A method and system for calculating the influence degree of extreme rainfall on the turbidity of a lake shore zone

Abstract

The application discloses a kind of methods and systems for calculating the influence of extreme rainfall on lake shore zone turbidity, belong to the technical field of environmental science and monitoring technology field.The method comprises: identifying extreme rainfall event through meteorological observation data;Based on multi-source remote sensing data and measured data, a lake turbidity remote sensing inversion model is constructed;The lake shore zone area is defined by combining shoreline extraction and spatial buffer method;Before and after the extreme rainfall event, the turbidity is calculated in time sequence using the inversion model, and the turbidity change intensity is obtained, so as to quantify the influence of extreme rainfall on lake shore zone turbidity.The application can provide scientific basis for lake water environment monitoring and watershed hydrological management, with the advantages of high precision, high degree of automation, wide applicability.

Description

Technical Field

[0001] This invention belongs to the technical field of environmental science and monitoring technology, specifically relating to a method and system for calculating the impact of extreme rainfall on turbidity in lake shore zones. Background Technology

[0002] As vital freshwater reservoirs, lakes directly impact regional water security and the stability of aquatic ecosystems. Lakeshore zones serve as transitional areas between land and water, playing crucial roles in material cycling, nutrient retention, and ecological buffering. However, these areas are also highly sensitive to external disturbances, particularly the impact of extreme weather events.

[0003] In recent years, the frequency and intensity of extreme rainfall events have been increasing against the backdrop of global climate change. Heavy rainfall not only leads to a surge in runoff within watersheds but also carries large amounts of sediment, nutrients, and organic debris into lakes, causing a significant increase in turbidity in the lakeshore waters. This rapid change in turbidity weakens light conditions in the water, inhibits photosynthesis by benthic plants and algae, and may also lead to the resuspension and release of pollutants from lakes, thereby impacting the health of lake ecosystems and water quality safety.

[0004] Existing research on lake turbidity changes mainly focuses on long-term water quality monitoring and statistical analysis, or numerical simulations based on hydrological processes. However, these methods have the following limitations when dealing with extreme rainfall events: Insufficient timeliness: Traditional water quality monitoring methods rely heavily on periodic sampling and laboratory analysis, making it difficult to capture rapid changes in turbidity during extreme rainfall events.

[0005] Limited spatial resolution: The number of conventional monitoring points is limited, making it difficult to fully reflect the spatial heterogeneity of the lake shoreline.

[0006] The coupling mechanism is unclear: Most existing models focus on watershed sediment transport and runoff processes, lacking a comprehensive analysis of the interaction mechanism between extreme rainfall and hydrodynamics, material transport and turbidity changes in the nearshore area of ​​lakes.

[0007] Therefore, there is an urgent need to propose a new method that can combine meteorological data, hydrological processes, and water quality changes to quantitatively and dynamically calculate the impact of extreme rainfall events on the turbidity of lake shorelines, and to develop a corresponding system to provide scientific support for early warning of lake water environment risks and ecological protection. Summary of the Invention

[0008] To address the technical problems mentioned above, this invention provides a method and system for calculating the impact of extreme rainfall on turbidity in lake shore zones.

[0009] This invention employs the following technical solution: a method for calculating the impact of extreme rainfall on turbidity in lake shore zones, comprising the following steps: The lake shoreline zone is determined based on the proportional internal buffer method of the lake's equivalent radius. The determined lake shoreline zone is used as a spatial constraint to select the target study area. Continuous historical precipitation data of the target area is obtained, and extreme rainfall thresholds are determined based on the continuous historical precipitation data. Obtain the daily rainfall in the target area. If the daily rainfall is greater than or equal to the extreme rainfall threshold, it is determined to be an extreme rainfall event. Based on the extreme rainfall event, a complete rainfall event is determined based on the entire time period of continuous historical precipitation data. A remote sensing inversion model for lake water turbidity was constructed, and the mean turbidity value before a complete rainfall event was calculated using the model. Mean turbidity of the post-state The mean turbidity of the previous state Mean turbidity of the post-state The influence of turbidity in the lake shore zone was calculated. and trends of change.

[0010] In a further embodiment, the process for determining the lake shoreline is as follows: Calculate the lake area based on the lake's vector boundary. Equivalent circle radius and the inward distance d; Applying a negative shape buffer operation to the lake polygon P yields a shrunken polygon. The coastal zone vector layer of the original polygon and the shrunken polygon is calculated by using the vector difference between the original polygon and the shrunken polygon. : This yields a ring-shaped region with a width of d extending inward from the shoreline, which is the lake shoreline zone. This is for vector interpolation operations.

