No abstract

The research has been made on the basis of an array of ice data for the maximums and minimums seasonal distribution of sea ice in the Northern Hemisphere. The probability range for the appearance of positive differences after the negative ones in March is greater than in September, and conversely. During the periods of minimum (March) and maximum (September) interannual changes, after positive differences with an increase in the number of differences in the group, the values of subsequent changes increase. There is an increase in the probability of a one sign occurrence interannual difference with an increase in the group of another sign previous differences. For the probability of zero differences (absence of variability), there is an increase in both the probability of zero values and the number of previous zero differences. The results indicate the presence of a stabilization mechanism in the stochastic interannual variability of the sea ice area, which is confirmed by an increase in the probability of the appearance of a one sign and the number of previous differences of another sign in the group. An increase in the probability of zero values with an increase in the number of previous zero differences confirms the presence of geographical areas in which there are no interannual changes in the area of sea ice.

This article examines the reliability of ice data presented in the HadISST1 database (Hadley Centre Sea Ice and Sea Surface Temperature data set) of Hadley Meteorological Center (Hadley Centre for Climate Prediction and Research, Met Office, UK). The criterion for the reliability of HadISST1 database on the Northern hemisphere serves the average multiyear annual course of insolation in the Northern Hemisphere and the average multiyear annual course of insolation in 5-degree latitudinal zones at the upper boundary of the atmosphere, phase-shifted by two and three months to the past. It is revealed that the representative criterion for assessing the reliability of the business database is the average multiyear course of insolation in 5-degree latitudinal zones shifted by two months to the past. Evaluation of the data array on the Northern Hemisphere on the basis of the selected criterion demonstrates the overall high reliability of ice data reflected therein. However, there are two areas that differ in the degree of reliability of the presented data: 1) pack, drift ice with high negative values of the correlation coefficient; 2) coastal zone, in which the correlation is virtually absent (shore ice and flaw leads). The weak (or absent) correlation is observed over a significant length of the coastal zone in the Arctic Ocean, as well as in the straits and bays of the Canadian Arctic Archipelago. It is established hat the correlation coefficient of the annual course of the ice flow in the cells of the array of size 1 with the annual course of insolation of the Northern Hemisphere and the corresponding 5-degree latitudinal zones increases from the past to the present. This indicates the need for the correction of ice data in the early time range of the array (1901–1978), which precedes the time of the beginning of satellite observations.