Lake chlorophyll-a concentration and total suspended matter concentration remote sensing inversion model and its application

Abstract

The application discloses a lake chlorophyll-a concentration and total suspended matter concentration remote sensing inversion model, comprising the following steps: data downloading, data preprocessing, extracting and removing cloud coverage range, water body range extraction, TSM inversion and Chla inversion. The application further discloses an application of the lake chlorophyll-a concentration and total suspended matter concentration remote sensing inversion model, which is applied to Gaoyou Lake to monitor the chlorophyll-a concentration and total suspended matter concentration of water bodies in the Gaoyou Lake. The remote sensing inversion model can long-term, periodic and simple and rapid monitor the chlorophyll-a concentration and total suspended matter concentration of water bodies in the Gaoyou Lake, and can further rapidly evaluate the eutrophication degree of the Gaoyou Lake, thereby providing suggestions for scientific management and treatment of the Gaoyou Lake.

Description

Technical Field

[0001] This invention belongs to the field of water quality parameter monitoring technology, specifically involving a remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes. Background Technology

[0002] Gaoyou Lake is located in the core area of ​​the Jianghuai Ecological Corridor, and contains several national-level aquatic germplasm resource protection areas for species such as the large silverfish, anchovy, and freshwater shrimp, giving it an important ecological position. However, with frequent human activities, such as net enclosure aquaculture and agricultural irrigation, Gaoyou Lake's ecology has been seriously threatened. Gaoyou Lake is currently in a state of mild eutrophication. According to the 2011 Surface Water Environmental Quality Assessment Method (Trial Implementation), Chlorophyll a (ChLA) is used as the benchmark parameter for evaluating eutrophication indicators, indicating that ChLA is a highly indicative indicator of lake eutrophication levels. Incorporating the monitoring of Gaoyou Lake's eutrophication level into the routine monitoring system, especially the routine monitoring of parameters closely related to eutrophication levels such as chlorophyll a concentration (ChLA) and total suspended solids (TSM), is crucial for the scientific assessment of changes in Gaoyou Lake's water quality and ecology.

[0003] Currently, routine monitoring of Gaoyou Lake mainly includes manual sampling monitoring and automatic station monitoring. These traditional monitoring methods have problems such as long analysis cycles, low monitoring frequency, few monitoring points, and poor sample representativeness. Moreover, the monitoring content mainly targets the first 9 of the 24 basic parameters of surface water environmental quality, namely water temperature, pH, dissolved oxygen, conductivity, turbidity, permanganate index, ammonia nitrogen, total phosphorus, and total nitrogen. They do not directly indicate the degree of eutrophication of the water body, and Chla and TSM measurements need to be supplemented.

[0004] Remote sensing monitoring technology is characterized by its wide coverage, stable cycle, and low cost. Compared with traditional monitoring methods, it can quickly reflect the distribution of monitored targets on both spatial and temporal scales, making it more suitable for long-term dynamic monitoring. Currently, remote sensing model estimation methods for water quality parameters are divided into three types: empirical methods, semi-analytical methods, and analytical methods. Empirical methods are only effective for Class I water bodies. However, for Class II turbid inland water bodies, the dominant optical factors are often jointly dominated by chlorophyll a, total suspended particles, and yellow substances, making empirical methods insufficient for general applicability. While semi-empirical methods may perform well in a given year, they are difficult to apply to different years. Analytical methods require a large amount of data on quasi-synchronous meteorological conditions and the inherent optical characteristics of water bodies, which are difficult to obtain. In comparison, semi-analytical methods are currently the most widely used. Commonly used semi-analytical models for chlorophyll a concentration include the two-band method, the three-band method, the NDCI method, and the FLH method. Regarding total suspended solids (TSS) concentration, previous studies have found that when the spectral wavelength range is between 700 nm and 850 nm, the absorbance of both pigmented and non-pigmented particulate matter in the water is relatively weak, and at this time, the correlation between single-band reflectance and TSM concentration is relatively good. However, the Chla and TSM inversion models obtained from the above studies do not have universality across different lakes, and there is a lack of inversion models for relevant water quality parameters of Gaoyou Lake. Summary of the Invention

[0005] In view of this, this invention proposes a remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes and its application. This model enables long-term, periodic, simple, and rapid monitoring of chlorophyll a concentration and total suspended matter concentration in Gaoyou Lake, thereby allowing for rapid assessment of the eutrophication level of Gaoyou Lake and providing suggestions for the scientific management and governance of Gaoyou Lake.

