In support of NOAAs sea ice forecasts in subseasonal to seasonal(S2S) range, the Climate Prediction Center (CPC) has been using an experimental sea ice prediction system, CFSm5, to provide weekly and seasonal Arctic sea ice predictions. The CFSm5 was developed based on the Climate Forecast System (CFS) with the MOM5 as oceanic component. Sea ice forecasts from CFSm5 initialized from CPC sea ice system (CSIS) have been shown to significantly improve over that from the operational CFS. In 2020, CPC started to prepare a transition from the use of CFSm5 to a new FV3-based Unified Forecast System (UFS) framework. This coupled UFS S2S model consists of the new FV3 atmospheric component, MOM6 oceanic component and CICE6 sea ice component. In this work, we evaluate the Arctic sea ice forecast skill in the UFS system in melt seasons and investigate impacts of cloud related parameters for an improved representation of sea ice in the coupled system. Analysis of the retrospective 45-day forecasts spanning 2012-2020 show the prediction skill of multi-week Arctic sea ice is generally (a) comparable to CFSm5, and (b) better than the operational CFSv2 hindcasts. The 9-month hindcasts spanning 2007-2020 show improved seasonal skill over the CFSm5 and CFSv2 hindcasts. The skill comparisons of CFSm5, CFSv2, UFS and Multi-Model Ensemble (MME) hindcasts will also be presented.

The warming Arctic is leading to dramatic sea ice loss, which is likely to continue. The resulting broader open-water area and thinning sea ice have already madetrans-Arctic navigation possible.However, navigability assessments of the entire Arcticat high temporal resolutionare still very limited and cannot meet the booming demand for Arctic transportation.To bridge this gap, we estimate the future potential of Arctic shipping by applying projections from state-of-the-art CMIP6 models to examine sea ice concentration and thickness at a daily frequency.The September navigable area will continue to increase through the 2060s for open-water (OW) ships and the2050s for Polar Class 6 (PC6) vessels. A quasi-equilibrium state will then ensue.The sailing time will be continuously shortened, especially for non-ice resistant OW ships, while the navigable days for both types of vessels will increase dramatically.PC6 ships will be able to smoothly sail both the European routes and Northwest Passage year-round starting in the 2070s when the decadal mean global surface temperature reaches approximately 18Cunder the SSP585 scenario.This systematic assessment of Arctic navigability shows a bright prospect of the emerging trans-Arctic shipping corridors, which should be of great interest to scientific as well as industrial communities.

The Arctic sea ice volume export exhibits a significant impact on the freshwater redistribution and energy balance. Benefiting from the multi-source sea ice data, including satellite retrievals and reanalysis, we firstly calculated the sea ice volume export through the two main outlets (i.e., the Fram Strait andthe Davis Strait) in both the freezing and melting season from 2010 to 2016. Our estimates show that the seasonal and interannualsea ice volume export through the Fram Strait vary from -240 (40) to -970 (60) km3 and -1970 (290) to -2490 (280) km3, respectively. Meanwhile, the maximum ice volume export through the Fram Strait usually occurs in spring and the average amount of the melt season ice volume export only occupies about one third of the yearly total amount. However, the sea ice volume export estimations through the Davis Strait indicate that there are nearly no ice outflows from June to November. Ice outflow through the Davis Strait reaches the maximum in spring with a mean value of 168 (46) km3while the long-term annual mean ice volume outflow is 312 (80) km3yr1. Analysis of the sea ice variables contributions demonstrates that the sea ice drift plays a much more important role on its monthly volume flux variability in the Fram Strait.

Forecast systems ranging from short term to climate predictions all aim at improving the resolution, physical parameterizations etc. These improvements are often demanding for the computer resources at hand, thus the details and quality of the forecast is often limited by processing power or the storage capacity of the system.

This presentation will focus on how to improve the models ability to utilize the system at hand. As a motivation, the presentation will show how the Elastic Viscous Plastic solver, within the sea-ice model CICE (https://github.com/CICE-Consortium/CICE), is refactored in order to optimize the scalability of the model. Two main issues was identified within the standard solver of the dynamics of CICE. These are the memory and the number of computations between each communication call that distributes information between the blocks. The solution builds on conversion of 2D arrays into 1D arrays within the dynamics in order to reduce the memory consumption and to eliminate communication within the dynamic component of CICE.

The result of these changes to the code is a physical model that significantly reduces the resources spend on calculating the forecast.

