Search

Global warming has deeply affected coastal and offshore marine ecosystems and is thought to have a significant impact on the fish stocks and their decline in the northwest Atlantic region in Canada. However, it is difficult for scientists and marine biologists to have a clear and precise insight of this impact. Studying fish aggregations and their evolution in time and space may help scientists to have a better understanding of the decline of fish stocks and elaborate potential solutions for it. Fish aggregations are not only 3D spatiotemporal phenomena which need to be represented and managed in 3D but also have fuzzy boundaries which makes it too difficult to clearly identify and delineate them in the space. Although, Geographic Information Systems (GIS) constitute a powerful tool for handling spatial information, they are prone to problems when dealing with 3D dynamic phenomena, especially, when those phenomena have fuzzy boundaries. This paper addresses these two problems at local and regional scales and proposes two new approaches for the representation and visualization of fish aggregations and their evolution through time and space. At the regional scale, the proposed approach combines fuzzy logic methods with spatial raster representation tools within GIS to provide a more realistic representation and visualization of fish aggregations and their evolution in time. At the local scale, we developed an integrated method based on 3D Delaunay triangulation and the 3D alpha-shapes algorithm to carry out the spatial modeling of fish aggregations in a true 3D space. The applications of these methods to the fisheries data reviled several potentials and limitations of the proposed methods which are discussed throughout this paper.

The effective management of disasters requires providing relevant and right information to the concerned decision makers. By its nature, disaster management involves multiple actors and organizations, potentially implying a significant volume of geospatial data coming from heterogeneous and autonomous geospatial data sources. Integration of these data sources is difficult not only because of the semantic heterogeneity of data but also because of the dynamic nature of the reality that is studied. The dynamic aspect of the reality has a direct impact in the conceptualisation of such a reality by adding different event categories to the domain ontology, thus making more complicated to apply existing methods for the mapping and integration of ontologies. In this article, we highlight some problems of heterogeneity that complicate the integration of ontologies composed of objects and events concepts; we also propose a similarity model designed to support mapping of these ontologies.

Modeling and simulation of dynamic phenomena helps specialists and decision makers to better understand, analyze, and predict natural disasters to reduce the related damages. Hence, several models and simulation approaches have been developed with their own strengths and limitations. One important factor which has to be taken into account when dealing with a disaster is its dynamic behavior and its geometrical and topological representation. Geographic information systems are very well adapted for spatial data organization, visualization, querying, and analysis, and may be helpful in the context of simulation and modeling of spatial phenomena (e.g. floods). However, regarding the three dimensional and dynamic nature of spatial data related to a disaster (which changes with respect to space and time), several complexities are added to the entire process of simulation using the existing GIS.