Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Projected Influences of Changes in Weather Severity on Autumn-Winter Distributions of Dabbling Ducks in the Mississippi and Atlantic Flyways during the Twenty-First Century

  • Michael Notaro ,

    mnotaro@wisc.edu

    Affiliation Nelson Institute Center for Climatic Research, University of Wisconsin-Madison, Madison, Wisconsin, United States of America

  • Michael Schummer,

    Affiliation Department of Biological Sciences, State University of New York at Oswego, Oswego, New York, United States of America

  • Yafang Zhong,

    Affiliation Space Science and Engineering Center, University of Wisconsin-Madison, Madison, Wisconsin, United States of America

  • Stephen Vavrus,

    Affiliation Nelson Institute Center for Climatic Research, University of Wisconsin-Madison, Madison, Wisconsin, United States of America

  • Lena Van Den Elsen,

    Affiliation Long Point Waterfowl/Bird Studies Canada, Port Rowan, Ontario, Canada

  • John Coluccy,

    Affiliation Ducks Unlimited, Ann Arbor, Michigan, United States of America

  • Christopher Hoving

    Affiliation Michigan Department of Natural Resources–Wildlife Division, Lansing, Michigan, United States of America

Abstract

Projected changes in the relative abundance and timing of autumn-winter migration are assessed for seven dabbling duck species across the Mississippi and Atlantic Flyways for the mid- and late 21st century. Species-specific observed relationships are established between cumulative weather severity in autumn-winter and duck population rate of change. Dynamically downscaled projections of weather severity are developed using a high-resolution regional climate model, interactively coupled to a one-dimensional lake model to represent the Great Lakes and associated lake-effect snowfall. Based on the observed relationships and downscaled climate projections of rising air temperatures and reduced snow cover, delayed autumn-winter migration is expected for all species, with the least delays for the Northern Pintail and the greatest delays for the Mallard. Indeed, the Mallard, the most common and widespread duck in North America, may overwinter in the Great Lakes region by the late 21st century. This highlights the importance of protecting and restoring wetlands across the mid-latitudes of North America, including the Great Lakes Basin, because dabbling ducks are likely to spend more time there, which would impact existing wetlands through increased foraging pressure. Furthermore, inconsistency in the timing and intensity of the traditional autumn-winter migration of dabbling ducks in the Mississippi and Atlantic Flyways could have social and economic consequences to communities to the south, where hunting and birdwatching would be affected.

Introduction

Waterfowl and their habitats are ecologically, economically, and socially valuable, providing food, income, recreation, and ecosystem stability [1]. It has been estimated that 1.8 million waterfowl hunters in the United States participate during nearly 30 million recreational days per year, with the economic activity leading to $2.3 billion in spending and over 21,000 jobs [2]. The total economic impact of waterfowl hunting in Mississippi alone was estimated at $27.4 million [3]. An estimated 47 million birdwatchers in the United States, over the age of 16, spent approximately $41 billion in total trip and equipment expenditures in 2011, with peak participation in the South and birding expenditures supporting 666,000 jobs [4]. The most watched type of bird for traveling birders is waterfowl [4].

Because waterfowl are relatively large-bodied migratory birds that number in the millions, they have considerable energy needs during migration and winter [5]. Redistribution of these birds under climate change would result in substantial changes in their latitudinal foraging pressure in aquatic and agricultural habitats [6]. Despite its importance to managers and conservation planners, there has been insufficient exploration of the meteorological factors influencing the movement and distribution of waterfowl during the autumn-winter non-breeding period, with impacts on waterfowl harvest, habitat use, and survival [5,713]. While the impact of climate change on birds has received considerable attention, these studies have primarily focused on the breeding season, and few studies have investigated changes to the distributions of waterfowl during autumn-winter [1317].

The autumn migration of dabbling ducks is thought to be controlled by changes in photoperiod; habitat suitability and management; food accessibility; weather severity, including effects from both temperature (thermoregulatory cost) and snow cover; feeding ecology and competition [7,12,1732]. The cost of migration is high [3334]. However, when temperature is below a critical threshold, elevated thermoregulatory demands cause it to become more energetically costly for Northern Hemispheric waterfowl to persist at higher latitudes than to migrate southward to warmer environments, according to energy conservation theory [5,3537]. Furthermore, the presence of snow and ice cover can restrict wetland food availability and foraging capacity, thereby increasing competition and time in flight and reducing nutrient acquisition among wetland-obligate waterfowl [3840]. [41], for example, demonstrated that winter temperature and snow cover are key regulators of waterfowl abundance in the Mississippi Alluvial Valley, which is the primary region for wintering Mallards (Anas platyrhynchos) [7,42]. Waterfowl, such as Mallards, typically migrate farther south in response to cold conditions in their northern wintering areas [7,4345], with movements often triggered by cold spells [19,30, 40,4648] or food depletion [30,4950].

Anthropogenic climate change, particularly rising air temperatures, longer growing seasons, and diminished snow accumulation at many locations [51], has already induced noteworthy phenological shifts in the timing of migration and breeding by waterfowl and other birds, thereby impacting population distributions, and poleward shifts in range margins [13,45, 5270]; however, attributing these range shifts to specific drivers is often not straightforward [3132]. For example, the abundance of American Black Duck (Anas rubripes), a short-distance migrant, has shown signs of delayed migration and wintering at more northern areas [18,67,7174]. A continued poleward shift in the distribution of wintering waterfowl will increase foraging pressure and, thus, habitat needs across mid-high latitudes, including central North America, where only 10–15% of historical wetlands remain [62,7578].

Recent studies by [6,12,21,23] have focused on the relationship between changes in the relative abundance of dabbling ducks and meteorological variables across mid-latitude North America. Based on waterfowl survey data from conservation areas, [12] developed a cumulative weather severity index (WSI) to explain changes in the relative abundance of Mallards during autumn-winter migration at their staging areas in Missouri, within the Mississippi Flyway. Their model represented the current and cumulative effects of air temperature on energy expenditure and snow cover and wetland icing on food availability by considering four variables: (1) the mean daily temperature, (2) the number of consecutive days with mean air temperature at or below freezing, (3) snow depth, and (4) the number of consecutive days with at least 2.54 cm of snowpack. The WSI is a valuable tool for resource managers and researchers to identify the days when Mallards are likely to begin leaving a location for more southern latitudes [21]. [23] later expanded the analysis to the meteorological and photoperiod regulators of the rate of change in the relative autumn-winter abundance of multiple dabbling duck species using aerial and ground-based survey data across the Mississippi and Atlantic Flyways for 25 locations in the United States and Canada. The statistical models that explained the greatest total variance included air temperature, snow depth, and latitude for all analyzed duck species, except for the Blue-Winged Teal (Anas discors), which is a long-distance migrant that largely responds to photoperiod. In general, the primary migration cue for long-distance migrants (i.e., neotropical migrants) wintering in stable habitats is photoperiod, with decreasing day length synchronizing their circadian and circannual rhythms, while weather cues have a greater impact on short-distance migrants that winter in unstable habitats [23,7980]. Most dabbling ducks that breed in North America winter at southern locations on the continent, where they are subject to variable weather conditions during the non-breeding period [22].

