Summer Habitat Selection of the Lower Colorado River Valley Population of Greater Sandhill Cranes

Understanding how wildlife species select habitats is among the most important aspects of wildlife science. It allows managers to identify geographic ranges, conserve critical resources, understand the consequences of management actions, and map current and potential distributions of animals on the basis of resource selection model outputs and predictions (Fielding and Bell 1997; McClean et al. 1998; McDonald and McDonald 2002; Millspaugh et al. 2006). For most bird species, resource selection varies seasonally on the basis of shifts in habitat availability and quality, behavior, and nutritional needs to support energetically costly behaviors (e.g., reproduction, molting, or migration; Cody 1985; McLoughlin et al. 2010; Conring 2016). Resource use and selection studies of sandhill cranes (Antigone canadensis) have mainly been conducted in staging areas such as the Platte River in Nebraska, overwintering areas such as the Texas Southern High Plains, Texas Coast, Southern California, and Arizona, or on breeding ranges in Oregon and the Arctic (Krapu et al. 1984; Iverson et al. 1985; Norling et al. 1992; Baker et al. 1995; Littlefield 1995; Conring 2016). However, little research has been conducted on greater sandhill cranes (A. c. tabida; hereafter crane[s]), particularly on their summer areas in the Intermountain West (Colorado, Idaho, Montana, Nevada, Oregon, Utah, Washington, and Wyoming) of the United States.

Within the Intermountain West, several populations of cranes are recognized under Pacific and Central Flyway management plans: the Rocky Mountain Population, Lower Colorado River Valley Population (LCRVP), and Central Valley Population (Tacha et al. 1992; August 2011; Collins et al. 2015). Cranes summering in southwest and south-central Idaho, northeast Nevada, northwest Utah (Ivey and Herziger 2006; August 2011; Collins et al. 2015), and likely west-central Idaho (J. M. Knetter, Idaho Department of Fish & Game, personal communication) are associated with the LCRVP. The LCRVP is the least abundant (3-y average = 2,768; Dubovsky 2016), has the lowest reported recruitment (4.8%; Drewien et al. 1995), and is the least studied of any migratory crane population in North America (Dubovsky 2016). In many areas of the arid West where LCRVP cranes breed and summer, low-density rural home development is the fastest growing form of land use, and water availability is a major driver of this expanding human footprint (Gude et al. 2006). Agriculture and ranching traditionally accounted for >85% of western water use (National Research Council 1982; Brown et al. 2005); however, increased aesthetic and recreational value of this commodity has stimulated an exponential rise in rural development, and placed unprecedented pressure on scarce water resources (Hansen et al. 2002). As water demand shifts from agricultural to domestic and industrial uses, sustainability of flood-irrigated rangeland and biologically diverse wetland habitats are at risk, which could ultimately affect crane population persistence throughout the Intermountain West.

Disproportionally high private ownership of these resources (> 70%) in a largely public-land-dominated landscape inextricably links migratory bird conservation to private lands in the Intermountain West (Donnelly and Vest 2012). Despite encompassing only a small fraction of the landscape (< 2%), wetland habitats act as critical features that drive crane distribution and abundance. Rural development in significant portions of the LCRVP summer range has increased 350% in recent decades (Gude et al. 2006). Current population levels of the LCRVP are considered stable on the basis of wintering abundance estimates (Dubovsky 2016); however, crane longevity (up to 37 y; Drewien et al. 2010; Gerber et al. 2014) may mask temporal lag effects in future declines resulting from habitat loss and degradation that has already occurred. This trait is supportive of a K-selected life-history strategy adapted to exploit periodically favorable wetland conditions, and maintain long-term population viability (Bårdsen et al. 2011). During breeding periods, LCRVP cranes use palustrine and riparian wetlands in the mountain valleys of Idaho, Utah, Oregon, and Nevada that are characterized by dramatic climate-driven variation, a known factor influencing population recruitment (Ivey and Dugger 2008; McWethy and Austin 2009).

Despite these known conservation threats to regional wetland resources, no modern assessment of habitat selection for summering LCRVP has been attempted. Our objectives were to estimate home ranges using satellite transmitters for use in assessing habitat selection ratios of LCRVP cranes during the summering months to help identify habitats in need of conservation measures (i.e., enhancement, easements, acquisition, etc.). This analysis is one piece of a project that has looked at different variables throughout the entire annual cycle of LCRVP cranes: winter home range and resource use, migration strategy and timing, and discovered new summering areas of LCRVP cranes (Collins et. al 2015; Conring 2016).

