D3D: A potential 3D visualization tool for the National Weather Service

The NOAA Forecast Systems Laboratory (FSL) has been testing a software package that could potentially add the capability of 3D visualization to operational meteorological systems. The software package that has been the basis of D3D (Display Three Dimensions) at FSL is Vis5D, developed at the University of Wisconsin-Madison. In close cooperation with the University of Wisconsin, FSL has taken Vis5D and made a primary push towards changing the interface to make it appear as close as possible to the D2D interface that National Weather Service (NWS) forecasters are familiar with when using the Advanced Weather Interactive Processing System (AWIPS).

At the IIPS conference in 1999 we reported on our initial efforts in this development, which included a limited exercise testing D3D that involved local FSL staff and a few forecasters from the Denver (now Boulder) Weather Forecast Office (WFO). There were a number of suggestions for potential and/or improved products and displays that resulted from the exercise, but the overall response to D3D was very positive. This despite the use of a platform that sometimes resulted in rather slow response to D3D commands. Since that exercise the D3D interface has continued to evolve to where most of the Vis5D GUIs are no longer in place and the overall appearance is quite close to the D2D interface. We therefore felt that the next logical step in D3D development would be to expose the package to a wider variety of users via a more formal forecast exercise, much as was done in the early development days of D2D. On this basis, an exercise called RT99 was carried out from October to December 1999 at FSL, with one forecaster invited from each of the NWS five Regions, and from each National Center. All but one participant was able to attend RT99, which was a much more comprehensive exercise than the previous one, consisting of two weeks of training and D3D use with both real-time and delayed real-time data.

The exercise generated an extensive volume of evaluation material, from input on the D3D interface to potential value of 3D visualization to the forecast process. Our paper will summarize the results of RT99, and present our views on the next steps that might be undertaken with D3D.