Kansikuva

Resolving inherent constraints in eutrophication monitoring of small lakes using multi-source satellites and machine learning

Tallennettuna:
Julkaisussa:NPJ Clean Water vol. 8, no. 1 (2025), p. 99-111
Päätekijä: Si, Wei
Muut tekijät: Chen, Zhixiong, Jim, Chi Yung, Tan, Mou Leong, Liu, Dong, Yao, Yue, Wei, Lifei, Xu, Shangshang, Zhang, Fei
Julkaistu:
Nature Publishing Group
Aiheet:
Lakes
Accuracy
Deep learning
Phosphorus
Water shortages
Chemical oxygen demand
Normal distribution
Eutrophication
Water color
Remote sensing
Remote monitoring
Spatial distribution
Seasonal variations
Machine learning
Monitoring systems
Performance evaluation
Probability distribution
Algae
Nitrogen
Water quality
Regression
Trophic levels
Correlation analysis
Imagery
Environmental
Linkit:Citation/Abstract
Full Text
Full Text - PDF
Tagit: Lisää tagi
Ei tageja, Lisää ensimmäinen tagi!
Abstrakti:Remote sensing monitoring of small-lake eutrophication faces challenges such as sparse data, insufficient synergy of multi-source data, and limited model generalization performance. Hence, this study developed a scenario-aware modeling framework for the trophic level index (TLI) by integrating multi-source imagery data from Sentinel-2, GF-1, HJ-2, and PlanetScope, using Dongqian Lake in Zhejiang Province, China as the case study. The cross-sensor prediction accuracy was evaluated using algorithms such as CatBoost Regression (CBR), XGBoost Regression (XGBR), TabPFN Regression (TPFNR), and Linear Regression (LR). Meanwhile, the influence of input features was quantified by SHapley Additive exPlanations (SHAP). The main results found that : (1) Overall annual mean values of total nitrogen/total phosphorus ratio (TN/TP) and TLI were 22.13 and 37.36 ± 4.99, respectively, indicating a mesotrophic and phosphorus-limited state in Dongqian Lake. (2) TLI exhibited the strongest correlation with water color and algal spectral indices, including Normalized Difference Water Index (NDWI), Normalized Green–Red Difference Index (NGRDI), and Blue–Green Ratio (BGR). (3) CBR demonstrated the strongest cross-sensor generalization capability across different imagery, with only minor variations in prediction accuracy (ΔR ≈ 0.07–0.15). Feature attribution analysis identified NDWI, NGRDI, and BGR as primary contributing features for the CBR model. (4) Integrating high-frequency multi-source remote sensing imagery with 27 field surveys achieved seamless monitoring of the TLI. The spatial distribution of TLI showed distinct seasonal variations, with higher values observed in nearshore areas and lower values in the lake center. TLI values were relatively low in spring, but surged sharply and remained elevated in summer. This study provided a reference basis for detailed remote sensing monitoring and management of eutrophication in small lakes.
ISSN:2059-7037
DOI:10.1038/s41545-025-00525-8
Lähde:Publicly Available Content Database