A method for monitoring and evaluating long-term hydrological effects of grassed swales

Abstract

The application provides a kind of grassed swale long-term hydrological effect monitoring and evaluation method, representative rainfall events in future of to be evaluated grassed swale are obtained by screening using data availability evaluation model;Collect meteorological parameters and hydrological performance parameters of representative rainfall events;Grassed swale instantaneous saturated permeability coefficient value under representative rainfall events is calculated using grassed swale instantaneous saturated permeability coefficient inversion method, and meteorological parameter value constitutes second training data set;Based on second training data set, establish instantaneous saturated permeability coefficient dynamic prediction model;Grassed swale instantaneous saturated permeability coefficient prediction value under new rainfall events outside representative rainfall events is obtained by using instantaneous saturated permeability coefficient dynamic prediction model;Grassed swale instantaneous saturated permeability coefficient prediction value is input into grassed swale hydrological model, and hydrological performance parameter simulation result is obtained.The method of the application can accurately simulate the long-term hydrological performance of grassed swale under different climate conditions.

Description

Technical Field

[0001] This invention belongs to the field of municipal facility management and service technology, specifically, it relates to a method for monitoring and evaluating the long-term hydrological effects of planted swales. Background Technology

[0002] Vegetated swales are among the most widely used municipal facilities for stormwater management both domestically and internationally. Accurately predicting the changing patterns of their runoff control capacity (i.e., hydrological effects) is a crucial foundation for related planning, design, and operation and maintenance. The saturated permeability coefficient is a major controlling factor of the hydrological effects of vegetated swales, but existing monitoring or calibration methods cannot quantitatively assess their heterogeneity along the course and their long-term variation patterns. Current research on the long-term hydrological effects of vegetated swales faces the following challenges:

[0003] 1. Spatial uncertainty: Traditional grassed swales are generally long, and the heterogeneity of their filler material leads to variations in the saturated permeability coefficient along the swale. Single-point measurements cannot accurately characterize the permeability performance of the entire system. Currently, there is no reliable method to calculate the equivalent value of the saturated permeability coefficient of a single facility, resulting in a significant difference between conventional hydrological effect simulation results and actual facility performance.

[0004] 2. Time uncertainty: The saturated permeability coefficient of vegetated swales exhibits a long-term decline and periodic fluctuation due to runoff scouring, water evaporation, and vegetation root growth. However, existing infiltration theory models and hydrological effect simulation software do not take this into account, and there is a lack of quantitative analysis and prediction methods for the long-term changes in the saturated permeability coefficient. Therefore, it cannot meet the assessment needs of the hydrological effects of vegetated swales in the long term or even throughout the entire life cycle.

[0005] 3. Methodological uncertainty: Existing studies generally use the method of estimating the instantaneous saturated permeability coefficient of vegetated swales based on the inflow and outflow curves of vegetated swales during rainfall events to study the long-term hydrological effects of vegetated swales. However, in current practice, there is a lack of screening methods for relevant meteorological and hydrological performance data of vegetated swales under rainfall events, resulting in poor reliability of the research results. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method for monitoring and evaluating the long-term hydrological effects of vegetated swales, which can accurately simulate the long-term hydrological performance of vegetated swales under different climatic conditions.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] This invention provides a method for monitoring and evaluating the long-term hydrological effects of vegetated swales, comprising the following steps:

[0009] Step 10: Use the data availability assessment model to assess and screen future rainfall events in the area where the vegetated swale to be assessed is located, and obtain representative rainfall events;

[0010] Step 20: When a representative rainfall event occurs, collect meteorological parameters and hydrological performance parameters of the representative rainfall event; calculate the instantaneous saturated permeability coefficient value of the vegetated swamp under the representative rainfall event using the instantaneous saturated permeability coefficient inversion method, and form a second training dataset with the meteorological parameter values ​​of the representative rainfall event.

[0011] Step 30: Based on the second training dataset, establish an instantaneous dynamic prediction model for saturated permeability coefficient using the XGBoost algorithm;

[0012] Step 40: Use the instantaneous saturated permeability coefficient dynamic prediction model to predict the instantaneous saturated permeability coefficient of the vegetated swamp under new rainfall events other than representative rainfall events, and obtain the predicted value of the instantaneous saturated permeability coefficient of the vegetated swamp.

[0013] Step 50: Input the predicted value of the instantaneous saturated permeability coefficient of the grassed swale into the grassed swale hydrological model to obtain the simulation results of the hydrological performance parameters of the grassed swale to be evaluated.

[0014] As a further improvement of the present invention, the meteorological parameters include the number of sunny days before rainfall, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, and seasonal factors; the hydrological performance parameters include the water depth in the grassed swale and the outflow rate of the grassed swale.

[0015] As a further improvement of the present invention, the instantaneous saturated permeability coefficient dynamic prediction model is a dynamic prediction model of the instantaneous saturated permeability coefficient of the vegetated swale with respect to the number of sunny days in the previous period, temperature, humidity, rainfall, rainfall duration, cumulative rainfall and seasonal factors.

[0016] As a further improvement of the present invention, the construction process of the data collectability assessment model includes:

[0017] The instantaneous saturated permeability coefficient of the vegetated swamp was calculated using the inversion method of instantaneous saturated permeability coefficient of the vegetated swamp area under at least 100 historical rainfall events over at least 5 years. The instantaneous saturated permeability coefficient of the vegetated swamp area was then calculated and combined with the meteorological parameter values ​​of the historical rainfall events to form the first training data subset.

[0018] The first training data subset is resampled to randomly generate N*M short data sequences containing N rainfall event groups in M ​​classes; the average value of the meteorological parameters corresponding to each short data sequence is calculated, and the average values ​​of the meteorological parameters calculated from the N*M data sequences are mapped to three levels: low, medium, and high; where N is a multiple of 1000, and M is an integer not greater than the total number of historical rainfall events;

[0019] Based on N*M sets of short data sequences, the XGBoost algorithm is used to establish dynamic prediction models for instantaneous saturation permeability coefficients, and the relative error between the model and the model established based on the first training data subset is calculated. If the relative error is less than 10%, the set of short data sequences is considered representative; otherwise, the set of short data sequences is considered unrepresentative.

[0020] Based on the rank mapping results and the classification results of whether the N*M short data sequences are representative, a data acceptability assessment model is established using the Bernoulli Naive Bayes method.

[0021] As a further improvement of the present invention, step 10 specifically includes:

[0022] Calculate the average meteorological parameters of historical rainfall events in the area where the grassed swale is to be evaluated, map the calculated average meteorological parameters into three categories: low, medium and high, and update the mapping results into the data collectability assessment model.

