| Increase in computing power has led to a substantial increase in the size of scientific and engineering data sets. Often, research in high-dimensional spaces requires analysis of terabytes of data. This in turn has led to an increase in the demand for simplified representations of these large datasets for effective analysis. Scientific visualization facilitates visual interpretation of massive data sets. Thus, visualization is driven by the needs of a broad spectrum of research areas.;The first part of this dissertation describes use of topology to segment two-dimensional tensor fields. The second part describes a ray intersection method to generate dual isosurfaces for trivariate, volumetric data. The third part describes an interactive visualization system for visualizing water-resource data.;In the first two parts, we describe how topology can serve as foundation for two different areas: (a) In tensor field interpretation, to extract topology of the field based on different interpolation schemes to reduce complexity. (b) In volume visualization, to ensure topological correctness of isosurfaces, and using that as the underlying principle of the dual-isosurfacing algorithm.;In the third part, we present the visualization systems developed as a part of application-driven research concerned with management and planning of water resources. We describe two systems which support: (a) a global analysis of multi-parameter time-series data of different components of a large hydrological system, and (b) a localized statistical analysis of time-series data of a single parameter in a specific part of a hydrological system. |