[0011] In a further embodiment, the process for determining the complete rainfall event is as follows: Simultaneous recording of event windows based on extreme rainfall events ,in, The day that marks the beginning of the extreme rainfall event. The end date of the extreme rainfall event; will cover the event window. The image is marked as ; Starting date Using this as a boundary, trace back along the time frame and select the nearest qualified remote sensing image observation window as the previous state image, and mark it as... ; End date Using this as a boundary, we trace back along the time frame and select the nearest qualified remote sensing image observation window as the post-state image, marking it as... ; Will simultaneously contain the marked 、 and The corresponding event is defined as a complete rainfall event.

[0012] In a further embodiment, the construction process of the lake water turbidity remote sensing inversion model is as follows: Obtain the remote sensing reflectance of the spectrometer and acquire the blue light band. Green light band Red light band and near-infrared band Obtain the corresponding water-based remote sensing reflectance. , The water-reflectance The quality is processed to obtain the first type of input features; Based on the water-free remote sensing reflectivity Multiple binary ratio features are constructed by combining different bands, and these multiple binary ratio features are used as the second type of input features. Based on the blue light band Green light band Red light band Select the corresponding representative wavelengths respectively 、 and ; Representative wavelengths were calculated using the appropriate formulas. 、 and Corresponding total absorption coefficient , , and backscattering coefficient 、 、 , and serve as the third type of input feature; The machine learning method based on the random forest algorithm uses the first type of input features, the second type of input features, and the third type of input features as input data. The input data and the measured turbidity data are preprocessed to construct training and validation sets, and a remote sensing inversion model of lake water turbidity is trained and established.

[0013] In a further embodiment, the mean turbidity of the pre-state The calculation process is as follows: The turbidity of the lake shoreline at the corresponding time period in the pre-state image was obtained using a remote sensing inversion model of lake water turbidity. ; The mean turbidity of the initial state was calculated based on the median method. : , This is a vector layer representing the coastal zone.

[0014] In a further embodiment, the mean turbidity of the post-state... The calculation process is as follows: The turbidity of the lake shoreline at the corresponding time period in the post-state image was obtained using a remote sensing inversion model of lake water turbidity. ; The mean turbidity of the later state was calculated based on the median method. : , This is a vector layer representing the coastal zone.

[0015] In a further embodiment, the turbidity influence of the lake shoreline is... The calculation formula is as follows: ; The method for judging the trend of change is as follows: If The trend is determined to be an increase in turbidity in the coastal zone after extreme rainfall. The trend is determined to be a decrease in turbidity in the coastal zone after extreme rainfall.

[0016] In a further embodiment, the water-sensitive reflectance The acquisition process is as follows: Based on the acquired blue light band Green light band Red light band and near-infrared band Remove pixels with cloud cover, shadows, strong aerosols, and snow cover; The following formula is used to convert the bands to their corresponding water-free remote sensing reflectance. : In the formula, For satellite corresponding band Apparent reflectance.

[0017] In a further embodiment, the preprocessing of input data and measured turbidity data is spatiotemporal matching. The spatiotemporal matching principles include: spatial matching principle, temporal matching principle, multi-measured value matching processing principle, and matching data validity verification principle.

[0018] A system for calculating the impact of extreme rainfall on turbidity in lake shore zones, used to implement the calculation method described above, includes: The first module is set to determine the lake shore zone using a proportional internal buffer method based on the lake's equivalent radius, and select the target study area using the determined lake shore zone as a spatial constraint; acquire continuous historical precipitation data of the target area, and determine the extreme rainfall threshold based on the continuous historical precipitation data; The second module is set to obtain the daily rainfall in the target area. If the daily rainfall is greater than or equal to the extreme rainfall threshold, it is determined to be an extreme rainfall event. Based on the extreme rainfall event, the complete rainfall event is determined based on the entire time period of continuous historical precipitation data. The third module is configured to construct a remote sensing inversion model of lake water turbidity, and use this model to calculate the mean turbidity value before a complete rainfall event. Mean turbidity of the post-state The mean turbidity of the previous state Mean turbidity of the post-state The influence of turbidity in the lake shore zone was calculated. and trends of change.

