Hydrology affects shorebirds, waterfowl, and other waterbirds at Bear River Bay, a Globally Important Bird Area
2021; Association of Field Ornithologists; Linguagem: Inglês
10.1111/jofo.12381
ISSN1557-9263
AutoresBrian G. Tavernia, Tim Meehan, John D. Neill, John Luft,
Tópico(s)Fire effects on ecosystems
ResumoThe value of saline lakes and associated wetlands as habitats in the xeric Great Basin is dependent on having water of sufficient quantity and quality to support wetland-dependent birds. To inform conservation and management of these habitats, models are needed to link birds and hydrological changes due to climate and human water use. We modeled seasonal relationships between counts for 35 migratory shorebird, waterfowl, and other waterbird species or taxonomic groups and hydrological metrics at Bear River Bay, a globally Important Bird Area at Utah's Great Salt Lake. We found that increased fall surface flows to the bay increased counts of 13 species, including American Avocets (Recurvirostra americana), American White Pelicans (Pelecanus erythrorhynchos), American Wigeons (Mareca americana), Northern Pintail (Anas acuta), Redheads (Aythya americana), and Ruddy Ducks (Oxyura jamaicensis). Increased spring surface flows increased counts of Forster's Terns (Sterna forsteri) and the sandpiper group, whereas intermediate spring flows produced peak counts for American White Pelicans. Thus, conservation or management actions that increase seasonal flows to Bear River Bay are expected to increase bay use by diverse members of the avian community. Counts for 11 species or taxonomic groups responded positively or negatively to the seasonal elevation of Great Salt Lake, and these responses are hypothesized to reflect the relative availability of habitats within the bay versus the lake as a whole. Our models provide tools that allow managers to understand how hydrological changes associated with climate change and human water use will affect birds in Bear River Bay. Addressing lake-wide and regional population implications of changing hydrological conditions at Bear River Bay, Great Salt Lake, and other locations across the Great Basin will require a regionally coordinated assessment of hydrology, habitat, and bird movements in response to changing habitat conditions. La hidrología afecta aves playeras, anátidas y otras aves acuáticas en Bear River Bay, una Área de Importancia Global para las Aves El valor como hábitat de los lagos salinos y humedales asociados en el Great Basin xérico depende de tener suficiente calidad y cantidad de agua para dar soporte a las aves dependientes de humedales. Para informar la conservación y el manejo de estos hábitats, se necesitan modelos que vinculen a las aves con los cambios hidrológicos relacionados al clima y el uso humano del agua. Modelamos la relación estacional entre los conteos de 35 aves playeras, anátidas y otras especies acuáticas migratorias o grupos taxonómicos, y métricas hidrológicas en Bear River Bay, un Área de Importancia para las Aves en el Great Salt Lake de Utah. En otoño, encontramos que el incremento de flujos de superficie a la bahía aumentó los conteos de 13 especies, incluyendo a Recurvirostra americana, Pelecanus erythrorhynchos, Mareca americana, Anas acuta, Aythya americana y Oxyura jamaicensis. En primavera, el incremento de los fujos de superficie aumentó los conteos de Sterna forsteri y varios playeros, mientras que los flujos intermedios produjeron conteos pico de Pelecanus erythrorhynchos. Por ello, se espera que las acciones de conservación o manejo que incrementen los flujos estacionales a Bear River Bay a su vez aumente el uso de la bahía por diversos miembros de la comunidad de aves. Los conteos de 11 especies o grupos taxonómicos respondieron positiva o negativamente a la elevación estacional del Great Salt Lake y tenemos la hipótesis de que estas respuestas reflejan la disponibilidad relativa de hábitats al interior de la bahía versus la del lago completo. Nuestros modelos proveen herramientas que permiten a los gestores entender cómo los cambios hidrológicos asociados con el cambio climático y el uso humano del agua afectarán a las aves en Bear River Bay. Entender las implicaciones del cambio de condiciones hidrológicas en todo el lago y en poblaciones regionales de Bear River Bay, Great Salt Lake, así como en otras localidades a lo ancho del Great Basin, requerirá de una caracterización regionalmente coordinada de la hidrología, hábitat y movimientos de las aves en respuesta a las cambiantes condiciones de sus hábitats. Wetland birds rely on saline lakes and associated wetlands as wintering, breeding, and migratory stopover areas in xeric landscapes of the western United States (Aldrich and Paul 2002, Petrie and Vest 2013, Wilsey et al. 2017). For example, Great Salt Lake, Utah, and associated ecosystems are used annually by millions of wetland birds with records of 42 shorebird species, 38 waterfowl species, and 18 species of colonial-nesting waterbirds (Aldrich and Paul 2002, Cavitt et al. 2014, Sorensen and Hoven 2020, J. Neill, pers. observ.). Wetland birds have morphological, physiological, and behavioral traits whose evolution has been driven by water resource conditions (e.g., water depth and salinity) or other habitat resources (e.g., food resources and vegetation composition) affected by water resource conditions (Weller 1999, Sorensen et al. 2020). For example, declining water levels may eliminate islands that provide refuge from terrestrial predators for colonial-nesting birds (Knopf and Evans 2020), or increasing water levels may flood nesting areas (Aldrich and Paul 2002). Water levels also indirectly affect, through changes in salinity, the community composition and abundance of invertebrates (Dana and Lenz 1986, Ma et al. 2010), an important food resource for wetland birds. Securing water supplies of sufficient quantities and qualities at the right times of year and locations is key to maintaining or restoring the value of saline lake and wetland habitat to birds (Aldrich and Paul 2002, Petrie et al. 2013, Thomas and Andres 2013, Wilsey et al. 2017). Located in terminal basins, saline lakes have water supply budgets that include gains from surface water flows, groundwater flows, and precipitation, and losses through evaporation and seepage to sediments (Williams 2002). Natural short- and long-term changes in precipitation, temperature, and net evaporation affect water budgets with consequences for lake elevation, volume, salinity, and other physical and biological attributes (Alder 2002, Williams 2002). For example, an unusual wet cycle in the 1980s increased Great Salt Lake's elevation from 1280.3 to 1283.7 m, an historic high, by 1987 (Alder 2002). This flooding reduced or eliminated nesting opportunities for colonial waterbirds that nest on the ground or in emergent vegetation (Aldrich and Paul 2002) and presumably also influenced distributions of non-breeding birds (Kadlec and Smith 1989). Anthropogenic water diversions, such as those for irrigated agriculture or municipal and industrial uses, have modified natural climate-driven hydrology and reduced water inflows to, and water cover of, saline lakes and their associated wetlands (Wurtsbaugh et al. 2017, Donnelly et al. 2020, King et al. 2021). Due to anthropogenic consumptive water use, Great Salt Lake's volume in 2015 was 48% less and its elevation 3.6 m lower than values expected without consumptive use (Wurtsbaugh et al. 2017). Anthropogenic climate change is also expected to affect saline lake water budgets (Ficklin and Stewart 2013, Jeppesen et al. 2015). Seasonal minimum air temperatures have increased and hydroperiod durations have decreased in the Great Basin, a region containing Great Salt Lake and several other saline lakes, in recent decades (Haig et al. 2019), and increased evaporative demand has contributed to surface water loss from saline lakes and their associated wetlands (Donnelly et al. 2020). Some lakes may experience more frequent droughts and the potential for volume reductions (Ficklin et al. 2013, Moore 2016). Conservation and management plans have identified the effects of anthropogenic water use and climate change on saline lake water budgets as major threats to wetland birds and their habitats (Thomas et al. 2013, Zimmerman and Ivey 2013, Wilsey et al. 2017). Concerns about anthropogenic water use and effects of climate change on wetland birds and their habitats have led researchers to model relationships between hydrological and bird habitat use metrics (Sorenson et al. 1998, Timmermans and Badzinski 2008, Wang et al. 2013, Zhang et al. 2016). Hydrological metrics reflecting the magnitude, duration, frequency, timing, and change rate of surface water inputs and lake water levels are ecologically relevant and potentially related to bird use (Poff et al. 1997, Timmermans et al. 2008). Previous studies have revealed connections between wetland bird species and communities and water levels (Bolduc and Afton 2004, DesGranges et al. 2006, Cavitt 2013, Aharon-Rotman et al. 2017), water-level fluctuations (Kingsford and Jenkins 2004, DesGranges et al. 2006, Bellio and Kingsford 2013), and surface water flows to lakes (Cavitt 2013, Wang et al. 2013, Zhang et al. 2016). These studies have shown that bird responses to hydrological metrics vary across guilds, species, and seasons, and these differences may be attributed to differences in foraging habitats and techniques. Modeled relationships between hydrological and bird metrics can help assess potentially negative effects of human water use and climate change and evaluate the effectiveness of conservation and management actions intended to offset or redress negative effects (Jones and Lacey 2009, Wang et al. 2013, Zhang et al. 2016). To maintain or restore bird communities, the needs of species with contrasting and potentially conflicting responses to hydrological metrics must be addressed and balanced. Efforts are underway to assess impacts of