A method for predicting natural attenuation remediation of pollution in an underground water interaction zone

Abstract

The application discloses a kind of groundwater interaction zone pollution monitoring natural attenuation repair prediction method, it is related to groundwater pollution simulation and prediction technical field, the application includes: based on real-time water level monitoring data to construct unsteady flow model, two-way exchange flux is calculated by transient hydraulic gradient inversion algorithm;Establish three-dimensional solute displacement model, two-way exchange flux is input as dynamic boundary condition, synchronously simulate pollutant dissolved phase concentration and adsorption phase concentration, generate historical adsorption load distribution;Full true scene and anaerobic control scene are set in model, and the positive contribution of oxygen supply and the negative contribution of groundwater back discharge are quantitatively calculated by comparing output results;Based on positive and negative contribution data, the natural attenuation repair trend of interaction zone pollution under the condition of current water level fluctuation is predicted.The application realizes the accurate capture of dynamic exchange process and the contribution quantitative separation, solves the problem that the existing technology ignores the hysteresis effect, and the historical adsorption release risk is underestimated.

Description

Technical Field

[0001] This invention belongs to the field of groundwater pollution simulation and prediction technology, and in particular relates to a method for predicting the natural decay and remediation of pollution in groundwater interaction zones. Background Technology

[0002] The surface water-groundwater interaction zone is a critical transitional area between river ecosystems and groundwater systems, and also the most active zone for pollutant migration and transformation. Within this zone, river levels fluctuate frequently due to factors such as upstream reservoir regulation, seasonal rainfall, or artificial water replenishment, leading to periodic, alternating bidirectional exchanges between river water and groundwater. This dynamic exchange process has a complex dual impact on the natural decay of pollutants: on the one hand, the infiltration of clean river water rich in dissolved oxygen can supplement the interaction zone with electron acceptors, promoting aerobic microbial degradation and accelerating the natural decay of pollutants; on the other hand, the return of pollutant-laden groundwater to the river when the water level drops may cause secondary pollution, and when the river water itself is polluted, the infiltration of inferior water bodies will directly increase the pollution load of the interaction zone. Therefore, scientifically assessing the impact of dynamic water flow exchange on the natural decay of pollutants in the interaction zone is of great significance for the design of remediation schemes and environmental risk management for contaminated sites.

[0003] Currently, the main methods for predicting the natural degradation of groundwater pollution include numerical simulation methods based on physicochemical mechanisms and data-driven machine learning methods. Among them, numerical simulation methods can simulate the migration and transformation process of pollutants in the underground environment by constructing groundwater flow models and solute transport models. However, when dealing with the interaction boundary between surface water and groundwater, existing models usually adopt simplified constant head boundary or constant concentration boundary assumptions, treating the exchange process in the interaction zone as static or steady state. They cannot dynamically capture the instantaneous bidirectional water exchange volume caused by water level fluctuations and its real-time impact on the pollutant degradation rate.

[0004] Existing technologies struggle to quantitatively distinguish the complex relationship between the oxygen replenishment from river infiltration promoting degradation and dilution effects and the secondary migration of pollutants caused by groundwater backflow, leading to significant uncertainty in the assessment of the natural attenuation capacity of the interaction zone. Summary of the Invention

[0005] The purpose of this invention is to provide a method for monitoring and predicting the natural decay and remediation of pollution in groundwater interaction zones, which solves the problems in the prior art of making it difficult to dynamically capture the instantaneous bidirectional water exchange caused by water level fluctuations, making it impossible to quantitatively distinguish the oxygen supply promoting degradation and dilution effects brought by river water infiltration, and making it impossible to scientifically assess the risk of secondary pollution caused by groundwater backflow.

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0007] This invention relates to a method for monitoring and predicting the natural degradation and remediation of pollution in groundwater interaction zones, comprising the following steps:

[0008] S1. Based on real-time monitoring of river water level time series data and riverbank groundwater level data, an unsteady flow model is constructed. According to the transient hydraulic gradient inversion algorithm, the water level difference driving each time period is calculated, and the bidirectional interactive flux is obtained through the river water infiltration and groundwater discharge per unit area of ​​the riverbed.