[0019] The beneficial effects of this invention are significant: It boasts substantial advantages in automation and timeliness. Leveraging the batch processing capabilities of MODIS (Medium Resolution Imaging Spectroradiometer) remote sensing images and the real-time acquisition characteristics of meteorological data, this invention eliminates the need for manual on-site sampling and laboratory analysis, enabling rapid response to extreme rainfall events. From determining extreme rainfall thresholds and screening qualified images to coastal turbidity inversion and impact intensity calculation, the entire process can be automated, significantly shortening the turbidity impact assessment cycle and effectively solving the timeliness problem of traditional methods failing to capture rapid turbidity changes after extreme rainfall.

[0020] This invention differs from traditional monitoring methods that rely on limited monitoring points and struggle to reflect the spatial heterogeneity of the shoreline. It utilizes a shoreline vectorization definition method based on equivalent circle radius and indentation vector difference, combined with the spatial continuity of MODIS imagery, to achieve turbidity calculation across the entire lake shoreline. Furthermore, this invention ensures spatial resolution of turbidity data through pixel-level inversion, accurately characterizing turbidity differences in different areas of the shoreline (such as nearshore shallow waters and estuary areas), providing high-resolution data support for analyzing the spatial impact of extreme rainfall on the shoreline.

[0021] This invention overcomes the limitations of single spectral features in the construction of turbidity inversion models by introducing both MODIS multi-band raw reflectance and band ratios, and combining physical optical parameters such as absorption coefficient and scattering coefficient obtained from water color semi-analytical algorithms, thus enriching the model input dimensions. At the same time, it employs the nonlinear fitting capability of the random forest algorithm and grid search hyperparameter optimization to effectively reduce the impact of complex water optical properties (such as sediment content and chlorophyll interference) on the inversion results, ensuring the reliability of turbidity inversion results and providing an accurate data foundation for subsequent influence intensity calculations.

[0022] This invention outputs the intensity of the impact of extreme rainfall on turbidity in the coastal zone as a percentage, which not only intuitively reflects the trend of turbidity increase / decrease, but also quantifies the degree of impact through percentage values. In addition, this calculation method does not rely on the lake's own turbidity benchmark value, which can eliminate the benchmark differences between different lakes (such as high-turbidity lakes and low-turbidity lakes) and different periods of the same lake, and realize the comparison of the impact of extreme rainfall across lakes and across time scales, providing a unified evaluation standard for watershed-scale lake water environment management. Attached Figure Description

[0023] Figure 1 This is a schematic diagram illustrating the determination of the lake shoreline in Example 1.

[0024] Figure 2 This is a schematic diagram illustrating the definition of extreme rainfall in Example 1.

[0025] Figure 3 This refers to the training and testing accuracy of the global lake turbidity inversion model in Example 1. Detailed Implementation

[0026] The present invention will now be further described with reference to the accompanying drawings and embodiments.

[0027] Example 1 Combination Figures 1 to 3 This embodiment discloses a method for calculating the impact of extreme rainfall on the turbidity of lake shore zones, including the following steps: The lake shoreline zone is determined using a proportional internal buffering method based on the lake's equivalent radius. This method, grounded in lake morphology, adaptively sets the buffer distance according to the lake's equivalent radius, achieving parameterization and adjustability of the shoreline zone width. Using the determined lake shoreline zone as a spatial constraint, a target study area is selected; continuous historical precipitation data for the target area is acquired, and extreme rainfall thresholds are determined based on this data. Obtain the daily rainfall in the target area. If the daily rainfall is greater than or equal to the extreme rainfall threshold, it is determined to be an extreme rainfall event. Based on the extreme rainfall event, a complete rainfall event is determined based on the entire time period of continuous historical precipitation data. A remote sensing inversion model for lake water turbidity was constructed, and the mean turbidity value before a complete rainfall event was calculated using the model. Mean turbidity of the post-state The mean turbidity of the previous state Mean turbidity of the post-state The influence of turbidity in the lake shore zone was calculated. and trends of change.