[81] investigated future changes in the relative abundance and autumn-winter migratory timing of Mallards across central-eastern North America based on the WSI established by [12]. Projected changes in air temperature and snow depth by the mid- and late 21st century were based on statistically downscaled climate projections from nine global climate models (GCMs) within the Coupled Model Intercomparison Project Phase Three (CMIP3) and the application of an operational snow accumulation and ablation model. December-January WSI was projected to decline substantially during this century, leading to increased likelihood of delays in the timing and intensity of autumn-winter waterfowl migrations. Besides being restricted to only one dabbling duck species, the previous study was limited because the statistical downscaling approach used to derive snowfall projections did not consider projected changes in water temperature, ice cover, and evaporation for the Great Lakes and thus the earlier projections poorly represented future lake-effect snowfall dynamics [81]. This is particularly concerning, given the pronounced trends of declining ice cover [82] and increasing lake evaporation [83] and lake-effect snowfall [84,85] that have already been observed. To address this limitation, [86,87] produced dynamically downscaled climate projections for the Great Lakes Basin and broader region using a high-resolution regional climate model (RCM), interactively coupled to a one-dimensional lake model to represent changing water temperature and ice cover on the Great Lakes. They downscaled two of the Coupled Model Intercomparison Project Phase Five (CMIP5) GCMs. Here, we expanded the number of downscaled GCMs to six in order to better capture uncertainty of regional climate change projections.

The objective of the current study is to investigate the likely impacts of dynamically downscaled weather severity projections for the mid- and late 21st century on the relative abundance and autumn-winter migratory behavior of seven common dabbling duck species across the Mississippi and Atlantic Flyways of North America. While the current study focuses on dabbling ducks as an example of the application of downscaled weather severity projections, the general approach can be more broadly applied to other wildlife investigations.

Data and Methods

Species-specific weather severity indices

The following summarizes the statistical models developed by [23] for the relative rate of change of seven focal dabbling duck species as a function of current and cumulative weather severity. The onset of negative population rates implies southward migration. [23] applied an information-theoretic approach for model selection [88] based on Akaike’s Information Criterion (AIC), which measures the quality of statistical models while penalizing based on the number of predictors. All candidate models within 2.0 ΔAIC units of the top-ranked models were believed to exhibit biological significance, such that modeling averaging was applied to estimate parameters and the 85% confidence intervals for the top models [89]. All calculations are performed during September through March, focusing on the autumn-winter migration.

For the American Black Duck, principal component (PC) analysis of the individual elements of weather severity yields the following index for PC1: (1) where TEMP is the daily mean air temperature (°C) multiplied by -1, TEMPDAY is the number of consecutive days with mean air temperature less than or equal to 0°C, SNOW is the daily snow depth (cm) multiplied by 0.394, and SNOWDAY is the number of consecutive days with at least 2.54 cm of snow on the ground. The PC analysis aims to produce an index that explains the maximum variance among TEMP, TEMPDAY, SNOW, and SNOWDAY, which are seasonally correlated [12,90]. The empirical coefficients in Eq (1) represent the eigenvectors of the predictor variables that compose the first PC. Based on the PC1 index and latitude (degrees north), the following quadratic equation estimates the relative rate of change in American Black Duck abundance between two dates at a surveyed location: (2) Likewise, for Mallards, the PC1 index and resulting rate were as follows: (3) (4) Sample data for the calculation of Mallard population rate is presented in S1 Table to demonstrate the methodology. For the American Wigeon (Anas americana), the index, WSIMEAN, is introduced and defined as: (5) where TEMPMEAN is the mean air temperature during the last seven days multiplied by -1. The relative rate of change of population for the American Wigeon is estimated by: (6) The rates for Gadwall (Anas strepera, RateG), Green-Winged Teal (Anas crecca, RateGWT), and Northern Shoveler (Anas clypeata) are as follows: (7) (8) (9) For Northern Pintain (Anas acuta), the WSI index is computed as follows, (10) in which the maximum value of WSI over the last seven days is selected. Using this WSI index, the relative rate of change of population for the Northern Pintail is estimated by: (11) Among the numerous candidate models adapted from [12], [23] selected WSIMEAN to be retained in the best models for the American Wigeon, Gadwall, Green-Winged Teal, and Northern Shoveler only and selected WSI to be retained in the best model for the Northern Pintail only.

Given the quadratic fit of most of the aforementioned population rate formulas, biologically unrealistic values can be generated for PC1 and WSIMEAN by the aforementioned models when WSI is calculated for extremely mild or severe values not within the range of the collected data sample. As a result, positive population rates are assumed in the study region for the American Black Duck when PC1 falls below the threshold of -19 to -15 (depending on latitude) or for the Mallard when PC1 falls below the threshold of -23 to -11, indicative of mild, snow free conditions. Furthermore, negative population rates are assumed for the American Wigeon and Gadwall when WSIMEAN exceeds the thresholds of 85 to 150 and 80 to 185, respectively, indicative of cold, snow-covered conditions.

Dynamical downscaling

The output from six CMIP5 GCMs is dynamically downscaled using the Abdus Salam International Centre for Theoretical Physics (ICTP) Regional Climate Model version four (RegCM4) [91], interactively coupled to a one-dimensional, energy-balance lake model [92] and lake ice sub-model [9394] to represent the Laurentian Great Lakes. The GCMs include the Centre National de Recherches Meteorologiques Coupled Global Climate Model Version Five (CNRM-CM5), the Model for Interdisciplinary Research on Climate Version Five (MIROC5), the Institut Pierre Simon Laplace Coupled Model Version Five-Medium Resolution (IPSL-CM5-MR), the Meteorological Research Institute Coupled Global Climate Model Version Three (MRI-CGCM3), the Centre for Australian Weather and Climate Research, Australia GCM (ACCESS1-0), and the National Oceanic and Atmospheric Administration Geophysical Fluid Dynamics Laboratory model (GFDL-ESM2M). The domain, consisting of 217 by 141 grid cells, extends across most of the contiguous United States and southern Canada (Fig 1). The simulations apply 25-km grid spacing and 28 vertical sigma levels. Lateral boundary conditions from the GCMs are provided through a linear relaxation scheme to a 15 gridcell buffer zone, which surrounds the inner domain. Analyses for the late 20th, mid-21st, and late 21st centuries are limited here to 1980–1999, 2040–2059, and 2080–2099, according to the representative concentration pathway 8.5 (RCP8.5) [95], which is a high-end emission scenario for greenhouse gases. Further information on the dynamical downscaling and model performance is presented by [8687,9699]. The locations of the Mississippi and Atlantic Flyways within the United States are identified in Fig 1. Subsequent area-average calculations for the study region cover 30–50°N, 97–67°W, extending northward into the breeding zones across southern Ontario, Quebec, and Manitoba.

thumbnail
Fig 1. Model domain with elevation (shading, meters).

