After tillage operations were completed, tree saplings were planted on half of the treatment plots at each site. Trees were planted during March, 1995 at the Beaufort site and during February, 1995 at the Craven site. Yellow poplar (liriodendron tulipifera), swamp white oak (quercus bicolor), willow oak (quercus phellos), and cherrybark oak (quercus falcata var. pagodifolia) were planted at each site on the selected plots.
Flashboard riser type water control structures were installed at the Beaufort
and Craven sites during March, 1995 and April, 1995, respectively, in order
to restore wetland hydrology and provide a means for measuring outflow
from the experimental plots. Boards and weirs were installed during May,
1995 at both sites to begin the process of restoring wetland hydrology.
On two field ditches, the weirs were installed at an elevation approximately
15 centimeters above the average land surface along a transect across each
of the plots. Therefore, surface ponding to a depth of 15 centimeters would
be necessary in order for water to leave the site. For the remaining two
field ditches, weirs were installed at an elevation 15 centimeters below
the average land surface. Earthen berms were created around the outer perimeter
of the entire site to prevent ponded surface water from running off-site
before running into the field ditches. In addition, berms were constructed
between each experimental plot to prevent surface runoff from flowing between
treatment plots. This created eight experimental treatment plots at each
site. The dimensions of individual plots were approximately 34 by 300 m
at the Beaufort County site and 42 by 275 m at the Craven County site.
At each site, one conventionally drained hydrologic "control" plot was
established between a treatment plot and free drainage ditch. The control
plots provided a means of monitoring hydrologic conditions similar to the
conditions observed under agricultural drainage and prior to wetland restoration.
Combinations of the two water table management and two microtopography treatments made up four hydrologic treatments monitored at each site: high water table management and smooth topography, low water table management and smooth topography, high water table management and rough topography, and low water table management and rough topography. Two replicates of these treatments were established at each site, for a total of eight, approximately 0.9 ha, experimental treatment plots at each site.
Installation of field instruments began in the spring of 1995 and was completed during the summer of 1996. The following hydrologic data were collected at each site: water table depth at six locations within each experimental plot with one location being an automated stage recorder, level of water in field ditches, soil moisture conditions, and precipitation. All instruments, except the ditch stage recorders, were installed within two experimental transects delineated across each site approximately 40 meters from the site perimeters (see above figure). These 10 m wide transects were implemented to provide for measurements at the same location within each experimental plot. The ditch stage recorders were installed approximately 4 meters up stream from the control structures.
Four inch polyvinal chloride (PVC) water table observation wells
were installed (one per experimental plot) during December, 1995 and provided
a means of periodically recording water table depths within each plot.
Five, one and one quarter inch PVC wells were later installed across both
instrumentation transects to determine groundwater gradients across individual
treatment plots. All wells were perforated from the bottom to within 15
cm of the soil surface. Water table measurements from these observation
wells were recorded weekly using a steel tape. Water table recorders were
installed on the four inch PVC wells during the summer of 1996 in order
to provide a continuous record of water table depth. Continuous stage recorders
were installed at the outlets of the field ditches, to provide a continuous
record of water levels within the field ditches. Data from these recorders
were used to compute the amount of outflow from the experimental plots
by recording the depth of water flowing over the 60 degree, V-notch weirs
in the water control structures. The term "outflow", as used in this report,
is defined as the culmination of surface runoff and subsurface drainage
which flows into the field ditches and flows out the water control structures.
Soil moisture data was collected periodically at the restoration sites.
Three pronged, 15 cm long time domain reflectometry probes (TDR), similar
in design to those described by Zegelin et al. (1989) and Heimovaara (1993),
were installed across the instrumentation transect closest to the water
table control structures and measured on a weekly interval during the growing
season. On smooth microtopography plots, TDR probes were installed at two
locations (i.e. two replicates) within each experimental plot. At each
location, probes were installed at the following depth intervals: 0 to
15 cm, 15 to 30 cm, 30 to 45 cm, 45 to 60 cm, and 60 to 75. On rough microtopography
plots, probes were installed in such a way as to measure differences in
soil moisture on the high mounded areas, furrowed depressional areas, and
areas more representative of the soil surface elevation before soil roughing.
At two high mounded locations within each rough experimental plot, probes
were installed at depth intervals of 0 to 15 cm and 15 to 30 cm. The same
was done for two depressional locations. At two locations representative
of the original soil surface elevation, probes were installed at depth
intervals of 30 to 45 cm, 45 to 60 cm, and 60 to 75 cm. All soil moisture
field measurements were made with a Tektronix 1502C Metallic TDR Cable
Tester, and volumetric water content was calculated using the equation
given by Topp et al. (1980).
Rainfall
gauges were installed at both sites in the summer of 1996 and cumulative
rainfall measurements were recorded approximately every week to the nearest
0.01 inch. In order to provide more site specific rainfall data for the
remaining duration of the study, an automated tipping-bucket rainfall gauge
was installed at each site during the spring of 1997. For periods of record
before the installation of the rain gauges, rainfall data from nearby automated
weather stations were utilized. For the Beaufort County site, daily rainfall
data from an automated weather station located approximately eight km north
of the site were used. For the Craven County site, a weighted average rainfall
value, based on relative distance from the site, was calculated for each
day from automated weather stations located in the towns of Aurora, Greenville,
New Bern, Kinston, Trenton, and Washington, NC. For the period of time
after the installation of the manually read gauges and before the installation
of the automated rain gauges, the daily rainfall data from nearby weather
stations was used to scale the cumulative rainfall data collected from
the manually read gauges, in order to estimate daily rainfall values for
each site.
