ABSTRACT
Common data capture methods for geospatial data include remote sensing, Global Positioning System (GPS), eld surveys, and analog-to-digital conversion (e.g., digitizing and scanning) of existing maps. Each and every geospatial data are subject to some degree of error, a quantiable departure from the “ground truth” or the best-measured reference. During the process of data integration, the errors commonly originate from the sensing device, environmental characteristics, and the postprocessing operations conducted to transform the raw data into useful information for decision-making, such as radiometric correction, rectication, enhancement, conversion, classication, and so on (Jensen 2009). Uncertainty is commonly characterized using statistics based on the generalization of multiple observed errors (e.g., 90% condence interval). Formally dened, uncertainty is the unpredictability of an outcome due to a lack of knowledge about the appropriateness of parameters, models, or factors to be used in representing the complex and chaotic reality (Wang et al. 2005). Uncertainties in any components of the data integration would degrade the accuracy in describing the actual landscape of a geographic phenomenon.
