Search

As an environmental parameter, the chlorophyll-a concentration (Chl-a) is essential for monitoring water quality and managing the marine ecosystem. However, current mainstream Chl-a inversion algorithms have limited accuracy and poor spatial and temporal generalization in Case II waters. In this study, we constructed a quantitative model for retrieving the spatial and temporal distribution of Chl-a in the Bohai–Yellow Sea area using Geostationary Ocean Color Imager (GOCI) spectral remote sensing reflectance ( R r s λ ) products. Firstly, the GOCI  R r s λ  correction model based on measured spectral data was proposed and evaluated. Then, the feature variables of the band combinations with the highest correlation with Chl-a were selected. Subsequently, Chl-a inversion models were developed using three empirical ocean color algorithms (OC4, OC5, and YOC) and four machine learning methods: BP neural network (BPNN), random forest (RF), AdaBoost, and support vector regression (SVR). The retrieval results showed that the machine learning methods were much more accurate than the empirical algorithms and that the RF model retrieved Chl-a with the best performance and the highest prediction accuracy, with a determination coefficient R2 of 0.916, a root mean square error (RMSE) of 0.212 mg · m−3, and a mean absolute percentage error (MAPE) of 14.27%. Finally, the Chl-a distribution in the Bohai–Yellow Sea using the selected RF model was derived and analyzed. Spatially, Chl-a was high in the Bohai Sea, including in Laizhou Bay, Bohai Bay, and Liaodong Bay, with a value higher than 4 mg · m−3. Chl-a in the Bohai Strait and northern Yellow Sea was relatively low, with a value of less than 3 mg · m−3. Temporally, the inversion results showed that Chl-a was considerably higher in winter and spring compared to autumn and summer. Diurnal variation retrieval effectively demonstrated GOCI's potential as a capable tool for monitoring intraday changes in chlorophyll-a concentrations. [ABSTRACT FROM AUTHOR]

To characterize a community-scale urban functional area using geo-tagged data and available land-use information, several supervised and semi-supervised models are presented and evaluated in Hong Kong for comparing their uncertainty, robustness and sensitivity. The following results are noted: (i) As the training set size grows, models' accuracies are improved, particularly for multi-layer perceptron (MLP) or random forest (RF). The graph convolutional network (GCN) (MLP or RF) model reveals top accuracy when the proportion of training samples is less (greater) than 10% of the total number of functional areas; (ii) With a large amount of training samples, MLP shows the highest prediction accuracy and good performances in cross-validation, but less stability on same training sets; (iii) With a small amount of training samples, GCN provides viable results, by incorporating the auxiliary information provided by the proposed semantic linkages, which is meaningful in real-world predictions; (iv) When the training samples are less than 10%, one should be cautious using MLP to test the optimal epoch for obtaining the best accuracy, due to its model overfitting problem. The above insights could support efficient and scalable urban functional area mapping, even with insufficient land-use information (e.g., covering only ~20% of Beijing in the case study). [ABSTRACT FROM AUTHOR]

Dust storms are of great importance to climate change, air quality, and human health. In this study, a complete application frame of integrating hybrid methods based on multi-source data is proposed for remote sensing monitoring and process analysis of dust storms. In the frame, horizontal spatial distribution of dust intensity can be mapped by optical remote sensing products such as aerosol optical depth (AOD) from MODIS; the vertical spatial distribution of dust intensity by LIDAR satellite remote sensing products such as AOD profile from CALIPSO; geostationary satellite remote sensing products such as Chinese Fengyun or Japanese Himawari can achieve high-frequency temporal distribution information of dust storms. More detailed process analysis of dust storms includes air quality analysis supported by particulate matter (PM) data from ground stations and the dust emission trace and transport pathways from HYSPLIT back trajectory driven by meteorological data from the Global Data Assimilation System (GDAS). The dust storm outbreak condition of the source location can be proved by precipitation data from the WMO and soil moisture data from remote sensing products, which can be used to verify the deduced emission trace from HYSPLIT. The proposed application frame of integrating hybrid methods was applied to monitor and analyze a very heavy dust storm that occurred in northern China from 14–18 March 2021, which was one of the most severe dust storms in recent decades. Results showed that the dust storm event could be well monitored and analyzed dynamically. It was found that the dust originated in western Mongolia and northwestern China and was then transmitted along the northwest–southeast direction, consequently affected the air quality of most cities of northern China. The results are consistent with the prior research and showed the excellent potential of the integration of the hybrid methods in monitoring dust storms. [ABSTRACT FROM AUTHOR]

In many change detection applications, the focus is often on one specific change class. The one-class support vector machine (OCSVM)-based change detection method has been proved effective for dealing with such problems, which only requires samples from the change class of interest as the training data. However, this classical method only uses a single kernel which limits its separating capabilities in real-world applications. To further improve the efficacy of the OCSVM-based change detection method, this paper proposes an improved change detection method that uses a data-oriented composite-kernel-based one-class support vector machine. It utilizes the feature information entropy of the training data to determine the kernel weights in constructing a composite kernel. Experimental results on two data-sets demonstrate that the proposed method outperforms the existing classical OCSVM-based change detection method and the traditional composite-kernel-based method with relatively few false alarm errors, and shows good potential for further applications. [ABSTRACT FROM PUBLISHER]