Abstract
Water mass exchanges between land and ocean affect ocean mass and, consequently, sea level. The land contribution to ocean-mass change (OMC) originates from mass changes in multiple water storage compartments (WSCs). These compartments include glaciers, snow, canopy, soil, groundwater, and surface water bodies. The combined water mass in all these compartments constitutes terrestrial water storage (TWS). Changes in TWS are caused by an array of natural and anthropogenic drivers. Their accurate quantification remains a challenge due to uncertainties regarding the impact of climate conditions and human activities. Due to their importance for water mass exchange and sea level, an improved understanding of the processes underlying TWS changes is needed. To deepen our understanding, a special focus on climate-induced glacier mass loss is critical, as it is one of the major driving processes of TWS changes. Climate-induced glacier mass loss is also an important driver of streamflow variability in glacier-fed regions. Streamflow variability directly impacts human populations and ecosystems by affecting downstream water availability. A large portion of the world’s glaciers are projected to exceed “peak water” (i.e. maximum annual glacier runoff) within this century, signaling a subsequent decline in annual glacier runoff. Regions dependent on the delayed supply of glacier meltwater are particularly vulnerable to this impending change. A deeper comprehension of the key factors controlling the changing contribution of glacier runoff to streamflow is required to better predict the future evolution of glacier-fed regions. The research presented within this dissertation is driven by two major motivations: 1) to improve our understanding of the WSCs and drivers governing the contribution from continents to OMC and thus sea-level change at the global scale, and 2) to improve our understanding of past changes in glacier runoff and its contribution to streamflow across major glacier-fed regions globally. The dissertation comprises three standalone studies. The first study investigates the global net contribution from continental TWS changes (excluding Greenland and Antarctica) to OMC over January 1948–December 2016. This contribution is quantified by computing global TWS anomalies (TWSAs) with a modified version of the WaterGAP 2.2d global hydrological model (GHM) that includes the representation of glaciers. The model is validated against satellite-based time series of global monthly TWSAs over January 2003–August 2016. The assessment shows that continents lost 34–41 mm of sea level equivalent (SLE) to oceans during 1948–2016. The primary driver is identified as global glacier mass loss, which constitutes 81% of the overall loss. A decline in glacier-free land water storage is responsible for the remaining 19% of the overall loss. This decline is attributed to sustained water abstractions for human use (15–21 mm SLE) primarily from groundwater (13–19 mm SLE). This human-induced water mass loss was partially offset by a water mass gain of 8 mm SLE, resulting from the artificial impoundment of water through the construction of dams and reservoirs. Climate variability also resulted in a slight water mass gain on glacier-free land through increased precipitation. Moreover, climate variability is pinpointed as the primary factor driving the interannual variability of TWSAs. A considerable part of global climate variability originates from El Niño-Southern Oscillation (ENSO) teleconnections. Substantial uncertainty arises from the simulation of groundwater depletion, which is highly sensitive to irrigation water use, and interannual variability, which is highly sensitive to precipitation forcing. The second study examines the impact of ENSO events on water storage variability in 24 major river basins worldwide. The CWatM v1.04 and WaterGAP 2.2e GHMs are used to compute TWSAs, and anomalies in the soil, groundwater, and river WSCs. By removing the long-term trend and seasonal signals based on harmonic analysis, a residual signal containing the interannual variability is isolated. The models are validated against satellite-based global and basin-scale monthly TWSA residuals over April 2002–December 2016. The influence of ENSO is assessed by calculating the correlation between monthly residual anomaly time series and Multivariate ENSO Index (MEI) intensities over January 1979–December 2016. Models show a remarkable agreement concerning the correlation between TWSA residuals and MEI intensities. Most basins exhibit a negative correlation to the MEI. Strong correlations are observed in (sub-)humid basins in northern South America and Southeast Asia. The highest correlation is found in the Amazon basin (–0.56). The correlation’s direction reveals that water storage in basins such as the Amazon, Parana, and Yangtze is primarily influenced by the warm ENSO phase (i.e. El Niño) rather than its cold counterpart (La Niña). In many of the basins, the strongest correlation to MEI is found in one of the individual WSCs rather than in TWS. Model discrepancies regarding the correlation to MEI intensities are largest for the groundwater compartment.