LaBaugh, James W.;; Nelson, Richard D.; Euliss, Ned H. (2018). Long-term changes in pond permanence, size, and salinity in Prairie Pothole Region wetlands The role of groundwater-pond interaction. Journal of Hydrolo

Journal of Hydrology: Regional Studies 17 (2018) 1–23
Contents lists available at ScienceDirect
Journal of Hydrology: Regional Studies

journal homepage: www.elsevier.com/locate/ejrh
Long-term changes in pond permanence, size, and salinity in

T
Prairie Pothole Region wetlands: The role of groundwater-pond
interaction
⁎
James W. LaBaugha, , Donald O. Rosenberryb, David M. Mushetc, Brian P. Neffb,
Richard D. Nelsond, Ned H. Euliss Jr.e,1
a
U.S. Geological Survey, 411 National Center, 12201 Sunrise Valley Drive, Reston, VA 20192, USA
b
U.S. Geological Survey, W 6th Ave. Kipling St. MS 413, Lakewood, CO 80225, USA
c
U.S. Geological Survey, Northern Prairie Wildlife Research Center, 8711 37th Street SE Jamestown, ND 58401, USA, USA
d
U.S. Fish and Wildlife Service, 3425 Miriam Avenue, Bismarck, ND 58501, USA
e
U.S. Geological Survey, USA
A R T IC LE I N F O ABS TRA CT
Keywords: Study Region: Cottonwood Lake area wetlands, North Dakota, U.S.A.
Wetland hydrology Study Focus: Fluctuations in pond permanence, size, and salinity are key features of prairie-
Climate variability pothole wetlands that provide a variety of wetland habitats for waterfowl in the northern prairie
Groundwater solutes of North America. Observation of water-level and salinity fluctuations in a semi-permanent
Prairie-pothole wetlands
wetland pond over a 20-year period, included periods when the wetland occasionally was dry, as
well as wetter years when the pond depth and surface extent doubled while volume increased 10
times.
New hydrological insights for the study region: Compared to all other measured budget components,
groundwater flow into the pond often contributed the least water (8–28 percent) but the largest
amount (> 90 percent) of specific solutes to the water and solute budgets of the pond. In drier
years flow from the pond into groundwater represented > 10 percent of water loss, and in 1992
was approximately equal to evapotranspiration loss. Also during the drier years, export of cal-
cium, magnesium, sodium, potassium, chloride, and sulfate by flow from the pond to ground-
water was substantial compared with previous or subsequent years, a process that would have
been undetected if groundwater flux had been calculated as a net value. Independent quantifi-
cation of water and solute gains and losses were essential to understand controls on water-level
and salinity fluctuations in the pond in response to variable climate conditions.
1. Introduction
The northern prairie of North America contains a remarkable variety of wetlands known as prairie potholes or sloughs that serve
as an important resource for waterfowl, producing as much as 50–80 percent of these birds in the continent (Batt et al., 1989). Biota
in these wetlands change in response to dynamic fluctuations in water levels and salinity, thereby resulting in a mosaic of different
community types and vegetation cover (Stewart and Kantrud, 1971) that influence waterfowl use of the wetlands (Swanson and
⁎
Corresponding author.
E-mail addresses: jlabaugh@usgs.gov (J.W. LaBaugh), rosenber@usgs.gov (D.O. Rosenberry), dmushet@usgs.gov (D.M. Mushet), bneff@usgs.gov (B.P. Neff),
Richard_d_nelson@fws.gov (R.D. Nelson), eulissfamilyinnd@gmail.com (N.H. Euliss).
1
Retired.
https://doi.org/10.1016/j.ejrh.2018.03.003
Received 5 June 2017; Received in revised form 8 December 2017; Accepted 19 March 2018
2214-5818/ Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).
J.W. LaBaugh et al. Journal of Hydrology: Regional Studies 17 (2018) 1–23
Duebbert, 1989; Weller, 1978). Understanding factors controlling fluctuations in water levels and salinity in the ponds of these
wetlands, therefore, is of interest to managers of these natural resources (Euliss et al., 2004, 2014; Renton et al., 2015; Robarts and
Bothwell, 1992), as well as those who study Earth surface processes (Arndt and Richardson, 1989; Goldhaber et al., 2014), wetland
hydrological characteristics (Winter and Woo, 1990; Woo and Rowsell, 1993), wetland vegetation dynamics (van der Valk, 2005) and
waterfowl habitat (Swanson and Duebbert, 1989).
Prairie-pothole wetlands occupy depressions capable of storing water. These depressions are found within end moraines, ground
moraines, stagnation moraines, outwash plains, and lake plains formed during the advance and retreat of the continental ice sheet
(Eisenlohr et al., 1972; Winter 1989). Most of these depressions lack interconnection by natural surface channels, however, water
levels in the ponds can increase to points in their basins where water can spill overland to adjacent wetland depressions (Eisenlohr
et al., 1972; Rosenberry and Winter, 1997; Shaw et al., 2012).
The ponded portion of the wetlands, where surface water is present in the prairie pothole depressions, originates primarily from
snow that melts in the spring and rainfall; most of the water lost from the ponds is through evapotranspiration (Shjeflo, 1968).
Groundwater interaction with wetland ponds mainly involves localized, shallow groundwater flow (Eisenlohr et al., 1972; Sloan,
1972) within the depressions containing the wetlands (Lissey, 1971). In cases where groundwater flow is generally toward a wetland
pond, reversals in groundwater flow directions at the periphery of the pond can occur due to two processes, transpiration and water
input from intense rain and snowmelt. Transpiration by vegetation at the pond perimeter can cause a localized depression in the
water table immediately adjacent to the pond, thereby causing water to flow out of the pond into groundwater (Berthold et al., 2004;
Meyboom, 1966; Meyboom et al., 1966; Mills and Zwarich, 1986; Rosenberry and Winter, 1997). Episodes of intense rain and
snowmelt can raise the water level in the pond above the adjacent groundwater level also causing water to temporarily flow out of the
pond, particularly for temporary ponds (Johnson et al., 2004).
Water stored in prairie-pothole wetland ponds varies in salinity from less than 300 ppm (fresh water) to greater than 35,000 ppm
(saline water) (Barica, 1975; Bierhuizen and Prepas, 1985; Hammer, 1978; Last, 1992; Petri and Larson, 1973; Rozkowski,1969;
Swanson et al., 1988). Wetland ponds that have surface or groundwater outlets are relatively fresh; those that have no outlets are
more saline (Sloan, 1972). Solutes are concentrated in prairie-pothole wetland ponds in two ways. When evapotranspiration is
greater than precipitation, wetland pond water levels decline and solutes are concentrated in the ponds (Jones and Deocampo, 2003).
Solutes also are concentrated during ice formation. As pond water freezes, solutes are excluded from the ice and concentrated in the
remaining liquid water. When freezing continues into the sediments, solutes are concentrated in the pore water in sediments below
the maximum ice extent (Ficken, 1967). After the sediments thaw, diffusion of solutes from the sediment pore water into the pond
contributes to pond salinity (Ficken, 1967; Heagle et al., 2013).
Evapotranspiration by plants at the edge of ponds can lead to localized increased salinity of shallow groundwater at the pond
periphery (Arndt and Richardson, 1993; Miller 1969), and salinization of soils (Arndt and Richardson, 1989). Within the more
permanent zone of soil salinization, Arndt and Richardson (1993) noted the increased salinity in the shallow groundwater at the pond
periphery can be transient because periodic recharge events result in groundwater flow that moves the accumulated solutes from the
periphery into the pond. Solutes are also added to the ponds from groundwater flow originating from the adjacent uplands (Arndt and
Richardson, 1993; Eisenlohr et al., 1972; Hayashi et al., 1988b; Heagle et al., 2007; Miller, 1969; Rozkowski, 1969). Rain, snow, and
surface runoff are other hydrologic sources of solutes to the ponds (Eisenlohr et al., 1972; Hayashi et al., 1988b). Decomposing
vegetation also can contribute to solutes in pond water (Biondini and Arndt, 1993).
Solutes are removed from wetland ponds by several mechanisms. Water flows overland from a pond to an adjacent wetland basin
when the pond level rises to its basin’s “spill” point (Eisenlohr et al., 1972) thereby removing solutes from the pond (Leibowitz and
Vining, 2003; Nachshon et al., 2014; Leibowitz et al., 2016). When the pond level is below the surface spill point of the depression it
occupies, no loss of solutes by surface-water flow can occur. Movement of pond water into groundwater (outflow seepage), however,
can remove solutes from the pond when the water level in the pond is higher than the water levels in the adjacent groundwater
(Sloan, 1972). Processes within a pond can favor removal of solutes from pond water by chemical precipitation (Goldhaber et al.,
2014) or by gas flux (Biondini and Arndt, 1993; Mills et al., 2011). When wetlands are dry, salts can be removed from the wetland
sediment surface by wind (Eisenlohr et al., 1972; Jones and Deocampo, 2003; LaBaugh et al., 1996; Luba et al., 1988), similar to loss
of salts that occurs from closed lakes (Langbein, 1961).
The semi-arid region occupied by prairie-pothole wetlands is subject to considerable variation in climate (Laird et al., 2003; van
der Valk, 2005). Episodes of drought commonly recur; less common are episodes of very wet conditions and deluge (Winter and
Rosenberry, 1998). Observation of changes in pond water levels and salinity in the past few decades have begun to document how
variations in water level and salinity are related to the processes controlling the hydrologic and chemical characteristics of these
ponds, both during the onset of drought and transition to wetter conditions (LaBaugh et al., 1996), as well as during persistent wet
conditions and high water levels (LaBaugh et al., 2016; Nachshon et al., 2014). Simple water balance models (rainfall and runoff)
have been used to relate wet and dry episodes to fluctuations in water levels of prairie-pothole wetlands (Carroll et al., 2005; Huang
et al., 2013; Johnson et al., 2005, 2015; Liu and Schwartz, 2011; Poiani et al., 1996). Quantification of the contribution of each of the
various processes affecting variations in pond water levels and salinity in prairie-pothole wetlands, however, are rare and are only of
several years duration (Hayashi et al., 1988a, 1998b; Heagle et al., 2007, 2013; Shjeflo, 1968; Woo and Rowsell, 1993).
Published water and solute budgets provide an initial foundation regarding the quantification of the factors controlling fluc-
tuations in water levels and salinity in the ponds of these wetlands that influence use by waterfowl. The several years’ duration of
published budget studies (Hayashi et al., 1988b; Heagle et al., 2007; Shjeflo, 1968; Woo and Rowsell, 1993) have not yet quantified
the full cycle of wet and dry conditions that have such an important influence on wetland habitat dynamics (van der Valk, 2005).
Measurement of the components of the water and solute budgets of the pond of Wetland P1 in the Cottonwood Lake area, North
2
Dakota, U.S.A., a semi-permanent pond that has no surface outlet, during a period of drought and deluge between 1979 and 1998,
provides the opportunity to add to the foundation of the earlier studies, because the duration of the study encompassed hydrologic
extremes. Results of the study of Wetland P1 are the focus of this report.
1.1. Purpose and scope
Here we examine the water and chemical balances of the pond of a semi-permanent wetland in a groundwater discharge area,
Wetland P1, during a period of climatic contrasts. Examination of such balances provides information about the importance of
various water and solute sources, and losses, on the geochemical characteristics of the wetland, which are related to the biological
communities for which these wetlands are an important resource. Complete water-budget studies of ponds of prairie-pothole wet-
lands have appeared in only a few studies (Shjeflo, 1968; Hayashi et al., 1988a; Woo and Rowsell, 1993) and such water budgets have
only been associated with chemical budgets (sulfate and chloride) for a recharge wetland (Hayashi et al., 1988b; Heagle et al., 2007).
Presented here are water and solute budgets for Wetland P1 for the period of 1979–1998. This period includes years when the
wetland pond was dry, as well as when water levels increased to a point where the distinct ponds in the wetland merged with each
other and with the pond of an adjacent wetland. The focus is on budgets for open-water periods, placed in context of the contribution
from snow prior to the open-water period, which is consistent with analyses of other pond water and solute budgets of prairie-pothole
wetlands (Eisenlohr et al., 1972; Hayashi et al., 1988a, 1998b; Heagle et al., 2007; Shjeflo, 1968).
2. Study area
Located in Stutsman County, North Dakota, U.S.A., the 92-ha study site is a U.S. Fish and Wildlife Service Waterfowl Production
area (Winter, 2003). Native prairie and wetlands covers 80 percent of the site; the remainder was cultivated for brome grass and
alfalfa prior to purchase in 1963. The Cottonwood Lake area study site in east-central North Dakota is near the eastern edge of a large
glacial stagnation moraine known as the Missouri Coteau. Glacial drift that covers the site is mostly clayey, silty till, the upper 5–15 m
of which is oxidized and fractured (Goldhaber et al., 2014). Many depressions occupy this hummocky area of approximately 25 m
relief. Depressions include wetlands that contain water for weeks to months before becoming dry (seasonal ponds) and those that
contain water throughout most years (semi-permanent ponds), becoming dry only during very dry years. Terms used here to describe
the temporal characteristics of the ponds are in accordance with those used by Stewart and Kantrud (1971). Wetlands in the study
area include those whose ponds lose water to groundwater (recharge wetlands), both receive water from and lose water to
groundwater (flow-through wetlands), and primarily receive water from groundwater (discharge wetlands). Terms used here to
describe the relation of the ponds to groundwater are in accordance with those used by Sloan (1972). Water levels in some wetland
ponds have increased until adjacent ponds coalesce or water can spill out of one depression or basin into an adjacent depression.
This region of North America is semi-arid. Average annual precipitation in east-central North Dakota is 440 mm and average
annual evapotranspiration is 810 mm (Rosenberry, 2003) based on data from three National Weather Service Stations for 1961–1990.
Precipitation is greatest during July, averaging 90 mm, and least during November through February. Snowfall averages 865 mm and
the snowpack commonly extends from early December through late March. Average annual temperature is 4 °C, while extreme
temperatures range from −47 to +48 °C (Rosenberry, 2003).
Investigations at the site began in 1967 and focused on the relation of waterfowl to prairie-pothole wetlands. Included in that
work were intermittent analyses of chemical characteristics of water in the wetland ponds. Salinity in the ponds varies from fresh to
saline. Detailed study of the hydrological characteristics of the wetlands at the site was added in 1979 and accompanied by more
frequent analysis of chemical characteristics of water in the wetland ponds and shallow groundwater. The term “pond” refers to the
surface water within the depression containing the hydric soils and associated vegetation that define a wetland, following the
distinction between a pond and its wetland noted by Eisenlohr et al. (1972) and Hayashi et al. (2016).
The focus of this study was the semi-permanent pond of Wetland P1 (Fig. 1) − a closed basin receiving groundwater discharge –
including its episodic merger with a separate seasonal pond in the wetland (identified in past studies as Wetland T1, but hereafter
identified as Pond T1, and the seasonal pond of the adjacent wetland (Wetland T3) between 1979 and 1998. The two seasonal ponds
of Wetlands T1 and T3 were relatively shallower (0.48 m – T1, 0.3 m – T3) and smaller (853 m3 – T1, 25 m3 – T3) than the semi-
permanent pond of Wetland P1 at the point of merger (1.11 m, 16,673 m3 with T1 and 2.03 m, 58,164 m3 with T3). Changes in the
pond surface extent of these wetlands in response to changes between drier and wetter conditions were documented in Winter and
Rosenberry (1998). At times, the ponds were absent and the wetlands were dry. In wetter years, the surface of the ponds extended
into the uplands to the point where all wetland emergent vegetation was inundated, decomposed and thereafter was absent during
the study period. Changes in the relation between these ponds and groundwater are described in Winter and Rosenberry (1995) and
Rosenberry and Winter (1997); however, water and solute budgets for open-water periods were not examined in those previous
publications, and are presented here.
3. Material and methods
3.1. Hydrological instrumentation
3.1.1. Groundwater
Twenty shallow observation wells (depths between 5.2 and 14.6 m below land surface) were installed in the Cottonwood Lake
3
Fig. 1. Location of wetlands, groundwater observation wells, test holes, and climate instrumentation in the Cottowood Lake study area.
modification of Fig. 1 from Winter, 2003.
wetlands study area as described in Winter and Carr (1980), and many of these wells were in the upland of Wetland P1. Water levels
were measured with a chalked steel tape on a monthly basis during winter, and weekly to biweekly from spring through fall. Six
additional wells (depths between 1.5 and 2.1 m below land surface) were installed between Wetlands P1 and T3 in June of 1989
(Rosenberry and Winter, 1997) and water levels were recorded with potentiometer-float systems connected to a digital data logger
programmed to produce bihourly and daily averages. Measuring points of all wells were surveyed to a common datum.
3.1.2. Surface water

