Underground water level early warning system and method

Abstract

The invention discloses an underground water level early warning system and method, and belongs to the field of water level early warning, and the system comprises a water level early warning threshold determination module which is used for determining a water level early warning threshold of an evaluation area; the station network selection module is used for selecting a representative station based on the water level early warning threshold value and the underground water type; the water level change calculation module is used for calculating underground water level change conditions; and the prediction and early warning module is used for issuing a water level change early warning identifier and predicting the underground water level change.

Description

Technical Field

[0001] The present invention belongs to the technical field of water level early warning, and in particular relates to a groundwater level early warning system and method. Background Art

[0002] Groundwater early warning includes water quality early warning and water level early warning, which are of two types: spatial early warning and time early warning. Spatial early warning is also called state early warning, which is based on the main control factors and inducing factors of groundwater quality and water level deterioration. The control factors are basic conditions, and the inducing factors often show great differences in different regions or different sections of the same region. Time early warning is also called trend early warning, which is based on the intensity of inducing factors of groundwater quality and water level stability, as well as their duration and prediction accuracy. Generally, it is based on spatial early warning, through professional technical monitoring, systematic theoretical analysis and expert discussion, and reported to relevant management departments for early warning. Carrying out groundwater early warning work is of great significance for protecting and managing groundwater resources, responding to water and drought disaster risks, timely discovering and preventing groundwater pollution, supporting decision-making and policy formulation, and raising public awareness of participation.

[0003] Groundwater quality early warning requires continuous quantitative and qualitative analysis of groundwater quality status, determines its changing trend and gives different early warning information according to the monitoring results. However, the existing groundwater monitoring projects have failed to fully consider the impact of human pollution in the layout of monitoring points and the selection of monitoring indicators. There are many water quality monitoring indicators, which need to be selected in combination with regional groundwater problems. At present, monitoring is difficult, and the work currently being carried out mainly revolves around groundwater level early warning. Therefore, the present invention provides a groundwater level early warning system and method. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a groundwater level early warning system and method to solve the problems existing in the above prior art.

[0005] To achieve the above object, the present invention provides a groundwater level early warning system, comprising:

[0006] Wherein, the water level warning threshold determination module is used to determine the water level warning threshold of the overexploitation area;

[0007] The station network selection module is used to select representative stations based on the water level warning threshold and groundwater type;

[0008] The water level change calculation module is used to calculate the groundwater level change;

[0009] The prediction and warning module is used to issue water level change warning signs and predict groundwater level changes.

[0010] Optionally, the groundwater level early warning system further comprises an early warning information display module, and the early warning information display module is used to visually display early warning results in the form of a graph; the early warning results comprise a rainbow early warning mode and a four-color identification early warning schematic diagram.

[0011] Optionally, the water level change calculation module includes a data processing unit and a prediction unit;

[0012] Wherein, the data processing unit is used to collect groundwater level monitoring data

[0013] The prediction unit is used to establish a groundwater level depth prediction model, and predict groundwater level changes based on the groundwater level depth prediction model and the groundwater level monitoring data.

[0014] The present invention also provides a groundwater level early warning method for implementing a groundwater level early warning system, the method comprising the following steps:

[0015] Determine the water level warning threshold in the assessment area;

[0016] Based on the water level warning threshold, representative stations are selected according to shallow groundwater and deep groundwater respectively;

[0017] Based on the selected representative stations, calculate the water level changes to obtain early warning information;

[0018] The water level change is predicted to obtain groundwater level prediction information.

[0019] Optionally, the process of determining the water level warning threshold of the evaluation area includes:

[0020] Select the early warning method based on the duration of the groundwater level monitoring sequence in the evaluation area;

[0021] When the monitoring sequence is longer than 30 years, a third data set is directly constructed using statistical methods, and the water level warning threshold is determined based on the directly constructed third data set;

[0022] When the monitoring sequence is less than 30 years, the first data set is constructed based on the groundwater level factor to determine the upper limit water level; and the second data set is constructed based on the mining volume monitoring data to determine the lower limit water level;

[0023] The first data set and the second data set are merged to indirectly construct a third data set, and a water level warning threshold is determined based on the indirectly constructed third data set.

[0024] Optionally, the groundwater level factors include: groundwater level to prevent salinization, safe groundwater level for urban buildings for building stability and foundation bearing capacity, and groundwater recharge capacity.

[0025] Optionally, the process of predicting the water level change to obtain groundwater level prediction information includes:

[0026] The shallow groundwater prediction model and deep groundwater prediction model were established by using multivariate regression model and neural network RBF model respectively.

[0027] Determine the groundwater level prediction information of each city-level administrative area based on the shallow groundwater prediction model and the deep groundwater prediction model;

[0028] Comparing the groundwater level prediction information with the measured burial depth on a monthly basis to obtain an error value;

[0029] The shallow groundwater prediction model and the deep groundwater prediction model are optimized based on the error value to determine the model scheme and model parameters selected by various cities.

[0030] Optionally, the factors affecting the establishment of the shallow groundwater prediction model include: the groundwater level depth of the previous month, the precipitation of the current month, and the groundwater extraction volume of the current month;

[0031] The factors affecting the establishment of the deep groundwater prediction model include: the groundwater level depth in the previous month and the groundwater extraction volume in the current month.

