Search

We developed a new version of the Alfred Wegener Institute Climate Model (AWI-CM3), which has higher skills in representing the observed climatology and better computational efficiency than its predecessors. Its ocean component FESOM2 (Finite-volumE Sea ice–Ocean Model) has the multi-resolution functionality typical of unstructured-mesh models while still featuring a scalability and efficiency similar to regular-grid models. The atmospheric component OpenIFS (CY43R3) enables the use of the latest developments in the numerical-weather-prediction community in climate sciences. In this paper we describe the coupling of the model components and evaluate the model performance on a variable-resolution (25–125 km) ocean mesh and a 61 km atmosphere grid, which serves as a reference and starting point for other ongoing research activities with AWI-CM3. This includes the exploration of high and variable resolution and the development of a full Earth system model as well as the creation of a new sea ice prediction system. At this early development stage and with the given coarse to medium resolutions, the model already features above-CMIP6-average skills (where CMIP6 denotes Coupled Model Intercomparison Project phase 6) in representing the climatology and competitive model throughput. Finally we identify remaining biases and suggest further improvements to be made to the model.

Presentation on sea-ice drift forecasts from the SIDFEx database, given at the IGS symposium held in Bremerhaven from June 4th to June 11th, 2023.

We introduce the Sea Ice Drift Forecast Experiment (SIDFEx) database. SIDFEx is a collection of close to 180,000 lagrangian drift forecasts for the trajectories of specified assets (mostly buoys) on the Arctic and Antarctic sea ice, at lead times from daily to seasonal scale and 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 operationally to the Alfred-Wegener-Institute, where they are archived and evaluated. Many systems send forecasts in near-real time.One key purpose when starting SIDFEx in 2017 was to find the optimal starting position for the Multidisciplinary Drifting Observatory for the Study of Arctic Climate (MOSAiC). Over the years, more applications evolved: During MOSAiC, the SIDFEx forecasts were used for ordering high-resolution TerraSAR-X images in advance, with a hit rate of 80%. During the Endurance22 expedition, we supported the onboard team with near-real time forecasts, contributing to the success of the mission. Currently, we evaluate drift forecasts for several buoys of the MOSAiC Distributed Network (DN). We know that there is skill in predicting the location of single buoys. Now, we extend this to studying the deformation of the polygon spanned by the DN buoys. Deformation is derived from the spatial velocity derivatives of the buoy array. We find low correlation coefficients between the deformation in the models and the observed deformation for a small-scale DN configuration, but larger and significant correlations around 0.7 for larger configurations and an Arctic-wide buoy array.

Marine heatwaves are on the rise: their frequency, intensity, and duration are expected to increase in a warming world. Yet it remains unclear whether local feedback processes could amplify extreme ocean temperatures. A prominent marine heatwave recently occurred in the Northeast Pacific Ocean in summer 2019, exhibiting the highest sea surface temperatures ever recorded in this area since the availability of satellite observations in 1979. Here, we use fully-coupled model experiments, termed “nudged storylines”, in which the evolution of large-scale winds in the free troposphere is nudged to the observed (reanalysed) one before and during the summer 2019 event, to generate close analogues of this record-breaking marine heatwave for past, present, and plausible future climates. We show in particular that future climate analogues of the marine heatwave may warm 50% more than what is expected from the projected global-mean ocean warming. Together with the rapid Northeast Pacific mean warming, air-sea feedback processes lead to a projected warming amplification of 1°C above the 1.9°C global-mean ocean temperature rise. Primary drivers of this amplification are a reduction in clouds and ocean mixed-layer depth, as well as anomalous air advection from fast-warming subpolar regions. Our results show that marine heatwave temperatures may warm substantially faster than the global and regional background temperature, increasing the stress on local ecosystems and fishery resources.

Siberia experienced an exceptionally warm first half of 2020, with temperatures peaking on 20th June, when the weather station at Verkhoyansk registered 38ºC – the highest-ever temperature recorded north of the Arctic Circle. This event led to a state of emergency declaration due to a wide range of natural and human disasters, including permafrost collapse or wildfires. Several previous attribution studies have found anomalous synoptic atmospheric circulation and human-caused climate change as major drivers of this event. However, it remains unclear how such an event would unfold in warmer climates.In this work, we have constructed analogues of this extreme event in past and future climates, so-called “storylines”, employing spectral nudging experiments using the coupled climate model AWI-CM1. In these simulations, large-scale free-troposphere winds are constrained toward ERA5 data, and the model is run for different boundary and initial conditions (pre-industrial, present, +2K, +3K, +4K climates). By doing so, this approach focuses on the less uncertain thermodynamic influence of climate on extreme events, constraining the much more uncertain dynamical changes.Our present-climate simulations realistically reproduce the observed Siberian heatwave and the exceptionally warm conditions before the event. When the different background climates are considered, on the one hand, a robust global warming amplification is obtained until spring. On the other hand, a future global warming dampening or even local cooling is observed in phase with the heatwave peak. We examine the mechanisms that damp warming during the heatwave analogue, especially changes in soil moisture and radiative fluxes.

Climate is defined by the statistics of the weather. We are thus used to the notion that weather depends on climate in the sense that a weather state is a quasi random realization drawn from the climatological distribution of weather states. Consequently, weather obviously also depends on climate change. Here it is argued that it can be very instructive to consider this dependence the other way around and to investigate how climate change depends on the weather. More specifically, the question is how different parts of the climatological distribution change, depending on certain relevant characteristics of the weather. For example, one may ask how the climate at some time of the year and at some location changes between a pre-industrial and a globally +4K warmer climate, considering only days where the local winds at some height blow from a specific direction. Alternatively, one may investigate how the climate change pattern differs between certain large-scale atmospheric circulation regimes, such as NAO+ and NAO- situations. While such conditional climate change analyses can be based for example on reanalysis or CMIP-type climate model data, a more extreme variant are storyline simulations where the evolution of the large-scale circulation is imposed in a climate model using different climate backgrounds, allowing to assess climate change conditional on a specific evolution of the large-scale circulation. Storyline simulations inevitably ignore possible changes in the likelihood of circulation patters. In contrast, analyses based on sufficiently large samples of reanalysis or CMIP-type data also allow for quantifying changes in likelihoods and constitute a proper decomposition of the complete (unconditional) climate change signal. Here the concept of weather-dependent climate change is explained and its potential to help unravel the complexities of climate change is demonstrated.