Staff gages installed in each of the ponds were read weekly to biweekly or monthly, on the same day as water levels were
measured in the wells, beginning in 1979. Continuous stage (water level) data for the pond of Wetland P1 were recorded by a
potentiometer-float system installed in a stilling well near the center of the pond in July 1980. Data were recorded every minute from
which hourly and daily averages were computed by a digital data logger. Continuous stage data were collected during as much of the
ice-free period as was practical, typically from April through October. All pond staff gages were surveyed each year to a common
datum.
4
3.1.3. Atmospheric water

A tipping-bucket rain gage and digital data logger were installed in July 1979, and a standard non-recording rain and snow gage
was installed in January 1980, both in the upland adjacent to Wetland P1 (Winter and Carr, 1980). For periods when data from those
gages were missing or not available, data from a nearby National Weather Service gage at Woodworth, North Dakota (Station ID:
GHCND:US1NDSM0008 − located approximately 16 km west-northwest of the Cottonwood site − data accessed January 8, 2016)
were used.
Instrumentation needed to calculate evapotranspiration was installed beginning in 1979 (Fig. 1). Water-surface and air tem-
perature, relative humidity, and wind speed were measured at a raft near the center of Wetland P1 (Winter and Carr, 1980).
Atmospheric sensors were installed 2 m above the water surface. Vapor-pressure gradient was determined from relative humidity
data and by assuming that the atmosphere was saturated at the water surface. Additional instrumentation was installed in 1982
(Parkhurst et al., 1998). Thermistors thermometers were installed approximately 2 cm beneath the water surface and 20 cm above the
wetland sediments, and also 0.5 m and 1 m below the wetland sediment-water interface to determine change in heat stored in the
water column and underlying sediments. Solar and atmospheric radiation was measured in the upland near Pond T1. Wind speed was
measured at 2 m above land surface in the vicinity of the radiation instruments. Auxiliary air temperature and relative humidity data
were measured with a hygrothermograph located in an instrument shelter in the upland between Wetlands P1 and T3.
3.1.4. Measurement uncertainty

Measurements of the components of the water budgets of lakes and wetland ponds are associated with some uncertainty (Winter,
1981; LaBaugh, 1986). The water budget of the pond of Wetland P1 was determined as follows:
p + r + gi − et − go = dv (1)
Where
p = precipitation (snow, rain) falling directly on the pond surface,
r = runoff from rainfall,
gi = groundwater input,
et = evapotranspiration,
go = groundwater output, and
dv = change in pond volume.
Rainfall, groundwater in and out, evapotranspiration, and change in pond volume were measured independently. Runoff was
calculated using change in pond stage and rainfall data. No component, however, was estimated from the difference between inputs,
outputs, and change in storage. This approach enabled us to represent cumulative error of all of the measured components and any
water gain or loss unaccounted for in the budget as the residual of the water budget (R)
p + r + gi − et − go − dv = R (2)
The cumulative error in the budget (δ) was determined from a first order error analysis assuming errors in measurement of the
budget components are independent and randomly distributed.
δ = square root of (δ2p + δ2r + δ2gi + δ2et + δ2go + δ2dv) (3)
Where δ = error,
δ2
p = (precipitation falling directly on the pond surface × error estimate for precipitation)2,
δ2
r = (runoff from rainfall × error estimate for runoff)2,
δ2
gi = (groundwater input × error estimate for groundwater input)2,
δ2
et = (evapotranspiration × error estimate for evapotranspiration)2,
δ2
go = (groundwater output × error estimate for groundwater output)2, and
δ2
dv = (change in pond volume × error estimate for change in pond volume)2.
Estimates of measurement error used in this analysis primarily were based on Winter (1981) and were: 5 percent for precipitation
falling directly on the pond surface, 50 percent for groundwater input, 10 percent for evapotranspiration determined from the energy
budget and 15 percent for evapotranspiration determined from the Mather and Hamon methods (based on Rosenberry et al., 2004),
50 percent for pond water loss by flow to groundwater, and 10 percent for change in pond volume. Runoff error was not quantified in
Winter (1981). We assumed an estimate of error of 100 percent for runoff from rain based on Motz et al. (2001). In cases when the
cumulative error in the budget was larger than the budget residual, we assumed this indicated that all water gains and losses were
accounted for within the uncertainty in the measured budget. In cases when the cumulative error in the budget was smaller than the
budget residual, we assumed this indicated some water gain or loss was unaccounted for within the uncertainty in the measured
budget, whereby the difference between the cumulative error and the budget residual represented the part of the budget not mea-
sured.
3.2. Water budget calculation
3.2.1. Pond surface area and volume

Stage-area and stage-volume relations for the pond were determined from surveys of the wetland pond bathymetry and adjacent
5
upland, including a plane table survey made in 1983 and a lidar (light detection and ranging) survey of the entire waterfowl
production area in 2008. Daily change in volume of the pond was calculated by difference between the volume of the pond on one
day and the volume of the pond on the previous day. Summation of daily values was used to calculate monthly values.
3.2.2. Precipitation volume

Precipitation falling on the wetland pond was determined in two ways. First, the total volume contributed by snow at the end of
each snow-accumulation season was approximated by difference using the last pond volume in fall, based on the last recorded pond
water level, and the first pond volume in the following spring, based on the first recorded pond water level. This volume includes the
combination of snow that fell directly on the wetland or frozen pond, snow that drifted onto those surfaces, and snowmelt flowing
over frozen ground to the pond. Second, during the months when the pond surface was not covered by ice, the precipitation volume
was determined by multiplying daily rainfall by the pond water-surface area associated with the water-level elevation on the day the
rainfall was recorded. Monthly values of precipitation in cubic meters were determined from summation of daily values.
3.2.3. Runoff volume

On days when rainfall was measured, the volume of the pond was compared to the volume from the previous day. If the change in
volume of the pond was greater than the volume of rainfall falling on the surface of the pond, the difference was attributed to runoff
(Shjeflo, 1968). Monthly runoff in cubic meters was calculated by summation of daily values within the month.
3.2.4. Evapotranspiration volume

Evapotranspiration was calculated using the energy-budget method for years (1982–1985, 1987) when instrumentation at the site
enabled such calculations (Parkhurst et al., 1998). The Mather (1978) and Hamon (1961) equations were used to determine eva-
potranspiration for the period 1979–1998, based on an evaluation of results from 13 equations for determining evapotranspiration in
comparison with the energy budget (Rosenberry et al., 2004).
Both the Mather (commonly called Thornthwaite) and Hamon equations use air temperature data. For periods when daily average
air temperature was measured at the site, monthly average air temperature values were determined from daily data. For periods when
air temperature data were not available at the site, data from the Jamestown, North Dakota, National Weather Service site
GHCND:USW00014919 – located approximately 37 km southeast of the site (Mene et al., 2012, National Oceanic and Atmospheric
Administration(NOAA) National Climatic Data Center. http://doi.org/10.7289/V5D21VHZ access date January 8, 2016) were used
to determine monthly average air temperature values (1979 to spring 1982, fall to early spring 1982–1990, 1994–1995, April and
October 1996), based on the regression equation: monthly average air temp at the site in degrees C = −1.183 + monthly average air
temperature at Jamestown in degrees C × 0.9808 (n = 120, r2 = 0.99 F = 20815, p > F < 0.0001).
The Mather equation calculates potential evapotranspiration (PET) in millimeters per day as follows:
PET = [1.6 (10Ta/I)6.75×10−7I3−7.71×10−5I2+1.79×10−2I+0.49] (10/d)

(4)
Where Ta = air temperature in degrees C, I = heat index, and d = number of days in the month.
Monthly heat index values (I) were determined from the monthly average air temperature data (Ta) using the equation
I = (Ta/5)1.514. (5)
The Hamon equation calculates daily PET in mm as follows:
PET = 0.55 (D/12)2 (SVD/100) 25.4 (6)

−3
Where D = h of daylight, SVD = saturated vapor density at mean air temperature, in g m
The equation for calculating saturated vapor density is:
SVD = 5.018 + 0.32321 Ta + 0.0081847 Ta2 + 0.00031243 Ta3 (7)
The average hours of daylight for the month were obtained from the U.S. Naval Observatory at http://aa.usno.navy.mil/data/
docs/Dur_OneYear.php (access date January 22, 2016) for the location of Jamestown, North Dakota.
Values in millimeters per month determined from both the Mather and Hamon equations were multiplied by the surface area of
the pond on the last day of the month to obtain evapotranspiration in cubic meters for the month.
3.2.5. Groundwater volume

