Cross-scale chlorophyll-a monitoring in a canyon-type reservoir using Sentinel-2 and UAV hyperspectral observations

Abstract Monitoring chlorophyll-a (Chl-a) in morphologically complex reservoirs requires observations capable of capturing both reservoir-wide spatial patterns and fine-scale local variability. This study developed a cross-scale Sentinel-2–unmanned aerial vehicle (UAV) framework for Chl-a monitoring in Mianhuatan Reservoir, a canyon-type reservoir in southeastern China. Five field campaigns were conducted between July and October 2025, and field-measured Chl-a concentrations were matched independently with Sentinel-2 multispectral imagery and UAV hyperspectral observations. After quality control and spatial matching, 150 Sentinel-2–field samples and 47 UAV–field samples were retained for model development. Partial least squares regression, random forest, extreme gradient boosting, and support vector regression were evaluated separately for the two remote-sensing platforms using leave-one-out cross-validation. The optimal UAV support vector regression model achieved an R² of 0.81 and an RMSE of 4.72 µg L⁻¹, whereas the optimal Sentinel-2 XGBoost model achieved an R² of 0.65 and an RMSE of 5.93 µg L⁻¹. Sentinel-2 provided the reservoir-wide spatial background of Chl-a and revealed broad differences among reservoir sections, while UAV hyperspectral imagery resolved continuous concentration gradients and finer spatial heterogeneity within representative local waters. Under partial cloud-cover conditions, same-day UAV observations further provided spatially continuous high-resolution Chl-a information where the spatial completeness of Sentinel-2 retrievals was reduced. Overall, the proposed framework demonstrates the complementary value of Sentinel-2 and UAV observations across spatial scales and observation conditions. Rather than directly fusing the two products, the framework links reservoir-wide satellite monitoring with targeted UAV-based fine-scale characterization, providing a practical approach for multi-scale Chl-a monitoring in morphologically complex reservoirs.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-12
DOI
https://doi.org/10.21203/rs.3.rs-11488811/v1
Primary Topic
Water Quality Monitoring Technologies
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Cross-scale chlorophyll-a monitoring in a canyon-type reservoir using Sentinel-2 and UAV hyperspectral observations

Qiuting Zhang, Wenbin Pan, Chaojun Lin, Lvxin Qian et al.
Research Square
Water Quality Monitoring Technologies
preprint

Cross-scale chlorophyll-a monitoring in a canyon-type reservoir using Sentinel-2 and UAV hyperspectral observations

Qiuting Zhang, Wenbin Pan, Chaojun Lin, Lvxin Qian, Xuewen Chen, Miaohua Zhong, Kaixin Qiu, Fan Zhang
preprint en

Abstract

Abstract Monitoring chlorophyll-a (Chl-a) in morphologically complex reservoirs requires observations capable of capturing both reservoir-wide spatial patterns and fine-scale local variability. This study developed a cross-scale Sentinel-2–unmanned aerial vehicle (UAV) framework for Chl-a monitoring in Mianhuatan Reservoir, a canyon-type reservoir in southeastern China. Five field campaigns were conducted between July and October 2025, and field-measured Chl-a concentrations were matched independently with Sentinel-2 multispectral imagery and UAV hyperspectral observations. After quality control and spatial matching, 150 Sentinel-2–field samples and 47 UAV–field samples were retained for model development. Partial least squares regression, random forest, extreme gradient boosting, and support vector regression were evaluated separately for the two remote-sensing platforms using leave-one-out cross-validation. The optimal UAV support vector regression model achieved an R² of 0.81 and an RMSE of 4.72 µg L⁻¹, whereas the optimal Sentinel-2 XGBoost model achieved an R² of 0.65 and an RMSE of 5.93 µg L⁻¹. Sentinel-2 provided the reservoir-wide spatial background of Chl-a and revealed broad differences among reservoir sections, while UAV hyperspectral imagery resolved continuous concentration gradients and finer spatial heterogeneity within representative local waters. Under partial cloud-cover conditions, same-day UAV observations further provided spatially continuous high-resolution Chl-a information where the spatial completeness of Sentinel-2 retrievals was reduced. Overall, the proposed framework demonstrates the complementary value of Sentinel-2 and UAV observations across spatial scales and observation conditions. Rather than directly fusing the two products, the framework links reservoir-wide satellite monitoring with targeted UAV-based fine-scale characterization, providing a practical approach for multi-scale Chl-a monitoring in morphologically complex reservoirs.

Research Square
Longyan University (CN), Fuzhou University (CN)
Fujian Provincial Department of Science and Technology
Clean water and sanitation, Climate action
Water Quality Monitoring Technologies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon