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

Beginning of article

Introduction

Natural disasters cause suffering through the harm of people and infrastructure as well as enormous economical damage. Natural hazard management aims at minimizing these impacts by the tasks of prevention, event management, and rebuilding (Bezzola and Hegg, 2008). Assessments of natural hazards form the basis for all management tasks and are therefore a crucial component of hazard management. This fact has become apparent during the last years and consequently funds for the advancement of hazard assessments as well as the enhancement of management strategies have been allocated (e.g. by the Swiss Government).

Cartographic representations have proved suitable for the communication of hazard assessment results which is reflected in the fact that the generation of hazard maps as basis for land-use-planning is standard procedure in many countries and in some places even regulated by law (e.g. Switzerland (1), Colorado (2), and many more).

Motivation

Recent analyses of past flood events (Bezzola and Hegg, 2008) showed that the requirements towards hazard maps have increased over the last years: not only spatial planners for whom these maps were designed work with these visual representations of natural hazard assessment results, but also many other specialists involved in different tasks of hazard management. Bezzola and Hegg (2008) therefore suggest that hazard assessments should not be performed for a particular application anymore but as a general basis for various future uses. Once these multifaceted results exist, they can be visualized for specific users according to their requirements. These visualizations, however, have to be generated following cartographic design principles in order to produce clear and well balanced maps that are effortlessly readable.

An additional issue which is often discussed in different hazard management phases and tasks is the question of uncertainty inherent to hazard assessment results. Many important decisions that can have severe consequences for third parties (e.g. initiation of evacuation, construction bans, etc.) are based on these results. Information about the accuracy of the presented assessment results is therefore very important. However, until now, most hazard maps pretended absolute certainty by solid borders of hazard zones even though experts agree that the definition of hazard zones is associated with uncertainy. Apart from the difficulties of quantifying existing uncertainties, this issue also poses a cartographic challenge: there are no guidelines about suitable methods for uncertainty visualization in natural hazard maps and most existing recommendations are only of theoretical nature (Pang 2008).

Overview and References to Related Work

The Internet has evolved to one of the most relevant media to publish cartographic information, as it facilitates greater access to spatial information, increased levels of interactivity with maps, real-time locational information, and greater integration of multimedia content through pictures, sound, and video (Peterson, 2008). In recent years, web cartography shifted towards a distributed and service-oriented cartography, providing individual maps on-demand for specific purposes (Schnabel and Hurni, 2009). While early web maps were mostly raster-based and static, modern interactive applications allow for thematic as well as geographic navigation and offer visualization functionality to display available information according to the specific needs of the users. In addition, users can be guided through the map making process in order to avoid the violation of cartographic rules. Consequently, a web-based cartographic information system provides a well suited environment for the visualization and exploration of natural hazards assessment results as well as associated uncertainties.

Chesneau (2004) analyzed over two hundred hazard visualizations which were published in geographic journals and the Internet and observed that most maps are published in printed form; interactive or multimedia environments are rare. Her analysis also showed that most web-based maps offer little interactive functionality and consequently the implementation of animations and interactivity into natural hazard visualization environments is suggested. Research by Peterson (2007) confirms that it is generally believed that multimedia and interactive techniques can convey the multifaceted and dynamic character of the spatial environment much more effectively than static paper maps.

The lack of interactive functionality in web-based applications can also be observed in tools for the presentation of spatial data in general. In Switzerland for example, such tools have become common during the past years and every canton (= state or province) maintains its own system. These so called geoportals are designed for the general public and the typical application offers little interactivity: thematic content is available in a layer structure so that users can select the topics they want to have visualized in 2D maps and sometimes the query of attribute information is possible. Further interactions are generally limited to zooming and panning.

However, the need for interactive expert tools has been identified in recent research. Lienert et al. (2009) developed a web-based application for the real-time visualization of hydrological data. This application offers functionality to interactively monitor, retrace, and compare the available information. Romang et al. (2010) built on the experiences of snow avalanche tools and established an interactive early warning and information system for floods and debris flows.

Research in the field of uncertainty visualization has been ongoing for the last 30 years; after Buttenfield and Ganter (1990) analyzed Bertin's (1983) visual variables and suggested what characteristics of uncertainty can be illustrated by these variables, different studies have investigated and in some cases also evaluated potential visualization methods for the visualization of uncertainty in geospatial data sets (Buttenfield and Beard 1991; MacEachren 1992; McGranaghan 1993; Goodchild et al. 1994; Van der Wel et al. 1994; Wittenbrink et al. 1996; Leitner and Buttenfield 2000; Drecki 2002; Aerts et al. 2003; MacEachren et al. 2005; Zuk and Carpendale 2006). More recent research analyzed general uncertainty visualization methods and assessed their suitability for applications in the field of natural hazards: Trau and Hurni (2007) give an overview on existing methods and suggest visualizations suitable for the depiction of uncertainty in hazard and hazard index maps, Bostrom et al. (2008) present a review of research about the visualization of seismic risk and uncertainty and Pang (2008) finally discusses the issue of uncertainty associated to natural hazards data in detail and presents potential methods for visualizing uncertainty in natural hazards such as the application of blurriness, transparency, or fuzziness, the use of color hue, saturation, or value, the superimposition of a grid that is modified according to uncertainty values, the drawing of contour lines, the variation of the thickness, brightness, or connectedness of symbolization, the use of glyphs, histograms, or box plots, or the creation of complex 3D surfaces.

Aims

The objective of this research is to facilitate the interpretation of natural hazard data by implementing natural hazard assessment results into a web-based cartographic information system. Since …