Data Mining and Visualization of Climate and Air Quality Datasets
College: Liberal Arts and Sciences
Award year: 2005-2006
This project has two related purposes: (1) to continue and complete the first phase of implementing new visualization tools for climate research started under the first Faculty Fellows project, and (2) to implement data mining and visualization tools available at NCSA in a new project to enhance the understanding and evaluation of large modeling and measurement datasets coming out of new studies of air quality and dust storm issues in China. Although the datasets used in both of these projects are quite different and have different purposes, the software tools required for data mining and visualization of these datasets bear strong relationships, and the developments done for one project should in many cases also be useful for the other project. Two graduate students working with Dr. Wuebbles, each working on one of the primary objectives, will also interact closely to allow the full benefits of these relationships.
Climate change is one of the primary concerns currently confronting humanity. The contribution of visualization techniques to improve the understanding of this complex issue could be huge but has not been fully developed. Climate data, both historical observations and future model-based projections, tend to create unwieldy and complex datasets containing multiple dimensions and many climate variables. Due to the massive amounts of data involved, analyses of datasets have generally been limited to contour plots and other two-dimensional graphics. Substantial, but incomplete, progress was made in developing new visualization capabilities in my first year as a Faculty Fellow. In my 2nd year as a NCSA Faculty Fellow, I plan to continue exploration of the potential for the application of such visualization tools through partnering with the data mining and visualization groups at NCSA, specifically through interactions with Donna Cox and Michael Welge.
This project will also implement data mining and visualization techniques towards improving our analysis of measurement and modeling datasets to better understand air quality and climate issues in China, plus resulting effects of long-range transport of pollutants from China on North America. Datasets of meteorology and pollution observations are available through collaborations with the Chinese Academy of Sciences and other key organizations. Our modeling tools are used for regional to global processes in research studies to understand the effects resulting from urban to regional air pollution resulting from energy, industrial and transportation sources, from dust storms and from biomass burning. The complex nature of these datasets requires new data mining and visualization capabilities to fully understand the underlying effects on the environment.
This project will extensively use the 15 node display wall developed at the Department of Atmospheric Sciences and is also aimed at taking advantage of the GEOWall now installed in our department. We should be able to make extensive progress during this project, and be able to put data mining and visualization capabilities into full operation for research and presentation. As with the first year, we expect the interactions with NCSA developed as a Faculty Fellow to continue to grow and prosper in coming years through external support development.