Water environment cyanobacterial bloom early warning method and system

Abstract

The invention discloses a water environment cyanobacterial bloom early warning method and system, and relates to the technical field of environment monitoring, and the method comprises the steps: collecting multi-source monitoring data through an air-sky-ground integrated network composed of a liftable layered sensor, a satellite, an unmanned aerial vehicle and a meteorological station, and carrying out the fusion processing of the data, and generating a cyanobacteria space-time distribution thermodynamic diagram; further, introducing a cyanobacteria vertical migration factor and a zooplankton feeding factor, constructing an adaptive prediction model, and accurately calculating the cyanobacteria cell density and the water bloom coverage area; meanwhile, a regional feature library is formed based on climate zone and water body type clustering, and an optimal early warning threshold value is dynamically matched through ROC curve analysis; finally, an early warning coefficient is calculated according to a prediction result and a self-adaptive threshold value, a graded early warning signal is output, and the accuracy, reliability and practicability of early warning of cyanobacterial bloom are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental monitoring, in particular to a water environment blue-green algae bloom early warning method and system. BACKGROUND

[0002] Blue-green algae bloom is a visible algae aggregate formed by overgrowth of blue-green algae in freshwater bodies. Its outbreak can destroy the ecological balance of water bodies, produce algal toxins, cause the death of aquatic organisms, and threaten human drinking water safety and health, such as causing liver damage and skin allergy.

[0003] The prior art is usually artificial monitoring, in-situ online monitoring, satellite remote sensing monitoring, etc.

[0004] Among them, artificial monitoring collects water samples on site by artificial means, analyzes blue-green algae cell density, chlorophyll a and other core indicators in the laboratory, and combines on-site determination of water temperature, pH and other environmental factors as the core, relying on artificial microscope observation to complete algae classification and data statistics.

[0005] In-situ online monitoring collects real-time data such as chlorophyll a, blue-green algae fluorescence, and nutrient salts in water bodies by deploying buoys or underwater sensors, and automatically triggers an early warning when the data reaches a preset threshold, realizing automation and real-time of the monitoring process.

[0006] The prior art has the following technical defects, which are specifically embodied in:

[0007] 1. Single monitoring means, lack of three-dimensional data: traditional methods mainly rely on fixed points or single platforms (such as only ground sensors or only satellite remote sensing), which cannot synchronously obtain detailed data and multi-dimensional spatial information of the vertical section of the water body, resulting in data blind spots and difficulty in fully reflecting the real distribution and migration state of blue-green algae.

[0008] 2. Simple prediction model, ignoring key ecological processes: existing models are mostly based on surface data or average levels for prediction, without quantifying the vertical migration behavior of blue-green algae and key biological and physical regulation factors such as zooplankton feeding, resulting in insufficient mechanism of growth prediction model and large deviation of prediction results from actual situation.

[0009] 3. Rigid early warning threshold, poor universality: usually uses fixed and unified threshold standards, without considering the huge differences between different climate zones and water body types, resulting in false negatives in high-risk water bodies and false positives in low-risk water bodies, low accuracy and practicality. SUMMARY

[0010] In view of the above technical deficiencies, the purpose of the present application is to provide a water environment blue-green algae bloom early warning method and system.

[0011] To solve the above technical problems, the present application adopts the following technical solutions: The present application provides a water environment blue-green algae bloom early warning method, comprising: S1, multi-source data acquisition: through the "space-ground" integrated monitoring network composed of a liftable layered sensor array, a satellite, a UAV and a weather station, full profile hydrochemical data, remote sensing image data and weather data of the target water body are collected at a preset period.

[0012] S2, intelligent fusion of multi-source data: based on the full profile hydrochemical data, remote sensing image data and weather data of the target water body, data cleaning, noise reduction and space-time registration processing are performed to obtain a blue algae space-time distribution thermal map.

[0013] S3, adaptive prediction: based on the blue algae space-time distribution thermal map and historical data, by introducing a blue algae vertical migration factor and a plankton feeding factor, a weighted fusion is performed to calculate the prediction results of the blue algae cell density and the bloom coverage area.

[0014] S4, regional dynamic threshold judgment: based on the climate zone and the water body type clustering to form a regional characteristic library, the true positive rate and the false positive rate are calculated through ROC curve analysis to select the optimal early warning threshold, and the target water body threshold is automatically matched.

[0015] S5, early warning output: based on the prediction results and the target water body threshold, the early warning coefficient of the target water body is calculated to further judge the early warning level of the target water body.