Soils
Field measurements were conducted during the spring of 1997 to determine hydraulic soil parameters for each site. Specific parameters evaluated included saturated horizontal hydraulic conductivity with depth, soil water characteristic, and surface storage capacity.
Hydraulic conductivity was determined in-situ using the auger-hole method
as described by Van Beers (1970) and Bouwer and Jackson (1974). Auger-hole
conductivity measurements were made at three depths; 45, 90, and 130 cm.
In general, one measurement was made for each of the dominant soil horizons
found during preliminary augering. Measurements were made at two locations
within each experimental plot and the control plots. At each location,
three holes were dug to the specified depths and the water level in the
holes was allowed to equilibrate with the water table before beginning
measurements. At the time of testing, depth to the water table was generally
less than 5 cm for the Beaufort site and less than 20 cm for the Craven
site. Hydraulic conductivity values were calculated from these field measurements
for each bore hole. Conductivity values for the two deepest soil layers
were then back calculated and an average conductivity value was assigned
to each layer.
Within four of the experimental plots at each site, a soil pit was dug
to collect 7.6 cm diameter undisturbed soil cores from each defined soil
layer. Average sample depths were 5, 40, 90, and 200 cm at the Beaufort
site, and 5, 45, 90, and 150 cm at the Craven site. Two cores were collected
at each depth and plot location. The cores were partially prepared in the
field, placed in cardboard storage containers, and sealed in plastic bags
to prevent the cores from drying before reaching the laboratory. In the
laboratory, the cores collected from the first two soil layers at each
site were prepared and allowed to slowly saturate from the bottom over
a period of several days before beginning testing. The cores were then
placed in pressure plate apparatii for determination of soil water characteristic
data using the procedure described by Klute (1965a). For soil water tensions
up to 400 cm of water, testing was conducted on the entire undisturbed
soil core as removed from the field. At each incremental increase in tension,
cores were allowed to stabilize for 24 hours before reading the volume
of water drained. For soil water tensions from 600 to 15000 cm of water,
high pressure plate apparatii were used with sections of the 7.6 cm diameter
approximately 1.5 cm thick, using the method described by Klute (1965a).
These "slices" were placed directly on the pressure plates, pressure was
applied, and the volume of water drained was recorded after 72 hours.
Volume drained versus soil water tension relationships were then determined.
Estimates of surface storage for each experimental treatment
plot were calculated from detailed topographical surveys conducted during
the summer of 1996. Six topographical transects were run across each plot;
three parallel to the field ditches and three perpendicular to the field
ditches. Transects were centered within the experimental plots and spaced
one meter apart. Two surface storage values were estimated from these data;
1.) the maximum amount of surface storage which must be filled before surface
runoff from the plot could occur, and 2.) the amount of local depressional
storage that must be filled before movement of trapped water between depressions
begins to occur. Both values were calculated by first estimating the depth
of ponded water and then determining the areal extent associated with this
depth. The estimates given by this method are likely to be higher than
the actual surface storage on the treatment plots, since this method does
not take into account surface flow patterns and how surface depressions
are interconnected.
At the Beaufort County site, the high water table management, rough microtopography treatment showed the most consecutive and total number of wet days with the water table less than 30 cm deep, while the low water table, smooth microtopography treatment displayed the driest conditions. Wetland hydrology, as defined by the Army Corp of Engineers, was achieved at the Beaufort County site during both 1996 and 1997.
The Craven County site displayed much lower water table depths and drier soil conditions than the Beaufort County site throughout the study period. This was due in part to seepage around the water control structures in the field ditches and lateral seepage to a large channelized stream located adjacent to the site. The intended levels of water table control were not attained on all treatment plots. At the Craven County site, the high water table management, smooth microtopography treatment displayed the wettest conditions, while the low water table, smooth microtopography treatment displayed the driest conditions. Jurisdictional wetland hydrology, as defined by the Corp, was met on one hydrologic treatment (high water table, smooth microtopography) at the Craven site during 1996, a wetter than average year, but not during 1997, a drier than average year. These data indicate that the Craven County site was restored to a marginal wetland at best, and most likely continues to lean toward being an upland. Based on soil types at the two field sites, the Craven County site appears to historically be a drier site than the Beaufort County site. Though both sites contain hydric soils, the Roanoke soil of the Beaufort site is generally considered a wetter soil under undrained conditions than the Leaf soil at the Craven site. This indicates that even under natural, undrained conditions, the Craven County site may have been only a marginal wetland.
Microtopography treatments were found to influence the amount, intensity, and duration of outflow at both sites. The use of rough microtopography decreased total outflow volume, decreased outflow intensity, and typically increased the duration of outflow events over the smooth microtopography treatments at the Beaufort County site. Not enough outflow data was collected from the Craven site to quantify outflow results, although the same trends were indicated. Rough microtopography was shown to provide more variable soil moisture conditions in the surface soil layers (0 to 30 cm), especially during drier periods. Variations in soil moisture between rough and smooth microtopography treatments were not as apparent during wet periods when the water table was near the surface.
The water table hydrology data collected from
this research indicate that a similar water table response can be induced
by various combinations of water table control and surface roughing. For
example, water table hydrographs and hydroperiods were very similar between
the low water table management, rough microtopography treatments and the
high water table management, smooth microtopography treatments at the Beaufort
County site. Research has shown that vegetation diversity is greater when
the soil surface in uneven, due to more variable soil moisture and soil
chemical conditions. For this reason, some degree of surface roughing is
recommended for restoration of prior converted lands.