Hydrobiologia DOI 10.1007/s10750-016-2817-4

PRIMARY RESEARCH PAPER

Connecting the trophic dots: responses of an aquatic bird species to variable abundance of macroinvertebrates in northern boreal wetlands K. E. B. Gurney . R. G. Clark . S. M. Slattery . L. C. M. Ross

Received: 22 January 2016 / Revised: 5 May 2016 / Accepted: 6 May 2016 Ó Springer International Publishing Switzerland 2016

Abstract To evaluate variation in abundance of boreal wetland macroinvertebrates and test for effects of this variation on the diet and habitat use of a bird species that consumes aquatic invertebrates (lesser scaup, Aythya affinis), we collected macroinvertebrates and birds and conducted bird surveys (June–August) at two wetland complexes in northwestern Canada and assessed diet composition using an isotopic approach. At both study areas, for macroinvertebrate taxa reported to be key prey items for scaup, biomass varied intra-seasonally and annually, but patterns differed among taxa and between areas. Macroinvertebrate biomass varied strongly across wetlands within study areas, and isotopic mixing models indicated that local heterogeneity in macroinvertebrate biomass was reflected in duckling diets, which also

varied across wetlands, indicating a generalist foraging strategy for this species. Wetland habitats used by broodrearing female scaup had greater amphipod and gastropod biomasses. Our results show that boreal wetland macroinvertebrate abundances vary considerably across coarse and fine spatial scales. Female scaup and their ducklings appear well adapted to exploit this dynamic food resource, but overall productivity of scaup may depend on the abundances of certain taxa, suggesting that conservation efforts should focus on maintaining abundant populations of key wetland invertebrates. Keywords Breeding birds  Habitat use  Stable isotopes  Lesser scaup  Temporal variation  Diet

Handling editor: M. Power

Electronic supplementary material The online version of this article (doi:10.1007/s10750-016-2817-4) contains supplementary material, which is available to authorized users. K. E. B. Gurney  R. G. Clark Department of Biology, University of Saskatchewan, Saskatoon, Canada

L. C. M. Ross Native Plant Solutions, Ducks Unlimited Canada, Winnipeg, Canada

R. G. Clark Science and Technology Branch, Environment and Climate Change Canada, Saskatoon, Canada

Present Address: K. E. B. Gurney (&) Science and Technology Branch, Environment and Climate Change Canada, 115 Perimeter Road, Saskatoon, SK S7N 0X4, Canada e-mail: [email protected]

S. M. Slattery  L. C. M. Ross Institute for Wetland and Waterfowl Research, Ducks Unlimited Canada, Stonewall, Canada

123

Hydrobiologia

Introduction The boreal biome in North America has an estimated area of 5.9 9 106 km2 and is characterized by an abundance of diverse aquatic ecosystems, including approximately 8 9 105 km2 of lentic and lotic freshwater habitats. Rivers, streams, deltas, lakes, and—in many areas—open water wetlands, ranging in size from \1 hectare to complexes spanning several thousand km2, dominate this landscape (Wells et al., 2011). Wetlands in the North American boreal, as elsewhere, provide important ecosystem services (carbon storage, flood control), and the wetlands found in northern reaches of the western boreal, in particular, are highly productive as indexed by levels of chlorophyll-a, compared to eastern boreal wetlands (Zoltai, 1988). As such, these habitats support large populations of wildlife that forage heavily on aquatic macroinvertebrates (hereafter invertebrates) (Slattery et al., 2011). Unlike inland waters further south or in the boreal forest of Fennoscandia, however, there has been limited study of aquatic ecology in the North American boreal region (hereafter the boreal). Temporal and spatial patterns of abundance for boreal wetland invertebrate communities are poorly quantified, and there are scant data linking changes in these communities to responses in upper trophic levels. Such information is necessary to accurately predict how wetland-dependent wildlife will respond to continuing resource development and changing environmental conditions across the boreal (Webster et al., 2015). Climate, geology, and to lesser extent, hydrology exert the strongest controls on wetland invertebrate communities in non-boreal locations (Batzer & Ruhı´, 2013). Some evidence suggests that boreal wetland invertebrates may be similarly regulated (Hornung & Foote, 2006; Corcoran et al., 2009), but wetland invertebrates can be highly plastic in their life cycles, so that variation in invertebrate abundance in boreal wetlands might be observed only when changes in environmental conditions are extreme (Feuchtmayr et al., 2007; Batzer, 2013). Like other temperatebreeding organisms, invertebrates in northern boreal wetlands should respond strongly to the onset of favourable environmental conditions in spring. For some species, short growing seasons in these locations may constrain reproduction, with only one generation produced during the summer (i.e. univoltinism). Other

123

species are multivoltine and can produce multiple generations in one season, with recent evidence suggesting that this life history strategy is becoming more prevalent among invertebrates in northern regions as ambient temperatures increase (Rautio et al., 2008; Po¨yry et al., 2011). Changes in aquatic invertebrate abundance can influence their predators, including waterbirds, although among avian taxa the responses to changes in food resources are highly variable. In some species, fluctuating prey availability is associated with a shift in diet, but some species are less flexible, consistently consuming the same prey items, regardless of the relative abundances of those items in the environment (Robbins et al., 2007; Lewis et al., 2008). At a coarser geographic scale, variation in food resources can modify the spatial distribution of birds on the landscape. Northern breeding birds—adults and juveniles—tend to have relatively high nutritional and energetic demands, and densities of freshwater avifauna often increase with the increasing productivity of aquatic habitats (Schekkerman et al., 2003; Epners et al., 2010). Food resources, however, are not necessarily limiting in all freshwater systems, and top-down processes can also influence habitat use (Dessborn et al., 2011; Nummi et al., 2013). Temporal variation in food resources—within and among years—might also affect wetland use by aquatic bird species, particularly in northern locations, where birds have limited time to breed and late reproduction can have fitness consequences (Gurney et al., 2012; Hayden et al., 2015). In boreal wetlands, however, no studies have yet evaluated how strongly intraseasonal fluctuations in invertebrate prey impact upper trophic levels. To address questions related to invertebrate ecology and trophic connections in boreal wetlands, we collected field data over 6 years at two wetland complexes in northwestern Canada. Our specific objectives were to (i) evaluate changes in abundances of wetland invertebrates that are important food resources for wildlife, within season and inter-annually, and (ii) determine how such variation affects both the diet and habitat use of a wetland-dependent bird species, the lesser scaup, Aythya affinis (Eyton, 1838) (hereafter scaup). Scaup are ideal models for studying boreal wetland food webs as they breed primarily in the western boreal and rely extensively on freshwater

Hydrobiologia

invertebrates for successful reproduction (Baldassarre, 2014). In addition, during recent declines of the scaup population, juveniles from northern boreal areas were less likely to recruit to the breeding population than juveniles from southern locations, and—from a conservation standpoint—potential links between changes in aquatic invertebrate communities, dietary patterns of scaup ducklings and habitat use warrant further investigation (Corcoran et al., 2009; Hobson et al., 2009). We expected that invertebrates we sampled might include species with both uni- and multivoltine life histories—anticipating that, at a coarse taxonomic level, prey abundance in these aquatic systems would not have the characteristic seasonal prey peaks observed in other, primarily terrestrial, habitats (Strode, 2009; Arnold et al., 2010). In particular, repeated production of new cohorts by multivoltine species should dampen the seasonal peaks characteristic of univoltine species, weakening unimodal seasonal patterns of overall invertebrate abundance in boreal wetlands (Pehrsson & Nystrom, 1988; Foxi & Delrio, 2010). Furthermore, if reproduction and phenology of invertebrates in northern systems are flexible with respect to environmental conditions, then we predicted that abundances of invertebrates in our study would vary across years and that seasonal patterns of abundance would fluctuate annually (Hornung & Foote, 2006; Hanson et al., 2009). To contend with variable patterns of wetland invertebrate prey abundance, we propose that birds that feed on these organisms are dietary generalists that optimize their foraging effort by consuming different prey species as invertebrate abundances vary. We therefore expected that diets of scaup ducklings hatched later in the season would be distinct from those of earlier hatched individuals, reflecting seasonal changes in abundances of different invertebrate species. We also anticipated that wetlands used by brood-rearing female scaup would be those with greater abundances of invertebrate foods. Equally importantly, if female scaup select wetlands that provide consistent resources for their offspring throughout the period of brood-rearing, an idea that is mostly untested, then abundances of the specific food items consumed by ducklings should show less extreme seasonal fluctuations in wetlands used by broods (Po¨ysa¨ et al., 2000).