[0016] Preferably, the blue algae vertical migration factor is specifically analyzed as follows: the measured water temperature of each depth of the target water body is extracted from the full profile hydrochemical data of the target water body, the reference water temperature of each depth of the target water body is obtained from the database, the measured water temperature of each depth of the target water body is subtracted from the reference water temperature of each depth of the target water body to obtain the water temperature difference value of each depth of the target water body, the depth division threshold of the target water body is obtained from the database, and the water temperature difference value of each depth of the target water body is divided by the depth division threshold of the target water body to obtain the water temperature gradient coefficient of each depth of the target water body.

[0017] The dissolved oxygen concentration of each depth of the target water body is extracted from the full profile hydrochemical data of the target water body, the reference dissolved oxygen concentration of each depth of the target water body is obtained from the database, the dissolved oxygen concentration of each depth of the target water body is subtracted from the reference dissolved oxygen concentration of each depth of the target water body to obtain the dissolved oxygen concentration difference value of each depth of the target water body, and the dissolved oxygen concentration difference value of each depth of the target water body is divided by the depth division threshold of the target water body to obtain the dissolved oxygen concentration gradient coefficient of each depth of the target water body.

[0018] Based on the water temperature gradient coefficient of each depth of the target water body and the dissolved oxygen concentration gradient coefficient of each depth of the target water body, a weighted calculation is performed to obtain the blue algae vertical migration factor of each depth of the target water body.

[0019] Preferably, the zooplankton feeding factor is calculated by extracting the light intensity and the concentration of key nutrients at each depth of the target water body from the full profile hydrochemical data of the target water body, and inputting the light intensity and the concentration of key nutrients at each depth of the target water body into a blue algae self-growth rate calculation formula to obtain the blue algae self-growth rate at each depth of the target water body.

[0020] The time variation rate of the chlorophyll a concentration at each depth of the target water body is extracted from the full profile hydrochemical data of the target water body, and the zooplankton feeding factor at each depth of the target water body is calculated based on the blue algae self-growth rate at each depth of the target water body and the blue algae vertical migration factor at each depth of the target water body.

[0021] Preferably, the prediction result of the blue algae cell density is obtained by obtaining the basic blue algae growth rate at each depth of the target water body from the database, calculating the actual blue algae growth rate at each depth of the target water body based on the blue algae vertical migration factor at each depth of the target water body and the zooplankton feeding factor at each depth of the target water body, extracting the initial cell density at each depth of the target water body from the blue algae spatiotemporal distribution thermal map, and inputting the actual blue algae growth rate at each depth of the target water body and the initial cell density at each depth of the target water body into a blue algae cell density prediction model to obtain the prediction value of the blue algae cell density at each depth of the target water body.

[0022] Preferably, the prediction result of the water bloom coverage area is obtained by extracting the water bloom coverage at each depth of the target water body from the full profile hydrochemical data of the target water body, extracting the total area of the target water body from the remote sensing image data, and inputting the water bloom coverage at each depth of the target water body and the total area of the target water body into a water bloom coverage area prediction model to obtain the prediction value of the water bloom coverage area of the target water body.

[0023] Preferably, the optimal early warning threshold is obtained by extracting the sample labeled by historical data from the regional characteristic library formed based on the climate zone and the water body type, counting the number of samples corresponding to the true positive, the false positive, the true negative, and the false negative corresponding to each candidate threshold, dividing the number of samples corresponding to the true positive by the sum of the number of samples corresponding to the true positive and the false negative corresponding to each candidate threshold to obtain the true positive rate corresponding to each candidate threshold, dividing the number of samples corresponding to the false positive by the sum of the number of samples corresponding to the false positive and the true negative corresponding to each candidate threshold to obtain the false positive rate corresponding to each candidate threshold, drawing a ROC curve based on the true positive rate corresponding to each candidate threshold and the false positive rate corresponding to each candidate threshold, and obtaining the optimal early warning threshold.

[0024] Preferably, the automatic matching target water body threshold value, the specific analysis method is: obtaining the climate zone characteristics and water body type characteristics of the target water body from the database to form a target water body characteristic vector, obtaining the characteristic vectors of each sub-class water body from the database, calculating the Euclidean distance between the target water body characteristic vector and the characteristic vectors of each sub-class water body through the Euclidean distance calculation formula, screening the sub-class water body corresponding to the minimum Euclidean distance to extract the optimal early warning threshold value of the sub-class water body as the optimal early warning threshold value of the target water body.

[0025] Preferably, the early warning coefficient of the target water body, the specific analysis method is: extracting the cyanobacterial cell density prediction value of each depth of the target water body, summing to obtain the cyanobacterial cell density prediction value of the target water body, extracting the water bloom coverage area prediction value of the target water body, extracting the risk factor corresponding to the optimal early warning threshold value of the target water body from the database, and inputting the cyanobacterial cell density prediction value, the water bloom coverage area prediction value and the risk factor of the target water body into the target water body early warning coefficient calculation formula to obtain the early warning coefficient of the target water body.