[0028] In this embodiment, a custom lake shoreline zone was defined, primarily because traditional methods typically generate lake buffer zones using fixed or random distances, neglecting the differences in morphology, area scale, and shoreline complexity among different lakes, resulting in shoreline zone delineation lacking specificity and scientific rigor. This method constructs a lake shoreline zone with customizable parameters through morphological negative buffering operations, achieving accurate spatial characterization of the shoreline region while fully considering the lake's shape characteristics. This method not only improves the geometric accuracy and ecological representativeness of shoreline zone extraction, providing a unified and adjustable technical standard for shoreline region delineation of different types of lakes, but also provides a more reasonable and scientific spatial boundary for subsequent shoreline process simulation, remote sensing monitoring, ecological function assessment, and cross-lake comparative studies. The specific determination process is as follows: Calculate the lake area based on the lake's vector boundary. Equivalent circle radius and inward distance d The equivalent circle radius in this embodiment and inward distance d The calculation formula is as follows: , .

[0029] Applying a negative shape buffer operation to the lake polygon P yields a shrunken polygon. The coastal zone vector layer of the original polygon and the shrunken polygon is calculated by using the vector difference between the original polygon and the shrunken polygon. : This yields a ring-shaped region with a width of d extending inward from the shoreline, which is the lake shoreline zone. This is for vector interpolation operations.

[0030] It should be noted that the vector difference calculation described in this embodiment can be performed using the buffer tool in a GIS platform (such as ArcGIS, QGIS, or the GDAL / GEOS library). Furthermore, in conjunction with... Figure 1As shown, the obtained coastal zone vector layer L is stored in a standard GIS data format, such as Shapefile, GeoJSON, or Geopackage, and then visualized on a map platform for subsequent turbidity calculation and analysis.

[0031] In this embodiment, the continuous historical precipitation data is a 20-year (2001-2020) historical sequence, which is used as the baseline period. Daily precipitation data is used to average the precipitation over the target area to determine the extreme rainfall threshold. The specific implementation method is as follows: an empirical distribution is constructed using the baseline period data, and the 95th quantile of daily precipitation is calculated as the extreme rainfall threshold. .

[0032] make Rainfall for the day ,like If so, it is determined to be an extreme rainfall event.

[0033] Based on the above definition, the process for determining a complete rainfall event is as follows: Simultaneous recording of event windows based on extreme rainfall events ,in, The day that marks the beginning of the extreme rainfall event. The end date of the extreme rainfall event; will cover the event window. The image is marked as ; Starting date Using this as a boundary, trace back along the time frame and select the nearest qualified remote sensing image observation window as the previous state image, and mark it as... The qualified remote sensing image described in this embodiment is preprocessed using the MODIS image quality flag (QA flag) to remove pixels covered by clouds, cloud shadows, strong aerosols, and ice and snow. Based on this, the proportion of effective pixels within the lake shoreline area in the image is counted. When the proportion of effective pixels is ≥50%, it is judged as a qualified remote sensing image.

[0034] End date Using this as a boundary, we trace back along the time frame and select the nearest qualified remote sensing image observation window as the post-state image, marking it as... ; Will simultaneously contain the marked 、 and The corresponding event is defined as a complete rainfall event.

[0035] In a further embodiment, the construction process of the lake water turbidity remote sensing inversion model is as follows: Obtain the remote sensing reflectance of the spectrometer and acquire the blue light band. Green light band Red light band and near-infrared band Obtain the corresponding water-based remote sensing reflectance. , The water-reflectance Quality processing is performed to obtain the first type of input features; the blue light band described in this embodiment. To be close to 469nm wavelength, green light band To be close to 555nm wavelength, red light band Close to 645nm wavelength, near-infrared band To be close to 859 nm The wavelength.

[0036] Based on the water-free remote sensing reflectivity Four binary ratio features are constructed by combining different bands, and these four binary ratio features are used as the second type of input features. Further, the ratio features include: , , and The purpose of introducing a second type of input feature in this embodiment is to enhance the sensitivity differences between different spectral bands to changes in water turbidity. Compared to single-band reflectance, which can only reflect local spectral responses, by constructing... , The equal band ratio can amplify the differences in spectral characteristics caused by changes in turbidity (e.g., when turbidity increases, the reflectance of the red band increases significantly more than that of the green band, and the ratio can more clearly reflect this change), effectively reduce the interference of other optically active substances such as chlorophyll and dissolved organic matter in the water, improve the model's ability to identify and respond to turbidity signals, and lay the foundation for the subsequent random forest algorithm to accurately establish the mapping relationship between spectral characteristics and turbidity.

[0037] Based on the blue light band Green light band Red light band Select the corresponding representative wavelengths respectively 、 and Furthermore, such as , , .