[0032] The present invention also provides a computer terminal device, comprising:

[0033] one or more processors;

[0034] A memory, coupled to the processor, for storing one or more programs;

[0035] When the one or more programs are executed by the one or more processors, the one or more processors implement a groundwater level early warning system.

[0036] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, a groundwater level early warning system is implemented.

[0037] Compared with the prior art, the present invention has the following advantages and technical effects:

[0038] The present invention improves the groundwater early warning framework, proposes a process suitable for groundwater level early warning, and promotes scientific decision-making and precise management of groundwater management. The groundwater level early warning system effectively identifies the water level risk in over-exploited areas by accurately determining the water level early warning threshold. The system uses the station network selection module to intelligently select representative monitoring sites based on the early warning threshold and groundwater type to ensure the accuracy and representativeness of the monitoring data. The water level change calculation module monitors and calculates water level changes in real time to provide data support for prediction and early warning. Finally, the prediction and early warning module not only issues real-time early warning signs, but also predicts the trend of groundwater level changes, providing a scientific basis for water resources management and protection, and effectively preventing and reducing the environmental impact of groundwater over-exploitation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0040] Figure 1 A flow chart of selecting a water level warning threshold value according to an embodiment of the present invention;

[0041] Figure 2 This is the process of the water level warning method according to an embodiment of the present invention;

[0042] Figure 3 Schematic diagram of rainbow warning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0043] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0044] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0045] Embodiment 1

[0046] Ten key areas, including the Sanjiang Plain, Songnen Plain, Liaohe Plain and Northwest Liaoning, Xiliaohe River Basin, Huanghuai Region, Ordos Terrace, Fenwei Valley, Hexi Corridor, the southern and northern foothills of the Tianshan Mountains, Tuha Basin, and Beibu Gulf, involve 13 provincial administrative regions, 72 cities, and an area of ​​2.44 million km2, accounting for 25% of the country's land area. The corresponding over-exploitation area is 161,000 km2, accounting for 56% of the country's over-exploitation area, and the over-exploitation accounts for nearly 70% of the country's groundwater over-exploitation. Groundwater over-exploitation in North China and key areas and the resource and environmental problems it has caused have threatened regional water security and economic and social sustainable development. Therefore, we focus on key areas with prominent groundwater over-exploitation problems and systematically promote groundwater over-exploitation governance.

[0047] like Figure 2 As shown, this embodiment provides a groundwater level warning system, including a water level warning threshold determination module, a station network selection module, a water level change notification module and a prediction and warning module.

[0048] Furthermore, if Figure 1 As shown in the figure, the water level warning threshold determination module is used to determine the water level warning threshold: the key to groundwater level warning is to determine the water level threshold. The thresholds proposed in the current research include many concepts such as critical water level, warning water level, suitable water level, engineering safety water level, ecological safety burial depth, ecological threshold, etc. The degree of attention paid to groundwater level in various fields has different focuses. For example, the hydrogeological field focuses on the relationship between groundwater level and ecological and environmental problems such as vegetation growth, lake wetlands, desertification, and salinization. Environmental hydrogeology focuses on the relationship between groundwater level and environmental hydrogeological problems such as ground subsidence, seawater intrusion, water quality deterioration, groundwater drop funnel, and aquifer drainage. Agriculture focuses on the relationship between groundwater level and crop growth index and yield. The engineering field focuses on the corresponding groundwater burial depth and building safety groundwater level when tunnels and coal mines do not affect the surrounding ecological environment due to construction. According to different warning scopes, the selection of thresholds has different focuses. For example, the control target of over-exploitation areas is that the water level does not continue to decline, and the control target of key areas is to fluctuate within a reasonable water level range. After determining which indicators to use as the main considerations for water level thresholds, the warning values ​​can be determined using statistical and probability model methods, data-driven models or numerical model methods.

[0049] Furthermore, the workflow of the water level warning threshold determination module includes:

[0050] 1. Determine the early warning method according to the monitoring sequence duration of groundwater level (burial depth)

[0051] 1.1 If the continuous monitoring sequence is greater than 30 years (not less than the monthly scale), statistical methods (such as time series analysis, establishment of autoregressive model, etc.) can be used to directly construct the third data set and determine the water level warning threshold.

[0052] 1.1.1 Time series analysis is a method of predicting future values ​​by analyzing the trends, seasonal fluctuations, periodic changes, and random fluctuations of historical data series. For water level prediction, time series analysis can help identify the long-term trend of groundwater levels and predict future changes in water levels. The specific analysis steps include:

[0053] a) Data collection and preprocessing. Collect groundwater level data for at least 30 years and record them on a monthly or finer scale. Data preprocessing includes steps such as removing missing values, processing outliers, and stabilizing the data.

[0054] b) Data stabilization. Time series data usually needs to be stabilized to ensure that the data has a constant mean and variance. Common stabilization methods include differencing (such as first-order difference, second-order difference, etc.) and trend removal. After stabilization, time series are easier to model and analyze.

[0055] c) Model selection and fitting:

[0056] Autoregression (AR) model: If the time series is mainly affected by historical values, choose the autoregression model for modeling.

[0057] Moving average (MA) model: If the time series is greatly affected by random fluctuations, choose the moving average model for modeling.

[0058] ARMA model: For stable time series data, choose the ARMA model (autoregressive moving average model). The ARMA model combines the advantages of the AR and MA models, taking into account both the dependency of historical data and random noise.