We captured cranes on National Wildlife Refuges (NWR) along the Colorado River in Arizona (Cibola NWR), agricultural lands in the Imperial Valley, and Sonny Bono Salton Sea NWR in Southern California during the winters of 2014 and 2015. In the summer of 2014, we also captured cranes in the wet meadow wetland habitat in west-central Idaho. For complete capture area descriptions, please refer to Collins et al. (2015). Capture and transmitter attachment methods were approved by Texas Tech University Institutional Animal Care and Use Protocol #13108-12.

We tested habitat selection by cranes in Elko and White Pine counties of Nevada. Topography is characterized by north–south-oriented mountain ranges and associated watershed basins (Fiero 1986; August 2011). Elevation in the study area ranged from approximately 1,300 m at the edge of the Great Salt Lake Desert to nearly 4,000 m on Wheeler Peak. Lower-elevation areas in the study area are used primarily for cattle grazing and native hay production in pastures irrigated by geothermal springs and from intermittent mountain streams via diversion ditches. Much (86%) of the land area is in public ownership; however, > 85% of lowland meadow habitat is privately owned (McAdoo et al. 1986).

We tested habitat selection by cranes in Valley and Owyhee counties, Idaho. Valley County is in the west-central part of Idaho and cranes in this county are primarily located within Long Valley. The West Mountains and Lick Creek Range are Long Valley's western and eastern boundaries, respectively. The landscape on the western edge of the county is dominated by Cascade Reservoir, agricultural grasslands, wetlands, wet meadows, and lodgepole pine Pinus contorta stands (Van Daele and Van Daele 1982). Owyhee County is located in southwest Idaho. It is bounded on the north by the Snake River and Elmore County, on the east by Twin Falls County, on the south by Nevada, and on the west by Oregon. The landscape consists of undulating to rolling tablelands, structural benches, and foothills. The major watersheds in the area are the Bruneau and Owyhee rivers (Harkness 2003).

We captured cranes on wintering areas (Cibola NWR and Sonny Bono NWR) and in one summering area (Long Valley, Idaho). Trapping, banding, and ARGOS methods are described in Collins et al. (2015). Global positioning system (GPS) satellite platform terminal transmitters were attached to the tibiotarsus of captured individuals. Four GPS locations were recorded throughout the day (0000, 0700, 1000, and 1500 hours). We identified 0000 hours as nocturnal locations and only used 1000 hours for diurnal locations in the resource selection analysis to avoid pseudoreplication with the 1500 hours time-period locations. The locations at 0700 hours were not used in either time-period analysis because we could not be sure that the cranes had left their nocturnal location for diurnal activities or vice versa. Each crane was considered to be settled on its summering grounds and finished with its spring migration when it had stayed in an area for more than 1 wk. The fall migration, for each crane, started once the crane made a large movement southward out of the watershed in which it spent the summer. We established the summering period as April 1–August 15 and used that time period for all analyses (Figure 1; Data S1, Supplemental Material). Locations for summer areas were considered as general ranges because we were unable to determine the breeding status of cranes at time of capture and we did not conduct site visits on the summer range to determine if cranes were breeding (Collins et al. 2015).

Home ranges were developed using the Brownian bridge movement model (BBMM; Nielson et al. 2013) package within program R version 3.1.1 (R Core Team 2016). We used the BBMM method for estimating home ranges because it accounts for short time intervals between locations by modeling movement paths between sequential locations and eliminates problems that can arise with spatially and temporally correlated data with kernel density estimators (Horne et al. 2007; Walter et al. 2011; Kranstauber et al. 2012; Fischer et al. 2013). We used all locations (not just nocturnal and diurnal locations) between April 1 and August 15 with a location error size of 30 m and maximum time lag < 600 min between consecutive locations. The BBMM produced 50 and 99% isopleths for each crane-year, which were then imported into ArcMap 10.3 (hereafter ArcMap; ESRI 2015. ArcGIS Desktop: Release 1. Redlands, CA: Environmental Systems Research Institute) to perform subsequent analyses with habitat layers (Figure 2). We used the 50% isopleth (core area) for our habitat analysis, which we defined as the area within a home range (99% isopleth) where use exceeded the expected uniform distribution (Samuel et al. 1985; Nesbitt and Williams 1990; Fischer et al. 2013). This core area should be more spatially precise and informative for informing management for summering cranes. Cranes with multiple years of summering locations were considered unique and independent because of home ranges not overlapping completely. This allowed us to identify each year for individual cranes as unique and the sampling unit was identified as crane-year.

