Sajib, Abdul Majed, Uddin, Md. Galal, Rahman, Azizur, Ahmadian, Reza  ORCID: https://orcid.org/0000-0003-2665-4734 and Olbert, Agnieszka I.
2025.
Remote sensing applications for monitoring optically inactive water quality indicators: A comprehensive review.
Earth-Science Reviews
, p. 105259.
10.1016/j.earscirev.2025.105259
|
Abstract
Monitoring water quality (WQ) is crucial to ensure the safety and health of our water resources. Despite their importance, contemporary WQ monitoring programs are struggling with challenges such as high costs, limited spatio-temporal resolution, and data reliability issues. A promising solution to these challenges is the integration of remote sensing (RS) techniques with machine learning (ML) and artificial intelligence (AI) algorithms, which can significantly improve the efficiency and accuracy of WQ monitoring. Based on the literature, most of the studies have focused on optically active (OA)-WQ indicators like chlorophyll-a and colored dissolved organic matter, etc., while a few studies have been carried out focusing on optically inactive (OI)-WQ indicators. But WQ monitoring requires a number of OA- and OI-WQ indicators; for instance, the European Union Water Framework Directive (WFD) recommend 11 fundamental WQ indicators, which include both OA- and OI-WQ. Therefore, it is essential to consider both types of indicators in a regular WQ monitoring program to develop an effective water resources management plan. However, several recent studies have shown that the development of RS-based OI-WQ indicator retrieval model(s) introduces considerable uncertainty in the final retrieval results due to various factors. Additionally, these studies highlight that most of the retrieval models may not be suitable for global application. To highlight these challenges, the goal of the research is to conduct a comprehensive analysis of various RS data and existing techniques in order to more accurately retrieve OI-WQ indicators such as pH, dissolved oxygen (DOX), biological oxygen demand (BOD5), total phosphorus (TP), total nitrogen (TN), and dissolved inorganic nitrogen (DIN) in different waterbodies. To achieve the research objectives, this study conducted a critical review analysis of 105 research publications, including journal papers and conference papers, from 2005 to 2023. The study not only identified different types of satellite data, such as Landsat, Sentinel, and Aqua/Terra (MODIS), which are widely used, but also identified the advantages and disadvantages of different models, including empirical, semi-empirical, and ML/AI-based methods that are widely used in developing RS-driven retrieval model(s) for various OI-WQ indicators. Additionally, the study identified a range of opportunities (e.g., proposing a structural framework, reliable global model, etc.) and limitations (e.g., lack of in-situ data, structural framework, optimal RS wavelength for different OI-WQ indicators, etc.) in existing retrieval models. Moreover, the analysis suggests that advanced ML/AI approaches can be effective in retrieving OI-WQ indicators compared to other techniques in terms of retrieval data accuracy and reliability. The study also highlights current limitations of RS data and retrieval methods, such as spatial and temporal constraints, the need for improved calibration, and the demand for broader and more diverse training and testing datasets. Finally, the findings emphasize the significant potential of ML/AI algorithms in improving RS-based techniques for WQ monitoring, which may be more useful for water resource management and sustainable development strategies in the future.
| Item Type: | Article |
|---|---|
| Date Type: | Published Online |
| Status: | In Press |
| Schools: | Schools > Engineering |
| Publisher: | Elsevier BV |
| ISSN: | 0012-8252 |
| Date of Acceptance: | 24 August 2025 |
| Last Modified: | 15 Sep 2025 13:00 |
| URI: | https://orca.cardiff.ac.uk/id/eprint/181090 |
Actions (repository staff only)
 |
Edit Item |
Altmetric
Altmetric