Abstract
The 'Geovis' project was initiated by the researchers at the section for Meteorology and Oceanography (METOS) at the department of Geosciences. The goal was to develop a custom visualization tool that can be used for easy loading and presentation of climate datasets. The project also aimed at enhancing the communication of complex climate data to a diverse audience, including students, educators, the media, and policymakers. For this, users at the department of Geosciences, initially relied on a well-known NetCDF reader tool called “ncview”. However, “ncview” had several limitations, such as a lack of 3D visualization capabilities and frequent crashes while using it. For scientific dissemination and educational activities, “ncview” was not a great choice since it lacks the feature of geographic projection, e.g. visualization on the sphere-shaped Earth. Instead, the METOS community wanted a simple 3D visualization tool with customized workflows for visualizing netCDF datasets on the sphere. Technical requirements included the development of loading time series data and creating animated visualizations. The users also wanted to incorporate custom features like ability to easily choose, inspect, activate and compare certain data variables of interest with the help of visualizations.
Background
Methodology
The implementation of the project was done as an open-source standalone application using Python and 3D libraries like VTK. The tool with a simple user interface allowed users to load their netCDF data. Unlike ncview, the GeoVis tool offered a 3D interactive view of the dataset which gave the users a more natural context of the planet. With the custom workflows implemented, the users were allowed to select variables of interest, define the range of data and modify the color maps for an optimum visualization. Additional visualizations, such as the contour map enabled users to compare multiple variables changing over time. Finally, the animation export features in the tool helped to save the visualizations as a video file.
GeoVis can now be utilized in several environments, including locally, in a Silver+ virtual instance, on the GEO department’s server named “mimi,” and under EduCloud OnDemand for utilizing high-end GPUs. In future, we also plan to port some of the functionality as an extension in ParaView so that a user can even scale very large datasets utilizing multiple nodes and multiple GPUs.