[0009] S2. Establish a three-dimensional solute displacement model, input the calculated instantaneous bidirectional interactive flux as a dynamic boundary condition into the three-dimensional solute displacement model, and simultaneously calculate and simulate the concentration of pollutants in the dissolved phase in groundwater and the concentration of pollutants in the adsorbed phase in the soil particle layer. At the same time, after the simulation is completed, the historical adsorption load distribution is generated.

[0010] S3. Set up two parallel virtual scenarios in the three-dimensional solute displacement model, including a full-scale scenario and an anoxic control scenario. By comparing the output results of the two scenarios, calculate the positive contribution of oxygen replenishment from river water infiltration to the natural decay of pollutants, and the negative contribution of pollutant backflow flux from groundwater discharge.

[0011] S4. Based on the obtained positive and negative contribution data, predict the natural attenuation and remediation trend of groundwater pollution in the current water level fluctuation condition.

[0012] Furthermore, the unsteady flow model includes both one-dimensional and two-dimensional models. A one-dimensional model is used to simulate water flow along the direction perpendicular to the riverbank, while a two-dimensional model is used to simulate water flow simultaneously along both the direction perpendicular to and along the riverbank. The transient hydraulic gradient inversion algorithm includes: calculating the hydraulic gradient at each moment based on real-time monitored river and groundwater levels; combining this with the permeability coefficient of the riverbed medium; and using Darcy's law to invert the instantaneous infiltration rate of river water and the discharge rate of groundwater per unit area of ​​the riverbed.

[0013] The formula (1) for calculating the hydraulic gradient is: (1)

[0014] The formula (2) for calculating bidirectional interactive flux is: (2)

[0015] in, Let t be the river water level. Let t be the groundwater level at time t, L be the horizontal distance from the river channel to the monitoring well, and K be the riverbed permeability coefficient.

[0016] Furthermore, the three-dimensional solute displacement model in step S2 includes the decay reaction kinetics equations for the target pollutant, which include:

[0017] The kinetic equation for the nitration reaction, and its calculation formula (3) are as follows:

[0018] (3)

[0019] in, ammonia nitrification rate, For the maximum specific growth rate, This refers to the ammonia nitrogen concentration. The ammonia nitrogen half-saturation constant, Dissolved oxygen concentration, The dissolved oxygen half-saturation constant, This is a temperature correction factor, where T is the water temperature;

[0020] The adsorption-desorption kinetic equation, and its calculation formula (4) are as follows:

[0021] (4)

[0022] Where q is the concentration of the adsorbed phase and C is the concentration of the pollutant dissolved phase. For maximum adsorption capacity, The adsorption rate constant is . Let be the desorption rate constant. The adsorption amount q is the rate of change with time t;

[0023] The biodegradation kinetic equation, and its calculation formula (5) are as follows:

[0024] (5)

[0025] Where C is the concentration of the pollutant in the dissolved phase, and k is the first-order degradation rate constant. Let C be the rate of change of concentration C over time t.

[0026] Furthermore, the anoxic control scenario involves setting the dissolved oxygen concentration from river infiltration to zero in the three-dimensional solute displacement model calculation, while maintaining water exchange and other water quality parameters consistent with the full-scale real-world scenario. The positive contribution is calculated by comparing the pollutant concentration differences between the full-scale real-world scenario and the anoxic control scenario to determine the amount of pollutant degradation promoted by oxygen replenishment. The negative contribution is calculated by multiplying the groundwater discharge flux output by the three-dimensional solute displacement model with the groundwater pollutant concentration to obtain the total amount of pollutants discharged back into the river channel per unit time through groundwater discharge.

[0027] Furthermore, the predicted indicators in step S4 include: oxygen contribution rate, recharge contribution rate, and net benefit index. The oxygen contribution rate is the percentage of oxygen-promoted degradation to the total pollutant attenuation. The recharge contribution rate is the percentage of pollutant flux carried by groundwater recharge to the total load of the interaction zone. The net benefit index is the sum of oxygen contribution and dilution contribution minus the sum of recharge contribution and new pollution contribution.