The volume of water either flowing from groundwater into the wetland pond, or flowing from the pond to groundwater each
month was determined at multiple points around the pond perimeter using the Darcy equation,
Q = KIA, (8)
where
Q is groundwater flow (volume/time),
K is hydraulic conductivity (length/time)
I is the hydraulic gradient, and
6
A is the cross-sectional area through which water flows into or out of the wetland in relation to groundwater.
K was determined from 79 aquifer (slug) tests conducted at 49 observation wells in the study area and averaged 4.7 × 10−5 cm/
second. I was calculated by subtracting the water level in the pond from the water level in the wells located near the P1 pond and
dividing by the horizontal distance between the well and the P1 pond shoreline. A was determined by multiplying the length of the
shoreline by the average depth (15.2 m) of the local, shallow groundwater system (Parkhurst et al., 1998), which corresponds to the
approximate maximum depth of weathered and fractured till at the site (Goldhaber et al., 2014).
The shoreline of P1 was divided into segments, each attributed to a well that was used to calculate the hydraulic gradient near
that portion of the pond (Fig. 1). Segments initially were assigned to each of the following observation wells: 13, 14, 17, 20, 25, and
26 (Fig. 1). When the ponds of T1 and P1 coalesced, additional segments were assigned to observation wells 15 and 23. When the
combined pond of P1-T1 coalesced with the pond of T3, the segment assigned to well 13 was assigned to wells 12 and 52. Eq. (1) was
applied to each segment and, depending on the direction of the hydraulic gradient, the total groundwater flow to the pond, or flow
from the pond to groundwater, was calculated as the sum from each segment.
As the water level in P1 rose, and the shoreline expanded, the above calculations were adjusted in three ways. First, the horizontal
distance between each well and the shoreline was reduced as the shoreline moved closer to the well. Second, as the water level in P1
rose and approached to within 3 m of a piezometer used to calculate the hydraulic gradient to the pond, a more distant piezometer
was used to calculate the hydraulic gradient to the pond (Fig. 1): the gradient calculations were modified by substituting wells 12 and
52 for well 13, well 16 for well 15, well 18 for well 17, well 28 for well 25, and well 27 for well 26. Third, when the pond of Wetland
P1 merged with Pond T1, and eventually with the pond of Wetland T3, the shoreline changed. New shoreline segments were then
added as noted above (Fig. 1) and the shoreline length for each segment was adjusted for the shoreline configuration of the new,
merged pond.
Calculating monthly groundwater flow for each segment meant we could distinguish segments where groundwater was flowing
into the pond from segments where water was flowing from the pond into groundwater. Monthly groundwater exchange with the
pond therefore was not calculated as a net value. Instead, we determined separate monthly volumes for gi and go. Also, the si-
multaneous occurrence of groundwater flow into the pond and pond-water flow to groundwater did not indicate a flow-through
condition, where water lost to groundwater would eventually leave the local basin. Instead, water flowing out of the pond ended at
nearby depressions in the water table resulting from loss of shallow groundwater to evapotranspiration.
3.3. Chemical data collection
3.3.1. Groundwater
Groundwater was removed from the observation wells by use of a bailer or a peristaltic pump. In cases when removal of
groundwater left no water in the well, groundwater that seeped into the well overnight was collected by bailer or pump for sub-
sequent physical and chemical analysis. Otherwise, samples of groundwater were collected after three casing volumes of water had
been removed from the well. Around the periphery of Wetland P1 water was collected from wells 13, 14, 15, 16, 17, 20, 25, and 26.
Collection from wells 13 and 16 was initiated in 1980, from wells 17, 20, 25, 26 in 1983, and from well 15 in 1984. Sample collection
took place primarily between April and October, approximately monthly prior to 1983, and intermittently from 1983 through 1991.
In addition to the more detailed chemical analyses of groundwater collected from the wells, in-situ measurements were made. From
1987 to 1995 water temperature, pH, and specific conductance commonly were measured once in the summer in wells 17, 20, 25,
and 26 using an in-situ Hydrolab probe designed to fit into 2-inch-diameter wells.
Water collected prior to 1983 was analyzed at the U.S. Fish and Wildlife Service Northern Prairie Wildlife Research Center
laboratory in Jamestown, North Dakota, according to methods described in Swanson et al. (1988). Water collected between 1983 and
1991 was analyzed at the U.S. Geological Survey National Water Quality Laboratory in Lakewood, Colorado, according to methods
described in Fishman and Friedman (1989). Analysis of pH, specific conductance, and total alkalinity, was made on unfiltered
groundwater. Water filtered through a 0.45-micrometer filter was used for analysis of dissolved calcium (Ca), magnesium (Mg),
sodium (Na), potassium (K), chloride (Cl), sulfate (SO4), nitrate nitrogen, and ammonia nitrogen. For water analyzed at the U.S.
Geological Survey laboratory, filtrate used for cation analysis was preserved with 1 mL of nitric acid per 250 mL of filtrate, and water
used for nitrogen analyses was preserved with 1 mL of HgCl. The analysis for dissolved nitrite plus nitrate as nitrogen performed by
the U.S. Geological Survey laboratory was assumed to represent nitrate nitrogen, whereas the U.S. Fish and Wildlife Service la-
boratory determined nitrate individually as nitrate nitrogen.
3.3.2. Surface water

Water was collected from the pond of wetlands at the site using a tube water-column sampling device, typically 0.65–1.35 m long
with a minimum diameter of 10 cm, as described in Swanson (1978). Between 1979 and 1995 water was collected from the mid-point
of each of three transects (transects described in LaBaugh et al., 1987), and from 1996 to 1998, from a single transect within the
pond. When a pond was very shallow and almost dry, water was collected from the center of the pond. Water collected from each
transect was processed separately for subsequent analysis, thereby enabling determination of heterogeneity within the pond.
Water from the tube sampling device was transferred to a plastic bucket or large polyethylene bottle, then to either 2-L, 250-mL,
or 500-mL polyethylene sample bottles prior to transport to the laboratory, according to the protocols of the laboratories (Swanson
et al., 1988; Fishman and Friedman, 1989). The U.S. Fish and Wildlife Service laboratory was used for analysis of pond water
collected before 1983. The U.S. Geological Survey laboratory was used for analysis of pond water collected from 1983 to 1992, and
1994. For pond water collected from 1992 to 1998 water analysis also was conducted at the Bureau of Reclamation Laboratory in
7
Bismarck, North Dakota, using standard methods (Greenberg et al., 1992).
3.3.3. Atmospheric water

Results of chemical analyses of atmospheric deposition were obtained from the National Atmospheric Deposition Program/
National Trends Network gage at Woodworth, North Dakota, located approximately 16 km west-northwest of the Cottonwood site
(Station ND11 – data accessed July 8, 2016). Data for monthly average values from this site were available beginning December 1983
and continuing through 1998. The North Dakota State Department of Health collected atmospheric deposition at the Woodworth site
for chemical analysis between 1981 and 1984 (Deutschman and Ell, 1986). Data for the average values for May to October were
available for each of those years from the North Dakota State Department of Health.
3.4. Chemical budgets
3.4.1. Groundwater
To determine groundwater contribution to chemical flux entering the pond, the median concentrations in the wells within each
shoreline segment were multiplied by the monthly volumes of groundwater flow into the pond for their respective segments. The
median solute values used for groundwater input in each segment were those for the entire study period for that segment. When the
wetland surface expanded to encompass well 12 the gradient associated with nearby well 52 was used. This method of calculation is
consistent with other similar studies where solute concentrations in groundwater have been documented on decadal time scales
(LaBaugh et al., 1995; Buso et al., 2009).
To determine chemical flux from the pond to groundwater, the median concentration in the pond for the month was multiplied by
the total monthly volume of water flowing from the pond into groundwater. In some months detailed chemical analyses were not
available to calculate this flux (April 1982, 1984, 1986, 1996, May to September 1983, and October 1995 and 1996). For those
months with missing data the median specific conductance for the month was used with regression relations between Ca, Mg, Na, K,
Cl, SO4, and alkalinity in the pond of Wetland P1 for the years 1979–1998 to estimate concentrations in the wetland pond
[Ca = 25.55 + (μS/cm * 0.04347), n = 374, r2 = 0.70; Mg = −48.28 + (μS/cm * 0.11305), n = 374, r2 = 0.96;
Na = −11.18 + (μS/cm * 0.06849), n = 373, r = 0.74; K = −3.53 + (μS/cm * 0.02629), n = 373, r = 0.82; Cl = 3.05 + (μS/
2 2
cm * 0.00886), n = 374, r2 = 0.49; SO4 = −443 + (μS/cm * 0.69896), n = 374, r2 = 0.95; alkalinity = 4.28 + (μS/
cm * 0.00069405), n = 374, r2 = 0.11]. All regressions were statistically significant at the 0.01 level. For a few months (April 1989,
1990, 1991, October 1988, and November 1981) no specific conductance data were available, therefore chemical concentrations
from either the next month or the previous month were used in the calculations of the mass flowing out of the pond into groundwater.
3.4.2. Change in mass in the pond

On the dates for which water was collected for subsequent chemical analysis in the laboratory, the volume of water in the pond on
that date was multiplied by the solute concentration to determine the mass of solute in the pond. Because dates of pond water
collection for chemical analysis rarely coincided with the beginning or the end of a month, monthly changes in mass were not
calculated. Change in mass in the pond during the open-water period was determined by the difference between the mass on the
individual dates chemical analyses were available during the open-water period.
3.4.3. Rainfall and runoff from rain

Solute input from atmospheric precipitation during the open-water period was determined by summing daily volumes for rain
within a month, then multiplying the monthly total in cubic meter (m3) for the month by the average concentration for that month
obtained from the Woodworth National Atmospheric deposition site, for the years 1984–1998. Solute input during the open-water
periods of 1981–1983 was determined by multiplying the monthly total in m3 for each month by the average concentration at the
Woodworth site for the period May to October for each of those years published by Deutschman and Ell (1986). To approximate the
relative contribution from atmospheric deposition during the years 1979 and 1981, the average concentrations for the period
1981–1984 published by Deutschman and Ell (1986) were used in the calculation, assuming data from those years were a proxy for
atmospheric deposition chemical composition in 1979 and 1980.
Monthly totals for runoff from rain were multiplied by the concentrations in atmospheric deposition for the month to calculate
solute input to the pond each month of the open-water period. Open-water period values for solute mass input were calculated as the
sum of monthly values. As in the case of rain, average values for chemical constituents determined from the period when data were
available (1981–1984) were used to approximate the relative contribution of runoff in the open-water periods prior to December
1983.
4. Results
4.1. Water budgets
Changes in pond volume from spring to fall, and water entering the pond or leaving the pond during that period, are the main foci
of the budget analysis presented herein. On an annual basis, however, moisture deposited on the landscape and the frozen pond
surface by snow and rain between fall and spring plays a role in determining the water volume present in the pond at spring thaw.
Rain falling directly on the wetland pond during the open-water period commonly was the largest source of water for the pond in
8
Fig. 2. Water budget gain and loss components in millimeters for the pond of Wetland P1, 1979–1998, shown as box plots. Values are amounts
determined in the open-water period, except for snowmelt. The horizontal line within each rectangle is the median, the rectangle is bounded by the
25th and 75th percentiles, the horizontal lines at the end of the vertical lines indicate the 10th and 90th percentiles and the dots are all of the data
points, including the maximum and minimum values. SNOW is snowmelt, RAIN is rain falling directly on the pond surface, RUNOFF is rainfall
runoff from the watershed of the pond, G IN is groundwater input to the pond, ET is evapotranspiration, and G OUT is flow out the pond into
groundwater.
most years (Fig. 2, Table 1). In a few years (1983, 1987, 1997), the volume contributed to the pond from snowmelt in the spring
exceeded the contribution from rain falling on the pond surface during the subsequent open-water period. Runoff from rain con-
tributed less water to the pond than rain falling directly on the pond surface or snowmelt in many years, but was more important in
drier years when there was diminished or no snowmelt (1989–1992), and in the wetter year of 1993. Runoff from rain was the largest
Table 1
Water budget components for the pond of Wetland P1, 1979–1998.
Year All values are in cubic meters
snow Open-water period

melta
budget period rainfall runoff groundwater Evapotranspiration change in pond
volume
inflow outflow Energy budget Mather Hamon
method method method
1979 *b April–October 7385 380 659 11 * 8957 10468 −4609

