continuous period of more than 5% of the growing season (12 days in Carteret County, NC). From a USACE regulatory point of view, the restoration design clearly produced minimum wetland hydrology. Over the course of the study (2006–2008), the restoration and reference sites had an ADD of −3.8 cm, demonstrating that the average daily water table depth within the reference site was only 3.8 cm lower than in the restoration sites. The two sites also had an AAD of 10.8 cm meaning the average daily water table depth between the two sites only varied by that amount in either direction (higher or lower). Considering that the reference and restoration experienced an annual water table depth range of over 100 cm, the hydroperiod at both sites were very similar.
3.4. Evaluation of restoration treatment effects using hydrologic criteria Criterion 1, the longest continuous period when the water table was within 30 cm of the surface during the growing season, was used as an initial comparison since it was analogous to the minimum jurisdictional criterion used for wetland hydrology regulation and permitting in the United States (USACE, 1987). As shown in Fig. 6, all treatments produced hydrologic conditions that met the 5% threshold for jurisdictional hydrology. In the restored wetland, the water table in the CR treatment was consistently within 30 cm of the surface for the longest continuous period, while the continuous duration in the PLUG treatment was the shortest. Results of this comparison seemed to support our hypothesis that the crown removal process produced wetter conditions. The reference wetland exhibited a much wider range of hydrologic conditions; the wettest locations (Ref-Wet) experienced a much longer duration of high water table, while the driest location (Ref-Dry) only exceeded the 5% threshold in 2007 when rainfall exceeded the long-term average. The remaining criteria were all applied across the entire threeyear study period (Table 1). Comparisons using both Criterion 2 (average water table depth) and Criterion 3 (SEW30 ) yielded results similar to those observed using Criterion 1. For both of these criteria, no significant difference was found among observed water table conditions of PLUG, ROUGH, and Ref-Med (P > 0.05) which suggested that, based on Criterion 2 and 3, the ROUGH and PLUG treatments produced conditions most similar to the
Fig. 5. Mean daily water table profiles for the restoration site (12 continuous monitoring locations) and in the reference site (9 continuous monitoring locations). Mean soil surface is at 0 cm, and the 30 cm depth represents the threshold depth for USACE jurisdictional wetland hydrology.
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
41
Fig. 6. Criterion 1 – longest continuous period of water table within 30 cm of the surface during the growing season. Note: Horizontal dotted line indicates the 5% growing season threshold (12 days) for this location.
Table 1 Criteria 2 through 4 evaluated for the entire study period (2006–2008) for the full calendar year. Values followed by same value vertically are not significantly different (˛ = 0.05). Dataset
Criterion 2 (cm)
Criterion 3 (cm days)
PLUG ROUGH CR Ref-Dry Ref-Med Ref-Wet
42 (b) 42 (b) 34 (c) 51 (a) 44 (b) 35 (c)
7944 (c) 8865 (c) 13,777 (a) 2873 (d) 8750 (c) 10,562 (b)
Criterion 4 (days) 13 (b) 15 (b) 107 (a) 0 (b) 3 (b) 5 (b)
Note: Criterion 2, average water table depth for 2006–2008; Criterion 3, total SEW30 (sum of excess water above 30 cm of the soil surface) for 2006–2008; Criterion 4, total days of surface inundation for 2006–2008.
median hydrologic conditions observed within the reference wetland. Additionally these criteria suggested that the CR treatment and Ref-Wet were both significantly wetter than other areas of the restoration and reference wetlands, while Ref-Dry was significantly drier. Criterion 3, SEW30 , was considered to be a more robust metric for evaluating wetland hydroperiod since it captured both the
duration and degree for which the water table was higher than the threshold depth of 30 cm. The evaluation using this criterion generally ranked the treatment and reference areas in the same order of wetness as with Criterion 2. However, the SEW30 evaluation suggested that CR was significantly wetter than Ref-Wet, which was not demonstrated by the previous criteria. The final criterion (Criterion 4) used for the hydrologic evaluation was the number of days of surface inundation (water table at or above ground surface) during the study period. Unlike the previous criteria, the CR treatment was not only significantly wetter than other restoration and reference wetland areas, but wetter by almost an order of magnitude. The CR treatment experienced 105 days of inundation while the next wettest area, ROUGH only experienced 15 days of inundation. Only Ref-Dry failed to experience inundated surface conditions. Given the extreme difference in surface inundation for CR compared to all of the reference wetland areas, it clearly did not effectively mimic reference conditions for this criterion. However, this criterion did demonstrate that the CR treatment was more effective at surface water storage than predicted. For all of the hydrologic criteria, Ref-Dry was found to exhibit the driest conditions, which was not surprising since its corresponding
Fig. 7. Cumulative outflow from the restoration treatment plots during 2006–2008. Values with different letters are significantly different (˛ = 0.05).