[0023] Using a data availability assessment model and combining meteorological parameter values ​​under rainfall events from weather forecasts, the probability that future rainfall events in the area of ​​the grassy swamp to be assessed are representative is calculated one by one.

[0024] The first rainfall event of each month is considered representative; if it is an extreme rainfall event, the next one is postponed. If the probability of a rainfall event being representative increases after adding meteorological parameter values ​​for another rainfall event, then that rainfall event is considered representative; otherwise, it is not.

[0025] As a further improvement of the present invention, the construction method of the grassed ditch hydrological model is as follows: based on equations (1) and (2), considering the time-varying characteristics of the saturated permeability coefficient, the assumption of the consistency of the vertical change of soil moisture content in the ditch is proposed and the static pressure head parameter is introduced to establish the grassed ditch hydrological model.

[0026] Equation (1)

[0027] Equation (2)

[0028] In the formula, f Indicates the infiltration rate; K s ( x , t () represents the saturated permeability coefficient. x Indicate influencing factors, t Indicates time; Z f Indicates the length of the penetration path; ΔPThis represents the pressure difference across the permeation unit; K ( θ () represents the unsaturated permeability coefficient; K s Indicates the saturated permeability coefficient; θ Indicates the volumetric water content of the soil; θ r This indicates the remaining volumetric water content of the soil. θ s α represents the saturated volumetric water content of the soil; α represents the empirical fitting parameter or curve shape parameter. n This represents the empirical fitting parameters or curve shape parameters.

[0029] As a further improvement of the present invention, the calculation process of the wetting stage in the vegetated swale hydrological model includes:

[0030] As rainfall continues, surface runoff flows in continuously. The moisture content of the filler layer rises steadily under the combined action of top inflow and bottom drainage until it reaches saturation. The water effluent from the bottom of the filler layer serves as the inflow for the drainage layer. Specifically:

[0031] Equation (3)

[0032] In the formula, This represents the volume of surface water in the vegetated swale at the end of a unit of time when there is no infiltration at the current moment; This represents the volume of surface water in the grassy swale before the current moment began; Indicates rainfall intensity; Indicates a unit time interval;

[0033] Equation (4)

[0034] In the formula, This indicates the depth of water accumulation in the vegetated swale at the end of a unit of time when there is no infiltration at the current moment; This represents the surface area of ​​the filler layer when the water storage layer is rectangular;

[0035] Equation (5)

[0036] In the formula, This represents the infiltration rate of the filler layer at the end of a unit time when there is no infiltration. Indicates the saturated permeability coefficient of the packing layer; Indicates the thickness of the filler layer;

[0037] Equation (6)

[0038] In the formula, The infiltration rate represents the volume of water accumulated in the vegetated swale after infiltration per unit time, assuming no infiltration. This represents the overflow volume of the grassed swale per unit time.

[0039] make By setting the minimum error range and performing iterative calculations, the volume and depth of water accumulation in the grassy swale can be obtained.

[0040] Equation (7)

[0041] Equation (8)

[0042] In the formula, This indicates the infiltration rate of the filler layer at the end of the current time when the bottom is no longer drained; This indicates the amount of infiltration of the filler layer relative to the remaining moisture content before the current moment begins; This indicates the moisture content of the filler layer at the end of the current time period when the bottom is no longer drained.

[0043] Equation (9)

[0044] In the formula, This represents the unsaturated permeability coefficient of the packing layer when the bottom is no longer drained at the current moment; This indicates the remaining volumetric water content of the packing layer; This indicates the saturated volumetric water content of the packing layer; This represents the saturated permeability coefficient of the packing layer;

[0045] Equation (10)

[0046] In the formula, This indicates the water outflow rate at the bottom of the packing layer at the current moment;

[0047] Will As the inflow rate of the drainage layer, the same iterative method as that used for the filler layer was adopted to obtain the water volume and depth of the drainage layer surface at each moment;

[0048] Equation (11)

[0049] Equation (12)

[0050] In the formula, This represents the infiltration rate of the packing layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This represents the water content of the packing layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This represents the infiltration rate of the drainage layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. Indicates the surface area of ​​the drainage layer;

[0051] make Set the minimum error range and perform iterative calculations to obtain the infiltration rate and water content of the filler layer at the current moment;

[0052] The same iterative method as that used for the filler layer was employed to calculate the infiltration rate and moisture content of the drainage layer.

[0053] As a further improvement of the present invention, the calculation process for the drainage stage in the vegetated swale hydrological model includes:

[0054] When the rainfall process has not yet ended, but the drainage layer is already full, the vegetated swale begins to drain water outward through the perforated drainage pipe. Assuming that the drainage layer has sufficient permeability to drain the water from the filler layer in time, the outflow rate of the drainage pipe per unit time is equal to the inflow rate at the bottom of the filler layer minus the infiltration rate into the natural soil at the bottom of the vegetated swale. The average infiltration rate from the bottom to the natural soil is considered to be approximately equal to the saturated permeability coefficient of the natural soil. The outflow rate is calculated using equation (13):

[0055] Equation (13)

[0056] In the formula, This indicates the average outflow rate at the outlet of the drainage pipe in the grassy ditch. This represents the saturated permeability coefficient of the natural soil at the bottom of the vegetated swale; Indicates the surface area of ​​the drainage layer;

[0057] Calculate the pipe velocity along the friction and local head loss of the drainage blind pipe using equation (14):

[0058] Equation (14)

[0059] Equation (15)

[0060] In the formula, The flow velocity along the pipe and at local head loss in the drainage blind pipe; H0 represents the total head including the traveling head; This represents the correction factor, which is typically set to 1.0. This indicates head loss, including head loss along the friction path and local head loss; Indicates the friction coefficient; Indicates the length of the pipe; Indicates the pipe diameter; Indicates the cross-sectional length of the water flow in the pipe; Indicates the length of the pipe; Indicates the local drag coefficient; Indicates the flow velocity inside the pipe;

[0061] After the rainfall ends, surface runoff also stops, and the vegetated swales enter the drainage stage. Initially, there is still water on the surface of the vegetated swales, and the infiltration meets the saturation infiltration requirement. The infiltration rate of the filler layer is calculated using equation (7). After the surface water disappears, the filler layer begins to gradually drain in an unsaturated state. The drainage rate of the filler layer is calculated using equation (16).

[0062] Equation (16)

[0063] In the formula, This indicates the evacuation rate of the packing layer. This represents the unsaturated permeability coefficient of the packing layer. This indicates the current moisture content of the filler layer;

[0064] At this time, the drainage layer is still saturated. When the current moisture content of the drainage layer is less than the saturation moisture content of the drainage layer, the outlet stops discharging water, and the water stored in the drainage layer slowly seeps into the underground soil. When the moisture content of the filler layer and the drainage layer drops to their respective field moisture content, the infiltration process of this rainfall ends.