Methods Study areas We studied seasonal and annual variation in quantity of wetland invertebrates, as well as diets and distributions of pre-fledging scaup, at Cardinal Lake, Northwest Territories (67°360 N, 133°390 W; Gurney et al., 2012) during 2003 and 2005–2007. The Cardinal Lake Study Area (Fig. 1A) encompasses an area of approximately 1.5 9 104 hectares, where study wetlands were randomly selected to represent a range of wetland sizes and types, including those that were known to be used by scaup broods (brood ponds) as well as those where broods were never observed during regular brood surveys (non-brood ponds). Cardinal Lake wetlands were generally shallow (\2 m) and mesotrophic; values for mean total phosphorus were 11–19 lg/l, and the range of mean total nitrogen was 660–1405 lg/l. Characteristic wetland macrophytes included Equisetum palustre L. (marsh horsetail), Carex aquatilis Wahl (water sedge), Meyanthes trifoliata L. (buck-bean), and Potamogeton spp. (pondweed). We also examined temporal patterns in quantity of freshwater invertebrate food items at a second boreal wetland complex that surrounds Utikuma Lake in northcentral Alberta (55°540 N, 115°020 W) during 2001 and 2002. The Utikuma Research Study Area (Fig. 1B) covers approximately 1.0 9 104 hectares, and standing water bodies include lakes, shallow ponds between 0.01 and 0.05 km2, and poorly drained fens and bogs that are dominated by Potamogeton spp., Myriophyllum exalbescens Fernald (water milfoil), or Ceratophyllum demersum L. (hornwort) (Wiken et al., 1996; Sass et al., 2007). At this study area, wetlands, including both brood and nonbrood ponds, were selected to represent different geological landscapes: clay till plain, moraine, and outwash (Smerdon et al., 2005). Sampled wetlands were generally shallow (\2 m) and eutrophic (mean total phosphorus of 95 lg/l; mean total nitrogen of 2814 lg/l) (Hornung & Foote, 2006). Field work To evaluate temporal and spatial heterogeneity in wetland invertebrate abundance, between May and September, samples were collected from 30 wetlands at Cardinal Lake (2003, 2005–2007; n = 1835

123

Hydrobiologia Fig. 1 Aquatic invertebrates were collected from wetlands at two geographically distinct study areas in the boreal forest of western North America: A/Cardinal Lake and B/Utikuma Lake. Surveyed wetlands are coloured black, and the white triangles indicate those wetlands used by lesser scaup Aythya affinis broods

samples) and from 28 wetlands at Utikuma Lake (2001, 2002; n = 2136 samples). Sample collections at Cardinal Lake were limited to three distinct time intervals, reflecting specific periods during the reproductive period: nesting, brood-rearing and fledging. At Utikuma Lake, sampling occurred at 10 day intervals. At both study areas, sampling periods were annually variable, depending on nesting phenology, but within years, were fixed across wetlands. To the extent possible, we maintained a similar sampling effort across and within years, as identified by the total number of samples collected (see Tables S2 and S3).

123

To collect samples, we pulled a D-frame net (500 l mesh, surface area = 0.0604 m2) vertically through the water column—at randomly selected open water (stratified by depth) and emergent vegetation locations. All areas sampled were sufficiently shallow so that both nektonic and benthic species were collected. Water depth, as well as type and percentage of vegetation (emergent and submerged), was recorded for each sample. Invertebrates were identified to family level and enumerated in the field when possible (Thorp & Covich, 2001; Merritt et al., 2008). Each specimen was assigned a size class based on

Hydrobiologia

previously determined size distributions for each taxon and life stage (i.e. adult, larvae, nymph, pupae) (Murkin & Ross, 1999). When logistical constraints prevented processing of samples in the field, invertebrates were preserved in 70% ethanol and processed at a later date. At Cardinal Lake, between 2005 and 2007, we also collected samples of reported invertebrate dietary components for stable isotope analyses (SIA) by towing a D-frame net horizontally through the water at varying depths in multiple wetlands (n = 30) and preserving collections in 70% ethanol. To characterize assimilated diets of pre-fledging scaup at Cardinal Lake, in 2005 and 2006, we collected 49 ducklings over two sampling periods, one period in early to mid-August and the second in late August to early September of each year, depending on breeding chronology. We attempted to collect ducklings (one or two per brood) only after they had been observed feeding. Ducklings were shot, carcasses were immediately retrieved and ethanol (70%) was injected into the oesophagus to help preserve contents of the upper gastrointestinal tract (UGT, i.e. oesophagus, proventriculus and gizzard). We assigned an approximate age to each duckling based on plumage development (Gollop & Marshall, 1954) and then froze samples. At Cardinal Lake, to determine wetland use patterns, we conducted regular surveys for adult female scaup (and their broods) that were marked as part of a survival study, as well as for broods attended by unmarked females. For marked females with broods, surveys began the day after hatch and continued at approximately 7 day intervals until the brood reached 30 days old, was lost (no surviving ducklings), or became permanently part of a cre`che (see Gurney et al., 2012 for details). At both study areas, surveys for unmarked broods were also conducted every 7–10 days throughout the breeding season. Multiple wetlands were visited each day and surveyed for 1 h in the morning and then 1 h in the afternoon—two observers concealed themselves in the vegetation surrounding the wetland and used a 20–609 spotting scope to scan the wetland for broods. At Utikuma Lake, we also conducted two aerial surveys (using a Bell 206 L helicopter with bubble side windows) during the brood-rearing period (July to September) to determine if scaup broods were present in study ponds. Depending on the habitat-type and specific topography, flight altitude was 15–50 m

above the wetland, and flight speed varied from 60 to 100 km/h. Laboratory work For conventional dietary assessments, contents of the UGT were removed by dissection and dried to constant mass (60°C). We weighed dry contents to the nearest 0.1 mg and then calculated a mean percent aggregate dry biomass (PDMi) for each prey item, i (Strand et al., 2008). Results from UGT content analyses were used to guide selection of items for SIA; the five taxa with the highest PDMi (combined, accounting for 91.2% of total dry biomass) were considered important food items. To determine assimilated diet over different time periods, samples of two tissue types (leg muscle and liver) were removed from collected carcasses and processed for SIA (carbon and nitrogen) (Phillips & Eldridge, 2005). We rinsed all SIA samples (leg muscle, liver tissue and invertebrates) with a 2:1 chloroform–methanol solution to remove lipids and then dried the samples to constant mass at 60 C. Dried samples were ground to a homogenous powder and weighed (approximately 1 mg) into tin cups for analysis at the Department of Soil Science, University of Saskatchewan. Standard continuous flow–isotope ratio mass spectrometry techniques were employed, with measurements made using a Delta V mass spectrometer interfaced with a Costech (Milan, Italy) ECS4010 elemental analyzer. By convention, isotope values are reported per mil (%) in delta notation as d13C and d15N according to   Rsample dX ¼  1 &; Rstd where X = 13C or 15N, R = 13C:12C or 15N:14N, and standards (std) were Vienna PeeDee Belemnite (VPDB) for 13C and atmospheric N2 (AIR) for 15N. Two secondary isotopic reference materials (egg albumen and lyophilized bowhead whale baleen) were used as standards. Standards (in the same mass range as the test samples) were processed before and between every five tissue samples within each analytical run (for albumen) and at the start, midpoint and end of each run (for baleen). We used replicate standard measurements within a run to estimate analytical precision: mean d13C (± 1 standard deviation, SD) within runs = –22.4 ± 0.0%, and mean d15N (± 1 SD) within runs = 6.8 ± 0.1% for egg