Extreme weather events are triggered by atmospheric circulation patterns and shaped by slower components, including soil moisture and sea surface temperature, and by the background climate. This separation of factors is exploited by the storyline approach in which an atmospheric model is nudged toward the observed dynamics using different climate boundary conditions to explore their influence. The storyline approach disregards uncertain climatic changes in the frequency and intensity of dynamical conditions, focusing instead on the thermodynamic influence of climate on extreme events. Here we demonstrate an advanced storyline approach that employs a coupled climate model (AWI-CM-1-1-MR) in which the large-scale free-troposphere dynamics are nudged toward ERA5 data. Five-member ensembles are run for present-day (2017–19), preindustrial, +2-K, and +4-K climates branching off from CMIP6 historical and scenario simulations of the same model. In contrast to previous studies, which employed atmosphere-only models, feedbacks between extreme events and the ocean and sea ice state, and the dependence of such feedbacks on the climate, are consistently simulated. Our setup is capable of reproducing observed anomalies of relevant unconstrained parameters, including near-surface temperature, cloud cover, soil moisture, sea surface temperature, and sea ice concentration. Focusing on the July 2019 European heat wave, we find that the strongest warming amplification expands from southern to central Europe over the course of the twenty-first century. The warming reaches up to 10 K in the 4-K-warmer climate, suggesting that an analogous event would entail peak temperatures around 50°C in central Europe. Significance Statement This work explores a new storyline method to determine the impact of climate change on specific recent extreme events. The observed evolution of the large-scale atmospheric circulation is imposed in a coupled climate model. Variations in climate parameters, including ocean temperatures and sea ice, are well reproduced. By varying the background climate, including CO2 concentrations, it is demonstrated how the July 2019 European heat wave could have evolved in preindustrial times and in warmer climates. For example, up to 10°C warmer peak temperatures could occur in central Europe in a 4°C warmer climate. The method should be explored for other types of extreme events and has the potential to make climate change more tangible and to inform adaptation measures.

An ensemble-based data assimilation framework for a coupled ocean–atmosphere model is applied to investigate the influence of assimilating different types of ocean observations on the ocean and atmosphere simulation. The data assimilation is performed with the parallel data assimilation framework (PDAF) for the climate model AWI-CM. Observations of the ocean, namely satellite sea-surface temperature (SST) and temperature and salinity profiles, are assimilated into the ocean component. The atmospheric state is only influenced by the model dynamics. Different assimilation scenarios were carried out with different combinations of observations to investigate to what extent the assimilation into the coupled model leads to a better estimation of the state of the ocean as well as the atmosphere. The influence of the data assimilation is assessed by comparing the ocean prediction with dependent and independent ocean observations. For the atmosphere, the assimilation result is compared with the ERA-Interim atmospheric reanalysis data. The ocean temperature and salinity are improved by all the assimilation scenarios in the coupled system. The assimilation leads to a response of the atmosphere throughout the troposphere and impacts the global atmospheric circulation. Globally the temperature and wind speed are improved in the atmosphere on average.

Sea ice models have become essential components of weather, climate, and ocean models. A realistic representation of sea ice affects the reliability of process representation, environmental forecast, and climate projections. Realistic simulations of sea ice kinematics require the consideration of both large-scale and finescale geomorphological structures such as linear kinematic features (LKF). We propose a multiscale directional analysis (MDA) that diagnoses the spatial characteristics of LKFs. The MDA is different from previous analyses in that it (i) does not detect LKFs as objects, (ii) takes into account the width of LKFs, and (iii) estimates scale-dependent orientation and intersection angles. The MDA is applied to pairs of deformation fields derived from satellite remote sensing data and from a numerical model simulation with a horizontal grid spacing of ~4.5 km. The orientation and intersection angles of LKFs agree with the observations and confirm the visual impression that the intersection angles tend to be smaller in the satellite data compared to the model data. The MDA distributions can be used to compare satellite data and numerical model fields using conventional metrics such as a Euclidean distance, the Bhattacharyya coefficient, or the Earth mover’s distance. The latter is found to be the most meaningful metric to compare distributions of LKF orientations and intersection angles. The MDA proposed here provides a tool to diagnose if modified sea ice rheologies lead to more realistic simulations of LKFs.

The Sea Ice Drift Forecast Experiment (SIDFEx) database comprises more than 180,000 forecasts for trajectories of single sea-ice buoys in the Arctic and Antarctic, collected since 2017. SIDFEx is a community effort originating from the Year Of Polar Prediction. Forecasts are provided by various forecast centres and collected, and archived by the Alfred Wegener Institute (AWI). AWI provides a dedicated software package and an interactive online platform for analysing the forecasts. Their lead times range from daily to seasonal scales. Among the buoys targeted by SIDFEx are the buoys of the Distributed Network (DN) array which was deployed during the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition. In this contribution, we show to what extent the deformation (divergence, shear and vorticity) of the DN can be forecasted by the SIDFEx forecasts. We investigate the performance of single models as well as a consensus forecast which merges the single forecasts to a seamless best-guess forecast.

Under the ongoing climate change, extreme weather events are becoming more prolonged, intense, and frequent; and this trend is expected to continue in a future warmer climate. Several studies have found that the synoptic atmospheric circulation at the time of the event is the main contributing factor in most cases. Moreover, they are shaped by slower processes, including sea-surface temperature and soil moisture, in turn influenced by the history of preceding weather patterns, and by the background climate. The separation of influencing components is exploited by the storyline approach, where an atmosphere model is nudged toward the observed dynamics using different climate boundary conditions. Thus, the storyline approach focuses on the less uncertain thermodynamic influence of climate on extreme events, disregarding the somewhat controversial dynamical changes. This approach provides a very efficient way of making the impacts of climate change more tangible to experts and non-experts alike as events fresh in the people's memory are reproduced in different plausible climates with just moderate computational resources.Spectral nudging experiments have been run with two coupled climate models, AWI-CM-1 and AWI-CM-3. In these simulations, the large-scale free-troposphere dynamics are constrained toward ERA5 data and the model is run for different boundary conditions. Here, the ocean and sea-ice state are consistently simulated, unlike previous studies which employed atmosphere-only models. Our setups reasonably reproduce daily to seasonal observed anomalies of relevant unconstrained parameters, including near-surface temperature, soil moisture or cloud cover. In particular, our configurations showed satisfactory skills in reproducing two different extreme events: the July 2019 European heat wave, and the July 2021 European extreme rainfall. Therefore, this methodology has been applied to study several extreme events in different climates. To do so, nudged simulations are branched off CMIP6 historical and scenario simulations of the same model. For the particular July 2021 extreme rainfall event, we have run five ensemble members for AWI-CM-1-1-MR for dynamical conditions from 1st January 2017 to 31st July 2021 in pre-industrial, present-day, +2K, and +4K climates. These simulations are complemented with similar experiments for AWI-CM-3. The most outstanding finding of these studies is a global warming amplification associated with some events, which exacerbates their exceptionality, especially in a high emission scenario.

