基于LightGBM的全球海洋次表层温盐遥感预测

随着卫星遥感技术的发展，越来越多的卫星观测数据被应用于预测海洋内部温盐结构信息，而如何有效提高海洋内部温盐信息的预测精度仍是一个挑战。本文应用LightGBM算法结合随机森林算法构建全球海洋次表层（0—1000 m）温度异常（STA）与盐度异常（SSA）的预测模型，模型使用海表卫星观测数据（海表高度异常（SSHA）、海表温度异常（SSTA）、海表盐度异常（SSSA）和海表风场异常水平和垂直分量（USSWA、VSSWA），结合经纬度信息（LON、LAT）作为预测变量，使用Argo次表层温盐数据作为模型训练与测试标记。本文使用五参数模型（SSTA、SSHA、SSSA、USSWA、VSSWA）、带纬度六参数模型（LAT、SSTA、SSHA、SSSA、USSWA、VSSWA）、带经度六参数模型（LON、SSTA、SSHA、SSSA、USSWA、VSSWA）与带经纬度七参数模型（LON、LAT、SSTA、SSHA、SSSA、USSWA、VSSWA）来着重分析LON与LAT在STA、SSA遥感预测中发挥的作用。结果表明LON与LAT在STA、SSA各自预测中发挥不同的作用。在单时相和时序预测STA中LON与LAT对模型的贡献随着深度的增加逐渐增大，而在单时相和时序预测SSA中LON与LAT对不同深度预测模型始终保持较大的贡献。在单时相预测STA与SSA中LON较LAT对模型贡献更大，而在时序预测STA与SSA中LAT较LON对模型贡献更大。经纬度信息是全球海洋次表层温盐机器学习预测的重要参数，可以提高模型的预测精度。同时，LightGBM较随机森林在预测海洋次表层温盐异常时精度更高鲁棒性更强。     

全球海洋次表层温度异常遥感反演的季节时空变化特征

卫星遥感反演海洋内部多时相、大尺度热力结构信息对于了解海洋内部复杂、多维的动力过程有重要意义。本文采用随机森林回归模型，利用卫星遥感观测的海表参量（海表高度异常（SSHA）、海表温度异常（SSTA）、海表盐度异常（SSSA）和海表风场异常（SSWA）），反演不同季节、不同深度层位（1000 m深度以上）的海洋次表层温度异常（STA），并用Argo实测数据作精度验证，采用均方根误差（RMSE）、归一化均方根误差（NRMSE）以及决定系数（R²）评价模型在全球及洋盆尺度上的反演精度。结果显示，全球海洋16个深度层位的平均R²在春、夏、秋、冬四季分别为0.53、0.60、0.54、0.66，NRMSE分别为0.051、0.031、0.043、0.044。随着季节的变化，模型反演性能有所波动。模型在印度洋的反演效果最佳，不同季节、不同深度层位上的平均R²和RMSE分别为0.71和0.18 ℃，而大西洋的反演精度最低，平均R²和RMSE分别为0.46和0.25 ℃。研究表明随机森林模型适用于全球不同季节的STA遥感反演，且在不同洋盆上均有较好的反演效果；不同季节上，上层STA有明显变化信号，空间异质性显著，但300 m以深，STA信号较弱且基本不随季节变化。本研究可为长时序、大尺度海洋内部参量信息遥感反演与重建提供依据，有助于进一步发展深海遥感方法。     

海表温度对SMOS盐度遥感反演精度的影响

利用星载微波辐射计对全球海表盐度的卫星遥感探测,其精度会受到多种环境因子的影响。采用广义加性模型GAM和偏最小二乘法PLS分析了水温对海表盐度遥感反演精度的影响,同时,利用ARGO观测数据对SMOS卫星反演的赤道太平洋和西北太平洋海表盐度进行精度检验。结果表明,水温对海表盐度反演精度具有显著影响,且Stokes矢量第一参数(总辐亮度)是海表盐度反演的最佳亮温参数。在平均水温约16℃时的均方误差约为0.9 psu,23℃水温下的均方误差约为0.7 psu,30℃水温下的均方误差约为0.4 psu,即高水温下盐度反演精度相对较高。     

渤海环境遥感监测业务化应用技术

重点介绍利用遥感卫星数据进行海表温度、海面风场、海洋水色和大气参数等海洋环境要素反演技术以及数据产品真实性检验方法, 描述了渤海遥感监测数据产品和精度指标, 对于建立区域性海洋遥感监测业务化系统具有重要的科学价值。     