42
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43
wells were located in higher elevation areas. While neither PLUG nor ROUGH experienced water table conditions significantly different than Ref-Med, the ROUGH treatment appeared to be a closer match to median reference conditions in terms of actual numerical difference. 3.5. Effect of restoration treatments on surface outflow Wetlands such as these whose main inputs are precipitation will have surface outflow volumes that are inversely proportional to surface storage. Minimizing outflow was particularly important for this restoration, because this site discharged to an estuary important to the local shellfishing industry. During the study, only 25–32% of rainfall recorded at the site was exported from the restored wetlands as outflow. In comparison, long-term hydrologic simulations performed by Wright et al. (2006) showed that under the previous agricultural conditions, the site exported around 36% of annual rainfall volumes on average. Given that the restoration was still young, outflow from the site should gradually decrease as it reaches maturity. The simulations by Wright et al. (2006) showed that outflow volumes from this site when mature would be reduced by almost 30% compared to previous agricultural conditions. It was hypothesized that the PLUG treatment would have the greatest outflow, since the intact field crown and smoother surface conditions would provide less surface storage. The ROUGH and CR treatments were expected to both export significantly less outflow than the PLUG treatment. Given that the CR treatment produced the wettest soil–water conditions, it was not surprising that it exported significantly less total outflow (104 cm) than the ROUGH (136 cm) and the PLUG treatments (129 cm). An unexpected result was the outflow from the ROUGH treatment was significantly greater than CR or PLUG (Fig. 7). A post-construction topographic survey of site showed that some surface water conveyance paths may have unintentionally been created while constructing the topographic highs and lows within the ROUGH treatments. Since the field crown was not removed, surface flow through these conveyances may have also been enhanced due to the higher lateral gradient. 4. Conclusions This research evaluated hydrologic conditions at a restored wetland in coastal North Carolina that was constructed by plugging drainage ditches, controlling outflow with water control structures, and using three surface construction techniques that created varying degrees of microtopography and surface storage. The results of field monitoring and analysis showed that the CR treatment areas produced significantly wetter soil conditions and exported a significantly lower volume of water than the ROUGH and PLUG treatments areas. However, CR appeared to produce a hydroperiod that was wetter than the reference wetland for 3 out of 4 hydrologic evaluation criteria. As expected, hydrologic conditions within the ROUGH treatment areas closely matched median hydrologic conditions observed within the reference. For the majority of evaluation criteria, there was not a significant difference between hydrologic conditions observed in the ROUGH and PLUG treatments – though ROUGH did appear to be slightly wetter. Surprisingly, the ROUGH treatment exported the greatest volume of water. The roughening process, which was intended to create surface storage and reduce runoff, may have unintentionally created interconnected surface flow conveyances that actually increased surface drainage. While this research did not find clear evidence to support the application of any particular surface treatment, mostly because this was already a sufficiently wet site, surface roughening still presents
many wetland restoration benefits for a nominal expense. Plugging pre-existing field ditches may be adequate to restore jurisdictional wetland hydrology and match reference hydrologic conditions. However, this minimalistic approach does introduce a greater risk of restoration failure at sites that possessed borderline wetland characteristics prior to agricultural conversion. For most sites, surface roughening should be included in the wetland restoration design. Besides improved habitat diversity and enhanced biogeochemical conditions found in natural wetlands, the increased surface storage should also help ensure that a restoration site meets jurisdictional criteria for minimum wetland hydrology even in years with below normal rainfall. Care should be taken during construction to avoid the creation of unwanted surface water conveyances as was observed during this study. Crown removal will add significant costs to a wetland restoration project since it requires topsoil to be removed, stockpiled, and then reapplied to the field following crown excavation. The expense may be justified for sites that historically exhibited extremely marginal wetland hydrology, but proper site selection can help reduce the need for this extensive earthwork.
Acknowledgements The authors wish to thank the NC Department of Environment and Natural Resources – Ecosystem Enhancement Program for supporting this research. The North Carolina Coastal Federation was instrumental in supporting our efforts at their North River Farms restoration site. Special thanks to the NC State University Bio&Ag Engineering staff, graduate, and undergraduate students that supported well installation, site maintenance, and field data collection during the project.
