Choropleth maps are a common form of online cartographic visualization. They reveal patterns in spatial distributions of a variable by associating colors with data values measured at areal units. Although this capability of pattern revelation has popularized the use of choropleth maps, existing methods for their online delivery are limited in supporting dynamic map generation from large areal data. This limitation has become increasingly problematic in online choropleth mapping as access to small area statistics, such as high-resolution census data and real-time aggregates of geospatial data streams, has never been easier due to advances in geospatial web technologies. The current literature shows that the challenge of large areal data can be mitigated through tiled maps where pre-processed map data are hierarchically partitioned into tiny rectangular images or map chunks for efficient data transmission. Various approaches have emerged lately to enable this tile-based choropleth mapping, yet little empirical evidence exists on their ability to handle spatial data with large numbers of areal units, thus complicating technical decision making in the development of online choropleth mapping applications. To fill this knowledge gap, this dissertation study conducts a scalability evaluation of three tile-based methods discussed in the literature: raster, scalable vector graphics (SVG), and HTML5 Canvas. For the evaluation, the study develops two test applications, generates map tiles from five different boundaries of the United States, and measures the response times of the applications under multiple test operations. While specific to the experimental setups of the study, the evaluation results show that the raster method scales better across various types of user interaction than the other methods. Empirical evidence also points to the superior scalability of Canvas to SVG in dynamic rendering of vector tiles, but not necessarily for partial updates of the tiles. These findings indicate that the raster method is better suited for dynamic choropleth rendering from large areal data, while Canvas would be more suitable than SVG when such rendering frequently involves complete updates of vector shapes.

This dissertation advances spatial decision support system development theory by using a geodesign approach to evaluate design alternatives for such systems, including the impacts of the spatial model, technical spatial data, and user interface tools. These components are evaluated with a case study spatial decision support system for watershed management in the Niantic River watershed in Connecticut, USA. In addition to this case study, this dissertation provides a broader perspective on applying the approach to spatial decision support systems in general. The spatial model presented is validated, the impacts of the model are considered. The technical spatial data are evaluated using a new method developed to quantify data fitness for use in a spatial decision support system. Finally, the tools of the user interface are assessed by applying a conceptual framework and evaluating the resulting tools via user survey.

Citizen Science programs create a bi-directional flow of knowledge between scientists and citizen volunteers; this flow democratizes science in order to create an informed public (Bonney et al. 2014; Brown, Kelly, and Whitall 2014). This democratization is a fundamental part of creating a science that can address today’s pressing environmental, economic, and social justice problems (Lubchenco 1998). While citizen science programs create an avenue for sharing knowledge between the public and scientists, the exact program details and dynamics leading to different outcomes have not been studied in detail. The current shortcomings in the literature fall into three categories. First, the concept of ‘volunteer’ is used as a catch-all without considering how different

demographics (e.g. young, old, wealthy, poor, differently abled, local inhabitants, and visitors) affect both volunteer and scientific outcomes of citizen science. The second shortcoming: there are no standards to assess the quality of citizen science datasets. The third shortcoming: the volunteer and scientific outcomes of these programs are not routinely, or strategically, measured, or integrated into policy and planning (Brossard, Lewenstein, and Bonney 2005). This research advances the understanding of tourist volunteers in citizen science by examining these three shortcomings through a case-study in Denali National Park and Preserve. This case study included the development of the Map of Life-Denali citizen science program is a “tourist-friendly” program. Volunteers of the program use the Map of Life- Denali mobile application to record wildlife observations in the park. Research conducted on this program shows that tourists can be successful citizen science volunteers, and when compared to resident volunteers produce similar data, and have positive volunteer outcomes. The development of a fitness for use assessment, called STAAq is also a part of this research. This assessment is shown to be an effective method for assessing citizen science data quality. Throughout the development and launch of the program, stakeholders (the Park Service, and Aramark) were consulted. The Map of Life-Denali program will be integrated into the park’s shuttle and tour bus systems as an educational tool, however, the scientific merits of the program are still disputed.

Spatial uncertainty refers to unknown error and vagueness in geographic data. It is relevant to land change and urban growth modelers, soil and biome scientists, geological surveyors and others, who must assess thematic maps for similarity, or categorical agreement. In this paper I build upon prior map comparison research, testing the effectiveness of similarity measures on misregistered data. Though several methods compare uncertain thematic maps, few methods have been tested on misregistration. My objective is to test five map comparison methods for sensitivity to misregistration, including sub-pixel errors in both position and rotation. Methods included four fuzzy categorical models: fuzzy kappa's model, fuzzy inference, cell aggregation, and the epsilon band. The fifth method used conventional crisp classification. I applied these methods to a case study map and simulated data in two sets: a test set with misregistration error, and a control set with equivalent uniform random error. For all five methods, I used raw accuracy or the kappa statistic to measure similarity. Rough-set epsilon bands report the most similarity increase in test maps relative to control data. Conversely, the fuzzy inference model reports a decrease in test map similarity.