中国海洋卫星及应用进展

中国十分重视海洋遥感及其监测技术的发展,初步形成了具有优势互补的海洋遥感观测体系,并发挥了显著的经济和社会效益。其中,海洋一号(HY-1A/B)卫星已经广泛应用于中国海温预报业务系统、冬季海冰业务监测、夏季赤潮和绿潮监测、海岸带动态变化监测、近岸海水水质监测和渔业遥感监测等方面。海洋二号(HY-2A)卫星不仅填补了中国海洋动力环境卫星遥感的空白,也是目前国际上唯一在轨运行的集主被动微波遥感器于一身的综合型海洋动力环境卫星,具备同时获取风场、有效波高、海面高度和海面温度的能力。通过卫星获得的数据提高了中国海洋环境监测与灾害性海况预报的水平,为国民经济建设和国防建设、海洋科学研究、全球变化研究等提供了可靠的遥感数据,同时还在国际对地观测体系中发挥了重要作用,受到国内外用户的高度认可。海洋一号和海洋二号卫星系列为中国建立完善的海洋环境立体监测体系奠定了坚实基础。根据国家发展和"一带一路"建设的实施,在加快建设海洋强国、维护海洋权益和加快发展海洋经济的进程中对海洋遥感的发展也进一步提出了更高的要求和更紧迫的需求。为此,紧紧围绕国家海洋强国战略需求,在《国家民用空间基础设施中长期发展规划(2015年—2025年)》中专门规划了海洋观测卫星系列,服务于中国的海洋资源开发、环境保护、防灾减灾、权益维护、海域使用管理、海岛海岸带调查和极地大洋考察等方面,同时兼顾陆地和大气观测领域的需求。在充分继承已有HY-1A/B、HY-2A、高分三号(GF-3)和中法海洋卫星(CFOSAT)成功研制经验和应用成果的基础上,发展多种光学和微波遥感技术,建设新一代的海洋水色卫星和海洋动力环境卫星,具备卫星组网观测能力;发展海洋监视监测卫星,构建优势互补的海洋卫星综合观测体系。通过空间基础设施的建设,海洋遥感卫星必将在建设海洋强国的进程中发挥出重要作用。     

HY-1C/D海洋水色卫星产品体系及其典型应用

海洋一号C星(HY-1C)和海洋一号D星(HY-1D)是中国首次对地观测组网的海洋水色业务化应用卫星,其上搭载了海洋水色水温扫描仪COCTS (Chinese Ocean Color and Temperature Scanner)、海岸带成像仪CZI(Coastal Zone Imager)、紫外成像仪UVI (Ultra-Violet Imager)、星上定标光谱仪SCS (Satellite-based Calibration Spectrometer)和船舶自动识别系统AIS (Automatic Identification System),实现了全球海洋水色要素每天2次、海表温度每天4次的全球覆盖观测,中国近海及海岸带区域每3天2次的50 m高分辨率观测。本文介绍了HY-1C/D卫星及其载荷性能、产品处理流程、产品分级与分发等相关信息,HY-1C/D卫星产品体系完整且数据产品处理与分发高效。同时阐述了HY-1C/D卫星的叶绿素a浓度估算和海表温度反演产品,以及HY-1C/D卫星数据在浒苔绿潮、海冰/极冰、近海养殖、内陆水体和台风云图等方面的典型监测应用且具有较好的产品质量、...     

近海多种卫星测高联合数据处理技术

海域海况复杂多变 ,潮汐不但受外海能量输入的控制 ,而且在复杂的海岸线、浅海海底和内陆河流运输的作用下变得异常复杂。卫星测高由于受复杂的海洋动力环境和陆地反射的影响 ,数据质量普遍较深海差。将多种卫星测高数据联合处理 ,可以大大提高近海海域平均海面高的精度与分辨率 ,增强测高卫星监测近海复杂动力现象与反演近海复杂动力机制的能力。本文讨论近海多种卫星联合数据处理的技术与方法 ,分别从提取海平面稳态和时变信息的角度较系统地研究了多种卫星测高联合数据处理方法 ,通过提取不同测高卫星海平面观测数据中与时间无关的系统偏差 ,建立多种卫星测高数据的海平面时变基准 ,从而将多种测高卫星海面监测数据融合到一个动力系统中。大大提高测高卫星海平面监测的时空分辨率 ,为联合多种卫星测高数据在大地测量与近海海洋动力学研究中的应用创造基本条件。     