[0026] Preferably, the early warning level of the target water body is judged, and the specific analysis method is: extracting the early warning coefficient of the target water body, inputting it into the early warning level judgment model, and outputting the judgment result of the early warning level of the target water body.

[0027] The judgment result of the early warning level of the target water body contains the numerical values of 0, 1 and 2, if the judgment result of the early warning level of the target water body is 0, it is determined that the early warning level of the target water body is low, if the judgment result of the early warning level of the target water body is 1, it is determined that the early warning level of the target water body is medium, and if the judgment result of the early warning level of the target water body is 2, it is determined that the early warning level of the target water body is high.

[0028] The second aspect of the present application provides a water environment cyanobacterial bloom early warning method system, characterized in that it comprises: a multi-source data acquisition module: for collecting full-profile hydrological and biochemical data, remote sensing image data and meteorological data of a target water body in a preset period through a "space-air-ground" integrated monitoring network composed of a liftable layered sensor array, a satellite, a unmanned aerial vehicle and a weather station.

[0029] A multi-source data intelligent fusion module: for data cleaning, noise reduction and space-time registration processing based on the full-profile hydrological and biochemical data, remote sensing image data and meteorological data of the target water body to obtain a cyanobacterial space-time distribution heat map.

[0030] An adaptive prediction module: for calculating the prediction results of cyanobacterial cell density and water bloom coverage area by introducing a cyanobacterial vertical migration factor and a plankton feeding factor and weighted fusion based on the cyanobacterial space-time distribution heat map and historical data.

[0031] Regional dynamic threshold judgment module: used for forming a regional characteristic library based on climate zones and water body type clustering, calculating true positive rate and false positive rate through ROC curve analysis, selecting the optimal early warning threshold, and automatically matching the target water body threshold.

[0032] Early warning output module: used for calculating the early warning coefficient of the target water body based on the prediction result and the target water body threshold, and then judging the early warning level of the target water body.

[0033] The beneficial effects of the present application are:

[0034] 1. The present application realizes full-range data acquisition of water bodies from the surface to the bottom and from points to planes by constructing an air-space-ground integrated three-dimensional monitoring network, completely eliminating monitoring blind spots and providing a comprehensive and accurate data basis for early warning.

[0035] 2. The present application innovatively introduces and quantifies the blue-green algae vertical migration factor and the plankton feeding factor, and integrates them into the prediction model, making the model more in line with the actual ecological dynamics process, and significantly improving the accuracy and mechanism of blue-green algae cell density and water bloom area prediction.

[0036] 3. The present application dynamically matches the optimal early warning threshold for different climate zones and water body types by establishing a regional characteristic library and using ROC curve analysis, realizes the application of threshold according to local conditions, greatly reduces false positives and false negatives, and improves the accuracy and practicality of early warning. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0038] Figure 1 The flowchart of the method implementation steps of the present application is shown in the figure;

[0039] Figure 2 The connection diagram of the system structure of the present application is shown in the figure. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0041] According toFigure 1 As shown, the present application provides a water environment blue-green algae bloom early warning method, comprising:

[0042] S1, multi-source data acquisition: through the "space-air-ground" integrated monitoring network composed of liftable layered sensor array, satellite, unmanned aerial vehicle and weather station, the full profile hydrochemical data, remote sensing image data and meteorological data of the target water body are collected according to the preset period.

[0043] It should be noted that the full profile hydrochemical data of the target water body includes but is not limited to water temperature, flow rate and flow direction, transparency, dissolved oxygen concentration, conductivity, chlorophyll a concentration, nutrient salt concentration and blue algae cell density, the remote sensing image data includes but is not limited to water body basic information, such as water body boundary and range, water bloom intuitive distribution image, blue algae and water body optical parameter, such as blue algae specific index: such as blue algae index, remote sensing inversion chlorophyll a concentration suspended matter concentration and colored soluble organic matter, the meteorological data includes but is not limited to photosynthetic active radiation, sunshine duration, air temperature, wind speed and direction, precipitation and evaporation.

[0044] S2, intelligent fusion of multi-source data: based on the full profile hydrochemical data, remote sensing image data and meteorological data of the target water body, data cleaning, noise reduction and space-time registration processing are carried out to obtain a blue algae space-time distribution thermal map.

[0045] S3, adaptive prediction: based on the blue algae space-time distribution thermal map and historical data, by introducing blue algae vertical migration factor and plankton feeding factor, weighted fusion is carried out to calculate the prediction results of blue algae cell density and water bloom coverage area.

