10.5061/DRYAD.JH9W0VTBD
Sibilia, Carly
0000-0001-8838-4800
Environmental Resources Management (United Kingdom)
Aguirre-Gutiérrez, Jesús
0000-0001-9190-3229
Naturalis Biodiversity Center
Mowbray, Lauren
United States Fish and Wildlife Service
Malhi, Yadvinder
University of Oxford
Submerged aquatic vegetation, water quality (pH, salinity, and turbidity)
and waterfowl abundance data from 1991-2017 in Back Bay, Virginia
Dryad
dataset
2021
FOS: Other natural sciences
2022-02-02T00:00:00Z
2022-02-02T00:00:00Z
en
https://doi.org/10.5281/zenodo.5948770
https://doi.org/10.5281/zenodo.5948772
23839 bytes
6
CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Back Bay, Virginia, has been documented as an important foraging area for
waterfowl since at least the mid-1800s. Expansive submerged plant beds
historically supported diverse assemblages of non-breeding waterfowl,
however coastal development and other anthropogenic influences have since
led to fluctuations in submerged aquatic vegetation (SAV) and an
associated decline in waterfowl abundance in the bay. To gain insight into
the effects of environmental drivers on waterfowl foraging guilds, our
study explores the effects of SAV frequency and water quality on the
abundance of dabbling ducks, diving ducks, and swans and geese in Back
Bay. We use 8 years of SAV, water quality, and waterfowl monitoring data
collected by state and federal agencies to model the effects of salinity,
turbidity, pH, and percent frequency of SAV on the relative abundance of
waterfowl by foraging guild in Back Bay. The appropriateness of the data
and reasonability of the preliminary results were then evaluated through
semi-structured interviews with 11 local informants representing state,
federal, and non-governmental organizations. Quantitative results
indicated that dabbling ducks are affected differently than other guilds
by water quality and percent frequency of SAV. Thematic analysis of the
interview data revealed a number of potential explanations for the model
results, as well as highlighted areas of uncertainty in need of further
research. In a test of face validity, participants demonstrated a
significant degree of belief in turbidity, salinity, and SAV as drivers of
waterfowl abundance, but were not convinced by the potential effects of pH
as demonstrated by the model. This mixed methods study provides insights
that could potentially influence the management and conservation of
non-breeding waterfowl populations by challenging the assumption that
particular environmental conditions serve all foraging groups equally.
STUDY AREA Back Bay, Virginia, USA (36°36’49” N, 75°56’40” W) is narrowly
separated from the Atlantic Ocean by sand dunes and marsh impoundments to
the east, bordered by emergent wetlands and maritime forest to the north
and west, and connected to Currituck Sound through Knotts Island Channel
to the south (Morton and Kane, 1994; USFWS, 2010) (Fig. 1). As Back Bay is
situated approximately 128 km north of the nearest marine inlet, Oregon
Inlet, North Carolina, the waterbody experiences no lunar tidal action;
instead, water levels are dictated by the prevailing wind direction and
speed (Morton & Kane, 1994; USFWS, 2010). Wind tides coupled with
precipitation and watershed inputs influence the salinity levels of the
bay, which has been historically characterized as oligohaline, ranging 0-3
parts per thousands (ppt) (Norman, 1991; USFWS, 2010). WATERFOWL AERIAL
SURVEYS Aerial surveys designed to estimate wintering waterfowl abundance
and distribution were conducted by the VDWR, in January of each year
between 1998 and 2017, and by the Back Bay National Wildlife Refuge
biologist in 2011 (January, February, March, November, December), 2012
(January, February, March), and 2013 (November, December). In both the
VDWR and USFWS aerial surveys, waterfowl numbers were recorded by species
in generalized locations in and around Back Bay. Mid-winter inventory
flight transects were designed to provide complete aerial coverage of the
mid-winter inventory survey unit, however USFWS surveys placed a greater
focus on refuge and state park property (see Settle & Schwab,
1991). Count data from the refuge and state park impoundments were
excluded here to retain compatibility with the SAV and water quality
datasets. Waterfowl counts by species and location were spoken into a
handheld recorder (Sony ICD-BX140 4GB Digital Voice Recorder, San Diego,
USA, or similar), and later reproduced onto handwritten datasheets.
