Abstract
1. Both the quantification and modelling of harbour seal movement are required to predict the consequence of environmental change on population distribution and connectivity.
2. Two modelling approaches were considered. The first is an empirical InterHaulout Transition Rate (I-HTR) model which estimates the population probability of an individual moving from one haulout site to another. The second is a mechanistic Individual Based Model (IBM) of movement which uses seal physiology in a simulated quasi-realistic environment to predict movement patterns. The scope of the IBM development is to demonstrate its ‘proof of concept’.
3. In the I-HTR model, data from harbour seals tracked in Orkney and the Pentland Firth (22 GPS / GSM tags and 19 Argos tags) are used to populate a transition matrix (TM) showing the frequency of movement from one haulout site cluster to another. The TM is statically adjusted to represent the likely transitions of the local population of seals. It is then transformed to represent the probability of movement per unit time. The TM can now be used as the basis for a time-based simulation of seal movement and connectivity. However, the I-HTR model is empirical and as such does not consider the causes of movement, and so is not well suited to predict the consequence of environmental change.
4. A prototype individual based model (IBM) of harbour seal movement over time scales of days/weeks has been developed for the Pentland Firth / Orkney area. Many of the parameters in the model are place-holders (albeit realistic ones), with values that will be refined in the next stage of model development. The prototype IBM has a simple structure but it nonetheless captures the basic movement patterns and behaviour that are observed in harbour seal telemetry data.
5. To become a useful management tool, an IBM of appropriate complexity must be developed and tested, with the best possible estimates of parameters used to construct the model, e.g. from bioenergetic studies of captive seals or realistic estimates of habitat preference. The validation and checking of IBMs in general is an area of active research and for the seal IBM, appropriate checks may include comparison of model predictions in terms of summary properties and emergent properties of observational data, e.g. general patterns of spatial distribution.
6. The prototype IBM has proven the concept and development work should continue to test whether the available data (seal and environmental), statistical selection and fitting techniques can ultimately progress to producing a robust management tool. However, an important future challenge is to sufficiently map and quantify the dynamics of the geographical resources that seals require, such as haulout sites and foraging areas. Approaches could include the use of synoptic physical and biological data to predict those regions that may be preferred for foraging.