[0059] ARIMA model: For non-stationary time series, the stationary series is obtained through difference processing, and then the ARMA model is applied. The ARIMA model (autoregressive integrated moving average model) is modeled after difference processing and is suitable for time series data with trends.

[0060] d) Model diagnosis and evaluation. For the fitted model, residual analysis is required to check whether the model meets the assumptions (such as residuals follow a normal distribution, variance homogeneity, etc.). Use indicators such as AIC (Akaike Information Criterion) and BIC (Bayesian Information Criterion) to compare and select models. If the model diagnosis results are not ideal, you can try to adjust the model parameters or use a more complex model (such as seasonal ARIMA).

[0061] e) Prediction and threshold determination. After the model is established and verified, the model is used to predict the future water level. The warning threshold of the groundwater level is determined by combining the prediction results and the statistical characteristics of historical data. For example, a safe water level range is determined based on the predicted water level change trend and fluctuation range, and different thresholds are set according to the requirements of the warning level.

[0062] 1.1.2 Autoregressive Model (AR) is a linear model that predicts future values ​​based on past observations. This model assumes that the current groundwater level is a linear combination of the previous water levels and is suitable for predictions when there is obvious autocorrelation in time series data. The basic form of the autoregressive model is as follows:

[0063]

[0064] In the formula, H t represents the groundwater level at time point t; is the autoregressive coefficient; p is the model order, indicating how many periods of historical water level data are used to predict the current water level; ∈ t is a random error term.

[0065] The specific analysis steps include:

[0066] a) Data preparation and stabilization. Check whether the time series is stable. If the original data has trend or seasonal fluctuations, differential processing is required to make the data stable.

[0067] b) Determine the order of the AR model. Use the ACF (autocorrelation function) and PACF (partial autocorrelation function) graphs to determine the order of the AR model. Determine the appropriate order by observing the decay trend of the ACF and PACF graphs. For example, if the PACF graph is truncated at the pth order, select p as the order of the AR model.

[0068] c) Model fitting and estimation. Estimate the autoregressive coefficients of the model using the least squares method (OLS) or maximum likelihood estimation (MLE) method. Use statistical software (such as R, Python, etc.) for estimation.

[0069] d) Model diagnosis and optimization. Check whether the residuals of the model are white noise (i.e., the residuals have no significant autocorrelation). If the model diagnosis fails, consider adjusting the order or using a more complex model (such as ARMA, ARIMA, etc.).

[0070] e) Determine water level prediction and threshold. Predict the future water level based on the fitted AR model. In practical applications, the warning threshold is determined based on the historical change trend and fluctuation range of the water level. For example, a water level warning threshold is set, and when the future water level prediction result is lower than the threshold, the water level alarm is triggered.

[0071] 1.2 If the continuous monitoring series is less than 30 years (not less than the monthly scale), the following aspects should be considered:

[0072] 1.2.1 It is necessary to consider factors such as the groundwater level to prevent salinization, the safe groundwater level of urban buildings in terms of building stability and foundation bearing capacity, and the groundwater recharge capacity, and construct the first data set to determine the upper limit water level;

[0073] 1.2.2 If the mining volume monitoring data is sufficient, the prediction method is considered according to the duration of the continuous monitoring sequence:

[0074] 1.2.2.1 If the continuous monitoring sequence is longer than 10 years (not less than the monthly scale), the multivariate regression analysis prediction method is used;

[0075] 1.2.2.2 If the continuous monitoring sequence is less than 10 years (not less than monthly scale), the mining volume-drawdown curve method shall be used;

[0076] 1.2.2.3 The multivariate regression analysis prediction method uses the relationship between multiple independent variables (prediction factors) in historical data and groundwater levels (dependent variables) to predict water level changes by establishing a regression model between groundwater levels and multiple related factors (such as precipitation, extraction volume, temperature, etc.).

[0077] The specific analysis steps include:

[0078] a) Data preparation;

[0079] b) Regression model establishment;

[0080] c) Predictive applications.

[0081] In the data preparation stage, it is necessary to collect at least 10 years of historical water level data and collect corresponding time series data based on factors that may affect the water level (such as precipitation, mining volume, temperature, etc.). Use statistical software (such as SPSS, R, Python, etc.) to perform multiple regression analysis, calculate the regression coefficient of each factor by the least squares method, and finally establish a regression equation.

[0082] The regression equation is of the form:

[0083] Ht=β0+β1X1+β2X2+…+β n X n (2)

[0084] Where Ht represents the groundwater level at time t; X1, X2…X n Indicates various factors of groundwater level, such as precipitation, extraction volume, etc.; β0 is a constant term, indicating the groundwater level when all independent variables are 0; β1, β2…β nis the regression coefficient, which indicates the influence of each factor on the groundwater level.

[0085] After the model is built, the latest monitoring data is substituted into the regression equation for water level prediction. This method requires that the input data have high quality and time consistency, and the results of the regression analysis need to be verified.

[0086] Common input parameters include historical data of groundwater levels, precipitation, air temperature, extraction, soil type, evaporation, etc., which should be recorded and monitored on a monthly or finer scale.

[0087] 1.2.2.4 The extraction volume-drawdown curve method is a prediction method based on the relationship between groundwater extraction volume and water level drawdown. This method assumes that there is a certain linear or nonlinear relationship between groundwater level drawdown and extraction volume, and predicts water level changes by establishing a curve model of extraction volume and water level drawdown. The specific analysis steps include:

[0088] a) Collect historical groundwater level data and corresponding extraction data over different time periods. These data should include the changes in water levels under different extraction conditions, usually recorded on a monthly basis. The groundwater level drawdown value (ΔH) is calculated from the difference between the initial water level at a certain point in time and the current water level at the corresponding point in time. The collected data should cover multiple different extraction intensities to ensure that the changing relationship between extraction and water level drawdown can be reflected.

