Trophic state assessment of optically diverse lakes using Sentinel-3-derived trophic level index

Liu, Hui; He, Baoyin; Zhou, Yadong; Kutser, Tiit; Toming, Kaire; Feng, Qi; Yang, Xiaoqin; Fu, Congju; Yang, Fan; Li, Wen; Peng, Feng

doi:10.1016/j.jag.2022.103026

Cited by 11 publications

(14 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, further research is required to investigate the effect of the trophic level on the spectral separability of hill lakes from surrounding features since the water trophic state rules the irradiance water attenuation through the water column [87] and the water spectral reflectance [88] at the surface.…”

Section: Discussionmentioning

confidence: 99%

On the Choice of the Most Suitable Period to Map Hill Lakes via Spectral Separability and Object-Based Image Analyses

Maltese¹

2023

Remote Sensing

View full text Add to dashboard Cite

Technological advances in Earth observation made images characterized by high spatial and temporal resolutions available, nevertheless bringing with them the radiometric heterogeneity of small geographical entities, often also changing in time. Among small geographical entities, hill lakes exhibit a widespread distribution, and their census is sometimes partial or shows unreliable data. High resolution and heterogeneity have boosted the development of geographic object-based image analysis algorithms. This research analyzes which is the most suitable period for acquiring satellite images to identify and delimitate hill lakes. This is achieved by analyzing the spectral separability of the surface reflectance of hill lakes from surrounding bare or vegetated soils and by implementing a semiautomatic procedure to enhance the segmentation phase of a GEOBIA algorithm. The proposed procedure was applied to high spatial resolution satellite images acquired in two different climate periods (arid and temperate), corresponding to dry and vegetative seasons. The segmentation parameters were tuned by minimizing an under- and oversegmentation metric on surfaces and perimeters of hill lakes selected as the reference. The separability of hill lakes from their surrounding was evaluated using Euclidean and divergence metrics both in the arid and temperate periods. The classification accuracy was evaluated by calculating the error matrix and normalized error matrix. Classes’ reflectances in the image acquired in the arid period show the highest average separability (3–4 higher than in the temperate one). The segmentation based on the reference areas performs more than that based on the reference perimeters (metric ≈20% lower). Both separability metrics and classification accuracies indicate that images acquired in the arid period are more suitable than temperate ones to map hill lakes.

show abstract

Section: Discussionmentioning

confidence: 99%

On the Choice of the Most Suitable Period to Map Hill Lakes via Spectral Separability and Object-Based Image Analyses

Maltese¹

2023

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…The inputs of the model were MCI, B5/B4, and B3/B4, while TLI obtained from WQP field water quality monitoring data served as the output. This approach has been proven effective in inverting TLI and providing important support for water quality monitoring and management [33,78]. Therefore, we adopted the same methodology to model TLI in this study.…”

Section: Assisted Verification-tli-based Modelingmentioning

confidence: 99%

Using C2X to Explore the Uncertainty of In Situ Chlorophyll-a and Improve the Accuracy of Inversion Models

Zhou

Yang

et al. 2023

Sustainability

View full text Add to dashboard Cite

Quality water plays a huge role in human life. Chlorophyll-a (Chl-a) in water bodies is a direct reflection of the population size of the primary productivity of various phytoplankton species in the water body and can provide critical information on the health of water ecosystems and the pollution status of water quality. Case 2 Regional CoastColour (C2RCC) is a networked atmospheric correction processor introduced by the Sentinel Application Platform for various remote sensing products. Among them, the Extreme Case-2 Waters (C2X) process has demonstrated advantages in inland complex waters, enabling the generation of band data, conc_chl product for Chl-a, and kd_z90max product for Secchi Depth (SD). Accurate in situ data are essential for the development of reliable Chl-a models, while in situ data measurement is limited by many factors. To explore and improve the uncertainties involved, we combined the C2X method with Sentinel-2 imagery and water quality data, taking lakes in Wuhan from 2018 to 2021 as a case. A Chl-a model was developed and validated using an empirical SD model and a neural network incorporating Trophic Level Index (TLI) to derive the predicted correction result, Chl-a_t. The results indicated that (1) the conc_chl product measured by C2X and in situ Chl-a exhibited consistent overall trends, with the highest correlation observed in the range of 2–10 μg/L. (2) The corrected Chl-a_t using the conc_chl product had a mean absolute error of approximately 10–15 μg/L and a root-mean-square error of approximately 8–10 μg/L, while using in situ Chl-a had a root-mean-square error (RMSE) of approximately 15 μg/L and a mean absolute error (MAE) of approximately 20 μg/L; both errors decreased by double after correction. (3) The correlation coefficient (R) between Chl-a_t and each data point in the Chl-a model results was lower than that of SD-a_t with each data point in the SD model results. Additionally, the difference in R-value between Chl-a_t and each data point (0.45–0.60) was larger than that of SD-a_t with each data point (0.35–0.5). (4) When using corrected Chl-a_t data to calculate the TLI estimation model, both RMSE and MAE decreased, which were 1μg/L lower than those derived from uncorrected data, while R increased, indicating an improvement in accuracy and reliability. These findings demonstrated the presence of in situ errors in Chl-a measurements, which must be acknowledged during research. This study holds practical significance as some of these errors can be effectively corrected through the use of C2X atmospheric correction on spectral bands.