Summer habitat selection was evaluated using Gap Analysis Program Land Cover raster data for Idaho and Nevada (U.S. Geological Survey, Gap Analysis Program, May 2011, National Land Cover, Version 2, September 2016). Habitat classification data were filtered to the Formation level, which are combinations of dominant and diagnostic growth forms, of the National Vegetation Classification system (U.S. National Vegetation Classification 2016) to identify the various habitat categories. There were very few locations in many of the Formation categories, so we combined them into broader habitat categories: Agriculture, Cliff and Desert, Developed Areas, Grasslands, Forests, and Wetlands (see Table 1 for Formation level breakdown). To evaluate the distribution of marked cranes across a large landscape, we used an equivalent landscape-level boundary data set to distinguish those landscapes. We used the Watershed Boundary data set layer at the Watershed (HUC 10) level (Watershed Boundary Data Set for Idaho and Nevada, http://datagateway.nrcs.usda.gov, August 2016). The Watershed layer was ideal because summering cranes are known to nest and congregate in high montane valleys that contain palustrine and riparian wetlands near agricultural lands, which can be found in unique watersheds. We only used watersheds used by cranes in the summers of 2014–2016 (Figure 3).

We conducted a simple habitat analysis at multiple spatial scales (watershed boundaries used by cranes and 50% isopleth as our core area; i.e., second order and third order; Johnson 1980) using the different design levels dependent on the spatial scale (designs I, II, and III; Manly et al. 2002) using the wides module in the AdehabitatHS R package (Calenge 2015). We acknowledge that not every land cover type was used, even if it was considered available on the basis of spatial scales and design within the watershed. Within the guidelines of traditional habitat terminology, land cover that is not used, regardless of availability, is not considered crane habitat (Hall et al. 1997). Therefore we defined habitat as land cover types that were selected for in proportion to their availability for each spatial scale and design level. We hypothesized that cranes did not select habitat types disproportionately to what was available at different spatial scales. We used the spatial join tool in ArcMap to obtain the habitat category for each GPS location to use in the habitat used portion of the module. The Extract by Mask tool in ArcMap was used to determine hectares of each available habitat type at the watershed and 50% core area scales. The design I selection ratio determines if selection of habitat for diurnal and nocturnal locations at the population level is equivalent to the available habitat across all watersheds. We calculated percent habitat type used for all cranes, grouped by daily time period, and divided it by the percentage of habitat type available for all watersheds used by the population. The design II selection ratio was used to determine if individual cranes selected habitat in different daily time periods in the same proportion as was available across all the watersheds. We calculated the percent habitat type used by each crane-year by time period, then divided by the percent habitat type available over the entire watershed layer to determine the selection ratio. Finally, we investigated individual crane habitat selection within the core area (50% isopleth) for design III selection ratio. Habitat used was obtained for each location per crane-year (by daily time period) within a core area and habitat available was calculated within the core area. The available habitat proportions were obtained from the core area for each crane-year. All analyses for the three design levels were conducted separately for diurnal and nocturnal locations.

Following Manly et al. (2002), we calculated the habitat selection proportions (w_ij; used_j/available_i), selection ratio estimate (ŵ_i), SE, and 95% confidence intervals by habitat type. If the confidence interval did not include 1, then the habitat type was selected or avoided. A selection ratio > 1 indicated a selection for that habitat type; < 1 indicated an avoidance; and there was no selection if it equaled 0. A standardized selection ratio (B_i) was calculated by dividing the specific habitat selection ratio by the summed selection ratios for all habitat types to measure the strength of the selection. We also tested for random resource use (Pearson statistic) in the design I analysis. For design II, we tested for equal use of habitat by all animals, overall habitat selection differences, and if, on average, the animals are using resources in proportion to availability using a χ² test. In design III, we conducted a χ² test to examine differences in overall habitat selection (Manly et al. 2002).

