Ecosphere (Apr 2020)
Calibrating and adjusting counts of harbor seals in a tidewater glacier fjord to estimate abundance and trends 1992 to 2017
Abstract
Abstract Long‐term monitoring for understanding status and trend of species of conservation concern is undeniably valuable, yet monitoring methods often evolve over time due to the development of new technology, fluctuations in funding, logistical constraints, and innovations in sampling methods or analytical approaches. Consequently, valuable insights into annual or decadal‐scale trends can be lost unless calibration between historical and current methods is developed. Glacier Bay National Park, in southeastern Alaska, hosts an important regional population of harbor seals, with the majority of seals pupping and molting on icebergs calved from a tidewater glacier in Johns Hopkins Inlet. Monitoring efforts to assess abundance and trends of harbor seals used counts of seals by shore‐based observers from 1992 to 2002, but transitioned to aerial photographic surveys in 2007 through 2017. To produce a rigorous long‐term evaluation of abundance and trends of harbor seals, we (1) conducted concurrent shore‐based counts and aerial photographic surveys in 2007 and 2008; (2) developed an analytical calibration between the two monitoring methods; (3) developed a haul‐out model to estimate the number of harbor seals in the water at the time of counts; and (4) estimated abundance and trends of harbor seals from 1992 to 2017 from the adjusted counts. Our calibration analysis revealed that during the pupping season in June, counts of harbor seals by observers from shore were consistently lower than counts from aerial surveys. During the molting season, counts by shore‐based observers were only slightly less than aerial photographic surveys, and there was an interaction between survey method and season. After calibrating methods, we found important decadal‐scale changes in trend. Over the 26‐yr period (1992–2017), the estimated trend was negative; however, trends computed for rolling 10‐yr time intervals showed steep and significant declines ending around 2011, with leveling off and possibly some subsequent recovery. The most recent shorter‐term (2013–2017) trends are negative again, rivaling the steepest decreases over the 26‐yr period. Our calibration between two monitoring methods improved continuity for long‐term monitoring for a species of conservation concern by taking advantage of new sampling methods and innovations in analytical approaches.
Keywords