1980 3331 April-October 7945 1395 1005 377 * 7880 8727 −3984
1981 1691 March- November 5945 1801 1846 176 * 7223 8154 −4145
1982 7716 April-October 8939 3254 1622 181 8716 9033 10185 −2275
1983 8076 April-September 6575 1227 802 313 11057 9680 11705 −9730
1984 6797 April-October 9026 1771 1404 279 13078 9850 11323 −6967
1985 2014 April-September 3310 646 817 490 8962 5726 6444 −7209
1986 2950 April-September 7750 4109 1980 170 * 6911 7878 2476
1987 10519 April-October 9444 3063 2683 86 19350 12498 14058 −6726
1988 2034 April-September 3461 195 1456 273 * 6428 8000 −11296
1989 1474 April-October 3557 1648 1383 978 * 3486 4076 −1336
1990 0 April-September 2590 3300 1359 901 * 3810 4547 0
1991 0 April-July 1735 831 907 1048 * 2115 2638 113
1992 669 April-July 1102 1176 494 1603 * 1371 1650 −669
1993 2360 April-October 13819 15684 4520 378 * 12812 13384 19337
1994 2283 April-October 21554 6899 4214 123 * 21411 24420 10810
1995 2223 April-October 25512 12552 4750 250 * 22975 26212 19957
1996 17512 April-October 27454 8475 4024 611 * 23228 26617 −10332
1997 33407 April-October 27722 3857 5483 11 * 28334 32524 −16659
1998 13503 April-October 24725 6009 4984 375 * 29722 33321 −20492
Snowmelt represents water accounting for the difference in pond volume when the pond depth was last measured in the previous year and first
a
measured in the next year when the pond was free of ice.
b
*Indicates no data.
9
Table 2
Water budget component percentages for the pond of Wetland P1, open-water periods, 1979–1998, and including snowmelt percentage of open-
water inputs for reference.
Year Percent of total open water period gains or (losses)a
snow Open-water period

meltb
Water inputs Water losses
Energy budget Mather method Hamon method
rainfall runoff groundwater Evapotranspiration groundwater Evapotranspiration groundwater Evapotranspiration groundwater
c d
1979 * 88 5 8 * * (100) (0 < ) (100) (0 < )
1980 32 77 13 10 * * (95) (5) (96) (4)
1981 18 62 19 19 * * (98) (2) (98) (2)
1982 56 65 24 12 (98) (2) (98) (2) (98) (2)
1983 94 76 14 9 (97) (3) (97) (3) (97) (3)
1984 56 74 15 12 (98) (2) (97) (3) (98) (2)
1985 42 69 14 17 (95) (5) (92) (8) (93) (7)
1986 21 56 30 14 * * (98) (2) (98) (2)
1987 69 62 20 18 (100) (0) (99) (1) (99) (1)
1988 40 68 4 28 * * (96) (4) (97) (3)
1989 22 54 25 21 * * (78) (22) (81) (19)
1990 0 36 46 19 * * (81) (19) (83) (17)
1991 0 50 24 26 * * (67) (33) (72) (28)
1992 24 40 42 18 * * (46) (54) (51) (49)
1993 7 41 46 13 * * (97) (3) (97) (3)
1994 7 66 21 13 * * (99) (1) (99) (1)
1995 5 60 29 11 * * (99) (1) (99) (1)
1996 44 69 21 10 * * (97) (3) (98) (2)
1997 90 75 10 15 * * (100) (0 < ) (100) (0 < )
1998 38 69 17 14 * * (99) (1) (99) (1)
a
Percentages not totaling 100 are due to rounding.
b
Snowmelt is not included in the total water input determined for the open water period: value shown is for reference relative to total open-
water period water inputs.
c
*indicates no data.
d
*indicates absence of energy budget evapotranspiration data for that year: total water loss was not calculated.
source of water to the pond in 1990, 1992, and 1993. Groundwater generally was the smallest source of water for the pond in
comparison to rainfall or runoff from rain, yet did contribute volumes similar to runoff from rain in many years.
Evapotranspiration was the largest water loss from the pond (Fig. 2), representing greater than 90 percent of the total water lost in
all but the drier years of 1989–1992 (Table 2). Only in those drier years did pond water lost by flow into groundwater represent more
than 10 percent of the total water loss. As the pond volume and periphery diminished between 1989 and 1992, the flow of pond water
to groundwater became more important as a mechanism removing water from the pond, being approximately equivalent to water loss
from evapotranspiration in 1992. The importance of evapotranspiration as the primary mechanism removing water from the pond
was similar among the three methods used to determine evapotranspiration (Table 2) even though the absolute volumes determined
from each method were not identical (Table 1).
In most years the pond volume declined during the open-water period (Table 1) and the pond was shallower in the late summer or
fall than it was in the spring (Fig. 3). In some years (1986, and 1993–1995) pond volume increased during the open-water period
(Table 1) and the pond was deeper at the end of the open-water period than at the beginning of the period. In other years (1990 and
1991) change in volume was negligible because in both years the pond was dry in April and in late summer. The budget computation
periods for 1990 and 1991 included data from the recording stage gage prior to the installation of the staff gage. Differences in water
depths between the beginning and end of the open-water periods for 1990 and 1991 based on staff gage data shown in Fig. 3 therefore
do not reflect the water depths recorded during the entire budget computation period.
Change in pond volume during the open-water period was not accounted for completely by the difference between total water
gains and losses, based on the budget residual – the difference between the change in pond volume and total water gains and losses
(Table 3). A value of the budget residual other than zero indicates some water gains or losses may have been underestimated,
overestimated or unaccounted for by the measured components of the water budget. The budget residual, however, includes un-
certainty in measurement in addition to the difference between measured water gains and losses. Cumulative uncertainty in the
budget was obtained from a first order error determination. In some years (1986, 1990–1995), the value of the cumulative error in
the budget was larger than the absolute value of the budget residual, therefore indicating all water gains and losses were accounted
for within the uncertainty in the measured budget. In other years (1979–1985, 1987–1989– except 1987 when energy-budget
evapotranspiration was used – and 1996–1998) the value of the cumulative error was less than the absolute value of the budget
residual, indicating some water gain or loss from the pond was not measured. Hydraulic gradients based on measurements at wells
10
Fig. 3. Changes in pond water depth in the pond of Wetland P1, 1979–1998. Gaps in the line representing water depth represent periods when the
pond was covered with ice.
Table 3
Cumulative water budget error and budget residual for the pond of Wetland P1, open-water periods 1979–1998.
All values are cubic meters for the open-water period
Budget residual (Total input − total output) − Cumulative water budget error Cumulative error − absolute value of the budget
change in pond volume residual
Year with energy with Mather with Hamon with energy with Mather with Hamon with energy with Mather with Hamon
budget method method method budget method method method budget method method method
1979 *a 4065 2554 * 1551 1751 * −2514 −803

