PCA study of the interannual variability of the GPS height and environmental parameters

• Main Author: Elia, Letizia
• Other Authors: Zerbini, Susanna
• Format: Master Thesis
• Language: English
• Published: Alma Mater Studiorum - Università di Bologna 2020
• Subjects: Geodesy,PCA,SVD,GPS,Climate indices,ENSO,TWS,Surface pressure,NAO,EA,SCAND,AO,TNA; Fisica del sistema terra [LM-DM270]; North Atlantic; North Atlantic oscillation
• Online Access: https://amslaurea.unibo.it/id/eprint/20438/; https://amslaurea.unibo.it/id/eprint/20438/1/Master_Thesis_Letizia_Elia%20_Finale.pdf

The objective of this study is to investigate a large network of GPS stations to identify and analyze spatially coherent signals present in the Up coordinate time series of the stations and, at the same locations, to identify and analyze common patterns in the series of environmental parameters and climate indices. The study is confined to Europe and the Mediterranean area, where 107 GPS sites were selected from the archive of the Nevada Geodetic Laboratory (NGL) on the basis of the completeness and length of the data series. The parameters of interest the Up coordinate of the GPS stations, the surface pressure (SP), the terrestrial water storage (TWS) and various climate indices: NAO (North Atlantic Oscillation), EA (East Atlantic), AO (Artic Oscillation), SCAND (Scandinavia), TNA (Tropical North Atlantic) and MEI v2 (Multivariate ENSO Index version 2). The Principal Component Analysis (PCA) is the methodology adopted to extract the main patterns of the space/time variability of these parameters. The work also focused on the coupled modes of space/time interannual variability between pairs of variables using the Singular Value Decomposition (SVD) methodology. The coupled variability between all the aforementioned parameters is investigated. This study has identified, over Europe and the Mediterranean, main modes of variability in the time series of the GPS Up coordinate, SP and TWS. The SVD analysis of coupled parameters, namely GPS Up-SP and GPS Up-TWS, showed that most of the common variability is explained by the first 3 modes. Moreover, the correlation between the GPS Up coordinate and the climate indices was estimated to investigate the possible influence of climate variability on the GPS Up behaviour. More than 30 stations, over the total of 107, show significant correlations with the AO, TNA and SCAND indices. The correlation coefficients with MEI v2 turn out to be significant and up to 0.5 for about half of the stations.