Water body floating algae biomass scale monitoring and early warning system and method

Abstract

The invention discloses a water body floating algae biomass scale monitoring and early warning system and method. The system is an algae biomass scale measurement scheme based on an ecological mathematical model, and comprises a sample acquisition module 100, a water quality type identification module 200, an initial model matching module 300, a target model calibration module 400, an algae biomass scale prediction module 500 and an algal bloom information processing module 600. Through layered classification technology logic of triggering response and matching models according to water quality types, the system has stage jump type response to high-organic-matter water quality and has emergency management significance. By means of dynamically constructing a prediction model around a population, putting the importance of the organic matter content in front of the nitrogen and phosphorus content, classifying environmental factors to construct a mathematical function, representing the hydrodynamic characteristics of a water body by using surface runoff variables and the like, the data cost and the operation overhead are saved, and the popularization adaptability of the technical scheme is reduced. The invention further provides a water body floating algae biomass scale monitoring and early warning method based on the system.

Description

Technical Field

[0001] The present invention relates to a water environment monitoring technology, specifically to monitoring the biomass scale of phytoplankton in water bodies and implementing early warning based on the biomass scale change trend. The present invention belongs to the technical fields of measuring the biomass scale characteristics of phytoplankton in water bodies, measuring the trend characteristics of water blooms in water bodies, and water environment monitoring. Background Art

[0002] The measurement of the biomass scale of phytoplankton in water bodies, especially the monitoring and measurement of the trend of biomass scale changes, is the core content of water bloom outbreak monitoring and is also one of the important tasks of water body environmental management. When the signs of algal bloom outbreak appear, predicting the changing trend of algal bloom, that is, the growth trend of phytoplankton in water bodies becomes an important project of water body monitoring. With the integration of the feedback management concept into the environmental crisis response management mechanism, the scientific prediction of algal bloom occurrence has become an equally important task in water body management.

[0003] At present, the key issue of the monitoring and early warning system for phytoplankton growth in water bodies is to build an algae biomass scale prediction model that expresses environmental factors and algae content together. The algae growth model is established from the perspective of growth and death dynamics, and the algae movement model is established from the perspective of floating with water. The mathematical model is used to simulate the activity of phytoplankton in the water body. The change in the scale of phytoplankton in the water body is reflected through the mathematical process of biological population dynamics, so as to predict the occurrence or change trend of algae bloom. This is an ecological mathematical model method for algae bloom monitoring research. After the ecological mathematical model is established, the algae growth and death rate, as well as the movement speed with water, which can be calculated based on nutrients, nutrient absorption and other limiting factors, can be used to measure the current algae biomass scale (or concentration) in the water body.

[0004] Prior art ZL2022107809933 discloses an intelligent early warning system for water blooms based on the prediction of the nutrient status of water areas. The system includes a multi-scale information collection module, a nutrient status identification module, an early warning module for water blooms and a digital visualization module; specifically, the following steps are performed: collecting multi-scale water body information of the target water area, predicting the nutrient concentration of the target water area based on the multi-scale water body information, determining the nutrient status of the target water area based on the nutrient concentration, predicting the probability of water blooms based on the nutrient status of the water body, and generating corresponding warning feedback notification information based on the probability of water blooms; visualizing the nutrient status of the water body, the probability of water blooms and the warning feedback notification information. The defects of this technology are: first, taking multi-scale water body information as input, using machine learning algorithms to learn and construct an algae density prediction model to predict the algae density of the target water area, the original data items are too complicated, and it is difficult to meet the sample data requirements in actual production. In addition, the data collection task is heavy, the amount of calculation is large, and the cost of technology promotion and implementation is high. Second, algae blooms are typical r-strategies, which are not sensitive to subtle changes in environmental factors and do not require overly complex and precise models to predict their biomass. Therefore, a large number of data collection and calculation tasks actually waste productivity. Third, the comprehensive nutrient status index TLI of water bodies obtained after a large number of calculations only uses a single eutrophication index to segment and classify water body characteristics. In essence, it uses a unified model to predict algae growth and reproduction without distinguishing the characteristics of water body nutrient components. Its technical concept is forced. Moreover, under the technical concept of unified model prediction limitations, even if extremely rich sample data can be used to optimize and select the optimal solution of model parameters, there is still a problem: the cost of sample data collection is increased, and the "optimality" of the obtained target model may not be optimal. The large amount of computing cost is contrary to the original intention of the feedforward management of prediction. Summary of the invention

[0005] The purpose of the present invention is to provide a water body phytoplankton biomass scale monitoring and early warning system in view of the deficiencies of the prior art.

[0006] The present invention first provides a water body phytoplankton biomass scale monitoring and early warning system, and its technical solution is as follows.

[0007] A water body phytoplankton biomass scale monitoring and early warning system, characterized by: comprising a sample collection module 100, a water quality type identification module 200, an initial model matching module 300, a target model calibration module 400, and an algae biomass scale prediction module 500;

[0008] The sample collection module 100 acquires water quality monitoring data RD, meteorological data, and hydrological data of the monitored water body to form monitoring samples MS, sample MS1 is used to generate a prediction target model, and sample MS2 is used to predict phytoplankton biomass;

[0009] The water quality type identification module 200 reads the sample MS1 of the module 100, identifies the water quality type of the monitored water body and adds a water quality model label, and adds a water quality label E to the water quality type rich in organic matter;

[0010] The initial model matching module 300 matches G to predict the initial model according to the water quality model tag; if the water quality tag E is read, match G to predict the initial model I, and the G prediction initial model I is expressed as formula 1,

