A method for extracting driving factors of groundwater level based on variogram combined with transfer function

Through the combined transfer function method of variation analysis, the problem of insufficient in-depth exploration of the long-term variation characteristics of groundwater level and impact distinction in the existing technology is solved, and the accurate analysis of groundwater level changes is achieved. It is suitable for a wide range of areas and provides a scientific basis for groundwater resource management. .

CN118364039BActive Publication Date: 2024-11-01CAPITAL NORMAL UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410581120.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-11
Publication Date
2024-11-01
Estimated Expiration
2044-05-11

AI Technical Summary

Technical Problem

The existing technology has failed to explore the long-term variation characteristics of groundwater level, and it is difficult to distinguish the impact of rainfall and groundwater extraction on groundwater level. Mathematical statistical methods are difficult to accurately quantify the contribution of the two to groundwater level changes. Traditional physical models are driven by climate change and human activities. The groundwater mining area below is difficult to apply.

Method used

The method of variance analysis combined transfer function is adopted to obtain the regional long-time series hydrological data, calculate the interannual variation coefficients of groundwater level and rainfall, identify the variation points under different time scales, construct the rainfall-groundwater level transfer function, and strip off the influence of artificial mining. , quantitatively describe the response time of groundwater levels to rainfall.

Benefits of technology

It improves the accuracy and accuracy of groundwater level change analysis, can more effectively distinguish the impact of rainfall and groundwater extraction on groundwater level, and is suitable for a wide range of time and space, providing scientific basis for groundwater resource management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118364039B_ABST
    Figure CN118364039B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for extracting driving factors of groundwater level by using a variation analysis combined transfer function, and relates to the technical field of groundwater resource protection; the method comprises the following steps: S1, obtaining regional long-term series hydrological data; S2, calculating the interannual variation coefficient of groundwater level and the interannual variation coefficient of rainfall, and comparing and judging the interannual variability of rainfall and groundwater level; S3, using the Pettitt method and the sliding T test method to respectively identify groundwater level and rainfall variation points, and using the jump variation classification method to describe the variation degree of the variation points; S4, performing stabilization processing on the rainfall data and groundwater level data of each subsequence of each monitoring point; S5, constructing a rainfall-groundwater level transfer function, and extracting the main driving factors affecting the change of groundwater level in different subsequences according to the response time of different subsequences of each monitoring point, and combining the overall groundwater extraction data of the region. The present invention helps to clearly distinguish the impact of rainfall and groundwater extraction on groundwater level.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of groundwater resource protection, and in particular relates to a method for extracting groundwater level driving factors by using a variation analysis combined with a transfer function. Background Art

[0002] Groundwater is the main source of fresh water. Excessive exploitation of groundwater leads to a sharp drop in groundwater levels, which in turn triggers geological and environmental problems such as ground subsidence, ground fissures, and seawater intrusion, affecting the healthy development of regional economy and ecology. Understanding the long-term dynamic changes of groundwater levels and revealing the causal mechanism of groundwater level evolution are crucial for the scientific management of groundwater resources. Previous studies have shown that rainfall and groundwater exploitation are the main reasons affecting the dynamic changes of groundwater, and climate change and human activities affect the changes in groundwater levels to varying degrees. In this context, identifying long-term groundwater level variation points, extracting subsequences whose trends show the same behavior, and quantifying the response time of groundwater levels in different subsequences to rainfall and groundwater exploitation can more effectively manage groundwater resources.

[0003] Generally, the analysis of groundwater level changes mainly includes three aspects: the first method is to use trend detection to explore the trend of groundwater level changes and compare it with the trends of rainfall and groundwater exploitation. However, this method cannot accurately and quantitatively determine the impact of rainfall and exploitation on groundwater level changes. In addition, the evolution process of groundwater is often complex, and the impact of trend changes may only be part of the long-term evolution of hydrological series. Moreover, trend analysis may have different results for different time periods, which is not suitable for long-term groundwater level analysis. The second method is to use variation point detection to identify groundwater level variation points and evaluate the impact of rainfall on groundwater levels by the time interval with rainfall variation points. However, in areas with severe artificial exploitation, this calculation method may have large errors. In addition, the calculation accuracy and test results of variation points obtained by different detection methods may be inconsistent, and the accuracy of variation points calculated in areas with severe artificial exploitation is difficult to guarantee. The third method is to use physical models to simulate the recharge and infiltration process of groundwater, but the scope of application of the model is limited due to the complex representation of model parameters and the influence of subjective assumptions.