中国近海绿潮生物量的卫星光学遥感估算

生物量是精准量化海洋大型漂浮藻类的关键参数,是反映海洋生态环境变化的有效指标。光学遥感卫星能为绿潮的精细化监测与评估提供数据支持,实现绿潮的精准识别与量化估算。针对中国海洋一号C/D卫星（HY-1C/D）海岸带成像仪CZI (Coastal Zone Imager)、美国中分辨率成像光谱仪MODIS (Moderate-resolution Imaging Spectroradiometer)、欧洲空间局哨兵2号卫星多光谱成像仪MSI(Multi Spectral Instrument)等光学遥感数据特点,基于绿潮生物量变化模拟与观测验证数据,本研究提出了适用于不同光学卫星数据的绿潮生物量估算模型与计算方法,开展了中国近海绿潮生物量光学遥感估算与交叉验证。结果表明：相较于绿潮像元面积和覆盖面积,绿潮生物量估算结果的不确定性最小,该参数能有效减少面积参数所内含的尺度效应差异,能更准确地用于海洋绿潮的量化与评估。此外,基于CZI和MODIS数据开展2021年中国近海绿潮生物量协同监测应用,有效提高了绿潮生物量监测的精度,详细量化了2021年中国近海绿潮生物量的年内变化,展现了绿潮生物量的精...     

河北省卫星遥感海洋应用平台建设研究

随着卫星遥感技术在海洋监测中的作用日益凸显,为了加快地方海洋监测能力建设,构建省级卫星遥感海洋应用平台十分必要。介绍了河北省卫星遥感海洋应用平台的总体设计思路、总体架构、系统建设和应用情况。该平台将国家遥感数据分发单位、地方业务化监测单位、现场观测单位有机联合在一起,形成了自卫星遥感数据与现场观测数据的收集、数据处理、产品生产、数据管理、成果发布至精度评价的全业务闭环流程,实现了海洋环境常规监测业务系统后台全自动化运行,极大程度地提升了河北省海洋遥感监测能力和服务能力,为地方海洋遥感监测平台的建设提供参考。     

随机森林反演卫星遥感海表面盐度研究

海表面盐度是描述海洋状态、模拟海洋循环和检测气候变化的重要指标,对海洋研究意义重大。土壤湿度与海水盐度(soil moisture and ocean salinity,SMOS)卫星为全球海表面盐度分析提供了重要数据,但其整体精度尚未达到预期要求。基于海表面盐度遥感机理和SMOS卫星盐度反演基础理论,选取海表面盐度敏感因子,建立随机森林(random forest, RF)模型,并基于网格搜索算法优化模型参数,辅助提高SMOS卫星产品精度。其中基础RF得到的海表面盐度与Argo (array for real-time geostrophic oceanography)数据之间的平均绝对误差为0.08,均方根误差为0.15。而经网格搜索算法优化后的随机森林模型精度稍有所提升,其与Argo数据的绝对平均误差为0.08,均方根误差仅为0.14,且误差分布范围较小。两种模型均显著优于SMOS卫星Level 2级盐度产品。从机器学习与统计学理论出发,建立的高精度、高适应性的随机森林海表面盐度反演模型大幅提高了盐度精度,能够为相关海洋研究提供数据支撑。     

The operational method of filling information gaps in satellite imagery using numerical models

Occurrence of cloud cover over remotely sensed area is a significant limitation in the ocean colour and infra-red remote sensing applications, especially when operational use of such a data is considered. A method for the reconstruction of missing data in remote sensing images has been proposed. It is based on complementing satellite data with the corresponding information from other sources of data, in our tested case it was the ecohydrodynamic model. The method solves the problem the presence of a cloud cover also during an extended period. Unlike in many other similar methods, emphasis has been put on retaining remotely sensed information to a high degree and preserving local phenomena that are usually difficult to capture by other methods than satellite remote sensing. The method has been tested on the Baltic Sea. Sea surface temperature and chlorophyll a concentration estimated from satellite data, ecohydrodynamic models and merged product were compared with in situ data. The algorithm was optimized for the two parameters that are crucial for e.g. creating algae bloom forecasts. The root mean square error (RMSE) of the final product of sea surface temperature was 0.73 °C, whereas of the input satellite images 1.26 °C or 1.33 °C and of model maps 0.89 °C. The error factor of chlorophyll a concentration product was 1.8 mg m⁻³, in comparison to 2.55 mg m⁻³ for satellite input source and 2.28 mg m⁻³ for the model one. The results show that the proposed method well utilizes advantages of both satellite and numerical simulation data sources, at the same time reducing the errors of estimation of merged parameters compared to similar errors for both primary sources. It would be a valuable component of fuzzy logic and rule-based HABs prediction.     