[0011] G=A×g max ×G(T)×G(I)×G(COD)×G(Q) Formula 1

[0012] Where, G is the growth rate of phytoplankton biomass, dimensionless, G(T) is the water temperature function, G(I) is the sunshine function, G(COD) is the COD Cr Function, G(Q) - Hydrodynamic function

[0013]

[0014]

[0015] In the formula, T is water temperature in °C, I is sunshine hours in h, q is hydrodynamic variable in units determined by the selected variable, COD is COD Cr Chemical oxygen demand, unit mg / L; both are model independent variables;

[0016] A - overgrowth coefficient, dimensionless, g max - Maximum growth rate of phytoplankton, dimensionless, T opi - Optimum water temperature, unit ℃, I opi - Optimal sunshine hours, unit: h, k COD - Half-saturated COD Cr Concentration, unit mg / L, k q - Hydrodynamic variable influence coefficient, dimensionless, q0 - hydrodynamic variable critical value, unit is determined by the selected variable; both are model parameters;

[0017] The target model calibration module 400 reads the sample MS1 of the module 100 and the G prediction initial model of the module 300, inputs the data RD into the G prediction initial model, determines the optimal solution set of the model parameters through parameter calibration, and generates the corresponding G prediction target model;

[0018] The algae scale prediction module 500 reads the sample MS2 and uses the G prediction target model to calculate the growth rate G of the phytoplankton biomass on the predicted day i. i .

[0019] The above-mentioned water body phytoplankton biomass scale monitoring and early warning system collects monitoring water body and environmental monitoring samples MS through the sample collection module 100. After the water quality type is identified by the water quality type identification module 200, the corresponding G prediction initial model is matched in the initial model matching module 300, and the parameters of the G prediction initial model are calibrated using the sample MS1 in the target model calibration module 400 to obtain the optimal solution set of the model parameters, thereby generating the G prediction target model for algae biomass prediction. In the algae scale prediction module 500, the sample MS2 (generally the real-time monitoring data of the i-th day of the prediction period) is input into the G prediction target model, and the growth rate G of the phytoplankton biomass on the i-th day can be measured. i . Measurement results Growth rate G i is the percentage value of the phytoplankton biomass scale on the i-th day of the monitoring water body prediction period. The specific meaning is that the phytoplankton biomass scale in the monitoring water body will be G i Value growth and expansion.

[0020] The key concepts of the above-mentioned water body phytoplankton biomass scale monitoring and early warning system of the present invention mainly include three aspects.

[0021] First, the water quality type rich in organic matter is used as the primary prediction model matching factor. Once module 200 identifies the water quality type rich in organic matter, it adds the water quality tag E, and module 300 reads it and starts matching G to predict the initial model. The previous research of the present invention found that although various water quality types are related to algal blooms, water quality with high organic matter content is more closely related to algal blooms. Maximum growth rate g max In this water environment, the algae reproduction capacity is characterized. In this water environment, since various biological and non-biological environmental factors are in a dynamic balance of comprehensive health, the algae reproduction capacity is mainly limited by mineral nutrition factors represented by nitrogen and phosphorus in the water. Therefore, g max The substantive meaning is the maximum growth rate under the condition of nutrient factors. However, when the water body is rich in organic matter, various environmental factors in the water body break through the comprehensive dynamic balance state, the algae reproduction restriction factors become invalid, and the organic-rich water body provides a more favorable ecological niche space for algae growth than the healthy water environment. The algae are actually in a state of g max The most rapid overgrowth state. Therefore, the present invention places the identification of water quality types rich in organic matter in a higher level concept than the identification of other water quality types, so that the prediction method can make the fastest feedback to the most urgent situation. For the identification of water quality types rich in organic matter, COD Cr The concentration ≥30 mg / L is the identification threshold. Through the cooperation of module 200 and module 300, a "first-level alert" can be formed for water bodies rich in organic matter, and the measurement and calculation of algae biomass growth rate can be triggered in time.

[0022] Second, various environmental factor functions G(T), G(I), G(COD), and G(Q) are reasonably constructed. The previous research of this invention found that for climate factors (water temperature T, sunshine I), the range and fluctuation law of the cosine function can fit the optimal value X opi (X represents climate factors) dynamics, so the mapping relationship between algae growth and such environmental factors can be accurately captured. For nutritional factors (chemical oxygen demand COD Cr ), through the half-saturation concentration variable k Y (Y represents nutritional factors), the sine function can accurately characterize the range, fluctuation law, and infinite approach characteristics between algae growth and such environmental factors. For hydrodynamic factors, the focus of their impact on algae growth lies in the "positive and negative" effects of low hydrodynamic promotion and high hydrodynamic hindrance, as well as the mitigation characteristics of the changes in this effect. The use of the natural logarithm function supplemented by the critical value q0 can effectively characterize this impact.

[0023] Third, design effective hydrodynamic variables. In the G prediction model, the hydrodynamic variable q is a variable that characterizes the flow velocity, flow rate or pressure of the monitored water body, and can also be a variable that characterizes the flow velocity, flow rate or pressure of the surface runoff that flows into the monitored water body. In the prior art, variables that characterize the flow velocity, flow rate or pressure of the monitored water body are usually used, but since such monitoring indicators are not routine monitoring items of hydrological monitoring stations. In particular, most natural small watersheds, river sections, and lakes are not equipped with corresponding hydrological stations, and long-term continuous hydrodynamic observation data cannot be obtained. The preliminary study of the present invention found that since the water area where algal blooms occur is usually close to the water bank, the dynamic characteristics of the water body are mainly affected by the influx of surface runoff. By using the relationship between the two, the hydrodynamic variable q of the G prediction model that can characterize the flow velocity, flow rate or pressure of the surface runoff that flows into the monitored water body can be used as an effective model independent variable. Runoff data can be continuously obtained through meteorological data, which ensures the continuity and stability of the data required for building the target model, greatly reduces the cost of data acquisition, and improves the applicability of the prediction method.