123

Hydrobiologia

albumen; mean d13C (±1 SD) within runs = -18.5 ± 0.1%, and mean d15N (±1 SD) within runs = 14.8 ± 0.2%, for bowhead whale baleen. Measurement error estimates (±1 SD) were ±0.1% for d13C and ±0.3% for d15N, based on measurements of laboratory standards across multiple runs (Bond & Hobson, 2012). Statistical analyses For each D-frame net sample, we estimated dry biomass (mg) by multiplying the number of individuals for each taxon, size class and life stage by a corresponding mean dry mass (Murkin & Ross, 1999). Total sample counts and biomass values were obtained by summing all estimated values for a given taxon. These data followed a negative binomial distribution, but tests for zero inflation were negative (Warton, 2005). Therefore, to evaluate temporal and spatial patterns in abundance of selected invertebrates, we used generalized linear models with generalized estimating equations (i.e. marginal models) implemented in SASÒ Version 9.4 (PROC GENMOD)(SASÒ Inc., Cary, NC). These models account for covariance in the response variable from dependencies among repeated measures from the same wetland while treating other covariates as fixed effects, and allow inference at the population level (Stroup, 2012). However, because the quasi-likelihood-based model selection criterion (QIC) for marginal models is biased and often untrustworthy (Hardin & Hilbe, 2003; Koper & Manseau, 2009), instead of using model selection, we developed a tightly focused set of candidate models for each study area and evaluated the influence of selected predictor variables in a step-wise fashion. For Cardinal Lake, each taxon-specific candidate set included 6 models; for Utikuma Lake, 13 models were included in each set—model combinations for each study area are available as Supporting Information (Table S1). Response variables were biomass or count of selected coarse-level taxonomic groups, based on previous dietary assessments for scaup (Sugden, 1973), as well as our UGT findings. Predictor variables included only known factors of influence, such as habitat zone (i.e. emergent vegetation versus open water) and factors related to our hypotheses of interest: year, timing of collection, and presence or absence of broods (BroodPA). To evaluate seasonal patterns at

123

Cardinal Lake and test for non-linear relationships, we included timing of collection as a categorical variable (i.e. BroodPeriod—nesting, brood-rearing, fledging). Because non-brood ponds were not sampled in every year at this wetland complex, we also included a nested term for brood use within year (BroodPA (Year)) in Cardinal Lake abundance models. For Utikuma Lake (where non-brood ponds were sampled in both years), a BroodPA*Year interaction term was included in model sets to test whether the relationship between brood use and invertebrate abundance was annually variable. More regular invertebrate sampling throughout the season at Utikuma Lake facilitated a finer scale assessment of non-linear seasonal patterns; collection day (CDay) and the quadratic term (CDay2) were included as continuous covariates in these models. All numerical predictor variables were standardized (converted to z scores) to deal with potential issues related to differing scales. In addition, because invertebrate abundance increased as sample volume increased, for each sample, we calculated volume as the surface area of the net times sample depth and included log(volume) as an offset in all models. The same covariance structure, i.e. wetland nested within year (Wetland (Year), compound symmetry), was maintained for all models to ensure that parameter estimates were comparable across models (Stroup, 2015). Confidence intervals (95%, CI) for regression coefficients (b) were calculated based on empirically based standard errors, which are more robust to violations of independence than model-based error estimates (Koper & Manseau, 2009). To evaluate simultaneously whether d13C and d15N values for invertebrate sources varied by taxonomic group and to ascertain other factors that might influence isotopic composition (d13C and d15N) of aquatic invertebrates or ducklings (leg muscle and liver tissues), we analysed raw isotopic data in SASÒ Version 9.4, using a multivariate analysis of variance (PROC GLM; MANOVA) (SASÒ Inc., Cary, NC). In these descriptive analyses, we ran one global model each for invertebrates and ducklings that evaluated the effects of selected predictor variables simultaneously. Four predictor variables (year, timing of collection, wetland, taxon) were assessed in our multivariate isotope model for invertebrates, and three predictor variables (year, timing of collection, and wetland of origin) were included for ducklings. Data comprised isotope values from 347 invertebrate samples (19 taxa;

Hydrobiologia

5 functional groups) and from 43 (leg muscle) or 38 (liver tissue) ducklings with estimated ages between 10 and 38 days old in these analyses. Our isotopic dietary assessment—focusing on wetland invertebrate taxa with the highest PDMi—included 197 samples of possible food items. To determine whether the different food sources we selected for our dietary assessment should be grouped or considered separately, we used a Bayesian mixing model framework to evaluate the optimal number of source contributions (Ward et al., 2011). Scaup duckling leg tissues were used to represent the consumer isotopic signature, as sample sizes were higher than for liver tissues. Using hierarchical Bayesian isotopic mixing models, implemented in the open-source R package, MixSIAR Version 2.1 (Stock & Semmens, 2013), we then estimated the relative contributions of invertebrate groups to pre-fledging scaup tissues. Within MixSIAR, Gibbs sampling is conducted for each model in JAGS— sampling parameters were set to 3 parallel chains of 50,000 vectors (burn-in = 25,000, thinning interval = 25). Because wetland was an important source of variation in invertebrate abundance, our mixing models evaluated variation in diet among birds from different wetlands, as well as among individuals, between years and throughout the brood-rearing period, using hatch date classes (i.e. early, peak or late) to index seasonal diet patterns. To account for a potential confounding effect of age with hatch date, we used only individuals estimated to be between 17 and 38 days old in these analyses, and two ducklings from a wetland where invertebrate isotope values were distinct were also censored so that our final sample of consumers included thirty-three ducklings from three different wetlands (sample sizes ranged from 6 to 14 ducklings per wetland). To determine whether mixing models converged, we examined trace plots for relatively constant mean and variance, as well as for small fluctuations in the estimated parameters—results from the Gelman-Rubin diagnostic test were also considered (Stock & Semmens, 2013). Different specifications of mixing models that converged were ranked using the Deviance Information Criterion (Spiegelhalter et al., 2014). Trophic enrichment factors used in mixing models were determined during controlled feeding studies of captive scaup ducklings: -0.44, SD = 0.4 for d13C and 4.1, SD = 0.2 for d15N (K. Gurney, unpublished data).

