Customized Visualization of Natural Hazards Assessment Results and Associated Uncertainties through Interactive Functionality

Communication of natural hazard assessment results is crucial to protect people and infrastructure from devastating impacts of extreme events. While hazard maps provide important information on potential impacts, their interpretation and the general knowledge exchange between stakeholders is often difficult. Web-based information systems contain the potential to support hazard management tasks by fast distribution and customization of hazard visualizations through interactive functionality. Cartographic principles are, however, often ignored in existing web-based visualizations which leads to poor graphical results and consequently to an impairment of the information flow. While these issues need to be solved, a new task is already waiting: the integration of uncertainty information into hazard visualizations. Since many hazard management activities rely on hazard assessment results, communication of associated uncertainties among experts is vital.

The challenge of this research is to overcome these existing shortcomings by combining high quality cartographic visualizations of natural hazard data as well as associated uncertainties with interactive functionality. The resulting web-based cartographic information system will convene the needs of natural hazard specialists by offering a high level of customization: the suggested visualizations include various cartographic techniques such as the application of textures, bars, and interpolated surfaces. The possibility to interactively select particular data sets, customize colors, choose dimensions, query attribute data, and include uncertainty information facilitates the interpretation of complex data and finally the communication among natural hazard specialists.

In this article we summarize requirements that have to be considered, suggest functionalities necessary to perform natural hazards management tasks, and present a prototype of an expert system for the visualization and exploration of natural hazards assessments results and associated uncertainties.     

Reducing Linear and Perimeter Measurement Errors in Raster-based Data

An important issue in cartography and GIS is determining the appropriate resolution or cell size when converting vector data to raster. The general consensus is to make the cell size as small as possible to resolve geographic features and provide the most accurate estimates of measurements. Finer resolution results in more accurate estimates of polygon area; however, the raster data structure introduces an artifact that causes errors in the estimation of the length of linear features and of the perimeter of polygon features to increase with increasing resolution. Over-estimation as high as 41 percent is theoretically possible and was found to be around 26 percent for representative polygon maps. A method is described that uses a correction coefficient to reduce overestimation error to less than 3 percent.     

Interface Design for Geographic Visualization: Tools for Representing Reliability

Digital "softcopy" maps are becoming the norm—replacing static paper maps in applications from wayfinding to scientific research. As a result, the design of interface tools that allow users to manipulate map parameters effectively and efficiently is likely to become as fundamental to cartography as the design of maps themselves. This article presents some principles for the design of interfaces to geo-referenced data. These principles are summarized in a hierarchical approach to interface design with conceptual, operational, and implementational levels. This hierarchical approach leads designers from questions about the goals of the system and the users of that system to the creation of tools to accomplish those goals and interface controls that allow effective interaction with the tools. The article goes on to describe the application of these principles to a prototype geographic visualization system designed for exploration of spatial data sets and visualization of reliability of both data and data abstractions associated with environmental change. The prototype involves a synthesis of concepts and methods derived from cartography, scientific visualization, and exploratory data analysis into a system for exploratory spatial data analysis and spatial decision support.     

Highlighting in Geovisualization

Coordinated view geovisualizations allow users to interactively pick and attend to data observations across multiple views. This is frequently supported by the transient application of a visual effect to an observation during a mouse selection or rollover. This technique, known as highlighting, is typically implemented using a dedicated bright and saturated color to outline observations. In this paper we present a range of possibilities for alternative approaches to color highlighting, beginning with examples from the range of available visual variables and moving beyond those options to other, non-visual variable methods such as the use of lines to connect highlighted observations. We also describe design criteria for highlighting methods that can be used to predict and test the suitability of different approaches, and apply those criteria to our set of proposed methods to identify potential good candidates for implementation in future systems. Next, we present a set of highlighting types that define basic ways in which highlighting methods can be combined, modified, and driven directly by data values. We conclude by outlining several broad research challenges for future work on the development and evaluation of highlighting methods in geographic visualization.     

Exploration of Information Theoretic Arguments for the Limited Amount of Information in a Map

From the application of information theory capacity limits of maps are derived and some cartographic rules are formulated—based on theoretical models and perception studies. The perception study shows that for a relief map with coloured height intervals the channel capacity, as defined in information theory, is reached at seven～eight height classes. Generally, the length of a visual variable may be characterized by the channel capacity of the map. This theoretical study shows how increased attention should be paid to perceptual separation as the number of categories increases. The requirement becomes more and more crucial as the number of categories exceeds five. The principle of group visibility is formulated and demonstrated. Group visibility considers the visibility of groups of map features and may be utilized in a visual search for geographical patterns. The quantitative measures presented offers computational methods to control group visibility for example in an interactive visualization system.