The solution can be expanded to other parts of the sea-ice model and other geophysical models. Especially the memory part is a general issue for many model geophysical models

The sea-ice thermodynamics included in the Regional Ice-Ocean Prediction System (RIOPS, Dupont et al. 2015) is based on the parameterizations of Bitz and Lipscomb (1999). As the sea-ice model in the Canadian Operational Network of Coupled Environmental Prediction Systems (CONCEPTS) is being updated to the Los Alamos sea-ice model (CICE) version 6, recent developments in sea-ice thermodynamics, such as the mushy layer thermodynamics, are becoming available and implemented as a column thermodynamic model package (Icepack). Here, Icepack (v.1.1.0) is used to reproduce observations from Sea Ice Mass BAlance (SIMBA) buoys deployed in the landfast ice close to Nain and Pond Inlet. We evaluate the performance of the conventional RIOPS thermodynamics and investigate new physics that improve the simulations. Results show that the sea-ice growth using the RIOPS configuration is smoother than in observations, where significant variations in growth rates are present. The inclusion of brine processes in the mushy layer thermodynamics improves the simulations, with periods of rapid ice growth that coincide with observations. We find that the snow-ice parameterizations do not reproduce the observed negative freeboards and delayed snow-ice formation, calling for added dependencies on brine fraction or ice porosity. Pond Inlet observations are further used to investigate the sensitivity of sea-ice melt to the Ice Thickness Distribution and melt pond parameterizations. Finally, we compare results from Pond Inlet to outputs from RIOPS and find that the local ice thickness was already overestimated in RIOPS prior to the buoy deployment, indicating discrepancies in the early winter dynamics.

Sea-ice is characterized by a coherent spatial structure,with sharp discontinuities and linear features (e.g. leads and ridges), and a multi-scale spatial structure (e.g. agglomerates of floes of different sizes). Traditional point-by-point verification approaches do not account for this complex spatial structure and the intrinsic spatial correlation existing between nearby grid-points. This leads to issues (such as double penalties), and an overall limited diagnostic power (e.g. traditional scores are insensitive to distance errors).

In this study we verify sea-ice forecasts using binary image distance metrics.Distance metrics provide a complementary set of verification measures that account for the spatial nature of sea-ice, and that overcome some of the limitations of traditional point-by-point verification. Distance metrics provide distance errors, e.g. of observed versus predicted sea-ice edge, in physical terms (km), thereby being informative and meaningful for user-relevant applications (such as navigation). Moreover, distance metrics are sensitive to similarities in the shape (e.g. for the MIZ) and overlap (e.g. for the sea-ice extent) of the predicted versus observed sea-ice features. Finally, distance metrics do not require interpolation to a common grid and are suitable for comparing models with different resolutions. We present results for the Canadian Global and Regional Ice Ocean Prediction Systems (GIOPS and RIOPS) and the ECMWF high resolution (HRES) global system, evaluated against the Ice Mapping System analysis during the Year of Polar Prediction.

Based on the measurements conducted over the landfast sea-ice in Prydz Bay, East Antarctica during the sea-ice growth season in 2016, various parameterization schemes in the High-resolution Thermodynamic Snow/Ice (HIGHTSI) model are evaluated to determine their applicability in the present study. The parameterization scheme of turbulent fluxes produces the largest errors compared with the parameterization schemes for other surface heat fluxes. However, the sea-ice thickness simulation is most sensitive to the differences between the parameterized and observed upward longwave radiation at the surface. In addition, the sea-ice thickness simulation during the growth season is highly sensitive to the oceanic heat flux, and a new oceanic heat flux parameterization scheme based on the bulk method is proposed. This scheme is verified in another year. Finally, the seasonal variation in the heat budget and its influence on the sea-ice thickness variation are analyzed. The net shortwave radiation, sensible heat flux, and conductive heat flux (the net longwave radiation and latent heat flux) are found to be the heat sources (heat sinks) at the surface during the growth season. The larger surface conductive heat flux and the smaller oceanic heat flux can intensify the growth of sea-ice.