The thick black rectangle indicates the buffer zone. Dots indicate the 25-km grid spacing. The orange, blue, red, and purple polygons identify the Pacific, Central, Mississippi, and Atlantic Flyways, respectively, based on a shapefile obtained from Ducks Unlimited.

https://doi.org/10.1371/journal.pone.0167506.g001

Debiasing methodology

Prior to calculating species-specific weather severity indices, RegCM4-simulated output of daily 2-meter air temperature and snow depth is debiased against observational data, both in terms of the daily mean and interannual standard deviation for each day. The source of observed daily mean air temperature for 1984–2013 is the 1-km, gridded Daily Surface Weather and Climatological Summaries (Daymet) product [100101]. For the purpose of debiasing daily snow depth, a gridded product is created using data from 7,360 meteorological stations within the Global Historical Climate Network (GHCN) [102] across the area of 26–54°N, 101–63°W; this station list is reduced from the original 12,911 stations in that region by requiring at least 16.7% daily data availability. For a given day, debiasing is performed by subtracting the simulated climatological mean for that day, multiplying by the ratio of the observed interannual standard deviation to the simulated interannual standard deviation for that day, and then adding the observed climatological mean for that day. [103] explored multiple debiasing techniques, including linear bias correction of the mean and standard deviation as applied here, and concluded that the relative performance of each method varies by region and season, such that no universally superior method could be identified. Subsequent analyses and plots apply the debiased temperature and snow data.

Results

Projected climate change

According to the debiased dynamical downscaling, the Mississippi and Atlantic Flyways are projected to warm in autumn-winter (September through February) by 2.4°C by the mid-21st century, ranging from +1.9°C for GFDL to +3.1°C for MIROC5, compared to the late 20th century (Fig 2A–2C). The mean projected warming by the late 21st century is 4.8°C, ranging from 3.9°C for MRI to 5.7°C for MIROC5 (Fig 2D–2F). The uncertainty in projected warming, represented by the spread among models, increases substantially later in the century. The projected warming is approximately 10% greater in the debiased data than the original RCM data, due to the model’s modest underestimation of the interannual variability in temperature. For all six models and both time periods, the peak projected warming occurs over southern Ontario, Quebec, and Manitoba and the Upper Midwest United States, representing the northern and western portions of the study region. There are also local minima in warming in close proximity to the Great Lakes and Atlantic Coast, due to the buffering effect of these large water bodies and their substantial heat capacity.

thumbnail
Fig 2.

Projected change in 2-m air temperature (°C) for autumn-winter (September through February) by the (a-c) mid-21st and (d-f) late 21st century, compared to the late 20th century. Results are shown for the (a,d) six-model mean, (b,e) the model with the least warming, and (c,f) the model with the greatest warming.

https://doi.org/10.1371/journal.pone.0167506.g002

The frequency of extremely cold episodes across the Great Lakes region (40–50°N, 95–70°W) during autumn-winter (September through February) is expected to decrease substantially during the 21st century, based on an analysis of the frequency of days within specific mean air temperature bins, between -35°C to -30°C and 20°C to 25°C (Fig 3). During the late 20th century, 35% of days exhibit a mean temperature within 5°C of the freezing point, with -5°C to 0°C and 0°C to 5°C representing the most active bins. The frequency of days below freezing is expected to decrease and the frequency of days above freezing is expected to increase during the 21st century within the Great Lakes region as the probability density function of daily air temperatures experiences a pronounced shift towards higher temperatures. As a result of this warming, the projected increase in autumn-winter precipitation across the Great Lakes region among all six models is characterized by greater rainfall and reduced snowfall. Specifically, when averaged across the six models, the largest projected decline in frequency is -3.6 days by the mid-21st century and -7.7 days by the late 21st century for the -15°C to -10°C bin, and the largest projected increase in frequency is +4.3 days by the mid-21st century and +10.2 days by the late 21st century for the 20°C to 25°C bin.

thumbnail
Fig 3. Projected change in the frequency of days during autumn-winter (September-February) with daily mean air temperatures lying within specified bins between -35°C to -30°C and 20°C to 25°C across the Great Lakes region (40–50°N, 95–70°W).

The frequency for the late 20th century is shown in blue bars. Projected changes in frequency by the mid- and late 21st century, compared to the late 20th century, are shown with green and red lines, respectively, with one line for each of the six downscaled models.

https://doi.org/10.1371/journal.pone.0167506.g003

In response to the aforementioned warming in autumn-winter, ice cover on the Great Lakes is projected to experience pronounced reductions during the 21st century, especially in February-March, with declines during the two-month period ranging from -31% in GFDL to -50% in MIROC5 by the late 21st century (not shown). Declines of this magnitude would mean that the Great Lakes are projected to become mostly open water during winter by the late 21st century, with a dramatically shortened ice season. As described by [86], future changes in lake-effect snowfall are uncertain; declining lake ice cover would support more evaporation and thus more lake-effect snowfall, but the reduced frequency of strong cold air outbreaks out of Canada would trigger fewer lake-effect snow events. [86,96] introduced an objective methodology for identifying heavy lake-effect snowstorm days in RCM output, based on proximity to lakeshore, wind direction, lake ice cover, local snowfall amount, and enhancement of snowfall near the lakeshore. Based on the application of these criteria, the frequency of heavy lake-effect snowstorms within the Great Lakes Basin is projected to decline by -1% in GFDL to -20% in ACCESS by the mid-21st century and by -17% in GFDL to -45% in ACCESS by the late 21st century.

Due to reduced snowfall and accelerated snowmelt, the mean number of days with at least 2.54 cm (1 inch) of snow on the ground is projected to decline across the Mississippi and Atlantic Flyways (Fig 4). Reductions of -8.5 days (-27%, compared to 31.1 days in the late 20th century) are simulated by the mid-21st century, ranging from -4.3 days in GFDL to -10.9 days in ACCESS, and of -15.0 days (-48%) are simulated by the late 21st century, ranging from -10.4 days in GFDL to -17.8 days in MIROC5. Within the study region, the zone of most pronounced reductions in the duration of snowpack is identified around 44–45°N, across the Great Lakes region, partly associated with the projected decline in lake-effect snowfall.

thumbnail
Fig 4.

Projected change in the mean number of days during autumn-winter (September through February) with at least 2.54 cm of snow on the ground by the (a-c) mid-21st and (d-f) late 21st century, compared to the late 20th century. Results are shown for the (a,d) six-model mean, (b,e) the model with the least loss of snowpack, and (c,f) the model with the greatest loss of snowpack.

https://doi.org/10.1371/journal.pone.0167506.g004

Projected responses of dabbling ducks

The mean migration date from the Great Lakes region for seven dabbling ducks species is estimated according to the aforementioned weather severity equations across the Great Lakes zone of 40–50°N (Table 1). Mean migration date is defined as the initial date during autumn-winter in which there is at least a 50% chance (computed across years) of achieving a negative population rate due to low temperatures and snowpack. Among the seven species, the American Black Duck and Mallard are the latest migrants, with mean migration dates out of the Great Lakes region of 10 December and 9 December, respectively, during the late 20th century (Figs 5 and S1). In contrast, the Northern Shoveler is the earliest migrant for the same region, with a mean migration date of 2 October during the late 20th century (Fig 6).

thumbnail
Fig 5. Probability of a negative population rate of the Mallard across the Mississippi and Atlantic Flyway, within latitudinal bands from 30°N to 50°N, for the late 20th, mid-21st, and late 21st centuries (first three columns).

The probabilities are shown for each day from 1 September through 31 March, with the x-axis labeled as days since 1 September. Projected changes in these probabilities are shown for the mid-21st and late 21st century, compared to the late 20th century (fourth and fifth columns). Results are shown for CNRM, MIROC5, IPSL, MRI, ACCESS, and GFDL, from top to bottom row.

https://doi.org/10.1371/journal.pone.0167506.g005

thumbnail
Table 1. Summary for seven analyzed dabbling duck species.

https://doi.org/10.1371/journal.pone.0167506.t001

Both the timing and the distance traveled (implied if species do not have to leave northern latitudes due to weather severity) during migration are projected to shift for each of the dabbling duck species. Projected changes in the probability (among 20 years) of negative population rates across the Mississippi and Atlantic Flyways, and thus the tendency for autumn-winter migration, are presented in Figs 5 and 6, for the Mallard (late migrant) and Northern Shoveler (early migrant) as contrasting examples, and S1S5 Figs, for the remaining species, through zonal averages (by latitude from 30°N to 50°N). The probabilities are shown for each day from 1 September through 31 March (e.g. 1 Jan is day 122), with the x-axis labeled as days since 1 September. In terms of the two rightmost columns in Figs 5 and 6, which display projected changes in the probability of a negative population rate, a shift to the right of the colored shading indicates a delayed phenological response in migration, while a shift upward indicates the tendency for more southerly populations to overwinter and not migrate.

In general, the greatest (least) projected delays in migration are expected for modern-day late (early) migrant species. The Northern Pintail is projected to experience the least delay in migration out of the Great Lakes region (40–50°N), on the order of +12 days and +23 days by the mid- and late 21st century, when averaged across the six models (Table 1, S5 Fig). Its mean migration date is 4 November in the late 20th century, ranges from 14 November in GFDL to 20 November in MIROC5 for the mid-21st century, and ranges from 21 November in GFDL and 7 December in MIROC5 for the late 21st century. The greatest delay in migration out of the Great Lakes region is expected for the Mallard, on the order of +19 days by the mid-21st century and +40 days, or the possibility of overwintering, by the late 21st century (Table 1, Fig 5). Its mean migration date is 9 December in the late 20th century, ranges from 18 December in GFDL to 5 January in MIROC5 for the mid-21st century, and ranges from 5 January in GFDL to never in MIROC5 for the late 21st century. Overwintering in the Great Lakes region could become increasingly likely by the late 21st century for both American Black Ducks and Mallards (Table 1, Figs 5 and S1). The models with the least warming, such as CNRM and MRI, generally produce more modest delays in migration than the models with the greatest warming, such as MIROC5. For example, by the late 21st century, the mean migration date of the Northern Shoveler in the Great Lakes region may become delayed by 21 days according to CNRM or 37 days according to MIROC5.

Discussion and Conclusions

Projected changes in the relative abundance and timing of autumn-winter migration for seven dabbling duck species are investigated for the mid- and late 21st century across the Mississippi and Atlantic Flyways. Based on aerial and ground-based survey data of waterfowl abundance, species-specific empirical relationships are established, specific to the Mississippi and Atlantic Flyways, between the rate of change in abundance of dabbling ducks and cumulative weather severity indices, based on daily mean air temperature, the number of consecutive days with mean air temperature below freezing, daily snow depth, and the number of consecutive days with at least 2.54 cm of snow on the ground. These weather severity indices reflect both energy conservation theory and energy acquisition theory by capturing the impacts of cold atmospheric conditions on the energy expenditure of ducks and of snow cover, lake ice cover, and wetland icing on food availability to ducks [12]. Of the seven species, the Northern Shoveler is the earliest migrant, as it is a wetland-obligate foraging species, which consumes invertebrates on or near the surface in shallow wetlands and must migrate once wetlands freeze. In contrast, the American Black Duck and Mallard are the latest migrants, acting as opportunistic foragers that can switch to waste agricultural grains when wetland foods are restricted, thereby enabling a delay in their autumn migration [22,104106].

Statistical downscaling can be a powerful and efficient way to translate GCM output to spatial scales that are more relevant for planners and resource managers. However, a key assumption of statistical downscaling is that the drivers of local spatial variation in climate remain unchanged over time. This stationarity assumption has not held true for the Great Lakes Basin, where lake ice cover has rapidly declined in recent decades, leading to enhanced evaporation and lake-effect snowfall [8285]. Past spatial variations in snowfall across the basin were largely driven by lakes that were significantly ice covered during the winter, but that is not expected in the future. This was a weakness of the study by [81], which was rectified in a small initial pool of dynamically downscaled models presented by [86,87]. The clear advantage of the dynamical downscaling approach is that it addresses projected changes in Great Lakes’ water temperatures, ice cover, and lake evaporation and resulting impacts on lake-effect snowfall. Changes in snowpack, lake ice cover, and wetland icing in the Great Lakes Basin have critical implications to food availability for dabbling ducks within the Mississippi and Atlantic Flyways. Here, the output from six CMIP5 global climate models is dynamically downscaled over much of the contiguous United States and southern Canada for the late 20th, mid-21st, and late 21st centuries using a high-resolution regional climate model, RegCM4, interactively coupled to a one-dimensional lake model to represent the Great Lakes. This effort expands upon the initial pool of downscaled models presented by [86,87].

Dramatic reductions in weather severity are projected for the mid- and late 21st century across the Mississippi and Atlantic Flyways, including substantial mean warming during autumn-winter, fewer days below freezing, and diminished lake ice cover, lake-effect snowfall, and snow depth. The most pronounced reductions in the number of days with snow on the ground is identified around 44–45°N, across the Great Lakes region, consistent with the findings of [81]. As discussed by [81], this zone is located close to the -5°C isotherm of the late 20th century climatological near-surface air temperature for December-March, such that as the isotherm shifts northward in the future, projected trends in declining snowfall accelerate in this zone. Further to the north, mean temperatures remain low enough, even later this century, for much of the wintertime precipitation to still occur as snow and for the snowpack to persist. From a climatological perspective, we recommend that future work expand the number of applied RCMs, to better capture the spread of climate change uncertainty for the study region, and couple the RCM to a three-dimensional lake model, in order to represent the circulation of deep lakes and reduce biases in lake temperatures, timing of stratification, and ice cover.

The projected reduction in weather severity should lead to delayed autumn-winter migration for all seven dabbling duck species, with results suggesting that the delays will be least for the Northern Pintail and greatest for the Mallard (greatest delays for the ducks with the largest mean weight, namely American Black Duck and Mallard). By the late 21st century, overwintering in the Great Lakes region may become increasingly likely for the American Black Duck and Mallard. This appears to be supported by the study of [67], which noted observed mid-winter trends toward an increased abundance of American Black Ducks in Ontario and decreased abundance in the Mississippi Flyway of the United States. The earlier study by [12], which lacked consideration of future changes in snow depth, likewise concluded that the Mallard would experience a greater northward shift in distribution than most other dabbling duck species, which generally migrate south during milder conditions than the Mallard. Future declines in wetland icing will also provide greater foraging opportunities for all species, but may benefit wetland obligates, like Gadwall and Northern Shoveler, to a greater degree [2122]. Projected reductions in weather severity and delays in autumn-winter duck migration will increase foraging pressures on remaining wetland habitats in the Great Lakes Basin and Upper Midwest United States, but reduce energy needs in southerly locations [12].

Large economic losses might be expected for the southern flyway states due to diminished hunter and bird watching opportunity, especially because the most common species of duck in North America is projected to overwinter in some years in the Great Lakes Basin. This could reduce the number of migrating waterfowl in more southerly states significantly. As emphasized by [107], in light of ongoing climate change and northward shifts in duck distributions, it will be critical to protect and restore wetlands across the mid-latitudes of North America, especially given the loss of historic wetlands and ongoing stresses from development and pollution to existing wetlands across the Great Lakes region [1,7576,108].

Supporting Information

S1 Fig. Probability of a negative population rate of the American Black Duck across the Mississippi and Atlantic Flyway, within latitudinal bands from 30°N to 50°N, for the late 20th, mid-21st, and late 21st centuries (first three columns).

The probabilities are shown for each day from 1 September through 31 March, with the x-axis labeled as days since 1 September. Projected changes in these probabilities are shown for the mid-21st and late 21st century, compared to the late 20th century (fourth and fifth columns). Results are shown for CNRM, MIROC5, IPSL, MRI, ACCESS, and GFDL, from top to bottom row.

https://doi.org/10.1371/journal.pone.0167506.s001

(EPS)

S2 Fig. Probability of a negative population rate of the American Wigeon across the Mississippi and Atlantic Flyway, within latitudinal bands from 30°N to 50°N, for the late 20th, mid-21st, and late 21st centuries (first three columns).

The probabilities are shown for each day from 1 September through 31 March, with the x-axis labeled as days since 1 September. Projected changes in these probabilities are shown for the mid-21st and late 21st century, compared to the late 20th century (fourth and fifth columns). Results are shown for CNRM, MIROC5, IPSL, MRI, ACCESS, and GFDL, from top to bottom row.

https://doi.org/10.1371/journal.pone.0167506.s002

(EPS)

S3 Fig. Probability of a negative population rate of the Gadwall across the Mississippi and Atlantic Flyway, within latitudinal bands from 30°N to 50°N, for the late 20th, mid-21st, and late 21st centuries (first three columns).

The probabilities are shown for each day from 1 September through 31 March, with the x-axis labeled as days since 1 September. Projected changes in these probabilities are shown for the mid-21st and late 21st century, compared to the late 20th century (fourth and fifth columns). Results are shown for CNRM, MIROC5, IPSL, MRI, ACCESS, and GFDL, from top to bottom row.

https://doi.org/10.1371/journal.pone.0167506.s003

(EPS)

S4 Fig. Probability of a negative population rate of the Green-Winged Teal across the Mississippi and Atlantic Flyway, within latitudinal bands from 30°N to 50°N, for the late 20th, mid-21st, and late 21st centuries (first three columns).

The probabilities are shown for each day from 1 September through 31 March, with the x-axis labeled as days since 1 September. Projected changes in these probabilities are shown for the mid-21st and late 21st century, compared to the late 20th century (fourth and fifth columns). Results are shown for CNRM, MIROC5, IPSL, MRI, ACCESS, and GFDL, from top to bottom row.

https://doi.org/10.1371/journal.pone.0167506.s004

(EPS)

S5 Fig. Probability of a negative population rate of the Northern Pintail across the Mississippi and Atlantic Flyway, within latitudinal bands from 30°N to 50°N, for the late 20th, mid-21st, and late 21st centuries (first three columns).

The probabilities are shown for each day from 1 September through 31 March, with the x-axis labeled as days since 1 September. Projected changes in these probabilities are shown for the mid-21st and late 21st century, compared to the late 20th century (fourth and fifth columns). Results are shown for CNRM, MIROC5, IPSL, MRI, ACCESS, and GFDL, from top to bottom row.

https://doi.org/10.1371/journal.pone.0167506.s005

(EPS)

S1 File. Netcdf file containing underlying data from Fig 2A–2C.

https://doi.org/10.1371/journal.pone.0167506.s006

(NC)

S2 File. Netcdf file containing underlying data from Fig 2D–2F.

https://doi.org/10.1371/journal.pone.0167506.s007

(NC)

S3 File. Excel file containing underlying data from Fig 3.

https://doi.org/10.1371/journal.pone.0167506.s008

(XLSX)

S4 File. Netcdf file containing underlying data from Fig 4A–4C.

https://doi.org/10.1371/journal.pone.0167506.s009

(NC)

S5 File. Netcdf file containing underlying data from Fig 4D–4F.

https://doi.org/10.1371/journal.pone.0167506.s010

(NC)

S6 File. Netcdf file containing underlying data from Fig 5.

https://doi.org/10.1371/journal.pone.0167506.s011

(NC)

S7 File. Netcdf file containing underlying data from Fig 6.

https://doi.org/10.1371/journal.pone.0167506.s012

(NC)

S1 Table. Sample daily data for a location in Wisconsin (44.79°N, 89.95°W) to demonstrate the calculation of cumulative weather severity indices.

Here, the data includes the day since 1 September, daily mean air temperature (°C), daily snow depth (cm), four components of weather severity (TEMP, TEMPDAY, SNOW, and SNOWDAY), the PC1 index, and computed population rate of change for Mallards. Grey shading indicates conditions favorable for declining populations related to migration. It might be expected that Mallards would migrate south from this location around early December that year.

https://doi.org/10.1371/journal.pone.0167506.s013

(DOCX)

Acknowledgments

The research was funded by the Northeast Climate Science Center, National Oceanic and Atmospheric Administration, National Science Foundation, Great Lakes Restoration Initiative through the Environmental Protection Agency, Gulf Coastal Plains and Ozarks Landscape Conservation Cooperative, Long Point Waterfowl, Ontario Federation of Anglers and Hunters, and the University of Western Ontario through the Western Graduate Research Scholarship. Modeling contributions from Dr. Val Bennington are acknowledged. We acknowledge the World Climate Research Programme’s Working Group on Coupled Modelling, which is responsible for CMIP, and we thank the climate modeling groups for producing and making available their model output. For CMIP, the U.S. Department of Energy’s Program for Climate Model Diagnosis and Intercomparison provides coordinating support and led development of software infrastructure in partnership with the Global Organization for Earth System Science Portals. Computational resources were provided through the National Center for Atmospheric Research and the Teragrid from the University of Texas at Austin and the University of Illinois at Urbana–Champaign. Comments from Dr. Matthew Guillemain and an anonymous reviewer are greatly appreciated. Publication of the Nelson Institute Center for Climatic Research.

Author Contributions

  1. Conceptualization: MN YZ SV.
  2. Formal analysis: MN MS LVDE.
  3. Funding acquisition: MN MS.
  4. Investigation: YZ.
  5. Methodology: MN YZ SV MS LVDE.
  6. Project administration: MN.
  7. Visualization: MN.
  8. Writing – original draft: MN MS LVDE.
  9. Writing – review & editing: YZ SV JC CH.