show abstract

“…The use of NNs in satellite-based water quality monitoring goes beyond predictive modeling, demonstrating the algorithm's broad technical appeal. For instance, New Neural Network version.1 (NNv1) and New Neural Network version.2 (NNv2) have been used specifically to develop neural network-based algorithms such as C2RCC, multilayer neural networks (MLNNs), and Case2extreme or Case2complex (C2X) [180,253,[280][281][282][283][284][285][286][287][288][289][290]. These algorithms correct and retrieve water reflectances as well as water inherent optical properties (IOPs).…”

Section: Machine or Deep Learning Model Choicementioning

confidence: 99%

“…Tree-based models, such as DTs [25,51,[221][222][223], RFs [25,30,[32][33][34]39,46], BST [157,193,201], and XGBoost [23,36,180,193], are effective methods for capturing complex patterns and relationships in datasets. They are particularly adept at handling both linear and non-linear associations between features and the target variable.…”

Section: Machine or Deep Learning Model Choicementioning

confidence: 99%

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

Mukonza,

Chiang

2023

Environments

View full text Add to dashboard Cite

This review paper adopts bibliometric and meta-analysis approaches to explore the application of supervised machine learning regression models in satellite-based water quality monitoring. The consistent pattern observed across peer-reviewed research papers shows an increasing interest in the use of satellites as an innovative approach for monitoring water quality, a critical step towards addressing the challenges posed by rising anthropogenic water pollution. Traditional methods of monitoring water quality have limitations, but satellite sensors provide a potential solution to that by lowering costs and expanding temporal and spatial coverage. However, conventional statistical methods are limited when faced with the formidable challenge of conducting pattern recognition analysis for satellite geospatial big data because they are characterized by high volume and complexity. As a compelling alternative, the application of machine and deep learning techniques has emerged as an indispensable tool, with the remarkable capability to discern intricate patterns in the data that might otherwise remain elusive to traditional statistics. The study employed a targeted search strategy, utilizing specific criteria and the titles of 332 peer-reviewed journal articles indexed in Scopus, resulting in the inclusion of 165 articles for the meta-analysis. Our comprehensive bibliometric analysis provides insights into the trends, research productivity, and impact of satellite-based water quality monitoring. It highlights key journals and publishers in this domain while examining the relationship between the first author’s presentation, publication year, citation count, and journal impact factor. The major review findings highlight the widespread use of satellite sensors in water quality monitoring including the MultiSpectral Instrument (MSI), Ocean and Land Color Instrument (OLCI), Operational Land Imager (OLI), Moderate Resolution Imaging Spectroradiometer (MODIS), Thematic Mapper (TM), Enhanced Thematic Mapper Plus (ETM+), and the practice of multi-sensor data fusion. Deep neural networks are identified as popular and high-performing algorithms, with significant competition from extreme gradient boosting (XGBoost), even though XGBoost is relatively newer in the field of machine learning. Chlorophyll-a and water clarity indicators receive special attention, and geo-location had a relationship with optical water classes. This paper contributes significantly by providing extensive examples and in-depth discussions of papers with code, as well as highlighting the critical cyber infrastructure used in this research. Advances in high-performance computing, large-scale data processing capabilities, and the availability of open-source software are facilitating the growing prominence of machine and deep learning applications in geospatial artificial intelligence for water quality monitoring, and this is positively contributing towards monitoring water pollution.

show abstract

Trophic state assessment of optically diverse lakes using Sentinel-3-derived trophic level index

Cited by 11 publications

References 47 publications

On the Choice of the Most Suitable Period to Map Hill Lakes via Spectral Separability and Object-Based Image Analyses

On the Choice of the Most Suitable Period to Map Hill Lakes via Spectral Separability and Object-Based Image Analyses

Using C2X to Explore the Uncertainty of In Situ Chlorophyll-a and Improve the Accuracy of Inversion Models

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

Contact Info

Product

Resources

About