The Viscous-Plastic (VP) rheology with an elliptical yield curve and normal flow rule is implemented in a Lagrangian modelling framework using the Smoothed Particle Hydrodynamics (SPH) meshfree method. Results show, from perturbation analysis of SPH sea-ice dynamic equations, that the classical SPH particle density formulation expressed as a function of sea-ice concentration and mean ice thickness, leads to incorrect plastic wave speed. We propose a new formulation for particle density that gives a plastic wave speed in line with theory. In all cases, the plastic wave in the SPH framework is dispersive and depends on the smoothing length (i.e., the spatial resolution) and on the SPH kernel employed in contrast with its finite difference method (FDM) implementation counterpart. The steady-state solution for the simple 1D ridging experiment is in agreement with the analytical solution within an error of 1%. SPH is also able to simulate a stable upstream ice arch in an idealized domain representing the Nares Strait in low wind regime (5.3 m · s−1) with an ellipse aspect ratio of 2, an average thickness of 1 m and free-slip boundary conditions in opposition to the FDM implementation that requires higher shear strength to simulate it. In higher wind regime (7.5 m · s−1) no stable ice arches are simulated — unless the thickness is increased — and the ice arch formation showed no dependence on the size of particles contrary to what is observed in the discrete element framework. Finally, the SPH framework is explicit, can take full advantage of parallel processing capabilities and show potential for pan-arctic climate simulations.

Sea-ice behaves like a fractal, that is, similar patterns repeat themselves at different scales. For sea-ice, these repeating patterns are the deformations. At large scales, there are long lines of fractures, or Linear Kinematic Features (LKF), and at small scales, finer lines appear. LKFs are strongly localized shear deformations together with divergence (leads) or convergence (pressure ridges). This multifractal property of LKFs can be seen in both data and models. Therefore, it becomes a way of comparing how well a model fits the satellite data, and ultimately, how good is the model at taking into account the fluxes between the ocean and the atmosphere. However, fine-scale LKFs happen at high resolution (0-2 km), and high-resolution models are extremely expensive to run, hence the importance of parametrizing these sub-grid scale phenomena. Different models perform differently, and one model, the Maxwell-Elasto-Brittle model (MEB) claims to perform better than the standard viscous plastic model (VP) at fine scales. One reason could be the presence of a brittle fracture parametrization, that is not present in the VP model. Therefore, we include a suitable damage parametrization in the VP model. Preliminary results show that the deformation statistics in the VP setting are largely influenced by the inclusion of damage in the model. This corresponds to finer, more defined high-intensity LKFs, which underlines the multifractal nature of sea-ice.

The anticyclonic winds blowing over the Beaufort Gyre lead to lateral Ekman transport, resulting in the accumulation of fresh water and a deepening halocline. Traditionally, studies suggest the equilibrium state of the gyre and the associated halocline is determined by a balance between atmospheric forcing and lateral mesoscale eddy fluxes. However, in recent years a third contender has been proposed to balance the Beaufort Gyre system, and it has been named the Ice-Ocean Governor. The Ice-Ocean governor was introduced in order to include the effects of sea ice on the ocean circulation. It is active when sea ice is present, and is believed to act as a regulating mechanism which brings the gyre to equilibrium. The governor minimizes the necessity for mesoscale eddies in order for the mass budget to remain balanced. To further investigate the role of the governor further, NASA ECCO data at 2km resolution with hourly outputs will be used to analyze the divergence of sea ice. Future work will focus on how eddies effect the relative velocity in a variety of sea ice regimes, and thus with different strengths of the Ice-Ocean Governor.

In recent decades, the Arctic minimum sea ice extent has transitioned from a predominantly thick multiyear ice cover to a thinner seasonal ice cover. In this contribution, we quantify the summer thermodynamic (lateral and basal melt) and dynamic (advection, compaction and export) sea ice area loss in the Arctic Ocean during the satellite era from 1979 to 2021 using passive microwave sea ice concentration and a Lagrangian sea ice tracking model driven by satellite-derived sea ice velocities. Results show that the thermodynamic signal dominates the total summer ice area loss and the dynamic signal remains small (~15%), even in years when the export is largest. Sea ice loss by compaction dominates the dynamic area loss except in the 1980s and early 1990s when the pack is thick and less mobile and in the 2010s when ridging and export become more important. Results from a simple (Ekman) free-drift sea ice model, supported by results from the Lagrangian model, suggest that non-linear effects between dynamic and thermodynamic area loss are key for all-time record minimum sea ice extent years. A detailed analysis of two all-time record minimum years (2007 and 2012) — one with a semi-permanent high in the southern Beaufort Sea and the other with a short-lived but extreme storm in the Pacific sector of the Arctic — shows that compaction by Ekman convergence amplified by the ice-albedo feedback dominated the sea ice loss signal in 2007 while Ekman divergence (together with an early melt onset) amplified by sea-ice albedo feedback was a major cause of 2012 minimum. We argue that Ekman divergence from more numerous and earlier intense summer storms when the sun is high above the horizon is a likely mechanism for a "first time" ice-free Arctic.

