海洋白帽的高空间分辨率光学遥感估算分析

Estimation of oceanic whitecaps using high spatial-resolution optical remote sensing

海洋白帽是海面波浪破碎的直观表现,反映了海面风速等海洋动力环境情况,是海气交换的重要通道,也是海洋研究关注的重点目标之一。海洋白帽对入射光具有较强的反射散射作用,随着海洋光学遥感空间分辨率的提高,海洋白帽能被有效识别与区分。在空间分辨率为10 m的哨兵2号卫星(Sentinel-2)多光谱成像仪MSI(Multi-Spectral Instrument)数据的近红外波段(842 nm)中,不同风速下的海洋白帽影像特征清晰;基于海洋白帽对入射光的辐射传输过程与图像特征,本研究利用一种白帽覆盖率的光学遥感估算方法,实现了MSI弱耀光反射图像上的白帽覆盖率估算。研究表明:(1)基于MSI数据估算的海面白帽覆盖率,在无其他海洋环境动力的影响下,能进一步用于海面风速的遥感反演,且具有较高的空间分辨率;(2)其他海洋动力环境要素,能对海面波浪的破碎程度产生调制作用,因此海洋白帽的识别提取也有助于海洋环境动力遥感监测;(3)海面强耀光反射会对海洋白帽遥感带来较大的干扰,导致海面白帽覆盖率的高估,有效剔除耀光反射是提高海洋白帽识别提取精度的重要前提。对中国首个海洋水色业务卫星星座(海洋一号C/D卫星,HY-1C/D)获取的海岸带成像仪CZI(Coastal Zone Imager)数据开展分析,结果表明,在无其他海洋环境动力的调制下,当风速约高于9 m/s时,海洋白帽能为CZI载荷探测并区分,白帽覆盖率随着风速的增加而增加;CZI瑞利校正反射率图像的统计分析还表明,白帽反射率会比背景海水反射率高约5.8%—8.3%;对CZI图像中海洋白帽的深入研究,不仅为大气校正提供数据参考,也有助于光学遥感对海洋环境的深入理解与应用开发。随着海面白帽光学遥感的不断深入,今后将能为海面风场反演、海洋环境动力监测、海气交换评估提供新的技术方法参考。

Oceanic whitecaps, generated from wind-wave breaking process, are the medium of air-sea exchange and the indicator of sea surface state. Due to the intense reflection and scattering of incident light, whitecaps can be effectively recognized and distinguished in insitu photos or videos. Whitecap coverage(W), defined as the proportion of ocean surface covered by whitecaps, is an important parameter for the quantification of whitecaps. Of course, oceanic whitecaps can also be discriminated in high spatial resolution optical remote sensing images, such as Sentinel-2 MSI and Landsat-8 OLI data. This can provide a new research direction in marine environment observation, and may be further used for wind speed monitoring. However, how to estimate oceanic whitecap coverage from these optical remote sensing images is still a challenge. In this study, the formula of whitecap coverage is obtained by converting the form of radiative equation related with the constant and image reflectance of Sentinel-2 MSI data under some assumptions. The background signal of seawater and atmosphere was eliminated by data optimization, and the signal of whitecaps can be distinguished using regional filtering method. Then, whitecap coverage can be estimated. The identification and estimation results indicate that whitecap coverage derived from Sentinel-2 MSI images are consistent with previous studies using in situ observations in the order of magnitude, and can invert sea surface wind speed using a statistical model. Coarse spatial resolution wind speed images converted from MSI inversion were validated by ERA5 wind speed products from European Centre for Medium-Range Weather Forecasts(ECMWF). In addition, whitecap coverage can imply the modulation of other marine environmental dynamic factors(e.g., water mass, ocean fronts, ocean eddies and internal waves). Moreover, it should be noted that sunglint reflection is a non-negligible issue for optical remote sensing of oceanic whitecaps whose signal should be effectively eliminated.Examining the MSI-derived results to HY1-C/D CZI, it indicates that whitecaps can be identified in CZI images when wind speed is greater than 9 m/s, and the reflectance difference between whitecaps and background seawater is 5.8%—8.3%. We hope the above results can be used to improve the accuracy of atmospheric correction, and provide new reference for using high spatial resolution optical remote sensing in sea surface wind speed estimation and marine environmental dynamic factors monitoring. This will hopefully expand the research and application fields of ocean color remote sensing.