Optimization of water quality monitoring programs by data mining

Monitoring network optimization and impact of fish farming upon water quality in the Três Marias Hydroelectric Reservoir, Brazil

Hydroelectric power is the main source of electrical energy in Brazil. Electrical energy providers have the duty to monitor water quality in reservoirs to preserve water quality and support best management practices that enable multiple water uses, including fish production. In this context, the objectives of this study were (i) to perform a historical evaluation of water quality in Três Marias Reservoir, (ii) to present an optimization of the water quality monitoring network, and (iii) to evaluate the evolution and impact of fish farming upon surface water quality by using secondary data measured in situ and remote sensing. A systematic approach was applied to analyze historical water quality data. Principal component analysis (PCA) and cluster analysis (CA) were applied to identify the most important parameters and monitoring points. Images obtained from Sentinel 2 were treated by contrast to quantify simple and weighted densities of fish farming activities in the region while regression analysis was performed to verify correlations between these densities and water quality parameters. Results showed that the pH and total suspended solids were the most important parameters for characterizing water quality, especially near tributaries, and that monitoring points could be grouped into three clusters (upstream, central, and downstream regions) with distinct water quality conditions. The PCA indicated that there is no redundance among parameters nor monitoring stations and that areas near tributaries must be prioritized for monitoring as these are important sources of suspended solids. Remote sensing images showed that the area occupied by fish farms has increased in the reservoir from 2016 to 2022 and the methodology used for this purpose in this study may be applied to other bodies of water. Chlorophyll-a showed a direct relationship with the density of fish farms indicating a possible influence of nutrient input to the reservoir by this activity. These results provide valuable information to support decision-making related to water management in the reservoir.

Deep-Learning-Based Automated Tracking and Counting of Living Plankton in Natural Aquatic Environments

Plankton are widely distributed in the aquatic environment and serve as an indicator of water quality. Monitoring the spatiotemporal variation in plankton is an efficient approach to forewarning environmental risks. However, conventional microscopy counting is time-consuming and laborious, hindering the application of plankton statistics for environmental monitoring. In this work, an automated video-oriented plankton tracking workflow (AVPTW) based on deep learning is proposed for continuous monitoring of living plankton abundance in aquatic environments. With automatic video acquisition, background calibration, detection, tracking, correction, and statistics, various types of moving zooplankton and phytoplankton were counted at a time scale. The accuracy of AVPTW was validated with conventional counting via microscopy. Since AVPTW is only sensitive to mobile plankton, the temperature- and wastewater-discharge-induced plankton population variations were monitored online, demonstrating the sensitivity of AVPTW to environmental changes. The robustness of AVPTW was also confirmed with natural water samples from a contaminated river and an uncontaminated lake. Notably, automated workflows are essential for generating large amounts of data, which are a prerequisite for available data set construction and subsequent data mining. Furthermore, data-driven approaches based on deep learning pave a novel way for long-term online environmental monitoring and elucidating the correlation underlying environmental indicators. This work provides a replicable paradigm to combine imaging devices with deep-learning algorithms for environmental monitoring.

A general approach for evaluating of the coverage, resolution, and representation of streamflow monitoring networks

Streamflow monitoring networks provide information for a wide range of public interests in river and streams. A general approach to evaluate monitoring for different interests is developed to support network planning and design. The approach defines three theoretically distinct information metrics (coverage, resolution, and representation) based on the spatial distribution of a variable of interest. Coverage is the fraction of information that a network can provide about a variable when some areas are not monitored. Resolution is the information available from the network relative to the maximum information possible given the number of sites in the network. Representation is the information that a network provides about a benchmark distribution of a variable. Information is defined using Shannon entropy where the spatial discretization of a variable among spatial elements of a landscape or sites in a network indicates the uncertainty in the spatial distribution of the variable. This approach supports the design of networks for monitoring of variables with heterogeneous spatial distributions ("hot spots" and patches) that might otherwise be unmonitored because they occupy insignificant portions of the landscape. Areas where monitoring will maintain or improve the metrics serve as objective priorities for public interests in network design. The approach is demonstrated for the streamflow monitoring network operated by the United States Geological Survey during water year 2020 indicating gaps in the coverage of coastal rivers and the resolution of low flows.

Data mining from process monitoring of typical polluting enterprise

With the increasing volume of environmental monitoring data, extracting valuable insights from multivariate time series sensor data can facilitate comprehensive information utilization and support informed decision-making in environmental management. However, there is a dearth of comprehensive research on multivariate data analysis for process monitoring in typical polluting enterprises. In this study, an artificial neural network model based on back-propagation algorithm (BP-ANN) was developed to predict the wastewater and exhaust gas emissions using IoT data obtained from process monitoring of a typical polluting enterprise located in Taizhou, Zhejiang Province, China. The results indicate that the model constructed has a high predictive coefficient of determination (R2) with values of 0.8510, 0.9565, 0.9561, 0.9677, and 0.9061 for chemical oxygen demand (COD), potential of hydrogen (pH), electrical conductivity (EC), flue gas emission (FGE), and non-methane hydrocarbon concentration (NMHC) respectively. For the first time, the variable importance measure (VIM)-assisted BP-ANN was employed to investigate the internal and external correlations between wastewater and exhaust gas treatment, thereby enhancing the interpretability of mapping features in the BP-ANN model. The predicted errors for pH and FGE have been demonstrated to fall within the range of - 0.62 ~ 0.30 and - 0.21 ~ 0.15 m3/s, respectively, with average relative errors of 1.05% and 9.60%, which is advantageous in detecting anomalous data and forecasting pollution indicator values. Our approach successfully addresses the challenge of segregating data analysis for wastewater disposal and exhaust gas disposal in the process monitoring of polluting enterprises, while also unearthing potential variables that significantly contribute to the BP-ANN model, thereby facilitating the selection and extraction of characteristic variables.

The relationship among air pollution, meteorological factors and COVID-19 in the Brussels Capital Region

In great metropoles, there is a need for a better understanding of the spread of COVID-19 in an outdoor context with environmental parameters. Many studies on this topic have been carried out worldwide. However, there is conflicting evidence regarding the influence of environmental variables on the transmission, hospitalizations and deaths from COVID-19, even though there are plausible scientific explanations that support this, especially air quality and meteorological factors. Different urban contexts, methodological approaches and even the limitations of ecological studies are some possible explanations for this issue. That is why methodological experimentations in different regions of the world are important so that scientific knowledge can advance in this aspect. This research analyses the relationship between air pollution, meteorological factors and COVID-19 in the Brussels Capital Region. We use a data mining approach that is capable of extracting patterns in large databases with diverse taxonomies. Data on air pollution, meteorological, and epidemiological variables were processed in time series for the multivariate analysis and the classification based on association. The environmental variables associated with COVID-19-related deaths, cases and hospitalization were PM_2.5, O₃, NO₂, black carbon, radiation, air pressure, wind speed, dew point, temperature and precipitation. These environmental variables combined with epidemiological factors were able to predict intervals of hospitalization, cases and deaths from COVID-19. These findings confirm the influence of meteorological and air quality variables in the Brussels region on deaths and cases of COVID-19 and can guide public policies and provide useful insights for high-level governmental decision-making concerning COVID-19. However, it is necessary to consider intrinsic elements of this study that may have influenced our results, such as the use of air quality aggregated data, ecological fallacy, focus on acute effects in the time-series study, the underreporting of COVID-19, and the lack of behavioral factors.