human water use and climate change on the hydrology of Great Salt Lake (e.g., Jacobs Engineering Group Inc. 2019). Our objective was to complement these efforts by producing quantitative models linking hydrological metrics to habitat use by birds in Great Salt Lake's Bear River Bay. Bear River Bay is recognized as a globally Important Bird Area (IBA) given its use by breeding and migrating shorebirds, waterfowl, and other waterbirds (Birdlife International 2020). We modeled relationships between hydrological metrics and counts of wetland birds during spring and fall migration over a 14-year period. Cavitt (2013) reported a mix of negative, positive, and no responses by 11 bird species to increasing surface water flows and lake elevation for Willard Spur, an ecosystem representing 17.4% of our Bear River Bay study area. Here, we build on Cavitt's (2013) study by (1) modeling 35 species and taxonomic groups that included migrants of regional conservation concern, (2) considering the potential for non-linear relationships between bird counts and hydrological metrics, and (3) including the entire unimpounded area of Bear River Bay, an area that, except for water rights related to the Bear River Migratory Bird Refuge, does not currently have a perfected water right for propagation of wildlife and vegetation. Consequently, the area is vulnerable to upstream water development and climate change. We sought to identify hydrological metric(s) most related to the abundance of individual bird species and determine if modeled relationships were positive, negative, or non-linear. Great Salt Lake is a saline lake located in the cold desert environment of northeastern Utah (Fig. 1; Aldrich and Paul 2002). Inflows from the Bear, Jordan, Ogden, and Weber rivers are the dominant freshwater supply for the lake, with smaller sources including precipitation, groundwater, treated sewage flows, and smaller tributaries. Salinities in the lake vary from freshwater up to 280 parts per thousand, with gradients present at multiple spatial scales, including across major bays (Aldrich and Paul 2002). Plants found in freshwater wetlands include cattail (Typha spp.), sago pondweed (Stuckenia pectinata), and hardstem bulrush (Schoenoplectus acutus), whereas plants in relatively saline wetlands include muskgrass (Chara spp.) and alkali bulrush (Bolboschoenus maritimus) (Aldrich and Paul 2002). Invasive common reed (Phragmites australis) is the focus of control efforts at Great Salt Lake (Rohal et al. 2018) and can negatively affect bird habitat (Benoit and Askins 1999). Bear River Bay (41.3872°N, 112.3169°W) is located in northeastern Great Salt Lake and is geographically delimited by federal, state, and private lands to the north and east, the Promontory Mountains to the west, and industrial areas and Union Pacific Railroad's causeway to the south (Fig. 1) (Aldrich and Paul 2002, Birdlife International 2020). It occupies a position between actively managed wetlands on federal, state, and private lands and the open, hypersaline waters of Great Salt Lake. Designated as a Global IBA, Bear River Bay supports over 1% of the biogeographic populations (IBA criterion A4i) of 14 wetland bird species (Birdlife International 2020). The area also hosts ≥ 20,000 waterbirds on a regular basis (IBA criterion A4iii) (Birdlife International 2020). The salinity of Bear River Bay is close to that of the Bear River, its major source of surface water and, as a result, water in the bay is fresh enough to support plants such as sago pondweed (Ruppia spp.) (Aldrich and Paul 2002) as well as fish species including common carp (Cyprinus carpio) and Utah chubs (Gila atraria) (Armstrong and Wurtsbaugh 2019). Bird data were collected as part of a long-term monitoring program by the Great Salt Lake Ecosystem Program, part of Utah's Division of Wildlife Resources. For two survey units (28C: 19,870.9 ha; 37A: 17,955.2 ha) in Bear River Bay (Fig. 1), observers counted individuals of species in the families Gaviidae, Podicipedidae, Pelecanidae, Phalacrocoracidae, Ardeidae, Threskiornithidae, Anatidae, Rallidae, Gruidae, Charadriidae, Recurvirostridae, Scolopacidae, and Laridae. Biologists counted birds from the air along 463-m wide transects spaced 1852 m apart within survey units. The width and spacing of transects were selected to cover a quarter of each survey unit. During transect flights, two observers counted birds out to 231.5 m on each side of the plane with distances from the plane visually estimated. Flights were at altitudes ranging from 24 to 61 m and speeds of 129–161 km/h in a fixed-wing Cessna 185 Skywagon. Given that altitude varied within and between flights, observers did not use window markings to determine distances. As needed, observers calibrated their visual distance estimates against known landmarks. In aggregate, transects covered a quarter of each survey unit, so we multiplied total counts across all transects for each species and taxonomic group by a factor of four in each