We captured and fitted 21 sandhill cranes with platform terminal transmitters. Only those cranes that were active through a full summering season (April 1–August 15) and stayed within a discrete area (i.e., did not wander over multiple watersheds and states) were used for the habitat selection analyses. We also post hoc removed two radiomarked cranes that looked to be paired to another radiomarked crane on the basis of their overlapping locations to avoid pseudoreplication. We used data from 13 radiomarked cranes (10,211 locations) in the BBMM analysis. Each summer was treated as unique, which allowed for creation of 26 distinct summer ranges (crane-year) over 3 y (six cranes with three summers, one crane with two summers, and six with one summer). Average BBMM home range size at the 50% core area level and the 99% isopleth was 525.4 ha (SE = 155.6) and 6,476.5 ha (SE = 1,637.5), respectively (Table 2). For the habitat analysis, we limited the locations to those at 0000 and 1000 hours and subsequently had 5,212 and 4,394 unique locations for all watersheds and those within the 50% core area.

At the population level (design I), cranes did not select habitats in proportion to availability in either diurnal (χ² = 2,966.05, df = 5, P < 0.01) or nocturnal locations (χ² = 3,159.51, df = 5, P < 0.01). Crane diurnal selection was for Wetlands (ŵ_i = 5.65, 95% CI = 5.40–5.91) and Agriculture (ŵ_i = 1.88, 95% CI = 1.63–2.14), whereas Grasslands (ŵ_i = 0.78, 95% CI = 0.76–0.80), Forest (ŵ_i = 0.13, 95% CI = 0.11–0.15), Developed Areas (ŵ_i = 0.14, 95% CI = 0.00-0.34), and Cliff and Desert types (ŵ_i = 0.00, 95% CI = 0.00–0.00) were selected against (Table 3). The standardized selection ratio indicates that Wetlands (B_i = 0.66) were selected almost seven times more than any other category during the diurnal time period. Cranes selected Wetlands as nocturnal sites (ŵ_i = 6.93, 95% CI = 6.64–7.22) nearly eight times (B_i = 0.76) more frequently than any other habitat type. Cranes, at their nocturnal sites, also selected for Agriculture (ŵ_i = 1.28, 95% CI = 1.05–1.52), whereas Grasslands (ŵ_i = 0.65, 95% CI = 0.61–0.69), Forest (ŵ_i = 0.22, 95% CI = 0.18–0.26) and Developed Areas (ŵ_i = 0.08, 95% CI = 0.00–0.24) were selected against; Cliff and Desert types were not selected at all because of no locations within this habitat type (Table 3). Overall, habitat availability ranged from 67% (Grasslands) to ∼1% (Cliff and Desert). Wetlands account for 6.81% of the available habitat; however, 38% and 47% of all diurnal and nocturnal locations were within Wetlands. Specifically, within the National Vegetation Classification Formation level the Temperate Flooded and Swamp Forest accounted for 26.35% of all locations during diurnal periods and 34.18% for nocturnal locations (Table 1). Conversely, Grasslands (67% of available habitat) accounted for 52% and 43% of all locations during diurnal and nocturnal locations, respectively. Agriculture accounted for 3.78% of the available habitat and 6.10% of the used locations (7.15% for diurnal; 4.85% nocturnal), which were in the Herbaceous Agricultural Vegetation Formation of the National Vegetation Classification (Table 1).

Results from the design II analysis, by time period, suggest that individual cranes did not select habitat types equally (diurnal:χ² = 1,324.44, df = 125, P < 0.01; nocturnal:χ² = 1,601.92, df = 125, P < 0.01). There was strong evidence for selection of habitat disproportionately (diurnal:χ² = 4,290.52, df = 130, P < 0.01; nocturnal:χ² = 4,761.45, df = 130, P < 0.01) and cranes are not, on average, using habitat types in proportion to availability (diurnal:χ² = 2,966.08, df = 5, P < 0.01; nocturnal: χ² = 3,159.54, df = 5, P < 0.01). Results for crane diurnal locations indicated that Wetlands were the only habitat type selected for (ŵ_i = 5.65, 95% CI = 3.50–7.79), whereas Forest (ŵ_i = 0.13, 95% CI = 0.00–0.29) and Developed Areas (ŵ_i = 0.14, 95% CI = 0.00–0.45) were selected against. There was no selection for or against Agriculture or Grasslands (on the basis of confidence interval containing 1.0); Cliff and Desert types were not used at all (Table 4). Cranes also selected Wetlands for nocturnal locations (ŵ_i = 6.93, 95% CI = 4.37–9.48), whereas cranes selected against Forest (ŵ_i = 0.22, 95% CI = 0.00–0.57), Grasslands (ŵ_i = 0.65, 95% CI = 0.42–0.88), and Developed Areas (ŵ_i = 0.08, 95% CI = 0.00–0.35). Cranes did not select for or against Agriculture, and Cliff and Desert types were not selected because there were no locations in this type (Table 4).