1980 * 6072 5225 * 1987 2065 * −4085 −3160
1981 * 6338 5407 * 2353 2421 * −3985 −2986
1982 7193 6876 5724 3502 3653 3720 −3691 −3223 −2004
1983 6964 8341 6316 1992 2203 2414 −4972 −6138 −3902
1984 5811 9039 7566 2459 2554 2687 −3352 −6485 −4879
1985 2530 5766 5048 1412 1389 1458 −1118 −4377 −3590
1986 * 4282 3315 * 4377 4414 * 95 1099
1987 2480 9332 7772 3950 3921 4038 1470b −5411 −3734
1988 * 9707 8135 * 1680 1826 * −8027 −6309
1989 * 3460 2870 * 1938 1964 * −1522 −906
1990 * 2538 1801 * 3449 3469 * 911 1668
1991 * 197 −326 * 1131 1155 * 934 829
1992 * 467 188 * 1462 1468 * 995 1280
1993 * 1496 924 * 16095 16105 * 14599 15181
1994 * 323 −2686 * 8043 8233 * 7720 5547
1995 * −368 −3605 * 13442 13575 * 13074 9970
1996 * 26446 23057 * 9542 9740 * −16904 −13317
1997 * 25376 21186 * 6720 7134 * −18656 −14052
1998 * 26113 22514 * 8244 8548 * −17869 −13966
a
* indicates absence of energy budget evapotranspiration data for that year: budget residual and cumulative water budget error were not
calculated.
b
Bold font indicates absolute value of the budget residual was smaller than the cumulative water budget error.
may not be indicative of gradients closer to the pond. It is possible, perhaps likely, that the water table closer to the pond was lower
than at the more distant well, in which case more water could have been lost to groundwater due to evpotranspiration drawdown in
some years than was indicated based on gradients integrated over the distance between the pond shoreline and the well (Rosenberry
and Winter, 1997). The unaccounted water loss from the pond in some years also may represent water lost by transpiration not
adequately represented by the method used to calculate evapotranspiration. As the pond shrank in 1989, cattail germinated on the
exposed wetland sediments, with cattail increasing in areal extent from 1990 to 1993. In the drier years of 1990–1992, the extent of
evapotranspiration drawdown perhaps was sufficient to be adequately represented by the wells, more so than in other years. In 1994
and 1995 the pond primarily was open water with little living emergent vegetation present in the pond.
The variable and dynamic effect of water gains and losses on the depth of the pond is evident in Fig. 3. The amount of snowmelt
prior to pond thaw each year determined the starting depth and volume of the pond. In most years, a rise in pond depth occurred
11
Fig. 4. Open-water-period water gains and losses for the pond of Wetland P1 in millimeters, 1979–1998, using the evapotranspiration value from
the Mather equation.
sometime after the snowmelt and pond thaw in response to rainfall, as well as when runoff occurred. Runoff from rainfall to the pond
was episodic, occurring when rain fell on frozen ground in the spring and occasionally thereafter more commonly during periods
when rain occurred over two or more days. Runoff, however, was not evident during all periods when rain occurred on multiple
consecutive days. Consequently years with similar volumes of rainfall input, such as 1979 and 1986, had dissimilar volumes of runoff
(Table 1) varying by an order of magnitude. A general decline in pond depth following peaks that commonly occurred earlier in the
open-water period reflected the dominance of water loss by evapotranspiration relative to water inputs. In some years, rainfall on the
pond surface and associated episodic runoff in the latter part of the open-water period caused increases in pond depth prior to
formation of ice cover on the pond.
Changes in pond water depth as a result of water losses exceeding gains, or gains exceeding losses, during the open-water period
(Fig. 3) were associated with large changes in the pond surface area. At a water depth of 1.11 m, the pond of Wetland P1 merges with
Pond T1, and at a depth of approximately 2.03 m, that single pond merges with the pond of Wetland T3. When either of the adjacent
Wetland ponds merged with the pond of P1, the combination was identified as the pond of Wetland P1 and that volume was used in
all of the calculations in this budget analysis. Because of the variation between drier and wetter conditions observed between 1979
and 1998, the surface area of the pond ranged from zero when dry episodically in 1989–1992 to 63,761 m2 in late April and early
May of 1997. The surface area at the peak in 1997 was more than double the peak of 29,489 m2 in mid-April of 1987, the largest
value recorded prior to the onset of the wetter conditions that began in 1993.
Water gains and losses to wetlands when presented in terms of water fluxes per unit area, such as in Fig. 4, enable comparison of
year-to-year trends independent of the marked variations between years with respect to pond surface area. For the purpose of this
comparison, evapotranspiration values derived from the Mather equation were used because values were available for all years and
the distribution of those values was more similar to energy-budget values than those derived from the Hamon equation (Fig. 2). The
year 1993 was the wettest year of the study, when rainfall measured during the open-water period was 638 mm. Subsequent years
(1994–1998) were associated with consistently higher rainfall amounts during the open-water period than occurred in most other
years prior to 1993, when rainfall was more variable. During some open-water periods in years prior to 1993, the rainfall was similar
to that measured in the latter wetter years following 1993 (471 mm – 1980, 420 mm – 1982, 446 mm – 1986, 451 mm – 1997,
417 mm – 1998). These years of somewhat similar rainfall totals during the open-water period, however, differ substantially in the
volumetric contribution to the water budget of the pond (Table 1). Such differences highlight the importance of the size of the pond
surface area to the translation of those rainfall totals into the volumetric contribution to the water budget of the pond in different
years. This also highlights the importance of snowmelt because it affects the size of the pond at the beginning of the open-water
period upon which rain falls.
Changes in surface area of the pond also were associated with changes in the length of shoreline at the periphery of the pond.
Shoreline length is used in the calculation of groundwater flux to and from the pond. Consequently, as the pond surface and perimeter
increased substantially beginning in 1993, the volume of groundwater inflow to the pond was greater (> 4000 m3) from 1993 to
1998 (Table 1) than in earlier years, even though, when expressed on a linear basis, values for the open-water periods of 1993–1998
were similar to values in many of the other years (Fig. 4).
4.2. Solute budgets
Groundwater flowing into the pond of Wetland P1 was the largest external source of calcium, magnesium, sodium, potassium,
12
J.W. LaBaugh et al.
Table 4
Summary of solute budgets for the pond of Wetland P1, for open-water periods, 1979–1998. Values are in kilograms except for alkalinity (equivalents), hydrogen ion (milliequivalents), and specific
conductance (μS cm −1).
solute rainfall runoff groundwater in groundwater out change in pond residual I-O +/− dS Cumulative error
calcium 2.13a 0.253–6.41b 0.69 0.07–2.69 686 203–2110 35.5 0.80–397 0.11 (−3011)c–1524 668 (−141)–4252 365 151–1066
magnesium 0.42 0.06–1.14 0.13 0.01–0.36 1787623–4159 52.4 0.44–554 407 (−1318)–2340 1156 (−452)–2840 896 312–2085
sodium 0.52 0.08–1.9 0.14 0.02–0.52 1022 350–2300 34.2 0.73–340 160 (−822)–1554 675 (−303)–1623 524 178–1152
potassium 0.39 0.07–0.89 0.11 0.01–0.3 40.5 14–99.5 14.9 0.17–155 93.7 (−465)–598 (−114) (-521)–547 42.3 14–79
chloride 1.05 0.07–3.31 0.33 0.02–1.25 119 15.5–281 6.13 0.06–92.1 (−0.44) (-120)–286 74.1 (−74)–219 60.9 12–142
13
sulfate 7.37 1.16–25.8 1.28 0.21–14.4 10797 4108–24559 253 2.7–3417 682 (−7046)–12321 8351 (−3095)–19968 5437 2054–12288
alkalinity 0 0 14289 5368–43892 1883 11.7–5926 4252 (−155641)–140943 5755 (−109266)–193534 8315 3742–25553
hydrogen ion 51.8 1.39–120 8.62 0.106–73.4 0.186 0.026–0.578 0.0018 0.0007–0.0237 (−0.063) (-3.97)–0.261 66.6 1.97–194 8.64 0–73.6
nitrate N 9.31 1.21–38 2.79 0.31–13.9 0.123 0.003–0.410 0.003 0–0.043 0 (−0.63)–0.02 13.2 2.54–46.6 2.85 0.45–13.9
ammonium N 4.1 0.87–18.9 1.14 0.14–7.49 0.319 0.127–0.664 0.019 0–0.155 0 (−4.04)–12.3 5.23 1.79–17.8 1.18 0.2–7.58
specific conductance 79 11–247 27 2–124 12877 4433–30025 638 6.49–6080 1360 (−17183)–23778 7291 (−3411)–32508 6540 2265–15015
a
All values in bold are median values for open-water periods from 1979 to 1998.
b
All values in italics represent the range of open-water period values.
c
All values in parenthesis are negative numbers.
Fig. 5. Solute concentrations for the most abundant cations in the pond of Wetland P1 during the open-water periods, 1979–1998, expressed in
milliequivalents per liter to show relative contribution to total dissolved constituents in the pond. Data for each date include values for multiple
transects for dates when more than one sample was collected for laboratory analysis.
chloride, and sulfate input to the pond (Table 4). In general, in each open-water period, groundwater accounted for 99 percent of the
calcium, magnesium, sodium, and sulfate input, greater than 96 percent of potassium input, and greater than 90 percent of chloride
input. The remainder of the input to the pond of these solutes from external sources was provided by rainfall on the pond surface and
runoff. Open-water periods in which the volume of runoff was larger than the volume of rainfall input to the pond (1990, 1992, and
1993) were also the years when runoff contributed as much or more of the solute input relative to rainfall for these solutes. One
exception occurred in 1995 when runoff contributed more chloride than rainfall even though the volume of water contributed by
runoff was less than for rainfall. In April of 1995 the concentration of chloride in rainfall used to calculate chloride input from runoff
was nearly three times the concentration during subsequent months, and the volume of runoff in April was over four times the volume
of rainfall in the same month.
Rainfall was the major external source of input to the pond during open-water periods for nitrate nitrogen, ammonium nitrogen,
and hydrogen ions (Table 4). The input of rainfall was less than 50 percent of total input for nitrate nitrogen and ammonium nitrogen
only in 1990, 1992, and 1993 when runoff contributed the most volume to the pond. Only in 1992 and 1993 was less than 50 percent
of the hydrogen ion input accounted for by rainfall, the majority of the remainder coming from runoff. In years when rainfall was the
major external source of solute input of nitrogen and hydrogen ion, the relative contribution of rainfall input of nitrogen and
hydrogen ion from one open-water period to the next was variable: 59–95 percent for nitrate nitrogen, 59–92 percent for ammonium
nitrogen, and 52–95 percent for hydrogen ion (Supplementary material Table A1).
Solute concentrations in the pond result from the combination of solute gains and losses, coincident with changes in pond volume
Fig. 6. Solute concentrations for the most abundant anions (sulfate and chloride) and acid-neutralizing capacity (ANC) in the pond of Wetland P1
during the open-water periods, 1979–1998. Data are expressed in milliequivalents per liter to show relative importance of each. Data for each date
include values for multiple transects for dates when more than one sample was collected for laboratory analysis.
14
as the relation between water gain and loss vary. The seasonal and year-to-year variation in concentrations of solutes contributing the
most to salinity in the pond are shown in Figs. 5 and 6. Concentrations of the most abundant solutes generally increased during the
open-water periods as pond water levels declined. Concentrations overall were greater in years when pond levels were lower than 1 m
before 1993, before pond levels showed pronounced increases. Magnesium and sulfate were the most abundant cations and anions in
the pond throughout almost the entire study period, which reflects their relation in groundwater flowing into the pond. Also, data in
Figs. 5 and 6 include concentration values obtained from each transect within the pond on each day water was collected for analysis.
The fact there was some within-pond variability in concentration on a given day is more readily apparent for magnesium and sulfate
(Figs. 5 and 6), but occurred for all solutes. The gradual rise in major solute concentration after 1993 reflects increased solute input
from groundwater in succeeding years (Supplementary material Table A1).
Dissolved nitrate nitrogen and ammonia nitrogen concentrations varied the most of all the major solutes on a given day (median
standard deviations were 45 percent for nitrate nitrogen and 12 percent for ammonia nitrogen). Variation within a month and from
year-to-year exhibited no readily identifiable seasonal pattern. Nitrate nitrogen concentrations commonly were below detection
limits, and otherwise generally less than ammonia nitrogen concentration. These generally low concentrations of nitrate have been
observed in ponds of other prairie wetlands in response to uptake by biota (Barica, 1974; Neely and Baker, 1989).
Overall, the sum of external solute inputs was greater than solute mass loss by flow from the pond to groundwater (Table 4).
Measured change in solute mass in the pond during the open-water period commonly was less than the difference between all
external inputs and solute mass loss by flow from the pond to groundwater. Some of the difference between measured change in mass
of solute in the pond and external inputs and loss of solutes may simply be due to uncertainty in measurement of the solute budget.
Uncertainty in the solute budget was determined using the method used to calculate cumulative error in the water budget using Eq.
(1). Based on median values, the residuals of the solute balances were larger than the cumulative error in the solute balances for all
solutes except alkalinity (Table 4).
Based on the many days when water was collected for solute analysis from more than one transect, some of the uncertainty in the
solute budgets may be due to the variability of solute concentrations in the pond on a given day. The standard deviation per day
expressed in terms of percent of mean concentration on that day is one measure of variability in the mass in the pond. The medians of
those percentages for all days in open-water periods from 1979 to 1998 when data were available from more than one transect were:
Ca 3.3 percent, Mg 2.5 percent, Na 2.3 percent, K 2.6 percent, Cl 4.1 percent, SO4 1.9 percent, alkalinity 2.4 percent, hydrogen ion 25
percent, nitrate nitrogen 45 percent, and ammonium nitrogen 12 percent. In years when the solute budget residual was larger than
the cumulative error, some of the residual unaccounted by cumulative error might be due simply to variability in concentration. The
percentages noted for the solutes above, however, were all smaller than the percentage of variability estimated from the mass of
solute in excess of the cumulative error divided by the solute mass of the pond at the start of the open water period except for
alkalinity (1987, 1996, and 1997). Therefore variability within the pond, based on concentration values in each of the separate
transects, likely did not account for the budget residual mass of solute unexplained by uncertainty in the solute budgets.
Even though in general between 1979 and 1998 some imbalance in the solute budgets of the pond were unaccounted for by
measured inputs and outputs, this was not the case every year. For some years, the cumulative error in open-water period solute
budgets was larger than the absolute value of the solute budget residual, indicating all solute gains and losses were accounted for
within the uncertainty in the budget. However, years in which this occurred were not always coincident for all solutes: Ca
1990–1994, and 1996; Mg 1989, 1992, 1994–1997; Na 1980 and 1981, 1992, 1994–1997; K 1983 and 1987; Cl 1989–1997; SO4
1992, 1996 and 1997; alkalinity 1980 and 1981, 1990, 1992, 1995; ammonium nitrogen 1993 and 1995. In addition, not all of these
years coincided with years in which the water budget was accounted for within measurement uncertainty (1986, 1990–1995).
Closure of individual solute budgets within uncertainty therefore was not simply a function of closure of water budgets.
Among the possible mechanisms that might explain the imbalance in the solute budgets for the pond is that flow of water from the
pond into groundwater may be higher than calculated. This could occur if the observation wells used to estimate the groundwater
gradient around the wetland were not close enough to the pond to capture the transpiration-driven water table- depression around
the periphery of the wetland pond (Rosenberry and Winter, 1997). The fact that mass export occurs as the pond shrinks is evident
from year-to-year changes in the open-water period loss of sulfate by flow from the pond to groundwater (Fig. 7), and for calcium,
magnesium, sodium, potassium, and chloride (Supplementary material Table A1) being most pronounced in dry years. These were
the periods when the loss of water from the pond to groundwater represented approximately 20 percent or greater of the total loss of
water from the pond (1989–1992). This finding also is consistent with the fact that in the preceding years the likely imbalance in the
water budget was due to underestimation of evapotranspiration when emergent vegetation was present in the pond.
Another possible source of imbalances that were greater than could be accounted for within uncertainty in the budgets is
overestimation of solute contributions to the pond from groundwater. Median concentrations among groundwater collected from the
observation wells were not uniform (Table 5). Groundwater collected at the periphery of the Wetland P1 pond in June of 1988 by
Arndt and Richardson (1993) varied with respect to electrical conductivity between 4.6–8.7 dS m−1 (4600–8700 microSiemens
cm−1) in the shoreline segment associated with well 13, and between 3.4–6.3 dS m−1 (3400–6300 microSiemens cm−1) in the
shoreline segment associated with well 25. Specific conductance in water from well 25 varied from 4680 to 6320 microSiemens cm−1
(n = 9) (based on seven in-situ probe measurements and two laboratory analyses), values within the range observed at the pond
periphery by Arndt and Richardson (1993). Specific conductance in water from well 13 varied between 10600 and 15200 micro-
Siemens cm−1 (n = 17), values approximately double the values observed at the pond periphery by Arndt and Richardson (1993).
Shoreline segments other than the segment associated with well 13, however, contributed most of the solutes from groundwater to
the pond. Assuming concentrations in the shoreline segment associated with well 13 were half of what was observed did not sig-
nificantly decrease solute inputs from groundwater, therefore the budgets reported here did use the laboratory solute data from well
15
Fig. 7. Sulfate flux from the pond to groundwater, 1979–1998.
13. In addition, concentrations in groundwater from the wells, including well 13, were within the range observed in groundwater
