Alumni
PhD 2004

Supervisor
Dr Andrew Trites

Thesis
Modelling and mapping resource overlap between marine mammals and fisheries on a global scalemarine mammals and fisheries on a global scale.

Potential competition for food resources between marine mammals and fisheries has been an issue of much debate in recent years. Given the almost cosmopolitan distributions of many marine mammal species, investigations conducted at small geographic scales may, however, result in a distorted perception of the extent of the problem.

Unfortunately, the complexity of marine food webs and the lack of reliable data about marine mammal diets, abundances, food intake rates etc., currently preclude the assessment of competition at adequately large scales. In contrast, the investigation of global resource overlap between marine mammals and fisheries (i.e., the extent to which both players exploit the same type of food resources in the same areas) may, however, be easier to achieve and provide some useful insights in this context. Information about the occurrence of species is a crucial pre-requisite to assess resource overlap and also to address other marine mammal conservation issues. However, the delineation of ranges of marine mammals is challenging, due to the vastness of the ocean environment and the difficulties associated with surveying most species.

Consequently, existing maps of large-scale distributions are mostly limited to subjective outlines of maximum range extents, with little additional information about heterogeneous patterns of occurrence within these ranges. I developed a new, more objective approach to map global geographic ranges and the relative environmental suitability (RES) for 115 marine mammal species. This habitat suitability model is a rulebased environmental envelope model that can utilize not only quantitative data, but also alternative, nonquantitative and more readily available information about species habitat preferences (such as expert knowledge).

As a first step, I assigned each species to broad-scale ecological niche categories with respect to depth, sea surface temperature and ice edge association based on synopses of published qualitative and quantitative habitat preference information. Within a global grid with 0.5 degree latitude by 0.5 degree longitude cell dimensions, I then generated an index of the relative environmental suitability (RES) of each cell for a given species by relating quantified habitat preferences to locally averaged environmental conditions in a GIS modeling framework. RES predictions closely matched published distributions for most species, suggesting that this rule-based approach for delineating range extents represents a useful, less subjective alternative to existing sketched distributional outlines. In addition, raster-based predictions provided information about heterogeneous patterns of the relative suitability of the environment and potential core habitat for each species throughout its range.

I validated RES model outputs for four species (northern fur seals, harbor porpoises, sperm whales and Antarctic minke whales) from a broad taxonomic and geographic range using at-sea sightings from dedicated surveys. Observed relative encounter rates and species-specific predicted environmental suitability were significantly and positively correlated for all species. In comparison, observed encounter rates were positively correlated with < 3 % of 1000 simulated random data sets of species occurrence.

To validate the RES predictions for data-deficient species such as beaked whales (Ziphiidae, Cetacea), I developed a different evaluation approach using stranding records as an alternative type of test data. Ziphiids represent one of the least known families of mammals, primarily known from strandings, along with only a few known at-sea records for each species. Using a global data set of residual ocean currents, I developed a simulation model of ziphiid strandings and used this to generate relative probabilities of strandings along all coastlines. Predictions were generated based on two different input distributions: species-specific RES predictions and uniform distributions based on published information, which served as the null model. Large-scale patterns of simulated strandings based on RES predictions produced significant correlations with observed strandings for five times as many species (10 of 21 ziphiid species) as those generated based on the null model (2/21), suggesting that RES predictions represent an improvement over existing simple outlines.

The extensive validation provided support that RES predictions capture patterns of species occurrence sufficiently enough to be used as the basis for large-scale investigations of marine mammal-fisheries interactions. I therefore used the model to assess the importance of spatial aspects for the investigation of overlap between marine mammals and fisheries in terms of food resource exploitation. To assess spatially-explicit resource overlap, I first estimated global annual food intake (specified by food types) for each species based on a basic food consumption model. This model required information about global population abundances, sex-specific mean weights, standardized diet compositions, and weightspecific feeding rates, which was obtained through the compilation, screening and processing of more than 2000 publications. By linking species-specific RES predictions with estimated consumption for each species, I generated spatially-explicit food consumption rates (expressed as food intake per square-km per year).

Superimposing geographically disaggregated fisheries catches (generated by a similar model) allowed me to calculate overlap between catches and consumption with respect to both the composition of food types and areas where food / catches were taken. The model indicated that, in the 1990s, average consumption by all marine mammal species combined was several times higher than total fisheries catches during the same time period. However, effective spatial overlap and exploitation of the same food types was relatively low, suggesting that actual competition between fisheries and marine mammals may be much lower than proposed. I predicted the highest resource overlap in the temperate to subpolar shelf regions of both hemispheres, though overlap is more pronounced in the North. Overall, < 15 % of all fisheries catches and < 1% of all estimated marine mammal food consumption stem from areas of high predicted overlap.

Nevertheless, overlap between iv marine mammals and fisheries may be an issue of concern on smaller scales (especially for species with small feeding distributions) that requires more detailed, local investigations. I propose that mapping of suitable habitat for marine mammals using the new RES model is useful for evaluating current assumptions and knowledge about species occurrences, especially for data-poor species. Generated hypotheses about suitable habitat and species occurrences may help to focus research efforts on smaller geographic scales, and usefully supplement other, statistical habitat suitability models. Furthermore, the mapping of food consumption rates and geographical ‘hotspots’ of marine mammalfisheries interactions will help to identify potential areas of highest conflict and may aid the development of management approaches at appropriate scales.

Publications

Modelling and mapping resource overlap between marine mammals and fisheries on a global scalemarine mammals and fisheries on a global scale. Kaschner, K. 2004. PhD Thesis, University of British Columbia, Vancouver BC. 225 pages (PDF)

Mapping world-wide distributions of marine mammal species using a relative environmental suitability (RES) model. Kaschner, K., R. Watson, A. W. Trites and D. Pauly. 2006. Marine Ecology Progress Series 316:285-310.(PDF)

Modeling and mapping trophic overlap between marine mammals and commercial fisheries in the North Atlantic. Kaschner, K., R. Watson, V. Christensen, A.W. Trites and D. Pauly. 2001. In D. Zeller and R.D.Pauly Watson (eds), Fisheries impacts on North Atlantic ecosystems: catch, effort and national/regional datasets. Fisheries Centre Research. 9(3):35-45. (PDF)