NEWS
2026-01-05
Adjoint-Based Simultaneous State and Parameter Estimation in an Arctic Sea Ice-Ocean Model using MITgcm (c63m)
Editor: LYU Guokun (Polar Research Institute of China)
Polar region research has long been constrained by the scarcity of observational data and errors in ice-ocean coupled numerical simulations. Developing advanced ice-ocean coupled assimilation schemes can fully utilize observational data to correct model errors and improve the accuracy of model simulations and predictions.
The key challenges of ice-ocean coupled assimilation lie in constructing an error propagation model capable of quantifying the relationship between uncertain inputs and model simulation errors. To address this challenge, we have developed an adjoint-based scheme for simultaneous state and parameter estimation (adjoint-SPE). This scheme uses the adjoint model to project model-data misfits onto the initial condition field, atmospheric forcing field, and spatio-temporally varying ocean-sea ice coupling parameters (Figure 1).
The assimilation results show that the joint state and parameter estimation (SPE) substantially improves the sea ice concentration simulations. Particularly in October, when the ocean surface starts to refreeze, SPE reduces the lead closing parameter Ho (which determines the minimum ice thickness formed in the open water), thereby increasing the sea ice growth and facilitating the seasonal rapid sea ice recovery in the Arctic’s Pacific sector. Comparisons with sea ice thickness observations from the moored upward-looking sonars and Ice Mass Balance buoys demonstrate that incorporating optimized model parameters into the coupled model also leads to better sea ice thickness estimation.
Given that the optimal set of sea ice parameters may evolve alongside the thinning of Arctic sea ice, the adjoint-based SPE scheme has the potential to more accurately reconstruct the histical Arctic ocean and sea ice changes covering the satellite era, supporting research on Arctic sea ice and ocean variability.
Figure legends
Figure 1. Schematic diagram of the adjoint-based assimilation system. In this study, the control variable C includes the model initial condition field on January 1, 2012, the daily atmospheric forcing field, and the daily sea ice and ice-ocean coupling parameters.
Figure 2. Differences in sea ice concentration between simulations and satellite observations during October 12–22, 2012: (a) control experiment, (b) state estimation, and (c) state and parameter estimation (SPE). (d) shows the average sea ice concentration from the three experiments and satellite observations in the dashed region, where the SPE results are most consistent with the observations.
Reference: Lyu, G., Mu, L., Koehl, A., Lei, R., Liang, X., and Liu, C.: Adjoint-based simultaneous state and parameter estimation in an Arctic Sea Ice-Ocean Model using MITgcm (c63m), Geosci. Model Dev., 18, 9451–9468, https://doi.org/10.5194/gmd-18-9451-2025, 2025.