Version 1
: Received: 17 May 2024 / Approved: 20 May 2024 / Online: 20 May 2024 (12:12:30 CEST)
How to cite:
Costa Rocha, P. A.; Oliveira Santos, V.; Thé, J. V. G.; Gharabaghi, B. Enhancing Turbidity Modeling in the Mississippi River Using Machine Learning and Sentinel‐2 Remote Sensing Data: A Generalizability Analysis. Preprints2024, 2024051259. https://doi.org/10.20944/preprints202405.1259.v1
Costa Rocha, P. A.; Oliveira Santos, V.; Thé, J. V. G.; Gharabaghi, B. Enhancing Turbidity Modeling in the Mississippi River Using Machine Learning and Sentinel‐2 Remote Sensing Data: A Generalizability Analysis. Preprints 2024, 2024051259. https://doi.org/10.20944/preprints202405.1259.v1
Costa Rocha, P. A.; Oliveira Santos, V.; Thé, J. V. G.; Gharabaghi, B. Enhancing Turbidity Modeling in the Mississippi River Using Machine Learning and Sentinel‐2 Remote Sensing Data: A Generalizability Analysis. Preprints2024, 2024051259. https://doi.org/10.20944/preprints202405.1259.v1
APA Style
Costa Rocha, P. A., Oliveira Santos, V., Thé, J. V. G., & Gharabaghi, B. (2024). Enhancing Turbidity Modeling in the Mississippi River Using Machine Learning and Sentinel‐2 Remote Sensing Data: A Generalizability Analysis. Preprints. https://doi.org/10.20944/preprints202405.1259.v1
Chicago/Turabian Style
Costa Rocha, P. A., Jesse Van Griensven Thé and Bahram Gharabaghi. 2024 "Enhancing Turbidity Modeling in the Mississippi River Using Machine Learning and Sentinel‐2 Remote Sensing Data: A Generalizability Analysis" Preprints. https://doi.org/10.20944/preprints202405.1259.v1
Abstract
Turbidity is an important indicator of water quality in hydrology. More traditional ways to monitor turbidity can provide reliable results. However, they are prone to human error, have elevated costs, and lack real-time monitoring capacity. Addressing these hindrances, in this work we combine spectral bands and indices from Sentinel-2 with several machine learning paradigms, namely XGBoost, Random Forests, GMDH, Support Vector Regression, k-Nearest Neighbors and Least Absolute Shrinkage and Selection Operator to model turbidity, using data from twelve monitoring stations encompassing the Mississippi River, USA. Results show that considering the individual monitoring stations, the ML paradigms for turbidity modeling were satisfactory at locations with a larger range and standard deviation values, achieving a mean R2 value of 59.5%. Tree-based models were the best overall approach, often ranking as the best or second-best performing model. When all the samples from the monitoring stations were used, the XGBoost provided superior output for turbidity modeling, reaching an R2 equal to 75.7%. A comprehensive comparison with the literature found values showed that the models implemented using this study’s methodology could provide competitive results, deeming it as an alternative for turbidity modeling from remote sensing data.
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