[0046] In one specific embodiment, the blue algae vertical migration factor is specifically analyzed as follows: the measured water temperature of each depth of the target water body is extracted from the full profile hydrochemical data of the target water body, the reference water temperature of each depth of the target water body is obtained from the database, the measured water temperature of each depth of the target water body is subtracted from the reference water temperature of each depth of the target water body to obtain the water temperature difference value of each depth of the target water body, the depth division threshold value of the target water body is obtained from the database, and the water temperature difference value of each depth of the target water body is divided by the depth division threshold value of the target water body to obtain the water temperature gradient coefficient of each depth of the target water body.

[0047] It should be noted that the reference water temperature of each depth of the target water body is set by a professional, for example, according to the climate zone and water body type (such as subtropical lake, temperate reservoir) to which the target water body belongs, combined with the historical same period average stratified water temperature of the same type of water body in the regional characteristic library (for example, the historical same period average value is taken for the surface layer 0.5m, the average value is taken for the middle layer according to the water depth 1 / 2, and the average value is taken for the bottom layer 0.5m); then, according to the real-time stratified water temperature monitoring data of the target water body (such as the surface layer, middle layer and bottom layer data collected by the liftable sensor), combined with the key water layer of blue-green algae vertical migration (such as the surface layer with active photosynthesis, the middle layer of thermocline), the basic reference value is dynamically calibrated; the depth division threshold of the target water body is set by a professional, for example, according to the water body type (such as shallow lake, deep reservoir, river) and the characteristics of the climate zone, and referring to the typical depth structure of the same type of water body in the regional characteristic library (for example, the surface layer / bottom layer is divided according to 3m for shallow lake, and the surface layer / thermocline layer / bottom layer is divided according to the position of the thermocline layer for deep reservoir), the basic framework is determined; then, combined with the key functional layer of blue-green algae vertical migration (the surface layer with active photosynthesis, the thermocline layer inhibiting growth, and the bottom layer releasing nutrients), and according to the season (such as summer) and the real-time monitoring of the water body thermal stratification data, the division is dynamically fine-tuned to ensure that the division can accurately match the depth correlation law of blue-green algae growth and migration.

[0048] The dissolved oxygen concentration of each depth of the target water body is extracted from the full profile hydrological and biochemical data of the target water body, the reference dissolved oxygen concentration of each depth of the target water body is obtained from the database, the dissolved oxygen concentration of each depth of the target water body is subtracted from the reference dissolved oxygen concentration of each depth of the target water body, and the dissolved oxygen concentration difference of each depth of the target water body is obtained. The dissolved oxygen concentration difference of each depth of the target water body is divided by the depth division threshold of the target water body to obtain the dissolved oxygen concentration gradient coefficient of each depth of the target water body.

[0049] It should be noted that the reference dissolved oxygen concentration of each depth of the target water body is set by a professional, and the setting method refers to the reference water temperature of each depth of the target water body, which will not be described in detail.

[0050] Based on the water temperature gradient coefficient of each depth of the target water body and the dissolved oxygen concentration gradient coefficient of each depth of the target water body, weighted calculation is performed to obtain the blue-green algae vertical migration factor of each depth of the target water body.

[0051] In one specific embodiment, the specific analysis method of the zooplankton feeding factor is: the light intensity and key nutrient salt concentration of each depth of the target water body are extracted from the full profile hydrological and biochemical data of the target water body, and are brought into the blue-green algae self growth rate calculation formula to calculate the blue-green algae self growth rate of each depth of the target water body.

[0052] It should be noted that the blue-green algae self growth rate calculation formula is:

[0053]

[0054] Where G represents the cyanobacteria's own growth rate, μ max Indicates the maximum specific growth rate of cyanobacteria, I represents light intensity, and K represents the maximum specific growth rate of cyanobacteria. I Indicates half-saturated light intensity, N represents the concentration of key nutrients, and K represents the concentration of light in the light source. N Let f(T) represent the concentration of semi-saturated key nutrients, and f(T) represent the effect of temperature.

[0055] The time variation rate of chlorophyll a concentration at each depth of the target water body was extracted from the full profile hydrological and biochemical data of the target water body. Combined with the growth rate of cyanobacteria at each depth of the target water body, the zooplankton feeding factor at each depth of the target water body was calculated based on the formula for calculating zooplankton feeding factor.

[0056] It should be noted that the formula for calculating the zooplankton feeding factor is as follows:

[0057]

[0058] Where F represents the zooplankton feeding factor, α represents the time change rate of chlorophyll a concentration in the target water body, and γ represents the conversion factor between cyanobacterial biomass and chlorophyll a.

