Spectral Analysis of Sub-Daily GPS Positions in Volcanic Areas

Geodetic space techniques, mainly Global Positioning System (GPS), are important tools to study ground deformations caused by natural phenomena since they can provide continuous three dimensional positions in all weather conditions. In addition to conventional daily positioning approach, epoch-by-epoch processing of GPS data allows us to detect short term movements such as volcanic eruption. However, the characteristics of sub-daily noise in GPS position estimates should be analyzed for the proper interpretation of the deformation. In this study, sub-daily positions of GPS sites located in a volcanic area on Montserrat were estimated. In order to clean outliers and identify hidden periodicities within the data, a sliding two-sided-kth neighbour windowing approach and Least Squares Spectral Analysis (LSSA) were applied to the de-trended data. LSSA is an effective technique to detect periodic signals. One of the main advantages of LSSA is that it can handle missing data which is quite common for volcanic regions, especially during the eruptions. Preliminary results showed that the common peaks are located at or near diurnal and semidiurnal tidal constituents such as K1 and K2. The calculated amplitudes can reach up to 5-6 mm for some stations.