EOF Analysis Using Python Library

EOF Analysis Using Python Library

Postby yiboj » Thu Mar 31, 2016 11:09 am

Empirical Orthogonal Function (EOF) analysis is to simplify a spatial-temporal data set by transforming it to spatial patterns of variability and temporal projections of these patterns. The spatial patterns are the EOFs, and can be thought of as basis functions in terms of variance. The associated temporal projections are the pricipal components (PCs) and are the temporal coefficients of the EOF patterns.

In this article, we are using the EOF library created by Dawson to analysis SST data from AVHRR Level 4 dataset in central pacific ocean from 19982 to 2000. You can download the python eof python module and document from the link: http://ajdawson.github.io/eofs/. The results are shown in the following plots and you can download the sample code here: https://podaac-tools.jpl.nasa.gov/drive ... s_forum.py and data here: https://podaac-tools.jpl.nasa.gov/drive ... f/eof_data

Figure 1. EOF 1 Mode and Time Series
eof1.png (44.31 KiB) Viewed 10110 times

Figure 2. EOF 2 Mode and Time Series
eof2.png (42.36 KiB) Viewed 10110 times

Figure 3. EOF 3 Mode and Time Series
eof3.png (44.72 KiB) Viewed 10110 times

Figure 4. EOF 4 Mode and Time Series
eof4.png (45.92 KiB) Viewed 10110 times

Figure 5. EOF 5 Mode and Time Series
eof5.png (46.15 KiB) Viewed 10110 times

Figure 6. Fraction of EOF Modes
eof_fraction.png (14.61 KiB) Viewed 10110 times
Posts: 130
Joined: Mon Mar 30, 2015 11:22 am

Return to Numerical Analysis