نبذة مختصرة : This paper proposes an experimental model for predicting the intraseasonal variability of ice area in the Sea of Okhotsk, implemented as a computer program. The physical and statistical model was built on the basis of the following components: the relationship of the seasonal maximum ice area in the Sea of Okhotsk with surface air and water surface temperature parameters; the relationship between short-term fluctuations in ice area influenced by cyclonic activity during lunar syzygies; and the tendency of long-term changes in ice area series in the Sea of Okhotsk. Various data series for the period 1980–2018 were analyzed in order to build a predictive physical and statistical model. The main data sources for determining parameter relationships were the ERA-Interim and ERA5 reanalysis archives, data on air temperatures from coastal hydrometeorological stations, and five-day maps of the distribution of ice concentration published by the Japan Meteorological Agency. The predictive model algorithms were implemented in the Python programming language using additional software libraries. The model allows the calculation of predicted values of ice area during the period of its fall-winter growth from November/December until its seasonal maximum (March), with a lead time of up to four months. The input data for constructing regression equations were a long-term series of air and surface water temperature parameters, dates of lunar phase changes, and values of ice area in the Sea of Okhotsk. The output data of the model were the values of the intraseasonal variability of ice area. The model was tested for the period 2001– 2020, and the average relative error of the calculated and predicted values of ice area compared to the actual values was less than 7%. The accuracy of prediction during the period of maximum development growth and the seasonal maximum for temperate and mild winters in terms of the complexity of ice conditions was 79 %.
Processing Request
No Comments.