Comprehensive understanding of trends and variability in coastal mean sea level is vital for protecting shores under a changing climate. To understand the behavior of coastal relative sea level (RSL), it is crucial to identify all relevant processes. We combine data from various geophysical models and observations to determine whether the trends and decadal variability observed in relative sea level at tide gauges can be explained by the sum of all known contributors. A key contributor to RSL is vertical land motion, which is caused by glacial isostatic adjustment (GIA), solid earth response to surface loading, tectonics, and local effects. We explicitly model low-frequency loading effects to correct GPS records, which leads to a more consistent trend than only using GIA models. Secondly, we create sea level fingerprints based on estimates of ice melt and changes in land hydrology, which provide the RSL contribution due to large-scale mass transport. Since coastal areas are often located on shallow continental shelves, steric effects will generally be small, and a large fraction of the decadal sea level variability will have a remote steric origin. Therefore, we determine a relation between coastal sea level and deep sea steric variability. For the period 1950-2012, we find that for many locations, including the European coast, the observed and modeled RSL time series agree well on decadal and secular scales.

We estimate the total land water storage (LWS) change between 2003 and 2013 using a global water mass budget approach. Hereby we compare the ocean mass change (estimated from GRACE space gravimetry on the one hand, and from the satellite altimetry-based global mean sea level corrected for steric effects on the other hand) to the sum of the main water mass components of the climate system: glaciers, Greenland and Antarctica ice sheets, atmospheric water and LWS(the latter being the unknown quantity to be estimated). For glaciers and ice sheets, we use published estimates of ice mass trends based on various types of observations covering different time spans between 2003 and 2013. From the mass budget equation, we derive a net LWS trend over the study period. The mean trend amounts to+0.30 0.18mmyr1 in sea level equivalent. This corresponds to a net decrease of 108 64 km3 yr1 in LWS over the 2003–2013 decade. We also estimate the rate of change in LWS and find no significant acceleration over the study period. The computed mean global LWS trend over the study period is shown to be explained mainly by direct anthropogenic effects on land hydrology, i.e. the net effect of groundwater depletion and impoundment of water in man-made reservoirs, and to a lesser extent the effect of naturally-forced land hydrology variability. Our results compare well with independent estimates of human-induced changes in global land hydrology.