Modelling dynamics of Jakobshavn Isbræ and its contribution to sea level rise over the past and future century

Abstract

The rate of net ice mass loss from Greenland’s marine terminating glaciers has more than doubled over the past two decades highlighting their importance for future sealevel rise. Current projections are built upon observations from a short term record spanning only from several years to a decade. However, the last decade is dominated by anomalous dynamic changes and is therefore not representative of multi-decadal behaviour of glacier dynamics. A regional three-dimensional outlet glacier model developed as part of the PISM isused to simulate the behaviour of Jakobshavn Isbræ (JI; located in west Greenland) since the end of the Little Ice Age (LIA). The model is forced with different climate variables: near-surface air temperature, surface mass balance (SMB), sea-surface temperature and salinity. In order to accurately simulate and understand the longer term controls of dynamic changes, the model is constrained by observed terminus positions (1900-2014) and mass change estimates (1997-2014). The present study is the first that successfully simulated JI’s behaviour over the last century. For the period 1990-2014, the model simulated two major accelerations in 1998 and 2003that are consistent with observations of changes in glacier terminus. An initial, and most probably the first significant acceleration of JI after the end of LIA was modelled in ~1930. Overall, I found that the ocean influence in JI’s behaviour overthe last century is significant and most of the JI retreat during 1840–2014 is driven by the ocean parametrization and the glacier’s subsequent response, which waslargely governed by its own bed geometry. In my simulations, the retreat of the frontreduced the buttressing at the terminus and generated a dynamic response in theupstream region of JI which finally led to flow acceleration. This buttressing effecttends to govern JI’s behaviour. Consequently, the results showed that the dynamicchanges modelled at JI are triggered at the terminus. In a final phase, using this model that has been adjusted to the longer-term record, I performed experiments to the near future (i.e., 2013-2100) using five ocean temperature scenarios and two atmospheric scenarios (RCP 4.5 and RCP 8.5) asclimate forcing. In terms of mass change, I found that from the end of the LIA until the end of the 21st century JI’s mass change was and will remain predominately dynamically controlled (between 74 % and 86 % of the mass change is dynamic in origin). The study further indicates that the change in mass loss at JI is already triggered and that an eminent collapse of the terminus by the end of the century with retreat of up to 40 km along JI’s downward-sloping, marine-based bed is inevitable. A cooling experiment suggested that 0.9 °C of ocean cooling is needed to reverse the current retreat trend which however, will restart once cooling stops. The modelled cumulative mass change for the period 1840-2012 is ~1215 Gt or 3.4 mm SLRequivalent. For the 21st century, the best case scenario with regards to future warming (i.e., no increase in surface ocean temperature relative to present day)suggested mass loss estimates amount to ~1860 Gt by the year 2100 (67 % increase relative to 1840-2012) or 5.2 mm SLR equivalent. In the worst case scenario with regards to future warming expected mass loss of JI amounts to 3275 Gt by the year 2100 (192 % increase relative to 1840-2012) or 9.1 mm SLR equivalent. Overall, the study is unique both in approach and results obtained, and shows significant progress in modelling the temporal variability of the flow at JI. The study improves our quantitative understanding of the past and future of JI’s dynamics.