[0004] Considering the complexity of the long-term evolution of groundwater levels, the coefficient of variation can solve the problem that groundwater level changes cannot be detected by the average value of a single year; jump variation analysis can identify mutations in time series, thereby splitting time series into subsequences with the same trend pattern; based on the analysis of the impact of rainfall and groundwater exploitation on groundwater level evolution based on variation points, it can strip off the impact of groundwater exploitation and obtain the response time of groundwater level to rainfall. Cross-correlation analysis can reveal the random dependence of groundwater level and rainfall, limit the interference caused by subjective assumptions, and is suitable for quantitative evaluation of the response relationship between groundwater level and rainfall.

[0005] In summary, the research on groundwater level has limitations. First, the existing methods have not conducted in-depth research on the long-term groundwater level variation characteristics. Second, the existing methods have failed to effectively distinguish the impact of rainfall and groundwater exploitation on groundwater level, and mathematical statistics methods are difficult to accurately quantify the contribution of the two to groundwater level changes. In addition, in groundwater exploitation areas driven by climate change and human activities, traditional physical models are difficult to apply. Therefore, it is crucial to deeply explore the long-term variation characteristics of groundwater levels, quantify the response relationship of groundwater levels in different subsequences to rainfall and groundwater exploitation, and extract the main driving factors affecting groundwater level changes for the scientific management of groundwater resources. The present invention proposes a method to extract the main driving factors of groundwater level changes. The method can directly use groundwater and rainfall data, regard the aquifer as a filter, rainfall as an input signal, and groundwater level as an output signal to obtain the response time of groundwater level to rainfall. By determining the response time of different time periods, the impact of rainfall and human exploitation on groundwater level fluctuations is further analyzed. Provide a scientific basis for the decision-making of water resources management departments. Summary of the invention

[0006] The purpose of the present invention is to provide a method for extracting groundwater level driving factors by using variation analysis combined with transfer function, so as to solve the problems that the existing methods proposed in the above background technology have not conducted in-depth exploration of the long-term groundwater level variation characteristics, and the existing methods have failed to effectively distinguish the impact of rainfall and groundwater exploitation on groundwater levels, and mathematical statistics methods are difficult to accurately quantify the contribution of the two to groundwater level changes.

[0007] To achieve the above object, the present invention adopts the following technical solutions:

[0008] A method for extracting groundwater level driving factors by using variation analysis and transfer function, comprising the following steps:

[0009] S1. Obtain regional long-term series hydrological data, including groundwater depth data, groundwater level data of each monitoring point, and rainfall data of each monitoring point;

[0010] S2. Based on the hydrological data, the interannual maximum fluctuation amplitude sequence of groundwater level and the interannual maximum fluctuation amplitude sequence of rainfall at each monitoring point are calculated, and the interannual variation coefficient of groundwater level and the interannual variation coefficient of rainfall at each monitoring point are calculated, and the interannual variation coefficient of groundwater level and the interannual variation coefficient of rainfall are compared to determine the interannual variability of regional long-term series rainfall and groundwater level;

[0011] S3. The Pettitt method and sliding T test method were used to identify groundwater level variation points and rainfall variation points at different time scales, respectively. The hydrological series jump variation classification method based on correlation coefficient was used to quantitatively describe the degree of variation of the time series variation points. At the same time, the groundwater level data and rainfall data of each monitoring point were divided into subsequences with the same behavior based on the groundwater level variation points.

[0012] S4, performing stabilization processing on the rainfall data and groundwater level data of each subsequence of each monitoring point, wherein the stabilization processing includes using a linear regression equation combined with a power spectrum to remove the trend and seasonality of the subsequence, and using a harmonic analysis method to remove the periodicity of the subsequence;

[0013] S5. Construct a rainfall-groundwater level transfer function based on the smoothed rainfall data and groundwater level data, wherein the rainfall-groundwater level transfer function is used to describe the response time of rainfall and groundwater level in the time domain; extract the main driving factors affecting the groundwater level change in different subsequences according to the response time of different subsequences of each monitoring point and in combination with the overall groundwater extraction data of the region.

[0014] Preferably, S2 specifically includes the following steps:

[0015] S201. Based on the regional long-term series groundwater depth data, the difference between the groundwater depths of two consecutive years is calculated to evaluate the fluctuation range of the groundwater level and determine whether the groundwater level in the region is affected by rainfall and groundwater exploitation factors.