survey unit. From 2004 to 2017, numbers and the timing of surveys differed across two monitoring phases. Until 2007, annual surveys occurred three times in spring (15 April to 14 May) and nine times in fall (8 July to 5 September). Beginning in 2007, there were two spring (10 April to 9 May) and three fall (18 July to 31 August) annual surveys. To reconcile differences in survey numbers and timing, we defined five overlapping survey windows for both monitoring phases, including 10 - 24 April (period 1), 25 April to 9 May (period 2), 18 July to 1 August (period 3), 2 -16 August (period 4), and 17 - 31 August (period 5). We labeled periods 1 and 2 as spring migration counts and periods 3, 4, and 5 as fall migration counts. If multiple counts occurred during the same survey period, we averaged and rounded counts for the period for each survey unit and then summed counts by period and year across survey units. Finally, we took the maximum count across periods within a season to arrive at counts for spring and fall migration seasons. Several species or taxonomic groups were too rare (i.e., observed on <20% of counts across all years and seasons) to model robustly, so we eliminated their counts from the bird survey dataset. We considered 20 species difficult for observers to distinguish in the field and combined them into seven taxonomic groups (Table 1). Our final bird survey dataset contained 14-year seasonal counts for 28 species (Table S1) and seven taxonomic groups. To quantify surface water inflows, we combined river, reservoir release, and treated sewage flow datasets. We quantified Bear River surface water inflows using mean daily discharge (cubic feet per second) records from the U.S. Geological Survey's (USGS) gage near Corinne, Utah (site number 10126000). We accessed the discharge records using the R package dataRetrieval (De Cicco et al. 2018) in R 4.0.2 (R Core Team 2020). Releases from Willard Bay Reservoir represent an intermittent source of surface water inflow to Bear River Bay. Weber Basin Water Conservancy District provided daily flow release records for the reservoir (Chris Hogge, pers. comm.), and we added these to the USGS daily flows. Publicly owned treatment works (POTW), or sewage treatment facilities, also provide surface flows into Bear River Bay. Utah's Division of Environmental Quality makes available the locations of permitted POTWs via an interactive web interface (https://enviro.deq.utah.gov/). Using these spatial data, we identified three facilities (permit numbers UT0025721, UT0020931, and UT0022365) that (1) discharge directly into Bear River Bay or (2) discharge below the previously identified USGS gaging station into rivers or canals flowing into the bay. We accessed flow records (2004–2017) for these facilities from Discharge Monitoring Reports (DMR) available from the U.S. Environmental Protection Agency's Enforcement and Compliance History Online system (echo.epa.gov). All three facilities reported flow as monthly or 30-day averages. One monitored flow of gross effluent, one monitored flow of raw sewage influent, and the monitoring location of the final facility was unreported. Two facilities operated from 2004 to 2017 and the third began operation in March 2011. For each POTW, we calculated median monthly flow values by year and assumed that daily discharge equaled the median monthly value. When a monthly flow value was unreported in a given year, we used a monthly value equal to the median of available values for that month from 2004 to 2017. In some instances, facilities indicated that "No Discharge" prevented reporting of monthly discharge values and, in these cases, we assumed zero flow for the month. We joined and summed river, reservoir, and POTW daily discharge data to produce a final dataset of daily surface water inflows. Using this dataset, we calculated total cumulative flow from the start of the water year (October of the preceding calendar year) through the end of each migration season (spring: 9 May; fall: 31 August). This produced annual estimates of spring and fall cumulative flows. Like Wang et al. (2013), we assumed cumulative flow reflected or was correlated with overall supplies of water, nutrients, and sediments. We downloaded Great Salt Lake surface elevation data (National Geodetic Vertical Datum of 1929) from the USGS's gage at Saltair Boat Harbor, Utah (10010000) using the R package dataRetrieval (De Cicco et al. 2018) in R 4.0.2 (R Core Team 2020). We calculated the median Great Salt Lake surface elevation during the spring (10 April to 9 May) and fall (18 July to 31 August) survey periods for each year. We analyzed the effects of hydrological metrics on species and taxonomic group counts separately and seasonally. We modeled the expected count, µt, in year t as a function of standardized surface water inflows, Great Salt Lake elevation, and year using negative binomial regression. A variance inflation factor analysis revealed scores ≥ three indicating collinearity among