Estimation Of Thin Sea-Ice Thickness From AMSR-E Data In The Ross Ice Shelf Polynya

Due to the Speciality of geographical location and other aspects,the polar regions have become the amplifier of the global climate change.As a result of sea-ice-air interaction in the polar regions,polynya is the most sensitive sensor of local and global atmosphere change.It plays a crucial role in understanding the climate system,energy exchange and polar ecosystems.Coastal polynyas are isolated areas of open water and thin ice within the ice pack in a polar region which are driven by offshore-wind from the continental/ice shelf.Continuously wind from continental/ice shelf blows the ice away and makes the warm water exposed to cold air,resulting in the production of new ice,the process makes coastal polynya to be an important place for material and energy exchange among sea,ice and air,and ice-product as well.We can get a more accurate information about heat flux exchange result from specific thickness model to our study area,and get a accurate ice production further.In this paper,according to the ice thermodynamics principle and the relationship between ice thickness and emissivity,we built a model between passive microwave remote sensing data and thin ice thickness for Ross Ice Shelf Polynya(RISP),and analyzed it.The thermal growth model based thin ice thickness algorithm is an ice thickness method which based on a physical process.According to this method,a thin ice thickness of 0-0.5 m could be retrieved.In this paper,we chose the clear-sky days in 2003 and 2004 to calculate the MODIS thermal ice thickness,the result shows that the average thickness of thin ice is 0.23219m and the standard deviation is 0.130432m.We also explored the relationship among thermal ice thickness and input variables,and there is an strong relationship between the thickness and windspeed,R2 = 0.6373.After the resampling,uniformity analysis and ice shelf expansion processes,ice shelf expansion and resampling changed the distribution of ice thickness in the study area respectively:the expansions removed 10%and 6%of the 0-0.lm ice thickness at 6250m and 12500m resolution,respectively;the re-sampling removed 6.0%and 10.2%of the 0-0.lm ice thickness grid,respectively.The accuracy of the thermal ice thickness is based on correct choice of thermodynamic model and parameters,which is one of the difficulties in obtaining ice thickness using this method.At the same time,although this method reflects the physical mechanism of ice thickness formation,it is difficult to observe sea ice continuously due to the influence of cloud and fog in thermal infrared,and it is impossible to obtain a continuous observation ice thickness value of polynya.Passive microwave remote sensing has ability to observe the earth without influence of cloud,fog or night.Due to the different surface salinity of different ice thickness,the microwave radiometer will receive different levels of radiation signal.Based on the AMSR-E brightness temperature data released by NSIDC,comparing the thin ice thickness obtained from the MODIS,this paper established an AMSR-E thin ice thickness model for the RISP.Because the 89GHz data have a tendency to be influenced by water vapor in the atmosphere,before establishing the model,we use the 36GHz pixels to replace the vapored 89GHz pixels.After comparing with validation data,the average error of 36GHz model is-0.047m,RMSE is 0.082m,the average error of 89GHz model is-0.011m,RMSE is 0.0434m.The two models have a tendency to underestimate the ice thickness when the polarization is large.By comparison of the two models,there is high correlation between two models without water vapoured,and the result of 36GHz is slightly higher than 89GHz.In this paper,the method of extracting polynya from the ice thickness model which was constructed in this paper is proposed,and the threshold value of ice thickness for extracting was determined from the perspective of heat flux:0.2m.Using the ice-thickness threshold method and the concentration threshold method,the temporal-spatial changes of the polynya were analyzed on the changes of the daily,monthly,and annual scales,and it was found that the range of the polynya has a slight shrinking trend.In June 2007,an extreme value appeared in the scope of the polynya,and the expansion of the RISP is unlikely to be affected by temperature.Because in Chapter 3,the main factor affecting the ice thickness within the ice lake was analyzed as the wind speed,the method for extracting the range of the two polynya was analyzed in terms of wind speed.It was found that the scope of the RISP extracted by the ice thickness was more sensitive to wind speed.In summary,the AMSR-E thin ice thickness algorithm which built for Ross Ice Shelf Polynya works well in deriving ice thickness up to 0.2m.Comparing to the ice concentration method,the derived extent of our method is larger.