[0089] b) Plot the relationship between extraction and water level drawdown. Based on the historical data collected, plot the relationship between extraction and water level drawdown. It is assumed that there is a mathematical relationship between the groundwater level drawdown (ΔH) and the extraction volume (Q). Usually, the relationship is assumed to be linear or nonlinear, depending on the hydrogeological conditions and extraction method of the prediction area.

[0090] If the relationship is linear, use linear regression method to establish the linear equation:

[0091] ΔH=a·Q+b (3)

[0092] Where ΔH is the water level drawdown, Q is the mining volume, and a and b are regression coefficients, which are obtained by fitting with the least squares method.

[0093] If the relationship is nonlinear, more complex models may be required, such as exponential regression, logarithmic regression, or power function regression. These nonlinear relationships are obtained through nonlinear least squares or other curve fitting techniques.

[0094] ① When the data presents a gradually intensified nonlinear relationship, the power function regression model is selected. The power function regression is suitable for situations where there is a power relationship between the mining volume and the water level drawdown, especially when the amplitude of the change between the water level drawdown and the mining volume gradually increases or decreases within a certain mining volume range, showing a relatively gentle or accelerated nonlinear change.

[0095] The power function regression model form is:

[0096] ΔH=A·Q B (4)

[0097] Where A and B are fitting parameters, Q is the extraction volume, and ΔH is the water level drawdown.

[0098] ② When the data show a trend of rapid changes in the early stage and slow changes in the later stage, the exponential regression model is selected. It can be used in the initial exploitation stage of the groundwater system or in areas with limited recharge.

[0099] Exponential regression model form:

[0100] ΔH=A·e B·Q (5)

[0101] Where A and B are the parameters to be fitted, e is the base of the natural logarithm, Q is the extraction volume, and ΔH is the water level drawdown.

[0102] ③ When the data shows large changes in the initial stage and gradually stabilizes in the later stage, the logarithmic regression model is selected. It is suitable for some cases where the water level drop shows large changes in the initial stage when the mining volume increases, but as the mining volume continues to increase, the water level changes gradually stabilize.

[0103] The logarithmic regression model form is:

[0104] ΔH=A+B·ln(Q) (6)

[0105] Where A and B are fitting coefficients, ln(Q) is the logarithm of the extraction volume, and ΔH is the water level drawdown.

[0106] ④ When the data has multiple change stages (for example, the water level change trend is different when the mining volume is low and high), the segmented regression model is selected. This model fits different regression models in different mining volume intervals, thereby more accurately reflecting the changing law of water level drawdown under different mining intensities.

[0107] The piecewise regression model form is:

[0108]

[0109] Where A1 and B1 are fitting coefficients, Q1 is the threshold of mining volume, and A2 and B2 are fitting coefficients of the other segment.

[0110] c) Based on the current extraction volume, predict the future water level changes. Once the relationship curve between extraction volume and water level drop is established, the actual extraction volume Q current Predict future water level changes. Substitute the current mining volume data into the established regression equation and calculate the corresponding water level drop value ΔH future This value is used as the change in groundwater level at a certain point in the future, thereby obtaining the predicted water level:

[0111] H future =H initail -ΔH future (8)

[0112] In the formula, H initial is the current water level, ΔH future It is the water level drawdown predicted based on the amount of extraction.

[0113] d) Verify and optimize the model

[0114] In practical applications, the production-drawdown curve model needs to be verified with new monitoring data to ensure the accuracy of the prediction results. If there is a large deviation between the actual observed water level and the predicted water level, it may be necessary to refit the curve, adjust the model parameters, or try different regression methods. In addition, if the change pattern of production changes (for example, a sudden increase in production or a continuous low production), the model should be updated in time to improve the prediction accuracy.

[0115] Input parameters typically include:

[0116] a) Groundwater level data (H): groundwater level observation data at different times, usually recorded on a monthly or quarterly basis, covering a sufficient period of time (preferably not less than 10 years). The groundwater level data should be accurate and consistent over time.

[0117] b) Extraction volume data (Q): Groundwater extraction volume data at each time point, usually on a daily or monthly basis, in cubic meters per day or cubic meters per month.

[0118] c) Drawdown (ΔH): The change in water level, i.e. the drop in groundwater level, usually the difference between the initial water level and the current water level.

[0119] d) Extraction time (T): The period of time during which groundwater is extracted, which helps determine the time response of extraction activities to water level changes.

[0120] e) Morphological parameters of the water level drawdown curve: The slope of the fitting curve and the nonlinear morphological parameters of the curve are used to accurately describe the relationship between extraction volume and water level changes.

[0121] f) Groundwater recharge (S) (if known): In some cases, groundwater recharge may affect the rate of water level change, with larger recharge rates slowing the rate of water level decline.