our fall predictor variables (Zuur and Ieno 2010) caused by a negative correlation between year and Great Salt Lake elevation (adjusted r2 = 0.20). To address this correlation, we regressed year against lake elevation and used the resulting residuals as a predictor in lieu of year for both spring and fall seasons. We used year residuals as a predictor for both seasons to maintain a consistent modeling approach facilitating inter-season comparisons. With year residuals as a predictor, variance inflation factors for spring and fall did not exceed three. We used a Bayesian approach to fit models via the INLA 20.04.18 package (Rue et al. 2017) for R statistical computing software (R Core Team 2020). For all model fixed effects, we assigned a normal prior distribution with a mean of zero and precision equal to 0.001. We assumed a negative binomial distribution for expected counts to account for potential overdispersion (median dispersion parameter k = 0.65). To evaluate model fit, we examined distributions of conditional predictive ordinates looking for uniformity (Czado and Gneiting 2009). We also calculated correlations between observed and predicted counts per site, year, and season and found that models had reasonable predictive ability (median rs = 0.54). Models did not converge for Killdeer (Charadrius vociferus), yellowlegs, or Phalaropus in spring. Across years, surface flows varied from 0.45 to 1.75 km3 (median = 0.70 km3) in the spring and from 0.54 to 2.30 km3 (median = 0.88 km3) in the fall. Great Salt Lake elevation ranged from 1278.3 to 1279.8 m (median = 1278.9 m) in spring and from 1278.1 to 1279.6 m (median = 1278.6 m) in fall. Across seasons, the number of competitive models (Bayes factor <10) ranged from one to three (median = two). In spring and fall, the best models for the greatest number of species or taxonomic groups included either a linear effect of surface water inflows or a linear effect of Great Salt Lake elevation (Table 2). Parameter estimates for surface water inflows and lake elevation were not always significant (Figs. 2 and 3). Significant, linear responses to surface water were positive in spring and in fall (Fig. 2). There were positive and negative significant, linear responses to lake elevation in spring and fall (Fig. 3). The only best model that was non-linear showed counts of American White Pelicans responding in a significantly concave fashion to spring surface water flows (maximum abundance at an intermediate flow of 1.07 km3). Credible intervals for parameters of all other best models overlapped zero. For Bear River Bay, we found that increasing fall surface water flows increased counts for 13 species including American Avocets, American White Pelicans, American Wigeons (Mareca americana), Northern Pintails (Anas acuta), Redheads (Aythya americana), and Ruddy Ducks (Oxyura jamaicensis), all species of regional conservation focus (Petrie et al. 2013, Thomas et al. 2013, Zimmerman et al. 2013, Wilsey et al. 2017). These species have varied foraging habitats, techniques, and food resources, so increasing flows during fall benefits a broad segment of the avian community. Increased surface flows apparently create diverse and abundant habitats during fall, a season when bay water cover is relatively low. When water cover is greater in spring, we observed fewer relationships between bird counts and surface water flows. Counts for Forster's Terns (Sterna forsteri) and the sandpiper group increased with increased spring flows whereas counts of American White Pelicans peaked at intermediate spring flows. Increasing spring flows may create foraging habitat for Forster's Terns in deeper, unimpounded areas with fish and lead to additional shallow water and mudflat habitat for the sandpiper group. As with Forster's Terns, American White Pelicans benefit from increasing spring flows that may create foraging habitat in unimpounded areas of Bear River Bay, but, as spring flows continue to increase, pelicans may move to exploit emerging foraging opportunities in managed areas, reducing their counts in the unimpounded areas. Using our models, policy makers and resource managers can consider explicitly the degree to which water management and use decisions that reduce seasonal flows will negatively affect bird species and taxonomic groups in Bear River Bay. We found that increased Great Salt Lake elevation during spring decreased bay counts for five species and increased counts for one species, whereas increased lake elevation during the fall increased counts for one species and two taxonomic groups and decreased counts for two species. Eared Grebes (Podiceps nigricollis), a species of regional conservation focus during its migration (Zimmerman et al. 2013), were among those species that declined in numbers in Bear River Bay with increased lake elevation in fall. The elevation of Great Salt Lake reflects and affects the types, distributions, and areas of wetlands across the lake (Aldrich and Paul 2002, Roberts