海洋通报
海洋通報
해양통보
MARINE SCIENCE BULLETIN
2015年
4期
428-439
,共12页
路泽廷%朱江%韩君%元慧慧
路澤廷%硃江%韓君%元慧慧
로택정%주강%한군%원혜혜
SMOS%卫星海表盐度%Argo%误差分析
SMOS%衛星海錶鹽度%Argo%誤差分析
SMOS%위성해표염도%Argo%오차분석
SMOS%sea surface salinity%Argo%error analysis
选取SMOS任务的2个海洋盐度专家中心(法国的CATDS-CECOS和西班牙的BEC)的5种经过再处理的新版SSS L3/4产品作为研究对象,以Argo浮标资料及WOA09资料作为参考标准,对其误差特征进行了细致的分析比较,为将其同化到海洋模式中以及用于其它海洋学的分析应用研究,提供必要的参考。主要结论如下:SMOS年平均海表盐度场与WOA09资料很接近,一些已知的重要的分布形势都有所体现;大洋中部误差较小,近陆误差大;热带误差较小,高纬地区误差较大;三大洋中太平洋均方根误差最小。随着时空分辨率的降低,SMOS SSS资料的均方根误差显著减小。检验的几种资料中, CATDS/CEC-OS处理制作的月平均海表盐度L3级产品误差最小,全球平均均方根误差(RMSE)为0.314;另外几种高分辨率产品中,除由BEC制作的简单加权平均产品均方根误差最大,全球平均0.543以外,其他3种资料的均方根误差量级相当,差异不太明显,全球平均的RMSE为0.3~0.4;BEC的两种分析产品总体上RMSE更小。
選取SMOS任務的2箇海洋鹽度專傢中心(法國的CATDS-CECOS和西班牙的BEC)的5種經過再處理的新版SSS L3/4產品作為研究對象,以Argo浮標資料及WOA09資料作為參攷標準,對其誤差特徵進行瞭細緻的分析比較,為將其同化到海洋模式中以及用于其它海洋學的分析應用研究,提供必要的參攷。主要結論如下:SMOS年平均海錶鹽度場與WOA09資料很接近,一些已知的重要的分佈形勢都有所體現;大洋中部誤差較小,近陸誤差大;熱帶誤差較小,高緯地區誤差較大;三大洋中太平洋均方根誤差最小。隨著時空分辨率的降低,SMOS SSS資料的均方根誤差顯著減小。檢驗的幾種資料中, CATDS/CEC-OS處理製作的月平均海錶鹽度L3級產品誤差最小,全毬平均均方根誤差(RMSE)為0.314;另外幾種高分辨率產品中,除由BEC製作的簡單加權平均產品均方根誤差最大,全毬平均0.543以外,其他3種資料的均方根誤差量級相噹,差異不太明顯,全毬平均的RMSE為0.3~0.4;BEC的兩種分析產品總體上RMSE更小。
선취SMOS임무적2개해양염도전가중심(법국적CATDS-CECOS화서반아적BEC)적5충경과재처리적신판SSS L3/4산품작위연구대상,이Argo부표자료급WOA09자료작위삼고표준,대기오차특정진행료세치적분석비교,위장기동화도해양모식중이급용우기타해양학적분석응용연구,제공필요적삼고。주요결론여하:SMOS년평균해표염도장여WOA09자료흔접근,일사이지적중요적분포형세도유소체현;대양중부오차교소,근륙오차대;열대오차교소,고위지구오차교대;삼대양중태평양균방근오차최소。수착시공분변솔적강저,SMOS SSS자료적균방근오차현저감소。검험적궤충자료중, CATDS/CEC-OS처리제작적월평균해표염도L3급산품오차최소,전구평균균방근오차(RMSE)위0.314;령외궤충고분변솔산품중,제유BEC제작적간단가권평균산품균방근오차최대,전구평균0.543이외,기타3충자료적균방근오차량급상당,차이불태명현,전구평균적RMSE위0.3~0.4;BEC적량충분석산품총체상RMSE경소。
In this paper, 5 kinds of SSS L3/4 products from the two SMOS ocean salinity research centers (CATDS-CECOS in France and BEC in Spain) are selected as our study objects, which have been reprocessed and released recently. The error characteristics of these SSS products are analyzed carefully, with the WOA09 climatology and Argo SSS data as a reference for the"truth"value. The results are helpful to the assimilation of the SMOS SSS data, as well as other applications of these data. The following conclusions are drawn:SMOS annual mean SSS field is quite similar to the WOA09 annual field, showing many known important characteristics in the SSS distribution. The error is large near the land while it is small at the interior of the ocean. The error is large in the tropics while it is small in the high latitudes. Among the 3 oceans (Pacific, Atlantic and In-dian), the RMSE in Pacific is the smallest. The RMSE of SMOS SSS data decreases remarkably with the drop of the temporal and spatial resolution. The global mean RMSE of the monthly L3 SSS product from CEC/CATDS is the smallest(0.314);while the RMSE of the L3 SSS product from BEC with the simple method of weighted mean is the largest (global mean 0.543); the other three are comparative (0.3~0.4), with the two BEC analysis products seemingly better.