Remote estimation of phycocyanin concentration in inland waters based on optical classification

Area of Science:

• Environmental Science
• Remote Sensing
• Aquatic Ecology

Background:

• Cyanobacteria blooms threaten aquatic ecosystems due to climate change and human activities.
• Phycocyanin (PC) is a key pigment for detecting and warning of cyanobacteria blooms.
• Estimating PC concentration in turbid waters using remote sensing is challenging due to optical complexity.

Purpose of the Study:

• To develop and validate a remote sensing framework for accurate PC estimation in optically complex inland waters.
• To improve early warning systems for cyanobacteria blooms.

Main Methods:

• Collected 640 in situ pigment concentration and Ocean and Land Color Instrument (OLCI) reflectance data from Chinese lakes and reservoirs (2020-2022).
• Developed a water optical classification algorithm using remote sensing reflectance (Rrs) baseline height at 560, 647, and 709 nm.
• Compared three candidate algorithms (baseline height, band ratio, three-band) within five classified water types.

Main Results:

• The optical classification framework significantly improved PC estimation accuracy (RMSE = 72.6 μg L⁻¹, MAPE = 80.4%) compared to unclassified waters, especially for low PC concentrations.
• The band ratio algorithm showed broad applicability in medium turbid and clean waters.
• The three-band and line height algorithms were suitable for medium turbid and high PC waters, respectively.
• Classified water types correlated with PC/Chla ratios, enabling cyanobacteria risk assessment.

Conclusions:

• The proposed framework enhances PC estimation accuracy in optically complex inland waters.
• This study provides a novel strategy for water quality inversion and cyanobacteria bloom monitoring in inland aquatic environments.