Scanned copies of the original datasheets were provided by the respective
agencies and electronically transcribed. While pilots varied between
surveys, the VDWR and USFWS surveys were conducted by the same two
biologists, one state and one federal, respectively, thus reducing the
probability of observer bias (Pearse et al., 2008). While Huesmann (1999)
and Eggeman and Johnson (1989) caution against using aerial mid-winter
survey data, the latter identify Virginia as an Atlantic Flyway state
where surveys have been conducted with consistency in methods, personnel,
route coverage, and survey effort. Waterfowl were categorized into guilds
(dabbling ducks, swans and geese, and diving ducks) based foraging
behavior (Table S1). Dabbling ducks normally feed by dabbling their bills
or tipping forward in water ranging in depth from 5 to 30 cm (Guillemain
et al., 2000; Sibley, 2000; Nelms et al., 2007). Swans and geese feed
similarly, by tipping up or grazing, foraging at water depths of 0 to 10
cm (Fredrickson & Reid, 1988; Sibley, 2000; Tatu et al., 2007;
Gyimesi et al., 2011; Nelms et al., 2007). Diving ducks prefer water
greater than 25 cm in depth, where they can dive for SAV as well as animal
matter such as clams, fish, and various other invertebrates (Pöysä, 1983;
Sibley, 2000; Nelms et al., 2007). WATER QUALITY MONITORING STATIONS
Monthly water quality data for salinity, pH, and turbidity collected by
the Virginia Department of Environmental Quality was downloaded from the
National Water Quality Monitoring Council’s Water Quality Portal for nine
stations in and around the perimeter of Back Bay (NWQMC, 2021; Fig. 1).
Salinity was measured via the electrical conductivity method, and pH value
in water was measured by potentiometry using a standard hydrogen electrode
(NWQMC, 2020). Turbidity was measured in nephelometric turbidity units
(NTU) by the nephelometric method (NWQMC, 1995). While water quality
measurements were collected with consistent methodologies, variation in
the time of day that the samples were taken, and the speed and direction
of wind during sampling events may have introduced variability into
otherwise standardized measurement procedures. SUBMERGED AQUATIC
VEGETATION TRANSECTS Submerged aquatic vegetation surveys in Back Bay were
established in 1958 (Sincock et al., 1965) and have been surveyed annually
by the VDWR since 2009 following the methods of Schwab et al. (1991) (G.R.
Costanzo, VDWR, unpublished data). During each survey event, vegetation
samples were taken at 500 m intervals along eight transect lines
traversing Back Bay (Fig. 1). At each 500 m interval, three
two-square-foot bottom samples were taken using modified oyster tongs, and
species of SAV was recorded along with a visual estimate of percent cover
or density (trace, low, medium, and high) (Fig. S2). The one or two-day
sampling events were conducted annually between September and November.
Annual percent frequency of SAV in Back Bay was derived from dividing the
number of samples with any SAV by the total number of samples taken each
year. STATISTICAL ANALYSIS All quantitative analyses were conducted in the
R statistical platform with the ‘lme4’, ‘MuMIn’, ‘lsmeans’ and ‘multcomp’
packages (Version 1.1.463, 2009-2018) (Hothorn et al., 2008; Bates et al.,
2015; Lenth, 2016; Bartoń, 2019). As the data were not subject to
overdispersion, we used a Poisson distribution model to investigate the
impact of the selected environmental drivers on waterfowl abundance by
foraging guild. The water quality and SAV variables were centered and
standardized before analyses to allow for direct comparison of model
coefficients between variables with different units (z-scores; Gelman,
2008). A Bayesian information criterion (BIC) comparison, which penalizes
more complex models by excluding terms that explain only little
variability in the data, was used to evaluate three models with three-way
interactions between feeding guild, SAV percentage, and each water quality
characteristic (pH, salinity, and turbidity) (Aho et al., 2014; Table 1).
The first model comprises three, three-way interactions, including all
additive terms and lower order interactions. The second model builds on
the first by adding an offset for the number of waterfowl surveys per
year. In addition to the added offset, the third model incorporates year
as a random factor. The BIC comparison indicated that the first and second
models were equally parsimonious. The second model, which includes three,
three-way interactions between waterfowl guild, average water quality
measurements, and annual vegetation frequencies from 2010 to 2017, as well
as an offset for the number of waterfowl surveys per year, was chosen to
evaluate the data (R2Adjusted = 0.46; Table S2). A Tukey multiple
comparison of means analysis was subsequently run to assess differences
between abundances across feeding guilds.
