CR3.3 | Advances in sea-ice modelling: developments and new techniques
Orals |
Thu, 10:45
Posters on site |
Thu, 16:15
Posters virtual
Thu, 14:00
EDI
Advances in sea-ice modelling: developments and new techniques
Co-organized by NP1/OS1
Convener: Clara BurgardECSECS | Co-conveners: Carolin MehlmannECSECS, Adam BatesonECSECS, Lorenzo Zampieri, Einar Örn Ólason
Orals
| Thu, 18 Apr, 10:45–12:30 (CEST)
 
Room 1.34
Posters on site
| Attendance Thu, 18 Apr, 16:15–18:00 (CEST) | Display Thu, 18 Apr, 14:00–18:00
 
Hall X4
Posters virtual
| Attendance Thu, 18 Apr, 14:00–15:45 (CEST) | Display Thu, 18 Apr, 08:30–18:00
 
vHall X4
Orals |
Thu, 10:45
Posters on site |
Thu, 16:15
Posters virtual
Thu, 14:00

Orals: Thu, 18 Apr | Room 1.34

Chairpersons: Carolin Mehlmann, Einar Örn Ólason, Clara Burgard
10:45–10:55
|
EGU24-11288
|
On-site presentation
Towards improving numerical sea ice predictions with data assimilation and machine learning 
William Gregory, Mitchell Bushuk, Yongfei Zhang, Alistair Adcroft, and Laure Zanna

In this presentation we highlight recent developments in the implementation of Machine Learning (ML) algorithms into the large-scale sea ice model, SIS2. Specifically, we show how a Convolutional Neural Network (CNN) can be used to systematically reduce global sea ice biases during a 5-year ice-ocean simulation. The CNN has been trained to learn a functional mapping from model state variables to sea ice concentration Data Assimilation (DA) increments. Therefore, during model integration, the CNN ingests information about the numerical model's atmosphere, ocean, and sea ice conditions, and predicts the appropriate correction to the sub-grid category sea ice concentration terms (without seeing any actual sea ice observations). We also show how this combined DA+ML approach leads to a natural framework for augmenting training data for neural networks; one which can lead to significant improvements in online performance, without the need for direct online learning. The bias reductions over the 5-year simulation period for this CNN correction scheme are even competitive with the bias reductions achieved from DA. These findings therefore suggest that our approach could be used to reduce systematic sea ice biases in fully coupled climate model predictions on seasonal-to-climate timescales.

How to cite: Gregory, W., Bushuk, M., Zhang, Y., Adcroft, A., and Zanna, L.: Towards improving numerical sea ice predictions with data assimilation and machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11288, https://doi.org/10.5194/egusphere-egu24-11288, 2024.

10:55–11:05
|
EGU24-12912
|
On-site presentation
Improvements in September Arctic sea ice predictions via assimilation of summer CryoSat-2 sea ice thickness observations 
Yong-Fei Zhang, Mitch Bushuk, Michael Winton, Bill Hurlin, William Gregory, Jack Landy, and Liwei Jia

Because of a spring predictability barrier, the seasonal forecast skill of Arctic summer sea ice is limited by the availability of melt-season sea ice thickness (SIT) observations. The first year-round SIT observations, retrieved from CryoSat-2 from 2011 to 2020, are assimilated into the GFDL ocean–sea ice model. The model's SIT anomaly field is brought into significantly better agreement with the observations, particularly in the Central Arctic. Although the short observational period makes forecast assessment challenging, we find that the addition of May–August SIT assimilation improves September local sea ice concentration (SIC) and extent forecasts similarly to SIC-only assimilation. Although most regional forecasts are improved by SIT assimilation, the Chukchi Sea forecasts are degraded. This degradation is likely due to the introduction of negative correlations between September SIC and earlier SIT introduced by SIT assimilation, contrary to the increased correlations found in other regions.

How to cite: Zhang, Y.-F., Bushuk, M., Winton, M., Hurlin, B., Gregory, W., Landy, J., and Jia, L.: Improvements in September Arctic sea ice predictions via assimilation of summer CryoSat-2 sea ice thickness observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12912, https://doi.org/10.5194/egusphere-egu24-12912, 2024.

11:05–11:15
|
EGU24-11908
|
ECS
|
On-site presentation
A data-driven sea-ice model with generative deep learning 
Tobias Sebastian Finn, Charlotte Durand, Flavia Porro, Alban Farchi, Marc Bocquet, Yumeng Chen, and Alberto Carrassi

The current generation of sea-ice models with Brittle rheologies can represent the observed temporal and spatial scaling of the sea-ice dynamics at resolutions of around 10 km. However, running those models is expensive, which can prohibit their use in coupled Earth system models. The promising results of neural networks for the fast prediction of the sea-ice extent or sea-ice thickness offer an opportunity to remedy this shortcoming. Here, we present the development of a data-driven sea-ice model based on generative deep learning that predicts together the sea-ice velocities, concentration, thickness, and damage. Trained with more than twenty years of simulation data from neXtSIM, the model can extrapolate to previously unseen conditions, thereby exceeding the performance of baseline models.

Relying on deterministic data-driven models can lead to overly smoothed predictions, caused by a loss of small-scale information. This is why the ability to perform stochastic predictions can be instrumental to the success of data-driven sea-ice models. To generate stochastic predictions with neural networks, we employ denoising diffusion models. We show that they can predict the uncertainty that remains unexplained by deterministic models. Furthermore, diffusion models can recover the information at all scales. This resolves the issues with the smoothing effects and results in sharp predictions even for longer horizons. Therefore, we see a huge potential of generative deep learning for sea-ice modelling, which can pave the way towards the use of data-driven models within coupled Earth system models.