The Arctic sea ice cover and thickness have significantly declined since the 1970s, while exhibiting large interannual variability. Snow cover on sea ice, acting as an insulating barrier, was shown to be instrumental in driving the variability and trends in sea-ice thickness. Yet, the Arctic snow depth remains scarcely measured and overlooked in climate models, which translates to “very limited predictive skill” according to the IPCC (Special Report on the Ocean and Cryosphere in a Changing Climate). Moreover, sea-ice thickness initialization has been shown to be an important element for skilful sea-ice forecasts, and it appears plausible that the same holds for the snow layer on top.Here, we investigate the role of atmospheric circulation anomalies in shaping the Arctic snow-cover and sea-ice thickness anomalies. In this preparatory work, spectral nudging of the large-scale atmospheric circulation towards ERA5 reanalysis data is applied to the fully coupled AWI Climate Model (AWI-CM-3). We examine the variability and trends of Arctic snowfall, snow depth, sea ice cover and thickness over a 42-year period (1979-2021), and in particular the reproduction of observed anomalies. Two nudging configurations are used, differing in strength by their relaxation timescale τ and spectral truncation wavenumber T (namely τ=24 h, T20 and τ=1 h, T159). We demonstrate the importance of atmospheric circulation anomalies in shaping variations of snow and ice thickness at sub-seasonal to interannual scales, and discuss the potential of spectral nudging as a tool to improve the initialization of sea ice forecasts.

We compare strongly coupled data assimilation (SCDA) and weakly coupled data assimilation (WCDA) by analyzing the assimilation effect on the estimation of the ocean and the atmosphere variables. The AWI climate model (AWI-CM-1.1) is coupled with the parallel data assimilation framework (PDAF). Only satellite sea surface temperature data are assimilated. For WCDA, only the ocean variables are directly updated by the assimilation. For SCDA, both the ocean and the atmosphere variables are directly updated by the assimilation. Both WCDA and SCDA improve ocean state and yield similar errors. In the atmosphere, WCDA gives slightly smaller errors for the near-surface temperature and wind velocity than SCDA. In the free atmosphere, SCDA yields smaller errors for the temperature, wind velocity, and specific humidity than WCDA in the Arctic region, while in the tropical region, the errors are generally larger.

Accelerated loss of the sea-ice cover and increased human activities in the Arctic emphasize the need for skillful prediction of sea-ice conditions at subseasonal to seasonal (S2S) timescales. To assess the quality of predictions, dynamical forecast systems can be benchmarked against reference forecasts based on present and past observations of the ice edge. However, the simplest types of reference forecasts—persistence of the present state and climatology—do not exploit the observations optimally and thus lead to an overestimation of forecast skill. For spatial objects such as the ice-edge location, the development of damped-persistence forecasts that combine persistence and climatology in a meaningful way poses a challenge. We have developed a probabilistic reference forecast method that combines the climatologically derived probability of ice presence with initial anomalies of the ice-edge location, both derived from satellite sea-ice concentration data. No other observations, such as sea-surface temperature or sea-ice thickness, are used. We have tested and optimized the method based on minimization of the Spatial Probability Score. The resulting Spatial Damped Anomaly Persistence forecasts clearly outperform both simple persistence and climatology at subseasonal timescales. The benchmark is about as skillful as the best-performing dynamical forecast system in the S2S database. Despite using only sea-ice concentration observations, the method provides a challenging benchmark to assess the added value of dynamical forecast systems.

Using the depth (z) and density (ϱ) frameworks, we analyze local contributions to AMOC variability in a 900‐year simulation with the AWI climate model. Both frameworks reveal a consistent interdecadal variability; however, the correlation between their maxima deteriorates on year‐to‐year scales. We demonstrate the utility of analyzing the spatial patterns of sinking and diapycnal transformations through depth levels and isopycnals. The success of this analysis relies on the spatial binning of these maps which is especially crucial for the maps of vertical velocities which appear to be too noisy in the main regions of upwelling and downwelling because of stepwise bottom topography. Furthermore, we show that the AMOC responds to fast (annual or faster) fluctuations in atmospheric forcing associated with the NAO. This response is more obvious in the ϱ than in the z framework. In contrast, the link between AMOC and deep water production south of Greenland is found for slower fluctuations and is consistent between the frameworks.

Extreme weather events are triggered by atmospheric circulation patterns and shaped by slower components, including soil moisture and sea-surface temperature, and by the background climate. This se...

The transient climate response (TCR) is 20% higher in the Alfred Wegener Institute Climate Model (AWI‐CM) compared to the Max Planck Institute Earth System Model (MPI‐ESM) whereas the equilibrium climate sensitivity (ECS) is by up to 10% higher in AWI‐CM. These results are largely independent of the two considered model resolutions for each model. The two coupled CMIP6 models share the same atmosphere‐land component ECHAM6.3 developed at the Max Planck Institute for Meteorology (MPI‐M). However, ECHAM6.3 is coupled to two different ocean models, namely the MPIOM sea ice‐ocean model developed at MPI‐M and the FESOM sea ice‐ocean model developed at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI). A reason for the different TCR is related to ocean heat uptake in response to greenhouse gas forcing. Specifically, AWI‐CM simulations show stronger surface heating than MPI‐ESM simulations while the latter accumulate more heat in the deeper ocean. The vertically integrated ocean heat content is increasing slower in AWI‐CM model configurations compared to MPI‐ESM model configurations in the high latitudes. Weaker vertical mixing in AWI‐CM model configurations compared to MPI‐ESM model configurations seems to be key for these differences. The strongest difference in vertical ocean mixing occurs inside the Weddell and Ross Gyres and the northern North Atlantic. Over the North Atlantic, these differences materialize in a lack of a warming hole in AWI‐CM model configurations and the presence of a warming hole in MPI‐ESM model configurations. All these differences occur largely independent of the considered model resolutions., Plain Language Summary: The transient climate response (TCR) describes how strongly near‐surface temperatures warm in response to gradually increasing greenhouse‐gas levels. Here we investigate the role of the ocean which takes up heat and thereby delays the surface warming. Two models of the Coupled Model Intercomparison Project Phase 6 (CMIP6), the Alfred Wegener Institute Climate Model (AWI‐CM) and the Max Planck Institute Earth System Model (MPI‐ESM), which use the same atmosphere model but different ocean models are selected for this study. In AWI‐CM the upper ocean layers heat faster than in MPI‐ESM, while the opposite is true for the deep ocean. As a consequence, the TCR is 20% stronger in AWI‐CM compared to MPI‐ESM. We find that weaker vertical ocean mixing in AWI‐CM compared to MPI‐ESM, especially over the northern North Atlantic and the Weddell and Ross Gyres, is key for these differences. Our findings corroborate the importance of realistic ocean mixing in climate models when it comes to getting the strength and timing of climate change right., Key Points: The transient climate response in two coupled models with the same atmosphere but different ocean components differs by 20%. The upper (deeper) ocean heats faster (slower) in AWI‐CM compared to MPI‐ESM, independent of model resolution. Vertical mixing in the northern North Atlantic and the Weddell and Ross Gyres appears to be key for these differences., Bundesministerium für Bildung und Forschung (BMBF) http://dx.doi.org/10.13039/501100002347, German Climate Computing Centre (DKRZ), Federal Ministry of Education and Research of Germany, Helmholtz Association http://dx.doi.org/10.13039/501100009318, https://esgf-data.dkrz.de/projects/cmip6-dkrz/