[0024] Based on the above-mentioned water body phytoplankton biomass scale monitoring and early warning system, the present invention provides an optimization scheme thereof. The optimization schemes can be implemented simultaneously or separately without affecting the system logic.

[0025] Optimization 1: Enrich the prediction of water quality types with low organic matter content. In the water quality type recognition module 200, for water quality types with low organic matter content, water quality labels are added based on the water quality N / P ratio. Specifically, if N / P ≤ 10, the water quality label NP- is added; if N / P ≥ 22.6, the water quality label NP+ is added; if 10 < N / P < 22.6, the water quality label NP± is added. In the initial model matching module 300, if the water quality label NP- is read, the G prediction initial model II (Equation 2) is matched; if the water quality label NP+ is read, the G prediction initial model III (Equation 3) is matched; if the water quality label NP± is read, the G prediction initial model IV (Equation 4) is matched.

[0026] G = g max ×G(T)×G(I)×G(N)×G(Q) Equation 2

[0027] G = g max ×G(T)×G(I)×G(P)×G(Q) Equation 3

[0028]

[0029] In the formula, G(N) - total nitrogen function, G(P) - total phosphorus function, G(N / P) - N / P function;

[0030]

[0031] In the formula, N - total nitrogen, unit mg / L, P - total phosphorus, unit mg / L, N / P - nitrogen-phosphorus ratio, dimensionless; all are independent variables of the model; k N - half-saturation total nitrogen concentration, unit mg / L, k P - half-saturation total phosphorus concentration, unit mg / L, k N / P - half-saturation nitrogen-phosphorus ratio concentration, dimensionless; all are model parameters.

[0032] Optimization 2: The system includes a water bloom information processing module 600. The module 600 evaluates the occurrence / change trend of water bloom according to G i and the preset water bloom identification condition G (Equation 12). If the water bloom identification condition G is satisfied, a water bloom outbreak warning message is issued.

[0033] G i >> T G Equation 12

[0034] In the formula, T G - growth rate threshold of water bloom occurrence, dimensionless quantity.

[0035] By introducing the water bloom identification condition G and the corresponding judgment steps, it is possible to identify G iIn extremely critical situations, there is no need to consider the impact of natural death of algae biomass on total biomass, and algal bloom forecast and warning can be directly carried out.

[0036] Optimization 3: If the algal bloom identification condition G is not established in module 600, the effect of the natural death of algae biomass on the total biomass can be further considered, and the change in algae biomass under the action of life and death can be measured.

[0037] First, the target model calibration module 400 reads the monitoring sample MS of the module 100 and the G prediction initial model of the module 300, combines the G prediction initial model with the D prediction model (Formula 14), and constructs the c expressed in Formula 13 a Predict the initial model and input the data RD into c a Predict the initial model, determine the optimal solution set of model parameters through parameter calibration, and generate the corresponding c a Secondly, the algae scale prediction module 500 obtains the monitoring data MD of the water quality of the monitored water body on day i i , using c a The prediction target model estimates the size of phytoplankton biomass on day i. a,i Finally, the water bloom information processing module 600 is based on c a,i Evaluate the occurrence / change trend of algal bloom with the preset algal bloom identification conditions CA.

[0038] c a (t+1)=c a (t)+c a (t)·(GD) Formula 13

[0039]

[0040] In the formula, c a (t), c a (t+1) - the chlorophyll content of phytoplankton in the monitored water body on day t and day t+1, in μg / L; D - phytoplankton biomass mortality rate, dimensionless, d max - Maximum mortality rate of phytoplankton, dimensionless, a model parameter.

[0041] In this optimization scheme, the phytoplankton mortality model D uses water temperature T as the basic factor of the model independent variable and introduces the optimal water temperature T opi Compared with the existing technology, the expression of algae mortality rate D is more accurate.

[0042] By solving the problem of predicting the daily growth of phytoplankton in the monitored water body, the daily change trend of the scale of phytoplankton in the monitored water body can be predicted. By introducing the algae growth threshold T that induces algal bloom in the monitored water body, ca Indicators, constructed from T caExpression of c a The judgment condition function can realize the early warning of the occurrence of algae bloom in the monitored water body. The threshold T corresponding to the change of algae bloom stage is used. ca , then the stage warning can be further completed. Threshold T ca It can be the daily growth or the cumulative growth in the previous period. Therefore, further optimization is that the water bloom identification condition CA can be expressed by formula 15 to evaluate the different stages of the water bloom process.

[0043] c a,i ≤f(T ca,j ,j=1,2,3,…n) Formula 15

[0044] Where, T ca,j ,j=1,2,3,…n-the biomass threshold of phytoplankton in the jth stage or the jth degree of water bloom in the monitoring water body T ca .

[0045] The present invention also provides a method for monitoring and early warning of the biomass scale of phytoplankton in water bodies, which is specifically as follows.

[0046] A method for monitoring and warning the biomass scale of phytoplankton in a water body, which is implemented by using the above-mentioned monitoring and warning of the biomass scale of phytoplankton in a water body, is characterized by:

[0047] First, conduct on-site surveys to obtain background data on monitored water bodies;

[0048] Secondly, a monitoring sample MS1 is obtained through the sample collection module 100;

[0049] Thirdly, a prediction target model is constructed through the water quality type identification module 200, the initial model matching module 300, and the target model calibration module 400;

[0050] Again, the monitoring sample MS2 is obtained through the sample collection module 100;

[0051] Secondly, the monitoring sample MS2 is input into the prediction target model to obtain the prediction data of the biomass scale of phytoplankton in the monitored water body;

[0052] Finally, the risk of algal bloom in the monitored water bodies is evaluated by monitoring the predicted data of phytoplankton biomass size.