Results Taxonomically detailed summaries of invertebrate abundance (total estimated biomass) are available as Supporting Information (see Tables S2 and S3). Gammarid amphipods, Gammarus lacustris (Sars, 1864), were consistently the most abundant invertebrates in the Cardinal Lake wetlands, contributing 24% (2005) to 42% (2007) of the total measured biomass. Based on biomass, other key invertebrate fauna in these habitats included clam shrimp (Lynceus spp.)(26% in 2006), planorbid snails (21% in 2007), and fairy shrimp, dominated by Polyartemiella spp. (16% in 2003). At Utikuma Lake, our more southern wetland complex, in both years, the dominant invertebrate taxa were gammarid amphipods (Gammarus lacustris) and predacious beetles (Dytiscidae), at 26 and 17% of total biomass, respectively, in 2001, and 26 and 14% in 2002. Talitrid amphipods, Hyalella azteca (Saussure, 1858) also comprised a relatively large proportion of the total biomass at Utikuma Lake wetlands: 9% in 2001 and 14% for 2002, with the total combined amphipod biomass equal to 35% in 2001 and 40% in 2002. Temporal heterogeneity in abundance of wetland invertebrates At both study areas, the abundance of selected invertebrate taxa varied over time—and observed patterns were similar whether we considered count or biomass as a response variable. We therefore present results only for the latter. A complete listing of slope estimates for key covariates is available as Supporting Information (Tables S4–S13), and a summary of the main effects is presented in Table 1. Our data highlighted within season patterns of abundances for invertebrates at both study areas (with the exception of Diptera at Utikuma Lake) that varied across taxa. Contrary to our expectation of seasonally consistent biomass, there was evidence of seasonal peaks for some groups (Branchiopoda at both areas, Gastropoda and Hemiptera at Utikuma Lake), and others (Amphipoda at both areas, Gastropoda and Hemiptera at Cardinal Lake) increased throughout the period of avian reproduction (Figs. 2, 3). Both overall abundance and seasonal patterns varied across years for all taxa at Cardinal Lake, but in neither case were patterns of variation consistent among taxonomic

123

Hydrobiologia

groups. At Utikuma Lake, there was typically less annual variability—parameter estimates for year were only precisely estimated for Gastropoda and Hemiptera, and seasonal patterns were annually consistent, except for Amphipoda, which showed a greater lateseason increase in 2002. Heterogeneity in stable isotope signatures and diet composition Invertebrate d13C and d15N values were highly variable: MANOVA showed differences among wetlands (Wilks’ k = 0.26, F54,590 = 10.34, P \ 0.001), between years (Wilks’ k = 0.91, F4,590 = 7.23, P \ 0.001), and within seasons (Wilks’ k = 0.94, F6,590 = 2.85, P \ 0.01). Isotopic values also varied among taxonomic groups (Wilks’ k = 0.07, F36,590 = 45.19, P \ 0.001), driven primarily by the enriched 15N values of predatory larval Chaoborids and the enriched 13C values of snails and amphipods. Isotope values of duckling leg muscle were not significantly related to year (Wilks’ k = 0.85, F2,31 = 2.66, P \ 0.09), gender (Wilks’ k = 0.92, F2,31 = 1.26, P \ 0.30), or hatch day (Wilks’ k = 0.94, F2,31 = 2.66, P \ 0.39). Wetland of origin, however, had a strong influence on both d13C and d15N values of leg muscle (Wilks’ k = 0.19, F14,62 = 5.62, P \ 0.001). Isotope values of duckling liver tissue were similarly unaffected by gender (Wilks’ k = 0.87, F2,24 = 1.83, P \ 0.18), but varied by year (Wilks’ k = 0.67, F2,24 = 5.93, P \ 0.01), wetland of origin

(Wilks’ k = 0.01, F14,48 = 28.61, P \ 0.001), and hatch day (Wilks’ k = 0.72, F2,24 = 4.65, P \ 0.02)—liver 13C becoming depleted with later hatch dates (bHatch = –0.02, 85% CI -0.04 to -0.01). Of the 49 pre-fledging scaup collected at Cardinal Lake in 2005 and 2006, only 13 contained measurable food items in the UGT. On a dry weight basis, branchiopods and amphipods represented a substantial portion of consumed biomass (PDMi = 40.0 and 26.0% respectively). Chaoboridae larvae (8.7%), Corixidae (9.2%), and planorbid snails (7.3%) together accounted for 25.2% of aggregate dry biomass; aquatic beetles, larval Chironomidae and physid snails comprised the remaining 8.8%. For the larger sample of birds (n = 43) for which we had isotopic data, posterior probabilities of alternative combinations of putative food sources in mixing models indicated that the optimal number of food sources (i.e. wetland invertebrates) was four. Further, estimated contributions of amphipods and gastropods to scaup diets were identical (l = 0.01, r = 0.001). For these reasons, group assignment for sources was based on functional feeding groups (Fig. 4): collectors (Corixidae), grazers and scrapers (Amphipoda, Planorbidae), filter feeders (Branchiopoda), and predators (Chaoboridae). Using these groupings, our top-supported mixing model indicated that the diet of scaup ducklings was related most strongly to wetland (Table 2, posterior median of estimated variability among wetlands = 1.61, 95% credible interval, CI 0.52–4.67).

Table 1 Selected covariates had varying effects on the relative proportions (biomass) of key invertebrate taxa, as shown for Cardinal Lake (first line) and Utikuma Lake (second line) Zone

Year

Timing of collection

Brood use

: In emergent vegetation

: In 2007

Effect variable by year

: In used ponds

: In emergent vegetation

: In 2002

: With date (non-linear)

: In used ponds

Branchiopoda

: In open water

No effect detected

Effect variable by year

: In used ponds

: In emergent vegetation

No effect detected

Seasonal peak

No effect detected

Diptera

: In open water

: In 2007

Effect variable by year

Effect variable by year

: In emergent vegetation

No effect detected

: With date (non-linear)

: In unused ponds

Gastropoda

: In emergent vegetation

: In 2007

Effect variable by year

No effect detected

: In emergent vegetation

: In 2001

Seasonal peak

: In used ponds

: In emergent vegetation

: In 2007

Effect variable by year

No effect detected

: In emergent vegetation

: In 2002

Seasonal peak

Effect variable by year

Amphipoda

Hemiptera

123

Hydrobiologia Fig. 2 Biomass (dry, mg per sweep) (y-axis) of selected invertebrate taxa in wetlands at Cardinal Lake varied seasonally and among years (note variable scales for each taxon). All plots show model-based estimates (95% confidence interval) for a given sampling period (x-axis) and year, as indicated in the legend. Differences in abundance between wetlands used by lesser scaup Aythya affinis broods (black symbols) and unused wetlands (white symbols) are also included; the number of wetlands samples in each case is indicated in parentheses

Results were not consistent with our prediction of seasonal variation in diet; there was little support for an effect of hatch class within wetland (weight of evidence, wi = 9%). Similarly, no models that included either a year effect or an effect of individuals were supported, suggesting that diets of scaup ducklings did not vary among years and that within the wetlands, diets were similar among individuals.

Accounting for spatial variation, dominant dietary items for pre-fledging scaup on boreal wetlands in this study included branchiopods (estimated proportion, 95% CI 0.15–0.70), amphipods and snails (95% CI 0.08–0.57), as well as Corixidae (95% CI 0.03–0.49). Our results suggest that few Chaoboridae (95% CI 0.00–0.23) were consumed by scaup ducklings at Cardinal Lake (Fig. 5).

