![]() |
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ЦЭМИ РАН |
||
Passive microwave (SSMR, SSM/I, SSMIS etc) are covering the total area of the Arctic Ocean since October 1978. Various algorithms are used to derive sea ice concentration out of brightness temperatures. As the microwave data is the most widely-used source of information for estimation of the multiannual changes of ice extent, high attention is devoted to validation of existing and development of new more accurate algorithms. The two of the most popular NASA Team and ASI algorithms and recently developed Russian algorithm Vasia 2 data, were used for comparison with sea ice total concentration obtained from visual shipborne observations. Visual observations were carried out according to a special method, developed in the AARI: homogeneous ice zones are fixed continuously during the ship motion 24 hours a day and such parameters as concentration, age, ridges, snow depths, ice thickness, melting stage etc. are determined for each ice zone. The comparison shows significant differences between satellite and shipborne data observed onboard icebreakers during 15 scientific expeditions to the Barents, Kara, Laptev and East-Siberian Seas. The mostly pronounced differences are fixed in the ice edge areas, resulted to uncertain estimation of the total ice concentration and ice extent using in various practical and scientific tasks. Generally, satellite data underestimate sea ice concentration: average error is about 10% both in winter and in summer. Statistical analysis is presented separately for the summer and winter data and for total ice concentration with and without new ice. During the summer period in the area of rare ice the SSM/I data overestimate total ice concentration by 0.5-1 tenths, but in close ice – underestimate in average by 2 tenths. Excluding new ice from total ice concentration obtained onboard the icebreakers results in decreasing of average error. During the winter period in the area of rare ice the SSM/I data overestimate total ice concentration by 1-2 tenths, but in close ice - underestimate as in summer by 2 tenths. New ice in winter is better determined by remote sensing techniques, and, consequently, its excluding from the total ice concentration does not lead to decreasing of the average error. The most important parameter resulted to errors in calculation of ice concentration in summer – is melting stage of ice cover. Correlation of difference between SSM/I and field derived ice concentration with stage of ice melting was determined. The study was supported by the Russian Ministry of Education and Science, agreement №14.616.21.0078 (RFMEFI61617X0078).