[0053] The field survey referred to in this technology includes various surveys, investigations, mapping, and measurements of the small watershed and the two banks where the monitoring and early warning scheme is located, as well as existing simulation experiments, test experiments, observation experiments, and analysis experiments in the field, as well as the acquisition of historical water quality and hydrological records, relevant technical specifications, and empirical methods and data acquisition for reference. The data obtained from the field survey are collectively referred to as the background data of this technical scheme.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows: First, the water body phytoplankton biomass scale monitoring and early warning system of the present invention is an algae biomass scale measurement scheme based on the algae ecological mathematical model. The core module of the system adopts the growth and death dynamic perspective and the water-borne phytoplankton perspective to establish an algae quantity scale change measurement model, and reflects the scale change of phytoplankton in the water body through the mathematical process of biological population dynamics. Compared with the prior art technical solution that uses multi-scale information as the nutrient status of water bodies and water bloom early warning, the present invention strengthens the influence of environmental factors on algae biomass changes, greatly reduces the data requirements for system operation, saves data collection costs and system computing overhead, and reduces the promotion cost of technical solutions. Second, eutrophication of water bodies is the most critical factor inducing water blooms, but in different "nutritional type" environments, there are differences in the reproduction behavior of algae populations. The present invention constructs different prediction models according to the nutritional characteristics of water quality, which actually adjusts the importance of nutritional factors to algae growth according to the nutritional physiological metabolic characteristics in different nutritional environments, which is more in line with biological behavior characteristics. At the same time, by constructing a classification model, the mathematical expression of the model is placed in a purer environment with more concentrated variable influences, the function mapping relationship is easier to identify, and the calculation is more economical. Third, on the basis of the water quality matching prediction model, the present invention further superimposes the technical logic of hierarchical execution based on previous research. Since it is found that the high concentration of organic matter in the water body promotes algae reproduction much more than NP nutrients, the present invention designs the technical logic of triggering response according to water quality types, so that the core module of the system is always in a stage-jump response state to the organic-rich water quality type, shortening the system monitoring response time. Since organic-rich water has a strong promoting effect on algae reproduction, the rapid measurement and evaluation of the algae scale in this type of water quality state has special significance for emergency management. This is a new concept different from the prior art. Fourth, by respectively analyzing the range, change, rhythm and other characteristics of the mapping relationship between algae growth and different types of environmental factors, the present invention constructs factor functions with different mathematical characteristics from the environmental factor type. Fifth, based on the analysis of the location of algal blooms in the water area and the hydrological characteristics on both sides of the natural water body, surface runoff variables that are easy to obtain stably are introduced to replace the hydrodynamic variables of the monitored water body that require special measurements to obtain, which reduces the difficulty of data collection for system operation and improves the applicability of the implementation of the technical solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a schematic diagram of the framework structure of the water body phytoplankton biomass scale monitoring and early warning system in Example 1.

[0056] Figure 2 It is a schematic diagram of the framework structure of the water body phytoplankton biomass scale monitoring and early warning system according to the second embodiment.

[0057] Figure 3It is a schematic diagram of the framework structure of the water body phytoplankton biomass scale monitoring and early warning system in Example 3.

[0058] Figure 4 It is the comparison between the actual value and the predicted value of Example 4.

[0059] Figure 5 (a), (b), and (c) are comparisons between the actual value and the predicted value in Example 5.

[0060] The numbers in the accompanying drawings are:

[0061] Sample collection module 100; water quality type identification module 200; reading unit 210; labeling unit 220; initial model matching module 300; reading unit 310; model matching unit 320; matching output unit 330; target model calibration module 400; reading unit 410; D model storage unit 420; model combination unit 430; model calibration unit 440; target model unit 450; algae biomass scale prediction module 500; reading unit 510; prediction unit 520; result output unit 530; water bloom information processing module 600. DETAILED DESCRIPTION

[0062] The preferred embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0063] Embodiment 1

[0064] In order to build an ecologically clean small watershed, the algal bloom risk and dynamic monitoring of the small watershed are implemented, and the water body phytoplankton biomass scale monitoring and early warning system of the present invention is constructed.

[0065] Designate small watersheds as monitored water bodies, and conduct on-site surveys to obtain basic data on monitored water bodies.

[0066] Figure 1 This is a schematic diagram of the framework structure of the water body phytoplankton biomass scale monitoring and early warning system in Example 1.

[0067] The water body phytoplankton biomass scale monitoring and early warning system (hereinafter referred to as the system) includes a sample collection module 100, a water quality type identification module 200, an initial model matching module 300, a target model calibration module 400, and an algae biomass scale prediction module 500.

[0068] 1. Sample collection module 100

[0069] The sample collection module 100 collects monitoring samples MS required by the system, and the monitoring samples MS include water quality data, meteorological data, and hydrological data.

[0070] The sample collection module 100 collects monitoring samples MS1 during the monitoring period T1 for generating a prediction target model. The data items of each group of MS1 are shown in Table 1. Monitoring samples MS2 are collected during the monitoring period T2 for predicting the biomass characteristics of phytoplankton; the monitoring period T2 is set according to the target task of the monitoring project. The data items of monitoring sample MS2 do not contain the chlorophyll content c in the data items of monitoring sample MS1. a (t).

[0071] Table 1 MS1 data items of monitoring samples

[0072]

[0073] In Table 1, runoff depth is the hydrodynamic variable q specifically selected in this example.

[0074] 2. Water quality type identification module 200

[0075] The water quality type identification module 200 reads the monitoring sample MS1 of the module 100, identifies the water quality type of the monitored water body and adds a water quality model tag.

