Seasonal geodetic strain in the Himalaya induced by surface load variations and implications for shallow elastic structure of the Earth

Avouac, Jean-Philippe¹; Chanard, Kristel^1,2; Ito, Takeo¹; Genrich, Jeff¹; Galetzka, John¹; Flouzat, Mireille³; and NSC team^4
1) Tectonics Observatory, California Institute of Technology, Pasadena, CA;
²) Department de Geosciences, Ecole Normale Superieure, Paris;
³) Laboratoire de Detection Geophysique, Commissariat a l'Energie Atomique, Bruyeres-le Chatel, France;
⁴) National Seismological Center, Department of Mines and Geology, Kathmandu, Nepal

We analyze geodetic time series from continuous Global Positioning System (GPS) stations across the Nepal Himalaya. As previously reported, strong seasonal variations are observed on both horizontal and vertical components (Bettinelli, 2008). We confirm here that seasonal variations of surface loading, due mostly to continental water storage, is probably the primary cause for these geodetic seasonal variations. In addition we show that this effect can be used to constrain the shallow elastic structure of the Earth. The integrated land water mass determined from the global time variations of the Earth’s gravity field measured by the Gravity Recovery and Climate Experiment (GRACE) is used to estimate surface load variations. To test the proposed model we take advantage of a larger dataset of GPS time series in the India – Nepal – Tibet area and a longer time period of GRACE water storage data than previous studies. We model seasonal variations of geodetic surface displacements using first an elastic half-space approximation (Boussinesq, 1885) and find that the observed signal at a number of stations in Nepal and India can indeed be predicted reasonably well. The best fit, in the least squares sense, is obtained for an elastic modulus of 145 GPa. This model is however difficult to assess given that it ignores the spherical and internal structure of the Earth. We therefore show simulations based on a more realistic spherical layered Earth structure (Farrell, 1972; Guo, 2004) . We consider an initial model based on the Preliminary Reference Earth Model (PREM), which is found to underestimate seasonal displacements amplitude. We next determine a best fitting model, in the Bayesian sense (Fukuda, 2008), by adjusting the distribution of velocities and density with depth to best match the geodetic time series. Variations by up to 10%, relative to PREM in the upper 200 km are inferred. The correction of the effect of surface load variations allows estimating better secular geodetic rates. It also enhances detection of eventual transient strain events like slow earthquakes.