Topological Data Analysis of the NASA ATom-1 Data
Abstract
Big data sets such as the Atmospheric Tomography Mission (ATom) provide an exciting opportunity to test new data-scientific methods such as Topological Data Analysis (TDA). For atmospheric chemistry applications, TDA is an especially promising technique, as it is well-suited for identifying cycles in data -an invaluable capability for the highly cyclical chemical and meteorological pathways in the atmosphere. Additionally, TDA is performed in a coordinate-free environment, so it requires different (often less problematic) assumptions about variable relationships than traditional statistical approaches. Finally, TDA's ability to handle high-dimensional data removes the need to select only a few variables in advance of the analysis. We used the TDA package in R to analyze data from the ATom-1 mission which occurred during July-Aug 2016. Specifically, we employed density cluster trees, TDAmapper, and persistent homology. We found that TDA is an efficient tool for analyzing large data sets. Looking forward, TDA is a promising technique for developing new insights in the field of atmospheric chemistry.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.A11F2279K
- Keywords:
-
- 0322 Constituent sources and sinks;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 1910 Data assimilation;
- integration and fusion;
- INFORMATICSDE: 1986 Statistical methods: Inferential;
- INFORMATICSDE: 3260 Inverse theory;
- MATHEMATICAL GEOPHYSICS