collected from lysimeters along transects perpendicular to the shoreline in the vicinity between wells 13 and 25, and in the vicinity of
well 17 between 1988 and 1990 (Biondini and Arndt, 1993). Such information was consistent with the assumption that solute input
from groundwater to the pond based on the concentrations in Table 5 was representative of conditions at the site.
Chloride is considered to be unaffected by reactions within the pond at the concentrations observed during the study. In nine
years of the study (1989–1997) the residual of the open-water period budget for chloride was accounted for within the uncertainty in
the budget (Supplementary material Table A1). Based on this observation, no adjustments in the solute budgets were made to account
for years when there was a disparity between inputs, export, and change in mass in the pond. Instead, the disparity was assumed to
reflect some unaccounted for export. One possible export mechanism is evapotranspiration-induced movement of water and solutes
out of the pond, at the periphery of the pond as the pond surface shrinks, such as documented by Arndt and Richardson (1993). Such
variations were not quantified by the gradients evident from the wells located outside of the immediate pond periphery.
Loss of solute mass within the pond can occur due to chemical precipitation or gas flux out of the pond. One line of evidence for
loss by chemical precipitation is provided by the molar ratio of magnesium and calcium concentrations in the pond (Fig. 8). In many
years, the ratio generally increased through the open-water period, indicating a loss of calcium relative to magnesium within the
pond. Geochemical models of solute interaction indicated supersaturation conditions exist in the pond for calcite (Goldhaber et al.,
2014), therefore calcite precipitation could account for the loss of calcium relative to magnesium in the pond.
Magnesium and sulfate commonly were the most abundant cation and anion, respectively, based on their concentration in
milliequivalents per liter (Fig. 5 and 6). Their relation to one another, however, varied during the study (Fig. 9). The molar ratio of
magnesium to sulfate was generally less than one in the wells around the pond. The monthly ratios in atmospheric deposition varied
from 0.04 to 1.3; 95 percent of the values were less than 0.69; values greater than 1 were recorded in only three months between
1979 and 1998, and not after 1992. The noticeable increase in the ratio in the pond in 1993 likely indicates a change in sulfate
relative to magnesium outside of that expected from the solute sources of atmospheric deposition and groundwater flow to the pond.
Some change in magnesium does occur through precipitation associated with calcite deposition in the pond (Goldhaber et al., 2014).
Even so, the distinction between 1993 and years immediately thereafter, and years prior to 1993 may reflect inundation of all
emergent vegetation and subsequent decomposition. Loss of sulfate due to sulfate reduction occurs in the pond as documented in
Goldhaber et al. (2014), which requires a carbon source, such as decomposing vegetation. Decomposition of above-ground biomass in
Wetland P1 does occur, with a rate of 50 percent decomposition over 114 months (Biondini and Arndt 1993). The persistence of
ratios above 1 from 1993 onward may reflect the time involved to achieve complete decomposition of above-ground biomass in-
undated in 1993.
In general the budget residuals for nitrate and ammonia nitrogen were larger than could be accounted for by cumulative error
(Supplementary material Table A1). The budget imbalances for nitrate and ammonia nitrogen are consistent with biological removal
mechanisms within wetland ponds noted in other prairie wetland studies (Barica 1974; Neely and Baker, 1989), removal mechanisms
that were beyond the scope of this study. Similar large budget residuals for ammonia nitrogen also were reported by Murkin et al.
(2000) in experimental wetland ponds, a function of uptake by biota.
Calculating groundwater fluxes to and from the pond independently using multiple shoreline segments, rather than as a net value,
enabled unambiguous examination of changes in solute input from groundwater during the years after 1993. The volume of
groundwater flow into the pond in the open-water periods beginning in 1993 was at least twice that of years prior to 1993. This also
was reflected in solute input from groundwater to the pond for calcium, alkalinity (Fig. 10), and potassium (Supplementary material
Table A1). Other solutes also were elevated in input particularly in 1993, such as sulfate, but were not consistently elevated in
16
J.W. LaBaugh et al.
Table 5
Summary of concentrations in groundwater in wells used to determine groundwater input to the pond. All values are medians when n > 1.
Well number (number of observations)
12 (13) 13–18 (18) 14 (2) 15 (1) 16 (12) 17 (2) 18 (3) 20 (2) 23 (1) 25 (2) 26 (2) 27 (3) 28 (2)
solute units Median value
calcium mg L−1 112 483 450 500 521 365 400 525 460 475 230 98 215
magnesium mg L−1 58 1998 530 250 212 2900 300 745 440 495 135 34 86
sodium mg L−1 75 805 425 190 130 1650 310 335 150 345 38 11 58
17
potassium mg L−1 10.1 53 24 16 21.6 32 21 16.5 13 16 3.9 1.5 13.5
chloride mg L−1 5.05 10.5 29 7 30 193 55 115 9.8 17.5 12 1 7.95
sulfate mg L−1 280 12875 3613 2300 2500 15000 3000 3964 3000 3700 850 150 515
alkalinity meq L−1 4.98 12.25 9.59 8.15 5.9 12.04 9.05 5.92 6.61 6.97 5.8 5.22 9.53
nitrate N mg L−1 0.013 0.015 . . 0.04 0.116 0.012 . . 0.331 . 0.089 0.449
ammonium N mg L−1 0.262 0.77 . . 0.6 0.49 0.121 . . 0.21 0.107 0.039 0.132
pH 7.8 7.4 7.3 7.3 7.5 7.3 7.0 7.4 7.1 7.25 7.55 7.7 7.4
H ion meq L−1 3.98E-08 1.58E-08 5.01E-08 5.01E-08 3.57E-08 5.55E-08 0.0000001 3.98E-07 7.94E-08 5.96E-08 2.99E-08 2.00E-08 4.97E-08
specific conductance μS cm−1 1160 12500 4460 3720 3750 15750 4910 5140 4410 5365 1870 720 1670
Fig. 8. Magnesium–calcium molar ratio in the pond of Wetland P1 during the open-water periods, 1979–1998. Gaps in the line represent periods
when the pond was covered with ice.
Fig. 9. Magnesium–sulfate molar ratio in the pond of Wetland P1 during the open-water periods, 1979–1998. Gaps in the line represent periods
when the pond was covered with ice.
subsequent years. The difference in pattern, in part, reflected the contribution of solutes from wells with relatively smaller con-
centrations of those solutes, than the wells they substituted for in the calculations due to expansion of the pond (Table 5).
5. Discussion
5.1. Water budgets
The relative contribution of water sources to and water losses from the pond of Wetland P1 were similar to other water budgets
for closed-basin prairie-pothole wetland ponds (Hayashi et al., 1988a; Shjeflo, 1968; Woo and Rowsell, 1993) and similar to annual
water balances based on a coupled surface/subsurface model for the wetland pond (Carroll et al., 2005). Rainfall directly on pond
surfaces and evapotranspiration were the largest sources and losses of water during open-water periods. Snowmelt was important as a
source of water when ponds thawed in spring and provided sufficient water to enable the pond to persist, even though water losses
commonly exceeded gains through much of the open-water period. Yet, snowmelt contribution to ponds infrequently exceeded gains
during the open-water periods from rainfall and runoff. Runoff commonly contributed less water to ponds than rain falling directly on
the pond surface, and most often was associated with rain falling on frozen upland soil after the pond had thawed, or during episodes
of intense rainfall, or rain persisting for more than a day (Eisenlohr et al., 1972). Runoff, however, was variable such that years with
similar rainfall could differ substantially in the amount of runoff derived from rain falling on the uplands (Eisenlohr et al., 1972;
18
Fig. 10. Open-water period input of calcium and alkalinity to the pond, 1979–1998.
Heagle et al., 2007; Woo and Rowsell, 1993). Groundwater flow contributed the least to water gains and losses, and has been reported
as a net value in other water budget studies (Hayashi et al., 1988a; Shjeflo, 1968; Woo and Rowsell, 1993)
The few published water budgets of prairie-pothole wetland ponds all reported net seepage outflow for all ponds and years of
study (Hayashi et al., 1988a; Shjeflo, 1968; Woo and Rowsell, 1993), except in ponds intermittently connected by a stream in a year
of historically large snow accumulation (Brannen et al., 2015). Groundwater inflow to the pond of Wetland P1 exceeded loss of pond
water to groundwater in all years except 1991 and 1992, when snowmelt was absent or minimal. The predominance of water gain
from groundwater inflow, relative to pond outflow to groundwater, compared to other studies may be related to the topographic and
hydrogeological setting of Wetland P1 that favored groundwater discharge to the pond.
Few water budget studies of lakes and wetlands include determination of the budget residual, the difference between measured
water gains, losses, and change in storage, including measurement uncertainty or error. Hayashi et al. (1988a) noted uncertainty
existed in the water budget, but did not include a value for the budget residual. Only Shjeflo (1968) and Brannen et al. (2015)
included quantification of the budget residual in their water budgets of prairie-pothole wetland ponds. For the budgets provided by
Shjeflo (1968), Eisenlohr et al. (1972) noted the residual was an indicator of the uncertainty in the budget computations. Brannen
et al. (2015) used the relation between the budget residual and the water-table depth to indicate the residual may represent net
groundwater flow into the ponds, as well as error associated with the water budget. Our approach for the pond of Wetland P1, using
determination of the cumulative error, or uncertainty, in the measured budgets based on first order error analysis, was not used in
other published budgets. Such determination enabled the identification of periods when the budget residual was indistinguishable
from inherent uncertainty in measurements, and periods when the budget residual indicated an imbalance in the water budget larger
than could be accounted for by uncertainty in measurement alone. In the pond of Wetland P1, the budget imbalance in the latter case
19
was likely associated with under-representation of water loss by transpiration or periods when hydraulic gradients indicated by the
wells were not indicative of gradients closer to the pond. Also, as noted by Hayashi et al. (2016), at some point when the pond surface
area declines, change in pond water level exceeds that expected based on calculated evapotranspiration because of the increasing
influence of water lost at the pond perimeter relative to loss from the pond surface.
5.2. Solute budgets
The fact that change in sulfate mass in the pond of Wetland P1 was not always accounted for by the difference between external
solute inputs and losses from the pond to groundwater is consistent with other solute budgets for closed-basin prairie-pothole wetland
ponds (Hayashi et al., 1988b; Heagle et al., 2007). In the study of the pond of a recharge wetland (S109) at the St. Denis wetland site
in Saskatchewan, Canada, sulfate reduction accounted for most of the change in concentration of sulfate in the pond (Heagle et al.,
2007). Sulfate reduction also was an important process removing sulfate from the pond of nearby Wetland 50, a discharge wetland
(Heagle et al., 2013), based on observations of the change in mass in sulfate in the pond relative to change in volume. Changes in the
stable sulfur isotope of sulfate (δS34
SO4) measured in the pond of Wetland P1 (Goldhaber et al., 2014; Mills et al., 2011) indicate sulfate
reduction did occur in the pond. Sulfate reduction likely accounts for the loss of sulfate needed to reconcile the difference between
change in sulfate in the pond and the imbalance between external solute inputs and loss observed in the open-water period of all years
except 1992, 1996, and 1997.
Export of sulfate by movement of pond water into wetland sediments at the edge of the pond of discharge Wetland 50 as it dries
was documented by Heagle et al. (2013) over 19 years of observation. This export was a mechanism associated with long-term
accumulation within the wetland sediments. Dry periods also were associated with larger sulfate export from the pond of Wetland P1
to groundwater than in wetter periods. As the pond shrank, the export gradient was into wetland sediments rather than the uplands.
The fact that some segments of the pond shoreline could be locations of solute export at the same time other segments of shoreline
indicated solute import to the pond suggests the areal extent of solute accumulation in the subsurface may not be uniform.
Regarding other solutes, Heagle et al. (2007) found dissolution of magnesium-calcite in wetland soils added calcium and mag-
nesium to the pond of recharge wetland S109. In the case of Wetland P1, groundwater and runoff solute input to the pond was greater
than loss by export from the pond to groundwater. Loss of calcium and magnesium from the pond was indicated from the budget
imbalance in the open-water period for many years, despite the overall increase in concentration within the pond from the beginning
to the end of most periods. Net precipitation of magnesium carbonate in the pond was documented by Goldhaber et al. (2014).
Heagle et al. (2013) used chloride to balance the water budget of the pond of S109. The resulting water budget was used to
determine the mass balance for other solutes. Because open-water period chloride budgets for the pond of Wetland P1 were ac-
counted for within budget measurement uncertainty in nearly half of the years in our study, we assumed the years of imbalance
indicated unmeasured chloride export, likely due to evaporation-induced flow of water out of the pond at the periphery. Hayashi
et al. (1988b) observed a period when chloride concentrations declined in the pond of wetland S109, a decline that was not consistent
with physical processes. Such changes were attributed to uptake by plants as described by Hutchinson (1957). Such uptake might also
be a factor in accounting for those open-water periods for Wetland P1 when the change in mass in the pond was less than expected
based on external solute input and export. Such occurrences were observed mainly prior to 1989 when living emergent vegetation
occupied some of the areal extent of the pond throughout the open-water period. Seasonal measurement of the chloride content of
emergent vegetation to examine uptake, however, was not part of our study.
Wetter conditions have resulted in increased solute mass in the ponds of wetlands such as Wetland 50 at the St. Denis site (Heagle
et al., 2013) and the pond of Wetland P1 because of solute input to the ponds. Yet, changes in salinity due to wetter conditions are not
uniform in prairie-pothole wetlands, being dependent on their hydrological settings (Nachshon et al., 2014; LaBaugh et al., 2016)
either as recharge, flow-through, or discharge ponds, as well as the presence or absence of surface-water flow between ponds (Shaw
et al., 2012). For ponds with increased solute mass during wetter conditions, such as in Wetland P1 (LaBaugh et al., 2016), declining
water levels under drier conditions have the potential to increase salinity of the pond in the absence of a substantial loss mechanism,
as hypothesized by Nachshon et al. (2014). At the Cottonwood site, however, previous fluctuations in pond levels of Wetland P1 have
not been associated with higher salinities as the pond declined from water levels similar to that encountered after 1993 (LaBaugh
et al., 2016). Based on the solute budgets for the pond, as the pond dries some solutes are lost by flow from the pond to groundwater
and wetland sediments, and some are lost due to chemical precipitation or gaseous flux out of the pond. The increased loss of water at
the periphery of declining ponds noted by Hayashi et al. (2016) and the increased salinization and storage at the pond periphery as
pond levels decline as observed by Arndt and Richardson (1993) also must have been sufficient in the past to moderate the salinity of
the pond through episodes of deluge and drought.
6. Conclusions
Determination of groundwater input to a wetland pond and flow of water from the pond to groundwater revealed dynamic aspects
of groundwater interaction with prairie-pothole wetland ponds important to the water and solute budgets of the pond during the
open-water period. In drier years, as the pond shrank, loss of water from the pond by flow to groundwater became more important
than in other years. Such flow out of the pond also provided for export of solutes to wetland sediments when the pond edge receded
from the upland. Groundwater flow into the pond represented the largest source of many solutes (Ca, Mg, Na, K, Cl, SO4, alkalinity)
even though such flow commonly represented the smallest contribution to water input to the pond relative to rainfall or runoff.
During the wetter periods, as the pond increased in size thereby expanding the area of groundwater input to the pond, the relatively
20
greater influx of Ca and alkalinity was consistent with the presence of readily weathered calcite in the soils (Arndt and Richardson,
1993) and subsequent flow into the pond. The importance of solute input to the pond and solute loss from the pond as independent
processes when examining solute dynamics indicates determination of groundwater input and output separately has benefits over
determination of groundwater interaction with the pond as a net water flux alone. Measurement uncertainty in the solute budgets
also enabled distinction between years when change in mass in the pond was accounted for by external inputs and outputs, and when
change in mass in the pond was not accounted for by external inputs and outputs, thereby indicating some unmeasured loss in mass
took place.
Pond permanence, size and salinity all affect wetland habitats in the Prairie Pothole Region. It is important to understand how
these factors change under varying climate conditions ranging from drought to deluge, as documented here. Our study indicates the
importance of taking into account groundwater as a factor affecting wetland salinity, particularly in closed-basin wetlands receiving
groundwater discharge. Therefore it is important to know the relation of the pond to the adjacent groundwater system before
projecting what salinity-dependent habitat changes might occur in individual wetland basins as conditions either become drier or
wetter in the future.
Acknowledgements
The ability to examine processes over such a long term were made possible because of the work of our colleagues George Swanson
and Vyto Adomaitis, who began ecological studies of the Cottonwood Lake area wetlands site in 1967, and the detailed hydrological
studies started by Tom Winter in 1978. Data from Cottonwood Lake area wetlands (for example, Mushet et al., 2016a, 2016b; Mushet
and LaBaugh, 2017; Mushet et al., 2017a, 2017b) are available through the Missouri Coteau Wetland Ecosystem Observatory
(https://www.sciencebase.gov/catalog/item/52f0ffd9e4b0f941aa181fc6) and maintained through funding received from the U.S.
