Global Vision Media Focus
The company was featured on CCTV a total of 15 times, and its reach through Party media, central media, and local official media reached a total of 50 million people.
Quantitative inversion of chlorophyll a in lake water using remote sensing
Overview
Chlorophyll a Chlorophyll a is the main photosynthetic pigment of phytoplankton and is often used as an indicator of phytoplankton biomass, making it an important component of water environment management. Chlorophyll a concentration (Chla) is a key indicator for assessing the degree of eutrophication in water bodies and studying carbon cycling in aquatic environments; its rapid and accurate acquisition is crucial for lake ecological management.
Traditional methods for assessing lake trophic status primarily rely on in-situ sampling and analysis. However, this method is susceptible to local weather and environmental influences, and the sampling and testing process is time-consuming, labor-intensive, and costly, making it difficult to monitor lake eutrophication at fine spatiotemporal scales. In contrast, remote sensing technology offers advantages such as wide coverage, rapid data acquisition, and short processing times, overcoming many shortcomings of traditional water quality sampling. Therefore, it has been widely applied to the monitoring of lake water environment and aquatic ecology.
However, the optical characteristics of inland water bodies are highly regional and seasonal, and various algorithms for chlorophyll a inversion are still limited by seasons and geographical locations, resulting in inconsistencies in the inversion models established based on water quality parameters in different study areas. Therefore, this paper uses the previously established inversion models for different lakes to evaluate their universality.
data
Inversion model:
In the preliminary work, a binary linear function inversion model was established by using data from 20 lake sampling points on January 15, 2026, as well as Sentinel-2 satellite data passing over during the same period, through different combinations of wavebands.
Verification Lake
Using Sentinel-2 satellite data passing through during the same period, the data was analyzed for Luoma Lake and Chaohu Lake.Taiping LakeThe study primarily focused on Longgan Lake, Daguan Lake, Gucheng Lake, and Wabu Lake, collecting Chl-a concentration data from various stations as of January 15, 2026 (Table 1). The Sentinel-2 multispectral analyzer (MSI) Level-2A product was used, covering 10 bands (B1~B9, including B8A). Data was obtained from...Google Earth EngineObtain.
Evaluation methods
In the spatial matching process between remote sensing inversion results and measured station data, a 3×3 pixel window is used for collaborative extraction to improve verification accuracy and reduce the uncertainty caused by scale effects.
The specific implementation process is as follows:
(1) WindowLocation and ExtractionCentered on the coordinates of the monitoring station, locate its nearest neighbor pixel in the image, and extract all chlorophyll a inversion values within a 3×3 neighborhood window centered on that pixel.
(2) Quality control and statistics:
Remove invalid values within the window (such as cloud mask markers, land pixels, inverted values that exceed the physically reasonable range, etc.);
The number of valid pixels must be ≥5; otherwise, the data for that site will be marked as "match failed" and removed.
The arithmetic mean of the remaining effective pixels is calculated and used as the final remote sensing inversion representative value for the measured station;
The optimal model derived from previous work was directly applied to other lakes, and its general applicability was evaluated using the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE).
Coefficient of determination (R²):
Root Mean Square Error (RMSE):
Mean Absolute Error (MAE):
Results and Analysis
The previous inversion model was applied to each lake, and the following inversion maps were obtained:
Figure 1.Images of Sentinel-2 in Chaohu
Figure 2.Distribution map of chlorophyll a inversion results in Chaohu
Figure 3.Sentinel-2 images of Gucheng Lake
Figure 4.Distribution map of chlorophyll a inversion results in Gucheng Lake
Figure 5.Sentinel-2 images of Luoma Lake
Figure 6.Distribution map of chlorophyll a inversion results in Luoma Lake
Figure 7.Sentinel-2 images of Taiping Lake
Figure 8.Distribution map of chlorophyll a inversion results in Taiping Lake
The coefficient of determination (R²) was 0.7589, the root mean square error (RMSE) was 0.002143, and the mean absolute error (MAE) was 0.001189, reflecting that the inversion model exhibited "moderate fit and regional differentiation" characteristics in cross-lake validation. The inversion bias was relatively small in the estuary zone and low-concentration areas. The model at three stations showed insufficient sensitivity in the high-biomass area on the western shore of Chaohu Lake, possibly related to winter algal bloom residue and the spectral coupling effect of suspended matter-chlorophyll. Spatial distribution bias and...Optical properties of waterThe gradients are highly consistent—the high turbidity inversion system is systematically low, while the deep, clean water body (Taiping Lake) is relatively stable, confirming the remote sensing inversion rule that "water body type dominates model response." Overall, the model is highly adaptable to relatively homogeneous regions with water body optical backgrounds, and has a certain degree of regional applicability for inversion. The results are generally of practical value.
Because this inversion is based on single-day station data, and the concentration of chlorophyll a in water is low in winter and the image data is greatly affected by cloud cover, multiple periods of data will be collected in the future to increase the sample size for multiple inversions.