Abstract
Abstract. Estimating the contribution of Antarctica and Greenland to sea-level rise is a hot topic in glaciology. Good estimates rely on our ability to run a precisely calibrated ice sheet evolution model starting from a reliable initial state. Data assimilation aims to provide an answer to this problem by combining the model equations with observations. In this paper we aim to study a state-of-the-art ensemble Kalman filter (ETKF) to address this problem. This method is implemented and validated in the twin experiments framework for a shallow ice flowline model of ice dynamics. The results are very encouraging, as they show a good convergence of the ETKF (with localisation and inflation), even for small-sized ensembles.
Highlights
Antarctica and Greenland account for a significant fraction of today’s sea-level rise (Shepherd et al, 2012)
Calibrating the model means identifying as best we can the unknown parameters of the model, and finding a reliable initial state means of identifying the model variables that vary in time
To the best of our knowledge, neither Kalman filter (KF) nor ensemble Kalman filter (EnKF) has ever been used in glaciology, and we propose to implement and investigate its performance to tackle the sea-level rise problem
Summary
Antarctica and Greenland account for a significant fraction of today’s sea-level rise (Shepherd et al, 2012) Good estimates of their future contribution are crucial to producing sea-level change forecasts, as underlined by Hanna et al (2013). Producing pertinent estimates of polar ice sheet contribution to sea-level change relies on our ability to run a precisely calibrated ice sheet evolution model starting from a reliable initial state. Calibrating the model means identifying as best we can the unknown parameters of the model (bedrock topography, basal friction law), and finding a reliable initial state means of identifying the model variables that vary in time (ice thickness). In those cases, localisation and inflation are mandatory. Obtained performances were good even for ensembles with a size as small as 30
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
