一种使用卫星反射信号探测海冰分布新方法

A New Sea Ice Distribution Detection Method Using GNSS Reflected Signals

利用TechDemoSat-1低轨道卫星采集的卫星反射信号生成的时延多普勒图(delay-doppler map,DDM)探测北极格陵兰岛(Greenland)地区的海冰分布。由于卫星散射信号在海冰和海水上分别遵循相干散射和漫散射模型,与海水相比,海冰的时延多普勒图扩散较少,所以从海水到海冰的过渡会导致时延多普勒图中高于一定能量值的像素数量减小。提出了一种基于相邻DDM像素数量比值的检测方法,通过采用一定阈值来区分海冰和海水的时延多普勒图,用于分析格陵兰岛地区海冰的分布情况。还利用多期数据分析了格陵兰岛地区海冰分布随时间的变化。通过与美国国家冰雪数据中心提供的数据比对,利用数学统计方法对本文提出的方法性能进行了评估。结果显示,该方法的判断成功率可以达到98.76%~99.21%,整体评估成功率约为99.09%。

Global navigation satellite system reflectometry(GNSS-R)delay-doppler map(DDM)obtained with TechDemoSat-1(TDS-1)is used to detect ice presence over Greenland region of Arctic.DDM of sea ice shows less spreading than those of open water as the scattered GNSS signals follow the coherenty scattering model and diffuse scattering model over ice and water respectively.The transition from seawater to sea ice could lead to decrease of the pixel number of DDM.A pixel number ratio of adjacent DDM based detection scheme is proposed through employing a threshold method to distinguish sea ice and sea water,and to analyze sea ice distribution of Greenland region.Sea ice distribution of Greenland with time is analyzed through using multi-period data from TDS-1.The performance of the proposed method is assessed with a mathematical statistical approach through comparing with ground-truth sea ice data from the National Snow and Ice Data Center,USA.The effectiveness of this approach is validated with a probability of detection up to 98.76%-99.21%,and the overall probability of detection is 99.09%.