[0076] After the reading unit 210 reads the sample MS1 of the module 100 , the label adding unit 220 first identifies whether the water is rich in organic matter. If it is rich in organic matter, a water quality label E is added to the sample MS1 .

[0077] 3. Initial model matching module 300

[0078] The initial model matching module 300 predicts the initial model according to the water quality model label matching G.

[0079] The reading unit 310 detects the label filling status of the unit 220 in real time. Once the water quality label E is read, the G prediction initial model I is retrieved from the model matching unit 320 and sent to the matching output unit 330.

[0080] 4. Target model calibration module 400

[0081] The target model calibration module 400 reads the monitoring sample MS1 of the module 100 and the G prediction initial model of the module 300, inputs the sample MS1 into the G prediction initial model, determines the optimal solution set of the model parameters through parameter calibration, and generates the corresponding G prediction target model.

[0082] The reading unit 410 of module 400 reads the monitoring sample MS1 of module 100 and the G prediction initial model in unit 330. The model calibration unit 440 inputs the monitoring sample MS into the G prediction initial model, determines the optimal solution set of model parameters through parameter calibration, generates the corresponding G prediction target model and sends it to the target model unit 450.

[0083] 5. Algae scale prediction module 500

[0084] The algae scale prediction module 500 obtains the monitoring sample MS2 of the monitored water body on the predicted day i, and calculates the growth rate G of the phytoplankton biomass on the day i by using the G prediction target model. i .

[0085] The reading unit 510 of the module 500 reads the monitoring sample MS2 in the module 100 and the G prediction target model in the unit 450. The sample MS2 is the monitoring sample MS2 of the monitored water body on the predicted day i. i , and the prediction unit 520 inputs the monitoring sample MS2 into the G prediction target model to calculate the growth rate G of the phytoplankton biomass on the day i. i , and the result output unit 530 outputs the prediction result.

[0086] Embodiment 2

[0087] Improve the monitoring and early warning system for the scale of phytoplankton biomass in the water body of Embodiment 1. The same parts will not be repeated. The differences lie in adding the processing function for water quality types that are not rich in organic matter and the algal bloom information processing module 600.

[0088] Figure 2 is a schematic diagram of the framework structure of the monitoring and early warning system for the scale of phytoplankton biomass in the water body of Embodiment 2.

[0089] 1. Water quality type identification module 200

[0090] After the reading unit 210 reads the sample MS1 of the module 100, the label adding unit 220 first identifies whether it is a water quality type rich in organic matter. If it is a water quality type rich in organic matter, the water quality label E is added to the sample MS1; otherwise, the water quality type is further identified according to the water quality N / P ratio. Specifically: if N / P ≤ 10, the water quality label NP- is added; if N / P ≥ 22.6, the water quality label NP+ is added; if 10 < N / P < 22.6, the water quality label NP± is added.

[0091] By adopting a double-layer water quality type identification logic, the unit 220 first determines whether to add the label E. Only when the label E does not need to be added, the NP type labels (NP-, NP+, NP±) are added. Therefore, there is always and only one water quality label in the unit 220, and the label E belongs to the optimization level.

[0092] 2. Initial model matching module 300

[0093] The reading unit 310 detects the label filling status of the unit 220 in real time. Once the newly generated water quality label is read, the corresponding type of G prediction initial model is retrieved from the model matching unit 320 and sent to the matching output unit 330. Due to the characteristics of label filling and storage in the unit 220, the initial model matching module 300 can ensure that the relevant operations of organic-rich water quality can always be in a real-time priority state, which also ensures that the system can model and predict the hierarchical classification of different water quality types. The corresponding relationship between water quality type, label, and G prediction initial model is shown in Table 2.

[0094] Table 2 Correspondence between water quality type, label, and G prediction initial model

[0095] Serial number Water quality type Identification conditions Water quality label G predicts the initial model 1 Rich in organic matter <![CDATA[COD Cr Concentration ≥ 30 mg / L]]> E Model I (Formula 1) 2 Low N / P N / P≤10 NP- Model II (Formula 2) 3 High N / P N / P≥22.6 NP+ Model III (Formula 3) 4 Medium N / P 10<N / P<22.6 NP± Model IV (Formula 4)

[0096] 3. Water bloom information processing module 600

[0097] Water bloom information processing module 600 according to G i The occurrence / change trend of algal bloom is evaluated with the preset algal bloom identification condition G. If the algal bloom identification condition G is met, the algal bloom outbreak forecast information is issued.

[0098] In this example, the water bloom identification condition G is expressed according to formula 12.

[0099] Embodiment 3

[0100] The water body phytoplankton biomass scale monitoring and early warning system of the second embodiment is improved. The similarities are not repeated, and the differences are that the natural death factor of algae biomass is introduced into the total biomass prediction, and the module 600 is improved.

[0101] Figure 3 It is a schematic diagram of the framework structure of the water body phytoplankton biomass scale monitoring and early warning system in Example 3.

[0102] 1. Target model calibration module 400

[0103] The reading unit 410 of the target model calibration module 400 reads the monitoring sample MS of the module 100 and the G prediction initial model of the module 300, and reads the D prediction model (Formula 14) in the D model storage unit 420 at the same time. The model combination unit 430 constructs c according to Formula 13 a Predict the initial model, the model calibration unit 440 inputs the monitoring sample MS1 into c a Predict the initial model, determine the optimal solution set of model parameters through parameter calibration, and generate the corresponding c a The predicted target model is sent to the target model unit 450.

[0104] 3. Algae scale prediction module 500

[0105] The algae scale prediction module 500 obtains the monitoring sample MS2 of the monitored water body on the prediction day i i , using c a The prediction target model estimates the size of phytoplankton biomass on day i. a,i .