[0059] In one specific embodiment, the prediction result of the cyanobacterial cell density is analyzed by the following method: obtaining the basic cyanobacterial growth rate at each depth of the target water body from the database; calculating the actual cyanobacterial growth rate at each depth of the target water body based on the vertical migration factor of cyanobacteria and the zooplankton feeding factor at each depth of the target water body; extracting the initial cell density at each depth of the target water body from the cyanobacterial spatiotemporal distribution heat map; and inputting the actual cyanobacterial growth rate and the initial cell density at each depth of the target water body into the cyanobacterial cell density prediction model to obtain the predicted value of the cyanobacterial cell density at each depth of the target water body.

[0060] It should be noted that the baseline cyanobacteria growth rate at each depth of the target water body is set by professionals. For example, the initial baseline value is determined based on key environmental factors such as measured water temperature, light intensity, and nitrogen and phosphorus nutrient concentrations (e.g., total nitrogen and total phosphorus) at each depth, combined with historical monitoring data of cyanobacteria growth at the same depth during the same period. Then, the influence coefficient of each environmental factor on growth is quantified by referring to classic cyanobacteria growth models such as the Monod model and the Logistic model. At the same time, it is dynamically adjusted according to the water body type (e.g., lakes and reservoirs) and the season to ensure that it conforms to the actual growth pattern.

[0061] The calculation yields the actual cyanobacteria growth rate at each depth in the target water body, and the calculation formula is as follows:

[0062] G actual,z =G base,z·(1+V z )-F z ;

[0063] Among them, G actual,z G represents the actual cyanobacteria growth rate at depth z in the target water body. base,z V represents the basic cyanobacterial growth rate at a target water depth z. z F represents the vertical migration factor of cyanobacteria at depth z in the target water body. z This represents the zooplankton feeding factor at a target water depth z.

[0064] The cyanobacterial cell density prediction model is expressed as follows:

[0065]

[0066] Among them, C pred,z (t+Δt) represents the predicted cyanobacterial cell density at depth z in the target water body, C init,z (t) represents the initial cell density at the target water depth z, t represents the initial monitoring time, and Δt represents the prediction time interval.

[0067] In one specific embodiment, the prediction result of the algal bloom coverage area is analyzed by the following method: extracting the algal bloom coverage at each depth of the target water body from the full profile hydrological and biochemical data of the target water body, extracting the total area of ​​the target water body from the remote sensing image data, and inputting the algal bloom coverage at each depth of the target water body and the total area of ​​the target water body into the algal bloom coverage area prediction model to obtain the predicted value of the algal bloom coverage area of ​​the target water body.

[0068] It should be noted that the prediction model for the algal bloom coverage area is expressed as follows:

[0069]

[0070] Among them, A pred A represents the predicted area of ​​algal bloom coverage in the target water body. total H represents the total area of ​​the target water body, and C represents the total depth of the target water body. z h represents the algal bloom coverage at depth z of the target water body. z This represents the depth threshold for the target water body, and n represents the total number of depths to be divided.

[0071] S4. Regionalized Dynamic Threshold Judgment: Based on the clustering of climate zones and water body types to form a regional feature database, the true positive rate and false positive rate are calculated through ROC curve analysis, the optimal warning threshold is selected, and the target water body threshold is automatically matched.

[0072] In one specific embodiment, the method for selecting the optimal early warning threshold is as follows: Historically labeled samples are extracted from a regional feature database formed by clustering based on climate zones and water body types. The number of samples corresponding to true positives, false positives, true negatives, and false negatives for each candidate threshold is statistically calculated. The number of true positive samples corresponding to each candidate threshold is divided by the sum of the number of true positive and false negative samples to obtain the true positive rate for each candidate threshold. The number of false positive samples corresponding to each candidate threshold is divided by the sum of the number of false positive and true negative samples to obtain the false positive rate for each candidate threshold. Based on the true positive rate and the false positive rate for each candidate threshold, an ROC curve is plotted to obtain the optimal early warning threshold.

[0073] It should be noted that obtaining the optimal warning threshold specifically involves selecting the threshold closest to the upper left corner (0,1) of the curve as the optimal warning threshold. The core logic for a true positive is: it is actually a cyanobacteria warning event, and the judgment result based on the threshold is "triggered warning" (correct judgment, no omissions). The core logic for a false positive is: it is not actually a cyanobacteria warning event, but the judgment result based on the threshold is "triggered warning" (incorrect judgment, a "false alarm"). The core logic for a true negative is: it is not actually a cyanobacteria warning event, and the judgment result based on the threshold is "not triggered warning" (correct judgment, no interference). The core logic for a false negative is: it is actually a cyanobacteria warning event, but the judgment result based on the threshold is "not triggered warning" (incorrect judgment, a "missed report").

