IDEAS home Printed from https://ideas.repec.org/a/epw/ejece0/v10y2026i4id70436.html

When High Accuracy Misleads: Hidden Bias in Satellite Water Quality Models

Author

Listed:
  • Stefani D. W. Yates
  • Yanzhen Qu

Abstract

Harmonized Landsat–Sentinel-2 (HLS) imagery combined with machine learning (ML) enables continuous turbidity monitoring beyond sparse gauge networks, yet reported performance often relies on validation protocols that ignore spatial autocorrelation. This study quantifies that bias by benchmarking five ML algorithms (LightGBM, XGBoost, Gradient Boosting, Random Forest, Linear Regression) on an identical turbidity-regression task across the regulated Tennessee River and free-flowing Flint River basins near Huntsville, Alabama. The dataset includes 2,140 cloud-free spectral–turbidity matchups from 12 U.S. Geological Survey (USGS) stations (2018–2025). Models were evaluated using random 80/20 splits and five-fold HUC-12 spatially blocked cross-validation (SBCV). Across nonlinear models, random splits yielded R2 = 0.27–0.30, while SBCV produced R2 = −0.11 to −0.15, indicating ∼0.40 R2 inflation (Mann–Whitney U, p

Suggested Citation

  • Stefani D. W. Yates & Yanzhen Qu, 2026. "When High Accuracy Misleads: Hidden Bias in Satellite Water Quality Models," European Journal of Electrical Engineering and Computer Science, European Open Science, vol. 10(4), pages 9-16, July.
    • Handle: RePEc:epw:ejece0:v:10:y:2026:i:4:id:70436
    DOI: 10.24018/ejece.2026.10.4.70436
    as

    Download full text from publisher

    File URL: https://eu-opensci.org/index.php/ejece/article/view/70436
    File Function: Abstract page
    Download Restriction: no
    File URL: https://eu-opensci.org/index.php/ejece/article/download/70436/14583
    File Function: Full text
    Download Restriction: no
    File URL: https://libkey.io/10.24018/ejece.2026.10.4.70436?utm_source=ideas
    LibKey link: if access is restricted and if your library uses this service, LibKey will redirect you to where you can use your library subscription to access this item
    ---><---

