Monitoring and quantifying suspended sediment concentrations (SSC) in estuaries such as the Pearl River Estuary (PRE) provide crucial information for environmental processes, hydrological infrastructure, and navigation. Traditional SSC mapping based on in situ investigations lacks the spatial coverage necessitated by detailed analysis. In this study, based on in situ spectral data and SSC measurements, we developed models to quantify SSC based on Landsat TM/OLI and Sentinel-2 imagery. The models were then used to map the SSC distribution in the PRE for the period 1986–2020. The results indicated that SSC in the entire PRE displayed an overall decreasing trend over the past 35 years, but an abnormal increase was also observed in shallow waters near Hengmen. Geographically, the concentration along the west coast is higher than on the east coast, and the central part of Lingding Bay underwent a slight increase. The decreased sediment output from the Pearl River Basin is the major cause of the decreasing changes in SSC. However, the natural factors and human activities such as channel dredging and the construction of artificial facilities can also disturb the spatial distribution of SSC. Regions with a significant decrease in SSC should have special concerns about retaining the PRE’s health and sustainability.

Mulberry-dyke-fish pond ecosystems are a representative traditional eco-agriculture in the Guangdong–Hong Kong–Macao Greater Bay Area (GBA). Investigations about the changes in the systems and their relevant water environments under the background of rapid urbanization can provide valuable information to formulate sustainable protection and development strategies. Using the Landsat images obtained after 1986, this study combined supervised classification and visual interpretation approaches, as well as water intensity index and synthesized index to identify the spatial patterns of changes in the ponds in the GBA over the past 40 years. The results indicated that during the period 1986–2013, the total surface area of the ponds in the GBA increased significantly and peaked in 2013 with a total increase of 84.63%; After that, the total surface area showed a downward trend with a total decrease of approximately 31.34%. The year of 2013 was identified as the milestone of the changes. The results proved that human activities have continuously influenced the spatial distribution and size of fish ponds in the past 40 years. The fish ponds had transformed from near-natural ponds with different sizes and a near-natural random distribution in the early stage into an artificial distribution and an artificial shape. Land use changes, industrial transfer, Government guidance and financial motives were the major drivers to the changes. If no effective measures are taken, this shrinking trend in the ponds will remain in the future.

Chlorophyll-a (Chla) is an important parameter for water quality. For remote sensing-based methods for the measurement of Chla, in-situ hyperspectral data is crucial for building retrieval models. In the Pearl River Estuary, we used 61 groups of in-situ hyperspectral data and corresponding Chla concentrations collected in July and December 2020 to build a Chla retrieval model that takes the two different seasons and the turbidity of water into consideration. The following results were obtained. (1) Based on the pre-processing techniques for hyperspectral data, it was shown that the first-derivative of 680 nm is the optimal band for the estimation of Chla in the Pearl River Estuary, with R² > 0.8 and MAPE of 26.03%. (2) To overcome the spectral resolution problem in satellite image retrieval, based on the simulated reflectance from the Sentinel-2 satellite and the shape of the discrete spectral curve, we constructed a multispectral model using the slope difference index method, which reached a R² of 0.78 and MAPE of 35.21% and can integrate the summer and winter data. (3) The slope difference method applied to the Sentinel-2 image shows better performance than the red-NIR ratio method. Therefore, the method proposed in this paper is practicable for Chla monitoring of coastal waters based on both in-situ data and images.