How to cite: Finn, T. S., Durand, C., Porro, F., Farchi, A., Bocquet, M., Chen, Y., and Carrassi, A.: A data-driven sea-ice model with generative deep learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11908, https://doi.org/10.5194/egusphere-egu24-11908, 2024.

11:15–11:25
|
EGU24-13528
|
ECS
|
Virtual presentation
A MAGICC Arctic Sea Ice Emulator 
Sian Chilcott, Malte Meinshausen, and Dirk Notz

CMIP6 models present our best understanding of the Earth system, yet they currently fail to simulate a plausible evolution of sea ice area to changes in the global-mean temperature. We aim to assess whether correcting the temperature and Arctic Amplification biases between CMIP6 models and observations can simulate a sensitivity of sea ice loss to global warming that is within the plausible range. To do this, we develop an emulator that is calibrated to physically-based CMIP6 models and then constrained to observations. Such a tool efficiently translates the global-mean temperature of a specific year into a physically-based and observationally constrained probabilistic ensemble of SIA in each month. This setup allows our emulator to capture the core physical processes of CMIP6 projections, while capturing the observed sensitivity of sea ice loss to global warming through the observational constraint of Arctic Amplification. While there are many application possibilities of our emulator, we use our model here to probabilistically diagnose the timing of an ice-free Arctic Ocean. We find that under a high (SSP5-8.5), medium (SSP2-4.5) and low (SSP1-2.6) emission scenario, an ice-free September Ocean is ‘likely’ at 1.73 of global warming above the pre-industrial level, however we note that the probability in the lower emission scenario reduces to ‘unlikely’ in the late 21st century as the global temperature partially recovers. Our projections suggest that the probability of an ice-free summer ocean rises rapidly from ‘unlikely’ at 1.5 of global warming to ‘likely’ at 2 of global warming, stressing the importance of preventing global temperatures rising above 1.5, as the probability of losing sea ice coverage in September rises sharply thereafter. For March, we also find that the observational constraints increase the probability of an ice-free ocean under SSP5-8.5, becoming ‘likely’ in early 2200, while the probability remains very low under SSP2-4.5 and SSP1-2.6 as less than 5% of models reach ice-free conditions. Our projections suggest an ice-free summer ocean could occur at 0.5 cooler levels than the CMIP6 multi-model ensemble mean implies. Likewise, our approach suggests the probability of an ice-free Arctic Ocean year-round is increased when constraining the Arctic Amplification to observations.

How to cite: Chilcott, S., Meinshausen, M., and Notz, D.: A MAGICC Arctic Sea Ice Emulator, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13528, https://doi.org/10.5194/egusphere-egu24-13528, 2024.

11:25–11:35
|
EGU24-3367
|
ECS
|
On-site presentation
Perscribing Antarctic landfast sea ice in a sea ice-ocean model. 
Noé Pirlet, Thierry Fichefet, Martin Vancoppenolle, Clément Rousset, Pierre Mathiot, Alexander Fraser, Antoine Barthélemy, and Christoph Kittel

The coastal polynyas of the Southern Ocean play a crucial role in the formation of dense water and have an impact on the stability of ice shelves. Therefore, it is important to accurately simulate them in climate models. To achieve this goal, the relationship between grounded icebergs, landfast ice and polynyas appears to be central. Indeed, grounded icebergs and landfast ice are believed to be key drivers of coastal polynyas. However, ESMs do not simulate Antarctic landfast ice. Moreover, at a circumpolar scale, there are no observations of grounded icebergs available. Hence, we must seek model representations that can overcome these issues. To address these gaps, we conducted a study using an antarctic circumpolar configuration of the ocean–sea ice model NEMO4.2-SI–3 at the 1/4° resolution. We ran two simulations for the period 2001–17, with the only difference being the inclusion or exclusion of landfast ice information based on observations. All other factors, including initial conditions, resolution and atmospheric forcings, were kept the same. We then compared the results of these simulations with observations from the advanced microwave scanning radiometer to evaluate the performance of the new simulation. Our analysis allowed us to determine the extent to which prescribing the distribution of landfast ice and setting the sea ice velocity to zero on landfast ice regions influenced various aspects of the sea ice, such as polynyas, landfast ice and sea ice distribution in the model. In the future, we plan to look at the impact on the ocean and to develop a physical parameterization in order to model landfast ice and consequently polynyas on a permanent basis.

How to cite: Pirlet, N., Fichefet, T., Vancoppenolle, M., Rousset, C., Mathiot, P., Fraser, A., Barthélemy, A., and Kittel, C.: Perscribing Antarctic landfast sea ice in a sea ice-ocean model., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3367, https://doi.org/10.5194/egusphere-egu24-3367, 2024.

11:35–11:45
|
EGU24-19333
|
On-site presentation
Recent progress in nesting a DEM- based regional sea ice model within a continuum model 
Wenjun Lu, Andrei Tsarau, Yuan Zhang, Raed Lubbad, and Sveinung Løset

Understanding sea-ice dynamics at the floe scale is crucial to improve regional ice forecast and comprehend the polar climate systems. Continuum models are commonly used to simulate large-scale sea-ice dynamics. However, they have both theoretical and computational limitations in accurately representing sea-ice behaviour at small scales. Discrete Element Models (DEMs), on the other hand, are well-suited for modelling the behaviour of individual ice floes but face limitations due to computational constraints. To address the limitations of both approaches while combining their strengths, we explored the feasibility of nesting a DEM within a continuum model. This paper reports recent progresses in addressing two challenges associated with this method: 1) how to couple a discrete element method (DEM) – based model (a Lagrangian model explicitly tracking each element in space) into a continuum model (a Eulerian model with fixed spatial mesh transferring state variables within); 2) how to explicitly model fracture of sea ice at large scales. Based on our assessment, integrating DEM and continuum model simulations showed potential for offering accurate, high-resolution predictions of sea ice, particularly in coastal areas and near islands. Simulating fracture of sea ice still poses great computational challenges. However, we see a potential in a data-driven approach to accelerate the computational efficiency in resolving floe-scale ice fractures.  

