Satellite technology is an efficient tool, which can provide valuable observations for coastal areas from space. Compared with conventional bathymetric surveying approaches, remote sensing-based shallow water bathymetry retrieval methods have been widely used in recent years. Various empirical models have been proposed for deriving bathymetry of coastal shallow water, and prior topographic information is required to construct models. Traditional studies tend to select a cloud-free remote sensing image to map the coastal shallow water topography. As a result, in addition to the selection of empirical models, the high-quality remote sensing image and accurate prior topographic data are also of importance. This study aims to propose a method for mapping coastal shallow water bathymetry from multitemporal remote sensing imagery. Here, Sentinel-2 imagery time series are composited to produce a clear image, which can effectively avoid the contamination of clouds, water turbidity and other noises. ICESat-2 lidar altimeter data that contain accurate underwater elevations are used to provide topographic information. Moreover, Sentinel-2-based multispectral information and ICESat-2-based topographic information are combined for the coastal bathymetry retrieval by five empirical models (i.e., linear band model, ratio band model, support vector machine, neural network, and random forest). This proposed method is tested in Dongsha Atoll in South China Sea, and achieve a good performance [training: root mean square error (RMSE): 0.97 m ± 0.76 m, mean absolute percentage error (MAPE): 4.07% ± 0.046%, R-square (R ² ): 0.90 ± 0.14; validation: RMSE: 1.22 m ± 0.43 m, MAPE: 5.43% ± 0.035%, R ² : 0.86 ± 0.089]. The comparison confirms that machine learning methods perform better than traditional methods, and the deep learning techniques can be further introduced in estimating shallow water bathymetry in the future, which is expected to achieve an excellent accuracy in bathymetry inversion.

New evidences provided that the tropical cyclone (TC) Linfa in 2015 induced looping path of Kuroshio invasion into the northeastern South China Sea (NESCS) through the northwestern Luzon Strait (LS), based on the in-situ measurements, satellite data and model output data. This TC-enhanced Kuroshio invasion with low nutrients and denser waters suppressed the TC "Wind Pump" induced upwelling and nutrients uptake, and therefore inhibited the Chlorophyll a concentration (Chl-a) increase from surface to ~50 m in the open ocean of the NESCS. The TC-induced Kuroshio invasion promoted the generation of the strong cyclonic eddy to its left side where weak Ekman Pumping Velocity was observed. This enhancing cyclonic eddy then dominated the nutrients uplift and increased the surface and subsurface (0-50 m) Chl-a through eddy pumping rather than Ekman Pumping. The TC-declined anti-cyclonic eddy, which shoaled the Mixed Layer Depth (MLD), benefited to the nutrient uptake through TC-induced upwelling and thereby increased the surface Chl-a and raised the Chl-a Maximum Layer (CML) to ~20 m over the southwestern LS. The temporal Chl-a variations were also influenced by TC intensities and biochemical processes. The air-sea heat budget analysis indicated that, the air-sea heat exchange contributed to nearly 80% of the sea surface cooling (SST cooling) over the northwestern LS with Kuroshio invasion, while eddy-induced upwelling dominated the SST cooling over the western LS, and the wind-driven upwelling (and mixing) controlled the SST cooling over the southwestern LS. These different formations of SST cooling then played important role in Chl-a variations. This study is the first case of TC "Wind Pump" induced Chl-a variations considering air-sea heat exchange, Kuroshio invasion and mesoscale eddies over LS, which would help to evaluate the influence of TCs over the other major heat transport arteries of the world ocean: The Gulf Stream area.

Phytoplankton blooms induced by the interaction between typhoons and oceanic mesoscale eddies in the South China Sea were investigated using the composite averaging method based on cyclone‐centric coordinates from nine years (2002–2010) of satellite observations. Increases in the chlorophyll (Chl) concentration following the passage of a typhoon differed greatly among the pre‐existing cyclonic eddy (PCE), pre‐existing anticyclonic eddy (PACE) and non‐eddy (NE) ocean conditions (with Chl increases of approximately 35%, 6% and 12%, respectively), but the magnitude of sea surface temperature (SST) cooling was similar under the three conditions (∼1°C). Under the same level of SST decreases, the greater increase in the Chl concentration in the PCE condition was attributed to the longer persistence of both SST cooling and nutrient input caused by the typhoon‐strengthened PCE (associated with upwelling), which stopped the re‐warming of the typhoon‐induced cold water.