[0074] The automatic matching of optimal warning thresholds is essentially a process of "classification followed by association": first, the target water body is classified into a pre-defined "climate zone + water body type" subclass in the regional feature database based on its regional characteristics (climate zone, water type); then, the optimal warning threshold determined by ROC analysis for this subclass is directly associated with the threshold of the target water body. The core principle is to leverage the characteristic that "similar water bodies have similar cyanobacterial growth patterns" to address the problem of insufficient historical data for individual target water bodies, while ensuring the regional adaptability of the thresholds.

[0075] In one specific embodiment, the automatic matching of the target water body threshold is specifically analyzed as follows: obtaining the climate zone characteristics and water type characteristics of the target water body from the database to form a target water body feature vector; obtaining the feature vectors of each sub-type of water body from the database; calculating the Euclidean distance between the target water body feature vector and the feature vectors of each sub-type of water body using the Euclidean distance calculation formula; selecting the sub-type of water body corresponding to the minimum Euclidean distance; extracting the optimal warning threshold of the sub-type of water body; and using it as the optimal warning threshold of the target water body.

[0076] It should be noted that the Euclidean distance calculation formula is existing technology and will not be elaborated upon here.

[0077] S5. Early Warning Output: Based on the prediction results and the target water body threshold, the early warning coefficient of the target water body is calculated, and then the early warning level of the target water body is determined.

[0078] In one specific embodiment, the early warning coefficient of the target water body is specifically analyzed as follows: extract the predicted values ​​of cyanobacterial cell density at each depth of the target water body, sum them to obtain the predicted values ​​of cyanobacterial cell density of the target water body, extract the predicted values ​​of algal bloom coverage area of ​​the target water body, extract the risk factors corresponding to the optimal early warning threshold of the target water body from the database, and substitute the predicted values ​​of cyanobacterial cell density, algal bloom coverage area, and risk factors of the target water body into the calculation formula of the early warning coefficient of the target water body to obtain the early warning coefficient of the target water body.

[0079] It should be noted that the formula for calculating the target water body early warning coefficient is as follows:

[0080]

[0081] W represents the target water body warning coefficient, ε represents the cyanobacterial cell density weighting coefficient, and C total,pred T represents the total predicted value of cyanobacterial cell density in the target water body. den The threshold for the optimal early warning of cyanobacterial cell density in the target water body is represented by β, which represents the weighting coefficient for algal bloom coverage area. cover λ represents the optimal warning threshold for the algal bloom coverage area of ​​the target water body, and λ represents the risk factor corresponding to the optimal warning threshold of the target water body.

[0082] The three optimal warning thresholds for the target water body—optimal warning threshold for cyanobacterial cell density, and optimal warning threshold for algal bloom coverage—are essentially subordinate to each other as "overall threshold - sub-thresholds." The optimal warning threshold for cyanobacterial cell density is the optimal judgment threshold selected through ROC curve screening for the single indicator of "total predicted value of cyanobacterial cell density in the target water body." The optimal warning threshold for algal bloom coverage is the optimal judgment threshold selected through ROC curve screening for the single indicator of "predicted value of algal bloom coverage in the target water body." The optimal warning threshold for the target water body is a "comprehensive risk threshold" ultimately determined through ROC curve screening, based on the optimal warning thresholds for cyanobacterial cell density and algal bloom coverage, combined with the regional characteristics of the target water body (climate zone, water body type, from a regional feature database) and historical risk cases (true positive / false positive samples). This threshold measures the overall risk level under the synergistic effect of "cell density + coverage area."

[0083] In one specific embodiment, the method for determining the warning level of the target water body is as follows: extract the warning coefficient of the target water body, input it into the warning level judgment model, and output the judgment result of the warning level of the target water body.

[0084] It should be noted that the model expression in the aforementioned warning level judgment model is:

[0085]

[0086] Where L represents the warning level judgment result, W represents the warning coefficient of the target water body, W1 represents the critical value of the low-to-medium warning coefficient, and W2 represents the critical value of the medium-to-high warning coefficient.

[0087] The judgment result of the warning level of the target water body includes the values ​​of 0, 1 and 2. If the judgment result of the warning level of the target water body is 0, the warning level of the target water body is determined to be a low warning. If the judgment result of the warning level of the target water body is 1, the warning level of the target water body is determined to be a medium warning. If the judgment result of the warning level of the target water body is 2, the warning level of the target water body is determined to be a high warning.

[0088] according to Figure 2 As shown, the present invention provides a system for early warning of cyanobacterial blooms in aquatic environments, comprising: a multi-source data acquisition module: used to collect full-profile hydrological and biochemical data, remote sensing image data and meteorological data of the target water body at preset cycles through an integrated "air-space-ground" monitoring network consisting of a liftable layered sensor array, satellites, drones and meteorological stations.

[0089] Multi-source data intelligent fusion module: used to perform data cleaning, noise reduction and spatiotemporal registration processing based on full-profile hydrological and biochemical data, remote sensing image data and meteorological data of the target water body to obtain a spatiotemporal distribution heat map of cyanobacteria.