[0006] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0007] On one hand, this invention relates to a remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes, comprising the following steps:

[0008] S1: Data download. Download Sentinel L2A level data from the Sentinel satellite website according to the time and range of the required image. This data is satellite remote sensing monitoring image that has undergone atmospheric correction.

[0009] S2: Data preprocessing. The L2A level data is preprocessed by using SNAP software to perform multi-size morphic stitching, subset, and Rsampling on the L2A level data. Finally, the data is converted into TIFF format files with a spatial resolution of 10m and bands including Band2, Band3, Band4, Band5, Band6, Band7, Band8, Band8A, and cloudconfidence from the L2A level data.

[0010] S3: Extract and remove cloud coverage range. Mark all pixels with values ​​greater than 0 in the cloud confidence band as cloud coverage and remove them in the subsequent inversion model.

[0011] S4: Water Body Extent Extraction. First, log in to the National Geographic Information Resources Catalog Service System and download vector map data covering the lake area from the 1:250,000 national basic geographic database to obtain the lake coverage area. Then, use the NDWI index to extract the water body portion within the lake area. When NDWI > 0, it indicates that the area is a water body.

[0012] The calculation method is: NDWI = (R B3 -R B8 ) / (R B8 +R B3 (1):

[0013] S5: TSM Inversion: TSM inversion is performed separately for the high-water season and the normal-water season of the lake. The high-water season refers to May to October each year, and the normal-water season refers to November to April of the following year. The inversion results are in mg / L. The inversion method is as follows:

[0014] Normal water level:

[0015] TSM = 60.7805 × (R) B4 / R B3 )+98.20857×(R B4 / R B2 )-1.8474×(R B3 / R B8 )-92.93144 (2)

[0016] High water season:

[0017] TSM = 3786.21606 × R B8 +167.9372×(R B6 / R B5 )+3.5764×(R B3 / R B8 -90.84456(3);

[0018] S6: Chla Inversion: Based on the TSM inversion results from the previous step, Chla is inverted using TSM = 20 mg / L as the threshold. The inversion results are in μg / L. The inversion method is as follows:

[0019] TSM < 20 mg / L:

[0020]

[0021] TSM ≥ 20 mg / L:

[0022]

[0023] The R used in formulas (1) to (5) above Bx This refers to the reflectance of the bandx in L2A-level data.

[0024] On the other hand, the present invention also relates to the application of a remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes, applied to Gaoyou Lake, to monitor the chlorophyll a concentration and total suspended matter concentration in the water of Gaoyou Lake.

[0025] The beneficial effects of this invention are as follows:

[0026] 1) The satellite data selected in this invention is Sentinel-2 MSI data, which is an Earth observation satellite in the European Space Agency's Copernicus program (GMES). This data is available free of charge globally, and the cost of acquiring the original data is almost zero, resulting in low costs.

[0027] 2) This invention provides a remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes with a high monitoring frequency. The Sentinel-2 satellite consists of two satellites, A and B, each with a revisit period of 10 days. The two satellites complement each other, with a revisit period of 5 days. Inversion monitoring and analysis of water quality parameters of Gaoyou Lake can be completed every 5 days.

[0028] 3) This invention is highly representative. It can be used to perform inversion analysis of Chla and TSM in the entire water body of Gaoyou Lake, with a spatial resolution of 10 meters, which can fully reflect the spatial distribution characteristics of the water quality parameters of the entire lake.

[0029] 4) The remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes proposed in this invention has higher universality, is applicable to different lakes and different years, and has high accuracy. Attached Figure Description

[0030] Figure 1 A technical roadmap for a remote sensing inversion model of lake chlorophyll a concentration and total suspended matter concentration provided by the present invention;

[0031] Figure 2 This is a map showing the distribution of the research area and sampling points in this invention.

[0032] Figure 3 This is a scatter plot for verifying the accuracy of the total suspended solids concentration inversion model during the normal water period in this embodiment of the invention.

[0033] Figure 4 This is a scatter plot for verifying the accuracy of the total suspended solids concentration inversion model during the high-water season in an embodiment of the present invention.

[0034] Figure 5 This is a scatter plot verifying the accuracy of the chlorophyll a concentration inversion model when TSM < 20 mg / L in an embodiment of the present invention.