[0065] As a further improvement of the present invention, in step 20, the instantaneous saturated permeability coefficient value of the vegetated swale for a representative rainfall event is calculated using the instantaneous saturated permeability coefficient inversion method, specifically including:

[0066] Using the water depth or outflow rate of vegetated swales under representative rainfall events as observed variables, the initial hypothesized instantaneous saturated permeability coefficient of the vegetated swales is substituted into the vegetated swales hydrological model to obtain the simulation results of the observed variables. The ensemble Kalman filter algorithm is used to assimilate the observed and simulated results to generate an estimated value of the instantaneous saturated permeability coefficient, which is then substituted into the vegetated swales hydrological model for simulation calculation. After multiple iterations, the converged instantaneous saturated permeability coefficient value of the vegetated swales is obtained.

[0067] As a further improvement of the present invention, step 20 employs an ensemble Kalman filter algorithm for data assimilation, specifically including:

[0068] Based on the state transition equation shown in equation (17) and the observation equation shown in equation (18):

[0069] Equation (17)

[0070] Equation (18)

[0071] In the formula, This represents the predicted value of the k-th set number of the parameters of the grassy swale hydrological model at time i+1; This represents the updated value of the k-th set number of the parameters of the grassy ditch hydrological model at time i; This represents the predicted value of the k-th set number of the state variables at time i+1; This represents the updated value of the k-th set number of the state variables at time i; Represents the prediction operator; This represents the data driven by the hydrological model of the vegetated swale; This represents independent white noise representing the parameters of the grassy swale hydrological model. This represents independent white noise representing the state variables of the grassy swale hydrological model. This represents the kth set value of the simulated outflow or water accumulation depth of the grassy ditch at time i+1. h Represents the observation operator; The error term is represented by a mean of 0 and a variance of 0. The normal distribution;

[0072] Assimilation is performed using equations (19) to (21):

[0073] Equation (19)

[0074] Equation (20)

[0075] Equation (21)

[0076] In the formula, This represents the predicted value of the k-th set number at time i+1; This represents the updated value of the k-th set at time i; Represents the white noise of the k-th set; This represents the observation value of the k-th set; This represents the observation error of the k-th set; The Kalman gain represents the weighted relationship between the predicted and observed values, calculated using equations (22) to (24):

[0077] Equation (22)

[0078] Equation (23)

[0079] Equation (24)

[0080] In the formula, Represents the covariance matrix of the predicted state variables; This represents the covariance matrix of the predicted values ​​of the observed variables; This represents the set mean of the predicted state variables; N represents the set mean of the predicted values ​​of the observed variables; N represents the set number.

[0081] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0082] (1) The present invention provides a method for monitoring and evaluating the long-term hydrological effects of grassed swales. It uses a data availability assessment model to sample data, improves the predictability of sampling, and realizes lightweight monitoring from more to less. It uses the grassed swale instantaneous saturated permeability coefficient inversion method to invert the instantaneous saturated permeability coefficient, realizing lightweight monitoring from difficult to easy. It substitutes the monitored meteorological data and the inverted value of the instantaneous saturated permeability coefficient into the instantaneous saturated permeability coefficient dynamic prediction model to calibrate the model variable parameters, and uses the calibrated model to predict the instantaneous saturated permeability coefficient under future rainfall events, realizing simulation from static to dynamic. It substitutes the predicted value of the instantaneous saturated permeability coefficient under future rainfall events into the grassed swale hydrological model to simulate the hydrological performance, realizing the long-term hydrological effect evaluation of grassed swales.

[0083] (2) The present invention provides a method for monitoring and evaluating the long-term hydrological effects of grassed swales. It establishes a data availability assessment model based on Bayesian networks and uses the data availability assessment model to select representative measured data sets for training an instantaneous saturated permeability coefficient prediction model. This achieves predictability and lightweight data sampling, effectively reducing the workload of actual measurements while ensuring prediction accuracy, thereby improving the practicality of the method for evaluating the long-term hydrological effects of grassed swales.

[0084] (3) The present invention provides a method for monitoring and evaluating the long-term hydrological effects of grassed swales. It proposes the concept of the instantaneous saturated permeability coefficient of grassed swales under rainfall events, and establishes an instantaneous saturated permeability coefficient inversion method that couples the grassed swale hydrological model with the ensemble Kalman filter algorithm, using the relatively easy-to-measure water depth or outflow in grassed swales under rainfall events as the observation variables. The instantaneous saturated permeability coefficient inversion method is used to calculate the instantaneous saturated permeability coefficient value of grassed swales under representative rainfall events, which solves the problems that the saturated permeability coefficient of grassed swales varies greatly along the path under rainfall events, that the measured value of a single local point cannot represent the overall performance, and that the error of simulation calculation using the simple average value of measured values ​​from multiple points is large. This improves the accuracy of the instantaneous saturated permeability coefficient dynamic prediction model obtained by training the instantaneous saturated permeability coefficient of grassed swales.

[0085] (4) The method for monitoring and evaluating the long-term hydrological effects of planted swales provided by the present invention uses the inversion result of the instantaneous saturated permeability coefficient as the dependent variable and the measured data of meteorological parameters as the independent variable. The XGBoost algorithm is used to establish a dynamic prediction model of the instantaneous saturated permeability coefficient. Compared with the traditional machine learning model, the required amount of training samples is smaller while ensuring the performance of the model, which is suitable for situations where the sample size of the present invention is not large. Secondly, the XGBoost model has stronger interpretability. The decision-making process of the model can be understood through feature importance and tree structure visualization. In the method of the present invention, the importance of each feature can be given intuitively, and some features with weak importance can be deleted as appropriate without affecting the overall predictive ability of the model.

[0086] (5) The method for monitoring and evaluating the long-term hydrological effects of grassed gullies provided by the present invention is based on the Darcy formula and the van Genuchten-Mualem soil-water characteristic curve equation. Considering the time-varying characteristics of the saturated permeability coefficient, a grassed gully hydrological model is established. The static pressure head parameter is introduced in the coupling process of the Darcy formula and the soil-water characteristic curve equation, so that the grassed gully hydrological model can simulate the infiltration when there is water accumulation in the grassed gully under rainfall events. The assumption that the soil moisture content of the gully changes in the vertical direction is proposed, so that the grassed gully hydrological model can simulate the infiltration when the soil of the gully is unsaturated in the initial stage of rainfall, reducing the calculation difficulty of the position of the wetting front under unsaturated state. Attached Figure Description

[0087] Figure 1 A flowchart of the long-term hydrological effect monitoring and evaluation method for vegetated swales provided by the present invention;

[0088] Figure 2 This is a schematic diagram of the data acceptability assessment model in the method of the present invention;

[0089] Figure 3 This is a flowchart of the instantaneous saturated permeability coefficient inversion method of the grassed swale in this invention;

[0090] Figure 4 This is a flowchart illustrating the calculation process of the vegetated ditch hydrological model in the method of this invention. Detailed Implementation

[0091] The technical solution of the present invention will be described in detail below.

