Abstract
The Western Alps and the Pyrenees are mountain ranges that can almost be qualified asintraplate domain given the horizontal deformation rate so low that we cannot currentlyquantify it with geodesy. However, present-day tectonic deformation and seismicity inWestern Europe is essentially concentrated in these mountain ranges. Nocquet (2012) showedthat the deformation rates across Western Europe are so low that they remain belowmeasurement uncertainties, with horizontal motion across the Pyrenees and the Western Alpsof ~0 ± 0.5 mm/yr.To study tectonic deformation in the Western Alps and Pyrenees region, we have analyzed thetimes-series of 166 GPS permanent stations of RGP, RENAG, EUREF and IGS networkswith times-series length from 1,8 to 16,0 years. Data were processed using a Precise PointPositioning (PPP) approach. We have examined the influence of different corrections on thehorizontal and vertical velocities: (1) Earth Rotation Parameters (ERP), (2) Global MappingFunction (GMF) and the Vienna Mapping Function (VMF1) tropospheric delays, (3) absoluteantenna phase center (APC). In general, the influence of the troposphere model, the ERP andAPC corrections are negligible in terms of horizontal velocities. In contrast, ERP and APCcorrections affect vertical velocities with differences at ~0.5 mm/yr level. We have alsoanalyzed the effects of surface mass loading due to changes in continental water storage(GLDAS), atmospheric pressure (ATML), and non-tidal ocean loading (NTOL). On average,the combination of loading corrections (GLDAS + ATML + NTOL) result in an increase inthe seasonal signals: annual amplitudes (estimated using a best-fit sine function) are increasedby 0.10, 1.55 and 0.50 mm for the north, east, and vertical components, respectively. Hence,this combination of loading models does not seem to be appropriate to correct the time-series.Loading corrections have a significant influence on horizontal and vertical velocities(horizontal and vertical average differences of 0.24 and 0.55 mm/yr compared to uncorrectedvelocities). Therefore, the surface loading models must be improved before they can be usedto improve the GPS velocity estimates.We estimated the minimum time spans of GPS continuous data required to achieve 7 differentlevels of velocity precisions. The uncertainty of velocity estimates from GPS time-seriesstrongly depend on the length of time-series data. We examine the stability and uncertainty ofvelocity estimates by a convergence analysis (i.e., estimation of necessary observation time toobtain a velocity close to that calculated for the complete time series). To obtain an unbiasedand realistic comparative analysis, we have compared a synthetic solution of forward andbackward time for velocity and uncertainty estimates. On the basis of this analysis, weestimate that the precision of 0.5 mm/yr in velocity solution is achieved after an average timespan of 4,43 and 4,78 years of continuous GPS data for the horizontal and verticalcomponents, respectively.In this study, the Euler rotation pole for the stable Western European plate in the ITRF2008reference frame was defined from a 62-sites subset with an RMS of residual horizontalvelocities of 0.29 mm/yr level. The Euler pole is located at 53.730°N, and -101.856°E and hasa rotation rate of 0.256°/Myr. We also present the present-day velocity field with precisionsbetter than 0.5 mm/yr in the Western Alps and the Pyrenees region. The results show nosignificant vertical movements in the Pyrenees, in contrast with the vertical velocities of theWestern Alps that can reach up to 2.49 mm/yr.