Late winter coastal divergence along the Eurasian coastline (referred to as the ice factory) or the Fram Strait ice export, a proxy for coastal divergence in the ice factory is a skillful predictor of the minimum sea ice extent (Williams et al., 2016; Brunette et al., 2018; Sesternikov and, 1979). Coastal divergence leads to the formation of coastal polynya where new ice grows but to a thickness that is not large enough to survive the following summer melt. This signal is then amplified by the ice albedo feedback and leads to more open water at the end of the summer melt season. In this thesis project, we will identify if this source of predictive skill in the seasonal forecast of the minimum sea ice extent is present in General Circulation Models. Since even small biases in the large-scale atmospheric circulation simulated by a model can short-circuit this coupling between dominant modes of atmospheric variability (NAO and AO), coastal divergence along the Eurasian coastline, Fram Strait ice export and therefore seasonal forecasting skill of the model, failure to reproduce this coupling observed in the real Arctic will be used to identify biases in GCM. Preliminary results show that subtle changes in the large-scale atmospheric circulation leads to opposite statistical relationship between Fram Strait ice area export, coastal divergence along the Eurasian coastline, and seasonal predictability of the minimum sea ice extent.

The second version of the seamless sea ice prediction system is developed based on the AWI-CM, v3, by which the computation time of the coupled model is reduced by a factor of 5. More ocean observations can be used for initialization with the feature of online coupled data assimilation. A larger ensemble of the coupled model with a size of 30 that is more than twice of its ancestor (12) could now run simultaneously while using nearly the same computation resources but having a higher resolution atmosphere model. Results from the data assimilation and the ensemble spread demonstrate a robust constrain on both sea ice and ocean states. The one-year sea ice forecasts initialized on January, April, July and October from 2003 to 2019 are conducted with sea ice concentration forecast calibration applied from 2011 to 2019. The sea ice raw forecasts generally meet the skill of most forecast centers with lead time of more than 14 days outperforming the climatology forecasts in both hemispheres, which is better in the Antarctic with lead time about one week longer than that in the Arctic. The skill in the calibrated sea ice forecasts can even persist for about 45 days in the Arctic and 27 days in the Antarctic. The calibration in the Antarctic however is not such promising as that in the Arctic. Both the raw and calibrated forecasts demonstrated strong initial-date dependence.

Arctic cyclones are the dominant type of hazardous weather system affecting the Arctic environment in summer. In late summer the marginal ice zone is extensive and wind forcing can move the ice readily; in turn, the dynamic sea ice distribution is expected to feedback on the developing weather systems. The interaction presents a major challenge to coupled forecasts of the Arctic environment from days out to a season ahead.

In summer 2022, in concert with ONR-THINICE, we aim to fly two research aircraft from Svalbard into Arctic cyclones passing over the marginal ice zone. We will measure the turbulent exchange fluxes, flying low above the interface between atmosphere and ice, at the same time as measuring the wind and cloud structure of the cyclones above and the properties of the ice below. Combining the observations with numerical modelling experiments using the Met Office NWP model, we aim to deduce the dominant physical processes acting and test theoretical mechanisms for the influence of sea ice on Arctic cyclone dynamics, with a particular focus on form drag and momentum exchange in the boundary layer.

Met Office and ECMWF forecasts that are coupled, or uncoupled, with a dynamic sea ice distribution have been investigated initially for systematic differences in the representation of boundary layer and surface fluxes, composited relative to the warm and cold sectors of Arctic cyclones and conditional on the surface beneath (ice, ocean, land). These findings inform the flight plans for the field experiment and focus of the observations.

The sources of aerosol particles during the Arctic summer and their ultimate effect on climate through their interactions with clouds has large uncertainties. Although aerosols in the winter are dominated by long range transport from more southerly latitudes, aerosols in the summer are expected to be more influenced by local sources, although the identify of these sources, whether natural or anthropogenic, remain poorly characterized. To improve our understanding of summer Arctic aerosol sources, properties, and their effect on climate through droplet formation, measurements of aerosol size distribution (0.3-10 micron diameter), aerosol chemical composition and fog droplet size distribution (2-50 micron diameter) were conducted at Tuktoyaktuk, Northwest Territories, Canada (69.4 N, 133.0 W) as part of the Year of Polar Predictions second Special Observing Period in the Arctic. The particle data suggested that ocean, mineral and/or road dust and combustion sources were the primary aerosol sources, including local activities in Tuktoyaktuk. The fog droplet data showed nine periods with reduced visibility, three of which were simulated with the Polar Advanced Weather and Research Forecasting model. However, comparisons between the simulations and observations were limited by the short nature of the events and the inability to represent the complex coastal features. This work highlights the ongoing challenges of using in-situ observations to improve modelling efforts.

Mixed-phase clouds are very important in Antarctica because of their role in precipitation and radiative forcing, which consequently impact surface energy and mass balance. However, mixed-phase clouds containing supercooled liquid water (SLW) are one of the main sources of bias in the climate models. Therefore, advanced measurements of cloud microphysical properties are crucial to improve our knowledge about their impact on the Antarctic climate system. In-situ measurements of cloud SLW content were done for the first time in Antarctica during the Antarctic Circumnavigation Expedition (ACE, December 2016- March 2017). During ACE, almost 100 radiosonde launches were conducted, of which 10 were equipped with cloud SLW content sensors manufactured by Anasphere. The ACE measurements showed that geometrically thin SLW layers were detected at heights from just below 1 km up to 5km above sea level. We analyze the vertical distribution of cloud SLW content during precipitation events associated with extra-tropical cyclones and warm/moist intrusions causing mixed-phase cloud formation without precipitation. We found cloud SLW content ranging between 0.05 and 0.3 g/m3, with relative humidity of the cloud liquid layers within the range of 80-100%. Some SLW layers correspond to significant temperature inversions. The shortwave and longwave cloud radiative forcing is estimated based on the SLW and other measured parameters. These high-resolution in-situ measurements of cloud SLW content are used to evaluate ERA5 reanalysis. Further measurements of mixed-phase clouds using SLW content sensors are planned at the Antarctic Peninsula in summer 2021/2022 and during the winter YOPP-SH targeted observing periods (April-July 2022).

On 14 August 2021, the rain fell on the peak of Greenland for the first time on record. The atmospheric circulation and water vapour transport responsible for the rain were investigated. Results show that a high-pressure ridge favoured the southerly advection of warm and moist air, the intrusion of which contributed to the rainfall. At the same time, Summit station observed above-freezing temperatures, which is the third time ever recorded after 2012 and 2019. There were also influxes of moisture, but no rainfall in the prior two events. Comparison between them and the 2021 event show different pressures and water vapour transports. In 2021, the moisture from the southwest ascended the sloping ice sheet and condensed, while in prior events, less moisture was originated from the southeast. The sufficient supply of warm and moist air was the key factor in the 2021 rain event.

Determination of the second-year ice local strength on MOSAiC expedition was carried out using the Borehole jack installation. Borehole jack was putted to the hole drilled in ice cover. The test was carried out with a step of 30 cm from the ice surface along the entire depth of the hole.

Before tests, ice physical properties were determined. During tests, the pressure in hydraulic system, the loading time, and the jacks displacement in the ice cover were recorded.

Determinations of the local strength of second-year ice in all year periods have been carried out. In the temperature range (Тi) from 18 to 1 С, the linear approximation of the local strength (sloc) is: sloc = 15,53 1,56Тi, R2 = 0,79.

Dependences of the second-year ice local strength from the ice liquid phase volume and density were obtained. They show that the local ice strength increases with decrease of the liquid phase volume and increase of density.

The obtained time dependence of local strength shows that the maximum local strength was observed in March, the minimum in August-September.

Aircraft observations from two Arctic field campaigns are used to characterise and model surface heat and moisture exchange over the marginal ice zone (MIZ). We show that the surface roughness lengths for heat and moisture over unbroken sea ice vary with roughness Reynolds number (R✶; itself a function of the roughness length for momentum, z0, and surface wind stress), with a peak at the transition between aerodynamically smooth (R✶<0.135) and aerodynamically rough (R✶>2.5) regimes. The conceptual model of Andreas (1987) accurately reproduces this peak, in contrast to the simple parameterizations currently employed in two state-of-the-art numerical weather prediction models, which are insensitive to R✶. We propose a new, simple parameterization for surface exchange over the MIZ that blends the Andreas (1987) conceptual model for sea ice with surface exchange over water as a function of sea ice concentration. In offline tests, this Blended A87 scheme performs much better than the existing schemes for the rough conditions observed during the IGP field campaign. The bias in total turbulent heat flux across the MIZ is reduced to only 13 W m-2 for the Blended A87 scheme, from 48 and 80 W m-2 for the Met Office Unified Model and ECMWF Integrated Forecast System schemes, respectively. It also performs marginally better for the comparatively smooth conditions observed during the ACCACIA field campaign. However, the benefit of this new scheme is dependent on the representation of sea ice topography via z0; a key remaining source of uncertainty in surface exchange parameterization over sea ice.