123

Hydrobiologia Fig. 3 Biomass (dry, mg per sweep) (y-axis) of selected invertebrate taxa in wetlands at Utikuma Lake varied seasonally and among years (note variable scales for each taxon). All plots show model-based estimates (including a 95% confidence interval, dotted lines) for each year, as indicated in the legend. Differences in abundance between wetlands used by lesser scaup Aythya affinis broods (black lines) and unused wetlands (grey lines) are also included; the number of wetlands samples in each case is indicated in parentheses. Solid vertical lines divide the sampling period into nesting, broodrearing and fledging intervals

Habitat use and heterogeneity in biomass of wetland invertebrates consumed by scaup As we anticipated, wetlands used by brood-rearing female scaup at Cardinal Lake were characterized by higher abundances of key food items—Amphipoda

123

and Branchiopoda—but the opposite was true for Diptera. Brood use of ponds at Cardinal Lake was unrelated to abundances of either Gastropoda or Hemiptera (Fig. 2). Although we did not have dietary data for Utikuma Lake, relationships between invertebrate abundances and brood occurrence also suggest

Hydrobiologia Fig. 4 Lesser scaup Aythya affinis duckling leg muscle tissue stable isotope values (d13C, d15N) and isotope values of wetland macroinvertebrates based on four functional feeding groups. Invertebrate values include trophic enrichment factors (see text), and sample sizes are indicated in parentheses (invertebrate samples, number of wetlands)

Table 2 Diets of lesser scaup Aythya affinis ducklings in Canada’s western boreal forest are substantially different among wetlands and are less likely to vary over time Mixing model specification

DIC

DDIC

wi

Wetland

243.3

0.0

0.82

HatchClass (Wetland)

247.9

4.6

0.08

Year (Wetland)

249.1

5.8

0.05

Null (Intercept only)

249.7

6.4

0.03

HatchClass

252.2

8.9

0.01

Year

253.3

10.0

0.01

Bayesian stable isotope mixing models are ranked by differences in the Deviance Information Criterion (DDIC), wi = model weight Models with weights \0.01 are not shown

that Amphipoda are important for scaup broods at this study area. In addition, data indicated that brood use of Utikuma Lake wetlands was associated with a greater abundance of gastropods (Fig. 3). For branchiopods, dipterans (2005 only) and gastropods at Cardinal Lake, intra-seasonal patterns of abundance were different between wetlands used by broods and those that were not, but these differences were not consistent. For the remaining area/taxa combinations, changes in abundance throughout the sampling period were similar for used and unused ponds (Figs. 2, 3).

Discussion In the boreal wetlands that we studied—as in other wetlands where invertebrate communities have been

Fig. 5 Posterior estimates of the proportional contributions of macroinvertebrate functional feeding groups to the diet of prefledging lesser scaup Aythya affinis at Cardinal Lake generated by the isotopic mixing model with the lowest Deviance Information Criterion (DIC) value. Box plots are based on 50,000 vectors (see text): centre line = median, lower and upper hinges = first quartile (Q1) and third quartile (Q3), whiskers = Q1–1.5 (Q3–Q1) and Q3 ? 1.5 (Q3–Q1). Ducklings ate predominantly filter feeders (Branchiopoda), grazers/ scrapers (Amphipoda, Gastropoda), and collectors (Corixidae), with predatory dipteran larvae being a less probable food item

assessed—abundance (biomass and numbers) of selected invertebrate taxa varied substantially, across time and space, and patterns of variability were not consistent among taxonomic groups (Miller et al., 2008; Cobbaert et al., 2010). Temporal patterns of

123

Hydrobiologia

wetland invertebrate abundance did not impact either duckling diets or—in most cases—the habitats used by brood-rearing female scaup. Changes in scaup duckling diets, however, seem to suggest that use of food resources shifts to exploit differences in invertebrate communities among wetlands. Further, changes in the distribution of these scaup broods were linked to spatial variation in biomass of invertebrate food items; at both our study sites, broods were most commonly found on wetlands with higher amphipod abundance. At our more northern study area, Cardinal Lake, abundance of branchiopods was also an important predictor of brood use, consistent with dietary findings. By identifying connections between trophic levels, our results confirm that avian wetland fauna in boreal regions—as in locations further south—are dependent on invertebrate prey (Longcore et al., 2006; Horva´th et al., 2012). As environmental conditions and freshwater resources in the boreal continue to change, either due to changes in climate or extraction of natural resources (Schindler & Lee, 2010; Mitsch & Hernandez, 2013), these findings will provide critical baseline data describing ecological structure of important wildlife habitats. Temporal heterogeneity in abundance of key wetland invertebrates Our results provide some of the first empirical evidence that numbers and biomass of key invertebrate fauna in boreal wetlands are highly temporally variable, within and among years. In low latitude freshwater wetlands, especially those that tend to be eutrophic, resource supply or physical factors appear to influence invertebrate populations only during limited time frames and may have a reduced effect, relative to top-down control of invertebrates by aquatic predators, such as fish or salamanders (Hanson et al., 2009; Florencio et al., 2013; Knorp & Dorn, 2014). Although fish presence can also affect invertebrate communities in boreal wetlands, wetland biota at higher latitudes face harsher time constraints and environmental factors, such as nutrients or temperature, likely have a stronger influence in such ecosystems (Hornung & Foote, 2006; Rautio et al., 2011). For example, previous studies have demonstrated that water temperature plays a major role in regulating life histories of aquatic invertebrates and is particularly important in determining growth, fecundity or survival

123

for several freshwater species (Panov & McQueen, 1998; Gardner et al., 2011), such that increasing water temperatures for boreal wetlands could impact invertebrate populations. Indirect effects of climate change, such as increasing fire intensity or changes in wetland surface area, can also affect productivity of northern wetlands (Scrimgeour et al., 2001; Lewis et al., 2014). In our analyses, we did not specifically test for influences of non-avian aquatic predators or environmental variables on biomass of wetland invertebrates. Small fish—primarily stickleback (Gasterosteidae)— were observed in both study systems and may partially explain the patterns we saw. A post hoc evaluation using data from weather stations close to the study areas (www.climate.weatheroffice.gc.ca/climateData/ canada_e.html) showed that our annual estimates of abundance did not correlate in any consistent way with antecedent or current climate indices. However, given the substantial temporal variability in invertebrate abundance and variation in patterns among taxa that we observed, we recommend that investigating relationships between climate change and invertebrate productivity in northern areas is a critical area for future research in freshwater biology, particularly because of demonstrated links between climate and aquatic bird populations (Ross et al., 2015). Studies that evaluate the effects of specific environmental parameters on abundances of boreal wetland invertebrates during the avian breeding season will also be especially valuable. Scaup duckling diets and stable isotopes in freshwater food webs Taken together, our dietary analyses (gut contents and stable isotopes) suggest that pre-fledging scaup at Cardinal Lake consume a diverse diet of aquatic invertebrates, comprised predominantly of crustaceans (branchiopods and amphipods). Similar diets are noted for scaup in other locales—in southwestern Manitoba, juvenile scaup feed mostly on amphipods and snails, and in the southern Northwest Territories, Conchostraca (clam shrimp), amphipods, and Chaoboridae larvae are key diet items (Bartonek & Hickey, 1969; Bartonek & Murdy, 1970). Based on stable isotope analyses, Corixidae are also a relatively important food item for juvenile scaup at Cardinal Lake, whereas Chaoboridae are not. These findings are generally in agreement with findings from the gut

Hydrobiologia

contents, with the exception of Corixidae, which may have been under-represented in gut contents due to the younger age of birds with items in the UGT, relative to birds used in the mixing models (Bartonek & Murdy, 1970). On the basis of gut content analyses, scaup at Cardinal Lake also consume snails, counter to Sugden’s (1973) conclusion that young scaup have a low preference for such prey. Unfortunately, isotope analyses were unable to resolve this apparent contradiction—we could not determine the extent to which snails were assimilated into scaup tissues. The limited taxonomic resolution of our isotopic assessment, due to substantial variation in the isotopic signatures within invertebrate prey groups and overlap among groups (Fig. 4), is likely a consequence of the dietary flexibility of macroinvertebrates in freshwater systems (Kelly et al., 2013). The high level of spatial and temporal variability in the d13C and d15N values of wetland invertebrates highlighted by this study and others (Cremona et al., 2010) provides important insight regarding isotopic assessments of freshwater food webs. Although our capacity to isotopically differentiate scaup food items was limited by heterogeneity in invertebrate isotope values, our findings indicate that such heterogeneity should be carefully considered in future studies. Incorporating spatial and temporal aspects of isotope variation into sampling design will be critical for isotope studies attempting to make reliable inference about diet of freshwater invertivores at increasingly fine scales of taxonomic resolution (Phillips et al., 2014), particularly if these consumers are expected to have specialized diets. Including an additional isotope, such as sulphur, in dietary mixing models, or using alternative approaches for investigating diet, such as quantitative fatty acid signature analysis, or DNA barcoding, might also be helpful (Iverson, 2009; Valentini et al., 2009). To our knowledge, this is one of the first studies to assess spatial variation in diet of wetland birds at the local scale and to show that dietary variation among wetlands is substantial. Although some evidence suggests that ducklings feed selectively, this finding is more in accord with the concept that wetlanddependent birds are flexible consumers and well adapted to exploit fine-scale spatial heterogeneity in their invertebrate food resources (Beck et al., 2013; Nummi et al., 2013). As observed in captive scaup,