[0106] 3. Water bloom information processing module 600

[0107] Water bloom information processing module 600 according to c a,i The occurrence / change trend of algal bloom is evaluated with the preset algal bloom identification conditions CA. If the algal bloom identification conditions CA are met, algal bloom outbreak forecast information is issued.

[0108] In this example, the algal bloom identification condition CA is expressed according to Formula 15, thereby realizing the forecast and early warning of different stages of algal bloom changes.

[0109] Embodiment 4

[0110] The water body phytoplankton biomass scale monitoring and early warning system of Example 3 is applied to an ecological clean construction project in a small watershed of the JQ River in a certain place to predict the changing trend of the phytoplankton biomass in the water body and, when necessary, to forecast water blooms.

[0111] 1. Collect monitoring samples MS1

[0112] The sample collection module 100 collects monitoring samples MS1 during the monitoring period T1, and obtains 97 groups of samples in total (each sample is a monitoring day data, the same below). Table 3 is part of the data of the 97 groups of samples MS1.

[0113] Table 3 Partial data of MS1 of 97 samples in JQ River small watershed

[0114] Serial number <![CDATA[COD Cr ]]> N P N / P T I q <![CDATA[c a (t)]]> 1 23.00 0.90 0.55 1.65 10.90 0.58 0.13 20.28 2 26.00 0.84 0.42 1.99 10.80 0.58 0.13 54.25 3 26.00 0.84 0.42 1.99 11.10 0.58 0.13 35.21 4 23.00 0.95 0.47 2.01 23.30 0.53 0.09 31.22 5 29.00 0.37 0.18 2.04 12.50 3.98 0.04 203.90 6 22.00 0.30 0.14 2.14 12.50 3.98 0.04 79.63 7 18.00 0.25 0.11 2.27 12.50 3.98 0.04 94.79 8 25.00 0.33 0.15 2.28 12.50 3.98 0.04 296.20 9 17.00 0.39 0.17 2.28 13.00 1.75 0.06 165.60 10 24.00 0.34 0.14 2.43 12.50 3.98 0.04 174.90 11 22.00 0.47 0.19 2.53 12.50 3.98 0.04 39.32 12 22.00 0.36 0.14 2.57 12.50 3.98 0.04 122.60 13 21.00 0.29 0.11 2.59 12.50 3.98 0.04 234.90 14 24.00 0.30 0.11 2.65 22.70 4.94 0.08 34.47 15 25.00 0.42 0.16 2.66 12.50 3.98 0.04 234.60 16 18.00 0.60 0.22 2.74 12.20 1.85 0.18 104.27 17 22.00 0.49 0.18 2.74 12.30 2.70 0.04 80.55 18 21.00 0.49 0.17 2.83 12.50 3.98 0.04 28.31 19 24.00 0.59 0.20 2.96 11.90 1.89 0.04 107.80 20 20.00 0.52 0.17 3.00 11.10 0.58 0.13 34.77 …… …… …… …… …… …… …… …… ……

[0115] 2. Water quality type identification

[0116] According to the sample MS1 water quality classification index (COD Cr ≤30mg / L, N / P≤10,) water quality type identification module 200 identifies the water quality type of sample MS1 according to Table 2, and adds a water quality label NP-.

[0117] 3. Initial model matching

[0118] The initial model matching module 300 reads the water quality label NP- and retrieves G to predict the initial model II.

[0119] 4. Target model rate

[0120] The target model calibration module 400 reads the sample MS1 of the module 100, the G prediction initial model II output by the module 300, and the D prediction model stored in itself, and constructs ca Predict the initial model (Formula 13-2), and input the sample MS1 into Formula 13-2. Through parameter calibration (the optimization algorithm in this example specifically uses a genetic algorithm) to determine the optimal solution set of model parameters (Table 4), generate the corresponding c a Prediction target model (Formula 13-2-1).

[0121]

[0122] Table 4 Optimal solution of model parameters in formula 13-2

[0123] Model parameters <![CDATA[g max ]]> <![CDATA[T opi ]]> <![CDATA[I opi ]]> <![CDATA[k N ]]> <![CDATA[d max ]]> <![CDATA[k q ]]> <![CDATA[q0]]> Optimal solution 0.24 29.23 7.24 0.10 0.78 0.14 0.34

[0124]

[0125] 5. Algae scale prediction

[0126] The sample collection module 100 collects 19 groups of monitoring samples MS2 during the monitoring period T2, each group being a monitoring sample for the predicted day i. The algae scale prediction module 500 reads the samples MS2 and the c generated by the module 400. a Predict the target model and input the sample MS2 into c a The prediction target model calculates the scale c of phytoplankton biomass on day i a,i .

[0127] To test the prediction accuracy of the system, 19 groups of samples MS2 were used. a The actual detection value verification module 500 calculates the predicted value. The results show that the determination coefficient R of the degree of fit between the predicted value of the verification model and the true value is 2 =0.77, which meets the prediction accuracy requirements. Figure 4 It is the comparison between the actual value and the predicted value of Example 4.

[0128] 6. Water bloom information processing

[0129] In this example, the threshold value of phytoplankton biomass T for the jth degree of algal bloom is determined by referring to the Technical Specifications for Algal Bloom Degree Classification and Monitoring (DB44 / T 2261-2020). ca,j ,j. The prediction results were compared with the T ca , which can determine the evolution of the degree of algal bloom.

[0130] Embodiment 5

[0131] The water body phytoplankton biomass scale monitoring and early warning system of Example 3 is applied to an ecological clean construction project in a small watershed of DL Creek in a certain place to predict the changing trend of phytoplankton biomass in the water body and, when necessary, to forecast water blooms.

