Whooping crane use of riverine stopover sites

Abstract

Migratory birds like endangered whooping cranes (Grus americana) require suitable nocturnal roost sites during twice annual migrations. Whooping cranes primarily roost in shallow surface water wetlands, ponds, and rivers. All these features have been greatly impacted by human activities, which present threats to the continued recovery of the species. A portion of one such river, the central Platte River, has been identified as critical habitat for the survival of the endangered whooping crane. Management intervention is now underway to rehabilitate habitat form and function on the central Platte River to increase use and thereby contribute to the survival of whooping cranes. The goal of our analyses was to develop habitat selection models that could be used to direct riverine habitat management activities (i.e., channel widening, tree removal, flow augmentation, etc.) along the central Platte River and throughout the species' range. As such, we focused our analyses on two robust sets of whooping crane observations and habitat metrics the Platte River Recovery Implementation Program (Program or PRRIP) and other such organizations could influence. This included channel characteristics such as total channel width, the width of channel unobstructed by dense vegetation, and distance of forest from the edge of the channel and flow-related metrics like wetted width and unit discharge (flow volume per linear meter of wetted channel width) that could be influenced by flow augmentation or reductions during migration. We used 17 years of systematic monitoring data in a discrete-choice framework to evaluate the influence these various metrics have on the relative probability of whooping crane use and found the width of channel unobstructed by dense vegetation and distance to the nearest forest were the best predictors of whooping crane use. Secondly, we used telemetry data obtained from a sample of 38 birds of all ages over the course of seven years, 2010-2016, to evaluate whooping crane use of riverine habitat within the North-central Great Plains, USA. For this second analysis, we focused on the two metrics found to be important predictors of whooping crane use along the central Platte River, unobstructed channel width and distance to nearest forest or wooded area. Our findings indicate resource managers, such as the Program, have the potential to influence whooping crane use of the central Platte River through removal of in-channel vegetation to increase the unobstructed width of narrow channels and through removal of trees along the bank line to increase unforested corridor widths. Results of both analyses also indicated that increases in relative probability of use by whooping cranes did not appreciably increase with unobstructed views ≥200 m wide and unforested corridor widths that were ≥330 m. Therefore, managing riverine sites for channels widths >200 m and removing trees beyond 165 m from the channel's edge would increase costs associated with implementing management actions such as channel and bank-line disking, removing trees, augmenting flow, etc. without necessarily realizing an additional appreciable increase in use by migrating whooping cranes.

Fig 1. Associated Habitat Reach of the central Platte River extending from Lexington downstream to Chapman, NE.

Fig 2. Great Plains study area including riverine use locations (points) included in our analysis.

Fig 3. Example of how unobstructed channel width (UOCW; yellow lines), nearest forest (NF; red lines) and unforested channel width (UFCW; blue lines) were measured at whooping crane use and available locations.

Fig 4. Predicted, relative selection ratio of unobstructed channel width (UOCW) based on all systematically collected whooping crane (n = 235).

Tick marks indicate actual data (use points are presented at y = 1 and available points are presented at y = 0). Data is displayed from the 10^th to the 90^th percentile of use locations with 90% confidence intervals.

Fig 5. Predicted, relative selection ratio of nearest forest (NF) based on all systematically collected whooping crane roost locations (n = 235).

Tick marks indicate actual data (use points are presented at y = 1 and available points are presented at y = 0). Data is displayed from the 10^th to the 90^th percentile of use locations with 90% confidence intervals.

Fig 6. Predicted relative selection ratio (solid line), with 90% confidence intervals (dashed lines), between the 10^th and 90^th percentiles of unobstructed channel widths (UOCW).

Tick marks display response data (use locations are plotted at y = 1; available locations are plotted at y = 0). One use and 56 available locations that ranged in width from 1,054 m to 2,189 m are not included in the plot.

Fig 7. Predicted relative selection ratios (solid line) with 90% confidence intervals (dashed lines) between the 10^th and 90^th percentiles of distance to nearest forest.

Tick marks indicate response data (use locations are at y = 1, available locations are at y = 0).