Under the current global warming trend, heatwaves are becoming more intense, frequent, and longer-lasting; and this trend will continue in the future. In this context, the recent 2019 summer was exceptionally hot in large areas of the Northern Hemisphere, with embedded heatwaves, as for example the June and July 2019 European events, redrawing the temperature record map in western Europe. Large-scale dynamics (associated with blockings or subtropical ridges) play a key role in explaining these-large scale events.Conceptually, global warming can be split into two different contributions: Dynamic and thermodynamic changes. Whereas dynamic changes remain highly uncertain, some thermodynamic changes can be quantified with higher confidence. We exploit this concept by studying how these recent European heatwaves would have developed in a pre-industrial climate and how it would develop in the future for 1.5, 2 and 4 ºC warmer climates (storyline scenarios). To do so, we employ the spectral nudging technique with AWI-CM (CMIP6 model, a combination of ECHAM6 AGCM + FESOM Sea Ice-Ocean Model). Large-scale dynamics are prescribed by reanalysis data (ERA5). Meanwhile, the model is run for different boundary conditions corresponding to preindustrial and future climates along the SSP370 forcing scenario. This approach can be useful to help understand and communicate what climate change will mean to people’s life and hence facilitate effective decision-making regarding adaptation to climate change, as we are quantifying how recent outstanding events would be modified by our climate action. Temperatures during the heatwaves often increase twice as much as global mean temperatures, especially in a future 4 ºC warmer climate. In this future climate, maximum temperatures can locally reach 50ºC in many western Europe countries. Nighttime temperatures would be similar to the daytime temperatures in a preindustrial world. The global warming amplification can be partly explained by a robust soil drying in the future 4 ºC warmer climate (exacerbated due to the June 2019 heatwave) which is transmitted to a robust increase in Bowen ratio. Importantly, by design of our study, this response occurs without any changes in atmospheric circulation.

We compare the results of strongly coupled data assimilation and weakly coupled data assimilation by analyzing the assimilation effect on the prediction of the ocean as well as the atmosphere variables. The AWI climate model (AWI-CM), which couples the ocean model FESOM and the atmospheric model ECHAM, is coupled with the parallel data assimilation framework (PDAF, http://pdaf.awi.de). The satellite sea surface temperature is assimilated. For the weakly coupled data assimilation, only the ocean variables are directly updated by the assimilation while the atmospheric variables are influenced through the model. For the strongly coupled data assimilation, both the ocean and the atmospheric variables are directly updated by the assimilation algorithm. The results are evaluated by comparing the estimated ocean variables with the dependent/independent observational data, and the estimated atmospheric variables with the ERA-interim data. In the ocean, both the WCDA and the SCDA improve the prediction of the temperature and SCDA and WCDA give the same RMS error of SST. In the atmosphere, WCDA gives slightly better results for the 2m temperature and 10m wind velocity than the SCDA. In the free atmosphere, SCDA yields smaller errors for the temperature, wind velocity and specific humidity than the WCDA in the Arctic region, while in the tropical region, the error are larger in general.

This presentation will give an overview about an ongoing collaboration between the European Centre for Medium-Range Weather Forecasts (ECMWF) and the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI). Our recent development is a single-executable coupled configuration of the Integrated Forecasting System (IFS) and the Finite Volume Sea Ice-Ocean Model, FESOM2. This configuration is set up to participate in the DYAMOND project alongside ECMWF’s default IFS-NEMO configuration. IFS-FESOM2 and IFS-NEMO are tentative models to generate “Digital Twin” storm-scale, coupled simulations as envisioned in the European Destination Earth (DestinE) and Next Generation Earth Modelling Systems (NextGEMS) projects.FESOM2 has a novel dynamical core that supports multi-resolution triangular grids. The model and its predecessor FESOM1 have been used in many studies over the last decade, with a focus on the role of the polar regions in global ocean circulation. The impact of eddy-permitting and locally eddy-resolving resolution has been addressed in CMIP6 and HighResMIP simulations as part of the AWI-CM-1-1 global climate model, while simulations with up to 1km resolution in the Arctic Ocean have been performed in stand-alone mode.Initially, two coupled IFS-FESOM2 configurations have been tested: A coarse-resolution setup with a nominal 1° ocean, and a DYAMOND-II configuration with 0.25° ocean and IFS at 4.5km global resolution on average. For the latter configuration, FESOM2 is mimicking the “ORCA025” tri-polar curvilinear grid of the NEMO model, whose grid boxes have been split into triangles. Initialisation is from ECMWF’s analysis for IFS and NEMO, and from an ERA5-forced ocean spin-up for FESOM2. We discuss technical challenges with respect to the hybrid OpenMP and MPI parallelization in a single-executable context, describe a novel strategy for resource-efficient writing of model output, and summarise future applications such as exploring the impact of flexible FESOM2 grid configurations on the atmosphere - with ocean simulations that resolve leads in sea ice and ocean eddies almost everywhere.

The climate sensitivity is known to be mainly determined by the atmosphere model but here we discover that the ocean model can change a given transient climate response (TCR) by as much as 20% while the equilibrium climate sensitivity (ECS) change is limited to 10%. In our study, two different coupled CMIP6 models (MPI-ESM and AWI-CM) in two different resolutions each are compared. The coupled models share the same atmosphere-land component ECHAM6.3, which has been developed at the Max-Planck-Institute for Meteorology (MPI-M). However, as part of MPI-ESM and AWI-CM, ECHAM6.3 is coupled to two different ocean models, namely the MPIOM sea ice-ocean model developed at MPI-M and the FESOM sea ice-ocean model developed at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI). A reason for the different TCR is different ocean heat uptake through greenhouse gas forcing in AWI simulations compared to MPI-M simulations. Specifically, AWI-CM simulations show stronger surface heating than MPI-ESM simulations while the MPI-M model accumulates more heat in the deeper ocean. The vertically integrated ocean heat content is increasing stronger in MPI-M model configurations compared to AWI model configurations in the high latitudes. Strong vertical mixing in MPI-M model configurations compared to AWI model configurations seems to be key for these differences. The strongest difference in vertical ocean mixing occurs inside the Weddell Gyre, but there are also important differences in another key region, the northern North Atlantic. Over the North Atlantic, these differences materialize in a lack of a warming hole in AWI model configurations and the presence of a warming hole in MPI-M model configurations. All these differences occur largely independent of the considered model resolutions.