At the finer spatial scale (design III; 50% core area), there was a clear selection by cranes for both diurnal (χ² = 329.96, df = 47, P < 0.01) and nocturnal (χ² = 729.49, df = 38, P < 0.01) locations. At this scale, results were similar to designs I and II, but the selection ratio was not as high as for the other designs, indicating that there is more individual variation in habitat selection at the 50% core area. Cranes selected Wetlands (ŵ_i = 1.49, 95% CI = 1.17–1.81) during diurnal time periods, whereas Grasslands (ŵ_i = 0.85, 95% CI = 0.72–0.98), Developed Areas (ŵ_i = 0.16, 95% CI = 0.00–0.47), and Forest (ŵ_i = 0.41, 95% CI = 0.17–0.65) were selected against; Agriculture was not selected for or against (Table 5). For nocturnal locations, cranes selected Wetlands (ŵ_i = 1.87, 95% CI = 1.32–2.42), Grasslands (ŵ_i = 0.72, 95% CI = 0.54–0.90) were selected against, and Agriculture and Forest were not selected for or against; no locations occurred in Developed Areas (Table 5).

Wetland-dependent birds, such as cranes, select summering habitat on the basis of different needs and behaviors (e.g., nesting, loafing, foraging, communal, etc.), and at different scales (Burger 1985; Baker et al. 1995). Hayes (2015) found that two unpaired (presumed subadults) Eastern Population (A.c. tabida subspecies that primarily migrate in the Mississippi and Atlantic flyways) cranes during summer presumably did not have established breeding territories because they had very large home ranges at the 95% isopleths of 25,810 ha and 14,410 ha respectively. These are much greater home range estimates than reported in our study for the majority of our cranes, which indicates that either most of our cranes were paired and had established breeding territories or cranes in the LCRVP are restricted to the Intermountain West river valleys where wetland habitat availability is limited and potentially significant to LCRVP cranes. Additionally, cranes in Wyoming used wet meadows and grain fields between 69 and 100% of the time in summer (Rowland et al. 1992). Breeding Eastern Population cranes were also found to use wetlands, which is an important component for territorial cranes (Safina 1993; Lacy et al. 2015; Miller and Barzen 2016). It has also been hypothesized that paired cranes remain in territories and visit fewer habitat types (Johnsgard 1991; McIvor and Conover 1994). Cranes in this study showed a strong selection for wetland habitat, which lends itself to support Johnsgard's (1991) hypothesis, and potentially suggests that they were paired and breeding because they were using fewer habitat types, or space was a limiting factor.

Baker et al. (1995) found that emergent wetlands within a crane territory benefit cranes by increasing potential foraging areas. Cranes also typically nest in emergent wetlands, where standing water provides security from predators and persistent vegetation provides material to build nests (Urbanek and Bookhout 1992). Greater use of one habitat may reflect better habitat conditions to meet daily resource needs such as loafing, drinking, foraging, or conducting pair formation activities (Krapu et al. 1984; Iverson et al. 1987; Tacha 1988; Davis 2001). Cranes in this study exhibited a strong selection for wetland habitats over other habitats available during the diurnal hours. We speculate that the most plausible explanation for habitat selection by LCRVP cranes in the summer is that wetlands within their core area supply sufficient resources throughout the course of their summer season. Another factor that may be influencing diurnal habitat use selection is current land-use practices. Flood-irrigated hay meadows in the Intermountain West mimic once naturally occurring wet meadow complexes that traditionally existed in these watersheds. Cranes selected these wetland and upland complexes (in our analysis these two types were grouped into “wetlands”) with close association to privately owned, working rangelands throughout the Intermountain West. Wet meadows and flood-irrigated pastures are especially important for providing nesting, foraging, and colt-rearing habitat (Donnelly and Vest 2012). It has been reported that overwintering cranes select habitats such as hayed native grasslands because of unobstructed views of the surrounding area and freedom of movement (Lovvorn and Kirkpatrick 1982; VerCauteren 1998; Davis 2001).