[0028] Furthermore, based on the historical adsorption load distribution generated in step S2, and using the adsorption and desorption hysteresis characteristic parameters in the interaction zone, a long-term pollutant release early warning model is established. The early warning model includes:

[0029] A. Obtain historical adsorption load characteristic parameters. The historical adsorption load characteristic parameters are extracted based on the historical adsorption load distribution generated in step S2. The historical adsorption load distribution includes the cumulative adsorption amount of the target pollutant by soil particles at each spatial grid point, as well as the frequency of water level fluctuations and hydrochemical change characteristics experienced by each grid point during the simulation period.

[0030] B. Based on historical adsorption load characteristic parameters, and according to the hysteresis effect of the adsorbed pollutant release process, predict the future release behavior of adsorbed pollutants from the interaction zone to groundwater and surface water under the condition of improved surface water quality. The hysteresis effect refers to the delay of the desorption process relative to the adsorption process, that is, the release rate of adsorbed pollutants depends on their adsorption history.

[0031] C. Output risk prediction indicators that characterize future release behavior.

[0032] Furthermore, predicting future release behavior includes: determining the adsorption saturation based on the ratio of the extracted cumulative adsorption amount to the preset upper limit of adsorption capacity; determining the degree of historical disturbance based on the frequency of water level fluctuations and the characteristics of water chemical changes; and predicting the release process of adsorbed pollutants based on the influence of adsorption saturation and the degree of historical disturbance on the release rate. The upper limit of adsorption capacity is predetermined based on the soil type and the characteristics of the target pollutant.

[0033] Furthermore, risk prediction indicators include: release rate, i.e., the total amount of pollutants released from the interaction zone per unit time; release duration, i.e., the time required from the start of surface water quality improvement to the basic completion of release; peak concentration and occurrence time, i.e., the maximum value of pollutant concentration during release and the time of its occurrence; source-sink status, i.e., whether the interaction zone is a pollution source or a pollution sink at the current moment; and degree of lag effect, i.e., the relative proportion of the release duration extension caused by the lag effect.

[0034] Furthermore, the early warning model also includes the following steps:

[0035] D. After outputting the risk prediction indicators, obtain real-time monitoring data of the interaction zone area. The real-time monitoring data includes groundwater pollutant concentration, water level fluctuation data and water chemical indicators of the monitoring wells.

[0036] E. Employ a data assimilation algorithm to assimilate real-time monitoring data into the lag-based prediction process in step B, and dynamically update the key parameters on which the prediction depends.

[0037] F. Based on the updated parameters, re-execute steps B and C to output the revised risk prediction indicators.

[0038] Furthermore, data assimilation algorithms include ensemble Kalman filtering, particle filtering, and Bayesian dynamic updates.

[0039] The present invention has the following beneficial effects:

[0040] 1. This invention constructs an unsteady flow model and introduces a transient hydraulic gradient inversion algorithm to dynamically calculate the bidirectional interactive flux driven by water level fluctuations at each moment, overcoming the limitations of traditional static boundary assumptions. Furthermore, by setting up parallel scenario comparisons, it quantitatively separates the oxygen replenishment degradation and dilution effects brought about by river water infiltration, and simultaneously calculates the negative contribution of groundwater backflow, solving the problem that existing technologies cannot distinguish the contributions of different processes to natural decay.

[0041] 2. Based on the adsorption load distribution generated by historical simulation, this invention introduces adsorption saturation and historical disturbance degree to jointly characterize the hysteresis effect, and realizes accurate prediction of key risks such as release rate, duration and peak concentration of the interaction zone as a long-term secondary pollution source after the improvement of river water quality, making up for the shortcomings of traditional methods that only focus on the current hydrodynamic conditions.

[0042] 3. This invention achieves an organic connection between current dynamic assessment and long-term release prediction; at the same time, it introduces a data assimilation algorithm to dynamically calibrate the prediction parameters and output the uncertainty range of the prediction results, so that the prediction model can self-optimize as new monitoring data is acquired, which significantly improves the reliability of long-term prediction and decision support capabilities. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart of the method of the present invention.