During the winter of 2019/2020, as the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) project started its work, the Arctic Oscillation (AO) experienced some of its largest shifts, ranging from a highly negative index in November 2019 to an extremely positive index during January–February–March (JFM) 2020. The permanent positive AO phase for the 3 months of JFM 2020 was accompanied by a prevailing positive phase of the Arctic Dipole (AD) pattern. Here we analyze the sea ice thickness (SIT) distribution based on CryoSat-2/SMOS satellite-derived data augmented with results from the hindcast simulation by the fully coupled Regional Arctic System Model (RASM) from November 2019 through March 2020. A notable result of the positive AO phase during JFM 2020 was large SIT anomalies of up to 1.3 m that emerged in the Barents Sea (BS), along the northeastern Canadian coast and in parts of the central Arctic Ocean. These anomalies appear to be driven by nonlinear interactions between thermodynamic and dynamic processes. In particular, in the Barents and Kara seas (BKS), they are a result of enhanced ice growth connected with low-temperature anomalies and the consequence of intensified atmospherically driven sea ice transport and deformations (i.e., ice divergence and shear) in this area. The Davies Strait, the east coast of Greenland and the BS regions are characterized by convergence and divergence changes connected with thinner sea ice at the ice borders along with an enhanced impact of atmospheric wind forcing. Low-pressure anomalies that developed over the eastern Arctic during JFM 2020 increased northerly winds from the cold Arctic Ocean to the BS and accelerated the southward drift of the MOSAiC ice floe. The satellite-derived and simulated sea ice velocity anomalies, which compared well during JFM 2020, indicate a strong acceleration of the Transpolar Drift relative to the mean for the past decade, with intensified speeds of up to 6 km d−1. As a consequence, sea ice transport and deformations driven by atmospheric surface wind forcing accounted for the bulk of the SIT anomalies, especially in January 2020 and February 2020. RASM intra-annual ensemble forecast simulations with 30 ensemble members forced with different atmospheric boundary conditions from 1 November 2019 through 30 April 2020 show a pronounced internal variability in the sea ice volume, driven by thermodynamic ice-growth and ice-melt processes and the impact of dynamic surface winds on sea ice formation and deformation. A comparison of the respective SIT distributions and turbulent heat fluxes during the positive AO phase in JFM 2020 and the negative AO phase in JFM 2010 corroborates the conclusion that winter sea ice conditions in the Arctic Ocean can be significantly altered by AO variability.

We have equipped the unstructured‐mesh global sea‐ice and ocean model FESOM2 with a set of physical parameterizations derived from the single‐column sea‐ice model Icepack. The update has substantially broadened the range of physical processes that can be represented by the model. The new features are directly implemented on the unstructured FESOM2 mesh, and thereby benefit from the flexibility that comes with it in terms of spatial resolution. A subset of the parameter space of three model configurations, with increasing complexity, has been calibrated with an iterative Green's function optimization method to test the impact of the model update on the sea‐ice representation. Furthermore, to explore the sensitivity of the results to different atmospheric forcings, each model configuration was calibrated separately for the NCEP‐CFSR/CFSv2 and ERA5 forcings. The results suggest that a complex model formulation leads to a better agreement between modeled and the observed sea‐ice concentration and snow thickness, while differences are smaller for sea‐ice thickness and drift speed. However, the choice of the atmospheric forcing also impacts the agreement of the FESOM2 simulations and observations, with NCEP‐CFSR/CFSv2 being particularly beneficial for the simulated sea‐ice concentration and ERA5 for sea‐ice drift speed. In this respect, our results indicate that parameter calibration can better compensate for differences among atmospheric forcings in a simpler model (i.e., sea‐ice has no heat capacity) than in more realistic formulations with a prognostic sea‐ice thickness distribution and sea ice enthalpy., Plain Language Summary: The role of model complexity in determining the performance of sea‐ice numerical simulations is still not completely understood. Some studies suggest that a more sophisticated description of the sea‐ice physics leads to simulations that agree better with sea‐ice observations. Others, however, fail to establish a link between complex model formulations and improved model performance. Here, we investigate this open question by analyzing a set of sea‐ice simulations performed with a revised and improved sea‐ice model that features substantial modularity in terms of model complexity. Ten model parameters in three different model configurations are optimized to improve the agreement between model results and observations, allowing a fair comparison between model configurations with varying complexity. The model optimization is repeated for two different atmospheric forcings to shed light on the relationship between model complexity and other sources of uncertainty in the sea‐ice simulations, such as those associated with the atmospheric conditions. The results suggest that a more complex formulation of our model can lead to a more appropriate representation of sea ice concentration and snow thickness, while it is less relevant for sea‐ice thickness and drift., Key Points: Increased sea‐ice model complexity can improve the simulated sea‐ice concentration and snow thickness Sea‐ice thickness and drift are only weakly affected by model complexity Parameter calibration can better compensate for differences between atmospheric forcings in a simpler model, Bundesministerium für Bildung und Forschung (BMBF) http://dx.doi.org/10.13039/501100002347, European Commission (EC) http://dx.doi.org/10.13039/501100000780, US Department of Energy (DOE)

Saisonale Meereisvorhersagen haben sich lange auf die gesamt-arktische Meereisausdehnung im Spätsommer konzentriert. Auch heute noch widmet der weithin bekannte "Sea Ice Outlook", zu dem ein breites Spektrum an Forschungsgruppen beiträgt, dieser Zielgröße viel Aufmerksamkeit und kommt zu dem etwas ernüchternden Schluss, dass die Vorhersagen lediglich in solchen Jahren gut sind, wenn die Ausdehnung in etwa dem linearen Trend entspricht. In den letzten 5-10 Jahren hat sich jedoch viel getan. Immer mehr operationelle Wetter-Vorhersagezentren erweitern ihre Vorhersagesysteme mit dynamischen Ozean- und Meereis-Komponenten, für saisonale und sub-saisonale wie auch auch für "Wetter"-Zeitskalen, und liefern somit auch Meereisvorhersagen in Nahe-Echtzeit und während des ganzen Jahres. Es wurden gezielt Verifikationsmetriken entwickelt, die deutlich aussagekräftigere Aussagen über Meereis-Vorhersagegüte erlauben, vor allem bezüglich der Lage der Eiskante. Und schließlich wurde unser Wissen über die Grenzen der Vorhersagbarkeit von Meereis erweitert. Eine wichtige Erkenntnis dabei ist, dass wir es nach wie vor mit einer deutlichen Lücke zwischen derzeit realisierter Vorhersagegüte und den Grenzen des prinzipiell Machbaren zu tun haben. In dieser Präsentation wird die angedeutete Entwicklung im Bereich der Meereisvorhersagen rekapituliert, mit einem gewissen Schwerpunkt auf Arbeiten, die am Alfred Wegener Institut zu diesem Thema durchgeführt wurden, werden und geplant sind.