[0038] Representative wavelengths were calculated using the appropriate formulas. 、 and Corresponding total absorption coefficient 、 、 and backscattering coefficient 、 、 This is used as a third type of input feature; through the third type of input feature, the model's ability to interpret turbidity changes and the accuracy of inversion are improved.

[0039] The machine learning method based on the random forest algorithm uses the first type of input features, the second type of input features, and the third type of input features as input data. The input data and the measured turbidity data are preprocessed to construct training and validation sets, and a remote sensing inversion model of lake water turbidity is trained and established.

[0040] It should be noted that the total absorption coefficient in this embodiment... 、 、 and backscattering coefficient 、 、 Different calculation methods are used depending on the wavelength.

[0041] First, the overall absorption coefficient and backscattering coefficient The calculation formula is as follows: ; ; In the formula, and Wavelength and wavelength The overall absorption coefficient at that location, wavelength The water absorption coefficient of pure water at that location, and respectively wavelength and wavelength Remote sensing reflectance at the location, For pure water Position backscattering coefficient, For intermediate process quantities, the calculation formula is as follows: Correspondingly, , You can directly substitute it.

[0042] In a further embodiment, the other two wavelength positions, such as , Backscattering coefficient 、 The calculation formula is as follows (in words). For example: ; On this basis The formula for calculating the total absorption coefficient at the location is as follows: In the formula For intermediate process quantities, the calculation formula is as follows: , For pure water Position backscattering coefficient.

[0043] Based on the above description, the water-based remote sensing reflectance of this embodiment The acquisition process is as follows: Based on the acquired blue light band Green light band Red light band and near-infrared band Remove pixels with cloud cover, shadows, strong aerosols, and snow cover; The following formula is used to convert the bands to their corresponding water-free remote sensing reflectance. : In the formula ， For satellite corresponding band Apparent reflectance. Therefore, after determining the input data, this embodiment also includes preprocessing of the input data and measured turbidity data, specifically spatiotemporal matching. The spatiotemporal matching principles include: spatial matching principle, temporal matching principle, multi-measured value matching processing principle, and matching data validity verification principle.

[0044] Furthermore, the spatial matching principle in this embodiment uses the coordinates of the monitoring points of the measured turbidity data as a reference. It spatially associates the pixels of the input MODIS remote sensing image (image data used to construct the turbidity inversion model) with the monitoring points, sets the spatial matching radius threshold to 250–500m, and retains only the image pixels within this radius around the monitoring point as valid pixels. This ensures that the water body area represented by the image pixels has spatial consistency with the water environment of the measured point, and eliminates the interference of spatial heterogeneity on the reliability of the matching data.

[0045] The time matching principle is understood as follows: based on the sampling time of the measured turbidity data, MODIS remote sensing images with a time difference of ≤3 days from the sampling time are selected. If a single set of measured data corresponds to multiple images that meet the time difference requirement, the images with a lower proportion of cloud, cloud shadow, and snow pixels (effective pixel proportion of coastal zone ≥50%) are selected for matching. If multiple sets of measured data correspond to the same image, it is necessary to ensure that the time difference between the sampling time of each set of measured data and the image acquisition time meets the requirement of ≤3 days.

[0046] The principle of multi-measured value matching means that when a single valid image pixel (meeting the spatial and temporal matching principle) corresponds to the measured turbidity data of multiple monitoring points, or when there are multiple samples of measured turbidity data of the same monitoring point within a time difference of ≤3 days, the median calculation method is used to determine the final matching value for such multiple sets of measured values. This avoids the impact of single sample outliers on the quality of subsequent turbidity inversion model training data and ensures the stability and representativeness of the matching data.

[0047] The principle of validating the matching data is to verify the validity of the matching data after completing the spatiotemporal matching, remove outliers in the measured turbidity data that exceed the reasonable turbidity range of the water environment (such as significantly deviating from the conventional turbidity range of similar water bodies), and ensure that the matched image pixels have passed quality control (removing cloud, cloud shadow, strong aerosol, and snow pixels). Finally, a spatiotemporally consistent and qualified input data matching dataset with the measured turbidity data is formed, which is used for the training and validation of the lake water turbidity remote sensing inversion model.

