Site occupancy of focal shorebird species at Whiskey Island and Caminada Headland, Louisiana 2012–2020
Coastal restoration through island construction and augmentation is an increasingly common management method in the northern Gulf of Mexico, but evaluating the impacts to shorebird species is difficult. Shorebirds are mostly migratory and many aspects of their life history, including reproduction in some species, occur in other places. In addition, counts or observations of shorebirds made at any given time represent only a portion of the population and that proportion may change with site conditions such as time of day and weather. Dynamic occupancy models can account for imperfect detection and produce estimates of the proportion of area occupied over time as a way to track trends in bird utilization over time. In this chapter we report on occupancy trends for five focal shorebird species from Caminada Headland and Whiskey Island: American Oystercatcher ( Haematopus palliatus ), Piping Plover ( Charadrius melodus ), Red Knot ( Calidris canutus ), Snowy Plover ( Charadrius nivosus ), and Wilson’s Plover ( Charadrius wilsonia ). We examined up to nine years of surveys at the two sites to model long-term trends in occupancy rates from a period spanning before, during, and after restoration. Our objective was to determine if there was change in occupancy over time (trend) during the restoration period. Field sampling was conducted as described in Chapter 1 for the five focal species in this chapter. To create spatial units for occupancy analysis we used a grid of 53 unique cells at Caminada Headland and 26 unique cells at Whiskey Island. All observations of the species were located into spatial units using GIS and presence of each species in each cell was determined for each visit within a sampling season. Sampling seasons were defined for species as the period of time where they were most likely to be using the study area, and corresponded with wintering (August–May), breeding (April–August), or staging (March–October). We did not include covariates for initial occupancy or colonization and site survival because these were small sites with homogenous habitat. We did account for time of year within a season by treating survey date as a covariate of detection probability and allowed it to vary throughout the season. We tested for a trend in occupancy over time by determining if the estimated slope through the series of annual occupancy estimates was significantly different from 0. We conducted 153 total surveys at Caminada Headland from 11 January 2013 to 5 June 2019, and 213 total surveys at Whiskey Island from 7 August 2012 to 19 August 2020. The number of surveys per year varied by species and site and range from 6–23 per species annually at Caminada Headland and 11–28 per species annually at Whiskey Island. Occupancy trends were able to be assessed for all species at both sites, with the exception of the American Oystercatcher, which were only observed in sufficient numbers to estimate occupancy at Whiskey Island. We found a significant positive trend in occupancy from 0.325 to 0.741 for American Oystercatcher at Whiskey Island. There was no significant trend in Piping Plover occupancy at Caminada Headland where occupancy was consistently high (0.910 to 0.947). At Whiskey Island, Piping Plover occupancy estimates increased from 0.574 to 0.908 during the study period which was a significant increase. At Caminada Headland and Whiskey Island Red Knot occupancy varied from 0.482 to 0.891 and 0.350 to 0.742, respectively, but showed no significant trend over the study period. Snowy Plover occupancy at Caminada Headland increased significantly from 0.295 to 0.785 over the study period. Snowy Plover occupancy also increased significantly at Whiskey Island from 0.442 to 0.906. Wilson’s Plover occupancy declined slightly during the study period from 0.946 to 0.935 at Caminada Headland, and from 0.858 to 0.736 at Whiskey Island, but the decline was not significant at either site. We found no evidence that occupancy declined significantly for any of the species during the period prior to, during, and after restoration. We did find a significant increasing trend in occupancy for Snowy Plover at Caminada Headland and a significant increasing trend for American Oystercatcher, Piping Plover, and Snowy Plover at Whiskey Island. Our modeling results indicate that sampling such as this is sufficient for occupancy modeling and can provide a robust metric for comparison over time or among sites. In future research we plan to investigate the sample size needed to detect a trend. We are currently conducting a power analysis to determine the minimum amount of sampling necessary to have power to detect a trend based on the detection probabilities from this study.