Event Exploration in Spatiotemporal Data

Large amounts of spatially and temporally referenced data have been collected or generated for the monitoring and investigation of environmental systems. While the snapshot organization of data is conceptually simple and straightforward, it doesn’t represent and capture the dynamic characteristics in hydrometeorological systems. By defining the clusters of attributes and their changes in space and time as spatiotemporal events and by making the events explicit from spatiotemporal snapshot data, our research develops computational methods and tools to identify, represent, characterize, and explore events to support the investigation of dynamic behaviors of the hydrometeorological systems. For practical purposes, the identification and representation of events provide a content index to the raw time series of spatial snapshots, and allow queries and access of the spatiotemporal data based on the dynamic characteristics of the hydrometeorological systems (Tucker and Li, 2009). More information is available here.

Remote Sensing of Snow/Glaciers and Snowmelt Runoff Modeling in Mountain Watersheds

About one-sixth of the world’s population is dependent on glaciers and seasonal snow packs for their water supply (Barnett et al., 2005). The Tarim River basin, which is the largest continental river basin in the world, is located at the center of Eurasia in western China and surrounded by mountain watersheds dominated by snow and glacier cover. Here, mountain-fed rivers are the only available water resources to cover the needs of the oases scattered along its waterway into the Taklimakan Desert for public supply, agriculture irrigation, hydropower and other uses. Snow and glaciers in those mountain watersheds play an important role in forming the flow regime which depends on snow and glacier melt rather than the timing of precipitation. Change in snow and glacier cover in those mountainous regions is likely to disrupt the downstream flow pattern. Our research in this area includes a recent study on the changes of snow and glacier cover using MODIS, Landsat, and Corona imagery and stream flow simulation in the Tizinafu River watershed (Li and Williams, 2008). More information is available here.

Map Algebra

This research started with my dissertation which proposed data models and map algebra operations for vector fields, where each raster cell stores a vector instead of a scalar value. Moving beyond the dissertation, I began to examine how to model spatial dispersion where movement cost varies with direction. I developed an efficient data structure (a combination of a min-heap and a hash table) and algorithms to calculate the least cost path with anisotropic frictions (Li et al., 2005). Both the original map algebra and various extensions are exclusive to the raster data model. This limitation is manifested by the distinct ways of performing spatial analysis in the vector and raster data models. To overcome this limitation, we have sought to extend the map algebra operations to the vector data model (Keith and Li, 2010), which is the first to provide a conceptual framework and prototype syntax for the vector data model. Out most recent research in this area includes developing neighborhood analysis operations for stream networks and the parallelization of map algebra operations. More information is available here.

Impacts of Sea Level Rises

Nearly a quarter of the world’s population lives at elevations below 100 m from mean sea level and within 100 km from a coast (Nicholls and Small, 2002). Coastal regions have the greatest concentration of economic activities. Flooding caused by sea level rise will likely disrupt the physical processes, economic activities, and social systems in those regions. Our research in this topic focuses on development methods and tools to delineate potentially inundated areas resulting from projected sea level rises (Li et al., 2009).  More information is available here.