[0048] Based on this, the constructed matching dataset is divided into a training set and a validation set, with an optimal ratio of 7:3. A random forest algorithm is used to train the turbidity inversion model on the three types of input data and the matched turbidity data. During training, the model hyperparameters (including the number of decision trees, maximum depth, minimum number of samples per leaf node, etc.) are optimized using a grid search method. Model performance is evaluated using various statistical metrics, including the coefficient of determination (R²), root mean square error (RMSE), mean absolute error (MAE), and bias. (See reference for more information.) Figure 3 When the model accuracy reaches the preset standard (e.g., validation set R² ≥ 0.7, RMSE ≤ 15%), the model parameters are frozen. The model parameter tuning and optimal values ​​are shown in Table 1.

[0049] Table 1 Model Hyperparameter Tuning Using the frozen lake water turbidity remote sensing inversion model, the turbidity of the lake shoreline zone at the corresponding time period in the previous state image was calculated. Turbidity of the corresponding time period in the post-image The mean turbidity of the initial state was calculated based on the median method. : , This is a vector layer representing the coastal zone. The mean turbidity of the later state was calculated based on the median method. : , This is a vector layer representing the coastal zone.

[0050] Finally, the impact of turbidity on the lake shoreline. The calculation formula is as follows: ; The method for judging the trend of change is as follows: If The trend indicates that the turbidity of the coastal zone increased after extreme rainfall, meaning that the extreme rainfall increased the turbidity of the coastal zone through leaching; if The trend indicates that the turbidity of the coastal zone decreased after extreme rainfall, suggesting that the extreme rainfall had a dilution effect and reduced the turbidity of the coastal zone.

[0051] Example 2 This embodiment discloses a calculation system for the impact of extreme rainfall on the turbidity of lake shorelines, used to implement the calculation method described in Embodiment 1, including: The first module is set to determine the lake shore zone based on the proportional inner buffer method of the lake's equivalent radius, and select the target study area using the determined lake shore zone as a spatial constraint; obtain continuous historical precipitation data of the target area, and determine the extreme rainfall threshold based on the continuous historical precipitation data; The second module is set to obtain the daily rainfall in the target area. If the daily rainfall is greater than or equal to the extreme rainfall threshold, it is determined to be an extreme rainfall event. Based on the extreme rainfall event, the complete rainfall event is determined based on the entire time period of continuous historical precipitation data. The third module is configured to construct a remote sensing inversion model of lake water turbidity, and use this model to calculate the mean turbidity of the pre-state of a complete rainfall event. Mean turbidity of the post-state The mean turbidity of the previous state Mean turbidity of the post-state The influence of turbidity in the lake shore zone was calculated. and trends of change.

Claims

1. A method for calculating the impact of extreme rainfall on turbidity in lake shore zones, characterized in that, Includes the following steps: The lake shoreline zone is determined based on the proportional internal buffer method of the lake's equivalent radius. The determined lake shoreline zone is used as a spatial constraint to select the target study area. Continuous historical precipitation data of the target area is obtained, and extreme rainfall thresholds are determined based on the continuous historical precipitation data. Obtain the daily rainfall in the target area; if the daily rainfall is greater than or equal to the extreme rainfall threshold, it is determined to be an extreme rainfall event. Taking the aforementioned extreme rainfall events as the object, complete rainfall events are determined based on the overall time period of continuous historical precipitation data; A remote sensing inversion model for lake water turbidity was constructed, and the mean turbidity value before a complete rainfall event was calculated using the model. Mean turbidity of the post-state The mean turbidity in the previous state Mean turbidity of the post-state The influence of turbidity in the lake shore zone was calculated. and trends; The process for determining the lake shoreline is as follows: Calculate the lake area based on the lake's vector boundary. Equivalent circle radius and inward distance d Among them, the equivalent circle radius and inward distance d The calculation formula is as follows: ， ； Applying a negative shape buffer operation to the lake polygon P yields a shrunken polygon. The coastal zone vector layer of the original polygon and the shrunken polygon is calculated by using the vector difference between the original polygon and the shrunken polygon. : This yields a ring-shaped region with a width of d extending inward from the shoreline, which is the lake shoreline zone. This is a vector interpolation operation; The process for determining the complete rainfall event is as follows: Simultaneous recording of event windows based on extreme rainfall events ,in, The day that marks the beginning of the extreme rainfall event. The end date of the extreme rainfall event; will cover the event window. The image is marked as ; Starting date Using this as a boundary, trace back along the time frame and select the nearest qualified remote sensing image observation window as the previous state image, and mark it as... ; End date Using this as a boundary, we trace back along the time frame and select the nearest qualified remote sensing image observation window as the post-state image, marking it as... ; Will simultaneously contain the marked 、 and The corresponding event is defined as a complete rainfall event.