References Ahn, C., Dee, S., 2011. Early development of plant community in a created mitigation wetland as affected by introduced hydrologic design elements. Ecol. Eng. 37 (2011), 1324–1333. Alsfeld, A.J., Bowman, J.L., Jacobs, A.D., 2009. Effects of woody debris microtopography, and organic matter amendments on the biotic community of constructed depressional wetlands. Biol. Conserv. 142, 247–255. Cashin, G.E., Dorney, J.R., Richardson, C.J., 1992. Wetland alteration trends on the North Carolina coastal plain. Wetlands 12 (2), 63–71. Courtwright, J., Findley, S.E.G., 2011. Effects of microtopography on hydrology physiochemistry and vegetation in a tidal swamp of the Hudson River. Wetlands 31, 239–249. Evans, R.O., Skaggs, R.W., Sneed, R.E., 1991. Stress day index models to predict corn and soybean relative yield under high water table conditions. Trans. ASAE 34 (5), 1997–2005. Goodwin, R.A., 1984. Soil Survey of Carteret County. United States Department of Agriculture, Soil Conservation Service, North Carolina. Holman, R.E., Childres, W.S., 1995. Wetland restoration and creation: Development of a handbook covering six coastal wetland types. Report No. 289. North Carolina State University, Water Resources Research Institute, Raleigh, NC. Jarzemsky, R.D., 2009. Hydrologic Evaluation of a Restored Wetland in Eastern North Carolina. Biological and Agricultural Engineering, North Carolina State University, Raleigh, NC (M.S. Thesis). Lilly, J.P.,1981. The Blackland soils of North Carolina: Their characteristics and management for agriculture. Technical Bulletin No. 270. Raleigh, NC, North Carolina State University, North Carolina Agricultural Research Service. Moser, K.F., Ahn, C., Noe, G.B., 2009. The influence of microtopography on soil nutrients in created mitigation wetlands. Restorat. Ecol. 17 (5), 641–651. Natural Resource Conservation Service (NRCS), 2003. Wetland Restoration, Enhancement, and Management. United States Department of Agriculture (USDA), Washington, DC. SAS Institute Inc., 1985. SAS/STAT Guide for Personal Computers, Version 6 Edition. SAS Institute, Cary, NC. Schafale, M.P., Weakley, A.S., 1990. Classification of the Natural Communities of North Carolina North Carolina Natural Heritage Program. Department of Environment and Natural Resources, NC. Simmons, M.E., Wu, X.B., Whisenant, S.G., 2011. Plant and soil responses to created microtopography and soil treatments in bottomland hardwood forest restoration. Rest Ecol. 19 (1), 136–146.
R.D. Jarzemsky et al. / Ecological Engineering 58 (2013) 35–43 Simmons, M.E., Wu, X.B., Whisenant, S.G., 2007. Bottomland hardwood forest species response to flooding regimes along an urbanization gradient. Ecol. Eng. 29, 223–231. Skaggs, R.W., Amatya, D., Evans, R.O., Parsons, J.E., 1994. Characterization and evaluation of proposed hydrologic criteria for wetlands. J. Soil Water Conserv. 49 (5), 501–510. Skaggs, R.W., Gilliam, J.W., Sheets, T.J., Barnes, J.S., 1980. Effect of agricultural land development on drainage waters in the North Carolina Tidewater Region. Water Resources Research Institute of the University of North Carolina Rep. No. 159. Skaggs, R.W., Youssef, M.A., Chescheir, G.M., Gilliam, J.W., 2005. Effects of drainage intensity on nitrogen losses from drained lands. Trans. ASAE 48 (6), 2169– 2177. Smith, M.L., Meiman, P.J., Brummer, J.E., 2012. Characteristics of hummocked and non-hummocked Colorado riparian areas and wetlands. Wetlands Ecol. Manage. 20, 409–418.
43
Tweedy, K.L., Evans, R.O., 2001. Hydrologic characterization of two prior converted wetland restoration sites in Eastern North Carolina. Trans. ASAE 44 (5), 1135–1142. U.S. Army Corps of Engineers, 1987. Corps of Engineers Wetlands Delineation Manual. Technical Report Y-87-1. US Army Corps of Engineers Waterways Experiment Station – Environmental Laboratory, Vicksburg, MS. Wolf, K.L., Ahn, C., Noe, G.B., 2011. Microtopography enhances nitrogen cycling and removal in created mitigation wetlands. Ecol. Eng. 37 (9), 1398–1406. Wright, J.D., 2005. The evaluation and modeling of the effects of surface treatments on the hydrology of a restored wetland in the Coastal Plain of North Carolina. Biological and Agricultural Engineering, North Carolina State University, Raleigh, NC (M.S. Thesis). Wright, J.D., Burchell, M.R., Evans, R.O., Shelby, J.D., 2006. Hydrologic effect of three restoration techniques on a prior converted wetland in eastern NC. In: Hydrology and Management of Forested Wetlands: Proceedings of the International Conference, New Bern, NC, April 8–12.