Name
Abstract | ||
|---|---|---|
We present applications of uncertainty visualization methods to a global meteorological model, allowing better understanding of the composition of the local environment of developing hurricanes. Our work enables efficient visual pruning of unlikely results, especially in regions of atmospheric shear. We derive bounds on advection uncertainty due to interpolation and incorporate this uncertainty into our visualization of trajectories, facilitating visual pruning. By identifying trajectories that indicate a protection of storm's core from outside influence, we also attempt to corroborate the viability of a recently devised meteorological theory that hurricanes develop in protected, "marsupial" pouches. Index Terms—Uncertainty visualization, multi-field visualization, flow visualization, time-varying data, meteorological visualization techniques. This paper describes the application of uncertainty visualization methods to meteorological flow. The driving goal of this work is to improve hurricane prediction through a better understanding of the interaction of a storm and its local environment. Some important factors in this understanding are the air's source, path, and composition. These properties can be neatly encapsulated and represented by trajectories, otherwise known as pathlines or particle paths. These trajectories begin, or are "seeded," as a set of particles at a user-specified location in space and time, then travel through a time-varying vector field. In meteorological studies, this initial volume of seeds is known as an air parcel; the vector field is the air-parcel 3-dimensional velocity. The concept of backward trajectories (1) is applied to this problem since we know the "final" position of air we are interested in: that which is near developing hurricanes. Therefore we seed the air parcel trajectories near the storm and work backward through time and space to determine the source and path the air took to its final position. Along the way, we can also maintain "snapshots" in time of the path's properties, such as its relative humidity. While these analyses provide a good approximation of the air parcels' paths, the oft-overlooked problem of identifying and representing sources of potential uncertainty (2) is addressed. The pathlines are a derived quantity, but the accuracy of their derivations (3) is rarely represented. We provide an upper bound and visual representation for the uncertainty of these calculations. The aim is to prevent the user from reaching erroneous conclusions about the storm and its environment based upon trajectories with a high degree of uncertainty resulting from passage of air parcels through regions of sharp velocity gradients or errors associated with the underlying interpolation. It also facilitates the efficient visual pruning of unlikely results by allowing the user to disregard trajectories with higher uncertainty. |
Year | DOI | Venue |
|---|---|---|
2010 | 10.1007/s12145-010-0052-5 | Earth Science Informatics |
Keywords | DocType | Volume |
uncertainty visualization.multi-field visualization.flow visualization.time-varying data. meteorological visualization techniques,3 dimensional,vector field,relative humidity,upper bound,flow visualization,indexing terms | Journal | 3 |
Issue | ISSN | Citations |
1-2 | 1865-0481 | 3 |
PageRank | References | Authors |
0.38 | 6 | 4 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ryan A. Boller | 1 | 3 | 0.38 |
| Scott A. Braun | 2 | 3 | 0.72 |
| Jadrian Miles | 3 | 27 | 1.40 |
| David H. Laidlaw | 4 | 1781 | 234.58 |