[0132] A total of three prediction projects were designed and carried out for the change trend of phytoplankton biomass in the DL Creek small watershed using the system of the present invention, and each prediction project was designed for a different research period in the DL Creek small watershed. The description of the content of the system application operation process that is the same or highly similar to that of Example 4, as well as the repeated content descriptions in the three prediction projects are omitted here, and only the main data charts related to biomass measurement in each prediction project are listed. The work content of module 600 is omitted in all three prediction projects.

[0133] Table 5 Description of the main contents of the three prediction projects in the DL Creek small watershed

[0134]

[0135] Table 6 Partial data of 139 samples MS1 in the DL Creek small watershed prediction project

[0136]

[0137]

[0138] Table 7 Partial data of 60 samples MS1 in DL Creek small watershed prediction project 2

[0139] Serial number P N <![CDATA[COD Cr ]]> T N / P I q <![CDATA[c a (t)]]> 1 0.08 1.55 17 8.1 19.32 2.52 0.075 21.11 2 0.07 1.06 16 11.5 15.53 4.37 0.067 45.49 3 0.05 1.16 17 12.1 22.47 0.91 0.075 34.12 4 0.06 1.03 15 13.7 17.79 0.91 0.074 43.03 5 0.06 1.00 17 11.1 16.04 1.95 0.061 45.98 6 0.07 1.08 15 11.8 14.76 1.15 0.097 39.80 7 0.10 1.08 17 15.3 10.79 2.11 0.065 148.00 8 0.06 0.96 17 19.5 17.05 1.87 0.059 32.53 9 0.06 0.95 18 19.1 16.40 3.12 0.12 35.08 10 0.06 1.00 19 18.4 17.31 1.47 0.059 37.89 11 0.06 1.35 17 15.6 21.30 2.17 0.074 29.88 12 0.05 1.13 20 21.7 21.32 5.75 0.061 37.11 13 0.10 1.22 20 23.4 12.20 4.08 0.119 35.58 14 0.06 1.10 22 28.2 18.33 6.76 0.047 43.96 15 0.05 1.08 21 25.2 21.60 3.01 0.116 30.34 16 0.07 1.02 19 24.1 14.57 2.52 0.166 54.09 17 0.08 1.08 21 25.1 13.50 3.44 0.053 38.11 18 0.06 1.00 20 28.9 16.67 5.72 0.122 46.56 19 0.13 1.38 16 29.5 10.62 7.43 0.034 43.10 20 0.07 1.03 19 27.9 14.71 4.15 6.228 30.29 …… …… …… …… …… …… …… …… ……

[0140] Table 8 Partial data of 330 samples MS1 of DL Creek small watershed prediction project

[0141]

[0142]

[0143] Table 9 Optimal solutions for model parameters of three prediction projects in the DL Creek watershed

[0144]

[0145]

[0146]

[0147]

Claims

1. The water body phytoplankton biomass scale monitoring and early warning system is characterized by: It includes a sample collection module 100, a water quality type identification module 200, an initial model matching module 300, a target model calibration module 400, and an algal biomass scale prediction module 500; The sample collection module 100 obtains water quality monitoring data RD, meteorological data, and hydrological data of the monitored water body to form a monitoring sample MS. The sample MS1 is used to generate a prediction target model, and the sample MS2 is used for predicting the phytoplankton biomass; The water quality type identification module 200 reads the sample MS1 of the module 100, identifies the water quality type of the monitored water body and adds a water quality model label. The water quality type rich in organic matter adds a water quality label E; The initial model matching module 300 matches and predicts an initial model G according to the water quality model label; if the water quality label E is read, it matches and predicts the initial model Ⅰ of G. The initial model Ⅰ of G prediction is expressed as Equation 1, G = A × g max × G(T) × G(I) × G(COD) × G(Q) Equation 1 Where, G is the growth rate of phytoplankton biomass, dimensionless, G(T) is the water temperature function, G(I) is the sunshine function, G(COD) is the COD Cr Function, G(Q) - Hydrodynamic function In the formula, T is water temperature in °C, I is sunshine hours in h, q is hydrodynamic variable in units determined by the selected variable, COD is COD Cr Chemical oxygen demand, unit mg / L; both are model independent variables; A - overgrowth coefficient, dimensionless, g max - Maximum growth rate of phytoplankton, dimensionless, T opi - Optimum water temperature, unit ℃, I opi - Optimal sunshine hours, unit: h, k COD - Half-saturated COD Cr Concentration, unit mg / L, k q - Hydrodynamic variable influence coefficient, dimensionless, q0 - hydrodynamic variable critical value, unit is determined by the selected variable; both are model parameters; The target model calibration module 400 reads the sample MS1 of the module 100 and the initial model Ⅰ of G prediction of the module 300, inputs the data RD into the initial model Ⅰ of G prediction, determines the optimal solution set of the model parameters through parameter calibration, and generates the corresponding target model Ⅰ of G prediction; The algae scale prediction module 500 reads the sample MS2 and uses the G prediction target model to calculate the growth rate G of the phytoplankton biomass on the predicted day i. i .