ducklings in this study likely consume the most abundant invertebrate prey and may also prefer relatively large and conspicuous items, consistent with our observed link between high amphipod biomass and wetland use by broods (Sugden, 1973; Fast et al., 2004). That scaup ducklings are rarely observed on turbid wetlands further supports the idea of a visual foraging strategy (Walsh et al., 2006). Recent work indicates that submersed aquatic vegetation helps maintain clear water states in boreal waters, suggesting that conserving boreal wetlands containing submerged macrophytes will help preserve the value of these habitats for breeding waterbirds (Lemelin et al., 2010; Bayley et al., 2012). Contrary to our expectation, results were not consistent with annual or intra-seasonal changes in scaup diet despite considerable temporal variation in the abundance of the invertebrate species they consume. We do not know whether ducklings incurred energetic costs associated with maintaining a consistent diet during periods of lower prey abundance. Ducklings may have been able to maintain adequate nutrition without increasing effort if food abundance exceeded minimum energetic thresholds, but low prey abundance or patchy distributions could force ducklings to spend more time searching for food and foraging (Lourenc¸o et al., 2010). If such a behavioural shift resulted in less energy available for growth or greater vulnerability to predation, broods would be maintaining diet composition at the expense of growth or survival—indeed, our previous work in this system showed that late-hatched ducklings were nutrient limited (Gurney et al., 2012). It is also important to note that the extent to which wetland birds can adjust foraging efforts may be constrained by other stressors, such as predation or inclement weather, and that survival could be affected if birds cannot meet their energy requirements in the face of these limitations (Varo et al., 2011). Although scaup ducklings appear to be fooditem generalists, our results identify certain invertebrate taxa (primarily Crustacea) that seem to affect their distribution. Additional studies are needed to fully understand the consequences of local variation in diet, and, as the boreal region continues to change, monitoring waterfowl productivity will provide important information about the stability of freshwater ecosystems in the region.

123

Hydrobiologia

Use of boreal wetlands by scaup ducklings in relation to food

of prey or prey diversity (Halupka et al., 2008)—a hypothesis that warrants further investigation.

Use of freshwater habitats by brood-rearing female scaup was linked to quantity of specific food resources at the wetland level. In almost all cases, amphipods were more abundant in the wetlands that broods used, although this was not observed at Utikuma Lake in 2002. Abundance of branchiopods and gastropods, at Cardinal and Utikuma, respectively, were also important predictors of brood use. Selecting wetlands with abundant food resources may be a strategy by which female ducks can increase their fitness, either by laying larger clutches or by enhancing offspring survival through effects on duckling size (Afton & Ankney, 1991; Dawson & Clark, 2000). An alternative explanation is that the habitats in which these invertebrates are abundant, i.e. deep, permanent wetlands with rooted vegetation in the littoral zone and open water zones (Menon, 1969; Murkin et al., 1991), are also habitats that are preferred by brood-rearing females for improved predator evasion. We did not evaluate differences in detection of scaup broods across different types of wetland habitats, and thus, we cannot state how our index of habitat use relates to actual presence or absence of this species on surveyed wetlands. However, the tendency of female scaup with broods to move to open areas of water and dive when disturbed or threatened does suggest that probability of detection should be relatively high, and recent occupancy modelling of boreal waterbird broods supports this assumption (Lewis et al., 2015). Unlike previous studies of habitat use in aquatic birds, we sampled invertebrates at regular intervals across the breeding season and were able to evaluate relationships between seasonal patterns of invertebrate abundance and brood use. We anticipated that females attending broods might use wetlands where key invertebrate prey were more seasonally stable, as to ensure consistent access to adequate food resources for their offspring. However, we detected no consistent differences in seasonal patterns between wetlands used by broods and those that were not. Similar conclusions have been made for other wetland systems where food resources are diverse and abundant— reproductive success for birds using these habitats may be less affected by shifts in invertebrate phenology and more affected by factors that affect overall abundance

Acknowledgments We thank numerous individuals (notably Karen Petkau, Olivier Mongeon, Melanie Wilson, Wally Price, and Steve Leach) for help in the field, as well as the Gwich’in Renewable Resource Board (funding and logistical support), Gwich’in Tribal Council (land access), and the Gwichya Gwich’in Renewable Resource Council (Tsiigehtchic Renewable Resource Council) for permission to work on their land. Studies were conducted under licences from the Canadian Wildlife Service (Permit NWT-SCI-04-04) and the University of Saskatchewan Animal Care and Use Committee (Permit 20050038). We are also grateful to Hollie Remenda for tireless work in the lab. Funding was provided by Ducks Unlimited Canada (DUC) and Environment Canada, personal support for K. Gurney was through a Natural Sciences and Engineering Research Council (NSERC) Discovery Grant (to RGC), NSERC Industrial Postgraduate Scholarship in conjunction with DUC, University of Saskatchewan Graduate Fellowship, MBNA Canada Bank Conservation Fellowship (DUC), Northern Scientific Training Program (Aboriginal Affairs and Northern Development Canada) and Dennis Raveling Scholarship for Waterfowl Research, California Waterfowl Association. Finally, we thank Mark Miller and an anonymous reviewer for constructive review comments.

123

References Afton, A. D. & C. D. Ankney, 1991. Nutrient-reserve dynamics of breeding lesser scaup - a test of competing hypotheses. Condor 93: 89–97. Arnold, K. E., S. L. Ramsay, L. Henderson & S. D. Larcombe, 2010. Seasonal variation in diet quality: antioxidants, invertebrates and blue tits Cyanistes caeruleus. Biological Journal of the Linnean Society 99: 708–717. Baldassarre, G., 2014. Ducks, Geese, and Swans of North America. JHU Press, Baltimore. Bartonek, J. C. & J. J. Hickey, 1969. Food habits of canvasbacks redheads and lesser scaup in Manitoba. Condor 71: 280–290. Bartonek, J. C. & H. W. Murdy, 1970. Summer foods of lesser scaup in subarctic taiga. Arctic 23: 35–44. Batzer, D. P., 2013. The seemingly intractable ecological responses of invertebrates in North American wetlands: a review. Wetlands 33: 1–15. Batzer, D. P. & A. Ruhı´, 2013. Is there a core set of organisms that structure macroinvertebrate assemblages in freshwater wetlands? Freshwater Biology 58: 1647–1659. Bayley, S. E., A. S. Wong & J. E. Thompson, 2012. Effects of agricultural encroachment and drought on wetlands and shallow lakes in the boreal transition zone of Canada. Wetlands 33: 17–28. Beck, M. L., W. A. Hopkins & B. P. Jackson, 2013. Spatial and temporal variation in the diet of tree swallows: implications for trace-element exposure after habitat remediation.