    References listed on IDEAS

    as
    1. Pierre Ploton & Frédéric Mortier & Maxime Réjou-Méchain & Nicolas Barbier & Nicolas Picard & Vivien Rossi & Carsten Dormann & Guillaume Cornu & Gaëlle Viennois & Nicolas Bayol & Alexei Lyapustin & Syl, 2020. "Spatial validation reveals poor predictive performance of large-scale ecological mapping models," Nature Communications, Nature, vol. 11(1), pages 1-11, December.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Ali Ismaeel & Amos P. K. Tai & Erone Ghizoni Santos & Heveakore Maraia & Iris Aalto & Jan Altman & Jiří Doležal & Jonas J. Lembrechts & José Luís Camargo & Juha Aalto & Kateřina Sam & Lair Cristina Av, 2024. "Patterns of tropical forest understory temperatures," Nature Communications, Nature, vol. 15(1), pages 1-10, December.
    2. Christine Tedijanto & Solomon Aragie & Zerihun Tadesse & Mahteme Haile & Taye Zeru & Scott D Nash & Dionna M Wittberg & Sarah Gwyn & Diana L Martin & Hugh J W Sturrock & Thomas M Lietman & Jeremy D Ke, 2022. "Predicting future community-level ocular Chlamydia trachomatis infection prevalence using serological, clinical, molecular, and geospatial data," PLOS Neglected Tropical Diseases, Public Library of Science, vol. 16(3), pages 1-22, March.
    3. Haider, Saira M. & Benscoter, Allison M. & Pearlstine, Leonard & D'Acunto, Laura E. & Romañach, Stephanie S., 2021. "Landscape-scale drivers of endangered Cape Sable Seaside Sparrow (Ammospiza maritima mirabilis) presence using an ensemble modeling approach," Ecological Modelling, Elsevier, vol. 461(C).
    4. Marsh, Charles J. & Gavish, Yoni & Kuemmerlen, Mathias & Stoll, Stefan & Haase, Peter & Kunin, William E., 2023. "SDM profiling: A tool for assessing the information-content of sampled and unsampled locations for species distribution models," Ecological Modelling, Elsevier, vol. 475(C).
    5. Gustavo Larrea‐Gallegos & Ian Vázquez‐Rowe, 2022. "Exploring machine learning techniques to predict deforestation to enhance the decision‐making of road construction projects," Journal of Industrial Ecology, Yale University, vol. 26(1), pages 225-239, February.
    6. Zou, Lidong & Chen, Jing & Wang, Yahui & Zhao, Anzhou & Dai, Ming & Sanchez-Azofeifa, Arturo, 2025. "Simulating tropical forest change using a cellular automata-agent based model," Ecological Modelling, Elsevier, vol. 509(C).
    7. Wen, Bingbin & Landuyt, Dries & Verheyen, Kris & Waller, Donald M. & Blondeel, Haben, 2025. "Do European models of temperate forest ecological change apply in North America?," Ecological Modelling, Elsevier, vol. 509(C).
    8. Wan-Yi Zhao & Zhong-Cheng Liu & Shi Shi & Jie-Lan Li & Ke-Wang Xu & Kang-You Huang & Zhi-Hui Chen & Ya-Rong Wang & Cui-Ying Huang & Yan Wang & Jing-Rui Chen & Xian-Ling Sun & Wen-Xing Liang & Wei Guo , 2024. "Landform and lithospheric development contribute to the assembly of mountain floras in China," Nature Communications, Nature, vol. 15(1), pages 1-13, December.
    9. Xuejun Yang & Carol C. Baskin & Jerry M. Baskin & Robin J. Pakeman & Zhenying Huang & Ruiru Gao & Johannes H. C. Cornelissen, 2021. "Global patterns of potential future plant diversity hidden in soil seed banks," Nature Communications, Nature, vol. 12(1), pages 1-8, December.
    10. Katerina Georgiou & Robert B. Jackson & Olga Vindušková & Rose Z. Abramoff & Anders Ahlström & Wenting Feng & Jennifer W. Harden & Adam F. A. Pellegrini & H. Wayne Polley & Jennifer L. Soong & William, 2022. "Global stocks and capacity of mineral-associated soil organic carbon," Nature Communications, Nature, vol. 13(1), pages 1-12, December.
    11. David Mouillot & Laure Velez & Camille Albouy & Nicolas Casajus & Joachim Claudet & Vincent Delbar & Rodolphe Devillers & Tom B. Letessier & Nicolas Loiseau & Stéphanie Manel & Laura Mannocci & Jessic, 2024. "The socioeconomic and environmental niche of protected areas reveals global conservation gaps and opportunities," Nature Communications, Nature, vol. 15(1), pages 1-13, December.
    12. Laura G. van Galen & Justin D. Stewart & Clara Qin & Adriana Corrales & Bethan F. Manley & E. Toby Kiers & Thomas W. Crowther & Michael E. Van Nuland, 2025. "Global divergence in plant and mycorrhizal fungal diversity hotspots," Nature Communications, Nature, vol. 16(1), pages 1-12, December.
    13. Zhang, Yuxi & Wimalathunge, Niranjan & Haan, Sebastian & Wang, Jie & Zou, Xinglong & Bishop, Thomas, 2025. "High spatio-temporal resolution soil moisture nowcasting at multiple depths with data-driven approaches," Agricultural Water Management, Elsevier, vol. 312(C).
    14. Emre Tepe, 2024. "A random forests-based hedonic price model accounting for spatial autocorrelation," Journal of Geographical Systems, Springer, vol. 26(4), pages 511-540, October.
    15. Hanna Meyer & Edzer Pebesma, 2022. "Machine learning-based global maps of ecological variables and the challenge of assessing them," Nature Communications, Nature, vol. 13(1), pages 1-4, December.
    16. Shuai Zhao & Yixin Hua & Fang Yan, 2025. "Cyberspace Skeleton Map Visualization Considering Node Quality Correction and Spatial Distribution Characteristics," Complexity, John Wiley & Sons, vol. 2025(1).
    17. Anton M. Potapov & Carlos A. Guerra & Johan Hoogen & Anatoly Babenko & Bruno C. Bellini & Matty P. Berg & Steven L. Chown & Louis Deharveng & Ľubomír Kováč & Natalia A. Kuznetsova & Jean-François Pong, 2023. "Globally invariant metabolism but density-diversity mismatch in springtails," Nature Communications, Nature, vol. 14(1), pages 1-13, December.
    18. Temesgen Alemayehu Abera & Janne Heiskanen & Eduardo Eiji Maeda & Mohammed Ahmed Muhammed & Netra Bhandari & Ville Vakkari & Binyam Tesfaw Hailu & Petri K. E. Pellikka & Andreas Hemp & Pieter G. Zyl &, 2024. "Deforestation amplifies climate change effects on warming and cloud level rise in African montane forests," Nature Communications, Nature, vol. 15(1), pages 1-10, December.
    19. Wadoux, Alexandre M.J.-C. & Heuvelink, Gerard B.M. & de Bruin, Sytze & Brus, Dick J., 2021. "Spatial cross-validation is not the right way to evaluate map accuracy," Ecological Modelling, Elsevier, vol. 457(C).
    20. Diana Koldasbayeva & Polina Tregubova & Mikhail Gasanov & Alexey Zaytsev & Anna Petrovskaia & Evgeny Burnaev, 2024. "Challenges in data-driven geospatial modeling for environmental research and practice," Nature Communications, Nature, vol. 15(1), pages 1-16, December.

    More about this item

    Keywords

    ;
    ;
    ;
    ;

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:epw:ejece0:v:10:y:2026:i:4:id:70436. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: support (email available below). General contact details of provider: https://eu-opensci.org/index.php/ejece .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.