We present an application of uncertainty visualization to air parcel trajectories generated from a global meteorological model. We derive an approximation of advection uncertainty due to interpolation and incorporate this uncertainty into our visualization of trajectories. Our work enables efficient visual pruning of unlikely results, especially in regions of atmospheric shear, potentially reducing erroneous interpretations. Finally, we apply these methods to a real-world meteorological problem to demonstrate its use.
This paper examines 3D meterological trajectories and the uncertainty incurred from interpolation errors. The data lives in a spatio-temporal sampling grid which is a 4-dimensional hypercube. The "main" trajectory is computed by interpolating the trajectory sample points at each point in the hypercube. Because the true value within the hypercube is unknown, the extreme bound of the trajectory is found by integrating a constant vector field using each of the extreme velocity terms and using the difference as an upper bound. The uncertainty is encoded in the thickness of the pathlines - thinner lines are less certain. Color is used to encode a particular variable, such as humidity, over time, as well as the the time step by using dark blue to show steps that have not yet been reached. An underlying synchornized 2D colorplot also shows relative humidity. The goal of this work is to answer questions relating to the development of hurricanes.
Data: Vector Uncertainty: 4D Visualization: 3D
Uncertainty visualization - Multi-field visualization - Flow visualization - Time-varying data - Meteorological visualization techniques
@Article{ boller:2010:AUMT,
author = {Ryan A. Boller and Scott A. Braun and Jadrian Miles and David H. Laidlaw},
title = {Application of Uncertainty Visualization Methods to
Meteorological Trajectories},
journal = {Earth Science Informatics},
year = {2010},
volume = {3},
number = {1-2},
pages = {119--126},
}