Hydrobiologia Archives of Environmental Contamination and Toxicology 65: 575–587. Bond, A. L. & K. A. Hobson, 2012. Reporting stable-isotope ratios in ecology: recommended terminology, guidelines and best practices. Waterbirds 35: 324–331. Cobbaert, D., S. E. Bayley & J.-L. Greter, 2010. Effects of a top invertebrate predator (Dytiscus alaskanus; Coleoptera: Dytiscidae) on fishless pond ecosystems. Hydrobiologia 644: 103–114. Corcoran, R. M., J. R. Lovvorn & P. J. Heglund, 2009. Longterm change in limnology and invertebrates in Alaskan boreal wetlands. Hydrobiologia 620: 77–89. Cremona, F., D. Planas & M. Lucotte, 2010. Influence of functional feeding groups and spatiotemporal variables on the d15N signature of littoral macroinvertebrates. Hydrobiologia 647: 51–61. Dawson, R. D. & R. G. Clark, 2000. Effects of hatching date and egg size on growth, recruitment, and adult size of lesser scaup. Condor 102: 930–935. Dessborn, L., J. Elmberg & G. Englund, 2011. Pike predation affects breeding success and habitat selection of ducks. Freshwater Biology 56: 579–589. Epners, C., S. Bayley, J. Thompson & W. Tonn, 2010. Influence of fish assemblage and shallow lake productivity on waterfowl communities in the Boreal Transition Zone of western Canada. Freshwater Biology 55: 2265–2280. Fast, P. L. F., R. G. Clark, R. W. Brook & J. E. Hines, 2004. Patterns of wetland use by brood-rearing lesser scaup in northern boreal forest of Canada. Waterbirds 27: 177–182. Feuchtmayr, H., D. McKee, I. F. Harvey, D. Atkinson & B. Moss, 2007. Response of macroinvertebrates to warming, nutrient addition and predation in large-scale mesocosm tanks. Hydrobiologia 584: 425–432. Florencio, M., C. Go´mez-Rodrı´guez, L. Serrano & C. Dı´azPaniagua, 2013. Competitive exclusion and habitat segregation in seasonal macroinvertebrate assemblages in temporary ponds. Freshwater Science 32: 650–662. Foxi, C. & G. Delrio, 2010. Larval habitats and seasonal abundance of Culicoides biting midges found in association with sheep in northern Sardinia, Italy. Medical and Veterinary Entomology 24: 199–209. Gardner, J. L., A. Peters, M. R. Kearney, L. Joseph & R. Heinsohn, 2011. Declining body size: a third universal response to warming? Trends in Ecology & Evolution 26: 285–291. Gollop, J. B. & W. H. Marshall, 1954. A guide for aging duck broods in the field. US Department of the Interior, Bureau of Sport Fisheries and Wildlife. Gurney, K. E. B., R. G. Clark & S. M. Slattery, 2012. Seasonal variation in pre-fledging survival of lesser scaup Aythya affinis: hatch date effects depend on maternal body mass. Journal of Avian Biology 43: 68–78. Halupka, L., A. Dyrcz & M. Borowiec, 2008. Climate change affects breeding of reed warblers Acrocephalus scirpaceus. Journal of Avian Biology 39: 95–100. Hanson, M. A., S. E. Bowe, F. G. Ossman, J. Fieberg, M. G. Butler & R. Koch, 2009. Influences of forest harvest and environmental gradients on aquatic invertebrate communities of seasonal ponds. Wetlands 29: 884–895. Hardin, J. W. & J. M. Hilbe, 2003. Generalized Estimating Equations, 1st ed. Chapman & Hall/CRC, Washington, DC.

Hayden, B., C. Harrod, E. Sonninen & K. K. Kahilainen, 2015. Seasonal depletion of resources intensifies trophic interactions in subarctic freshwater fish communities. Freshwater Biology 60: 1000–1015. Hobson, K. A., M. B. Wunder, S. L. Van Wilgenburg, R. G. Clark & L. I. Wassenaar, 2009. A method for investigating population declines of migratory birds using stable isotopes: origins of harvested lesser scaup in North America. PLoS One 4: e7915. Hornung, J. P. & A. L. Foote, 2006. Aquatic invertebrate responses to fish presence and vegetation complexity in western boreal wetlands, with implications for waterbird productivity. Wetlands 26: 1–12. Horva´th, Z., M. Ferenczi, A. Mo´ra, C. F. Vad, A. Ambrus, L. Forro´, G. Szo¨ve´nyi & S. Andrikovics, 2012. Invertebrate food sources for waterbirds provided by the reconstructed wetland of Nyirkai-Hany, northwestern Hungary. Hydrobiologia 697: 59–72. Iverson, S. J., 2009. Tracing aquatic food webs using fatty acids: from qualitative indicators to quantitative determination. In Arts, M. T., M. J. Kainz & M. T. Brett (eds), Lipids in Aquatic Ecosystems. Springer, New York: 281–308. Kelly, H. A. W., E. J. Rosi-Marshall, T. A. Kennedy, J. Robert, O. Hall, W. F. Cross & C. V. Baxter, 2013. Macroinvertebrate diets reflect tributary inputs and turbidity-driven changes in food availability in the Colorado River downstream of Glen Canyon Dam. Freshwater Science 32: 397–410. Knorp, N. E. & N. J. Dorn, 2014. Dissimilar numerical responses of macroinvertebrates to disturbance from drying and predatory sunfish. Freshwater Biology 59: 1378–1388. Koper, N. & M. Manseau, 2009. Generalized estimating equations and generalized linear mixed-effects models for modelling resource selection. Journal of Applied Ecology 46: 590–599. Lemelin, L.-V., M. Darveau, L. Imbeau & D. Bordage, 2010. Wetland use and selection by breeding waterbirds in the boreal forest of Quebec, Canada. Wetlands 30: 321–332. Lewis, T. L., D. Esler & W. S. Boyd, 2008. Foraging behavior of Surf Scoters (Melanitta perspicillata) and White-winged Scoters (M-fusca) in relation to clam density: inferring food availability and habitat quality. Auk 125: 149–157. Lewis, T. L., M. S. Lindberg, J. A. Schmutz, P. J. Heglund, J. Rover, J. C. Koch & M. R. Bertram, 2014. Pronounced chemical response of Subarctic lakes to climate-driven losses in surface area. Global Change Biology 21: 1140–1152. Lewis, T. L., M. S. Lindberg, J. A. Schmutz, M. R. Bertram & A. J. Dubour, 2015. Species richness and distributions of boreal waterbird broods in relation to nesting and broodrearing habitats. The Journal of Wildlife Management 79: 296–310. Longcore, J. R., D. G. McAuley, G. W. Pendelton, C. R. Bennatti, T. M. Mingo & K. L. Stromborg, 2006. Macroinvertebrate abundance, water chemistry, and wetland characteristics affect use of wetlands by avian species in Maine. Hydrobiologia 567: 143–167. Lourenc¸o, P. M., F. S. Mandema, J. C. E. W. Hooijmeijer, J. P. Granadeiro & T. Piersma, 2010. Site selection and resource depletion in black-tailed godwits Limosa l. limosa