Geological Survey’s Climate Research and Development Program. We thank Greg Noe, Sandy Cooper and two anonymous reviewers
for constructive comments and suggestions on previous versions of the manuscript. Any use of trade, firm, or product names is for
descriptive purposes only and does not imply endorsement by the U.S. Government.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.ejrh.2018.03.
003.
References
Arndt, J.L., Richardson, J.L., 1989. Hydrology, salinity, and hydric soil development in a recharge-flowthrough-discharge prairie-pothole wetland system. Soil Sci. Soc.
Am. J. 53, 848–855. http://dx.doi.org/10.2136/sssaj1989.03615995005300030037x.
Arndt, J.L., Richardson, J.L., 1993. Temporal variations in the salinity of shallow groundwater from the periphery of some North Dakota wetlands (USA). J. Hydrol.
141, 75–105. http://dx.doi.org/10.1016/0022-1694(93)90045-B.
Barica, J., 1974. Some observations on internal cycling, regeneration and oscillation of dissolved nitrogen and phosphorus in shallow self-contained lakes. Arch.
Hydrobiol. 73, 334–360.
Barica, J., 1975, Geochemistry and nutrient regime of saline eutrophic lakes in the Erickson-Elphinstone district of southwestern Manitoba, Environment Canada,
Fisheries and Marine Service Technical Report 511, 82 p.
Batt, B.D., Anderson, M.G., Anderson, C.D., Caswell, F.D., 1989. In: van der Valk, A.G. (Ed.), The Use of Prairie Potholes by North American Ducks. Iowa State
University Press, Northern Prairie Wetlands, Ames, Iowa, pp. 204–227.
Berthold, S., Bentley, L.R., Hayashi, M., 2004. Integrated hydrogeological and geophysical study of depression-focused groundwater recharge in the Canadian prairies.
Water Resour. Res. 40 (W0605), 14. http://dx.doi.org/10.1029/2003WR002982.
Bierhuizen, J.F.H., Prepas, E.E., 1985. Relationships between nutrients, dominant ions, and phytoplankton standing crop in prairie saline lakes. Can. J. Fish. Aquat. Sci.
42, 1588–1594. http://dx.doi.org/10.1139/f85-199.
Biondini, M.E., Arndt, J.L., 1993. The Biogeochemistry of Carbon, Nitrogen, and Sulfur Transformations in Seasonal and Semipermanent Wetlands. North Dakota
Water Resources Research Institute Research Project Technical Completion Report ND90-05, 83 p.
Brannen, R., Spence, C., Ireson, A., 2015. Influence of shallow groundwater-surface water interactions on the hydrological connectivity and water budget of a wetland
complex. Hydrol. Process. 29, 3862–3877. http://dx.doi.org/10.1002/hyp.10563.
Buso, D.C., Likens, G.E., LaBaugh, J.W., Bade, D., 2009. Nutrient dyamics. In: Winter, T.C., Likens, G.E. (Eds.), Mirror Lake: Interactions Among Air, Land, and Water.
University of California Press, pp. 69–203.
Carroll, R., Pohlll, G., Tracy, J., Winter, T., Smith, R., 2005. Simulation of a semipermanent wetland basin in the Cottonwood Lake Area, East-Central North Dakota. J.
Hydrol. Eng. 10, 70–84. http://dx.doi.org/10.1061/(ASCE)1084-0699(2005)10:1(70).
Deutschman, M.R., Ell, M.J., 1986. Ambient Air Quality, Precipitation Chemistry and Atmospheric Deposition in North Dakota, 1980-1984, Research Contract Number
YA-553-CTI-1084 Final Report, October 1986. U.S. Department of the Interior Bureau of Land Management 214 p.
Eisenlohr, W.S., et al., 1972. Hydrologic Investigations of Prairie Potholes in North Dakota, 1959–68, U.S. Geological Survey Professional Paper 585-A. 102 p, 3 plates.
Euliss, N.H., LaBaugh, J.W., Fredericksen, L.H., Mushet, D.M., Laubhan, M.K., Swanson, G.A., Winter, T.C., Rosenberry, D.O., Nelson, Richard D., 2004. The wetland
continuum: a conceptual framework for interpreting biological studies. Wetlands 24, 448–458. http://dx.doi.org/10.1672/0277-5212(2004)024[0448:TWCACF]
2.0.CO;2.
Euliss, N.H., Mushet, D.M., Newton, W.E., Otto, C.R.V., Nelson, Richard D., LaBaugh, J.W., Scherff, E.J., Rosenberry, D.O., 2014. Placing prairie-pothole wetlands
along spatial and temporal continua to improve integration of wetland function in ecological investigations. J. Hydrol. 513, 490–503. http://dx.doi.org/10.1016/
j.hydrol.2014.04.006.
Ficken, J.H., 1967. Winter loss and spring recovery of dissolved solids in two prairie-pothole ponds in North Dakota, in Geological Survey Research 1967. U.S.
Geological Survey Professional Paper 575-C. pp. C228–C231.
Methods for determination of inorganic substances in water and fluvial sediments. In: Fishman, M.J., Friedman, L.C. (Eds.), Techniques of Water Resource
Investigations. U.S. Geological Survey, Book 5, (Chapter A1).
Goldhaber, M.B., Mills, C.T., Morrison, J.M., Stricker, C.A., Mushet, D.M., LaBaugh, J.W., 2014. Hydrogeochemistry of prairie pothole region wetlands: role of long-
term critical zone processes. Chem. Geol. 387, 170–183. http://dx.doi.org/10.1016/j.chemgeo.2014.08.023.
21
Greenberg, A.E., Clesceri, L.S., Eaton, A.D. (Eds.), 1992. Standard Methods for the Examination of Water and Wastewater, 18th edition. American Public Health
Association Washington, D.C.
Hammer, U.T., 1978. The saline lakes of saskatechewan, III: chemical characterization. Int. Revue Gesamten Hydrobiol. 63, 311–335. http://dx.doi.org/10.1002/iroh.
19780630303.
Hamon, W.R., 1961. Estimating potential evapotranspiration. Proc. Am. Soc. of Civil Eng. 87, 107–120.
Hayashi, M., van der Kamp, G., Rudolph, D.L., 1988a. Water and solute transfer between a prairie wetland and adjacent uplands, 1. Water balance. J. Hydrol. 207,
42–55. http://dx.doi.org/10.1016/S0022-1694(98)00098-5.
Hayashi, M., van der Kamp, G., Rudolph, D.L., 1988b. Water and solute transfer between a prairie wetland and adjacent uplands, 2 Chloride cycle. J. Hydrol. 207,
56–67. http://dx.doi.org/10.1016/S0022-1694(98)00099-7.
Hayashi, M., van der Kamp, G., Rosenberry, D.O., 2016. Hydrology of prairie wetlands: understanding the integrated surface-water and groundwater processes.
Wetlands 36 (2), 237–254. http://dx.doi.org/10.1007/s13157-016-0797-9.
Heagle, D.J., Hayashi, M., van der Kamp, G., 2007. Use of solute mass balance to quantify geochemical processes in a prairie recharge wetland. Wetlands 27, 806–818.
http://dx.doi.org/10.1672/0277-5212(2007)27[806:UOSMBT]2.0.CO;2.
Heagle, D.J., Hayashi, M., van der Kamp, G., 2013. Surface-subsurface salinity distribution and exchange in a closed basin prairie wetland. J. Hydrol. 478, 1–14.
http://dx.doi.org/10.1016/j.jhydrol.2012.05.054.
Huang, S., Young, C., Abdul-Aziz, O.I., Dahal, D., Feng, M., Liu, S., 2013. Simulating the water budget of a prairie potholes complex from LiDAR and hydrological
models in North Dakota, U.S.A. Hydrol. Sci. J. 58, 1434–1444. http://dx.doi.org/10.1080/02626667.2013.83419.
Hutchinson, G.E., 1957. A Treatise on Limnology: Volume 1 Geography, Physics, and Chemistry. John Wiley and Sons, New York 1015 p.
Johnson, W.C., Boettcher, S.E., Poiani, K.A., Guntenspergen, G., 2004. Influence of weather extremes on the water levels of glaciated prairie wetlands. Wetlands 24,
385–398. http://dx.doi.org/10.1672/0277-5212(2004)024[0385:IOWEOT]2.0.CO;2.
Johnson, W.C., Millett, B.V., Gilmanov, T., Voldseth, R.A., Guntenspergen, G.R., Naugle, D.E., 2005. Vulnerability of northern prairie wetlands to climate change.
Bioscience 55, 863–872. http://dx.doi.org/10.1641/bio.2005.55?.
Johnson, W.C., Werner, B., Guntenspergen, G.R., Voldseth, R.A., Millett, B., Naugle, D.E., Tulbure, M., Carroll, R.W.H., Tracy, J., Olawsky, C., 2015. Prairie wetland
complexes as landscape functional units in a changing climate. Bioscience 60, 128–140. http://dx.doi.org/10.1525/bio.2010.60.2.7.
Jones, B., Deocampo, D., 2003. Geochemistry of saline lakes. In: In: Drever, J.I., Holland, H.D., Turkian, K.K. (Eds.), Treatise on Geochemistry, vol. 5. Elsevier, pp.
393–424. http://dx.doi.org/10.1016/B0-08-043751-6/05083-0.
LaBaugh, J.W., Winter, T.C., Adomaitis, V.A., Swanson, G.A., 1987. Hydrology and chemistry of selected prairie wetlands in the Cottonwood Lake area, Stutsman
County, North Dakota, 1979-82. U.S. Geological Survey Professional Paper 1431. 26 p. https://pubs.er.usgs.gov/publication/pp1431.
LaBaugh, J.W., Rosenberry, D.O., Winter, T.C., 1995. Groundwater contribution to the water and chemical budgets of Williams Lake, Minnesota, 1980-1991. Can. J.
Fish. Aquat. Sci. 52, 754–767. http://dx.doi.org/10.1139/f95-075.
LaBaugh, J.W., Winter, T.C., Swanson, G.A., Rosenberry, D.O., Nelson, R.D., Euliss, N.H., 1996. Changes in atmospheric circulation patterns affect mid-continent
wetlands sensitive to climate. Limnol. Oceanogr. 41, 864–870. http://dx.doi.org/10.4319/lo.1996.41.5.0864.
LaBaugh, J.W., Mushet, D.M., Rosenberry, D.O., Euliss, N.H. Jr., Goldhaber, M.B., Mills, C.T., Nelson, R.D., 2016. Changes in pond water levels and surface extent due
to climate variability alter solute sources to closed-basin prairie-pothole wetland ponds 1979-2012. Wetlands 36 (Suppl. 2), 343–355. http://dx.doi.org/10.1007/
s13157-016-0808-x.
LaBaugh, J.W., 1986. Wetland ecosystem studies from a hydrologic perspective. Water Resour. Bull. 22, 1–10. http://dx.doi.org/10.1111/j.1752-1688.1986.
tb01853.x.
Laird, K.R., Cumming, B.F., Wunsam, S., Rusak, J.A., Oglesby, R.J., Fritz, S.C., Leavitt, P.R., 2003. Lake sediments record large-scale shifts in moisture regimes across
the northern prairies of North America during the past two millennia. Proc. Nat. Acad. Sci. 100, 2483–2488. http://dx.doi.org/10.1073/pnas.0530193100.
Langbein, W.B., 1961. Salinity and hydrology of closed lakes; a study of the long-term balance between input and loss of salts in closed lakes. U.S. Geological Survey
Professional Paper 412. 20 p.
Last, W.M., 1992. Chemical composition of saline and subsaline lakes of the northern Great Plains, western Canada. Int. J. Salt Lake Res. 1, 47–76.
Leibowitz, S.G., Vining, K.C., 2003. Temporal connectivity in a prairie pothole complex. Wetlands 23, 13–25. http://dx.doi.org/10.1672/0277-5212(2003)
023[0013:tciapp]2.0.co;2.
Leibowitz, S.G., Mushet, D.M., Newton, W.E., 2016. Intermittent surface water connectivity: fill and spill vs. fill and merge dynamics. Wetlands 26 (2), 323–342.
http://dx.doi.org/10.1007/s13157-016-0830-z.
Lissey, A., 1971. Depression-focused groundwater flow patterns in Manitoba. Geological Association of Canada Special Paper 9. pp. 333–341.
Liu, G., Schwartz, F.W., 2011. An integrated observational and model-based analysis of the hydrologic response of prairie pothole systems to variability in climate.
Water Resour. Res. 47 (W02504). http://dx.doi.org/10.1029/2010WR009804. 15 p.
Luba, L.D., Henry, J.L., Hogg, T.J., Harder, H., 1988. Wind transport of salts in the Old Wives Lake area of Saskatchewan. Proceedings of Soils and Crops Workshop,
Saskatoon, February 1988 409–423.
Mather, J.R., 1978. The Climatic Water Budget in Environmental Analysis. Lexington Books, Lexington Massachusetts 239 p.
Mene, M.J., Durre, I., Korzeniewski, B., McNeal, S., Thomas, K., Yin, X., Anthony, S., Ray, R., Vose, R.S., Gleason, B.E., Houston, H.G., 2012. Global Historical
Climatology Network-Daily (GCHND-Daily), Version 3.
Meyboom, P., van Everdingen, R.O., Freeze, R.A., 1966. Patterns of groundwater flow in seven discharge areas in Saskatchewan and Manitoba. Canada Geol. Survey
Bull. 147 57 p.
Meyboom, P., 1966. Unsteady groundwater flow near a willow ring in hummocky moraine. J. Hydrol. 4, 38–62. http://dx.doi.org/10.1016/0022-1694(66)90066-7.
Miller, R.F., 1969. Differences in soil chemistry induced by evaporation and flow of ground water. U.S. Geological Survey Professional Paper 650-D. pp. D255–D259.
Mills, J.G., Zwarich, M.A., 1986. Transient ground water flow surrounding a recharge slough in a till plain. Can. J. Soil. Sci. 66, 121–134. http://dx.doi.org/10.4141/
cjss86-012.
Mills, C.T., Goldhaber, M.B., Stricker, C.A., Holloway, J.M., Morrison, J.M., Ellefsen, K.J., Rosenberry, D.O., Thurston, R.S., 2011. Using stable isotopes to understand
hydrochemical processes in and around a Prairie Pothole wetland in the Northern Great Plains, USA. Ap. Geochem. 26, S97–S100. http://dx.doi.org/10.1016/j.
apgeochem.2011.03.040.
Motz, L.H., Sousa, G.D., Annable, M.D., 2001. Water budget and vertical conductance for Lowry (Sand Hill) Lake in north, central Florida, USA. J. Hydrol. 250,
134–148. http://dx.doi.org/10.1016/S0022-1694(01)00434-6.
Murkin, H.R., van der Valk, A.G., Clark, W.R., 2000. Prairie Wetland Ecology: the Contribution of the Marsh Ecology Research Program, Ames, Iowa. Iowa State
University Press 413 p.
Mushet, D.M., LaBaugh, J.W., 2017. Cottonwood Lake Study Area – Water Chemistry – Wells – In Situ Measurements, U.S. Geological Survey Data Release. http://dx.
doi.org/10.5066/F7N58JJ2.
Mushet, D.M., Rosenberry, D.O., Euliss Jr, N.H., Solensky, M.J., 2016a. Cottonwood Lake Study Area – Water Surface Elevation, U.S. Geological Survey Data Release.
http://dx.doi.org/10.5066/F7707ZJ6.
Mushet, D.M., Solensky, M.J., Euliss Jr, N.H., 2016b. Cottonwood Lake Study Area – Specific Conductance, U.S. Geological Survey Data Release. http://dx.doi.org/10.
5066/F7BP00XQ.
Mushet, D.M., LaBaugh, J.W., Nelson, R.D., Euliss Jr, N.H., 2017a. Cottonwood Lake Study Area – Water Chemistry – Wetlands, U.S. Geological Survey data release.
http://dx.doi.org/10.5066/F7DN437S.
Mushet, D.M., LaBaugh, J.W., Winter, T.C., 2017b. Cottonwood Lake Study Area – Water Chemistry – Wells, U.S. Geological Survey data release. http://dx.doi.org/10.
5066/F78W3BH2.
Nachshon, U., Ireson, A., van der Kamp, G., Davies, S.R., Wheater, H.S., 2014. Impacts of climate variability on wetland salinization in the North American prairies.
Hydrol. Earth Syst. Sci. 18, 1251–1263. https://doi.org/10.519/hess-18-1251-2014.
22
Neely, R.K., Baker, J.L., 1989. In: van der Valk, A.G. (Ed.), Nitrogen and Phosphorus Dynamics and the Fate of Agricultural Runoff. Iowa State University Press,
Northern Prairie Wetlands, Ames, Iowa, pp. 92–131.
Parkhurst, R.S., Winter, T.C., Rosenberry, D.O., Sturrock, A.M., 1998. Evaporation from a small prairie wetland in the Cottonwood Lake area, North Dakota–an energy
budget study. Wetlands 18, 272–287. http://dx.doi.org/10.1007/BF03161663.
Petri, L.R., Larson, L.R., 1973. Quality of water in selected lakes of eastern South Dakota, South Dakota Water Resources Commission Report of Investigations, No.1.
55 p.
Poiani, K.A., Johnson, W.C., Swanson, G.A., Winter, T.C., 1996. Climate change and northern prairie wetlands: simulations of long-term dynamics. Limnol. Oceanogr.
41, 871–881. http://dx.doi.org/10.4319/lo.1996.41.5.0871.
Renton, D.A., Mushet, D.M., DeKeyser, E.S., 2015. Climate Change and Prairie Pothole Wetlands-Mitigating Water-level and Hydroperiod Effects Through Upland
Management, U.S. Geological Survey Scientific Investigations Report 2015-5004. http://dx.doi.org/10.3133/sir20155004. 21 p.
Aquatic ecosystems in semi-arid regions: implications for resource management. In: Robarts, R.D., Bothwell, M.L. (Eds.), National Hydrology Research Institute
Symposium Series 7. Saskatoon, Environment Canada. 375 p.
Rosenberry, D.O., Winter, T.C., 1997. Dynamics of water-table fluctuations in an upland between two prairie-pothole wetlands in North Dakota. J. Hydrol. 191,
266–289. http://dx.doi.org/10.1016/S0022-1694(96)03050-8.
Rosenberry, D.O., Stannard, D.I., Winter, T.C., Martinez, M.L., 2004. Comparison of 13 equations for determining evapotranspiration from a prairie wetlands,
Cottonwood Lake Area, North Dakota, U.S.A. Wetlands 24, 483–497. http://dx.doi.org/10.1672/0277-5212(2004)024[0483:COEFDE]2.0.CO;2.
Rosenberry, D.O., 2003, Climate of the Cottonwood Lake area, in Winter, T.C., ed., Hydrological, chemical, and biological characteristics of a prairie pothole wetland
complex under highly variable climate conditions – the Cottonwood Lake Area, east-central North Dakota: Denver, U.S. Geological Survey Professional Paper
1675, p. 25–34.
Rozkowski, A., 1969. Chemistry of ground and surface waters in the Moose Mountain area, southern Saskatchewan. Geological Survey of Canada Paper 67-9. 111 p.
Shaw, D.A., van der Kamp, G., Conly, F.M., Pietroniro, A., Martz, L., 2012. The fill-spill hydrology of prairie wetland complexes during drought and deluge. Hydrol.
Process. 26, 3147–3156. http://dx.doi.org/10.1002/hyp.8390.
Shjeflo, J.B., 1968. Evapotranspiration and the Water Budget of Prairie Potholes in North Dakota, U.S. Geological Survey Professional Paper 585-B. 49 p.
Sloan, C.E., 1972. Ground-water Hydrology of Prairie Potholes in North Dakota, U.S. Geological Survey Professional Paper 585-C. 28 p, 2 plates.
Stewart, R.E., Kantrud, H.A., 1971. Classification of natural ponds and lakes in the glaciated prairie region. Bureau of Sport Fisheries and Wildlife, Fish and Wildlife
Service, Resource Publication 92, Washington D.C. 57 p.
Swanson, G.A., Duebbert, H.F., 1989. In: van der Valk, A.G. (Ed.), Wetland Habitats of Waterfowl in the Prairie Pothole Region. Iowa State University Press, Northern
Prairie Wetlands, Ames, Iowa, pp. 228–267.
Swanson, G.A., Winter, T.C., Adomaitis, V.A., LaBaugh, J.W., 1988. Chemical Characteristics of Prairie Lakes in South-Central North Dakota–their Potential for
Impacting Fish and Wildlife. United States Department of the Interior Fish and Wildlife Service Technical Report 18, Washington D.C 44 p.
Swanson, G.A., 1978. A water column sampler for invertebrates in shallow wetlands. J. Wildl. Manag. 42, 670–671. http://dx.doi.org/10.2307/3800841.
van der Valk, A.G., 2005. Water-level fluctuations in North American prairie wetlands. Hydrobiology 539, 171–188. http://dx.doi.org/10.1007/s10750-004-4866-3.
Weller, M.W., et al., 1978. Wetland habitats. In: Greeson, P.E. (Ed.), Wetland Functions and Values: the State of Our Understanding. American Water Resources
Association, Minneapolis, Minnesota, pp. 210–234.
Winter, T.C., Carr, M.R., 1980. Hydrologic Setting of Wetlands in the Cottonwood Lake Area, Stutsman County, North Dakota, U.S. Geological Survey Water Resources
Investigations Report 80-99. 42 p.
Winter, T.C., Rosenberry, D.O., 1995. The interaction of ground water with prairie pothole wetlands in the Cottonwood Lake area, east-Central North Dakota. Wetlands
15, 193–211. http://dx.doi.org/10.1007/BF03160700.
Winter, T.C., Rosenberry, D.O., 1998. Hydrology of prairie pothole wetlands during drought and deluge: a 17 year study of the Cottonwood Lake wetland complex in
North Dakota in the perspective of longer term measured and proxy hydrological records. Clim. Change 40, 189–209. http://dx.doi.org/10.1023/
A:1005448416571.
Winter, T.C., Woo, M.-K., 1990. Hydrology of lakes and wetlands. In: In: Wolman, M.G., Riggs, H.C. (Eds.), Surface Water Hydrology: Boulder, Colorado, Geological
Society of America, vol. O-1. The Geology of North America, pp. 159–187.
Winter, T.C., 1981. Uncertainties in estimating the water balance of lakes. Water Resour. Bull. 17, 82–115. http://dx.doi.org/10.1111/j.1752-1688.1981. tb02593.x.
Winter, T.C., 1989. In: van der Valk, A.G. (Ed.), Hydrologic studies of wetlands in the northern prairie. Iowa State University Press, Northern Prairie Wetlands, Ames,
Iowa, pp. 16–54.
Winter, T.C., (ed.), 2003, Hydrological, chemical, and biological characteristics of a prairie pothole wetland complex under highly variable climate conditions – the
Cottonwood Lake area, east-central North Dakota, U.S. Geological Survey Professional Paper 1675.
Woo, M.-K., Rowsell, R.D., 1993. Hydrology of a prairie slough. J. Hyrdrol. 146, 175–207. http://dx.doi.org/10.1016/0022-1694(93)90275-E.
23