[0122] 1.2.3 If the mining volume monitoring data is insufficient, the prediction method shall be considered according to the basic monitoring type zoning:

[0123] 1.2.3.1 If the basic type area is a plain area and the continuous monitoring sequence is greater than 10 years (not less than the monthly scale), the water level forecast statistics shall be carried out in combination with the frequency statistics of rainfall (such as precipitation-water level regression analysis method, precipitation distribution method, etc.);

[0124] 1.2.3.2 If the basic type area is a plain area and the continuous monitoring sequence is less than 10 years (not less than the monthly scale), the sequence statistics method is used for water level prediction;

[0125] 1.2.3.3 If the basic type area is a non-plain area (such as a hilly area, etc.), the sequence statistical method (such as the five-color early warning method, time series method, extreme value analysis method, etc.) is directly used to predict the water level;

[0126] 1.2.3.4 Precipitation-water level regression analysis is a method to predict water level changes by establishing a statistical regression relationship between precipitation and water level. This method is based on the assumption that precipitation has a significant impact on groundwater level changes, especially in plain areas where precipitation is an important source of groundwater recharge. The specific analysis steps include:

[0127] a) Data collection and collation. Collect historical precipitation and groundwater level data, usually recorded monthly or quarterly. These data should cover a longer time frame (e.g. more than 10 years) to ensure that they can reflect the long-term trend of precipitation and water level.

[0128] b) Establish a regression analysis model. Based on the collected data, a relationship model between precipitation and water level is established through regression analysis methods (such as linear regression, polynomial regression, etc.). Commonly used model forms are as follows:

[0129] H t =a·P t +b (9)

[0130] In the formula, H t is the groundwater level at a certain point in time, P t is the precipitation at the corresponding time point, a and b are regression coefficients, representing the intensity and offset of the impact of precipitation on the water level.

[0131] c) Perform model testing and optimization. The results of regression analysis need to be tested to check the goodness of fit of the model (such as R 2If the model fit is poor or the residuals are large, consider using a more complex regression model, such as polynomial regression or piecewise regression.

[0132] d) Water level prediction. Once the regression model is established and tested, the groundwater level is predicted based on future precipitation. By inputting the future precipitation P future , calculate the predicted water level H future .

[0133] 1.2.3.5 The precipitation distribution method is a method for predicting water level changes based on the statistical distribution characteristics of precipitation. This method assumes that the change of precipitation in a certain period of time presents a certain probability distribution (such as normal distribution, exponential distribution, etc.), and predicts future water levels by analyzing the statistical characteristics of precipitation and combining historical water level data. Unlike regression analysis, the precipitation distribution method relies more on the probability distribution and statistical characteristics of precipitation, rather than directly modeling the specific functional relationship between precipitation and water level.

[0134] The specific analysis steps include:

[0135] a) Collect and analyze data. Collect historical precipitation data and record them by month or season. Calculate the distribution characteristics of precipitation, including the mean, standard deviation, skewness, kurtosis and other statistics of precipitation. By analyzing the frequency distribution of precipitation, determine the probability distribution type of precipitation, such as whether it conforms to normal distribution, gamma distribution or lognormal distribution.

[0136] b) Determine the distribution model of precipitation. Fit the probability distribution model of precipitation based on historical data. For example, if the precipitation data presents a normal distribution, use the normal distribution model for fitting; if the precipitation shows a right-skewed distribution (such as more extreme precipitation events), consider using the gamma distribution or lognormal distribution.

[0137] c) Predicting future precipitation. Once the distribution model of precipitation is determined, the model is used to predict future precipitation. Based on the fluctuations and trends of historical precipitation, the probability distribution is used to predict the possible range of future precipitation and its probability of occurrence. For example, Monte Carlo simulation and other methods are used to generate multiple possible future precipitation sequences based on the known distribution model.

[0138] d) Combine the water level change characteristics for prediction. Based on the relationship between historical water level data and precipitation, combined with the distribution prediction results of precipitation, predict future water level changes. This is done by calculating the water level response under different precipitation amounts. If there is a relationship between water level and precipitation (such as obtained through empirical formulas or historical regression analysis), substitute the prediction results of future precipitation into it to obtain the corresponding water level prediction value.

[0139] 1.2.4 Based on the water level prediction results, the second data set is constructed to determine the lower limit water level, taking into account the ecological warning groundwater level, the groundwater level to contain seawater intrusion, the groundwater level to prevent land subsidence (deep layer) and other head drop constraints;

[0140] 1.2.5 After determining the first data set and the second data set, merge them to construct the third data set and determine the water level warning threshold.

[0141] Furthermore, the station network selection module is used to select representative stations based on the water level warning threshold and groundwater type. The basic principle of station network selection: no matter what method is adopted for early warning, representative stations should be selected according to the warning range according to the plain area, hilly area, shallow groundwater and deep groundwater. The monitoring layer of the monitoring station used is clear and can reflect the main mining layer. If the density of monitoring stations in the early warning unit exceeds the upper limit of each level specified in the "Technical Standards for Groundwater Monitoring Engineering", the selected stations should be evenly distributed in the unit and moderately encrypted in the funnel area; if the density in the early warning unit does not reach the upper limit of each level, the evaluation station network should be determined according to the principle of selecting as much as possible. At the same time, the timeliness of monthly reporting should be considered. The selection of the evaluation station network is based on the principle of making full use of automatic monitoring stations, with high data quality, no continuous measurement loss for more than half a year, and the data sequence should be able to cover the evaluation period. After the representative station is selected, it should remain relatively fixed.

