No abstract

No abstract

Understanding groundwater storage (GWS) changes is vital to the utilization and control of water resources in the Tibetan Plateau. However, well level observations are rare in this big area, and reliable hydrology models including GWS are not available. We use hydro-geodesy to quantitate GWS changes in the Tibetan Plateau and surroundings from 2003 to 2009 using a combined analysis of satellite gravity and satellite altimetry data, hydrology models as well as a model of glacial isostatic adjustment (GIA). Release-5 GRACE gravity data are jointly used in a mascon fitting method to estimate the terrestrial water storage (TWS) changes during the period, from which the hydrology contributions and the GIA effects are effectively deducted to give the estimates of GWS changes for 12 selected regions of interest. The hydrology contributions are carefully calculated from glaciers and lakes by ICESat-1 satellite altimetry data, permafrost degradation by an Active-Layer Depth (ALD) model, soil moisture and snow water equivalent by multiple hydrology models, and the GIA effects are calculated with the new ICE-6G_C (VM5a) model. Taking into account the measurement errors and the variability of the models, the uncertainties are rigorously estimated for the TWS changes, the hydrology contributions (including GWS changes) and the GIA effect. For the first time, we show explicitly separated GWS changes in the Tibetan Plateau and adjacent areas except for those to the south of the Himalayas. We find increasing trend rates for eight basins: +2.46 ± 2.24 Gt/yr for the Jinsha River basin, +1.77 ± 2.09 Gt/yr for the Nujiang-Lancangjiang Rivers Source Region, +1.86 ± 1.69 Gt/yr for the Yangtze River Source Region, +1.14 ± 1.39 Gt/yr for the Yellow River Source Region, +1.52 ± 0.95 Gt/yr for the Qaidam basin, +1.66 ± 1.52 Gt/yr for the central Qiangtang Nature Reserve, +5.37 ± 2.17 Gt/yr for the Upper Indus basin and +2.77 ± 0.99 Gt/yr for the Aksu River basin. All these increasing trends are most likely caused by increased runoff recharges from melt water and/or precipitation in the surroundings. We also find that the administrative actions such as the Chinese Ecological Protection and Construction Project help to store more groundwater in the Three Rivers Source Region, and suggest that seepages from the Endorheic basin to the west of it are a possible source for GWS increase in this region. In addition, our estimates for GWS changes basically confirm previous results along Afghanistan, Pakistan, north India and Bangladesh, and clearly reflect the excessive use of groundwater. Our results will benefit the water resource management in the study area, and are of particular significance for the ecological restoration in the Tibetan Plateau.

S U M M A R YDuring the last glacial maximum (LGM), large ice sheets covered Scandinavia, the Barents Sea and the Northern British Isles. Subsequent to the LGM, the ice sheets disappeared and the solid Earth readjusted towards a new isostatic equilibrium. The glacial isostatic adjustment process is documented in numerous observations, for example, palaeoshorelines covering the last deglaciation phase, and ongoing crustal deformations monitored by GPS stations, for example, the BIFROST project. In this study, we use palaeoshoreline data from Scandinavia, the Barents Sea and northwestern Europe (NW) as well as radial crustal velocities from the BIFROST campaign to infer the radial viscosity structure of the Earth's mantle underneath Scandinavia and NW Europe. A global inverse procedure based on the Neighbourhood Algorithm allows us to explore the hypothesis of a low-viscosity zone in the upper mantle, which has been proposed in the literature. Our results indicate a low-viscosity zone underneath the Barents Sea, with viscosities between 10 19 and 10 20 Pa s in a depth interval of 160-200 km. No such low-viscosity zone is found underneath Scandinavia, and no clear indication for such a zone underneath NW Europe. The thickness of the rheological lithosphere increases from 60 to 70 km underneath NW Europe and the Barents Sea towards values exceeding 120 km underneath Scandinavia.

Our GPS velocity field is directly realized in a GIA reference frame. Using this method (named the GIA frame approach) we are able to constrain GIA models with minimal influence of errors in the reference frame or biasing signals from plate tectonics. The drawbacks are more degrees of freedom that might mask real but unmodeled signals. Monte Carlo tests suggest that our approach is robust at the 97% level in terms of correctly separating different models of ice history but, depending on deformation patterns, the identified Earth model may be slightly biased in up to 39% of cases. We compare our results to different one-and three-dimensional GIA models employing different global ice-load histories. The GIA models generally provide good fit to the data but there are still significant discrepancies in some areas. We suggest that these differences are mainly related to inaccuracies in the ice models and/or lateral inhomogeneities in the Earth structure under Fennoscandia. Thus, GIA models still need to be improved, but the GIA frame approach provides a base for further improvements.