[0045] Figure 2This is a flowchart of the early warning model in this invention. Detailed Implementation

[0046] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0047] See Figure 1 This invention provides a method for predicting the natural degradation and remediation of pollution in groundwater interaction zones, comprising the following steps:

[0048] S1. Based on real-time monitoring of river water level time series data and riverbank groundwater level data, an unsteady flow model is constructed. According to the transient hydraulic gradient inversion algorithm, the water level difference driving each time period is calculated, and the bidirectional interactive flux is obtained through the river water infiltration and groundwater discharge per unit area of ​​the riverbed.

[0049] S2. Establish a three-dimensional solute displacement model, input the calculated instantaneous bidirectional interactive flux as a dynamic boundary condition into the three-dimensional solute displacement model, and simultaneously calculate and simulate the concentration of pollutants in the dissolved phase in groundwater and the concentration of pollutants in the adsorbed phase in the soil particle layer. At the same time, after the simulation is completed, the historical adsorption load distribution is generated.

[0050] S3. Set up two parallel virtual scenarios in the three-dimensional solute displacement model, including a full-scale scenario and an anoxic control scenario. By comparing the output results of the two scenarios, calculate the positive contribution of oxygen replenishment from river water infiltration to the natural decay of pollutants, and the negative contribution of pollutant backflow flux from groundwater discharge.

[0051] S4. Based on the obtained positive or negative contribution data, predict the natural attenuation and remediation process of groundwater pollution in the interaction zone over a future period under the current water level fluctuation conditions.

[0052] It should be further noted that the unsteady flow model can be a one-dimensional or two-dimensional model. A one-dimensional model is used to simulate water flow along the direction perpendicular to the riverbank, while a two-dimensional model is used to simulate water flow simultaneously along both the direction perpendicular to and along the riverbank. The transient hydraulic gradient inversion algorithm includes: calculating the hydraulic gradient at each moment based on real-time monitored river and groundwater levels; combining this with the permeability coefficient of the riverbed medium; and using Darcy's law to invert the instantaneous infiltration rate of river water or the discharge rate of groundwater per unit area of ​​the riverbed.

[0053] The formula (1) for calculating the hydraulic gradient is: (1)

[0054] The formula (2) for calculating bidirectional interactive flux is: (2)

[0055] in, Let t be the river water level. Let t be the groundwater level at time t, L be the horizontal distance from the river channel to the monitoring well, and K be the riverbed permeability coefficient.

[0056] The three-dimensional solute displacement model in step S2 includes the decay reaction kinetic equations of the target pollutant, which include:

[0057] The kinetic equation for the nitration reaction, and its calculation formula (3) are as follows:

[0058] (3)

[0059] in, ammonia nitrification rate, For the maximum specific growth rate, This refers to the ammonia nitrogen concentration. The ammonia nitrogen half-saturation constant, Dissolved oxygen concentration, The dissolved oxygen half-saturation constant, This is a temperature correction factor, where T is the water temperature;

[0060] The adsorption-desorption kinetic equation, and its calculation formula (4) are as follows:

[0061] (4)

[0062] Where q is the concentration of the adsorbed phase and C is the concentration of the pollutant dissolved phase. For maximum adsorption capacity, The adsorption rate constant is . Let be the desorption rate constant. The adsorption amount q is the rate of change with time t;

[0063] The biodegradation kinetic equation, and its calculation formula (5) are as follows:

[0064] (5)

[0065] Where C is the concentration of the pollutant in the dissolved phase, and k is the first-order degradation rate constant. Let C be the rate of change of concentration C over time t.

[0066] The anoxic control scenario involves setting the dissolved oxygen concentration from river infiltration to zero in the three-dimensional solute displacement model calculation, while maintaining water exchange and other water quality parameters consistent with the full-scale real-world scenario. The positive contribution is calculated by comparing the difference in pollutant concentration between the full-scale real-world scenario and the anoxic control scenario, or by comparing the difference in the total amount of pollutants between the anoxic control scenario and the full-scale real-world scenario, to determine the amount of pollutants degraded by oxygen supply. The negative contribution is calculated by multiplying the groundwater discharge flux output by the three-dimensional solute displacement model by the groundwater pollutant concentration, to determine the total amount of pollutants discharged back into the river channel per unit time through groundwater discharge.