123

Hydrobiologia eating rice during northward migration. Journal of Animal Ecology 79: 522–528. Menon, P. S., 1969. Population ecology of Gammarus-lacustris in Big Island Lake Canada. I. Habitat preference and relative abundance. Hydrobiologia 33: 14–32. Merritt, R. W., K. W. Cummins & M. B. Berg, 2008. An Introduction to the Aquatic Insects of North America, 4th ed. Kendall Hunt Publishing, Dubuque, IA. Miller, A. T., M. A. Hanson, J. O. Church, B. Palik, S. E. Bowe & M. G. Butler, 2008. Invertebrate community variation in seasonal forest wetlands: implications for sampling and analyses. Wetlands 28: 874–881. Mitsch, W. & M. Hernandez, 2013. Landscape and climate change threats to wetlands of North and Central America. Aquatic Sciences 75: 133–149. Murkin, H. R., J. A. Kadlec & E. J. Murkin, 1991. Effects of prolonged flooding on nektonic invertebrates in small diked marshes. Canadian Journal of Fisheries and Aquatic Sciences 48: 2355–2364. Murkin, H. R. & L. Ross, 1999. Macroinvertebrate responses to a simulated wet/dry cycle. In Batzer, D. P., R. B. Rader & S. A. Wissinger (eds), Invertebrates in Freshwater Wetlands of North America: Ecology and Management. John Wiley & Sons, New York, New York: 543–569. Nummi, P., A. Paasivaara, S. Suhonen & H. Po¨ysa¨, 2013. Wetland use by brood-rearing female ducks in a boreal forest landscape: the importance of food and habitat. Ibis 155: 68–79. Panov, V. E. & D. J. McQueen, 1998. Effects of temperature on individual growth rate and body size of a freshwater amphipod. Canadian Journal of Zoology 76: 1107–1116. Pehrsson, O. & K. G. K. Nystrom, 1988. Growth and movements of oldsquaw ducklings in relation to food. Journal of Wildlife Management 52: 185–191. Phillips, D. L. & P. M. Eldridge, 2005. Estimating the timing of diet shifts using stable isotopes. Oecologia 147: 195–203. Phillips, D. L., R. Inger, S. Bearhop, A. L. Jackson, J. W. Moore, A. C. Parnell, B. X. Semmens & E. J. Ward, 2014. Best practices for use of stable isotope mixing models in foodweb studies. Canadian Journal of Zoology 92: 823–835. Po¨yry, J., R. Leinonen, G. So¨derman, M. Nieminen, R. K. Heikkinen & T. R. Carter, 2011. Climate-induced increase of moth multivoltinism in boreal regions. Global Ecology and Biogeography 20: 289–298. Po¨ysa¨, H., J. Elmberg, K. Sjo¨berg & P. Nummi, 2000. Nesting mallards (Anas platyrhynchos) forecast brood-stage food limitation when selecting habitat: experimental evidence. Oecologia 122: 582–586. Rautio, M., I. Bayly, J. Gibson & M. Nyman, 2008. Zooplankton and zoobenthos in high-latitude water bodies. High Latitude Lake and River Ecosystems Oxford University Press, Oxford: 231–247. Rautio, M., F. Dufresne, I. Laurion, S. Bonilla, W. F. Vincent & K. S. Christoffersen, 2011. Shallow freshwater ecosystems of the circumpolar Arctic. Ecoscience 18: 204–222. Robbins, C. T., J. K. Fortin, K. D. Rode, S. D. Farley, L. A. Shipley & L. A. Felicetti, 2007. Optimizing protein intake as a foraging strategy to maximize mass gain in an omnivore. Oikos 116: 1675–1682. Ross, B., M. B. Hooten, J.-M. DeVink & D. N. Koons, 2015. Combined effects of climate, predation, and density

123

dependence on Greater and Lesser Scaup population dynamics. Ecological Applications 25: 1606–1617. Sass, G. Z., I. F. Creed, S. E. Bayley & K. J. Devito, 2007. Understanding variation in trophic status of lakes on the Boreal Plain: a 20 year retrospective using Landsat TM imagery. Remote Sensing of Environment 109: 127–141. Schekkerman, H., I. Tulp, T. Piersma & G. H. Visser, 2003. Mechanisms promoting higher growth rate in arctic than in temperate shorebirds. Oecologia 134: 332–342. Schindler, D. W. & P. G. Lee, 2010. Comprehensive conservation planning to protect biodiversity and ecosystem services in Canadian boreal regions under a warming climate and increasing exploitation. Biological Conservation 143: 1571–1586. Scrimgeour, G. J., W. M. Tonn, C. A. Paszkowski & C. Goater, 2001. Benthic macroinvertebrate biomass and wildfires: evidence for enrichment of boreal subarctic lakes. Freshwater Biology 46: 367–378. Slattery, S. M., J. L. Morissette, G. G. Mack & E. W. Butterworth, 2011. Waterfowl conservation planning: science needs and approaches. In J. V. Wells (ed), Boreal Birds of North America: A Hemispheric View of Their Conservation Links and Significance. Studies in Avian Biology 41. University of California Press, Berkeley, CA: 23–40. Smerdon, B. D., K. J. Devito & C. A. Mendoza, 2005. Interaction of groundwater and shallow lakes on outwash sediments in the sub-humid Boreal Plains of Canada. Journal of Hydrology 314: 246–262. Spiegelhalter, D. J., N. G. Best, B. P. Carlin & A. Linde, 2014. The deviance information criterion: 12 years on. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 76: 485–493. Stock, B. C. & B. X. Semmens, 2013. MixSIAR GUI User Manual, version 1.0. Strand, K. A., S. R. Chipps, S. N. Kahara, K. F. Higgins & S. Vaa, 2008. Patterns of prey use by lesser scaup Aythya affinis (Aves) and diet overlap with fishes during spring migration. Hydrobiologia 598: 389–398. Strode, P. K., 2009. Spring tree species use by migrating Yellow-rumped warblers in relation to phenology and food availability. The Wilson Journal of Ornithology 121: 457–468. Stroup, W. W., 2012. Generalized Linear Mixed Models: Modern Concepts, Methods and Applications. CRC Press, Boca Raton. Stroup, W. W., 2015. Rethinking the analysis of non-normal data in plant and soil science. Agronomy Journal 107: 811–827. Sugden, L. G., 1973. Feeding ecology of pintail, gadwall, American wigeon and lesser scaup ducklings in southern Alberta Canadian Wildlife Service Report Series. Winnipeg, MB: 1–45. Thorp, J. H. & A. P. Covich, 2001. Ecology and classification of North American freshwater invertebrates, 2nd ed. Academic Press, San Diego, California. Valentini, A., F. Pompanon & P. Taberlet, 2009. DNA barcoding for ecologists. Trends in Ecology & Evolution 24: 110–117. Varo, N., A. J. Green, M. I. Sa´nchez, C. Ramo, J. Go´mez & J. A. Amat, 2011. Behavioural and population responses to changing availability of Artemia prey by moulting black-

Hydrobiologia necked grebes, Podiceps nigricollis. Hydrobiologia 664: 163–171. Walsh, K. A., D. R. Halliwell, J. E. Hines, M. A. Fournier, A. Czarnecki & M. F. Dahl, 2006. Effects of water quality on habitat use by lesser scaup (Aythya affinis) broods in the boreal Northwest Territories, Canada. Hydrobiologia 567: 101–111. Ward, E. J., B. X. Semmens, D. L. Phillips, J. W. Moore & N. Bouwes, 2011. A quantitative approach to combine sources in stable isotope mixing models. Ecosphere 2: art19. Warton, D. I., 2005. Many zeros does not mean zero inflation: comparing the goodness-of-fit of parametric models to multivariate abundance data. Environmetrics 16: 275–289. Webster, K. L., F. D. Beall, I. Creed & D. Kreutzweiser, 2015. Impacts and prognosis of natural resource development on

water and wetlands in Canada’s boreal zone. Environmental Reviews 23: 53. Wells, J. V., D. Roberts, R. Cheng, P. Lee & D. M., 2011. A forest of blue - Canada’s boreal forest: the world’s waterkeeper. International Boreal Conservation Campaign and Canadian Boreal Initiative, Seattle, WA, and Ottawa, ON. Wiken, E. B., D. Gauthier, I. K. Marshall, K. Lawton & H. Hirvonen, 1996. A perspective on Canada’s ecosystems: an overview of the terrestrial and marine ecozones Occasional paper. Ottawa. Zoltai, S. C., 1988. Wetland environments and classification. In Rubec, C. D. A. (ed.), Wetlands of Canada. Polyscience, Montreal, Quebec: 1–26.

123