Radar altimetry for ice mass balance - impact of melting and refreezing in the snowpack

Climate change has caused increased mass loss from glaciers, ice sheets and sea ice over the last few decades, and that is affecting both sea level and ocean circulation. Satellite radar altimetry is one of few techniques that is capable of monitoring these changes through repeated ranging of Earth's surface. The precision of elevation measurements is not only dependent on the satellite and its instruments, but also the physical properties of Earth's surface, especially snow and ice which are partly penetrated by radar waves. This project has used detailed field observations of surface elevation and snowpack properties from Austfonna ice cap on Svalbard to determine the magnitude and variability of the penetration bias between altimetry-derived elevations from CryoSat-2 and the physical surface of a glacier. The results show that the winter-time penetration bias is near zero at the first point of signal reflection, known as the point-of-closest-approach (POCA), and then increases to 1-1.5 m further to the side of the satellite swath. Similar results were found for the colder snowpack of our second validation site on two ice rises in Dronning Maud Land, Antarctica. On Austfonna, where surface melting occurs every summer, the swath penetration is comparable to the snow depth down to the refrozen melt-layer of previous summer season and is relatively consistent from year to year. As the next melt season starts, the dominant reflection horizon jumps up to the wettened surface which later melts down during the summer and becomes the reference layer for the subsequent winter. This seasonality in snow penetration makes it difficult to monitor sub-annual glacier mass variations, but for longer time scales we have shown that robust trends of glacier thickening or thinning can be derived. After validating CryoSat-2 over Austfonna, we analysed all available data over Svalbard to derive glacier elevation trends for the period 2011-2017. We were able to determine elevation trends for 58% of the glacierized area by employing a linear regression technique to all repeated elevation measurements within 1 km grid cells. The results show widespread glacier thinning, particularly at lower elevations towards the fronts along the western and southeastern coasts of Svalbard, where thinning rates were typically 1-3 meters per year. Several glacier basins show evidence of glacier surging recognized by rapid ice-dynamical drawdown of ice from the interior towards the fronts, causing glacier advance or increased calving of ice to the ocean. The potential for short-lived glacier surging is an inherent property of many Svalbard glaciers at decades-to-century time scales, but the recent frequency and size of surges appear to be more substantial than in previous decades and could be linked with climate conditions. Apart from local impacts of glacier surges, significant thickening was only observed for the summit dome of Austfonna ice cap and some high-elevation areas of the main icefield in northeast Spitsbergen. We also analysed CryoSat-2 data similarly over Dronning Maud Land in East Antarctica, and here we found that some parts of the ice sheet have thickened slightly due to increased snowfall, but generally the elevation changes are much smaller than on Svalbard. Elevation change measurements can further be used to estimate glacier mass change and contribution to global sea level. To do that for Svalbard, we integrated the measured elevation trends over the glacier areas within selected regions and surge basins while assuming an average glacier density to convert from volume to mass change. This calculation shows that mass losses dominate in all glacier regions and are significantly larger than previous estimates, particularly in the southeastern parts of Svalbard bordering the Barents Sea. Over the 7-year period, we find that Svalbard glaciers have lost 110 +/- 30 gigatons of ice, which has contributed to a global sea level rise of about 0.3 millimeters. Time series analysis indicate moderate mass losses in the first few years of the record and higher losses since 2013, which agrees well with independent gravity measurements from the GRACE satellites. We find that glacier surges account for about one third of the total mass loss, whereas the remainder is due to a combination of more melting than snowfall at the surface and increased ice discharge to the ocean. Compared to earlier data we observe an apparent spread of glacier mass loss from the west coast of Svalbard to the Barents Sea margins coincides with reduced sea ice cover and upper ocean warming in the Barents Sea over the past decade, suggesting that this has played an important role in recent glacier mass loss on Svalbard and may continue to do so in the future.

I prosjektet har vi validert høydemålinger fra CryoSat-2 med detaljerte feltdata fra Austfonna på Svalbard over en periode på 7 år. Dette har gitt en unik innsikt i satellittdataenes nøyaktighet og presisjon, samt den variable signalpenetreringen i snødekket som avhenger av værforhold og klimatiske sesonger. Disse faktorene er avgjørende for å kunne gjøre pålitelige beregninger av isbreers massebalanse på regional og global skala, med direkte implikasjoner for globale havnivåendringer. Basert på de validerte CryoSat-2 dataene har vi kvantifisert massebalansen for breene på Svalbard for de ti siste årene. Resultatene viser at massetapet fra breene på Svalbard har økt markant, spesielt i de sørøstlige områdene mot Barentshavet. Dette har ført til et økt ferskvannsbidrag fra breene, noe som igjen har en betydelig påvirkning på lokale økosystem, sjøisdannelse og havsirkulasjon.

Earth observation from space is the only reliable method to determine accurate ice-mass changes and contribution to sea level rise at a global scale. Satellite SAR altimeters like CryoSat-2 are one of the most promising tools for this purpose, but precise applications over glaciers and ice sheets are hampered by variable penetration and backscatter in snow and ice, particularly during the transition from a cold winter snowpack to a melt-affected summer snowpack. This project will use an unique set of detailed ground-truth data, collected in the areas with densest CryoSat-2 data coverage, to calibrate and validate CryoSat-2 data locally over a full seasonal cycle. The synthesized results will be used to derive optimum techniques for elevation-change estimation over glaciers and ice sheets in different climatic regimes. Finally, we will apply the techniques to determine regional ice-mass changes in Svalbard and Dronning Maud Land, and their recent contribution to sea-level change.