[0067] The predicted indicators in step S4 include: oxygen contribution rate, recharge contribution rate, and net benefit index. The oxygen contribution rate is the percentage of oxygen-promoted degradation to the total pollutant attenuation. The recharge contribution rate is the percentage of pollutant flux carried by groundwater recharge to the total load of the interaction zone. The net benefit index is the sum of oxygen contribution and dilution contribution minus the sum of recharge contribution and new pollution contribution.

[0068] In this embodiment, the surface water infiltration rate is the amount of surface water flowing back into groundwater per unit time, while the groundwater discharge rate is the amount of groundwater flowing into surface water per unit time. The combination of the two is the bidirectional interactive flux. The hydraulic gradient is the height difference between the surface water level and the groundwater level, i.e., the pressure difference driving the water flow. The change of the pressure difference over a unit distance is the hydraulic gradient. The larger the gradient, the faster the water flow. The gradient direction determines the water flow direction. The inversion algorithm refers to the process of reversing the unknown exchange water volume based on known water level changes. Step S1 continuously calculates the instantaneous hydraulic gradient caused by water level changes, and combines it with the permeability coefficient of the riverbed medium to invert the river water infiltration rate and groundwater discharge rate that change over time, reflecting the water exchange process driven by water level fluctuations. Boundary adjustment refers to the area calculated in the model, which is divided into grids according to the calculated area. The calculation process of the inversion algorithm includes the following specific contents:

[0069] First, input the data, which includes the river water level time series, i.e., the water level height recorded at fixed times, the riverbank groundwater level time series, records with the same frequency as the river water level time series, and riverbed geological parameters.

[0070] Then, a one-dimensional or two-dimensional model is selected based on the direction of water flow;

[0071] The interaction flux is then calculated using the formula (2) for bidirectional interaction flux. The final output is a curve graph that changes over time. The horizontal axis represents time, and the vertical axis represents the exchange flux. Positive values ​​represent river water infiltration, and negative values ​​represent groundwater discharge.

[0072] The three-dimensional solute displacement model in step S2 is used to calculate the movement position of pollutants with the water flow. The basic principle conforms to the convection-dispersion equation. The model operation includes the following specific contents:

[0073] Initialization: Set initial values ​​for each grid point in the entire computational domain, including initial water level, initial ammonia nitrogen concentration, initial adsorption capacity, etc.

[0074] Time loop: Starting from t=0, calculate once at each time point: read the boundary exchange flux q(t) at this time; calculate the water flow distribution at this time; calculate the pollutant migration location at this time; calculate the chemical reaction at this time, i.e., the degradation and adsorption of ammonia nitrogen; update the water level and concentration at each grid point;

[0075] After the cycle reaches the set total time, the model's running results are output, ultimately providing ammonia nitrogen concentration distribution map and adsorption amount distribution map at any time and location within that cycle time.

[0076] In step S3, the full-scale scenario simulates a completely realistic natural state, which includes the following: water exchange is completely in accordance with the bidirectional flux q(t) calculated in step S1, and water can freely enter and exit; water quality exchange includes dissolved oxygen and pollutants inside the river water during infiltration, and pollutants inside the groundwater during discharge; finally, the contribution data of this scenario is output.

[0077] The anaerobic control scenario is used for simulation and includes the following: water exchange is exactly the same as the real scenario. The difference in water quality exchange is that when surface river water infiltrates the interaction zone, the dissolved oxygen concentration in the river water is set to 0, but the pollutant concentration in the river water remains unchanged. Therefore, in this scenario, aerobic degradation cannot occur because no oxygen is involved. Other processes are consistent with the real scenario. The contribution data of this scenario is finally output.

[0078] The output results of the two scenarios were then compared. The oxygen contribution was the difference between the pollutant concentration in the real scenario and the pollutant concentration in the anoxic control scenario. This difference was the degradation amount contributed by oxygen alone. At the same time, the negative contribution was calculated. The ammonia nitrogen backflow flux in the negative contribution was the product of the basement backflow amount and the groundwater ammonia nitrogen concentration.