Water transparency, measured with Secchi disk depth (SDD), is an important parameter for describing the optical properties of a water body. This study evaluates variations of SDD and related impact factors in the Bohai and Yellow Seas (BYS). Based on a new mechanistic model proposed by Lee et al. (2015) applied to MODIS remote sensing reflectance data, climatological SDD variation from 2003 to 2019 was estimated. The annual mean images showed an increasing trend from the coastal zone to the deep ocean. Lower values were found in the Bohai Sea (BHS), while higher values observed in the center of the southern Yellow Sea (SYS). Additionally, the entire sea has shown a decreasing temporal tend, with the variation rate lowest in the BHS at 0.003 m y⁻¹, and highest in the SYS at 0.015 m y⁻¹. However, the weak increasing trend that appeared since 2017 suggests that water quality seems to have improved. Further, it displayed seasonal patterns of low in winter and spring and high in summer and autumn. The empirical orthogonal function (EOF) analysis of SDD variations over the BYS, shows that the first SDD EOF mode is the highest, strongly correlated with total suspended matter. With the high correlation coefficients of chromophoric dissolved organic matter, it illustrates that the SDD variation is mainly dominated by the optical components in the seawater, although correlation with chlorophyll-a is the weakest. The second and third EOF modes show that photosynthetically available radiation, sea surface temperature, sea surface salinity, and wind speed are the main covariates that cause SDD changes. Water transparency evaluation on a long-term scale is essential for water quality monitoring and marine ecosystem protection.

The abundance of phytoplankton is generally estimated by measuring the chlorophyll-a concentration (C_chla), which is an important factor in photosynthesis and can be used to analyze the density and biomass of phytoplankton in the ecosystem. The band-ratio-based empirical or semi-analytical algorithms are operationally applied to retrieve C_chla in global oceans, which generally experience difficulties from the diversity of optical properties and the complexity of the radiative transfer equations in analytical analyses, respectively. With an attempt to develop an accurate C_chla retrieval model for the optically complex coastal and estuarine waters, this study aimed to explore the deep learning (DL) methods in satellite retrieval of C_chla. A two-stage convolutional neural network (CNN), named C_chla-Net, was proposed, which utilized the spectral information of remote sensing reflectances at MODIS/Aqua’s visible bands. In the first-stage phase, the C_chla-Net was pretrained by a set of remote sensing patches, in which the C_chla was generated from an existing model (OC3M). The pretrained results were than used as the initial values to refine the network with the synthetic oversampled in-situ dataset in the second-stage training phase. Using in-situ samples for training with the new initial values has a higher probability to reach the global optimum. The quantitative analyses showed that the two-stage training was more likely to achieve a global optimum in the optimization than the one-stage training. Matchups of the in-situ C_chla measurements were used to evaluate the retrieval models. Results showed that the proposed C_chla-Net produced obvious better performance than the empirical and semi-analytical algorithms, implying the DL method was more effective for optically complex waters with extremely high C_chla. This study provided an applicable method for remote sensing retrieval of C_chla, which should be helpful for studying the spatial distribution and temporal variability in the productive Pearl River estuary (PRE) waters.

A major task of ocean colour observations is to determine the distribution of phytoplankton primary production. At present, the global coverage of the sea surface chlorophyll concentration, sea surface temperature, photosynthetically available radiation (PAR) can nominally be achieved every 1 to 2 days with standard algorithms from satellite data. From these standard products, a variety of bio‐optical algorithms has been developed to estimate ocean primary productivity. In this communication, we have investigated the possibility of using a novel universal approximator‐support vector machine (SVM) as the nonlinear transfer function between ocean primary productivity and the information that can be retrieved from satellite data, including chlorophyll concentration, PAR, maximum carbon fixation rate and day length, which is the same as the vertically generalized production model (VPGM). The VGPM dataset was used to evaluate the proposed approach. The primary production algorithm round robin 2 (PPARR2) dataset was used to further compare the precision between the VGPM and the SVM model. The results suggest that the SVM model is more accurate than the VGPM. Using the SVM model to calculate the global ocean primary productivity, the result is 45.5 Pg C yr⁻¹, which is a little higher than the VGPM result.