and Conover 2016). Use of a wetland by migrating birds depends on local and landscape habitat conditions (Farmer and Parent 1997, Webb et al. 2010); thus, we hypothesize that bird use of Bear River Bay reflects the influence of lake elevation on bay and lake-wide wetland habitats. For example, with increasing lake elevation during fall and additional habitat available, migrating Eared Grebes may use other areas of the lake thereby reducing numbers in Bear River Bay. During fall migration, Eared Grebes tend to be found in more saline waters of the lake with abundant brine shrimp on which they forage while molting and preparing for further migratory movements (Caudell and Conover 2006). Utah's Division of Forestry, Fire and State Lands (UDNR FFSL 2013), in cooperation with partners, identified lake elevations ranging from 1279.6 to 1281.7 m as benefitting birds foraging in salt or fresh water, nesting on islands, and foraging or nesting in wetlands, in shallow water, or on shorelines/mudflats. Although useful for lake-wide assessments, this elevation range does not apply to specific areas, including Bear River Bay. This is evidenced by the fact that, despite lake elevations in our study generally falling below 1279.6 m in spring and fall, increasing lake elevation resulted in a decrease in Bear River Bay counts for species and taxonomic groups that use varied foraging habitats, techniques, and food resources. Understanding how the elevation of Great Salt Lake affects seasonal bird use of specific bays will require future studies that explicitly model how lake elevation affects the types, distributions, and areas of wetland habitats across the entire lake. Notwithstanding the need for such wetland modeling, relationships reported in our study can be used to evaluate and consider how changes in lake elevation affect Bear River Bay counts for species and taxonomic groups. Cavitt (2013) analyzed relationships between counts of 11 bird species or taxonomic groups and surface water inflows and lake elevation for Willard Spur, an ecosystem comprising 17.4% of our Bear River Bay study area. Using Willard Spur bird counts (1999–2012) from the Great Salt Lake Ecosystem Program, Cavitt (2013) reported negative relationships between Great Salt Lake elevation and fall counts for the Aechmophorus group, Black-necked Stilt (Himantopus mexicanus), California Gull (Larus californicus), Cinnamon Teal, and White-faced Ibis (Plegadis chihi) and a positive relationship for American Avocets. We do not attempt to compare these lake elevation results to those in our study primarily because Cavitt (2013) did not include survey year as a model predictor. We found that including survey year—specifically, its residual after controlling for correlations with lake elevation—as a predictor affected which species and taxonomic groups responded to lake elevation and changed the relative numerical balance of positive and negative relationships (B. Tavernia, unpubl. data). Given that our dataset spanned 14 years (2004–2017), we believed it appropriate to include survey year to account for temporal changes in counts. Using surface flows from the USGS gage near Corrine, Cavitt (2013) reported a positive effect of flows on fall counts of American Coots (Fulica americana), a negative effect on spring counts of White-faced Ibis, and no effect of fall or spring flows on counts of American Avocets, American White Pelicans, or Forster's Terns in Willard Spur. For our Bear River Bay study area, we found more responses to surface water flows, including positive responses for American Avocets, American Coots, American White Pelicans, Forster's Terns, and White-faced Ibis to increased fall surface water flows, a positive response of Forster's Tern to increased spring surface flows, and maximum counts of American White Pelicans at intermediate spring surface flows. In many years, Willard Spur retains some standing surface water when other areas of Bear River Bay have none (B. Tavernia, pers. observ.), and this may, in part, explain the few relationships between surface water flows and bird counts reported by Cavitt (2013). Nonetheless, surface flows and water connections among Willard Spur, Bear River Bay, and Great Salt Lake are important for the long-term ecological health of Willard Spur and its bird habitat (CH2M Hill Inc. 2016, Low 2018). Long-term bird monitoring programs do not always collect a suite of local habitat data concurrent with bird counts (e.g., O'Connor et al. 2000), and without habitat data, explaining bird distributions, abundances, and trends is impossible. In this paper, we developed hydrological measures—surface water inflows and Great Salt Lake elevation—for Bear River Bay hydrology by relying largely on long-term USGS hydrological monitoring data. Although we found relationships between bird counts and these hydrological metrics, our ability to model bird counts would be enhanced by temporally and spatially comprehensive data representing potentially