[0079] Assuming the initial ammonia nitrogen concentration at a simulated monitoring point is 100 units, after one year of simulation, the final ammonia nitrogen concentration in the anaerobic control scenario is 70, while the final ammonia nitrogen concentration in the full-scale real scenario is 40. That is, the oxygen contribution is the difference between the full-scale real scenario and the anaerobic control scenario, with a total attenuation of 60. Therefore, water level fluctuations generally help with the restoration. If the difference is negative, it will generally hinder the restoration.

[0080] See Figure 2In the above embodiments, the adsorption and desorption processes of pollutants on the interaction zone medium are shown to be instantaneously reversible during model construction. That is, it is assumed that when hydrochemical conditions recover and river water quality improves, pollutants adsorbed on soil particles will be released into the water immediately. However, due to frequent wet-dry cycles and continuous changes in hydrochemical conditions (such as pH and ionic strength), the adsorption and desorption processes of pollutants on soil particles exhibit a lag effect. This results in the adsorbed pollutants not being released immediately after environmental conditions recover, but rather being released slowly and with a lag. Therefore, pollutants accumulated in the sediments of the interaction zone may continue to be released over a long period in the future, becoming a secondary pollution source. To address this problem, in this embodiment, based on the historical adsorption load distribution generated in step S2 and the lag characteristic parameters of adsorption and desorption in the interaction zone, a long-term pollutant release early warning model is established. The early warning model includes:

[0081] A. Obtain historical adsorption load characteristic parameters. The historical adsorption load characteristic parameters are extracted based on the historical adsorption load distribution generated in step S2. The historical adsorption load distribution includes the cumulative adsorption amount of the target pollutant by soil particles at each spatial grid point, as well as the frequency of water level fluctuations and hydrochemical change characteristics experienced by each grid point during the simulation period.

[0082] B. Based on historical adsorption load characteristic parameters, and according to the hysteresis effect of the adsorbed pollutant release process, predict the future release behavior of adsorbed pollutants from the interaction zone to groundwater or surface water under the condition of improved surface water quality. The hysteresis effect refers to the delay of the desorption process relative to the adsorption process, that is, the release rate of adsorbed pollutants depends on their adsorption history.

[0083] C. Output at least one risk prediction indicator that characterizes future release behavior.

[0084] It should be further explained that predicting future release behavior further includes: determining the adsorption saturation based on the ratio of the extracted cumulative adsorption amount to the preset upper limit of adsorption capacity; determining the degree of historical disturbance based on the frequency of water level fluctuations and the characteristics of water chemical changes; and predicting the release process of adsorbed pollutants based on the influence of adsorption saturation and the degree of historical disturbance on the release rate. The upper limit of adsorption capacity is predetermined based on the soil type and the characteristics of the target pollutant.

[0085] Risk prediction indicators include: release rate, which is the total amount of pollutants released from the interaction zone per unit time; release duration, which is the time required from the start of surface water quality improvement to the basic completion of release; peak concentration and occurrence time, which is the maximum value of pollutant concentration during release and the time of its occurrence; source-sink status, which is whether the interaction zone is a pollution source or a pollution sink at the current moment; and degree of lag effect, which is the relative proportion of the release duration extension caused by the lag effect.

[0086] In this embodiment, the historical adsorption load distribution is used as the input condition. The problem of predicting the accumulation of pollutants in soil particles is solved based on the hysteresis effect. The historical adsorption load distribution records the cumulative adsorption amount of each grid point, as well as the frequency of water level fluctuations and water chemical changes experienced during the simulation period. Based on the historical data, the release behavior of adsorbed pollutants after the improvement of river water quality is predicted through the hysteresis effect information.

