Antarctic And Arctic Sea Ice Concentration Estimation And Sea Ice Extent Time Series Analysis

The change of Antarctic and Arctic which are important components of the earth has close relationship with change of global sea, atmospheric dynamics and ocean circulation patterns, has great significance in global climate change studies. Sea ice is the most active and variable composition that can affect the heat exchange of atmosphere and ocean through insulating them in polar. It plays an important role in regional and global climate change. Sea ice concentration and extent are important indicators in research on polar sea ice, and are also the most intuitive parameters to indicate polar sea ice surface change.Therefore, accurate estimation of sea ice concentration, time series analysis of sea ice extent and its relationship with the climate parameters(such as Southern Oscillation Index, the Arctic Oscillation and Global Mean Sea Level) are research emphasises at this stage.The microwave can penetrate clouds and almost has no influence from weather, data can be available freely and has wide coverage, so microwave remote sensing has become the main tool of large scale polar sea ice monitoring. There are more than ten kinds of sea ice concentration estimation methods based on passive microwave radiometer data, some methods are based on the sea ice concentration estimating equation and combined with the features and differences for different land types on passive microwave radiometer data.Such methods can obtain sea ice concentration for different types of sea ice, but the results are often overestimated or underestimated.This thesis built a method named Fully Constrained Least Squares (FCLS) algorithm based on sea ice concentration estimating equation. By constraining sea ice concentration value according to actual meaning of parameters based on sea ice concentration estimating equation and introducing FCLS to the solution procedure of sea ice concentration estimating equation, this method finally overcomes the problem of sea ice concentration overestimation or underestimation.Some sea ice parameters(such as area, extent and sea ice edge lines and so on) can be obtained from sea ice concentration results, therefore the accuracy of sea ice concentration product will affect the following study. The validation for sea ice concentration is limited although it has made some achievements by now. The limitation displays in two aspects: rather few validation studies for several algorithms utilizing multi-source data (in-situ data and estimated results obtained from different remote sensing data) simultaneously have been done, and also rather few cross comparison and analysis for different algorithms applied to different kinds of passive microwave radiometer datahave been done in this regard.Therefore, this thesis validates and analyzes the sea ice concentrations from Bootstrap, NASA Team, and FCLS algorithm which were applied to two kinds of passive microwave radiometer data (SSM/I and AMSR-E) utilizing multi-source data. The results indicated that FCLS generates a more accurate sea ice concentration with smaller RMSE (5%?18.4%) than Bootstrap and NT, the result compared with MODIS generates RMSE (5%?6.26%) and RMSE (7.7%? 18.4%) compared with ASPeCt.Bootstrap, NT and FCLS applied to the SSM/I and AMSR-E comparison with ASPeCt shows that FCLS also generates a more accurate sea ice concentration with smaller bias (-0.19%?0.84%) and RMSE(5.96%?6.21%) than Bootstrap and NASA Team.Time series analysis of sea ice and its anomaly under global climate change is the ultimate goal of this thesis. This thesis studies on time series analysis of sea ice and its relationship with the Southern Oscillation Index, the Arctic Oscillation and the global mean sea level from 1979 to 2014 from the perspective of the polar region and regional ultlizing a variety of mathematical methods. The results show that Antarctic sea ice concentration changed from-1.4% to 1.03%, except for the West Antarctic ice concentrations decreased, the remaining regional ice concentrations increased slightly, the Arctic sea ice concentration changed from -2.03% to 1.06%, sea ice concentration in most regions has decreased significantly. The sea ice extent in other regions and also the whole Antarctic increased, besides that of Amundsen slightly reduced, while Arctic sea ice extent decreased. Antarctic and Arctic sea ice extent has certain correlation with SOI, AO and GMSL, has obvious yearly cycle with SOI and AO.