[0142] Furthermore, the water level change notification module is used to notify the groundwater level changes and issue a four-color logo for the water level changes in the over-exploitation area. The water level change notification module includes a data processing unit and a prediction unit; wherein the data processing unit is used to collect groundwater level monitoring data; the prediction unit is used to establish a groundwater level depth prediction model, and predict the groundwater level changes based on the groundwater level depth prediction model and groundwater level monitoring data. Notification of water level changes in groundwater over-exploitation areas: The national groundwater level change notification is based on the city (prefecture) as a unit, and the technical process includes six steps: determining the notification scope - determining the notification representative station - determining the water level benchmark value - determining the water level fluctuation and ranking scheme - notification of water level changes - application of notification results. After determining that the notification scope is 108 shallow and 37 deep prefecture-level administrative regions with over-exploitation problems, the representative station list is preliminarily determined by the method of map demarcation, combined with the development and utilization of the stratum, and then through data process line analysis, the stations with higher data quality are selected to finally determine the representative stations for notification. The arithmetic average of the daily average water level values ​​of all representative stations in the shallow or deep city (region) in the first four quarters of the previous year is taken to obtain the average water level of the reported city (region) in the first four quarters of the previous year, which is used as the benchmark value for groundwater level reporting.

[0143]

[0144] Where, X iis the quarterly average water level value of the ith representative station, and n is the number of representative stations in the reporting city (region). The same method is used to calculate the average water level of the 1st to 4th quarters of the reported city (region). The quarterly water level fluctuation is the average water level of the ×th quarter of the reported city (region) minus the water level reference value of the same quarter of the previous year, or the depth reference value of the ×th quarter of the previous year minus the average depth of the same quarter of the current year, to obtain the water level fluctuation of the ×th quarter of the reported city (region). A positive fluctuation indicates a rise in water level, and a negative fluctuation indicates a drop in water level. A four-color mark for water level changes in over-exploitation areas is issued based on the groundwater level fluctuation.

[0145] The advantages of the first generation notification scheme are simple methods, convenient review, and the ability to quickly obtain water level changes in cities. After three years of operation, five major problems have been found. First, the density of station network layout in some areas is not enough, the monitoring stations are not representative enough, and the scientific layout of the notification station network needs to be improved; second, the deep and shallow division standards are not unified, the notification calculation method is still relatively simple, and there may be calculation deviations in areas where representative stations are unevenly distributed; third, local monitoring data is not shared enough, the timeliness of compilation is poor, and the data quality needs to be further improved; fourth, the impact of annual actual precipitation on groundwater level changes is not fully considered. Annual precipitation has a direct impact on shallow groundwater levels and groundwater extraction. In particular, in extremely abundant years, the groundwater level rises significantly. Once the groundwater level is dry in the next year, the water level drop is inevitable under the same exploitation conditions, which cannot reflect the inter-annual regulation of groundwater; fifth, the current notification mechanism uses a water level drop of more than 1 meter as a unified standard for naming, which will cause some areas with small annual water level drops but large cumulative drops to miss the net.

[0146] After research, the fourth and fifth problems were solved by adjusting the technical plan for the notification. While maintaining the advantages of the original notification in terms of strong operability, easy understanding and ease of use, the following two problems were focused on. First, the impact of inter-annual precipitation was reduced, and the water level drop caused by the sudden change from wet to dry season was not included in the scope of consultation. Second, the cities and counties with less drop in water level each quarter but with cumulative water level drop problems were identified.

[0147] Based on the original method of "water level changes between the current quarter and the same period of the previous year", after comprehensively comparing the "base year year-on-year changes", "cumulative water level changes", "sliding cumulative water level changes" and "quarterly average annual water level changes", starting from the first quarter of 2024, the "three-year sliding cumulative water level changes" indicator will be added to enhance the scientificity and accuracy of the reported results, and fully reflect the regulating role of groundwater.

[0148]

[0149] Where: m is the current year, n = m-2; It is the sum of the year-on-year changes in water levels from notification year n to notification year m. In order to reduce the impact of station network changes on the data, only the same station network is used for calculation.

[0150] In terms of the application of the results of the report, the naming standard was adjusted from "quarterly water level dropped by more than 1 meter year-on-year" to "quarterly water level dropped by more than 1 meter year-on-year or the cumulative water level dropped by more than 2 meters in the past three years", and the consultation conditions were adjusted to being named for two consecutive quarters or the cumulative water level dropped by more than 3 meters in the past three years, and the cities where the water level rose by more than 1 meter in the past three years were excluded. Cities where the trend of groundwater level decline has not improved or even worsened after the consultation, or where serious ecological and geological environmental problems have occurred, will be promptly included in the scope of the interview.

[0151] Groundwater level early warning information in key areas: The early warning scope is the Beijing-Tianjin-Hebei over-exploitation control area. Beijing and Tianjin use district (county) level administrative areas as the basic unit, and Hebei Province uses prefecture-level administrative areas as the basic unit for early warning. Beijing does not participate in deep water level early warning. The calculation units with monitoring stations in other regions participate in early warning. There are 33 shallow administrative areas and 23 deep administrative areas. The year-on-year change in groundwater level is used as the early warning indicator, and a two-level early warning mode is adopted, including over-warning and normal. A water level drop of more than 0.5 meters is considered over-warning. Later, during the actual early warning process, three adjustments were made. First, considering that the areas of the analysis units of the district (county) level administrative regions in Beijing and Tianjin and the prefecture-level administrative regions in Hebei are quite different, it is not appropriate to analyze water level changes at the same level. Therefore, each district in Beijing and Tianjin was merged into one calculation unit, and Hebei Province still used the prefecture-level administrative region as the basic unit. Second, considering the consistency of the results, the station network was adjusted to be consistent with the station network used for notification or assessment in various provinces. Third, in order to explain the reasons for the changes in groundwater levels, a comparative analysis of monthly precipitation and cumulative precipitation with last year and the same period of many years was added to illustrate the impact of precipitation on groundwater levels.

