Calibrating Glacier Surface Mass Balance Using Remote Sensing and Ensemble Kalman Filter

Surface mass balance (SMB) models are critical for understanding glacier evolution and projecting changes in response to climatic variations. This study presents a novel framework for calibrating SMB parameters using remotely sensed observations, incorporating the timing of data acquisition to improve accuracy and temporal relevance. The framework leverages the Ensemble Kalman Filter (EnKF), a robust data assimilation method, to iteratively refine model parameters based on incoming observations.

In our implementation, we decided on the Instructed Glacier Model (IGM) and embed it into the EnKF data assimilation approach. Before the transient ensemble simulations are started, a built-in stationary inversion is pursued to constrain ice-dynamic parameters and infer the basal topography. This stationary step relies on surface velocity, surface topography, and if available ice thickness measurements. For the transient evolution, a simple SMB model is calibrated using satellite-derived surface elevation changes. The calibration focuses on three primary parameters: the equilibrium line altitude (ELA) and two SMB elevation gradients for accumulation and ablation. This simplified SMB approach serves as a proof-of-concept, balancing simplicity with efficiency to showcase the effectiveness of the proposed method.

Initial results show that the method performs well for a synthetic glacier setup for which the target SMB is a-priori known. A sensitivity analysis highlights the importance of the key EnKF parameters. For real-world applications reasonable agreement is achieved with in-situ measurements - partially owing to the simple SMB approach. In summary, we are convinced that the approach could help improve our understanding of SMB processes, especially in regions with limited in-situ measurements.