[0035] Figure 6 This is a scatter plot verifying the accuracy of the chlorophyll a concentration inversion model when TSM ≥ 20 mg / L in an embodiment of the present invention.

[0036] Figure 7 This is a scatter plot comparing satellite-estimated values ​​and ground-measured values ​​of chlorophyll a concentration in Gaoyou Lake according to an embodiment of the present invention.

[0037] Figure 8 This is a scatter plot comparing satellite-estimated values ​​and ground-measured values ​​of total suspended matter concentration in Gaoyou Lake, as described in an embodiment of the present invention. Detailed Implementation

[0038] To provide a more detailed understanding of the features and technical content of this invention, the implementation of this invention will be described in detail below with reference to the accompanying drawings. The accompanying drawings are for reference only and are not intended to limit this invention.

[0039] This invention provides a remote sensing inversion model for lake chlorophyll a concentration and total suspended matter concentration, such as... Figure 1 As shown, the remote sensing inversion model includes the following steps:

[0040] S1: Data download. Download Sentinel L2A level data, which is atmospherically corrected satellite remote sensing imagery, from the Sentinel satellite website (https: / / scihub.copemicus.eu / ) according to the time and range of the required imagery.

[0041] S2: Data preprocessing. The L2A level data is preprocessed by using SNAP software to perform multi-size morphic stitching, subset, and Rsampling on the L2A level data. Finally, the data is converted into TIFF format files with a spatial resolution of 10m and bands including Band2, Band3, Band4, Band5, Band6, Band7, Band8, Band8A, and cloudconfidence from the L2A level data.

[0042] S3: Extract and remove cloud coverage range. Mark all pixels with values ​​greater than 0 in the cloud confidence band as cloud coverage and remove them in the subsequent inversion model.

[0043] S4: Water Body Extent Extraction. First, log in to the National Geographic Information Resources Catalog Service System and download vector map data covering the lake area from the 1:250,000 national basic geographic database to obtain the lake coverage area. Then, use the NDWI index to extract the water body portion within the lake area. When NDWI > 0, it indicates that the area is a water body.

[0044] The calculation method is: NDWI = (R B3 -R B8 ) / (R B8 +R B3 (1);

[0045] S5: TSM Inversion: TSM inversion is performed separately for the high-water season and the normal-water season of the lake. The high-water season refers to May to October each year, and the normal-water season refers to November to April of the following year. The inversion results are in mg / L. The inversion method is as follows:

[0046] Normal water level:

[0047] TSM = 60.7805 × (R) B4 / R B3 )+98.20857×(R B4 / R B2 )-1.8474×(R B3 / R B8 )-92.93144 (2)

[0048] High water season:

[0049] TSM = 3786.21606 × R B8 +167.9372×(R B6 / R B5 )+3.5764×(R B3 / R B8 -90.84456(3);

[0050] S6: Chla Inversion: Based on the TSM inversion results from the previous step, Chla is inverted using TSM = 20 mg / L as the threshold. The inversion results are in μg / L. The inversion method is as follows:

[0051] TSM < 20 mg / L:

[0052]

[0053] TSM ≥ 20 mg / L:

[0054]

[0055] The R used in formulas (1) to (5) above Bx This refers to the reflectance of the bandx in L2A-level data, R. B2 R is the reflectivity of band 2. B3 R represents the reflectivity of band 3. B4 R represents the reflectivity of band 4. B5 R represents the reflectivity of band 5. B8 This represents the reflectivity of band 8.

[0056] Specifically, taking the remote sensing monitoring of Gaoyou Lake as an example, the specific implementation of the present invention will be further explained.

[0057] The remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in Gaoyou Lake includes the following steps:

[0058] S1: Data download. Download Sentinel L2A level data, which is satellite remote sensing monitoring imagery that has undergone atmospheric correction, from the Sentinel satellite website (https: / / scihub.copernicus.eu / ) according to the time and scope of the required imagery.

[0059] S2: Data preprocessing. The L2A level data is preprocessed by using SNAP software to perform multi-size morphic stitching, subset, and Rsampling on the L2A level data. Finally, the data is converted into TIFF format files with a spatial resolution of 10m and bands including Band2, Band3, Band4, Band5, Band6, Band7, Band8, Band8A, and cloudconfidence from the L2A level data.

[0060] S3: Extract and remove cloud coverage range. Mark all pixels with values ​​greater than 0 in the cloud confidence band as cloud coverage and remove them in the subsequent inversion model.

