Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Prediction of Dissolved Organic Carbon in Inland Waters Using Machine Learning Methods : A Case Study of Lake Erken, Sweden

Wei, Xufeng LU ; Xu, Wenbo ; Wang, Lanhui LU orcid ; Guo, Renkui LU orcid and Duan, Zheng LU (2025) 2025 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2025 In IEEE International Symposium on Geoscience and Remote Sensing (IGARSS) p.4486-4490
Abstract

Dissolved Organic Carbon (DOC) is a key component of the inland water carbon cycle, playing a critical role in water quality, aquatic ecosystem dynamics, and carbon flux assessments. In this study, we combine Sentinel-2 Multi-Spectral Instrument (MSI) data with in-situ water quality, meteorological, and temporal variables to predict DOC concentrations in Lake Erken, Sweden. Three Machine Learning (ML) models, eXtreme Gradient Boosting (XGBoost), Random Forest Regression (RFR), and Gaussian Process Regression (GPR), are evaluated. XGBoost achieved the best performance with a MAPE of 3.46% and RMSE of 0.45 mg C/L. To enhance interpretability, we applied SHapley Additive exPlanations (SHAP) analysis, which identified that the year (Y),... (More)

Dissolved Organic Carbon (DOC) is a key component of the inland water carbon cycle, playing a critical role in water quality, aquatic ecosystem dynamics, and carbon flux assessments. In this study, we combine Sentinel-2 Multi-Spectral Instrument (MSI) data with in-situ water quality, meteorological, and temporal variables to predict DOC concentrations in Lake Erken, Sweden. Three Machine Learning (ML) models, eXtreme Gradient Boosting (XGBoost), Random Forest Regression (RFR), and Gaussian Process Regression (GPR), are evaluated. XGBoost achieved the best performance with a MAPE of 3.46% and RMSE of 0.45 mg C/L. To enhance interpretability, we applied SHapley Additive exPlanations (SHAP) analysis, which identified that the year (Y), total phosphorus (TP), total nitrogen (TN), and Sentinel-2 Band 2 (B2) were the most influential predictors of DOC. These results demonstrate the potential of integrating remote sensing and ML techniques for accurate DOC prediction in inland waters, thereby improving our understanding of the carbon cycle and providing insights to support sustainable water management.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
DOC, Inland waters, ML, Sentinel-2 MSI, SHAP, XGBoost
host publication
IGARSS 2025 : 2025 IEEE International Geoscience and Remote Sensing Symposium - 2025 IEEE International Geoscience and Remote Sensing Symposium
series title
IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)
pages
5 pages
publisher
IEEE - Institute of Electrical and Electronics Engineers Inc.
conference name
2025 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2025
conference location
Brisbane, Australia
conference dates
2025-08-03 - 2025-08-08
external identifiers
  • scopus:105033539615
ISSN
2153-7003
ISBN
979-8-3315-0810-4
DOI
10.1109/IGARSS55030.2025.11242290
language
English
LU publication?
yes
additional info
Publisher Copyright: ©2025 IEEE.
id
19ddd8bf-4229-4255-b1ee-baf15e852fb4
date added to LUP
2026-06-05 09:15:48
date last changed
2026-08-18 14:58:14
@inproceedings{19ddd8bf-4229-4255-b1ee-baf15e852fb4,
  abstract     = {{<p>Dissolved Organic Carbon (DOC) is a key component of the inland water carbon cycle, playing a critical role in water quality, aquatic ecosystem dynamics, and carbon flux assessments. In this study, we combine Sentinel-2 Multi-Spectral Instrument (MSI) data with in-situ water quality, meteorological, and temporal variables to predict DOC concentrations in Lake Erken, Sweden. Three Machine Learning (ML) models, eXtreme Gradient Boosting (XGBoost), Random Forest Regression (RFR), and Gaussian Process Regression (GPR), are evaluated. XGBoost achieved the best performance with a MAPE of 3.46% and RMSE of 0.45 mg C/L. To enhance interpretability, we applied SHapley Additive exPlanations (SHAP) analysis, which identified that the year (Y), total phosphorus (TP), total nitrogen (TN), and Sentinel-2 Band 2 (B2) were the most influential predictors of DOC. These results demonstrate the potential of integrating remote sensing and ML techniques for accurate DOC prediction in inland waters, thereby improving our understanding of the carbon cycle and providing insights to support sustainable water management.</p>}},
  author       = {{Wei, Xufeng and Xu, Wenbo and Wang, Lanhui and Guo, Renkui and Duan, Zheng}},
  booktitle    = {{IGARSS 2025 : 2025 IEEE International Geoscience and Remote Sensing Symposium}},
  isbn         = {{979-8-3315-0810-4}},
  issn         = {{2153-7003}},
  keywords     = {{DOC; Inland waters; ML; Sentinel-2 MSI; SHAP; XGBoost}},
  language     = {{eng}},
  pages        = {{4486--4490}},
  publisher    = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
  series       = {{IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)}},
  title        = {{Prediction of Dissolved Organic Carbon in Inland Waters Using Machine Learning Methods : A Case Study of Lake Erken, Sweden}},
  url          = {{http://dx.doi.org/10.1109/IGARSS55030.2025.11242290}},
  doi          = {{10.1109/IGARSS55030.2025.11242290}},
  year         = {{2025}},
}