[0092] This invention provides a method for monitoring and evaluating the long-term hydrological effects of vegetated swales, such as... Figure 1 As shown, it includes the following steps:

[0093] Step 10: Use the data availability assessment model to assess and screen future rainfall events in the area where the grassed swamp is located to obtain representative rainfall events.

[0094] Step 20: When a representative rainfall event occurs, collect meteorological and hydrological parameters for that event. Calculate the instantaneous saturated permeability coefficient of the vegetated swales under the representative rainfall event using the instantaneous saturated permeability coefficient inversion method. Combine this with the meteorological parameter values ​​for the representative rainfall event to form a second training dataset. Preferably, the meteorological parameters include the number of sunny days prior to rainfall, temperature, humidity, rainfall amount, rainfall duration, cumulative rainfall, and seasonal factors. The hydrological parameters include the water depth within the vegetated swales and the outflow rate.

[0095] Step 30: Based on the second training dataset, use the XGBoost algorithm to establish an instantaneous dynamic prediction model for the saturated permeability coefficient.

[0096] Step 40: Use the instantaneous saturated permeability coefficient dynamic prediction model to predict the instantaneous saturated permeability coefficient of the vegetated swamp under new rainfall events other than representative rainfall events, and obtain the predicted value of the instantaneous saturated permeability coefficient of the vegetated swamp.

[0097] Step 50: Input the predicted value of the instantaneous saturated permeability coefficient of the grassed swale into the grassed swale hydrological model to obtain the simulation results of the hydrological performance parameters of the grassed swale to be evaluated.

[0098] Preferably, the method for constructing the vegetated swale hydrological model in step 50 is as follows:

[0099] Based on the Darcy formula shown in Equation (1) and the van Genuchten-Mualem soil-water characteristic curve equation shown in Equation (2), considering the time-varying characteristics of the saturated permeability coefficient, the assumption of consistent vertical variation of soil moisture content in the ditch is proposed and the static pressure head parameter is introduced to establish a hydrological model for vegetated ditch.

[0100] Equation (1)

[0101] Equation (2)

[0102] In the formula, f Indicates the infiltration rate; K s ( x , t () represents the saturated permeability coefficient. x Indicate influencing factors, t Indicates time; Z f Indicates the length of the penetration path; ΔP This represents the pressure difference across the permeation unit; K ( θ () represents the unsaturated permeability coefficient; Ks Indicates the saturated permeability coefficient; θ Indicates the volumetric water content of the soil; θ r This indicates the remaining volumetric water content of the soil. θ s α represents the saturated volumetric water content of the soil; α represents the empirical fitting parameter or curve shape parameter. n This represents the empirical fitting parameters or curve shape parameters.

[0103] Constructing a hydrological model of a vegetated swale mainly includes wetting and drainage processes, such as... Figure 4 As shown.

[0104] The calculation process for the wetting process includes:

[0105] As rainfall continues, surface runoff flows in continuously. The moisture content of the filler layer rises continuously under the combined effect of top water inflow and bottom drainage until it reaches saturation. The water effluent at the bottom of the filler layer serves as the inflow for the drainage layer.

[0106] Specifically:

[0107] Equation (3)

[0108] In the formula, This represents the volume of surface water in the vegetated swale at the end of a unit of time when there is no infiltration at the current moment; This represents the volume of surface water in the grassy swale before the current moment began; Indicates rainfall intensity; Indicates a unit time interval.

[0109] Equation (4)

[0110] In the formula, This indicates the depth of water accumulation in the vegetated swale at the end of a unit of time when there is no infiltration at the current moment; This represents the surface area of ​​the filler layer when the water storage layer is rectangular.

[0111] Equation (5)

[0112] In the formula, This represents the infiltration rate of the filler layer at the end of a unit time when there is no infiltration. Indicates the saturated permeability coefficient of the packing layer; This indicates the thickness of the filler layer.

[0113] Equation (6)

[0114] In the formula, The infiltration rate represents the volume of water accumulated in the vegetated swale after infiltration per unit time, assuming no infiltration. This indicates the overflow volume of the grassed ditch per unit time.

[0115] make By setting a minimum error range and performing iterative calculations, the volume and depth of water accumulation in the grassed ditch can be obtained.

[0116] Equation (7)

[0117] Equation (8)

[0118] In the formula, This indicates the infiltration rate of the filler layer at the end of the current time when the bottom is no longer drained; This indicates the amount of infiltration of the filler layer relative to the remaining moisture content before the current moment begins; This indicates the moisture content of the filler layer at the end of the current time when the bottom is no longer drained.

[0119] Equation (9)

[0120] In the formula, This represents the unsaturated permeability coefficient of the packing layer when the bottom is no longer drained at the current moment; This indicates the remaining volumetric water content of the packing layer; This indicates the saturated volumetric water content of the packing layer; This represents the saturated permeability coefficient of the packing layer.

[0121] Equation (10)

[0122] In the formula, This indicates the water outflow rate at the bottom of the packing layer at the current moment.

[0123] Will As the inflow rate of the drainage layer, the same iterative method as that used for the filler layer was employed to obtain the water volume and depth at each moment on the surface of the drainage layer.

[0124] Equation (11)

[0125] Equation (12)

[0126] In the formula, This represents the infiltration rate of the packing layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This represents the water content of the packing layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This represents the infiltration rate of the drainage layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This indicates the surface area of ​​the drainage layer.

[0127] make By setting a minimum error range and performing iterative calculations, the infiltration rate and water content of the filler layer at the current moment can be obtained.

[0128] Using the same iterative method as the filler layer, the infiltration rate and water content of the drainage layer were calculated.