How to cite: Lu, W., Tsarau, A., Zhang, Y., Lubbad, R., and Løset, S.: Recent progress in nesting a DEM- based regional sea ice model within a continuum model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19333, https://doi.org/10.5194/egusphere-egu24-19333, 2024.

11:45–11:55
|
EGU24-4144
|
ECS
|
On-site presentation
A model for ice-mélange based on particle and continuums mechanics 
Saskia Kahl and Carolin Mehlmann

Ice mélange (a mixture of sea ice, bergy bits and icebergs) can have a strong influence on the sea-ice-ocean interaction. So far, ice mélange is not represented in climate models as numerically efficient realizations are missing. This motivates the development of an ice-mélange model based on the viscous-plastic sea-ice rheology, which is currently the most commonly used material law for sea ice in climate models. Starting from the continuum mechanical formulation, we modify the rheology so that icebergs are represented by thick, highly compact pieces of sea ice. These compact pieces of sea ice are held together by a modified tensile strength in the material law. In this framework, the ice mélange is considered as one single fluid, where the icebergs are realised by particles.
Using idealized test cases, we demonstrate that the proposed changes in the material law are crucial to represent icebergs with the viscous-plastic rheology. Similar to the viscous-plastic sea-ice model, the ice-mélange model is highly nonlinear. Solving the model at the resolution needed to represent the typical size of icebergs in ice mélange (< 300m) is therefore challenging. We show that the ice-mélange formulation can be approximated efficiently with a modified Newton's method. Overall, the simple extension of the viscous-plastic sea-ice model is a promising path towards the integration of ice mélange into climate models.

How to cite: Kahl, S. and Mehlmann, C.: A model for ice-mélange based on particle and continuums mechanics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4144, https://doi.org/10.5194/egusphere-egu24-4144, 2024.

11:55–12:05
|
EGU24-5374
|
ECS
|
On-site presentation
Floe-scale ocean / sea ice energy transfers in the marginal ice zone 
Mukund Gupta, Andrew Thompson, and Patrice Klein

Marginal ice zones are regions where individual sea ice floes interact mechanically and thermodynamically with turbulent ocean currents at the (sub-)mesoscale. Fine scale exchanges of momentum, heat and salinity at the interface between the ocean and the sea ice floes have important effects on upper-ocean energetics, under-ice tracer mixing, and the ice-pack melt rates. The dynamics of these moving floes remain poorly constrained, notably due to the challenge of numerically resolving sub-mesoscale processes and modelling the discrete behavior of sea ice in traditional climate models. 

Here, we use oceanic Large Eddy Simulations (LES), two-way coupled to a Discrete Element Model (DEM) of disk-shaped sea ice floes, to quantify the kinetic energy transfers between ocean and sea ice during summer-like conditions, varying sea ice concentration and floe size distribution. The damping of oceanic currents by floes is found to be important for a sea ice concentration as low as 40%, when the sizes of floes are comparable to the characteristic eddy size. This damping is largely compensated by the generation of kinetic energy due to melt-induced baroclinic instability at the edge of sea ice floes, leading to a net energy sink of approximately 15%, relative to a simulation with no floes. At higher sea ice concentrations, the oceanic kinetic energy production weakens, while energy loss due to ice/ocean damping and floe-floe collisions both increase. These energy fluxes are mediated by the spatial aggregation of sea ice floes that occurs within the high-strain regions surrounding ocean mesoscale eddies. Eddy-driven aggregation can also reduce the melt rate of small floes as they become shielded from warm waters by neighboring larger floes. These results highlight the need for scale-aware, and specifically floe-scale parameterizations of sea ice and its coupling to ocean turbulence, within global climate models.

How to cite: Gupta, M., Thompson, A., and Klein, P.: Floe-scale ocean / sea ice energy transfers in the marginal ice zone, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5374, https://doi.org/10.5194/egusphere-egu24-5374, 2024.

12:05–12:15
|
EGU24-348
|
ECS
|
On-site presentation
Ridge-formation simulations in three dimensions using discrete element methods 
Marek Muchow and Arttu Polojärvi

Sea-ice ridges form as a part of sea-ice deformation, while the ice is moved by winds and ocean currents. While ridging is a localized process, it is assumed to limit the compressive strength of sea ice in large scale. However, formulations of large-scale ice strength, as used in Earth System Models, do not consider individual ridge formation processes in detail. Thus, it is necessary to understand the energy spend in ridge formation and various processes related to generating ice rubble and redistributing it. To investigate ridge formation in detail, we use the Aalto University in-house discrete-element-method (DEM) model. This three-dimensional DEM model features deformable, multi-fracturing, ice floes, which can fail and form ridges when coming into contact, while recording the ridging forces. With this, we discuss why three-dimensional simulations are important to investigate ridge formation process.

How to cite: Muchow, M. and Polojärvi, A.: Ridge-formation simulations in three dimensions using discrete element methods, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-348, https://doi.org/10.5194/egusphere-egu24-348, 2024.