[0061] S4: Water Body Extent Extraction. First, log in to the National Geographic Information Resources Catalog Service System and download vector map data covering the Gaoyou Lake area from the 1:250,000 national basic geographic database to obtain the coverage area of ​​Gaoyou Lake. Then, use the NDWI index to extract the water body portion within the Gaoyou Lake area. When NDWI > 0, it indicates that the area is a water body.

[0062] The calculation method is: NDWI = (R B3 -R B8 ) / (RB8 +R B3 (1);

[0063] S5: TSM Inversion: TSM inversion was performed separately for the high-water season and the normal-water season of Gaoyou Lake. The high-water season refers to May to October each year, and the normal-water season refers to November to April of the following year. The inversion results are in mg / L, and the inversion method is as follows:

[0064] Normal water level:

[0065] TSM = 60.7805 × (R) B4 / R B3 )+98.20857×(R B4 / R B2 )-1.8474×(R B3 / R B8 )-92.93144 (2)

[0066] High water season:

[0067] TSM = 3786.21606 × R B8 +167.9372×(R B6 / R B5 )+3.5764×(R B3 / R B8 -90.84456(3);

[0068] S6: Chla Inversion: Based on the TSM inversion results from the previous step, Chla is inverted using TSM = 20 mg / L as the threshold. The inversion results are in μg / L. The inversion method is as follows:

[0069] TSM < 20 mg / L:

[0070]

[0071] TSM ≥ 20 mg / L:

[0072]

[0073] To verify the inversion accuracy of this invention, four sampling operations were conducted on Gaoyou Lake in March 2019, September 2019, April 2022, and October 2022. There were 24 sampling points in 2019 and 64 sampling points in 2022, resulting in a total of 176 sets of measured data. The sampling points are as follows: Figure 2 As shown in the figure. After removing some water samples with Chla concentrations below the standard detection limit and some sampling points where spectral measurements were affected by lake bottom aquatic plants, a total of 160 sampling points were available for model simulation and validation. Two-thirds of the above sampling data were used for model simulation, and one-third was used for model validation.

[0074] The root mean squared error (RMSE) and mean relative error (MRE) were used to validate and compare the model accuracy.

[0075]

[0076]

[0077] In the formula: y represents the inversion result; y represents the ground-based measured result; n represents the verification number.

[0078] Validation results for the TSM inversion model:

[0079] A total of 81 sets of measured data were used for TSM inversion during the normal water period, of which 55 sets were used for model simulation (14 sets in 2019 and 41 sets in 2022) and 26 sets were used for model validation (6 sets in 2019 and 20 sets in 2022). A total of 79 sets of measured data were used for inversion during the high water period, of which 53 sets were used for model simulation (15 sets in 2019 and 38 sets in 2022) and 26 sets were used for model validation (6 sets in 2019 and 20 sets in 2022). The validation results are as follows: Figure 3 , Figure 4 As shown.

[0080] The overall TSM inversion accuracy is RMSE = 16.2 mg / L and MRE = 23.3%, which meets the TSM inversion accuracy requirements.

[0081] Validation results for the Chla inversion model:

[0082] A total of 28 models met the TSM < 20 mg / L criteria, of which 19 were used for model simulation (8 in 2019, 11 in 2022) and 9 were used for model validation (4 in 2019, 5 in 2022). A total of 132 models met the TSM ≥ 20 mg / L criteria, of which 89 were used for model simulation (17 in 2019, 72 in 2022) and 43 were used for model validation (8 in 2019, 35 in 2022). Validation results are as follows: Figure 5 , Figure 6 As shown.

[0083] The overall statistical inversion accuracy is RMSE = 8.02 ug / L and MRE = 18.4%, which meets the Chla inversion accuracy requirements.

[0084] By comparing the sampling time with the measured data, nine sets of quasi-synchronous measured data from September 24, 2019, and April 6, 2022, were found to match the remote sensing inversion data. These nine sets of data were used to verify the model accuracy of the satellite inversion results. A comparison of the two sets of data is shown below. Figure 7 , Figure 8 As shown, the Chla inversion error RMSE = 14.5 ug / L, MRE = 25.7%, and the TSM inversion error RMSE = 23.4 mg / L, MRE = 31.9%. This inversion accuracy is acceptable.