The Iceland and Greenland Seas are a crucial region for the climate system, being the headwaters of the lower limb of the Atlantic Meridional Overturning Circulation. Investigating the atmosphere-ocean-ice processes in this region often necessitates the use of meteorological reanalyses a representation of the atmospheric state based on the assimilation of observations into a numerical weather prediction system. Knowing the quality of reanalysis products is vital for their proper use. Here we evaluate the surface-layer meteorology and surface turbulent fluxes in winter and spring for the latest reanalysis from the European Centre for Medium-Range Weather Forecasts ERA5. In situ observations from a meteorological buoy, a research vessel and a research aircraft during the Iceland-Greenland Seas Project provide unparalleled coverage of this climatically important region. They allow a comprehensive evaluation of the surface meteorology and fluxes of these subpolar seas and, for the first time, a specific focus on the marginal ice zone. Over the ice-free ocean, ERA5 generally compares well to the observations of surface-layer meteorology and turbulent fluxes. However, over the marginal ice zone the correspondence is noticeably less accurate: for example, the root-mean-square errors are significantly higher for surface temperature, wind speed and surface sensible heat flux. The primary reason for the difference in reanalyses quality is an overly smooth sea-ice distribution in the surface boundary conditions used in ERA5. Particularly over the marginal ice zone, unrepresented variability and uncertainties in how to parameterize surface exchange compromise the quality of the reanalyses.

We present an Arctic aerosol optical depth (AOD) climatology and trend analysis for 2003-2019 spring and summertime periods derived from a combination of multi-agency aerosol reanalyses, remote sensing retrievals, and ground observations. This includes the U.S. Navy Aerosol Analysis and Prediction System ReAnalysis version 1 (NAAPS-RA v1), the NASA MERRA-2, and the ECMWF CAMSRA. Space-borne remote sensing retrievals of AOD are considered from MODIS, MISR, and CALIOP. Ground-based data include sun photometer data from AERONET sites and oceanic Maritime Aerosol Network (MAN) measurements. Aerosol reanalysis AODs and space-borne retrievals show consistent climatological spatial patterns and trends for both spring and summer seasons over the sub-Arctic (60-70N). Consistent signs in the AOD trend are also found for the high Arctic (north of 70N) from reanalyses. The aerosol reanalyses yield more consistent AOD results than climate models, verify well with AERONET, and corroborate complementary climatological and trend analysis. Black Carbon (BC) AOD in the Arctic comes predominantly from biomass burning sources in both MAM and JJA, and biomass burning overwhelms anthropogenic sources in JJA for the study period.AOD exhibits a negative trend in the Arctic in MAM, and a positive trend in JJA during 2003-2019, due to an overall decrease in sulfate/anthropogenic pollutions, and a significant increase in biomass burning smoke in JJA. Interannual Arctic AOD variability is significantly large, driven by fine-mode, and specifically, biomass burning (BB) smoke, though more so in JJA than in MAM.

This presentation will give an overview of the work at ECMWF done during, and in support of, the Year of Polar Prediction. This will include the development of a multi-layer snow scheme and coupled atmosphere-ice-ocean forecasts and techniques to evaluate them against observations. Progress was also made in understanding the contribution of the Arctic observing system to forecast skill and the importance of polar-to-midlatitude linkages for medium-range weather forecasts. An important aspect was to make ECMWF forecasts openly available in order to facilitate collaboration with the academic and wider NWP community. This was done through the release of the ECMWF-YOPP dataset, which includes physical process tendencies, and the YOPP site Model-intercomparison Project (YOPPsiteMIP) dataset designed to facilitate process-oriented evaluation of forecasts at polar supersites.

All-sky is the assimilation of radiances in the clear, cloudy and precipitation conditions which have shown potentially positive impact on the weather forecast in the global NWP models. This work concerns the implementation of the All-sky method developed at ECMWF into the limited area NWP model Harmonie-Arome data assimilation system. We will show the first results where we activate the all-sky observation operator for the humidity sensitive microwave radiances from MHS sensors. We then evaluate the impact of assimilating the cloudy and rainy radiances on the analysis with means of observation minus background and observation minus analysis statistics. Further, we will present the results from the longer impact studies in the Arome-Arctic domain and gauge the impact on the short range forecasts compared to the currently operational clear-sky radiance assimilation.