Satellite mapping of Baltic Sea Secchi depth with multiple regression models

Secchi depth is a measure of water transparency. In the Baltic Sea region, Secchi depth maps are used to assess eutrophication and as input for habitat models. Due to their spatial and temporal coverage, satellite data would be the most suitable data source for such maps. But the Baltic Sea’s optical properties are so different from the open ocean that globally calibrated standard models suffer from large errors. Regional predictive models that take the Baltic Sea’s special optical properties into account are thus needed. This paper tests how accurately generalized linear models (GLMs) and generalized additive models (GAMs) with MODIS/Aqua and auxiliary data as inputs can predict Secchi depth at a regional scale. It uses cross-validation to test the prediction accuracy of hundreds of GAMs and GLMs with up to 5 input variables. A GAM with 3 input variables (chlorophyll a, remote sensing reflectance at 678 nm, and long-term mean salinity) made the most accurate predictions. Tested against field observations not used for model selection and calibration, the best model’s mean absolute error (MAE) for daily predictions was 1.07 m (22%), more than 50% lower than for other publicly available Baltic Sea Secchi depth maps. The MAE for predicting monthly averages was 0.86 m (15%). Thus, the proposed model selection process was able to find a regional model with good prediction accuracy. It could be useful to find predictive models for environmental variables other than Secchi depth, using data from other satellite sensors, and for other regions where non-standard remote sensing models are needed for prediction and mapping. Annual and monthly mean Secchi depth maps for 2003–2012 come with this paper as Supplementary materials.     

NOAA/AVHRR sea surface temperature accuracy in the East/Japan Sea

Sea surface temperature (SST) retrieved from Advanced Very High Resolution Radiometer (AVHRR) onboard National Oceanic and Atmospheric Administration (NOAA) polar orbiting environmental satellites were validated in the East/Japan Sea (EJS) using surface drifter measurements as ground truths from 2005 to 2010. Overall, the root-mean-square (rms) errors of multichannel SSTs (MCSSTs) and non-linear SSTs (NLSSTs) using global SST coefficients were approximately 0.85°C and 0.80°C, respectively. An analysis of the SST errors (satellite – drifter) revealed a dependence on the amount of atmospheric moisture. In addition, satellite-derived SSTs tended to be related to wind speeds, particularly during the night. The SST errors also demonstrated diurnal variations with relatively higher rms from 0.80°C to 1.00°C during the night than the day, with a small rms of about 0.50°C. Bias also exhibited reasonable diurnal differences, showing small biases during the daytime. Although a satellite zenith angle has been considered in the global SST coefficients, its effect on the SST errors still remained in case of the EJS. Given the diverse use of SST data, the continuous validation and understanding of the characteristic errors of satellite SSTs should be conducted based on extensive in-situ temperature measurements in the global ocean as well as local seas.     

SVR在平均海面高格网化中研究

研究了利用SVR进行平均海面高格网化的方法,并结合平均海面高数据的特点,确定了格网化半径;利用Surfer11软件中几种数据插值方法,对平均海面高进行了格网化处理,并对结果进行交叉验证;将SVR格网化方法所得平均海面高与surfer11格网化所得平均海面高进行了对比,结果表明利用SVR进行平均海面高格网化是切实可行的。     

Locating oceanic Tuna resources in the Eastern Arabian sea using remote sensing

Tuna fishery resources are currently under exploited. The resource potential of tunas in the Indian Exclusive Economic Zone (EEZ) beyond 50 m depths is around 2.09 lakh tonnes as estimated by Fishery Survey of India. The distribution and availability of the tuna are governed by environmental factors like temperature, thermocline depth, availability of prey, visibility etc. Remote sensing provides synoptic information on productivity in terms of chlorophyll and Sea Surface Temperature (SST). In the present paper, satellite remote sensing data from Indian Remote Sensing Satellite IRS- P4 Ocean Colour Monitor (OCM) sensor for chlorophyll-a and diffuse attenuation coefficient (K) and National Oceanic and Atmospheric Administration (NOAA) - Advanced Very High Resolution Radiometer (AVHRR) sensor data for sea surface temperature were analysed and correlated within situ catch data of oceanic tunas, Skipjack(Katsuwonus pelamis) and Yellowfin tuna(Thunnus albacares), off Maharashtra coast. Higher catches were found to be associated with moderate to good primary productivity and in the vicinity of thermal fronts. Relationship between Hooking rate and SST has shown that SST of 28–30°C range is optimum for skipjack and 28–31°C for yellowfin tuna. Besides satellite derived chlorophyll and SST for identification of potential tuna fishing zones, role of diffuse attenuation coefficient (K) for visibility factor is also discussed.     