The Alfred Wegener Institute Climate Model (AWI‐CM) participates for the first time in the Coupled Model Intercomparison Project (CMIP), CMIP6. The sea ice‐ocean component, FESOM, runs on an unstructured mesh with horizontal resolutions ranging from 8 to 80km. FESOM is coupled to the Max Planck Institute atmospheric model ECHAM 6.3 at a horizontal resolution of about 100km. Using objective performance indices, it is shown that AWI‐CM performs better than the average of CMIP5 models. AWI‐CM shows an equilibrium climate sensitivity of 3.2°C, which is similar to the CMIP5 average, and a transient climate response of 2.1°C which is slightly higher than the CMIP5 average. The negative trend of Arctic sea‐ice extent in September over the past 30years is 20–30% weaker in our simulations compared to observations. With the strongest emission scenario, the AMOC decreases by 25% until the end of the century which is less than the CMIP5 average of 40%. Patterns and even magnitude of simulated temperature and precipitation changes at the end of this century compared to present‐day climate under the strong emission scenario SSP585 are similar to the multi‐model CMIP5 mean. The simulations show a 11°C warming north of the Barents Sea and around 2°C to 3°C over most parts of the ocean as well as a wetting of the Arctic, subpolar, tropical, and Southern Ocean. Furthermore, in the northern middle latitudes in boreal summer and autumn as well as in the southern middle latitudes, a more zonal atmospheric flow is projected throughout the year.

Sea ice models have become essential components of weather, climate and ocean models. The reliability of process studies, environmental forecasts and climate projections alike depend on a realistic representation of sea ice. Developing and evaluating sea ice models requires methods for both large scales and fine-scale geomorphological structures such as linear kinematic features (LKF). We introduce a Multiscale Directional Analysis (MDA) method that diagnoses distributions of LKF orientation and intersection angles. The MDA method is different from previous methods in that it (a) takes into account the width of LKFs instead of estimating the orientation of centerlines; (b) separates curve-like features from point-like features providing the opportunity to reach a unified definition of LKF in both numerical and observational fields; (c) estimates scale-dependent intersection angles.

Despite ongoing global warming and strong sea ice decline in the Arctic, the sea ice extent around the Antarctic continent has not declined during the satellite era since 1979. This is in stark contrast to existing climate models that tend to show a strong negative sea ice trend for the same period; hence the confidence in projected Antarctic sea-ice changes is considered to be low. In the years since 2016, there has been significantly lower Antarctic sea ice extent, which some consider a sign of imminent change; however, others have argued that sea ice extent is expected to regress to the weak decadal trend in the near future.In this presentation, we show results from climate change projections with a new climate model that allows the simulation of mesoscale eddies in dynamically active ocean regions in a computationally efficient way. We find that the high-resolution configuration (HR) favours periods of stable Antarctic sea ice extent in September as observed over the satellite era. Sea ice is not projected to decline well into the 21st century in the HR simulations, which is similar to the delaying effect of, e.g., added glacial melt water in recent studies. The HR ocean configurations simulate an ocean heat transport that responds differently to global warming and is more efficient at moderating the anthropogenic warming of the Southern Ocean. As a consequence, decrease of Antarctic sea ice extent is significantly delayed, in contrast to what existing coarser-resolution climate models predict.Other explanations why current models simulate a non-observed decline of Antarctic sea-ice have been put forward, including the choice of included sea ice physics and underestimated simulated trends in westerly winds. Our results provide an alternative mechanism that might be strong enough to explain the gap between modeled and observed trends alone.

The need for reliable forecasts for the sea ice evolution from weeks to months in advance has substantially grown in the last decade. Sea ice forecasts are of critical importance to manage the opportunities and risks that come with increasing socioeconomic activities in the rapidly changing Arctic, which, despite the reduction of the sea ice cover, remains an extreme environment. The position of the sea ice edge is a key parameter for potential forecast users, such as Arctic mariners. However, little is known about the ability of current operational subseasonal forecast systems to predict the evolution of the ice edge. Therefore, we assess for the first time the skill of state‐of‐the‐art forecast systems, using a new verification metric that quantifies the accuracy of the ice edge position in a meaningful way. Our results demonstrate that subseasonal sea ice predictions are in an early stage, although skillful predictions 1.5months ahead are already possible. We argue that relatively modest investments into reducing initial state and model errors will lead to major returns in predictive skill.

Sea ice formation is accompanied by the rejection of salt which in nature tends to be mixed vertically by the formation of convective plumes. Here we analyze the influence of a salt plume parameterization in an atmosphere‐sea ice‐ocean model. Two 330‐yearlong simulations have been conducted with the AWI Climate Model. In the reference simulation, the rejected salt in the Arctic Ocean is added to the uppermost ocean layer. This approach is commonly used in climate modeling. In another experiment, employing salt plume parameterization, the rejected salt is vertically redistributed within the mixed layer based on a power law profile that mimics the penetration of salt plumes. We discuss the effects of this redistribution on the simulated mean state and on atmosphere–ocean linkages associated with the intensity of deepwater formation. We find that the salt plume parameterization leads to simultaneous increase of sea ice (volume and concentration) and decrease of sea surface salinity in the Arctic. The salt plume parameterization considerably alters the interplay between the atmosphere and the ocean in the Nordic Seas. The parameterization modifies the ocean ventilation; however, resulting changes in temperature and salinity largely compensate each other in terms of density so that the overturning circulation is not significantly affected.