References

  1. 1. Browne DM, Dell R, editors. Conserving waterfowl and wetlands amid climate change. Ducks Unlimited, Inc.; 2007.
  2. 2. Henderson E. Economic Impact of waterfowl hunting in the United States. Addendum to the 2001 survey of fishing, hunting, and wildlife-associated recreation. Report 2001–9. U.S. Fish and Wildlife Service. Div. of Economics. 13pp; 2005.
  3. 3. Grado SC, Kaminski RM, Munn IA, Tullos TA. Economic impacts of waterfowl hunting on public lands and at private lodges in the Mississippi Delta. Wildlife Society Bulletin. 2001;29: 846–855.
  4. 4. U.S. Fish & Wildlife Service. Birding in the United States: A demographic and economic analysis. Addendum to the 2011 National Survey of Fishing, Hunting, and Wildlife-Associated Recreation. Report 2011–1; 2013.
  5. 5. Baldassarre GA, Bolen EG. Waterfowl ecology and management. Second edition. Krieger Publishing, Malabar, FL; 2006.
  6. 6. Schummer ML, Cohen J, Kaminski RM, Brown ME, Wax CL. Atmospheric teleconnections and Eurasian snow cover as predictors of a weather severity index in relation to Mallards Anas platyrhynchos autumn-winter migration. Wildfowl. 2014;Special Issue No. 4: 451–469.
  7. 7. Nichols JD, Reinecke KJ, Hines JE. Factors affecting the distribution of mallards wintering in the Mississippi Alluvial Valley. The Auk. 1983;100: 932–946.
  8. 8. Johnson DH, Nichols JD, Schwartz MD. Population dynamics of breeding waterfowl. In: Ecology and Management of Breeding Waterfowl (ed. Batt BDJ), pp. 446–485, University of Minnesota Press, Minneapolis, MN; 1992.
  9. 9. Blums P, Nichols JD, Hines JE, Mednis A. Sources of variation in survival and breeding site fidelity in three species of European ducks. Journal of Animal Ecology. 2002;71: 438–450.
  10. 10. Mallory ML, Venier LA, McKenney D. Winter weather and waterfowl surveys in north-western Ontario, Canada. Journal of Biogeography. 2003;30: 441–448.
  11. 11. Greene AW, Krementz DG. Mallard harvest distributions in the Mississippi and Central Flyways. Journal of Wildlife Management. 2008;72: 1328–1334.
  12. 12. Schummer ML, Kaminski RM, Raedeke AH, Graber DA. Weather–related indices of autumn-winter dabbling duck abundance in middle North America. Journal of Wildlife Management. 2010;74: 94–101.
  13. 13. Gunnarsson G, Waldenström J, Fransson T. Direct and indirect effects of winter harshness on the survival of Mallards Anas platyrhynchos in northwest Europe. Ibis. 2012;154: 307–317.
  14. 14. Both C, Bouwhuis S, Lessells CM, Visser ME. Climate change and population declines in a long-distance migratory bird. Nature. 2006;441: 81–83. pmid:16672969
  15. 15. Thorup K, Tøttrup AP, Rahbek C. Patterns of phenological changes inmigratory birds. Oecologia. 2007;151: 697–703. pmid:17160398
  16. 16. Møller AP, Fiedler W, Berthold P (Eds.). Effects of climate change on birds.—Oxford University Press, Oxford, UK, 344 pp; 2011.
  17. 17. Guillemain M, Pöysä H, Fox AD, Arzel C, Dessborn L, Ekroos J, et al. Effects of climate change on European ducks: what do we know and what do we need to know? Wildlife Biology. 2013;19: 404–419.
  18. 18. Bellrose FC. Ducks, geese & swans of North America. Stackpole Books, Harrisburg, PA; 1980.
  19. 19. Ridgill SC, Fox AD. Cold weather movements of waterfowl in Western Europe. International Waterfowl and Wetlands Research Bureau Special Publication No. 13, Slimbridge, Gloucester, UK; 1990.
  20. 20. Švažas S, Patapavičius R, Dagys M. Recent changes in distribution of wintering populations of waterfowl established on the basis of Lithuanian ringing recoveries. Acta Zoologica Lituanica. 2001; 11: 235–242.
  21. 21. Schummer ML, Van Den Elsen L. Estimated effects of climate change on autumn-winter distributions of dabbling ducks in eastern North America: 2013 research update; 2013.
  22. 22. Baldassarre G. Ducks, geese, and swans of North America. Volume 1. Johns Hopkins University Press, Baltimore, MD; 2014.
  23. 23. Van Den Elsen LM. Weather and photoperiod indices of autumn and winter dabbling duck abundance in the Mississippi and Atlantic Flyways of North America. MS Thesis, University of Western Ontario, London, ONT; 2016.
  24. 24. Koskimies J, Lajti L. Cold-hardiness of the newly-hatched young in relation to ecology and distribution in ten species of European ducks. Auk. 1964;81: 281–307.
  25. 25. Talent LG, Jarvis RL, Krapu GL. Survival of Mallards broods in south-central North Dakota. Condo. 1983;85: 74–78.
  26. 26. Cox RR, Hanson MA, Roy CC, Euliss NH, Johnson DH, Butler MG. Mallard duckling growth and survival in relation to aquatic invertebrates. Journal of Wildlife Management. 1998;62: 124–133.
  27. 27. Anthes N. Long-distance migration timing of Tringa sand-pipers adjusted to recent climate change. Bird Study. 2004;51: 203–211.
  28. 28. Gunnarsson G, Elmberg J, Sjöberg K, Pöysä H, Nummi P. Why are there so many empty lakes? Food limits survival of Mallard ducklings. Canadian Journal of Zoology. 2004;82: 1698–1703.
  29. 29. Barshep Y, Meissner W, Underhill LG. Timing of migration of the Curlew Sandpiper (Calidris ferruginea) through Poland in relation to Arctic breeding conditions. Ornis Fenn. 2012;89: 120–129.
  30. 30. Dalby L, Fox AD, Petersen IK, Delany S, Svenning JC. Temperature does not dictate the wintering distribution of European dabbling duck species. Ibis. 2013;155: 80–88.
  31. 31. Guillemain M, Champagnon J, Massez G, Pernollet CA, George T, Momerency A, Simon G. Becoming more sedentary? Changes in recovery positions of Mallard Anas platyrhynchos ringed in the Camargue, France, over the last 50 years. Wildfowl. 2015;65: 51–63.
  32. 32. Guillemain M, Pernollet CA, Massez G, Cavallo F, Simon G, Champagnon J. Disentangling the drivers of change in Common Teal migration phenology over 50 years: Land use vs. climate change effects. Journal of Ornithology. 2015;156: 647–655.
  33. 33. Alerstam T. Bird migration. Cambridge University Press, Cambridge, United Kingdom; 1990.
  34. 34. Newton I. The migration ecology of birds. Oxford: Academic Press. 2008.
  35. 35. Calder WA. The consequences of body size for avian energetics. In: Avian Energetics (ed Paynter RA), pp. 86–151. Nuttall Ornithological Club, Cambridge, MA; 1974.
  36. 36. Bengtsson D, Avril A, Gunnarsson G, Elmberg J, Söderquist P, Norevik G, Tolf C, Safi K, Fiedler W, Wikelski M, Olsen B, Waldenström J. Movements, home-range size and habitat selection of Mallards during autumn migration. Plos One. 2014;9: 1–14.
  37. 37. Newton I. Weather-related mass mortality events in migrants. Ibis. 2007;149: 453–467.