[0087] See Figure 2 In the above embodiments, with continuous alternation of wet and dry conditions, further evolution of hydrochemical conditions, and aging or structural reorganization of soil particles, the intensity of the hysteresis effect may dynamically change. The prediction accuracy of the static prediction model established based on the initial parameters will gradually decrease over time, thereby affecting the prediction results and the reliability of the prediction. Therefore, in order to solve this problem, in this embodiment, the early warning model further includes the following steps:

[0088] D. After outputting the risk prediction indicators, obtain real-time monitoring data of the interactive zone area. The real-time monitoring data includes at least one of the following: groundwater pollutant concentration, water level fluctuation data, or water chemical indicators of the monitoring well.

[0089] E. Employ a data assimilation algorithm to assimilate real-time monitoring data into the lag-based prediction process in step B, and dynamically update the key parameters on which the prediction depends.

[0090] F. Based on the updated parameters, re-execute steps B and C to output the revised risk prediction indicators.

[0091] Data assimilation algorithms include at least one of ensemble Kalman filtering, particle filtering, or Bayesian dynamic updating.

[0092] The risk prediction index output in step C further includes the uncertainty interval of the prediction result. The uncertainty interval is obtained by setting the key parameters to probability distribution form during the prediction process in step B, generating the probability distribution of the risk prediction index through Monte Carlo simulation or ensemble forecasting methods, and outputting the confidence interval at the specified confidence level.

[0093] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting the natural degradation and remediation of pollution in groundwater interaction zones, characterized in that: Includes the following steps: S1. Based on real-time monitoring of river water level time series data and riverbank groundwater level data, an unsteady flow model is constructed. According to the transient hydraulic gradient inversion algorithm, the infiltration of river water and the discharge of groundwater per unit area of ​​the riverbed driven by the water level difference are calculated in each time period to obtain the bidirectional interactive flux. The unsteady flow model includes both one-dimensional and two-dimensional models. The one-dimensional model simulates water flow along the direction perpendicular to the riverbank, while the two-dimensional model simulates water flow simultaneously along both the direction perpendicular to and along the riverbank. The transient hydraulic gradient inversion algorithm includes: calculating the hydraulic gradient at each moment based on real-time monitored river and groundwater levels; combining this with the permeability coefficient of the riverbed medium; and using Darcy's law to invert the instantaneous infiltration rate of river water per unit area through the riverbed and the groundwater discharge rate. The formula (1) for calculating the hydraulic gradient is: (1); The formula (2) for calculating bidirectional interactive flux is: (2); in, Let t be the river water level. Let t be the groundwater level at time t, L be the horizontal distance from the river channel to the monitoring well, and K be the riverbed permeability coefficient; S2. Establish a three-dimensional solute displacement model, input the calculated instantaneous bidirectional interactive flux as a dynamic boundary condition into the three-dimensional solute displacement model, and simultaneously calculate and simulate the concentration of pollutants in the dissolved phase in groundwater and the concentration of pollutants in the adsorbed phase in the soil particle layer. At the same time, after the simulation is completed, the historical adsorption load distribution is generated. The three-dimensional solute displacement model includes the decay reaction kinetics equations for the target pollutant, which include: The kinetic equation for the nitration reaction, and its calculation formula (3) are as follows: （3）； in, ammonia nitrification rate, For the maximum specific growth rate, This refers to the ammonia nitrogen concentration. The ammonia nitrogen half-saturation constant, Dissolved oxygen concentration, The dissolved oxygen half-saturation constant, Here is the temperature correction factor, where T is the water temperature; The adsorption-desorption kinetic equation, and its calculation formula (4) are as follows: （4）； Where q is the concentration of the adsorbed phase and C is the concentration of the pollutant dissolved phase. For maximum adsorption capacity, The adsorption rate constant is . Let be the desorption rate constant. Let q be the rate of change of the adsorption amount with time t; The biodegradation kinetic equation, and its calculation formula (5) are as follows: （5）； Where C is the concentration of the pollutant in the dissolved phase, and k is the first-order degradation rate constant. The concentration C is the rate of change with time t; S3. Set up two parallel virtual scenarios in the three-dimensional solute displacement model, including a full-scale scenario and an anoxic control scenario. By comparing the output results of the two scenarios, calculate the positive contribution of oxygen replenishment from river water infiltration to the natural decay of pollutants, and the negative contribution of pollutant backflow flux from groundwater discharge. S4. Based on the obtained positive and negative contribution data, predict the natural attenuation and remediation trend of groundwater pollution in the interaction zone under the current water level fluctuation conditions.

2. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 1, characterized in that, The anaerobic control scenario is used in the three-dimensional solute displacement model calculation. The dissolved oxygen concentration brought by river water infiltration was set to zero, while water exchange and other water quality parameters were kept consistent with the real-world scenario. The positive contribution was calculated by comparing the pollutant concentration difference between the real-world scenario and the anoxic control scenario to obtain the amount of pollutant degradation promoted by oxygen supply. The negative contribution was obtained by multiplying the groundwater discharge flux output by the three-dimensional solute displacement model with the groundwater pollutant concentration to obtain the total amount of pollutants discharged back into the river channel per unit time through groundwater discharge.

3. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 1, characterized in that, The predicted indicators in step S4 include: oxygen contribution rate, recharge contribution rate, and net benefit index, wherein the oxygen contribution rate is the percentage of oxygen-promoted degradation to the total pollutant attenuation, the recharge contribution rate is the percentage of pollutant flux carried by groundwater recharge to the total load of the interaction zone, and the net benefit index is the sum of oxygen contribution and dilution contribution minus the sum of recharge contribution and new pollution contribution.

4. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 3, characterized in that, The historical adsorption load distribution generated in step S2, based on the adsorption and desorption hysteresis characteristic parameters in the interaction zone, establishes a long-term pollutant release early warning model, which includes: A. Obtain historical adsorption load characteristic parameters. The historical adsorption load characteristic parameters are extracted based on the historical adsorption load distribution generated in step S2. The historical adsorption load distribution includes the cumulative adsorption amount of the target pollutant by soil particles at each spatial grid point, as well as the frequency of water level fluctuations and hydrochemical change characteristics experienced by each grid point during the simulation period. B. Based on the aforementioned historical adsorption load characteristic parameters, and considering the hysteresis effect of the adsorbed pollutant release process, predict the future release behavior of adsorbed pollutants from the interaction zone to groundwater and surface water under conditions of improved surface water quality. The hysteresis effect refers to the desorption process relative to the adsorption process. The process delay phenomenon, that is, the release rate of adsorbed pollutants depends on their adsorption history; C. Output risk prediction indicators that characterize future release behavior.

5. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 4, characterized in that, Predicting the future release behavior further includes: determining the adsorption saturation based on the ratio of the extracted cumulative adsorption amount to a preset upper limit of adsorption capacity; determining the degree of historical disturbance based on the frequency of water level fluctuations and the characteristics of water chemical changes; and predicting the release process of adsorbed pollutants based on the influence of the adsorption saturation and the degree of historical disturbance on the release rate, wherein the upper limit of adsorption capacity is predetermined based on the soil type and the characteristics of the target pollutant.

6. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 5, characterized in that, The risk prediction indicators include: release rate, which is the total amount of pollutants released from the interaction zone per unit time; release duration, which is the time required from the start of surface water quality improvement to the completion of release; peak concentration and occurrence time, which is the maximum value of pollutant concentration during release and the time of its occurrence; source-sink status, which is whether the interaction zone is a pollution source or a pollution sink at the current moment; and degree of lag effect, which is the relative proportion of the release duration extension caused by the lag effect.

7. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 6, characterized in that, The early warning model also includes the following steps: D. After outputting the risk prediction indicators, obtain real-time monitoring data of the interactive zone area. The real-time monitoring data includes groundwater pollutant concentration, water level fluctuation data, and water chemical indicators of the monitoring well. E. Using a data assimilation algorithm, the real-time monitoring data is assimilated into the prediction process based on lag effect in step B, and the key parameters on which the prediction depends are dynamically updated. F. Based on the updated parameters, re-execute steps B and C to output the revised risk prediction indicators.

8. The method for predicting the natural attenuation and remediation of pollution in groundwater interaction zones according to claim 7, characterized in that, The data assimilation algorithm includes ensemble Kalman filtering, particle filtering, and Bayesian dynamic update.

Images