The Polar Prediction Project (PPP, www.polarprediction.net) is a 10-year (2013–2022) endeavour initiated by the World Meteorological Organization’s (WMO) World Weather Research Programme (WWRP). Aim of this wide international endeavour is to promote cooperative weather and sea-ice research enabling development of improved environmental prediction services for the polar regions, on time scales from hours to seasonal.The PPP flagship activity, the Year of Polar Prediction (YOPP), has been launched in mid-2017 as a coordinated two-year period of intensive observing, modelling, verification, user-engagement and education activities. Since then, scientists and operational forecasting centers worldwide have closely worked together to observe, model, and improve forecasts of the Arctic and Antarctic weather and climate systems. During three Special Observing Periods in the Arctic and Antarctic, routine observations such as radiosonde launches and buoy deployments were enhanced (in the Arctic: 1 February ­– 31 March 2018 and 1 July – 30 September 2018, in the Antarctic: 16 November 2018 – 15 February 2019), aiming to close gaps in atmospheric and sea-ice observations and to enable significant progress in environmental prediction capabilities for the polar regions and beyond.in mid-2019, PPP has moved into its Consolidation Phase which will be key for the success of the initiative. Central activities and projects such as the YOPPSiteMIP initiative or the EU-project APPLICATE will significantly contribute to improving forecasts of weather and sea-ice conditions in polar regions and to make them available to its user community. Data collected during YOPP are available for everyone through the YOPP Data Portal (https://yopp.met.no/) to feed into improved environmental forecasting systems.In this presentation, an overview of the main achievements accomplished during the three YOPP Special Observing Periods, current activities including two more Special Targeted Observing Periods (TOPs) as well as prospects for future evaluations of PPP are provided.

We implement multivariate data assimilation in a seamless sea ice prediction system based on the fully-coupled AWI Climate Model (AWI-CM, v1.1). AWI-CM has an ocean/ice component with unstructured-mesh discretization and smoothly varying spatial resolution, which aims for seamless sea ice prediction across a wide range of space and time scales. The assimilation uses a Local Error Subspace Transform Kalman Filter coded in the Parallel Data Assimilation Framework. To test the robustness of the assimilation system, a perfect-model experiment is configured to assimilate synthetic observations. Real observations from sea ice concentration, thickness, drift, and sea surface temperature are further assimilated in the system. The analysis results are evaluated against independent in-situ observations and reanalysis data. Further experiments that assimilate different combinations of variables are conducted to understand their individual impacts on the analysis step. Particularly we find that assimilating sea ice drift improves the sea ice thickness estimate in the Antarctic, and assimilating sea surface temperature is able to avert a circulation bias of the free-running model in the Arctic Ocean at mid-depth. We also test the performance of an extended experiment where the atmosphere is constrained by nudging toward reanalysis data. The second version of the system assimilating more observations also with a new atmospheric model is currently under development.

The Sea Ice Drift Forecast Experiment (SIDFEx) is a Year of Polar Prediction (YOPP) community effort to solicit, collect, and analyze sea ice drift forecasts, based on various methods, on a regular basis. SIDFEx is inspired by research and operational needs to forecast future positions of assets drifting in Arctic sea ice. Beside a number of sea-ice buoys of the International Arctic Buoy Programme, current targets include the MOSAiC drift campaign main site (and distributed network) for which consensus forecasts are delivered every six hours. A systematic assessment of real drift forecasting capabilities across operational and research forecast systems is meant to improve our physical understanding of sea ice and to identify and resolve model shortcomings.Since the launch of SIDFEx in 2017, thirteen groups have started contributing drift forecasts to SIDFEx on a regular basis. Most groups derive their 2-days to seasonal-range forecasts by means of diagnostic tracking based on prediction drift fields of coupled or uncoupled general circulation models. Some groups submit ensembles of drift trajectories instead of single (deterministic) trajectories, and several groups submit their forecasts in real-time. We present results from around 75,000 individual forecasts, how they have been used for real-time support of the MOSAiC Arctic drift campaign since autumn 2019, and what they reveal about current models' capabilities to forecast sea-ice drift and deformation.

A new global climate model setup using FESOM2.0 for the sea ice-ocean component and ECHAM6.3 for the atmosphere and land-surface has been developed. Replacing FESOM1.4 by FESOM2.0 promises a higher efficiency of the new climate setup compared to its predecessor. The new setup allows for long-term climate integrations using a locally eddy-resolving ocean. Here it is evaluated in terms of (1) the mean state and long-term drift under pre-industrial climate conditions, (2) the fidelity in simulating the historical warming, and (3) differences between coarse and eddy- resolving ocean configurations. The results show that the realism of the new climate setup is overall within the range of existing models. In terms of oceanic temperatures, the historical warming signal is of smaller amplitude than the model drift in case of a relatively short spin-up. However, it is argued that the strategy of ‘de-drifting’ climateruns after the short spin-up, proposed by the HighResMIP protocol, allows one to isolate the warming signal. Moreover, the eddy-permitting/resolving ocean setup shows notable improvements regarding the simulation of oceanic surface temperatures, in particular in the Southern Ocean.

Complex models of the atmosphere show that increased carbon dioxide (CO2) concentrations, while warming the surface and troposphere, lead to lower temperatures in the stratosphere and mesosphere. This cooling, which is often referred to as “stratospheric cooling”, is evident also in observations and considered to be one of the fingerprints of anthropogenic global warming. Although the responsible mechanisms have been identified, they have mostly been discussed heuristically, incompletely, or in combination with other effects such as ozone depletion, leaving the subject prone to misconceptions. Here we use a one-dimensional window-grey radiation model of the atmosphere to illustrate the physical essence of the mechanisms by which CO2 cools the stratosphere and mesosphere: (i) the blocking effect, associated with a cooling due to the fact that CO2 absorbs radiation at wavelengths where the atmosphere is already relatively opaque, and (ii) the indirect solar effect, associated with a cooling in places where an additional (solar) heating term is present (which on Earth is particularly the case in the upper parts of the ozone layer). By contrast, in the grey model without solar heating within the atmosphere, the cooling aloft is only a transient blocking phenomenon that is completely compensated as the surface attains its warmer equilibrium. Moreover, we quantify the relative contribution of these effects by simulating the response to an abrupt increase in CO2 (and chlorofluorocarbon) concentrations with an atmospheric general circulation model. We find that the two permanent effects contribute roughly equally to the CO2-induced cooling, with the indirect solar effect dominating around the stratopause and the blocking effect dominating otherwise.

