Using satellite remote sensing and numerical modelling for the monitoring of suspended particulate matter concentration during reclamation construction at Dalian offshore airport in China

Abstract

In this study, remote sensing and numerical modelling were used to monitor the distribution of suspended particulate matter (SPM) concentrations induced by a reclamation construction of an offshore artificial island airport in China's Bohai Sea. A hydrodynamic model and a water quality model were used to simulate the SPM diffusion under the actions of tides and currents. An artificial neural network (ANN) model was then developed based on HJ-1A/1B satellite and measured SPM concentrations. The obtained main results are as follows: (1) The dispersion area of SPM with a concentration higher than 10 mg/l obtained by the numerical model was approximately 105.15 km 2 ; (2) The developed ANN model provided an excellent relationship between the measured and retrieved SPM concentrations, with R 2 = 0.97 and mean relative error (MRE) of 30%, at the model validation phase; (3) The SPM distribution obtained by the numerical model showed a qualitatively similar pattern with that obtained by the ANN model, indicating the modelled SPM distribution was credible. Thus, when the distribution of SPM concentrations cannot be derived from the satellite images due to weather conditions, the numerical model can be used to simulate the SPM distribution in the sea area around the construction.

Meta-Analysis of Satellite Observations for United Nations Sustainable Development Goals: Exploring the Potential of Machine Learning for Water Quality Monitoring

“…In contrast, there are researchers who have also reported using R 2 alone or in combination with other evaluation metrics to assess model performance. Notable examples include Song et al [156], Balasubramanian et al [28], Bertani et al [157], Chang et al [158], Chen et al [159], Fan et al [160], and Kravitz et al [161]. These researchers have utilized a variety of metrics, including mean absolute percentage error (MAPE), mean absolute error (MAE), and bias relative error (RE), to evaluate model performance.…”

Coupled hydrodynamic and water quality modeling in the coastal waters off Chennai, East Coast of India

“…In situIn-situ sampling and laboratory analysis are more accurate than modeling and remote sensing data, while the remote sensing data provide better spatiotemporal resolution than the data obtained from in situ samplings (Miller and McKee, 2004;Wu et al, 2014), and numerical modeling infers spatiotemporal and water column information. However, satellite image data cannot consistently obtain the information due to the limitations of weather conditions, date of the pass, and its swath (Song et al, 2018); similarly, numerical modeling required accurate time series boundary information. Water temperature and salinity play a substantial role in regulating the processes in the aquatic system (Mogaddam et al, 2019).…”

Ambient air particulate matter (PM10) satellite monitoring and respiratory health effects assessment

“…Today spatial variability and longrange transport of air pollution can account for satellitemounted sensors, thus by measuring concentration gradients between stationary monitors it can provide global coverage of aerosols, this data provided better spatial distribution and temporal resolution than the data obtained from in-ground stations. It can be used especially in regions that lack ground monitoring capabilities [21,22]. For continuous air quality monitoring applications offering a wide coverage area satellite imagery such as Landsat, Moderate-resolution Imaging Spectroradiometer (MODIS), and SPOT can be used [18].…”