VISSYM '03 Proceedings of the symposium on Data visualisation 2003
| Hi-index | 0.00 |
A kinematics-based GIS methodology is applied to represent and analyze spatiotemporal patterns and pattern transitions in very large data sets. The study demonstrates that the kinematics approach is able to discern transitional patterns from a continuous field of geographic properties over time by defining objects through thresholds and analyzing the object's internal and external movement patterns in space and time. The kinematics approach quantifies divergence, rotation, and deformation about changes to precipitation patterns and enables the search for precipitation regions influenced primarily by local conditions or by general circulation patterns of water vapour transport. A use case is built from two precipitation data products projected for the A2 scenario by the International Panel for Climate Change (IPCC). The study takes a predefined threshold to delineate regions of interest, calculates shifts of the regions between years, and characterizes the pattern change. The study uses precipitation over 213 cm/year in 2001 and 2048 to illustrate the kinematics approach to comparing precipitation patterns predicted from the CCSM3 and CM3. Even though the precipitation data in 2001 and 2048 cannot be considered temporally continuous, the differential used here was to identify the patterns of precipitation shifts between the two years under the assumption that changes to spatial patterns of precipitation for 213 cm/year were gradual from 2001 to 2048. The 213 cm/year precipitation threshold is only met by a large number of precipitation events during the years of interest. Hence, this threshold appears stable from year to year although lesser thresholds would be discontinuous.