[0129] The calculation process for the drainage process includes:

[0130] When the rainfall process has not yet ended, but the drainage layer is already full, the vegetated swale begins to drain water outward through the perforated drainage pipe. Assuming that the drainage layer has sufficient permeability to drain the water from the filler layer in time, the outflow rate of the drainage pipe per unit time is equal to the inflow rate at the bottom of the filler layer minus the infiltration rate into the natural soil at the bottom of the vegetated swale. The average infiltration rate from the bottom to the natural soil is considered to be approximately equal to the saturated permeability coefficient of the natural soil. The outflow rate is calculated using equation (13):

[0131] Equation (13)

[0132] In the formula, This indicates the average outflow rate at the outlet of the drainage pipe in the grassy ditch. This represents the saturated permeability coefficient of the natural soil at the bottom of the vegetated swale; Indicates the surface area of ​​the drainage layer;

[0133] Calculate the pipe velocity along the friction and local head loss of the drainage blind pipe using equation (14):

[0134] Equation (14)

[0135] Equation (15)

[0136] In the formula, The flow velocity along the pipe and at local head loss in the drainage blind pipe; H0 represents the total head including the traveling head; This represents the correction factor, which is typically set to 1.0. This indicates head loss, including head loss along the friction path and local head loss; Indicates the friction coefficient; Indicates the length of the pipe; Indicates the pipe diameter; Indicates the cross-sectional length of the water flow in the pipe; Indicates the length of the pipe; Indicates the local drag coefficient; This indicates the flow velocity inside the pipe.

[0137] After the rainfall ends, surface runoff also stops, and the vegetated swales enter the drainage stage. Initially, there is still water on the surface of the vegetated swales, and the infiltration meets the saturation infiltration requirement. The infiltration rate of the filler layer is calculated using equation (7). After the surface water disappears, the filler layer begins to gradually drain in an unsaturated state. The drainage rate of the filler layer is calculated using equation (16).

[0138] Equation (16)

[0139] In the formula, This indicates the evacuation rate of the packing layer. This represents the unsaturated permeability coefficient of the packing layer. This indicates the current moisture content of the filler layer.

[0140] At this time, the drainage layer is still saturated. When the current moisture content of the drainage layer is less than the saturation moisture content of the drainage layer, the outlet stops discharging water, and the water stored in the drainage layer slowly seeps into the underground soil. When the moisture content of the filler layer and the drainage layer drops to their respective field moisture content, the infiltration process of this rainfall ends.

[0141] The saturated permeability coefficient of the filler layer in a vegetated ditch is a key factor affecting the hydrological effects of the ditch and is also the most important input parameter in vegetated ditch hydrological models. However, vegetated ditches are generally long, and their saturated permeability coefficient varies along the ditches. Therefore, the measured saturated permeability coefficient value at a single point cannot reflect the overall infiltration performance of the vegetated ditch. The commonly used method of averaging measurements from multiple points to assess the overall infiltration performance of the vegetated ditch also has a large error. Therefore, the results obtained by simulating the hydrological effects using the measured value of the saturated permeability coefficient are unreliable. Furthermore, existing vegetated ditch hydrological models typically assume that the saturated permeability coefficient is constant and does not change over time, or assign a constant decay coefficient to characterize the long-term decreasing characteristic of the saturated permeability coefficient. However, in changing environments, the saturated permeability coefficient of a vegetated ditch exhibits a dynamic changing trend under the influence of factors such as rainfall, temperature, humidity, and plant growth.

[0142] Preferably, in this embodiment of the invention, based on the variation and time-varying characteristics of the saturated permeability coefficient along the course of the vegetated swale, the saturated permeability coefficient value that can characterize the overall infiltration performance, which is derived from the measured hydrological performance data of the vegetated swale under a given rainfall event, is taken as the instantaneous saturated permeability coefficient under that rainfall event.

[0143] In step 20, the instantaneous saturated permeability coefficient values ​​of the vegetated swales under representative rainfall events are calculated using the instantaneous saturated permeability coefficient inversion method. Specifically:

[0144] Using the water depth or outflow rate of vegetated swales under representative rainfall events as observed variables, the initially assumed instantaneous saturated permeability coefficient of the vegetated swales is substituted into the vegetated swale hydrological model to obtain simulation results for the observed variables. An ensemble Kalman filter algorithm is used to assimilate the observed and simulated results to generate an estimated instantaneous saturated permeability coefficient, which is then substituted into the vegetated swale hydrological model for simulation calculation. After multiple iterations, a converged instantaneous saturated permeability coefficient value is obtained, which is the equivalent value of the overall infiltration performance of the vegetated swales under this rainfall event.

[0145] Among them, the ensemble Kalman filter algorithm is used for data assimilation, such as Figure 3 As shown, it specifically includes:

[0146] Construct the state transition equation shown in equation (17) and the observation equation shown in equation (18):

[0147] Equation (17)

[0148] Equation (18)

[0149] In the formula, This represents the predicted value of the k-th set number of the parameters of the grassy swale hydrological model at time i+1; This represents the updated value of the k-th set number of the parameters of the grassy ditch hydrological model at time i; This represents the predicted value of the k-th set number of the state variables at time i+1; This represents the updated value of the k-th set number of the state variables at time i; This refers to the prediction operator, specifically the grassy swale hydrological model. This refers to the data driven by the vegetated swamp hydrological model, which mainly refers to rainfall data. The parameters of the vegetated swale hydrological model are independent white noise, all conforming to a mean of 0 and a specific variance. The normal distribution; The state variables of the vegetated swamp hydrological model are independent white noise, all following a mean of 0 and a specific variance. The normal distribution; This represents the kth set value of the simulated outflow or water accumulation depth of the grassy ditch at time i+1. h This represents the observation operator, i.e., the transformation relationship between state variables and observation variables; The error term is represented by a mean of 0 and a variance of 0. It follows a normal distribution.

[0150] Assimilation is performed using equations (19) to (21):

[0151] Equation (19)

[0152] Equation (20)

[0153] Equation (21)

[0154] In the formula, This represents the predicted value of the k-th set number at time i+1; This represents the updated value of the k-th set at time i; Represents the white noise of the k-th set; This represents the observation value of the k-th set; This represents the observation error of the k-th set; The Kalman gain represents the weighted relationship between the predicted and observed values, calculated using equations (22) to (24):

[0155] Equation (22)

[0156] Equation (23)

[0157] Equation (24)

[0158] In the formula, Represents the covariance matrix of the predicted state variables; This represents the covariance matrix of the predicted values ​​of the observed variables; This represents the set mean of the predicted state variables; N represents the set mean of the predicted values ​​of the observed variables; N represents the set number; T represents the matrix transpose.

[0159] Preferably, step 10 specifically includes:

[0160] The average values ​​of historical rainfall parameters (number of sunny days, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, and seasonal factors) for the area of ​​the vegetated swale to be evaluated are calculated. These average values ​​are then mapped into three categories: low, medium, and high. The indicators are categorized using these three categories. Cluster analysis is performed on the data sequence characteristics based on the values ​​of each indicator. The category range is calculated by considering the amplitude of the indicator observed in all data sequences (maximum value minus minimum value). The overall minimum value is the lower limit of the "low" range, and the overall maximum value is the upper limit of the "high" range. Dividing the indicator amplitude by three and adding the minimum value gives the upper limit of the "low" range, while subtracting the maximum value gives the lower limit of the "high" range. The range from the upper limit of the "low" range to the lower limit of the "high" range constitutes the "medium" range.

