We consider simulation of the present day climate with the use of the climate model INM-CM48 in comparison with the result of the previous model INMCM4.0 which used different parameterizations of many physical processes and also in comparison with the model INM-CM5 which uses the same parameterizations, but with better spatial resolution. It is shown that the model INM-CM48 reproduces the modern climate better than the model INMCM4.0 in most indicators.

The Finite-Element Sea Ice Model (FESIM), used as a component of the Finite-Element Sea ice Ocean Model, is presented. Version 2 includes the elastic-viscous-plastic (EVP) and viscous-plastic (VP) solvers and employs a flux corrected transport algorithm to advect the ice and snow mean thicknesses and concentration. The EVP part also includes a modified approach proposed recently by Bouillon et al. (2013), which is characterized by an improved stability compared to the standard EVP approach. The model is formulated on unstructured triangular meshes. It assumes a collocated placement of ice velocities, mean thicknesses and concentration at mesh vertices, and relies on piecewise-linear (P₁) continuous elements. Simple tests for the modified EVP and VP solvers are presented to show that they may produce very close results provided the number of iterations is sufficiently high.

This paper completes a series of studies conducted with a numerical model of low spatial resolution for the area of the North Atlantic and Arctic Ocean north of 65° N without including Hudson Bay to reproduce the monthly mean state in 1948–2002. The paper gives a detailed description of the physical formulation of the problem, the approximations and parametrizations that were employed, and the parameters of atmospheric and oceanic external forcings. Generally, the author follows the requirements of the protocol of the Arctic Ocean Model Intercomparison Project (AOMIP). The paper analyzes the main monthly mean characteristics: ice thickness, ice-drift velocity, snow thickness, velocity of currents, and temperature and salinity, which were derived as multiyear averages over 1958–2002. A good reproduction of the following sea-ice characteristics is noted: ice thickness, propagation area, ice concentration, and drift velocity. For example, the annual mean ice-drift velocity is about 3.33 cm/s, which is consistent with the data obtained from drifting buoys, which yield an annual mean ice-drift velocity of 3.65 cm/s. At the same time, there are problems of reproducing the observed snow-thickness distribution. The use of parametrizations of horizontal turbulence of the Neptune-Effect type and eddy transfer for a scalar makes it possible to conserve the cyclonic circulation in the Atlantic water layer although the low spatial resolution gives no way of explicitly reproducing the narrow topographic jet. In conclusion, we briefly discuss the main unsolved problems and, primarily, the problem of reproducing the freshwater content. The model and further versions of it will be used in the European Commission project DAMOCLES (Developing Arctic Modeling and Observing Capabilities for Long-Term Environmental Studies) to assess the sensitivity of the Arctic Ocean to the parameters of external forcing and to evaluate the role of tides in the formation of the Arctic Ocean’s climate.

This paper presents the results of reconstructing the state of ice and snow covers on the Arctic Ocean from 1948 to 2002 obtained with a couplod model of ocean circulation and sea-ice evolution. The area of the North Atlantic and Arctic Ocean north of 65° N, excluding Hudson Bay, is considered. The monthly mean ice areas and extents are analyzed. The trends of these areas are calculated separately for the periods of 1970–1979, 1979–1990, and 1990–2002. A systematic slight underestimation by the model is observed for the ice extent. This error is estimated to fit the error of 100 km in determining the position of the ice edge (i.e., close to the model resolution). In summer the model fails to reproduce many observed polynias: by observational data, the ice concentration in the central part of the Arctic Ocean constitutes around 0.8, while the model yields around 0.99. The average trend for the area of ice propagation in 1960–2002 is 13931 km²/year (or approximately 2% per decade); the trend of the ice area is 17643 km²/year (or approximately 3% per decade). This is almost three times lower than satellite data. The calculated data for ice thickness in the late winter varies from 3.5 to 4.8 m, with a clear indication of periods of thick ice (the 1960s–1970s) and relatively thin ice (the 1980s); 1995 is the starting point for quick ice-area reduction. The maximum ice accumulation is in 1977 and 1988; here, the average trend is negative: −121 km³/year (or approximately 5.5% per decade). In 1996–2002, the average change in the ice thickness constituted +1.7 cm/year. This speaks to the relatively fast disappearance of thin-ice fractions. This model also slightly underestimates the snow mass with a trend of −2.5 km³/year (almost 0.35 mm of snow per year or 0.1 mm of liquid water per year). An analysis of the monthly mean ice-drift velocity indicates the good quality of the model. Data on the average drift velocity and the results of comparisons between the calculated and satellite data for individual months are presented. A comparison with observational data from 1990–1996 in the Fram Strait shows that the model yields 3.28 m for the average ice thickness against the observed value of approximately 3.26 m. For the same period, the model yields a monthly mean transport of 291.29 km³ as compared to the observed value of 237.17 km³. A comparison between the measured and calculated drift velocities in the Fram Strait indicates that the model value is around 9.78 cm/s, which is comparable to the measured value of 10.2 cm/s. The existing problems with describing the ice redistribution by thickness gradations are illustrated by comparing data on ice thickness in the Fram Strait.

This paper considers the formation of the thermohaline structure of the Arctic Ocean: the formation of the salinity field and a freshwater reservoir in the Beaufort Sea and the transport of warm Atlantic water into the central part of the Arctic Ocean. A new version of the Finite Element Model of the Arctic Ocean (FEMAO) with a low spatial resolution is used. The main distinctions of this version are the following features: a new equation of state, a more sophisticated parameterization of vertical turbulence, modified formulations for the boundary conditions on open boundaries (using satellite data on the sea level) and at the upper boundary of the ocean, and the use of a variable eddy diffusivity in the parameterization of the eddy transport of a scalar. Our experiments indicated that the use of the parameterization of the eddy transport of a scalar enhances the transport of warm Atlantic waters to the central part of the Arctic Ocean through the Fram Strait; the results are most realistic when a variable coefficient is used. The Neptune effect has a contradictory role and, in the future, a higher spatial resolution should be used instead of this parameterization. We revealed that a key factor in the thermohaline fields on a large time scale is the interaction with the Atlantic Ocean, which is the source of heat and saline water.