  38. 38. Albright JJ, Owen RB Jr, Corr PO. The effects of winter weather on the behavior and energy reserves of black ducks in Maine. Transactions of the Northeast Section of the Wildlife Society. 1983;40: 118–128.
  39. 39. Jorde DG, Krapu GL, Crawford RD. Feeding ecology of mallards wintering in Nebraska. The Journal of Wildlife Management. 1983;47: 1044–1053.
  40. 40. Lovvorn JR. Nutrient reserves, probability of cold spells, and the question of reserve regulation in wintering Canvasbacks. Journal of Animal Ecology. 1994;63: 11–23.
  41. 41. Pearse AT. Design, evaluation and applications of an aerial survey to estimate abundance of wintering waterfowl in Mississippi. PhD Thesis. Mississippi State University, Mississippi State, MS; 2007.
  42. 42. Reinecke KJ, Kaminski RM, Moorhead DJ, Hodges JD, Nassar JR. Mississippi Alluvial Valley. In: Habitat Management for Migrating and Wintering Waterfowl in North America (eds. Smith LM, Pederson RL, Kaminski RM), pp. 230–247. Texas Tech University Press, Lubbock, TX; 1989.
  43. 43. Lensink CJ. Distribution of recoveries from bandings of ducklings. U.S. Fish and Wildlife Service, Special Scientific Report, Wildlife No. 89; 1964.
  44. 44. Bellrose FC, Crompton RD. Migrational behavior of mallards and black ducks as determined from banding. Illinois Natural History Survey Bulletin. 1970;30: 167–234.
  45. 45. Sauter A, Korner-Nievergelt F, Jenni L. Evidence of climate change effects on within-winter movements of European Mallards Anas platyrhynchos. Ibis. 2010;152: 600–609.
  46. 46. Nilsson L. The impact of hard winters on waterfowl populations of south Sweden. Wildfowl. 1984;35: 71–80.
  47. 47. Scott DA, Rose PM. Atlas of Anatidae Populations in Africa and Western Eurasia. Oxford: Wetlands International; 1996.
  48. 48. Kershaw M. Common Pochard (Pochard) Aythya ferina. In: The Migration Atlas: Movements of the Birds of Britain and Ireland (eds. Werham C, Toms M, Marchant J, Clark J, Siriwardena G, Baillie S), pp. 204–207. London: T. & A. Poyser; 2002.
  49. 49. Suter W. Überwinternde Wasservögel auf Schweizer Seen: Welche Gewässereigenschaft bestimmen Arten- und Individuenzahl? Ornithologische Beobachter. 1991;88: 111–140.
  50. 50. Keller V. Entwicklung der Wasservogelbesta¨nde in den Wasservogelreservaten von internationaler Bedeutung 1992 ⁄ 93–2002 ⁄ 03: Eine Bilanz. Sempach: Schweizerische Vogelwarte; 2005.
  51. 51. Hartmann DL, Klein Tank AMG, Rusticucci M, Alexander LV, Brönnimann S, Charabi Y, Dentener FJ, Dlugokencky EJ, Easterling DR, Kaplan A, Soden BJ, Thorne PW, Wild M, Zhai PM. Observations: Atmosphere and Surface. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker T.F., Qin D., Plattner G.-K., Tignor M., Allen S.K., Boschung J., Nauels A., Xia Y., Bex V. and Midgley P.M. (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY; 2013.
  52. 52. Thomas CD, Lennon JL. Birds extend their ranges northwards. Nature. 1999;399: 213.
  53. 53. Hughes L. Biological consequences of global warming: is the signal already apparent? Trends in Ecology and Evolution. 2000;15: 56–61. pmid:10652556
  54. 54. Price JT, Root TL. Focus: effects of climate change on bird distributions and migration patterns. In: Preparing for a Changing Climate: the Potential Consequences of Climate Variability and Change (eds. Sousounis PJ, Bisanz JM), pp. 65–68, University of Michigan Atmospheric, Oceanic, and Space Sciences Department, Ann Arbor, MI; 2000.
  55. 55. Price JT, Root TL. No orioles in Baltimore? Climate change and neotropical migrants. Bird Conservation. 2002;17: 12.
  56. 56. McCarty JP. Ecological consequences of recent climate change. Conservation Biology. 2001;15: 320–331.
  57. 57. Sparks TH, Mason CF. Dates of arrivals and departures of spring migrants taken from the Essex Bird Reports 1950–1998. Essex Bird Report 1999, 154–164; 2001.
  58. 58. Hatzofe O, Yom-Tov Y. Global warming and recent changes in Israel’s avifauna. Israel Journal of Zoology. 2002;48: 351–357.
  59. 59. Walther G-R, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, Fromentin J.-M., Hoegh-Guldberg O, Bairlein F. Ecological responses to recent climate change. Nature. 2002;416: 389–395. pmid:11919621
  60. 60. Parmesan C, Yohe G. A globally coherent fingerprint of climate change impacts across natural systems. Nature. 2003;421: 37–42. pmid:12511946
  61. 61. Brommer JE. The range margins of northern birds shift polewards. Annual Zoological Fennici. 2004;41: 391–397.
  62. 62. Crick HQP. The impact of climate change on birds. Ibis. 2004;146: 48–56.
  63. 63. Hitch AT, Leberg PL. Breeding distributions of North American bird species moving north as a result of climate change. Conservation Biology. 2006;21: 534–539.
  64. 64. Sedinger JS, Ward DH, Schamber JL, Butler WI, Eldridge WD, Conant B, Voelzer JF, Chelgren ND, Herzog MP. Effects of El Niño on distribution and reproductive performance of Black Brant. Ecology 2006;87: 151–159. pmid:16634306
  65. 65. Huntley B, Green R, Collingham Y, Willis SG. A Climatic Atlas of European Breeding Birds.–Lynx Edicions, Barcelona, Spain, 521 pp; 2007.
  66. 66. La Sorte FA, Thompson FR III. Poleward shifts in winter ranges of North American birds. Ecology. 2007;88: 1803–1812. pmid:17645026
  67. 67. Brook RW, Kenyon Ross R, Abraham KF, Fronczak DL, Chris Davies J. Evidence for black duck winter distribution change. Journal of Wildlife Management. 2009;73: 98–103.
  68. 68. Švažas S, Meissner W, Serebryakov V, Kozulin A, Grishanov G. Changes in wintering sites of waterfowl in central and eastern Europe. Oiseaux Migrateurs du Palearctique Occidental (OMPO) Special Publication. OMPO Vilnius & Lithuanian Institute of Ecology, Vilnius, Luthuania. 2001.
  69. 69. Nilsson L. Changes in numbers and distribution of wintering waterfowl in Sweden during forty years, 1967–2006. Ornis Svec. 2008;18: 135–226.
  70. 70. Lehikoinen A, Jaatinen K, Vähätola AV, Clausen P, Crowe O, Deceunick B, Hearn R, Holt CA, Hornman M, Keller V, Nilsson L, Langendoen T, Tománkova I, Wahl J, Fox AD. Rapid climate driven shifts in wintering distributions of three common waterbird species. Global Change Biology. 2013;19: 2071–2081. pmid:23509023
  71. 71. Palmer RS. Handbook of North American birds. Volume 2: Waterfowl (part 1). Yale University Press, New Haven, CT; 1976.
  72. 72. Rusch DH, Ankney CD, Boyd H, Longcore JR, Montalbano F III, Ringelman JK, Stotts VD. Population ecology and harvest of the American black duck: a review. Wildlife Society Bulletin. 1989;17: 379–406.