12:15–12:25
|
EGU24-18506
|
ECS
|
On-site presentation
Direct Numerical Simulation of shear turbulence interacting with a melting-freezing ice layer 
Diego Perissutti, Francesco Zonta, Alessio Roccon, Cristian Marchioli, and Alfredo Soldati

When a turbulent flow of water interacts with an ice boundary at near-freezing temperature, the fluid can undergo freezing or melting, depending on the local temperature. The turbulence structures that develop in proximity to the ice layer can affect the convective heat transport patterns, leading to the formation of complex phase-boundary morphologies. The ice layer evolves as part of the solution and modifies the near-boundary fluid structures, resulting in heat transfer perturbations. We investigate these ice-water interactions at small scales by performing Direct Numerical Simulations of an open channel flow at shear Reynolds number in the range between 10^2 and 10^3. The upper section of the channel is occupied by ice, while free shear conditions are applied at the bottom. Temperature is imposed on both walls. The ice melting/freezing is simulated using a phase field method [1] combined with a volume penalization immersed boundary method. A pseudo-spectral scheme [2] is used to solve the equations for momentum and energy transport and for phase evolution. We investigated how the behavior of the system changes with the flow conditions (i.e. Reynolds number), with a specific focus on characterizing the features of the ice morphology. In particular, we observed a remarkable influence of turbulence intensity on the ice morphology: at low shear Reynolds, the typical streamwise-oriented canyons already reported in similar studies [3] are present. However, at higher shear Reynolds, spanwise instabilities are triggered, making the final ice morphology more complex.

FIgure1: Render view from below of the open channel flow at a low Reynolds number. On the top section of the channel, the corrugated ice layer is shown. On the ice boundary, the normalized heat flux passing through it is displayed (high heat flux is shown in red, low heat flux in blue). The local temperature field is reported on the side domain boundaries. The typical streamwise-oriented canyons at the ice interface are visible and the heat flux correlates well with those patterns (the heat flux is higher inside the canyons).

[1]R. Yang et al., Morphology evolution of a melting solid layer above its melt heated from below, Journal of Fluid Mechanics, 956, A23, 2023.

[2]F Zonta et al., Nusselt number and friction factor in thermally stratified turbulent channel flow under non-Oberbeck–Boussinesq conditions, International journal of heat and fluid flow, 44:489–494, 2013.

[3]L. A. Couston et al., Topography generation by melting and freezing in a turbulent shear flow, Journal of Fluid Mechanics, 911, A44, 2021.

How to cite: Perissutti, D., Zonta, F., Roccon, A., Marchioli, C., and Soldati, A.: Direct Numerical Simulation of shear turbulence interacting with a melting-freezing ice layer, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18506, https://doi.org/10.5194/egusphere-egu24-18506, 2024.

12:25–12:30
Poster express teaser

Posters on site: Thu, 18 Apr, 16:15–18:00 | Hall X4

Display time: Thu, 18 Apr, 14:00–18:00
Chairperson: Adam Bateson
X4.1
|
EGU24-22361
neXtSIM-DG – A next-generation discontinuous Galerkin sea ice model 
Einar Ólason, Timothy Spain, and Thomas Richter and the The neXtSIM team

We present neXtSIM-DG, the novel sea ice model created as part of the Scale Aware Sea Ice Project (SASIP). NeXtSIM-DG is a continuum sea ice model that combines several new model paradigms at once: besides established rheologies, we use the newly developed Brittle Bingham–Maxwell rheology. The discretization is based on higher-order continuous and discontinuous finite elements. We take advantage of the object orientation of the C++ implementation of the model to create a flexible, maintainable, and easily modifiable code base ready for adaptation and adaptation by the user. Finally, the C++ implementation uses modern data structures that allow for efficient shared-memory parallelization and are ready for GPU acceleration. These aspects reflect better the different scales of sea ice dynamics in space and time. In this poster, we review the basic modelling features and present some details of numerical realization. In particular, we study the effect of high-order discretization and the role of different rheologies. 

How to cite: Ólason, E., Spain, T., and Richter, T. and the The neXtSIM team: neXtSIM-DG – A next-generation discontinuous Galerkin sea ice model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22361, https://doi.org/10.5194/egusphere-egu24-22361, 2024.

X4.2
|
EGU24-5177
|
ECS
Improving the representation of snow over sea-ice in the SI3 model 
Theo Brivoal, Virginie Guemas, Clement Rousset, and Martin Vancoppenolle

Snow plays a crucial role in the formation and sustainability of sea ice. Due to its thermal properties, snow acts as an insulating layer, shielding the ice from the air above. This insulation reduces the heat transfer between the sea-ice and the atmosphere. Due to its reflective properties, the snow cover also strongly contributes to albedo over ice-covered region, which gives it a significant role in the Earth's climate system.

Current state-of-art climate models use over-simple representations of the snow cover. The snow cover is often represented with a one-layer scheme, assuming a constant density, no wet or dry metamorphism or assuming that no liquid water is stored in the snow. Here, we present the integration of a more advanced snow scheme (ISBA-ES) into the sea-ice model SI3, which serves as the sea-ice component for upcoming versions of the CNRM climate model (CNRM-CM). We compare 1D simulations over the Arctic using this new scheme with observational data and simulations utilizing the previous SI3 snow scheme. Overall, the snow simulated by the ISBA-ES scheme is realistic. We also present a sensitivity analysis of the snow and sea-ice in the SI3 model, exploring various options in the ISBA-ES scheme. Our findings reveal a strong sensitivity of both the snow and the sea-ice to the representation of liquid water in snow and the parameterization employed for calculating snowfall density.

How to cite: Brivoal, T., Guemas, V., Rousset, C., and Vancoppenolle, M.: Improving the representation of snow over sea-ice in the SI3 model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5177, https://doi.org/10.5194/egusphere-egu24-5177, 2024.

X4.3
|
EGU24-15156
Calibration of a  hybrid sea-ice model based on particle and continuums mechanics 
Carolin Mehlmann and Thomas Richter

Presently, climate models employ a continuum approach to describe sea ice. This approach assumes that statistical averages can be derived from a large number of ice floes. However, employing continuum rheological models at or below the scale of individual floes is only valid if the failure mode of a single floe aligns with that of an aggregate of floes. Initially, continuum models were designed for a grid resolution of 100 km. With recent advancements in computing power, sea-ice models are frequently operated at higher mesh resolutions, potentially leading to grid cells that no longer contains a representative sample of sea-ice floes.

We are addressing these shortcomings of current continuum sea-ice models by developing a hybrid model. The idea of the hybrid approach is to nest a particle model into a continuum sea-ice model in order to predict sea ice on fine spatial scales in a region of interest. An important component of particle models is a drag law to describe the influence of ocean and atmospheric currents on the floes. Measurements obtained onboard the Polarstern expedition PS 138 have shown that the correlation cannot be described fully locally, in regions with strongly heterogenous ice cover. Instead, larger surrounding flows have a substantial effect on the motion of small ones. Detailed numerical simulations of idealised test cases do confirm these findings.   

How to cite: Mehlmann, C. and Richter, T.: Calibration of a  hybrid sea-ice model based on particle and continuums mechanics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15156, https://doi.org/10.5194/egusphere-egu24-15156, 2024.

X4.4
|
EGU24-6441
|
ECS
Application of mixed least-squares FEM to study sea ice dynamics 
Sonja Hellebrand, Carina Schwarz, and Jörg Schröder

The behavior of sea ice has been studied for many decades. In order to model its viscous-plastic behavior at scales spanning several thousand kilometers, different numerical models have been proposed. Based on the established approach in [1], this contribution presents a simulation model for sea ice dynamics to describe the sea ice circulation and its evolution over one seasonal cycle. In course of that, the sea ice concentration and the sea ice thickness are considered, of which the physical behavior is governed by transient advection equations. Here, the sea ice velocity serves as coupling field.

Recently developed approaches base on a finite element implementation choosing a (mixed) Galerkin variational approach, see e.g. [2] and [3]. But therein, challenges may occur regarding the stability of the numerically complex scheme, especially when dealing with the first-order advection equations. Thus, we propose the application of the mixed least-squares finite element method, which has the advantage to be also applicable to first-order systems, i.e., it provides stable and robust formulations even for non-self-adjoint operators, such as the tracer equations (for sea ice thickness and sea ice concentration).

For solving the instationary sea ice equation the presented least-squares finite element formulation takes into account the balance of momentum and a constitutive law for the viscous-plastic flow. The considered primary fields are the stresses σ, the velocity v, the concentration Aice and the thickness Hice. In relation, four residuals are defined for the derivation of a first-order least-squares formulation based on the balance of momentum, the constitutive relation for the stresses, and two tracer-equations. Different approaches can be made with respect to the approximation functions of the primary fields, i.e., choosing e.g. conforming (H(div) interpolation functions) or non-conforming (Lagrangian interpolation functions) stress approximations, while Lagrangian interpolation functions are chosen for the remaining fields. In order to compare such approaches, the box test case is utilized, cf. [3], which is well described in literature.

References:

[1] W.D. Hibler III. A dynamic thermodynamic sea ice model. Journal of Physical Oceanography, 9(4):815-846, 1979.

[2] S. Danilov, Q. Wang, R. Timmermann, M. Iakovlev, D. Sidorenko, M. Kimmritz, T. Jung. Finite-Element Sea Ice Model (FESIM), Version 2. Geoscientific Model Development, 8:1747-1761, 2015.

[3] C. Mehlmann and T. Richter. A modified global Newton solver for viscous-plastic sea ice models. Ocean Modelling, 116:96-107, 2017.

How to cite: Hellebrand, S., Schwarz, C., and Schröder, J.: Application of mixed least-squares FEM to study sea ice dynamics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6441, https://doi.org/10.5194/egusphere-egu24-6441, 2024.

X4.5
|
EGU24-6569
|
ECS
Using discrete element methods to understand in-plane fragmentation of sea ice floes 
Adam Bateson, Daniel Feltham, David Schröder, Scott Durski, Jennifer Hutchings, Rajlaxmi Basu, and Byongjun Hwang

Sea ice floe size can impact several processes that determine the evolution of the Arctic sea ice, including lateral melt volume, momentum exchange, and rheology. Floe size distribution (FSD) models are applied within continuum sea ice models to capture the evolution of the FSD through parameterisations of the processes that modify floe size such as lateral melting and wave break-up of floes. FSD models do not yet adequately resolve in-plane fragmentation processes of floes such as the breakup of floes under wind forcing, through interactions between neighbouring floes, or through thermal weakening. It is challenging to characterise and therefore parameterise these in-plane floe breakup processes due to limited availability of in-situ observations. Discrete element models (DEMs) offer an alternative way to understand the different mechanisms of floe fragmentation. By resolving relevant properties such as shear and normal stress and sea ice strength at the sub-floe scale, it is possible to use DEMs as a virtual laboratory and directly simulate the break-up of floes into smaller fragments.