[0152] The warning information display module is used to visually display the warning results in the form of a diagram; the warning results include a rainbow warning mode and a four-color warning diagram.

[0153] Improved early warning scheme: Several Opinions on Establishing a Long-term Mechanism for Monitoring and Early Warning of Resource and Environmental Carrying Capacity clearly states that resource and environmental carrying capacity is divided into three levels: overload, critical overload, and no overload. According to the degree of aggravation and slowdown of resource and environmental depletion, the overload level is further divided into two warning levels: red and orange, the critical overload level is divided into two warning levels: yellow and blue, and the no overload level is determined as a green warning level. The warning levels are red, orange, yellow, blue, and green from high to low. According to this document, the improved early warning scheme monitoring and early warning adopts the red, orange, yellow, blue, and green five-color early warning method, and the indicators are determined by comparing with the extreme values ​​of the multi-year series, the average values ​​of the multi-year series, and the data of the same period last year.

[0154] The calculation method uses the monthly average of representative stations, that is, the arithmetic mean of the daily values ​​reported in the current month. The monthly average calculation method for prefecture-level cities can be the arithmetic mean of representative stations or the area-weighted method, and the consistency of the calculation method must be maintained. The establishment of the ledger comprehensively considers the official operation time of the national groundwater monitoring project, the operation of local stations, precipitation and data compilation, etc. According to the actual situation of various cities, 2018-2022, 2019-2023 or 2018-2023 are selected as the warning benchmark value sequence, and the monthly average of representative stations in each month of the sequence is calculated to form a multi-year series of mean and extreme value (maximum burial depth, minimum burial depth) ledgers for prefecture-level cities. The ledger can be adjusted annually as needed.

[0155] The preliminary revised warning plan adopts the "rainbow warning" model, such as Figure 3 As shown, the warning results are intuitively displayed in the form of a graph. Later, considering the continuity of the same period last year, the scheme was adjusted as follows: red indicates that the water level of the current month is lower than the lowest water level in the same period for many years, orange indicates that the water level of the current month is more than 1 meter lower than the average water level of the same period for many years, yellow indicates that the water level of the current month is 1 meter or less lower than the average water level of the same period for many years or the water level of the current month is more than 1 meter lower than the water level of the same period last year, blue indicates that the water level of the current month is higher than the average water level of the same period for many years, but lower than the highest water level of the same period for many years, and green indicates that the water level of the current month is higher than the highest water level of the same period for many years. Among them, red and orange are water level drop warnings, and yellow is a water level drop reminder. When calculating, first determine the warning level with the extreme values ​​and averages of the multi-year series, and then compare with the water level data of the same period last year. The water level of the current month is adjusted to yellow if it is more than 1 meter lower than the water level of the same period last year.

[0156] The prediction and early warning module is used to predict the change of groundwater level. Prediction and early warning scheme: In North China, comprehensive multivariate regression, neural network and other methods are used to increase the prediction and early warning content. Taking the Beijing-Tianjin-Hebei prefecture-level administrative regions as units, the trend of groundwater level decline or rise in each prefecture-level administrative region next month is predicted, and corresponding early warnings are issued. First, a groundwater level depth prediction model is constructed. For the groundwater level depth of the current month, three influencing factors are mainly considered for shallow groundwater: the groundwater level depth of the previous month, the monthly precipitation, and the groundwater extraction volume of the current month. Since deep confined water has no direct connection with precipitation, the main influencing factors of the deep layer are the groundwater level depth of the previous month and the groundwater extraction volume of the current month. The water level depth of the previous month is the average value of the observations at the end of the month of all stations in the study area; the groundwater extraction volume of the current month is allocated to each month according to the proportion of the annual groundwater extraction volume and the statistical data of agricultural irrigation, industrial production and domestic water use in previous years; the monthly precipitation adopts the results of medium- and long-term numerical forecasts of meteorology. After using the multivariate regression model and the neural network RBF model (Radial Basis Function) to establish the shallow and deep groundwater models respectively, the predicted burial depths of each municipal administrative area were compared with the measured burial depths on a monthly basis, and error analysis was performed to determine the model schemes and model parameters selected by each city.

[0157] The present invention selects a more suitable water level threshold determination method based on the monitoring length and the evaluation unit type, while considering the upper and lower water levels of the evaluation unit. In areas where there is generally no monitoring data and extraction data for more than 30 years, the selection of sequence statistics and rainbow warning signs can reduce the impact of inter-annual precipitation and make more reasonable warning results. The core module is highly operational and easy to understand, and can clearly and intuitively represent the current groundwater level status.

[0158] Embodiment 2

[0159] In this embodiment, a groundwater level warning method is provided, comprising the following steps: determining a water level warning threshold in an over-exploited area; selecting representative stations for shallow groundwater and deep groundwater respectively based on the water level warning threshold; calculating water level changes based on the selected representative stations to obtain warning information; and predicting water level changes to obtain groundwater level prediction information.