This study forms part II of two papers describing ECHAM6-FESOM, a newly established global climate model with a unique multi-resolution sea ice-ocean component. While part I deals with the model description and the mean climate state, here we examine the internal climate variability of the model under constant present-day (1990) conditions. We (1) assess the internal variations in the model in terms of objective variability performance indices, (2) analyze variations in global mean surface temperature and put them in context to variations in the observed record, with particular emphasis on the recent warming slowdown, (3) analyze and validate the most common atmospheric and oceanic variability patterns, (4) diagnose the potential predictability of various climate indices, and (5) put the multi-resolution approach to the test by comparing two setups that differ only in oceanic resolution in the equatorial belt, where one ocean mesh keeps the coarse ~1° resolution applied in the adjacent open-ocean regions and the other mesh is gradually refined to ~0.25°. Objective variability performance indices show that, in the considered setups, ECHAM6-FESOM performs overall favourably compared to five well-established climate models. Internal variations of the global mean surface temperature in the model are consistent with observed fluctuations and suggest that the recent warming slowdown can be explained as a once-in-one-hundred-years event caused by internal climate variability; periods of strong cooling in the model (‘hiatus’ analogs) are mainly associated with ENSO-related variability and to a lesser degree also to PDO shifts, with the AMO playing a minor role. Common atmospheric and oceanic variability patterns are simulated largely consistent with their real counterparts. Typical deficits also found in other models at similar resolutions remain, in particular too weak non-seasonal variability of SSTs over large parts of the ocean and episodic periods of almost absent deep-water formation in the Labrador Sea, resulting in overestimated North Atlantic SST variability. Concerning the influence of locally (isotropically) increased resolution, the ENSO pattern and index statistics improve significantly with higher resolution around the equator, illustrating the potential of the novel unstructured-mesh method for global climate modeling.

Coupled subseasonal forecast systems with dynamical sea ice have the potential of providing important predictive information in polar regions. Here, we evaluate the ability of operational ensemble prediction systems to predict the location of the sea ice edge in Antarctica. Compared to the Arctic, Antarctica shows on average a 30% lower skill, with only one system remaining more skillful than a climatological benchmark up to ∼30days ahead. Skill tends to be highest in the west Antarctic sector during the early freezing season. Most of the systems tend to overestimate the sea ice edge extent and fail to capture the onset of the melting season. All the forecast systems exhibit large initial errors. We conclude that subseasonal sea ice predictions could provide marginal support for decision‐making only in selected seasons and regions of the Southern Ocean. However, major progress is possible through investments in model development, forecast initialization and calibration.

The chapter presents a review of sea ice properties in relation to sub-seasonal to seasonal (S2S) predictions in the Arctic and the Antarctic. After a concise presentation of the main processes governing sea ice physics, the spatial distribution, seasonal cycle, and variability of sea ice in both poles are described. Using a variety of observations and model reconstructions of the four recent decades, the memory of the main descriptors of the sea ice state is quantified. In both the Arctic and the Antarctic, persistence of the sea ice areal properties emerges as the primarily source of sea ice sub-seasonal predictability, with strong dependence on season. Further memory can be obtained from reemergence mechanisms, implying processes internal to sea ice and coupling with the atmosphere and the ocean. In addition, lessons from modeling studies are addressed in terms of potential sea ice predictability and actual predictive skill. Finally, the chapter provides an overview of our understanding of the possible role of sea ice as a source of S2S atmospheric predictability, both in the polar regions and beyond.

A new global climate model setup using FESOM2.0 for the sea ice‐ocean component and ECHAM6.3 for the atmosphere and land surface has been developed. Replacing FESOM1.4 by FESOM2.0 promises a higher efficiency of the new climate setup compared to its predecessor. The new setup allows for long‐term climate integrations using a locally eddy‐resolving ocean. Here it is evaluated in terms of (1) the mean state and long‐term drift under preindustrial climate conditions, (2) the fidelity in simulating the historical warming, and (3) differences between coarse and eddy‐resolving ocean configurations. The results show that the realism of the new climate setup is overall within the range of existing models. In terms of oceanic temperatures, the historical warming signal is of smaller amplitude than the model drift in case of a relatively short spin‐up. However, it is argued that the strategy of “de‐drifting” climate runs after the short spin‐up, proposed by the HighResMIP protocol, allows one to isolate the warming signal. Moreover, the eddy‐permitting/resolving ocean setup shows notable improvements regarding the simulation of oceanic surface temperatures, in particular in the Southern Ocean.

Skillful sea ice forecasts from days to years ahead are becoming increasingly important for the operation and planning of human activities in the Arctic. Here we analyze the potential predictability of the Arctic sea ice edge in six climate models. We introduce the integrated ice-edge error (IIEE), a user-relevant verification metric defined as the area where the forecast and the “truth” disagree on the ice concentration being above or below 15%. The IIEE lends itself to decomposition into an absolute extent error, corresponding to the common sea ice extent error, and a misplacement error. We find that the often-neglected misplacement error makes up more than half of the climatological IIEE. In idealized forecast ensembles initialized on 1 July, the IIEE grows faster than the absolute extent error. This means that the Arctic sea ice edge is less predictable than sea ice extent, particularly in September, with implications for the potential skill of end-user relevant forecasts.

The field of Arctic sea ice prediction on “weather time scales” is still in its infancy with little existingunderstanding of the limits of predictability. This is especially true for sea ice deformation alongso-called Linear Kinematic Features (LKFs) including leads that are relevant for marine operations.Here the potential predictability of the sea ice pack in the wintertime Arctic up to ten days aheadis determined, exploiting the fact that sea ice-ocean models start to show skill at representing seaice deformation at high spatial resolutions. Results are based on ensemble simulations with a high-resolutionsea ice-ocean model driven by atmospheric ensemble forecasts. The predictability of LKFsas measured by different metrics drops quickly, with predictability being almost completely lost after4–8 days. In contrast, quantities such as sea ice concentration or the location of the ice edge retainhigh levels of predictability throughout the full 10-day forecast period. It is argued that the rapiderror growth for LKFs is mainly due to the chaotic behaviour of the atmosphere associated with thelow predictability of near surface wind divergence and vorticity; initial condition uncertainty for icethickness is found to be of minor importance as long as LKFs are initialized at the right locations.

We introduce a verification score for probabilistic forecasts of contours—the Spatial Probability Score (SPS). Defined as the spatial integral of local (Half) Brier Scores, the SPS can be considered the spatial analog of the Continuous Ranked Probability Score (CRPS). Applying the SPS to idealised ensemble forecasts of the Arctic sea-ice edge in a global coupled climate model, we demonstrate that the metric responds in a meaningful way to ensemble size, spread, and bias. When applied to individual forecasts or ensemble means (or quantiles), the SPS is reduced to the ’volume’ of mismatch, which in case of the ice edge corresponds to the Integrated Ice Edge Error (IIEE). By comparing initialised forecasts with climatological and persistence forecasts, we confirm earlier findings on the potential predictability of the Arctic sea-ice edge from a probabilistic viewpoint. We conclude that the SPS is a promising probabilistic verification metric, for contour forecasts in general and for ice-edge forecasts in particular.

Triggered by global climate change, the rapidly changing polar environments increasingly capture public awareness. In the north, Arctic sea-ice opening comes with many opportunities for economic development, transport and tourism but also bears substantial changes and high risks for humans and nature.