In this study, we describe how in-situ observations of sea ice can be combined with output from sea ice DEMs to develop parameterisations of in-plane breakup of floes that can then be applied in continuum models. We then discuss the necessary model developments in order to apply a sea ice DEM to floe fragmentation at smaller scales. We will also present results from a series of DEM simulations used to model the fracture of sea ice under different forcing conditions and with varying sea ice states to identify the important sea ice parameters and processes in determining the size of the floes that form from in-plane breakup events.

How to cite: Bateson, A., Feltham, D., Schröder, D., Durski, S., Hutchings, J., Basu, R., and Hwang, B.: Using discrete element methods to understand in-plane fragmentation of sea ice floes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6569, https://doi.org/10.5194/egusphere-egu24-6569, 2024.

X4.6
|
EGU24-4502
|
ECS
Understanding influence of ocean waves on Arctic sea ice simulation: A modeling study with an atmosphere-ocean-wave-sea ice coupled model 
Chao-Yuan Yang and Jiping Liu

Rapid decline of Arctic sea ice has created more open water for ocean wave development and highlighted the importance of wave-ice interactions in the Arctic. Some studies have made contributions to our understanding of the potential role of the prognostic floe size distribution (FSD) on sea ice changes. However, these efforts do not capture the full interactions between atmosphere, ocean, wave, and sea-ice. In this study, a modified joint floe size and thickness distribution (FSTD) is implemented in a newly-developed regional atmosphere-ocean-wave-sea ice coupled model and a series of pan-Arctic simulation is conducted with different physical configurations related to FSD changes, including FSD-fixed, FSD-varied, lateral melting rate, wave-fracturing formulation, and wave attenuation rate. Firstly, atmosphere-ocean-wave-sea ice coupled simulations show that the prognostic FSD leads to reduced ice area due to enhanced ice-ocean heat fluxes, but the feedbacks from the atmosphere and the ocean partially offset the reduced ice area induced by the prognostic FSD. Secondly, lateral melting rate formulations do not change the simulated FSD significantly, but they influence the flux exchanges across atmosphere, ocean, and sea-ice and thus sea ice responses. Thirdly, the changes of FSD are sensitive to the simulated wave parameters associated with different wave-fracturing formulations and wave attenuation rates, and the limited oceanic energy imposes a strong constraint on the response of sea ice to FSD changes. Finally, the results also show that wave-related physical processes can have impacts on sea ice changes with the constant FSD, indicating the indirect influences of ocean waves on sea-ice through the atmosphere and the ocean.

How to cite: Yang, C.-Y. and Liu, J.: Understanding influence of ocean waves on Arctic sea ice simulation: A modeling study with an atmosphere-ocean-wave-sea ice coupled model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4502, https://doi.org/10.5194/egusphere-egu24-4502, 2024.

X4.7
|
EGU24-12804
 A laboratory model of fragmentation of a 2D membrane by waves. Analogies and differences with sea ice. 
Michael Berhanu, Louis Saddier, Mathéo Aksil, Palotai Ambre, and Michel Tsamados

The marginal ice zone is the transition region between the dense floating ice pack and the open ocean. In this zone, the interaction of surface waves with sea ice is highly complex. The sea ice is broken up into fragments, the floes, which can split into smaller parts and drift under the action of waves and underwater current. Although the downscaling is challenging, laboratory model experiments can contribute to a better understanding of this process coupling fluid and solid mechanics on a large range of time and space scales. We propose to study the fragmentation of a floating membrane, made up of 10 µm graphite particles arranged in a monolayer, by gravity surface waves with a wavelength of around 15 cm [1]. For a sufficiently strong wave amplitude, the raft progressively breaks up, developing cracks and producing fragments whose sizes decrease over a time scale that is long relative to the wave period. We then study the distribution of the fragments produced during the fragmentation process. The visual appearance of the size-distributed fragments surrounded by open water bears a striking resemblance to the floes produced by the fracturing of sea ice by waves. The fragmentation concepts and morphological tools developed for sea ice floes can be applied to our macroscopic analog. Although the mechanics of the two systems differ in their physical properties and in the fracture process, our experiment provides a model laboratory system for studying the fragmentation of floating 2D materials

 

[1] Saddier, L., Palotai, A., Aksil, M., Tsamados, M., & Berhanu, M. (2023). Breaking of a floating particle raft by water waves. In arXiv preprint arXiv:2310.16188.

How to cite: Berhanu, M., Saddier, L., Aksil, M., Ambre, P., and Tsamados, M.:  A laboratory model of fragmentation of a 2D membrane by waves. Analogies and differences with sea ice., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12804, https://doi.org/10.5194/egusphere-egu24-12804, 2024.

X4.8
|
EGU24-1671
|
ECS
Monthly Arctic sea ice prediction based on a data-driven deep learning model  
Xiaohe Huan, Jielong Wang, and Zhongfang Liu

There is growing interest in sub-seasonal to seasonal predictions of Arctic sea ice due to its potential effects on midlatitude weather and climate extremes. Current prediction systems are largely dependent on physics-based climate models. While climate models can provide good forecasts for Arctic sea ice at different timescales, they are susceptible to initial states and high computational costs. Here we present a purely data-driven deep learning model, UNet-F/M, to predict monthly sea ice concentration (SIC) one month ahead. We train the model using monthly satellite-observed SIC for the melting and freezing seasons, respectively. Results show that UNet-F/M has a good predictive skill of Arctic SIC at monthly time scales, generally outperforming several recently proposed deep learning models, particularly for September sea-ice minimum. Our study offers a perspective on sub-seasonal prediction of future Arctic sea ice and may have implications for forecasting weather and climate in northern midlatitudes.

