Plotting the ensemble as multiple small postage stamp maps is common but makes it difficult to compare the locations of weather phenomena between the ensemble members.
All plots of aggregated ensemble data end up hiding some details. For example, probability plots make it easy to see the likelihood of a phenomenon at any map location but require choosing a single threshold, such as temperature below zero. Choosing a threshold discards information about the other possible values a parameter may have. It also requires mental effort to think about weather in terms of probabilities, rather than in the familiar terms of weather parameters, such as temperature or precipitation.
Our map visualization interleaves the data from the plots of all individual ensemble members into a single image and is able to show multiple values of a parameter from all the ensemble members at the same time, without hiding any details. This results in a crisp solid image where the ensemble members agree, but where the ensemble members diverge, the image becomes progressively fuzzier. The level of fuzziness intuitively conveys the level of uncertainty.
The main advantages of the method are: an intuitive visualization of evolving uncertainty, possibility to use the same color schemes in both deterministic and ensemble plots, and the ease and high performance of implementation.
In the taxonomy of ensemble visualizations by Wang et al., this method could be classified as a Point-oriented Composition after Visualization.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner