Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.

Bibliography of the Maurice Lamontagne Institute


HAMMILL, M.O., B. FERLAND-RAYMOND, L.-P. RIVEST, G.B. STENSON, 2009. Modelling Northwest Atlantic Harp Seal populations : modifying an Excel model to R ; Modélisation des populations de phoques du Groenland du Nord-Ouest de l'Atlantique : modification d'un modèle Excel en R. DFO, Canadian Science Advisory Secretariat, Research Document ; MPO, Secrétariat canadien de consultation scientifique, Document de recherche, 2009/108, 15 p .

Click to see all the text

The model currently used to describe the population dynamics of the Northwest Atlantic harp seal (Pagophilus groenlandicus) is a two parameter model that uses information on age specific reproductive rates, ice-related mortality of young seals, removals and estimate of pup production. Although the basic model has taken a number of forms, in recent years it has been run using Excel as a basis. However, the current framework is very slow reducing the opportunities for testing different scenarios and management approaches. In order to improve performance, the model was transferred to R which allowed a reduction in the simulation processing time. During the transfer process, the model was also modified slightly, including a change to the resampling process of pregnancy rates to include correlation among age classes within a year. The R model also builds a new removal matrix for each Monte Carlo simulation using the modeled population age structure rather than the fixed age structure used previously. The Excel and the R models produced similar population trends, but the R model consistently estimated higher populations with a slightly smaller variance. This resulted from the elimination of negative age classes and by the optimisation process that produced lower mortality rates with a smaller variance. The lower variance resulted in higher L20 projections in the R model predictions.