[0160] Furthermore, the process of predicting water level changes to obtain groundwater level prediction information includes: using a multivariate regression model and a neural network RBF model to respectively establish a shallow groundwater prediction model and a deep groundwater prediction model; determining the groundwater level prediction information of each municipal administrative area based on the shallow groundwater prediction model and the deep groundwater prediction model; comparing the groundwater level prediction information with the measured burial depth on a monthly basis to obtain an error value; and optimizing the shallow groundwater prediction model and the deep groundwater prediction model based on the error value to determine the model scheme and model parameters selected by each municipality.

[0161] As a specific implementation method of this embodiment, the forecast and early warning scheme: In North China, comprehensive multivariate regression, neural network and other methods are used to increase the forecast and early warning content. Taking the Beijing-Tianjin-Hebei prefecture-level administrative regions as units, the trend of the decline or rise of the groundwater level in each prefecture-level administrative region next month is predicted, and corresponding early warnings are issued. First, a groundwater level depth prediction model is constructed. For the groundwater level depth of the current month, three influencing factors are mainly considered for shallow groundwater: the groundwater level depth of the previous month, the precipitation of the current month, and the groundwater extraction volume of the current month. Since deep confined water has no direct connection with precipitation, the main influencing factors of the deep layer are the groundwater level depth of the previous month and the groundwater extraction volume of the current month. The water level depth of the previous month is the average value of the observations at the end of the month at all stations in the study area; the groundwater extraction volume of the current month is allocated to each month according to the proportion of statistical data such as agricultural irrigation, industrial production and domestic water use in previous years according to the total annual groundwater extraction volume; the precipitation of the current month adopts the results of medium- and long-term numerical forecasts of meteorology. After using the multivariate regression model and the neural network RBF model (Radial Basis Function) to establish the shallow and deep groundwater models respectively, the predicted burial depths of each municipal administrative area were compared with the measured burial depths on a monthly basis, and error analysis was performed to determine the model schemes and model parameters selected by each city.

[0162] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A groundwater level early warning system, characterized in that: It includes: water level warning threshold determination module, station network selection module, water level change notification module and prediction and warning module; Wherein, the water level warning threshold determination module is used to determine the water level warning threshold of the overexploitation area; The station network selection module is used to select representative stations based on the water level warning threshold and groundwater type; The water level change calculation module is used to calculate the groundwater level change; The prediction and warning module is used to issue water level change warning signs and predict groundwater level changes.

2. The groundwater level early warning system according to claim 1, characterized in that: The groundwater level early warning system also includes an early warning information display module, which is used to visually display early warning results in the form of graphics; the early warning results include a rainbow early warning mode and a four-color identification early warning schematic diagram.

3. The groundwater level early warning system according to claim 1, characterized in that: The water level change calculation module includes a data processing unit and a prediction unit; Wherein, the data processing unit is used to collect groundwater level monitoring data The prediction unit is used to establish a groundwater level depth prediction model, and predict groundwater level changes based on the groundwater level depth prediction model and the groundwater level monitoring data.

4. A groundwater level early warning method, characterized in that: For implementing the groundwater level early warning system according to claim 1, the method comprises the following steps: Determine the water level warning threshold in the assessment area; Based on the water level warning threshold, representative stations are selected according to shallow groundwater and deep groundwater respectively; Based on the selected representative stations, calculate the water level changes to obtain early warning information; The water level change is predicted to obtain groundwater level prediction information.

5. The groundwater level early warning method according to claim 4, characterized in that: The process of determining the water level warning threshold for the assessment area includes: Select the early warning method based on the duration of the groundwater level monitoring sequence in the evaluation area; When the monitoring sequence is longer than 30 years, a statistical method is used to directly construct a third data set, and the water level warning threshold is determined based on the directly constructed third data set; When the monitoring sequence is less than 30 years, the first data set is constructed based on the groundwater level factor to determine the upper limit water level; and the second data set is constructed based on the mining volume monitoring data to determine the lower limit water level; The first data set and the second data set are merged to indirectly construct a third data set, and a water level warning threshold is determined based on the indirectly constructed third data set.

6. The groundwater level early warning method according to claim 5, characterized in that: The groundwater level factors include: groundwater level to prevent salinization, safe groundwater level for urban buildings for building stability and foundation bearing capacity, and groundwater recharge capacity.

7. The groundwater level early warning method according to claim 4, characterized in that: The process of predicting the water level change to obtain groundwater level prediction information includes: The shallow groundwater prediction model and deep groundwater prediction model were established by using multivariate regression model and neural network RBF model respectively. Determine the groundwater level prediction information of each city-level administrative area based on the shallow groundwater prediction model and the deep groundwater prediction model; Comparing the groundwater level prediction information with the measured burial depth on a monthly basis to obtain an error value; The shallow groundwater prediction model and the deep groundwater prediction model are optimized based on the error value to determine the model scheme and model parameters selected by various cities.

8. The groundwater level early warning method according to claim 4, characterized in that: The factors affecting the establishment of the shallow groundwater prediction model include: the depth of groundwater level in the previous month, the precipitation in the current month, and the groundwater extraction volume in the current month; The factors affecting the establishment of the deep groundwater prediction model include: the groundwater level depth in the previous month and the groundwater extraction volume in the current month.

9. A computer terminal device, characterized in that: include: one or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the groundwater level early warning system as described in any one of claims 1-3.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the groundwater level early warning system according to any one of claims 1 to 3 is implemented.

Images