Whooping Crane Migrational Habitat Use Draft (2)

Home Browse Search

the study area. The sediment and river hydraulics group of the USGS in Denver evaluated three methods to incorporate the water surface differential in the flow dependent parameters. The HECRAS model was chosen as the best method based on their evaluation (Fotherby and Russell 2008). The HECRAS model used the mean daily flow values from the gages adjacent to the crane group location, divided the flow between the channels in the full cross section, and computed the water surface differential at each transect. The HECRAS model output included an estimated water surface differential, wetted top width, hydraulic depth, and flow for whooping crane transects on every day with a crane group observation. The HECRAS water surface differential was used to adjust the profile data to estimate sandbar elevation with simple linear interpolation For the models incorporating flow dependent characteristics, we defined the available set of points as the base set of 165 transects spaced every 0.50 mile in the HECRAS model (Fotherby and Russell 2008). This set included transects which were measured in 1989, 1998, or 2002, transects which were interpolated from the measured transects, and transects which were synthetically constructed from the 1998 colar infra-red photos. The spatial extent of this set of transects was from river mile 157 to river mile 239, limiting the model to inferences within this section of the river. The HECRAS model output included an estimated water surface differential, wetted top width, hydraulic depth, and flow for base transects on every day with a crane group observation. Sandbar elevation could not be obtained for the base set of transects. Means of Habitat Parameters Land cover variables, characteristics measured on the ground, and flow dependent habitat variables were summarized for use sites in the channel and upland strata separately and also combined. For the systematic sample of use locations, strata means and variances were estimated by the Horvitz-Thompson estimator for unequal probability sampling designs (Thompson 1992). Each observation had a predicted probability of inclusion in the sample based on the characteristics of the flight on which it was observed. Strata estimates were combined using standard estimators for stratified random sampling (Scheaffer et al. 1996). Strata weights were based on the number of possible transect days and the number of surveyed transect days. We also used large sample approximations of estimators for simple random samples to calculate means and variances for the systematic and opporiunistic combined sample of use locations. We also estimated these values for the available samples for each habitat model. Habitat Models We constructed seven habitat models to describe habitat selection with the three types of habitat parameters. • Four models were fit to evaluate the association of habitat use with land cover: o Habitat selection in the entire study area (in-channel and upland), o Habitat selection in-channel in the study area, o Habitat selection in the local area (in-channel and upland) and o Habitat selection in the local in-channel area.