As contribution to the Year of Polar Prediction (YOPP), Environment and Climate Change Canada has developed the Canadian Arctic Prediction System (CAPS, 3 km), that has been running in experimental mode since February 2018. The Meteorological Service of Canada runs also two other operational systems that cover the Arctic: the Regional Deterministic Prediction System (RDPS, 10 km) and the Global Deterministic Prediction System (GDPS, 25 km).

This work presents the surface variable objective verification for the Canadian Deterministic Prediction Systems during YOPP, focusing on the Arctic Special Observing Periods (Feb-March and July-Aug-Sept 2018). All three systems exhibit a diurnal cycle in the near-surface temperature biases. All three systems systematically over-predict weak winds and under-predict strong winds; CAPS outperforms RDPS and GDPS in predicting near-surface wind.

In order to mitigate representativeness issues, the model tile temperatures are adjusted to the station elevation by applying standard atmosphere lapse-rate: the lapse-rate adjustment reduce the temperature cold biases characterizing mountain terrains. Verification of winter precipitation is performed by adjusting solid precipitation measurements from the undercatch in windy conditions: the systematic over-forecast, which was artificially inflated by the undercatch, is reduced after the adjustment, to attain neuter bias.

These YOPP verification activities have identified some strengths, weaknesses and systematic behaviours of the Canadian deterministic prediction systems at high latitudes. Moreover, they have revealed some issues related to the verification of surface variables, and led to the development of better verification practices for the polar regions (and beyond).

Assessing the quality of precipitation forecasts requires observations, but all precipitation observations have associated uncertainties making it difficult to quantify the true forecast quality. One of the largest uncertainties is due to the wind-induced undercatch of solid precipitation gauge measurements.

This study discusses how this impacts the verification of operational precipitation forecasts for Norway. The forecasts are compared with high-quality reference measurements (less undercatch) and with more simple measurement equipment commonly available (substantial undercatch) at the Haukeliseter observation site in Norway. Then the verification is extended to include all Norwegian observation sites: 1) stratified by wind speed, since calm (windy) conditions experience less (more) undercatch; and 2) by applying transfer functions, which convert measured precipitation to estimates of what would have been measured with high-quality equipment with less undercatch, before the forecastobservation comparison is performed. Finally, precipitation from the high-resolution regional reanalysis CARRA and the global reanalysis ERA5 are examined in a similar way.

Results show that the wind-induced undercatch of solid precipitation has a substantial impact on verification results. Furthermore, applying transfer functions to adjust for wind-induced undercatch of solid precipitation gives a more realistic picture of true forecast capabilities. In particular, estimates of systematic forecast biases are improved, and to a lesser degree, verification scores like correlation, RMSE, ETS, and stable equitable error in probability space (SEEPS). However, uncertainties associated with applying transfer functions are substantial and need to be taken into account in the verification process.

In the Arctic, significant warming is occurring and the sea ice extent is rapidly declining, but whether the Arctic sea ice changes are modulating the midlatitude climate is a contentious issue. BarentsKara (BK) sea ice conditions are considered a potential source of seasonal predictability in the winter midlatitudes. In this study, we examined the link between BK sea ice and winter Eurasian temperature in hindcast data produced by six state-of-the-art seasonal prediction models provided by Copernicus Climate Change Service. Specifically, we examined whether anomalous BK sea ice in November, which is provided as the initial condition of seasonal prediction, is a precursor to interannual variability of several wintertime variables associated with the Warm ArcticCold Eurasia (WACE) pattern. A lead-lag correlation analysis revealed that the autumn BK sea ice anomaly is a nominal precursor of winter atmospheric conditions over Eurasia, and the WACE pattern did not rigorously reflect a sea iceEurasia causal link. In the models used, the winter pressure system in the Urals region is the primary driver of BK sea ice and Eurasian temperature. We also found that most seasonal prediction models slightly undervalue the magnitude of BK sea ice variations and their link to the Eurasian temperature variations associated with the WACE in winter. This underestimation might be reducing by improving the reproduction of the amplitude of sea ice variability by the models.

