Jonathan Bamber, Richard M Westaway, Ben Marzeion, Bert Wouters

Since 1992, there has been a revolution in our ability to quantify the land ice contribution to SLR using a variety of satellite missions and technologies. Each mission has provided unique, but sometimes conflicting, insights into the mass trends of land ice. Over the last decade, over fifty estimates of land ice trends have been published, providing a confusing and often inconsistent picture. The IPCC Fifth Assessment Report (AR5) attempted to synthesise estimates published up to early 2013. Since then, considerable advances have been made in understanding the origin of the inconsistencies, reducing uncertainties in estimates and extending time series. We assess and synthesise results published, primarily, since the AR5, to produce a consistent estimate of land ice mass trends during the satellite era (1992 to 2016). We combine observations from multiple missions and approaches including sea level budget analyses. Our resulting synthesis is both consistent and rigorous, drawing on i) the published literature, ii) expert assessment of that literature, and iii) a new analysis of Arctic glacier and ice cap trends combined with statistical modelling. We present annual and pentad (five-year mean) time series for the East, West Antarctic and Greenland Ice Sheets and glaciers separately and combined. When averaged over pentads, covering the entire period considered, we obtain a monotonic trend in mass contribution to the oceans, increasing from 0.31±0.35 mm of sea level equivalent for 1992-1996 to 1.85±0.13 for 2012-2016. Our integrated land ice trend is lower than many estimates of GRACE-derived ocean mass change for the same periods. This is due, in part, to a smaller estimate for glacier and ice cap mass trends compared to previous assessments. We discuss this, and other likely reasons, for the difference between GRACE ocean mass and land ice trends.