Assessing Similarity of Geographic Processes and Events

The increased availability of spatiotemporal data collected from satellite imagery and other remote sensors provides opportunities for enhanced analysis of geographic phenomena. Much of the new data includes regular snapshots of the environment. Comparison of these snapshots can provide information about changes to the phenomena of interest. However, conventional GIS data models and analytical tools lack capabilities to adequately handle massive multidimensional data. One of the fundamental tools necessary to meet such challenges is query support to retrieve and summarize data according to dynamic geographic phenomena, such as geographic events and processes, of interest. Such query support depends upon abilities to assess spatiotemporal similarity so that data representing geographic events that exhibit the spatiotemporal characteristics of interest can be identified in a GIS database. To this end, this paper introduces a method to assess similarity of geographic events and processes (such as storms) based on their spatiotemporal characteristics (such as distribution of precipitation). We developed six indices to capture static and dynamic characteristics of geographic events and applied the Dynamic Time Warping method to temporal sequences of the six indices to examine the similarity among these events. With a case study, we demonstrated the proposed indices and method capable of comparing spatiotemporal characteristics of events as recorded in a GIS database and categorizing spatiotemporal data into groups of events according to their behavior in space and time.