2. The monitoring and early warning system for the scale of phytoplankton biomass in water bodies according to claim 1, characterized in that: In the water quality type identification module 200, for water quality types that are not rich in organic matter, water quality labels are added according to the water quality N / P ratio. If N / P ≤ 10, the water quality label NP- is added. If N / P ≥ 22.6, the water quality label NP+ is added. If 10 < N / P < 22.6, the water quality label NP± is added; In the initial model matching module 300, if the water quality label NP- is read, it matches and predicts the initial model Ⅱ of G. If the water quality label NP+ is read, it matches and predicts the initial model Ⅲ of G. If the water quality label NP± is read, it matches and predicts the initial model Ⅳ of G; The initial model Ⅱ of G prediction, the initial model Ⅲ of G prediction, and the initial model Ⅳ of G prediction are respectively expressed as Equations 2, 3, and 4, G = g max ×G(T)×G(I)×G(N)×G(Q), Equation 2 G = g max × G(T) × G(I) × G(P) × G(Q) Equation 3 In the formula, G(N) - total nitrogen function, G(P) - total phosphorus function, G(N / P) - N / P function; In the formula, N - total nitrogen, unit mg / L, P - total phosphorus, unit mg / L, N / P - nitrogen-phosphorus ratio, dimensionless; all are model independent variables; k N - Half-saturated total nitrogen concentration, unit: mg / L, k P - Half-saturated total phosphorus concentration, unit: mg / L, k N / P - Half-saturated nitrogen-phosphorus ratio concentration, dimensionless; both are model parameters.

3. The water body phytoplankton biomass scale monitoring and early warning system according to claim 2 is characterized by: The reading unit 310 of the initial model matching module 300 reads the water quality label added by the module 200. If the water quality label E is read, it retrieves the initial model Ⅰ of G prediction from the model matching unit 320 and sends it to the matching output unit 330, and sends a command to read the matching output unit 330 to the module 400; if the water quality label E is not read, it continues to read other water quality labels, and then retrieves the initial model Ⅱ / Ⅲ / Ⅳ of G prediction that matches the label from the model matching unit 320 and sends it to the matching output unit 330, and sends a command to read the matching output unit 330 to the module 400.

4. The water body phytoplankton biomass scale monitoring and early warning system according to claim 3 is characterized by: The water bloom information processing module 600 is also included. The water bloom information processing module 600 is based on G i The occurrence / change trend of water bloom is evaluated with the preset water bloom identification condition G. If the water bloom identification condition G is established, the water bloom outbreak forecast information is issued; the water bloom identification condition G is expressed according to formula 12; G i >>T G Formula 12 Where, T G - Growth rate threshold for algal bloom occurrence, dimensionless.

5. The water body phytoplankton biomass scale monitoring and early warning system according to claim 4 is characterized by: If the algae bloom identification condition G is not met, the target model calibration module 400 reads the monitoring sample MS of the module 100 and the G prediction initial model of the module 300, combines the G prediction initial model with the D prediction model, and constructs the c expressed in formula 13. a Predict the initial model and input the data RD into c a Predict the initial model, determine the optimal solution set of model parameters through parameter calibration, and generate the corresponding c a Prediction target model; the D prediction model is expressed as formula 14; c a (t + 1)= c a (t)+ c a (t)·(G - D) Equation 13 In the formula, c a (t), c a (t+1) - the chlorophyll content of phytoplankton in the monitored water body on day t and day t+1, in μg / L; D - phytoplankton biomass mortality rate, dimensionless, d max - Maximum mortality rate of phytoplankton, dimensionless, a model parameter; The algae scale prediction module 500 obtains the monitoring data MD of the water quality of the monitored water body on day i i , using c a The prediction target model estimates the size of phytoplankton biomass on day i. a,i ; The water bloom information processing module 600 is based on c a,i Evaluate the occurrence / change trend of algal bloom with the preset algal bloom identification conditions CA.

6. The water body phytoplankton biomass scale monitoring and early warning system according to claim 5 is characterized by: The target model calibration module 400 has a reading unit 410 that reads the monitoring sample MS of the module 100, the G prediction initial model of the module 300, and the D prediction model in the D model storage unit 420, and a model combination unit 430 that reads the G prediction initial model and the D prediction model in the unit 420 to construct a D prediction model. a Predict the initial model, the model calibration unit 440 monitors the sample MS input c a Predict the initial model, determine the optimal solution set of model parameters through parameter calibration, and generate the corresponding c a The predicted target model is sent to the target model unit 450.

7. The water body phytoplankton biomass scale monitoring and early warning system according to claim 6 is characterized by: The water bloom identification condition CA in the water bloom information processing module 600 is expressed according to Equation 15, c a,i ≤f(T ca,j ,j=1,2,3,…n) Formula 15 Where, T ca,j ,j=1,2,3,…n-the biomass threshold of phytoplankton in the jth stage or the jth degree of water bloom in the monitoring water body T ca .

8. The water body phytoplankton biomass scale monitoring and early warning system according to any one of claims 1 to 7, characterized in that: The hydrodynamic variable q is a variable representing the flow velocity or flow rate or pressure of the monitored water body, or a variable representing the flow velocity or flow rate or pressure of the surface runoff flowing into the monitored water body.

9. The system for monitoring and early warning of phytoplankton biomass in water bodies according to any one of claims 1 to 7, characterized in that: The organic matter-rich water quality type is COD Cr The threshold value is concentration ≥30 mg / L.

10. A method for monitoring and warning the biomass scale of phytoplankton in a water body, which is implemented by using the system for monitoring and warning the biomass scale of phytoplankton in a water body according to any one of claims 1 to 7, characterized in that: First, conduct on-site surveys to obtain background data on monitored water bodies; Secondly, a monitoring sample MS1 is obtained through the sample collection module 100; Thirdly, a prediction target model is constructed through the water quality type identification module 200, the initial model matching module 300, and the target model calibration module 400; Again, the monitoring sample MS2 is obtained through the sample collection module 100; Secondly, the monitoring sample MS2 is input into the prediction target model to obtain the prediction data of the biomass scale of phytoplankton in the monitored water body; Finally, the risk of algal bloom in the monitored water bodies is evaluated by monitoring the predicted data of phytoplankton biomass scale.

Images