[0090] Adaptive prediction module: Based on the spatiotemporal distribution heatmap of cyanobacteria and historical data, it calculates the predicted results of cyanobacterial cell density and bloom coverage area by introducing cyanobacterial vertical migration factor and zooplankton feeding factor through weighted fusion.

[0091] The regional dynamic threshold judgment module is used to form a regional feature library based on climate zone and water body type clustering, calculate the true positive rate and false positive rate through ROC curve analysis, select the optimal warning threshold, and automatically match the target water body threshold.

[0092] Early warning output module: Based on the prediction results and the target water body threshold, it calculates the early warning coefficient of the target water body and then determines the early warning level of the target water body.

[0093] It should be noted that it also includes a database.

[0094] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A water environment cyanobacterial bloom early warning method, characterized in that, The application relates to a cyanobacteria bloom early warning method and device. S1, multi-source data acquisition: through an "air-space-ground" integrated monitoring network composed of a liftable layered sensor array, a satellite, a drone and a weather station, full-profile hydrochemical data, remote sensing image data and weather data of a target water body are collected according to a preset period; S2, intelligent fusion of multi-source data: based on the full-profile hydrochemical data, remote sensing image data and weather data of the target water body, data cleaning, noise reduction and space-time registration processing are carried out to obtain a cyanobacteria space-time distribution thermal map; S3, adaptive prediction: based on the cyanobacteria space-time distribution thermal map and historical data, a cyanobacteria vertical migration factor and a zooplankton feeding factor are introduced, weighted fusion is carried out, and prediction results of cyanobacteria cell density and water bloom coverage area are obtained; S4, regional dynamic threshold judgment: based on climate zones and water body type clustering to form a regional characteristic database, the true positive rate and the false positive rate are calculated through ROC curve analysis, the optimal early warning threshold is selected, and the target water body threshold is automatically matched; S5, early warning output: based on the prediction results and the target water body threshold, the early warning coefficient of the target water body is calculated, and the early warning level of the target water body is judged.

2. The water environment cyanobacterial bloom early warning method according to claim 1, characterized in that, The specific analysis method of the cyanobacteria vertical migration factor is as follows: measured water temperatures of each depth of the target water body are extracted from the full-profile hydrochemical data of the target water body, reference water temperatures of each depth of the target water body are obtained from a database, the measured water temperatures of each depth of the target water body are subtracted from the reference water temperatures of each depth of the target water body, water temperature difference values of each depth of the target water body are obtained, a depth division threshold value of the target water body is obtained from the database, the water temperature difference values of each depth of the target water body are divided by the depth division threshold value of the target water body, and water temperature gradient coefficients of each depth of the target water body are obtained; dissolved oxygen concentrations of each depth of the target water body are extracted from the full-profile hydrochemical data of the target water body, reference dissolved oxygen concentrations of each depth of the target water body are obtained from the database, the dissolved oxygen concentrations of each depth of the target water body are subtracted from the reference dissolved oxygen concentrations of each depth of the target water body, dissolved oxygen concentration difference values of each depth of the target water body are obtained, and the dissolved oxygen concentration difference values of each depth of the target water body are divided by the depth division threshold value of the target water body, thereby obtaining dissolved oxygen concentration gradient coefficients of each depth of the target water body; the water temperature gradient coefficients of each depth of the target water body and the dissolved oxygen concentration gradient coefficients of each depth of the target water body are weighted and calculated, and cyanobacteria vertical migration factors of each depth of the target water body are obtained.

3. The water environment cyanobacterial bloom early warning method according to claim 2, characterized in that, The specific analysis method of the zooplankton feeding factor is as follows: light intensities and key nutrient salt concentrations of each depth of the target water body are extracted from the full-profile hydrochemical data of the target water body, and are introduced into a cyanobacteria self-growth rate calculation formula, so that the cyanobacteria self-growth rate of each depth of the target water body is calculated; a time change rate of chlorophyll a concentration of each depth of the target water body is extracted from the full-profile hydrochemical data of the target water body, the cyanobacteria self-growth rate of each depth of the target water body is combined, and a zooplankton feeding factor calculation formula is used to calculate the zooplankton feeding factor of each depth of the target water body.

4. The water environment cyanobacterial bloom early warning method according to claim 2, characterized in that, The specific analysis method of the prediction result of the cyanobacteria cell density is as follows: The actual blue-green algae growth rate at each depth of the target water body is calculated based on the vertical migration factor of blue-green algae at each depth of the target water body and the plankton feeding factor at each depth of the target water body, the initial cell density at each depth of the target water body is extracted from the blue-green algae spatiotemporal distribution thermal map, and the actual blue-green algae growth rate at each depth of the target water body and the initial cell density at each depth of the target water body are brought into the blue-green algae cell density prediction model to obtain the blue-green algae cell density prediction value at each depth of the target water body.

