Modelling demographic rates for White-capped Albatrosses

Jim Roberts and colleagues (National Institute of Water & Atmospheric Research Ltd, Christchurch, New Zealand) have tabled a draft background report to last month’s meeting of the Conservation Services Programme (CSP) of New Zealand’s Department of Conservation that considers estimating demographic rates for White-capped Albatrosses Thalassarche steadi.

The report’s Executive Summary follows:

“Disappointment Island, within the Auckland Islands group, supports over 70,000 breeding pairs of white-capped albatrosses Thalassarche cauta steadi annually, the largest colony of New Zealand’s most abundant albatross species.  This species interacts with commercial fisheries and ranks highly within the Level 2 Seabird Risk Assessment process, but with a relatively high level of uncertainty around the estimate of adult survival.  A study was undertaken to assess the effect of alternative mark-recapture sampling approaches to a potential mark-resighting study of white-capped albatross on the estimation of demographic rates.

A data simulator was used to create dummy mark-resighting observations for a single banding year with alternative scenarios of: banded sample size (150, 300 or 600 breeding individuals); number of subsequent consecutive resighting years (2, 3, 4, 5 or 10 years); and resighting probability of breeders (0.6 or 0.4) and non-breeders (0.0 or 0.1).

The SeaBird demographic modelling software was then used to determine variability in the estimates of survival and breeding rate using the dummy mark-resighting observations.  This assessment assumed that demographic rates were constant with respect to year and age and variability of demographic rates of wild populations are likely to be greater than those obtained by this assessment.  Increasing the banded sample size from 150 to 600 individuals led to an increase in the precision (c.v.) of annual survival breeding rate estimates.

With an input survival rate of 0.95 and a banded population of 150 individuals, the range of survival estimates was wide with 5 years of resighting effort (range from 0.91-0.99, x̅ = 0.95), though was much narrower with 10 years of resighting effort (0.93-0.96, x̅ = 0.95).  With a banded sample size of 600 individuals, the range of survival estimates was narrow with 5 years of resighting effort (0.93-0.97, x̅ = 0.95).

The precision of demographic rate estimates was not greatly affected by reducing the resighting probability of breeders from 0.6 to 0.4, though reducing the resighting probability of non-breeders from 0.10 to 0.00 produced imprecise estimates that were for some samples very different from input values.

To produce estimates of demographic rates that would be suitably precise for risk assessment purposes, this data simulation approach indicates that resighting effort over 5-10 years would be required subsequent to banding of a population between 150-600 individuals.  In a wild population, demographic rates are likely to change through time, so that greater sampling effort (in terms of banded individuals, number of resighting years or even resighting effort) may be required.”

White-capped Albatross, photograph by Graham Parker

Reference:

Roberts, J., Doonan, I. & Thompson, D. 2015.  Demographic Rate Estimation of White-capped Albatross Simulation Modelling.  Draft Copy Prepared for Department of Conservation June 2015.  Wellington: National Institute of Water & Atmospheric Research Ltd.  12 pp.

John Cooper, ACAP Information Officer, 12 July 2015