Research Document - 2013/011

Updated stock-dynamic model for the Northern Hudson Bay narwhal population based on 1982-2011 aerial surveys

By M.C.S. Kingsley, N.C. Asselin, and S.H. Ferguson

Abstract

The Northern Hudson Bay narwhal population has been assessed from aerial surveys flown in the early 1980s, 2000, 2008, and 2011. The August 2011 survey provided information necessary for a full assessment of the population. A stock dynamic model using Bayesian methods and run on the OpenBUGS platform was developed in 2010 based on surveys up to 2008. Here, we update this model with the 2011 survey results, using adjustments for different survey methods (Asselin and Ferguson 2013) to assess the population size indices, and with the catch history to inform management of this population. To minimize differences due to the changes in survey methods, the 2011 data were re-analysed as though it had been collected and recorded using the recording methods of 2000. The valid information on stock dynamics comes from the differences between the 1982–84 surveys and those flown in 2000 and 2011. The dynamics of the population was modelled as a constant growth rate and limits to population growth at high numbers were not considered. In the present document, catch was the only mortality taken into account, and loss rates were considered to have the same distribution over the whole period. Three survey treatment options were run:

1. 2011 surveys not used in trend analysis;
2. original 2011 line-transect survey analysis used; and
3. re-analysis of 2011 according to 2000 survey methods (i.e., multi-strip survey) used.

The serial correlation of process errors is higher for the line-transect analysis of the 2011 observations, indicating that the results obtained by analysing the observations as a multi-strip survey (option 3) fit better with the earlier surveys, the catch sequence, and the assumption of a constant population growth rate. Under option 3, the population has annual growth of 1.2% and can barely support landed catches of 75 a year. The population trajectory has been more or less flat since the late 1990s with mean landed catches since then of about 84 a year; the odds of any decrease in numbers are estimated at 48% over 10 years. The estimated probability that the population will decline increases with time, even for catch levels that are associated with a slightly increasing population, because of the rising uncertainty of projections into the future. The results from this modelling exercise are uncertain and do not provide reliable estimates of future sustainable catches and further surveys are required. We conclude that management should continue to use Potential Biological Removal (PBR), rather than the risk-based approach reported in this document, until more surveys are undertaken.

Complete document (PDF, 333 KB)