How to cite: Huan, X., Wang, J., and Liu, Z.: Monthly Arctic sea ice prediction based on a data-driven deep learning model , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1671, https://doi.org/10.5194/egusphere-egu24-1671, 2024.

X4.9
|
EGU24-15487
Extended seasonal forecast of Antarctic Sea Ice using ANTSIC-UNet 
Ziying Yang, Jiping Liu, and Rune Grand Graversen

Antarctic sea-ice variability affects the ocean and atmosphere both locally through thermodynamic processes and beyond the Antarctic regions remotely through dynamic processes, which may all change due to global warming. In this study, we develop the ANTSIC-UNet, a deep-learning model trained on physically enriched climate variables, to predict the extended seasonal Antarctic sea ice concentration of up to 6 months in advance. We assess the predictive skill of ANTSIC-UNet as regards linear trend prediction and anomaly persistence prediction in the Pan- and regional Antarctic areas using comparative analyses with two baseline models. Our results exhibit superior performance of ANTSIC-UNet for the extended seasonal Antarctic forecast. The predictive skill of ANTSIC-UNet is notably season-dependent, showing distinct variations across regions. Optimal prediction accuracy is found in winter, while diminished skill found during the summer can be largely attributed to the ice-edge error. High predictive skills are found in the Weddell Sea throughout the year, which suggests that regional Antarctic sea-ice predictions beyond 6 months are possible. We further quantify variable importance through a post-hoc interpretation method which indicates that ANTSIC-UNet has learned the relationships between SIC and other climate variables and the method therefore provides information on the physics of the model. At short lead times, on timescales of up to two months, ANTSIC-UNet predictions exhibit heightened sensitivity to sea surface temperature, radiation conditions and vertical atmospheric circulation conditions in addition to the sea-ice itself. At longer lead times, predictions are dependent on stratospheric circulation patterns at 7-8 months lead in addition to sea-ice. Furthermore, we discuss the potential of implementing physical constraints to enhance sea-ice-edge predictability.

How to cite: Yang, Z., Liu, J., and Grand Graversen, R.: Extended seasonal forecast of Antarctic Sea Ice using ANTSIC-UNet, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15487, https://doi.org/10.5194/egusphere-egu24-15487, 2024.

X4.10
|
EGU24-2377
|
ECS
Multivariate state and parameter estimation using data assimilation in a Maxwell-Elasto-Brittle sea ice model 
Yumeng Chen, Polly Smith, Alberto Carrassi, Ivo Pasmans, Laurent Bertino, Marc Bocquet, Tobias Sebastian Finn, Pierre Rampal, and Véronique Dansereau

In an idealised setup, a dynamics-only sea ice model is used to investigate the fully multivariate state and parameter estimations that uses a novel Maxwell-Elasto-Brittle (MEB) sea ice rheology. In the fully multivariate state estimation, the level of damage, internal stress and cohesion are estimated along with the observed sea ice concentration, thickness and velocity. In the case of parameter estimation, we estimate the air drag coefficient and the damage parameter of the MEB model. The air drag coefficients adjust the strength of the forcing on the sea ice dynamics while the damage parameter controls the mechanical behaviour of the internal property of sea ice. We show that, with the current observation network, it is possible to improve all model state forecast and the parameter accuracy using data assimilation approaches even though problems could arise in such an idealised setup where the external forcing dominates the model forecast error growth.

How to cite: Chen, Y., Smith, P., Carrassi, A., Pasmans, I., Bertino, L., Bocquet, M., Finn, T. S., Rampal, P., and Dansereau, V.: Multivariate state and parameter estimation using data assimilation in a Maxwell-Elasto-Brittle sea ice model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2377, https://doi.org/10.5194/egusphere-egu24-2377, 2024.

X4.11
|
EGU24-10098
Best of SIDFEx: Highlights and lessons learned from six years of sea-ice drift forecasting 
Simon F. Reifenberg, Valentin Ludwig, and Helge F. Goessling and the SIDFEx Team

We showcase the Sea Ice Drift Forecast Experiment (SIDFEx) database. SIDFEx is a collection of close to 225,000 lagrangian drift forecasts for the trajectories of assets (mostly buoys) on the Arctic and Antarctic sea ice, at lead times from daily to seasonal with mostly daily resolution. The forecasts are based on systems with varying degrees of complexity, ranging from free-drift forecasts to forecasts by fully coupled dynamical general circulation models. Combining several independent forecasts allows us to construct a best-guess consensus forecast, with a seamless transition from systems with lead times of up to 10 days to systems with seasonal lead times. The forecasts are generated by 13 research groups using 23 distinct forecasting systems and sent regularly to the Alfred-Wegener-Institute, where they are archived and evaluated. Many groups send forecasts operationally in near-real time.

In our presentation, we will introduce the motivation behind and setup of SIDFEx, as well as an overview on the general forecast skill. We will focus on selected highlights, comprising the operational support of research cruises, short-term predictions of sea-ice deformation and regular contributions to the Sea Ice Outlook competition.

How to cite: Reifenberg, S. F., Ludwig, V., and Goessling, H. F. and the SIDFEx Team: Best of SIDFEx: Highlights and lessons learned from six years of sea-ice drift forecasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10098, https://doi.org/10.5194/egusphere-egu24-10098, 2024.