[0085] The remote sensing inversion model of this invention selects Sentinel-2 MSI satellite data, which is an Earth observation satellite in the European Space Agency's Copernicus Mission (GMES). This data is available free of charge globally, with almost zero cost in acquiring the original data, resulting in low overall costs. Furthermore, the monitoring frequency is high. Sentinel-2 consists of two satellites, A and B, each with a revisit period of 10 days. The two satellites complement each other, resulting in a revisit period of 5 days. This allows for the inversion monitoring and analysis of lake water quality parameters every 5 days, enabling long-term, periodic, simple, and rapid monitoring of chlorophyll a concentration and total suspended solids concentration in lake water. This allows for rapid assessment of lake eutrophication levels, providing insights for the scientific management and governance of lakes.

[0086] This invention is highly representative. It can be used to perform inversion analysis of Chla and TSM in the entire water body of Gaoyou Lake, with a spatial resolution of 10 meters, which can fully reflect the spatial distribution characteristics of the water quality parameters of the entire lake.

[0087] The remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes proposed in this invention has higher universality, is applicable to different lakes and different years, and has high accuracy.

Claims

1. A remote sensing inversion model for chlorophyll a concentration and total suspended matter concentration in lakes, characterized in that, Includes the following steps: S1: Data download. Download Sentinel L2A level data from the Sentinel satellite website according to the time and range of the required image. This data is satellite remote sensing monitoring image that has undergone atmospheric correction. S2: Data preprocessing. The L2A level data is preprocessed by using SNAP software to perform multi-size morphic stitching, subset, and Rsampling on the L2A level data. Finally, the data is converted into TIFF format files with a spatial resolution of 10m and bands including Band2, Band3, Band4, Band5, Band6, Band7, Band8, Band8A, and cloudconfidence from the L2A level data. S3: Extract and remove cloud coverage range. Mark all pixels with values ​​greater than 0 in the cloud confidence band as cloud coverage and remove them in the subsequent inversion model. S4: Water Body Extent Extraction. First, log in to the National Geographic Information Resources Catalog Service System and download vector map data covering the lake area from the 1:250,000 national basic geographic database to obtain the lake coverage area. Then, use the NDWI index to extract the water body portion within the lake area. When NDWI > 0, it indicates that the area is a water body. The calculation method is: NDWI = (R B3 -R B8 ) / (R B8 +R B3 (1); S5: TSM Inversion: TSM inversion is performed separately for the high-water season and the normal-water season of the lake. The high-water season refers to May to October each year, and the normal-water season refers to November to April of the following year. The inversion results are in mg / L. The inversion method is as follows: Normal water level: TSM=60.7805×(R B4 / R B3 )+98.20857×(R B4 / R B2 )-1.8474× (R B3 / R B8 )-92.93144(2) High water season: TSM=3786.21606×R B8 +167.9372×(R B6 / R B5 ))+3.5764× (R B3 / R B8 )-90.84456(3); S6: Chla Inversion: Based on the TSM inversion results from the previous step, Chla is inverted using TSM = 20 mg / L as the threshold. The inversion results are in μg / L. The inversion method is as follows: TSM < 20 mg / L: TSM ≥ 20 mg / L: The R used in formulas (1) to (5) above Bx This refers to the reflectance of the bandx in L2A-level data; The root mean square error and the mean relative error were used to verify and compare the model accuracy. In the formula: y represents the inversion result; y represents the ground-based measured result; n represents the validation number; Validation results for the TSM inversion model: A total of 81 sets of measured data were used for TSM inversion during the normal water period, of which 55 sets were used for model simulation and 26 sets were used for model validation; a total of 79 sets of measured data were used for inversion during the high water period, of which 53 sets were used for model simulation and 26 sets were used for model validation. The validation results are as follows: The overall TSM inversion accuracy is RMSE = 16.2 mg / L and MRE = 23.3%, which meets the TSM inversion accuracy requirements. Validation results for the Chla inversion model: A total of 28 models conforming to TSM < 20 mg / L were used for model simulation (19 groups) and model validation (9 groups). A total of 132 models conforming to TSM ≥ 20 mg / L were used for model simulation (89 groups) and model validation (43 groups). The validation results are as follows: The overall statistical inversion accuracy is RMSE = 8.02 ug / L and MRE = 18.4%, which meets the Chla inversion accuracy requirements.

2. The application of the remote sensing inversion model for lake chlorophyll a concentration and total suspended matter concentration according to claim 1, characterized in that, It was applied to Gaoyou Lake to monitor the concentrations of chlorophyll a and total suspended solids in the lake water.

Images