LaBaugh, James W.;; Nelson, Richard D.; Euliss, Ned H. (2018). Long-term changes in pond permanence, size, and salinity in Prairie Pothole Region wetlands The role of groundwater-pond interaction. Journal of Hydrolo

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LaBaugh, James W.;; Nelson, Richard D.; Euliss, Ned H. (2018). Long-term changes in pond permanence, size, and salinity in Prairie Pothole Region wetlands The role of groundwater-pond interaction. Journal of Hydrolo

Uploaded by

Copyright:

Available Formats

Journal of Hydrology: Regional Studies 17 (2018) 1–23

Contents lists available at ScienceDirect

Journal of Hydrology: Regional Studies

Long-term changes in pond permanence, size, and salinity in

1.1. Purpose and scope

3. Material and methods

3.1. Hydrological instrumentation

3.1.2. Surface water

3.1.3. Atmospheric water

3.1.4. Measurement uncertainty

δ = square root of (δ2p + δ2r + δ2gi + δ2et + δ2go + δ2dv) (3)

3.2. Water budget calculation

3.2.1. Pond surface area and volume

3.2.2. Precipitation volume

3.2.3. Runoﬀ volume

3.2.4. Evapotranspiration volume

PET = [1.6 (10Ta/I)6.75×10−7I3−7.71×10−5I2+1.79×10−2I+0.49] (10/d)

The Hamon equation calculates daily PET in mm as follows:

PET = 0.55 (D/12)2 (SVD/100) 25.4 (6)

SVD = 5.018 + 0.32321 Ta + 0.0081847 Ta2 + 0.00031243 Ta3 (7)

3.2.5. Groundwater volume

3.3. Chemical data collection

3.3.2. Surface water

Bismarck, North Dakota, using standard methods (Greenberg et al., 1992).

3.3.3. Atmospheric water

3.4. Chemical budgets

3.4.2. Change in mass in the pond

3.4.3. Rainfall and runoﬀ from rain

4.1. Water budgets

snow Open-water period

1979 *b April–October 7385 380 659 11 * 8957 10468 −4609

snow Open-water period

Energy budget Mather method Hamon method

rainfall runoﬀ groundwater Evapotranspiration groundwater Evapotranspiration groundwater Evapotranspiration groundwater

1979 *a 4065 2554 * 1551 1751 * −2514 −803

4.2. Solute budgets

Fig. 7. Sulfate ﬂux from the pond to groundwater, 1979–1998.

solute units Median value

5.1. Water budgets

5.2. Solute budgets

Appendix A. Supplementary data

You might also like

1979 b April–October 7385 380 659 11 8957 10468 −4609

1979 a 4065 2554 1551 1751 * −2514 −803