2. The method for calculating the impact of extreme rainfall on turbidity in lake shore zones according to claim 1, characterized in that, The construction process of the lake water turbidity remote sensing inversion model is as follows: Obtain the remote sensing reflectance of the spectrometer and acquire the blue light band. Green light band Red light band and near-infrared band Obtain the corresponding water-based remote sensing reflectance. , ; The water-free remote sensing reflectance The quality is processed to obtain the first type of input features; Based on the water-free remote sensing reflectivity Multiple binary ratio features are constructed by combining different bands, and these multiple binary ratio features are used as the second type of input features. Based on the blue light band Green light band Red light band Select the corresponding representative wavelengths respectively 、 and ; Representative wavelengths were calculated using the appropriate formulas. 、 and Corresponding total absorption coefficient , , and backscattering coefficient 、 、 , and serve as the third type of input feature; The machine learning method based on the random forest algorithm uses the first type of input features, the second type of input features, and the third type of input features as input data. The input data and the measured turbidity data are preprocessed to construct training and validation sets, and a remote sensing inversion model of lake water turbidity is trained and established.

3. The method for calculating the impact of extreme rainfall on turbidity in lake shore zones according to claim 1, characterized in that, The mean turbidity of the previous state The calculation process is as follows: The turbidity of the lake shoreline at the corresponding time period in the pre-state image was obtained using a remote sensing inversion model of lake water turbidity. ; The mean turbidity of the initial state was calculated based on the median method. : , This is a vector layer representing the coastal zone.

4. The method for calculating the impact of extreme rainfall on turbidity in lake shore zones according to claim 1, characterized in that, The mean turbidity of the post-state The calculation process is as follows: The turbidity of the lake shoreline at the corresponding time period in the post-state image was obtained using a remote sensing inversion model of lake water turbidity. ; The mean turbidity of the later state was calculated based on the median method. : , This is a vector layer representing the coastal zone.

5. The method for calculating the impact of extreme rainfall on turbidity in lake shore zones according to claim 1, characterized in that, The influence of turbidity on the lake shoreline The calculation formula is as follows: ； The method for judging the trend of change is as follows: If The trend is determined to be an increase in turbidity in the coastal zone after extreme rainfall. The trend is determined to be a decrease in turbidity in the coastal zone after extreme rainfall.

6. The method for calculating the impact of extreme rainfall on turbidity in lake shore zones according to claim 2, characterized in that, The water-reflectance The acquisition process is as follows: Based on the acquired blue light band Green light band Red light band and near-infrared band Remove pixels with cloud cover, shadows, strong aerosols, and snow cover; The following formula is used to convert the bands to their corresponding water-free remote sensing reflectance. : In the formula, For the corresponding band of the satellite Apparent reflectance.

7. The method for calculating the impact of extreme rainfall on turbidity in lake shore zones according to claim 1, characterized in that, The preprocessing of input data and measured turbidity data is spatiotemporal matching. The spatiotemporal matching principles include: spatial matching principle, temporal matching principle, multi-measured value matching processing principle, and matching data validity verification principle.

8. A system for calculating the impact of extreme rainfall on turbidity in lake shore zones, used to implement the calculation method as described in any one of claims 1 to 7, characterized in that, include: The first module is set to determine the lake shore zone based on the proportional inner buffer method of the lake's equivalent radius, and select the target study area using the determined lake shore zone as a spatial constraint; obtain continuous historical precipitation data of the target area, and determine the extreme rainfall threshold based on the continuous historical precipitation data; The second module is set to obtain the daily rainfall in the target area. If the daily rainfall is greater than or equal to the extreme rainfall threshold, it is determined to be an extreme rainfall event. Taking the aforementioned extreme rainfall events as the object, complete rainfall events are determined based on the overall time period of continuous historical precipitation data; The third module is configured to construct a remote sensing inversion model of lake water turbidity, and use this model to calculate the mean turbidity of the pre-state of a complete rainfall event. Mean turbidity of the post-state The mean turbidity in the previous state Mean turbidity of the post-state The influence of turbidity in the lake shore zone was calculated. and trends of change.

Images