Davis (2001) reported that crane abundance on roosts was due to availability of quality roosting habitat and proximity to lowland grassland habitat, which is similar to diurnal habitat selection previously reported by Anteau et al. (2011). As expected, LCRVP cranes selected wetlands as nocturnal sites and we suspect that nocturnal sites were based on individual needs such as nest site locations or a communal location for subadults and nonbreeders. Roost or nocturnal site quality and proximity to foraging areas is also important so cranes can obtain their energetic needs in one place and not expend energy acquiring resources elsewhere (Anteau et al. 2011).

It is clear that conservation of wetland habitats in the Intermountain West is essential for LCRVP cranes and the demand for water is only expected to increase because of expanding rural development. Private lands, where Temperate Flooded and Swamp Forests and flood-irrigated agriculture are found, are going to be an important factor in conserving these important habitats for LCRVP cranes, and there are many federal and nongovernmental organization programs that are available to protect or enhance important habitats. In our study area, many of the cranes used private lands throughout the summering season. Conservation and management of this summering population would benefit from an assessment of wetland habitats, especially Temperate Flooded and Swamp Forests and flood-irrigated agriculture, in the Intermountain West to determine where wetland habitats currently exist or are maintained artificially as a land management practice. These wetlands can then be evaluated for population expansion as well as give wetland managers in the Intermountain West strategic areas to concentrate on the ground conservation efforts.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Microsoft Excel file containing GPS data used for BBMM and resource analysis of Lower Colorado River sandhill cranes (Antigone canadensis tabida) summering in Nevada and Idaho during 2014–2016. Data are contained in one tab and the column abbreviations are explained in the second tab.

Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S1 (580 KB XLSX).

Reference S1. August CW. 2011. Demography of greater sandhill cranes in northeast Nevada. Master's thesis. Reno: University of Nevada Reno.

Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S2 (949 KB PDF).

Reference S2. Donnelly JP, Vest JL. 2012. Identifying Science Priorities 2013–2018: Wetland Focal Strategies. Missoula, Montana: Intermountain West Joint Venture Technical Series 2012. Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S3 (1.429 MB PDF); also available at https://iwjv.org/sites/default/files/iwjv_3_science_wetlands_2013-2018.pdf.

Reference S3. Dubovsky, JA. 2016. Status and harvests of sandhill cranes: Mid-Continent, Rocky Mountain, Lower Colorado River Valley and Eastern Populations. Denver, Colorado: U.S. Fish and Wildlife Service, Administrative Report.

Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S4 (2.053 MB PDF); also available at https://www.fws.gov/migratorybirds/pdf/surveys-and-data/Population-status/SandhillCrane/StatusandHarvestofSandhillCranes16.pdf.

Reference S4. Harkness AL. 2003. Soil survey of Owyhee County, Idaho. U.S. Department of Agriculture. Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S5 (2.216 MB PDF).

Reference S5. Ivey GL, Herziger CP. 2006. Intermountain West Waterbird Conservation Plan, Version 1.2. A plan associated with the Waterbird Conservation for the Americas Initiative. Portland, Oregon: U.S. Fish and Wildlife Service.

Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S6 (3.374 MB PDF); also available at https://www.fws.gov/pacific/migratorybirds/PDF/IWWCP.pdf .

Reference S6. VerCauteren TL. 1998. Local scale analysis of sandhill crane use of lowland grasslands along the Platte River, Nebraska. Master's thesis. Lincoln: University of Nebraska. Found at DOI: http://dx.doi.org/10.3996042017-JFWM-037.S7 (1.241 MB PDF).

Financial and logistical support for this research was provided by the U.S. Fish and Wildlife Service Webless Migratory Game Bird program and Texas Tech University. We thank G. Ivey with the International Crane Foundation, U.S. Fish and Wildlife Service Refuge Staff (Bosque del Apache, Cibola, and Sonny Bono Salton Sea NWRs): T. Anderson, S. Goehring, J. Grzyb, S. Rimer, R. Woody, J. Vradenburg, and B. Zaun for field and logistical support. We thank the Idaho Department of Fish and Game staff: R. Berkley, B. Bosworth, D. Evans-Mack, and C. White for field and logistical assistance. We appreciate the thorough reviews by D. Haukos, D. Brandt, and D. Fronczak that helped to greatly improve an earlier version of this manuscript. The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.