[0161] The mapping results are then updated in the data availability assessment model. The data availability assessment model is based on a Bayesian network structure, such as... Figure 2As shown, the input nodes represent the levels corresponding to the feature values ​​of the data sequence, including the number of event events, average number of sunny days in the preceding period, average temperature, average humidity, average rainfall, average rainfall duration, average cumulative rainfall, and seasonality index. The output nodes represent the probability that the data sequence is representative. Each node is discretized, and conditional probabilities are calculated. The category of node 1 is defined as the number of rainfall events. Nodes 2-8 are similarly divided into three categories: "low," "medium," and "high." The category of node 9 is "yes" or "no."

[0162] Using a data availability assessment model and combining meteorological parameter values ​​under rainfall events from weather forecasts, the probability that future rainfall events in the area of ​​the grassed swamp to be assessed are representative is calculated one by one.

[0163] The first rainfall event of each month is considered representative; if it is an extreme rainfall event, the next one is postponed. If the probability of a rainfall event being representative increases after adding meteorological parameter values ​​for another rainfall event, then that rainfall event is considered representative; otherwise, it is not.

[0164] The process of constructing the data availability assessment model includes:

[0165] The instantaneous saturated permeability coefficient of the vegetated swamp was calculated using the inversion method for the instantaneous saturated permeability coefficient of the vegetated swamp in the area where the modeled vegetated swamp is located, based on at least 100 historical rainfall events over at least 5 years. This value, along with the meteorological parameters of all historical rainfall events (number of sunny days before rainfall, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, and seasonal factors), constitutes the first training data subset.

[0166] The first training data subset is resampled to randomly generate N*M short data sequences containing N (multiples of 1000) rainfall event groups across M (integers not greater than the total number of historical rainfall events). The average value of the meteorological parameters corresponding to each short data sequence is calculated, and the average values ​​of the meteorological parameters calculated from the N*M data sequences are mapped to three levels: low, medium, and high.

[0167] Based on N*M sets of short data sequences, the XGBoost algorithm is used to establish dynamic prediction models for instantaneous saturated permeability coefficients. The relative error between the model and the model established based on the first training data subset is calculated. If the relative error is less than 10%, the set of short data sequences is considered representative; otherwise, the set of short data sequences is considered unrepresentative.

[0168] Based on the rank mapping results and the classification results of whether the N*M short data sequences are representative, an admissibility assessment model is established using the Bernoulli Naive Bayes method.

[0169] The main calculation process of the Bernoulli Naive Bayes method is as follows:

[0170] Data preprocessing: The features (average number of sunny days in the preceding period, average temperature, average humidity, average rainfall, average rainfall duration, average cumulative rainfall, and average seasonal factor) in the dataset (N*M groups) are converted into binary features (0 or 1), where 0 indicates that the feature value is not in the high, medium, or low value range, and 1 indicates that the feature value is in the high, medium, or low value range. The dataset is then divided into training and test sets.

[0171] Calculate the prior probability: for each class C k Calculate the prior probability P(C k ):

[0172]

[0173] In the formula, k This indicates a category index.

[0174] Calculate conditional probability: for each feature x i and each category C k Calculate the conditional probability P( x i | C k ):

[0175]

[0176]

[0177] In the formula, α represents the smoothing parameter, which is used to prevent the probability from being 0.

[0178] Calculate the prior probability:

[0179]

[0180] The category with the highest posterior probability is selected as the prediction category. For example, given a rainfall time group, if the probability of predicting representativeness is 0.8 and the probability of predicting non-representativeness is 0.2, then the prediction category is representativeness.

[0181] Preferably, in step 30, the XGBoost algorithm constructs the model as follows:

[0182] Data preparation: Convert the training data into a format suitable for XGBoost input. For example, an Excel spreadsheet or a TXT text file.

[0183] Data partitioning: The training data is divided into training and test sets, usually using a random partitioning method, such as partitioning it into a 70% training set and a 30% test set.

[0184] Model initialization: Select a number-based model (gbtree), set the objective function to minimize the squared error between the predicted and true values, and initialize the model parameters, including the learning rate, the maximum depth of the tree, and the L2 regularization parameter.

[0185] Model Training: The model is trained using the training set. During training, the model's parameters are gradually adjusted based on the objective function and optimization algorithm to minimize training error. Appropriate evaluation metrics (such as root mean square error, log loss, etc.) are selected by setting the `eval_metric` parameter, and the model's performance on the validation set is monitored during training to prevent overfitting. An early stopping mechanism is enabled during training; training is terminated prematurely if the performance on the validation set fails to improve for several consecutive rounds.

[0186] Parameter optimization: Model parameters are optimized using methods such as grid search, random search, or Bayesian optimization to improve model performance. Feature selection is performed by analyzing feature importance or using regularization methods to reduce model complexity and improve the model's generalization ability.

[0187] The following is a specific example.

[0188] A grassy swale in Utrecht, Netherlands, that has been in operation for more than 15 years, was selected as the application subject.

[0189] The data availability assessment model was used to screen 35 rainfall events in the Utrecht region of the Netherlands between 2020 and 2021. Seven meteorological data points were measured, including the number of sunny days before rainfall, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, and seasonal factors, as well as the outflow data of vegetated swales. The data sequence group composed of the measured data was determined by the data availability assessment model to be a representative data sequence with a representativeness probability of 92.4%.

[0190] Using the outflow from vegetated swales under 35 rainfall events as the observed variable, the instantaneous saturated permeability coefficient value for each rainfall event was calculated using the vegetated swales instantaneous saturated permeability coefficient inversion method. The training data consisted of the outflow from vegetated swales under the 35 rainfall events, the number of sunny days before rainfall, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, seasonal factors, and instantaneous saturated permeability coefficient.

[0191] Based on training data, an instantaneous dynamic prediction model for saturated permeability coefficient is established using the XGBoost algorithm.

[0192] The instantaneous saturated permeability coefficient was predicted for 12 measured rainfall events in 2023 using an instantaneous saturated permeability coefficient dynamic prediction model. The predicted values ​​of the instantaneous saturated permeability coefficient of the vegetated swales under the 12 rainfall events were obtained.

