Reconstruction of spatial distribution of sea ice in Bohai Sea from 1988 to 2015 based on space-time weighted KNN algorithm
This study used AVHRR and MODIS remote sensing interpretation data,combined with meteorological factors such as daily average temperature,accumulated temperature of 3 d-1.8℃,accumulated frozen ice days and accumulated melted ice days that are highly correlated with Bohai sea ice area,constructed a spatial completion model with spatial resolution of 1km based on space-time weighted KNN algorithm,and reconstructed the continuous daily data set of Bohai sea ice spatial distribution from 1988 to 2015.The complete mean square error of the spatial distribution of sea ice in Bohai Sea is 0.03 km2,the classification accuracy is above 87%,and the average accuracy in 28 years is 91.87%.The mean square error is positively correlated with the missing rate of sea ice remote sensing image data.As shown by results,the model has small mean square error and high classification accuracy,and can be used to complete the spatial distribution data of sea ice in Bohai Sea,with high spatial resolution and fast completion speed.It is of important practical value in the field of marine environmental safety management,especially in the risk management of sea ice disasters in icy waters.
Bohai sea iceweighted K-nearest neighbors(KNN)algorithmspatial distribution of sea icereconstruction of sea ice data
NETL
NSTL
万方数据