  73. 73. Nudds TD, Miller MW, Ankney CD. Black ducks: harvest, mallards or habitat. In: 7th International Waterfowl Symposium (ed. Ratti JT), pp. 50–60. Institute for Wetland and Waterfowl Research, Memphis, TN; 1996.
  74. 74. Longcore JR, McAuley DG, Hepp GR, Rhymer JM. American black duck (Anas rubripes). Account 481. In The birds of North America (eds. Pool A, Gill F). The Academy of Natural Sciences, Philadelphia, PA, and The American Ornithologists’ Union, Washington, D.C.; 2000.
  75. 75. Dahl TE. Wetland losses in the United States 1780’s to 1980’s. U. S. Department of the Interior, Fish and Wildlife Service, Washington, D.C.; 1990.
  76. 76. Dahl TE. Status and trends of wetlands in the conterminous United States 1998 to 2004. U.S. Department of the Interior, Fish and Wildlife Service, Washington, D.C.; 2006.
  77. 77. Newton I. Population limitation in birds. Academic Press, San Diego, California, USA; 1998.
  78. 78. Inkley DB, Anderson MG, Blaustein AR, Burkett VR, Felzer B, Griffith B, Price J, Root TL. Global climate change and wildlife in North America, Wildlife Society Technical Review 04–2. The Wildlife Society, Bethesda, MD; 2004.
  79. 79. Temple SA, Cary J. Climatic effects on year-to-year variations in migration phenology: a WSO research project. Passenger Pigeon. 1987;49: 70–75.
  80. 80. Gwinner E. Circannual rhythms in birds. Current Opinion in Neurobiology. 2003;13: 770–778. pmid:14662381
  81. 81. Notaro M, Lorenz D, Hoving C, Schummer M. Twenty-first-century projections of snowfall and winter severity across central-eastern North America. Journal of Climate. 2014;27: 6526–6550.
  82. 82. Wang J, Bai X, Hu H, Clites A, Colton M, Lofgren B. Temporal and spatial variability of Great Lakes ice cover, 1973–2010. Journal of Climate. 2012;25: 1318–1329.
  83. 83. Van Cleave K, Lenters JD, Wang J, Verhamme EM. A regime shift in Lake Superior ice cover, evaporation, and water temperature following the warm El Niño winter of 1997–1998. Limnology and Oceanography. 2014;59: 1889–1898.
  84. 84. Burnett AW, Kirby ME, Mullins HT, Patterson WP. Increasing Great Lake–effect snowfall during the twentieth century: A regional response to global warming? Journal of Climate. 2003;16: 3535–3542.
  85. 85. Kunkel KE, Ensor L, Palecki M, Easterling D, Robinson D, Hubbard KG, Redmond K. A new look at lake-effect snowfall trends in the Laurentian Great Lakes using a temporally homogeneous data set. Journal of Great Lakes Research. 2009;35: 23–29.
  86. 86. Notaro M, Bennington V, Vavrus S. Dynamically downscaled projections of lake-effect snow in the Great Lakes Basin. Journal of Climate. 2015a;28: 1661–1684.
  87. 87. Notaro M, Bennington V, Lofgren B. Dynamical downscaling-based projections of Great Lakes water levels. Journal of Climate. 2015b;28: 9721–9745.
  88. 88. Burnham KP, Anderson DR. Model selection and multimodel interference: a practical information-theoretic approach. Springer-Verlag, New York, NY; 2002.
  89. 89. Arnold TW. Uninformative parameters and model selection using Akaike’s Information Criterion. Journal of Wildlife Management. 2010;74: 1175–1178.
  90. 90. Rebetez M. Seasonal relationship between temperature, precipitation and snow cover in a mountainous region. Theoretical and Applied Climatology. 1996;54: 99–106.
  91. 91. Giorgi F, Coppola E, Solmon F, Mariotti L, Sylla MB, Bi X, et al. RegCM4: Model description and preliminary tests over multiple CORDEX domains. Climate Res. 2012;52: 7–29.
  92. 92. Hostetler SW, Bartlein PJ. Simulation of lake evaporation with application to modeling lake-level variations at Harney-Malheur Lake, Oregon. Water Resources Research. 1990; 26: 2603–2612.
  93. 93. Patterson JC, Hamblin PF. Thermal simulation of a lake with winter ice cover. Limnology and Oceanography. 1988;33: 323–338.
  94. 94. Hostetler SW. Simulation of lake ice and its effect on the late-Pleistocene evaporation rate of Lake Lahontan. Climate Dynamics. 1991;6: 43–48.
  95. 95. van Vuuren DP, Edmonds J, Kainuma M, Riahi K, Thomson A, Hibbard K, et al. The representative concentration pathways: an overview. Climatic Change. 2011;109: 5–31.
  96. 96. Notaro M, Zarrin A, Vavrus S, Bennington V. Simulation of heavy lake-effect snowstorms across the Great Lakes Basin by RegCM4: Synoptic climatology and variability. Monthly Weather Review. 2013a;141: 1990–2014.
  97. 97. Notaro M, Holman K, Zarrin A, Fluck E, Vavrus S, Bennington V. Influence of the Laurentian Great Lakes on regional climate. Journal of Climate. 2013b;26: 789–804.
  98. 98. Vavrus S, Notaro M, Zarrin A. The role of ice cover in heavy lake-effect snowstorms over the Great Lakes Basin as simulated by RegCM4. Monthly Weather Review. 2013;141: 148–165.
  99. 99. Bennington V, Notaro M, Holman KD. Improving climate sensitivity of deep lakes within a regional climate model and its impact on simulated climate. Journal of Climate. 2014;27: 2886–2911.
  100. 100. Thornton PE, Running SW. An improved algorithm for estimating incident daily solar radiation from measurements of temperature, humidity, and precipitation. Agricultural and Forest Meteorology. 1999;93: 211–228.
  101. 101. Thornton PE, Thornton MM, Mayer BW, Wilhelmi N, Wei Y, Devarakonda R, Cook RB. Daymet: Annual Climate Summaries on a 1-km Grid for North America, Version 2. ORNL DAAC, Oak Ridge, TN.
  102. 102. Menne MJ, Durre I, Vose RS, Gleason BE, Houston TG. An overview of the Global Historical Climatology Network-Daily Database. Journal of Atmospheric and Oceanic Technology. 2012;29: 897–910.
  103. 103. Räty O, Raisänen J, Ylhäisi . Evaluation of delta change and bias correction methods for future daily precipitation: Intermodel cross-validation using ENSEMBLES simulations. Climate Dynamics. 2014;42: 2287–2303.
  104. 104. Bossenmaier EF, Marshall WH. Field- feeding by Waterfowl in Southeastern Manitoba. Wildlife Monographs, 1, The Wildlife Society, Washington, D.C.; 1958.
  105. 105. Anderson HG. Food habits of migratory ducks in Illinois. Illinois Natural History Survey Bulletin. 1959;27: 289–344.
  106. 106. Havera SP. Waterfowl of Illinois: Status and Management. Special Publication 21. Illinois Natural History Survey, Urbana; 1999.
  107. 107. Carter Johnson W, Millett BV, Gilmanov T, Voldseth RA, Guntenspergen GR, Naugle DE. Vulnerability of northern prairie wetlands to climate change. BioScience. 2005;55: 863–872.
  108. 108. Fuller K, Shear H, Wittig J, editors. The Great Lakes: An Environmental Atlas and Resource Book. U.S. Environmental Protection Agency and Environment Canada. Great Lakes National Program Office, Chicago, IL. Available online at http://www.epa.gov/glnpo/atlas/glat-ch2.html (accessed July, 2007); 1995.