研究目的
Assessing the performance of sixteen schemes composed by QAA original and re-parameterized versions followed by models that use absorption coefficients as inputs for estimating Chl-a concentration in Ibitinga reservoir.
研究成果
QAAV5-based schemes provided reasonable Chl-a estimates, with CDOM dominating absorption. The reservoir showed eutrophic conditions, and models based on absorption coefficients were statistically equivalent. Future work should address temporal variability and improve algorithm robustness.
研究不足
High variability in water quality and optical properties limits algorithm applicability; atmospheric correction errors, especially at 412 nm; adjacency effects in satellite imagery; and the need for further validation under different environmental conditions.
1：Experimental Design and Method Selection:
The study involved assessing quasi-analytical algorithms (QAA) and absorption coefficient-based models for chlorophyll-a retrieval. Methods included field data collection, laboratory analysis, and remote sensing data processing.
2：Sample Selection and Data Sources:
Water samples and spectral data were collected from Ibitinga reservoir during two field campaigns (IBI1 and IBI2) in 2016 and 2017, with 29 and 6 sampling stations respectively. Satellite data from OLCI/Sentinel-3A were used.
3：List of Experimental Equipment and Materials:
Equipment included Secchi disk, turbidimeter (Hanna HI 98703), conductivity meter, spectrophotometer (Shimadzu 2600 UV-VIS), RAMSES sensors (TriOS), and filters (Whatman fiberglass and nylon membranes). Materials included acetone, hydrochloric acid, sodium hypochlorite, and water samples.
4：Experimental Procedures and Operational Workflow:
Procedures involved water sampling, filtration, spectrophotometric measurements for absorption coefficients, radiometric measurements for remote sensing reflectance, and application of QAA algorithms and models to estimate Chl-a. Image processing included atmospheric correction using empirical line method and application of retrieval schemes.
5：Data Analysis Methods:
Data were analyzed using linear regression for model calibration, leave-one-out cross-validation, and accuracy assessment with normalized root mean square error (nRMSE) and averaged unbiased absolute percentage difference (ε).
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
