Appropriate efficient access to weather data is critical to the development of scientific data stewardship (Trenberth et al. 2002). In today’s data-rich environment, search for data based on time, geographic location, and sensor of interest limits our ability to locate data of known meteorological significance. The vast amounts of observations create opportunities for us to discover weather events with similar characteristics and behaviors in space and time, but without efficient information access tools, the majority of weather observations will remain out of reach. In this paper, we introduce a methodology to formulate weather events as GIS database objects and develop a set of indices to measure spatiotemporal characteristics, such as movement, rotation, and growth. With these indices, we demonstrate efficient data access with examples to retrieve appropriate weather data based on specified spatiotemporal characteristics of weather events. We also show how these indices can be used to categorize weather events embedded in observations. Furthermore, we will show how to combine the indices to search for weather events of similar behaviors. Consequent analysis of events with similar behaviors may offer insights into environmental influences to weather development. A prototype is developed in ArcView GIS with precipitation data, but the conceptual framework can be easily adapted to other GIS platforms.

Trenberth, K. E., K. R. Thomas, et al. (2002). "The Need for a Systems Approach to Climate Observations." Bulletin of the American Meteorological Society 83(11): 1593-1602.