Search for Global Atmospheric Patterns through Principal Component Analysis on Radio Occultation Data

BURGOS FONSECA, Y., P. ALEXANDER, A. DE LA TORRE, R. HIERRO, AND P. LLAMEDO

Graz

Workshop; OPAC-IROWG 2016; 2016

Universidad de Graz

The GPS Radio Occultation (RO) technique provides diverse atmospheric variables of interest for the study of atmospheric patterns. We analyze the behavior of temperature and water vapor at 850 hPa (lower troposphere) obtained from global reprocessed GPS RO ?wet? retrievals for the period 2007-2013. Data were monthly averaged in 20°x12° (longitude x latitude) cells. The huge amount of data urges for the use of a specific method that may be able to locate and extract the most significant information. A principal component analysis (PCA) was applied in S-mode to detect given time variabilities and where they occurred. Five latitudinal bands have been defined to avoid the possible bias introduced by the increasing density of cells at higher latitudes in this kind of analysis. We searched for known variation patterns over time and others possibly unknown in the PCA scores. In this preliminary study we kept only the first component. We used wavelets to infer these issues. Percentages indicate the amount of variance explained. PCA loadings show the prevailing regions for these behaviors (deep blue or red show the highest coincidence areas).