The Western Pacific Coasts (WPC) are the outlets of many large Asian rivers. In recent years, the interplay of climate changes and human activities has persistently altered the suspended sediment concentrations (SSC) in the WPC, triggering substantial shifts in coastal ecosystems. However, the scarcity of coastal observation stations hampered comprehensive investigations at large scales. This study employed three representative SSC retrieval models and utilized Landsat images acquired from 1990 to 2020 to estimate the SSC in the WPC with a focused endeavor to dissect the intricate spatial and temporal variability of SSC in the region. The findings revealed the following insights: (1) The outcomes derived from the three distinct SSC models consistently manifested a discernible decreasing pattern in SSC changes over the past three decades across all six major estuaries (Liao River Estuary, Yellow River Estuary, Yangtze River Estuary, Hangzhou Bay, Pearl River Estuary, and Mekong River Estuary). (2) The seasonal attributes of the six major estuaries differed, primarily due to distinct dominant influencing factors like precipitation, upstream sediment load, wind, and tides. (3) Collectively, SSC tends to be relatively higher in the Yangtze River Estuary, Hangzhou Bay, and Yellow River Estuary, while the Pearl River and Mekong River Estuaries exhibit relatively lower levels. Notably, the SSC exhibited distinct spatial traits along the coastlines of different estuaries. (4) SSC in the non-estuarine regions along the WPC, a similar significant declining trend in SSC is observed as in the estuaries, albeit the rate of decline generally appeared to be less pronounced. Furthermore, regions with faster rates of SSC reduction are typically concentrated near major estuaries in the northern part of the Coasts. The decline in estuarine SSC plays an important role in the overall decrease in SSC across the WPC. These study outcomes held substantial significance for advancing the stability and sustainable evolution of the WPC.

Machine learning (ML) algorithms can accurately extract quantitative patterns from datasets without any prior knowledge or prerequisites, playing an increasingly important role in areas such as information extraction from remotely sensed data. This study attempts to use several ML algorithms, including Back Propagation Network (BPN), Support Vector Machine (SVM), Random Forest (RF), and Extreme Gradient Boosting (XGBoost), to build global oceanic chlorophyll-a (Chl-a) concentration inversion models for MODIS/Aqua (MODISA). The models are developed using the SeaWiFS Bio-optical Archive and Storage System (SeaBASS) dataset, which contains matching data of in situ Chl-a concentration measurements and MODISA remote sensing reflectance ( R _rs ) for various water color bands, spanning from 2002 to 2017. This study includes the following processing steps. Firstly, the SeaBASS dataset is randomly divided into training dataset and testing dataset. Secondly, the Chl-a concentration inversion models are established using the training dataset. The third step is to use the testing dataset to evaluate the established models. Finally, the inversion experiment is carried out with MODISA data. Experimental results show that all the ML algorithms achieve higher inversion accuracy compared to the traditional ocean color index (OCI) algorithm. Among them, the SVR model demonstrates the highest performance. In addition, to address the challenge of multi-dimensional parameter optimization in ML algorithms, this study introduces the differential evolution (DE) algorithm and constructs a DE-based SVM model (DE-SVM). Compared with the standard SVM model, the mean deviation, mean absolute deviation and root mean square error of the inverted Chl-a concentration from the DE-SVM model are significantly reduced, and the coefficient of determination ( R ² ) is increased from 0.87 to 0.926, indicating that the DE-SVM model can further improve the inversion accuracy of Chl-a concentration, and has great potential in satellite remote sensing inversion of Chl-a concentration.

Rain affects the wind measurement accuracy of the Ku-band spaceborne scatterometer. In order to improve the quality of the retrieved wind field, it is necessary to identify and flag rain-contaminated data. In this study, an HY-2A scatterometer is used to study rain identification. In addition to the conventional parameters, such as the retrieved wind speed, the wind direction relative to the along-track direction, and the normalized beam difference, the experiment expands the mean deviation of the backscattering coefficient, the beam difference between fore and aft, and the node number of the wind vector cell (WVC) as the sensitive parameters according to the microwave scattering characteristics of rain and the actual measurement situation of the HY-2A. Furthermore, a rain identification model for HY2 (HY2RRM) with the K-Nearest Neighborhood (KNN) algorithm was built. After several tests, the accuracy of the selected HY2RRM approach is found to about 88%, and about 70% of rain-contaminated data can be accurately identified. The research results are helpful for better understanding the characteristics of microwave backscattering and provide a possible way to further improve the wind field retrieval accuracy of the HY-2A scatterometer and other Ku-band scatterometers.