important habitat characteristics including water depth (specific to Bear River Bay), vegetation structure and composition, salinity, and food resources (Ma et al. 2010). Such characteristics may be significant correlates of counts including for those species that did not respond to either surface flows or lake elevation in our study. Some habitat data gaps could be addressed by pairing a feasible habitat monitoring protocol with ongoing Great Salt Lake Ecosystem Program bird surveys, and guidance exists to develop habitat survey protocols for nonbreeding season surveys (e.g., Program for Regional and International Shorebird Monitoring 2018). For isolated projects, remote sensing datasets and image classification techniques have shown promise in addressing habitat-related ecosystem attributes, including the invasive Phragmites australis (Long et al. 2017) and chlorophyll concentrations (Bradt 2012). Continued scientific refinement of these techniques and institutional support for their ongoing application at Great Salt Lake would provide important habitat data. Despite a desire for additional habitat metrics, we deliberately considered our hydrological metrics because they offer potential linkages to scenario-based projections of surface flows (Serago et al. 2020) and lake elevation at Great Salt Lake (Jacobs Engineering Group Inc. 2019). Our models can be used to evaluate bird responses to projections that are within the range of surface flows and lake elevations used in model construction. Projection efforts and associated models are still in their relative infancies, but these preliminary projections suggest that lake elevations will decline outside the range observed in the data used to build our models (Jacobs Engineering Group Inc. 2019). If our models are extrapolated to assess effects of surface flows and lake elevations not included in modeled data, such assessments should acknowledge that the models have been extrapolated. To obviate the need for extrapolation, continued monitoring and modeling can update our models to incorporate bird responses to lake elevations and surface flows that have no analog in currently modeled data. In our analyses, we assumed that a constant proportion of individuals was counted regardless of lake elevation, surface flows, or year. This requires that detection probability, or the chance of confirming an individual animal is present within a survey unit over a defined time (Thompson and White 1998), remains unchanged. Factors affecting detection probability include observer identity and experience, survey method, time of day, weather conditions, habitat conditions, and other variables (Thompson et al. 1998). In our study, wildlife biologists conducted counts, and most surveys (76%) were conducted by the same two observers (J. Neill, unpubl. data). Surveys were consistently conducted by air following a protocol that detailed the acceptable timing and weather conditions for flights. Because our study area consisted primarily of open water or mudflats, we assumed, as did Vest (2013), that habitat did not affect bird visibility. Therefore, we do not believe that detection probability biased observed relationships between counts, lake elevations, surface flows, and year. Bear River Bay is an integral part of a habitat network at Great Salt Lake and other western hemisphere saline lakes and wetlands, and this network supports wetland-dependent birds as they complete their annual cycles. Declining habitat quality or availability at critical locations within a habitat network can have disproportionate effects on migratory bird populations (Iwamura et al. 2013), so it is important to secure sufficient water resources for Bear River Bay and other locally, regionally, and internationally recognized sites for bird conservation. Conserving and increasing seasonal surface water flows now and in the future will increase numbers of shorebirds, waterfowl, and other waterbirds using Bear River Bay. Ultimately, successful bird conservation will require coordinated management and monitoring of hydrology, habitats, and bird populations across the network of saline lakes and their associated wetlands. We thank all the biologists, field technicians, and volunteers for collecting the field data used in these analyses. A. Manning, D. Paul, and C. Perschon provided key leadership in the early years of the monitoring program. P. Birdsey, G. Evans, F. Howe, K. Johnson, D. Mann, R. Norvell, and K. Poulsen helped with survey and database design and management. C. Hogge, C. Miller, J. Vander Laan, J. DenBleyker, J. Ostermiller, K. Cowan, L. Myers, and M. Garn provided guidance on use and interpretation of hydrological data. We thank A. Neville, C. Wilsey, E. Sorensen, H. Hoven, K. Stockdale, M. Malmquist, M. Shoop, S. Senner, G. Ritchison, and two anonymous reviewers for comments that improved the manuscript. The authors have no conflicts of interest to declare. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Referência(s)