[0193] The predicted values ​​of the instantaneous saturated permeability coefficient of the grassed swales under 12 rainfall events were substituted into the grassed swale hydrological model to simulate the outflow of the grassed swales under the 12 rainfall events.

[0194] The results show that the Nash efficiency coefficient (NSE) of the simulated outflow from the vegetated swales, compared with the measured values, ranges from 0.71 to 0.95, with an average of 0.83; the relative error (RE) ranges from 2.43% to 13.28%, with an average of 7.13%. It can be seen that the method of this invention has a good overall effect on the long-term simulation of the outflow from the vegetated swales and is feasible in practical applications.

[0195] The method of this invention provides a long-term hydrological effect monitoring and evaluation method for vegetated swales by combining long-term data monitoring with data assimilation and machine learning algorithms. It is applicable to vegetated swales in any region and any scenario, and improves the effectiveness of hydrological monitoring of vegetated swales and the accuracy of long-term hydrological model simulation.

[0196] 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 specific embodiments described above. The specific embodiments and descriptions in the specification are merely for further illustrating the principles of the invention. The basic principles, main features, and advantages of the present invention have been shown and described above without departing from the spirit and scope of the invention. Those skilled in the art should understand that various changes and modifications will be made, and all such changes and modifications fall within the scope of the present invention as claimed.

Claims

1. A method for long-term hydrological effect monitoring and evaluation of planted swales, characterized in that, Includes the following steps: Step 10: Use the data availability assessment model to assess and screen future rainfall events in the area where the vegetated swale to be assessed is located, and obtain representative rainfall events; Step 20: When a representative rainfall event occurs, collect meteorological parameters and hydrological performance parameters of the representative rainfall event; calculate the instantaneous saturated permeability coefficient value of the vegetated swamp under the representative rainfall event using the instantaneous saturated permeability coefficient inversion method, and form a second training dataset with the meteorological parameter values ​​of the representative rainfall event. Step 30: Based on the second training dataset, establish an instantaneous dynamic prediction model for saturated permeability coefficient using the XGBoost algorithm; Step 40: Use the instantaneous saturated permeability coefficient dynamic prediction model to predict the instantaneous saturated permeability coefficient of the vegetated swamp under new rainfall events other than representative rainfall events, and obtain the predicted value of the instantaneous saturated permeability coefficient of the vegetated swamp. Step 50: Input the predicted value of the instantaneous saturated permeability coefficient of the grassed swale into the grassed swale hydrological model to obtain the simulation results of the hydrological performance parameters of the grassed swale to be evaluated. The process of constructing the data collectability assessment model includes: The instantaneous saturated permeability coefficient of the vegetated swamp was calculated using the inversion method of instantaneous saturated permeability coefficient of the vegetated swamp area under at least 100 historical rainfall events over at least 5 years. The instantaneous saturated permeability coefficient of the vegetated swamp area was then calculated and combined with the meteorological parameter values ​​of the historical rainfall events to form the first training data subset. The first training data subset is resampled to randomly generate N*M short data sequences containing N rainfall event groups in M ​​classes; the average value of the meteorological parameters corresponding to each short data sequence is calculated, and the average values ​​of the meteorological parameters calculated from the N*M data sequences are mapped to three levels: low, medium, and high; where N is a multiple of 1000, and M is an integer not greater than the total number of historical rainfall events; Based on N*M sets of short data sequences, the XGBoost algorithm is used to establish dynamic prediction models for instantaneous saturation permeability coefficients, and the relative error between the model and the model established based on the first training data subset is calculated. If the relative error is less than 10%, the set of short data sequences is considered representative; otherwise, the set of short data sequences is considered unrepresentative. Based on the rank mapping results and the classification results of whether the N*M short data sequences are representative, a data acceptability assessment model is established using the Bernoulli Naive Bayes method. The construction method of the grassed ditch hydrological model is as follows: Based on Equations (1) and (2), considering the time-varying characteristics of the saturated permeability coefficient, the assumption of consistent vertical variation of soil moisture content in the ditch is proposed and the static pressure head parameter is introduced to establish the grassed ditch hydrological model. Equation (1) Equation (2) In the formula, f Indicates the infiltration rate; K s ( x , t () represents the saturated permeability coefficient. x Indicate influencing factors, t Indicates time; Z f Indicates the length of the penetration path; ΔP This represents the pressure difference across the permeation unit; K ( θ () represents the unsaturated permeability coefficient; K s Indicates the saturated permeability coefficient; θ Indicates the volumetric water content of the soil; θ r This indicates the remaining volumetric water content of the soil. θ s This indicates the saturated volumetric water content of the soil.

2. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 1, characterized in that, The meteorological parameters include the number of sunny days before rainfall, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, and seasonal factors; the hydrological performance parameters include the water depth in the grassed swales and the outflow rate of the grassed swales.

3. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 2, characterized in that, The instantaneous saturated permeability coefficient dynamic prediction model is a dynamic prediction model for the instantaneous saturated permeability coefficient of vegetated swales with respect to the number of sunny days in the previous period, temperature, humidity, rainfall, rainfall duration, cumulative rainfall, and seasonal factors.

4. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 1, characterized in that, Step 10 specifically includes: Calculate the average meteorological parameters of historical rainfall events in the area where the grassed swale is to be evaluated, map the calculated average meteorological parameters into three categories: low, medium and high, and update the mapping results into the data collectability assessment model. Using a data availability assessment model and combining meteorological parameter values ​​under rainfall events from weather forecasts, the probability that future rainfall events in the area of ​​the grassy swamp to be assessed are representative is calculated one by one. The first rainfall event of each month is considered representative; if it is an extreme rainfall event, the next one is postponed. If the probability of a rainfall event being representative increases after adding meteorological parameter values ​​for another rainfall event, then that rainfall event is considered representative; otherwise, it is not.

5. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 1, characterized in that, In the vegetated swale hydrological model, the calculation process for the wetting stage includes: As rainfall continues, surface runoff flows in continuously. The moisture content of the filler layer rises steadily under the combined action of top inflow and bottom drainage until it reaches saturation. The water effluent from the bottom of the filler layer serves as the inflow for the drainage layer. Specifically: Equation (3) In the formula, This represents the volume of surface water in the vegetated swale at the end of a unit of time when there is no infiltration at the current moment; This represents the volume of surface water in the grassy swale before the current moment began; Indicates rainfall intensity; Indicates a unit time interval; Equation (4) In the formula, This indicates the depth of water accumulation in the vegetated swale at the end of a unit of time when there is no infiltration at the current moment; This represents the surface area of ​​the filler layer when the water storage layer is rectangular; Equation (5) In the formula, This represents the infiltration rate of the filler layer at the end of a unit time when there is no infiltration. Indicates the saturated permeability coefficient of the packing layer; Indicates the thickness of the filler layer; Equation (6) In the formula, The infiltration rate represents the volume of water accumulated in the vegetated swale after infiltration per unit time, assuming no infiltration. This represents the overflow volume of the grassed swale per unit time. make By setting the minimum error range and performing iterative calculations, the volume and depth of water accumulation in the grassy swale can be obtained. Equation (7) Equation (8) In the formula, This indicates the infiltration rate of the filler layer at the end of the current time when the bottom is no longer drained; This indicates the amount of infiltration of the filler layer relative to the remaining moisture content before the current moment begins; This indicates the moisture content of the filler layer at the end of the current time period when the bottom is no longer drained. Equation (9) In the formula, This represents the unsaturated permeability coefficient of the packing layer when the bottom is no longer drained at the current moment; This indicates the remaining volumetric water content of the packing layer; This indicates the saturated volumetric water content of the packing layer; This represents the saturated permeability coefficient of the packing layer; Equation (10) In the formula, This indicates the water outflow rate at the bottom of the packing layer at the current moment; Will As the inflow rate of the drainage layer, the same iterative method as that used for the filler layer was adopted to obtain the water volume and depth of the drainage layer surface at each moment; Equation (11) Equation (12) In the formula, This represents the infiltration rate of the packing layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This represents the water content of the packing layer after drainage, calculated based on the current time when the bottom of the packing layer is not drained. This represents the infiltration rate of the drainage layer after drainage, calculated based on the current time when the bottom of the filler layer is not drained. This represents the surface area of ​​the drainage layer. make Set the minimum error range and perform iterative calculations to obtain the infiltration rate and water content of the filler layer at the current moment; The same iterative method as that used for the filler layer was employed to calculate the infiltration rate and moisture content of the drainage layer.

6. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 1, characterized in that, In the vegetated swale hydrological model, the calculation process for the drainage stage includes: When the rainfall process has not yet ended, but the drainage layer is already full, the vegetated swale begins to drain water outward through the perforated drainage pipe. Assuming that the drainage layer has sufficient permeability to drain the water from the filler layer in time, the outflow rate of the drainage pipe per unit time is equal to the inflow rate at the bottom of the filler layer minus the infiltration rate into the natural soil at the bottom of the vegetated swale. The average infiltration rate from the bottom to the natural soil is considered to be approximately equal to the saturated permeability coefficient of the natural soil. The outflow rate is calculated using equation (13): Equation (13) In the formula, This indicates the average outflow rate at the outlet of the drainage pipe in the grassy ditch. This represents the saturated permeability coefficient of the natural soil at the bottom of the vegetated swale. Indicates the surface area of ​​the drainage layer; Calculate the pipe velocity along the friction and local head loss of the drainage blind pipe using equation (14): Equation (14) Equation (15) In the formula, The flow velocity along the pipe and at local head loss in the drainage blind pipe; H0 represents the total head including the traveling head; This represents the correction factor, which is typically set to 1.

0. This indicates head loss, including head loss along the friction path and local head loss. Indicates the friction coefficient; Indicates the length of the pipe; Indicates the pipe diameter; Indicates the cross-sectional length of the water flow in the pipe; Indicates the length of the pipe; Indicates the local drag coefficient; Indicates the flow velocity inside the pipe; After the rainfall ends, surface runoff also stops, and the vegetated swales enter the drainage stage. Initially, there is still water on the surface of the vegetated swales, and the infiltration meets the saturation infiltration requirement. The infiltration rate of the filler layer is calculated using equation (7). After the surface water disappears, the filler layer begins to gradually drain in an unsaturated state. The drainage rate of the filler layer is calculated using equation (16). Equation (16) In the formula, This indicates the evacuation rate of the packing layer. This represents the unsaturated permeability coefficient of the packing layer. This indicates the current moisture content of the filler layer; At this time, the drainage layer is still saturated. When the current moisture content of the drainage layer is less than the saturation moisture content of the drainage layer, the outlet stops discharging water, and the water stored in the drainage layer slowly seeps into the underground soil. When the moisture content of the filler layer and the drainage layer drops to their respective field moisture content, the infiltration process of this rainfall ends.

7. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 1, characterized in that, In step 20, the instantaneous saturated permeability coefficient values ​​of the vegetated swales for representative rainfall events are calculated using the instantaneous saturated permeability coefficient inversion method. Specifically, this includes: Using the water depth or outflow rate of vegetated swales under representative rainfall events as observed variables, the initial hypothesized instantaneous saturated permeability coefficient of the vegetated swales is substituted into the vegetated swales hydrological model to obtain the simulation results of the observed variables. The ensemble Kalman filter algorithm is used to assimilate the observed and simulated results to generate an estimated value of the instantaneous saturated permeability coefficient, which is then substituted into the vegetated swales hydrological model for simulation calculation. After multiple iterations, the converged instantaneous saturated permeability coefficient value of the vegetated swales is obtained.

8. The method for long-term hydrological effect monitoring and evaluation of vegetated swales according to claim 7, characterized in that, In step 20, the ensemble Kalman filter algorithm is used for data assimilation, specifically including: Based on the state transition equation shown in equation (17) and the observation equation shown in equation (18): Equation (17) Equation (18) In the formula, This represents the predicted value of the k-th set number of the parameters of the grassy swale hydrological model at time i+1; This represents the updated value of the k-th set number of the parameters of the grassy swale hydrological model at time i; This represents the predicted value of the k-th set number of the state variables at time i+1; This represents the updated value of the k-th set number of the state variables at time i; Represents the prediction operator; This represents the data driven by the hydrological model of the vegetated swale; This represents independent white noise representing the parameters of the grassy swale hydrological model. This represents independent white noise representing the state variables of the grassy swale hydrological model. This represents the kth set value of the hydrological simulation outflow or water accumulation depth of the grassy ditch at time i+1; h Represents the observation operator; The error term is represented by a mean of 0 and a variance of 0. The normal distribution; Assimilation is performed using equations (19) to (21): Equation (19) Equation (20) Equation (21) In the formula, This represents the predicted value of the k-th set number at time i+1; This represents the updated value of the k-th set at time i; Represents the white noise of the k-th set; This represents the observation value of the k-th set; This represents the observation error of the k-th set; The Kalman gain represents the weighted relationship between the predicted and observed values, calculated using equations (22) to (24): Equation (22) Equation (23) Equation (24) In the formula, Represents the covariance matrix of the predicted state variables; This represents the covariance matrix of the predicted values ​​of the observed variables; This represents the set mean of the predicted state variables; N represents the set mean of the predicted values ​​of the observed variables; N represents the set number.

Images