X4.12
|
EGU24-12451
|
ECS
Development of ship navigation risk indicator in sea ice-infested water 
Xinfang Zhang

There's increasing transpolar shipping in both the Arctic and Antarctic as a result of the reduction of sea ice and the desire from social economics.  Sea ice is a hazard for shipping in ice-infested water, Ship navigability in ice-covered sea depends on sea ice concentration, ice thickness, fraction of pressure ridges, and multi-year ice as well as ice speed and compression, it also depends on the vessel ice class. IMO introduced Risk Index Outcome(RIO) to provide guidelines for safe navigation, calculation of RIO requires accurate sea ice information including sea ice concentration and thickness. We developed a method similar to RIO to calculate navigation risk indicators using forecasting models including ECMWF S2S data, Copernicus data, and DMI data. Other than conventional sea ice parameters sea ice concentration and sea ice thickness, ice salinity, and ice age are also taken into account in risk indicator calculation. We select the time March 2019 -Oct 2020 and adopt the initial condition of the model forecast for sea ice to demonstrate the capabilities of seasonal forecasting of this navigation risk indicator in different models. In future, the calculation method will be implemented within the ClimateDT environment.

How to cite: Zhang, X.: Development of ship navigation risk indicator in sea ice-infested water, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12451, https://doi.org/10.5194/egusphere-egu24-12451, 2024.

X4.13
|
EGU24-9554
|
ECS
Historical volcanic eruptions slowed down rapid decline in Arctic sea ice linked to global warming 
Xinyue Wang

Arctic sea ice has experienced a differential decline in speed due to the same anthropogenic greenhouse gas forcing, as evidenced by rapid decline after the end of the last century. Our convergent observations, last-millennium reanalysis, and model analyses have revealed that large tropical volcanic eruptions can lead to a decadal increase in Arctic sea ice, and the 1982 and 1991 large volcanic eruptions slowed down the decline of Arctic sea ice during the last century. The models, selected based on the observed sensitivity of Arctic sea ice to volcanic eruptions, suggest that the earliest ice-free summer year in the Arctic will be around 2040 in high-emission sceneria of SSP585. These findings emphasized the crucial need to incorporate volcanic influences when projecting future Arctic changes amid global warming.

How to cite: Wang, X.: Historical volcanic eruptions slowed down rapid decline in Arctic sea ice linked to global warming, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9554, https://doi.org/10.5194/egusphere-egu24-9554, 2024.

Posters virtual: Thu, 18 Apr, 14:00–15:45 | vHall X4

Display time: Thu, 18 Apr, 08:30–18:00
Chairpersons: Lorenzo Zampieri, Clara Burgard
vX4.1
|
EGU24-9802
|
ECS
The sea ice component of MUSE, the unstructured-mesh global ocean model of CMCC 
Francesco Cocetta, Lorenzo Zampieri, and Doroteaciro Iovino

The rapidly evolving sea ice cover requires novel modeling approaches and mathematical techniques to accurately simulate the sea ice dynamics, thermodynamics, and its interactions with the atmosphere and ocean at varying spatiotemporal resolutions. In this context, the CMCC is developing the Multiscale Unstructured model for Simulating the Earth’s water environment (MUSE), a novel global ocean-sea ice model on unstructured meshes.

MUSE employs a finite-element numerical discretization on unstructured meshes, aiming at offering flexibility in simulating the global ocean for various applications, ranging from physical process understanding to operational sea ice predictions. The ongoing implementation of the sea ice component utilizes the traditional continuous sea ice formulation and the 2+1 split assumption, meaning that the sea ice dynamics and advection are solved for horizontal motions while the thermodynamics and radiative processes are parameterized at the subgrid scale.   

MUSE employs a modified elastic-viscous-plastic (mEVP) solver for the sea ice dynamics and a Flux Corrected Transport (FCT) advection scheme, alongside the state-of-the-art column physics package "Icepack" maintained by the CICE consortium.

Here, we describe the global implementation of the sea ice component in MUSE and its coupling with the ocean. We present the resulting representation of vertical thermodynamic processes and horizontal dynamics of sea ice.

How to cite: Cocetta, F., Zampieri, L., and Iovino, D.: The sea ice component of MUSE, the unstructured-mesh global ocean model of CMCC, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9802, https://doi.org/10.5194/egusphere-egu24-9802, 2024.

vX4.2
|
EGU24-11413
|
Sea ice strength in SI3 
Emma Fiedler, Ed Blockley, Clement Rousset, and Martin Vancoppenolle

The NEMO sea ice model, SI3, includes the simple formulation of Hibler (1979; H79) to parameterise the compressive strength of sea ice. This assumes that thick and compact sea ice has more strength than thin and low concentration sea ice. However, the H79 strength scheme does not consider physical assumptions around energy conservation. The strength scheme of Rothrock (1975; R75) is based on the amount of potential energy gained and frictional energy dissipated during ridging, and has been introduced to SI3. Additionally, the option for a negative exponential redistribution of ridged ice among thickness categories, to better approximate observations and improve stability compared to the existing uniform redistribution when using R75, has been included. The R75 strength formulation is stable and works well in SI3 at version 4.2 with an EVP rheology, under a Met Office forced NEMO/SI3 model configuration. Sea ice strength is generally reduced for the R75 scheme compared to H79. The most notable effect on the model output is a greater number of, and sharper, features in the resulting modelled ice field when using the R75 scheme compared to the H79 scheme, which are particularly apparent in the ice thickness field. An increase in the model effective resolution is therefore demonstrated.

How to cite: Fiedler, E., Blockley, E., Rousset, C., and Vancoppenolle, M.: Sea ice strength in SI3, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11413, https://doi.org/10.5194/egusphere-egu24-11413, 2024.