5. The water environment cyanobacterial bloom early warning method according to claim 4, characterized in that, The prediction result of the bloom coverage area is specifically analyzed as follows: The bloom coverage at each depth of the target water body is extracted from the full-profile hydrochemical data of the target water body, the total area of the target water body is extracted from the remote sensing image data, and the bloom coverage at each depth of the target water body and the total area of the target water body are brought into the bloom coverage area prediction model to obtain the bloom coverage area prediction value of the target water body.

6. The water environment cyanobacterial bloom early warning method according to claim 5, characterized in that, The optimal early warning threshold is selected, and the specific analysis method is as follows: The sample labeled with historical data is extracted from the regional characteristic library formed based on the climate zone and water body type clustering, and the number of samples corresponding to true positives, false positives, true negatives and false negatives corresponding to each candidate threshold is counted. The number of samples corresponding to true positives corresponding to each candidate threshold is divided by the sum of the number of samples corresponding to true positives and false negatives to obtain the true positive rate corresponding to each candidate threshold. The number of samples corresponding to false positives corresponding to each candidate threshold is divided by the sum of the number of samples corresponding to false positives and true negatives to obtain the false positive rate corresponding to each candidate threshold. The ROC curve is drawn based on the true positive rate corresponding to each candidate threshold and the false positive rate corresponding to each candidate threshold, and the optimal early warning threshold is obtained.

7. The water environment cyanobacterial bloom early warning method according to claim 6, characterized in that, The optimal early warning threshold is selected, and the specific analysis method is as follows: The climate zone characteristics and water body type characteristics of the target water body are obtained from the database to form a feature vector of the target water body. The feature vector of each sub-class water body is obtained from the database. The Euclidean distance between the feature vector of the target water body and the feature vector of each sub-class water body is calculated by the Euclidean distance calculation formula to obtain the sub-class water body corresponding to the minimum Euclidean distance. The optimal early warning threshold of the sub-class water body is extracted and used as the optimal early warning threshold of the target water body.

8. The water environment cyanobacterial bloom early warning method according to claim 7, characterized in that, The early warning coefficient of the target water body is specifically analyzed as follows: The blue-green algae cell density prediction value at each depth of the target water body is extracted, and the blue-green algae cell density prediction value of the target water body is calculated by summation. The bloom coverage area prediction value of the target water body is extracted, and the risk factor corresponding to the optimal early warning threshold of the target water body is extracted from the database. The blue-green algae cell density prediction value, the bloom coverage area prediction value and the risk factor of the target water body are brought into the target water body early warning coefficient calculation formula to obtain the early warning coefficient of the target water body.

9. The water environment cyanobacterial bloom early warning method according to claim 8, characterized in that, The early warning level of the target water body is determined, and the specific analysis method is as follows: The early warning coefficient of the target water body is extracted and input into the early warning level judgment model to output the judgment result of the early warning level of the target water body. The judgment result of the early warning level of the target water body includes values of 0, 1 and 2, if the judgment result of the early warning level of the target water body is 0, it is determined that the early warning level of the target water body is low, if the judgment result of the early warning level of the target water body is 1, it is determined that the early warning level of the target water body is medium, and if the judgment result of the early warning level of the target water body is 2, it is determined that the early warning level of the target water body is high.

10. A system for performing the water environment cyanobacterial bloom early warning method according to any one of claims 1 to 9, characterized by, It comprises: A multi-source data acquisition module for acquiring full-profile hydro-biochemical data, remote sensing image data and meteorological data of the target water body by a "space-air-ground" integrated monitoring network composed of a liftable layered sensor array, a satellite, a drone and a weather station at a preset period; A multi-source data intelligent fusion module for performing data cleaning, noise reduction and space-time registration processing based on the full-profile hydro-biochemical data, remote sensing image data and meteorological data of the target water body to obtain a cyanobacteria space-time distribution heat map; An adaptive prediction module for calculating prediction results of cyanobacteria cell density and water bloom coverage area by introducing a cyanobacteria vertical migration factor and a zooplankton feeding factor and weighted fusion based on the cyanobacteria space-time distribution heat map and historical data; A regionalized dynamic threshold judgment module for automatically matching the threshold value of the target water body by selecting an optimal early warning threshold value through ROC curve analysis to calculate the true positive rate and the false positive rate based on a climate zone and a water body type cluster to form a regional feature library; An early warning output module for calculating the early warning coefficient of the target water body based on the prediction results and the threshold value of the target water body, and then judging the early warning level of the target water body.

Images