Arctic cyclones (ACs) are a severe atmospheric phenomenon that affects the Arctic environment in summer. Therefore, the accurate prediction of ACs is important in anticipating their associated environmental and societal costs. In addition, their long lifetime suggests that the ACs can be a source of the predictability of the Arctic atmosphere in summer on sub-seasonal timescales. However, the predictability of the position and development of extraordinary ACs at their mature stage is relatively low (about 2.5-4.5 days) compared with those of the mid-latitude cyclones, even in operational forecasts. Previous studies suggested that predictions for upper-level trough and ridge and tropopause polar vortex (TPV) were key factors for the accurate prediction of development and position of the AC in August 2012. The merging of upper-level warm cores associated with a migrating cyclone and an AC was essential for an accurate prediction of the development of surface ACs. This study investigated the processes of cyclone merging between AC and migrating cyclone (MC) by using potential vorticity (PV). The vertical cross-section of the PV on the AC and MC centers showed that the surface PV associated with MC merged with the upper-level PV associated with AC before their mature stage. Besides, the PV inversion analysis in the quasi-geostrophic framework for AC12 indicated that the vertical motion induced by the upper-level PV contributes to the persistence by absorbing the surface PV of MC. These results suggest the importance of accurate prediction of TPV and PV merging for the AC persistence.

Precipitation plays an important role for the Arctic climate system. It affects the atmospheric thermodynamics due to phase changes and the surface characteristics like snow depth, mass balance and albedo. In the Arctic, there are two processes which could affect the precipitation. These are the enhanced local evaporation and the poleward moisture transport. This poleward moisture transport is often associated with Atmospheric Rivers (ARs). ARs are long, narrow bands that transport anomalous huge amounts of water vapour and heat from the lower latitudes towards polar regions.

Preliminary results have shown that ARs produce significant amount of rain and snow. However, precipitation associated with an AR is not only concentrated within the AR itself. Precipitation also occurs within a wider area of the cyclones and frontal systems often connected to ARs. We developed a new method to distinguish precipitation within the AR shape and the precipitation related to cyclones and fronts. The methodology was applied to AR cases occurring during two field campaigns close to Svalbard. The early summer campaign 2017 included more cyclones that were responsible for more precipitation than ARs. Whereas in the late winter and early spring campaign 2019 the precipitation was more related to ARs. The presence of an AR close to a front strongly enhanced the frontal precipitation compared to the fronts without ARs. Applying the method to longterm ERA5 reanalysis data will shed light on the relative contribution of different weather systems on precipitation, its phase composition and possible trends.

Risk management for wind farms has become more standardized in terms of calculating acceptable risk criteria. Yet, communication of possible risks and their consequences for societal actors has not evolved into a validated set of best practices. The current state of knowledge about best practices for ice throw/fall risk communication is still in an exploratory phase, and empirical research is fragmented. The main attempt toward a consolidation of best practices in ice throw risk communication has been part of the IEA-Wind-TCP Task-19. Its mandate is to provide international guidelines for ice risk assessment. A report published in 2018 is currently being updated. The work reported in this presentation is part of the project Wind Energy in Icing Climates, funded by the Norwegian Research Council and wind farm operators in Norway. The Norwegian Meteorological Institute has executed a national survey with the specific aim to develop recommendations for communication of the risk of ice throw from turbines in Norwegian wind farms. The survey aimed at getting insight into perceptions of the general public in Norway about ice-throw risk, and the perceived value of different communication tools and formats of ice-throw risk information. We discuss findings on a range of topics, including peoples familiarity with wind turbine parks, their weather risk information seeking patterns, understanding of impact-based warnings for ice-throw risk, and behavioural capacity to mitigate negative impacts emerging from ice throw risks. We provide a systematic set of recommendations regarding communication and formatting of ice throw risk warning information.

The overall objective of Blue-Action was to actively improve our ability to describe, model, and predict Arctic climate change and its impact on Northern Hemisphere climate, weather and their extremes, and to deliver valued climate services of societal benefit. Blue-Action has taken a transdisciplinary approach, bridging scientific understanding within Arctic climate, weather and risk management research with key stakeholder knowledge of the impacts of climatic weather extremes and hazardous events, leading to the co-design of better services. This bridge has built on innovative statistical and dynamical approaches to predict weather and climate extremes. In dialogue with users, Blue-Action has taken stock in existing knowledge about cross-sectoral impacts and vulnerabilities with respect to the occurrence of these events and their prediction. Modeling and prediction capabilities have beenenhanced by targeting firstly, lower latitude oceanic and atmospheric drivers of regional Arctic changes and secondly, Arctic impacts on Northern Hemisphere climate and weather extremes. Coordinated multi-model experiments have been key to test new higher resolution model configurations, innovative methods to reduce forecast error, and advanced methods to improve uptake of new Earth observations assets have been developed and implemented. Blue-Action has demonstrated how such an uptake may assist in creating a better optimized observation system for various modelling applications. Here we report on main findings and their exploitation.