【作者】 李晓明；

【导师】 贺明霞；

【作者基本信息】 中国海洋大学 ， 海洋信息探测与处理， 2010， 博士

【摘要】 欧空局卫星计划中设置的SAR波模式数据,与所有宽刈幅图像数据不同,是一种常开通数据,从中可以获取高时空分辨率和全球覆盖的海浪参数。ERS和ENVISAT波模式SAR数据对于全球海浪数值预报、海洋灾害预报、海气相互作用研究和全球气候变化研究无疑是十分重要的。事实上欧洲中期天气预报中心(ECMWF)已经把SAR波模式数据同化入海浪数值预报业务化运行。本文主要利用ENVISAT和ERS卫星ASAR/SAR波模式数据,深入研究了海浪方向谱和海浪参数的反演算法。利用2006年12月至2007年2月三个月获得的全球ASAR波模式数据及其时空匹配的浮标数据和模式结果,对非线性反演算法PARSA和准线性反演算法WVW进行印证比较。PARSA非线性反演算法是广泛应用的MPI模式的扩展和改进,需要提供第一猜测谱和SAR多视交叉谱作为输入。WVW准线性反演算法是欧空局采用的ASAR波模式二级产品海浪谱的算法,该算法的优点是不需要第一猜测谱,而仅需要SAR多视交叉谱作为输入。印证结果表明：PARSA反演算法可以提供完整的二维海浪方向谱,反演的有效波高与浮标结果、模式同化结果、预报结果具有较好的一致性。该算法对第一猜测谱的依赖性较强,限制SAR对海浪的观测成为独立的数据源,但有利于反演结果应用于海浪模式的数据同化以及改善海浪高频信息。WVW算法存在相当大的问题。反演的二维海浪谱不完整,局限于长波范围；反演的有效波高存在较大负偏差；存在25%的情况该算法无法得到二维海浪谱。由于PARSA反演算法依赖于第一猜测谱,而ESA的标准产品又存在很大误差。基于CWAVE_ERS模型,本文提出适用于ENVISAT ASAR波模式数据的经验模式CWAVE_ENV,利用该经验模式可以直接从ASAR波模式图像提取海浪参数,而且不依赖于第一猜测谱。通过1270个数据对与浮标实测结果进行比较,偏差为0.05米,均方根误差为0.72米,散射指数为24%。利用该经验模式从ASAR波模式数据反演海浪参数,与非线性反演算法结果以及高度计测量结果相当,接近浮标结果。利用ASAR/SAR波模式数据对北太平洋和北大西洋的高纬度风暴,南太平洋的交错浪,以及印度洋的涌浪进行个例分析。重点比较ASAR/SAR波模式数据各种反演算法在复杂海况和高海况条件下的结果,发现本文提出的CWAVE ENV模式在这些海况条件下仍具有高可靠性,对海洋风暴监测和极端海况预警颇具应用价值。更多还原

【Abstract】 Over the ocean, the SAR and ASAR instruments onboard the ESA’s ERS and ENVISAT satellite are operated in wave mode whenever no other operation is requested. In wave mode, SAR collects data globally with high spatial resolution to form small images of 10 km x 5 km size every 100 km along the satellite’s orbit. To retrieve ocean wave parameters from ASAR or SAR wave mode data with high quality is essential important. It can benefit the numerical wave model forecast and hindcast, observations and forecast for extreme ocean weather, as well as the global wave climate analysis. Assimilation of SAR wave mode observations into the numerical wave model has been carried out operationally in the European Center for Medium-Range Weather Forecast (ECMWF).The main research demonstrated in the thesis focus on ocean wave information retrieval from SAR and ASAR wave mode data, including validation and intercomparison for two schemes, i.e. non-linear Partition Rescaling and Shift Algorithm (PARSA) and quasi-linear WVW algorithm for retrieving two-dimensional ocean wave spectrum and integral wave parameters.The PARSA algorithm needs the SAR look cross spectra and first guess spectra taken from numerical wave model as the input. It is an extension of improvement for the MPI scheme.Validation results indicate that the PARSA algorithm can yield the full two-dimensional ocean wave spectrum. The retrieved integral wave parameters have good agreement with buoy measurements, the ECMWF reanalysis wave model and the DWD forecast wave model results.While it seems that the PARSA algorithm has a strong dependence with the first guess information and therefore makes the SAR measurements not independent. The quasi-linear algorithm WVW has the advantage without using the prior information from numerical wave model. However, the retrieved spectra are limited to the SAR cut-off wavenumber domain. The validation of integral wave parameters derived from WVW spectra shows that there is a significant underestimation for this algorithm. And the trends increase along with sea state.Further it seems that the artificial effect of SAR imaging of ocean waves is not resolved in the inversion. Another important issue found in the validation is that around 25% ASAR wave mode cross spectra cannot be converted successfully by using this algorithm.Based on the empirical model CWAVE_ERS developed for reprocessed ERS-2 SAR wave mode data, the CWAVE_ENV model is proposed in this thesis and implemented in the ASAR wave mode data. Using the same three months ASAR wave mode data and its collocated dataset, the empirical algorithm is validated. Validation results show that a good agreement is achieved by comparing the retrieved significant wave height results to buoy measurements using 1270 data pairs with bias of 0.05 m, RMSE of 0.72 m and scatter index of 24%. This gives the confidence that CWAVE_ENV algorithm can be used for ASAR wave mode data to retrieve integral wave parameters without using any first prior information and its retrieval accuracy is comparable to non-linear inversion schemes and the measurements of Radar Altimeter.Four case studies, i.e.,Storms in the North Pacific and North Atlantic, cross sea in the Southern Pacific and extreme swell in the Indian Ocean, are analyzed by using SAR/ASAR wave mode data. The research interesting is addressed to evaluate different ocean wave retrieval algorithm in complex and extreme sea state.It is concluded that even in the extreme sea state, the CWAVE_ENV algorithm still can provide reliable sea state measurements, which can benefit the observation and forecast for the extreme sea state.更多还原