REFERENCES Aho, K., Derryberry, D. & Peterson, T. (2014). Model
selection for ecologists: The worldviews of AIC and BIC. Ecology, 95(3),
631–636. https://doi.org/10.1890/13-1452.1 Bates, D., Mächler, M., Bolker,
B., & Walker, S. (2015). Fitting Linear Mixed-Effects Models Using
lme4’, Journal of Statistical Software, 67(1), 1–48.
https://doi.org/10.18637/jss.v067.i01 Bartoń, K. (2019). Package ‘MuMIn’,
CRAN. Eggeman, D.R., & Johnson, F.A. (1989). Variation in effort
and methodology for the midwinter waterfowl inventory in the Atlantic
Flyway. Wildl Soc Bul,l 17(3), 227-233.
https://www.jstor.org/stable/3782374 Fredrickson, L.H., & Reid,
F.A. (1988). Nutritional values of waterfowl foods. In Waterfowl
Management Handbook. U.S. Fish and Wildlife Service. Gelman, A. (2008).
Scaling regression inputs by dividing by two standard deviations.
Statistics in Medicine, 27, 2865–2873. https://doi.org/ 10.1002/sim.3107
Guillemain, M., Fritz, H., & Blais, S. (2000). Foraging methods
can affect patch choice: an experimental study in Mallard (Anas
platyrhynchos). Behav Processes, 50(2–3), 123–129.
https://doi.org/10.1016/S0376-6357(00)00095-4 Gyimesi, A., de Vries, P.P.,
de Boer, T., & Nolet, B.A. (2011). Reduced tuber banks of fennel
pondweed due to summer grazing by waterfowl. Aquat Bot, 94(1), 24-28.
https://doi.org/10.1016/j.aquabot.2010.10.002 Hothorn, T., Bretz, F. and
Westfall, P. (2008) Simultaneous Inference in General Parametric Models.
Biometrical Journal, 50(3), 346–363. https://doi.org/
10.1002/bimj.200810425 Huesmann, W. (1999). Let’s get rid of the midwinter
waterfowl inventory in the Atlantic Flyway. Wild Soc Bull, 27(3), 559-565.
https://www.jstor.org/stable/3784074 Lenth, R. (2016). Least-Squares
Means: The R Package lsmeans. Version 2916. Journal of Statistical
Software, 61(1), 1–33. https://doi.org/10.18637/jss.v069.i01[NWQMC]
National Water Quality Monitoring Council. (1995). Standard methods: 2130
B: turbidity by nephelometry. National Environmental Methods Index.
https://www.nemi.gov/methods/method_summary/9645/ [NWQMC] National Water
Quality Monitoring Council. (1995). Standard methods: 2130 B: turbidity by
nephelometry. National Environmental Methods Index.
https://www.nemi.gov/methods/method_summary/9645/ [NWQMC] National Water
Quality Monitoring Council. (2020). Standard methods: 4500- H+B: pH in
water by potentiometry. National Environmental Methods Index.
https://www.nemi.gov/methods/method_summary/4707/ [NWQMC] National Water
Quality Monitoring Council. (2021). Water Quality Data.
https://www.waterqualitydata.us/portal/ Nelms, K.D., Ballinger, B.,
& Boyles, A. (2007). Wetland Management for Waterfowl Handbook.
Mississippi River Trust, Natural Resources Conservation Service, United
States Fish and Wildlife Service.
https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/nrcs142p2_016986.pdf
Pöysä, H. (1983). Resource Utilization Pattern and Guild Structure in a
Waterfowl Community. Oikos, 40(2), 295. https://doi.org/10.2307/3544594
Schwab, D., Settle, F.H., Halstead, O., & Ewell, R.L. (1991).
Submerged aquatic vegetation trends of Back Bay, Virginia. In Proceedings
of the Back Bay Ecological Symposium. Old Dominion University Digital
Commons, 265–269. https://digitalcommons.odu.edu/backbay1990_flora/6
Sibley, D.A. (2000). The Sibley Guide to Birds. Alfred A. Knopf, Inc.
Sincock, J.L., Johnston, K.H., Coggin, J.L., Wollitz, R.E., Kerwin, J.A.,
Dickson, A.W., Crowell, T., Grandy III, J., Davis, J.R., &
McCartney, R. (1965). Back Bay - Currituck Sound data report: Introduction
and vegetation studies, Volume I.
https://www.fws.gov/southeast/pdf/report/backbay-currituck-sound-vegatation.pdf Tatu, K.S., Anderson, J.T., Hindman, L.J., & Seidel, G. (2007). Mute swans’ impact on submerged aquatic vegetation in Chesapeake Bay. J Wildl Manage, 71, 1431–1439. https://doi.org/10.2193/2006-130