Comprehensive evaluation of soil moisture retrieval models under different crop cover types using C-band synthetic aperture radar data

In the present study, random forest regression (RFR), support vector regression (SVR) and artificial neural network regression (ANNR) models were evaluated for the retrieval of soil moisture covered by winter wheat, barley and corn crops. SVR with radial basis function kernel was provided the highest adj. R² (0.95) value for soil moisture retrieval covered by the wheat crop at VV polarization. However, RFR provided the adj. R² (0.94) value for soil moisture retrieval covered by barley crop at VV polarization using Sentinel-1A satellite data. The adj. R² (0.94) values were found for the soil moisture covered by corn crop at VV polarization using RFR, SVR linear and radial basis function kernels. The least performance was reported using ANNR model for almost all the crops under investigation. The soil moisture retrieval outcomes were found better at VV polarization in comparison to VH polarization using three different models.     

Upscaling plot-scale soil respiration in winter wheat and summer maize rotation croplands in Julu County,North China

Soil respiration (R_s) data from 45 plots were used to estimate the spatial patterns of R_s during the peak growing seasons of winter wheat and summer maize in Julu County, North China, by combining satellite remote sensing data, field-measured data, and a support vector regression (SVR) model. The observed R_s values were well reproduced by the model at the plot scale, with a root-mean-square error (RMSE) of 0.31 μmol CO₂ m⁻² s⁻¹ and a coefficient of determination (R²) of 0.73. No significant difference was detected between the prediction accuracy of the SVR model for winter wheat and summer maize. With forcing from satellite remote sensing data and gridded soil property data, we used the SVR model to predict the spatial distributions of R_s during the peak growing seasons of winter wheat and summer maize rotation croplands in Julu County. The SVR model captured the spatial variations of R_s at the county scale. The satellite-derived enhanced vegetation index was found to be the most important input used to predict R_s. Removal of this variable caused an RMSE increase from 0.31 μmol CO₂ m⁻² s⁻¹ to 0.42 μmol CO₂ m⁻² s⁻¹. Soil properties such as soil organic carbon (SOC) content and soil bulk density (SBD) were the second most important factors. Their removal led to an RMSE increase from 0.31 μmol CO₂ m⁻² s⁻¹ to 0.37 μmol CO₂ m⁻² s⁻¹. The SVR model performed better than multiple regression in predicting spatial variations of R_s in winter wheat and summer maize rotation croplands, as shown by the comparison of the R² and RMSE values of the two algorithms. The spatial patterns of R_s are better captured using the SVR model than performing multiple regression, particularly for the relatively high and relatively low R_s values at the center and northeast study areas. Therefore, SVR shows promise for predicting spatial variations of R_s values on the basis of remotely sensed data and gridded soil property data at the county scale.     

海洋涡旋遥感：进展与挑战

海洋涡旋数量大、分布广、含能高、裹挟强,是研究物质循环、能量级联和圈层耦合的理想载体。对涡旋的全生命周期追踪观测成为21世纪以来海洋遥感领域最重要的进展之一,并引发了新一轮涡旋研究的热潮。本文从涡旋的温度异常、物质示踪、旋转流场和闭合拓扑等特征出发,简述了红外辐射计、可见光扫描仪、微波高度计、合成孔径雷达等遥感技术在涡旋观测中的机理和方法,重点阐述了卫星高度计涡旋识别与追踪算法及其在涡旋形态学、运动学和动力学中的应用。基于虚拟星座下的多参数遥感,介绍了涡旋在海洋、大气、生态等交叉学科领域的前沿应用和最新进展。指出当前涡旋遥感发展面临的亚中尺度、垂直结构、跨学科研究等3大挑战,展望了新一代遥感技术在未来海洋科学特别是涡旋海洋学研究中的应用前景。