[0016] S202. If the groundwater level calculated in S201 fluctuates violently, it is determined that the groundwater level in the region is affected by rainfall factors and groundwater exploitation factors; based on the groundwater level data of each monitoring point in the regional long-term series, the interannual maximum fluctuation amplitude sequence of the groundwater level is calculated, and the interannual variation coefficient of the groundwater level at each monitoring point is calculated using the coefficient of variation calculation formula; based on the rainfall data of each monitoring point in the regional long-term series, the interannual maximum fluctuation amplitude sequence of rainfall is calculated, and the interannual variation coefficient of rainfall at each monitoring point is calculated using the coefficient of variation calculation formula;

[0017] S203. Compare the interannual variation coefficient of groundwater level and the interannual variation coefficient of rainfall. There are monitoring points where the interannual variation coefficient of groundwater level is greater than 0.3, indicating that the interannual variation of groundwater level in this area is relatively large. Correspondingly, there are monitoring points where the interannual variation coefficient of rainfall is less than 0.3 and is evenly distributed in the area, indicating that rainfall has little impact on the variation of groundwater level in this area.

[0018] Preferably, the inter-annual maximum fluctuation amplitude sequence of groundwater level is calculated in S202, and the inter-annual coefficient of variation of groundwater level at each monitoring point is calculated using the coefficient of variation calculation formula, which is specifically as follows:

[0019] Based on the groundwater level data of each monitoring point in the regional long-term series, the monthly groundwater level fluctuation amplitude of each monitoring point is calculated and its extreme value is subtracted, calculated year by year, and the annual maximum fluctuation amplitude sequence of the groundwater level of each monitoring point is obtained by merging.

[0020] The interannual coefficient of variation of groundwater level at each monitoring point is calculated. The coefficient of variation calculation formula is as follows:

[0021] (1)

[0022] in, is the coefficient of variation, is the average value of the values ​​in the sequence, is the standard deviation.

[0023] Specifically, based on the groundwater level data of each monitoring point in the regional long-term series, the monthly average groundwater level of each monitoring point is obtained, the monthly average groundwater levels of two adjacent months are subtracted to obtain the annual groundwater level variation, the maximum and minimum values ​​of the annual groundwater level variation of each monitoring point are obtained and the two are subtracted to obtain the maximum fluctuation amplitude of the groundwater level at each monitoring point in that year. The maximum fluctuation amplitude of the groundwater level of each monitoring point is calculated year by year, and the results are combined to obtain the inter-annual maximum fluctuation amplitude sequence of the groundwater level at each monitoring point.

[0024] Preferably, the interannual maximum fluctuation amplitude sequence of rainfall is calculated in S202, and the interannual variation coefficient of rainfall at each monitoring point is calculated using the variation coefficient calculation formula, which is obtained in the same way as the interannual variation coefficient of groundwater level, as follows:

[0025] Based on the rainfall data of each monitoring point in the regional long-term series, the monthly rainfall fluctuation amplitude of each monitoring point is calculated and its extreme value is subtracted, calculated year by year, and the interannual maximum fluctuation amplitude sequence of rainfall at each monitoring point is obtained; finally, the interannual variation coefficient of rainfall at each monitoring point is calculated using the coefficient of variation calculation formula.

[0026] Specifically, based on the rainfall data of each monitoring point in the regional long-term series, the monthly rainfall of each monitoring point is obtained, and the monthly rainfall of two adjacent months is subtracted to obtain the annual monthly rainfall variation. The maximum and minimum values ​​of the annual monthly rainfall variation of each monitoring point are obtained and the two are subtracted to obtain the maximum fluctuation range of rainfall at each monitoring point in that year. The maximum fluctuation range of rainfall each year is calculated year by year for each monitoring point, and the results are combined to obtain the inter-annual maximum fluctuation range sequence of rainfall at each monitoring point.

[0027] Preferably, S3 specifically includes the following steps:

[0028] S301. Use the Pettitt method to identify groundwater level variation points and rainfall variation points at different time scales. The calculation process of the Pettitt method test is as follows:

[0029] (2)

[0030] (3)

[0031] in, is a statistical variable; n is the sample size; time series ; is a sign function whose value is -1, 0 or 1;

[0032] The mutation point is defined as :

[0033] (4)

[0034] The significance level of the variation point is determined by:

[0035] (5)

[0036] When the significance level P When <0.05, the detected variation point is considered statistically significant;

[0037] S302, using the sliding T test method to identify groundwater level variation points and rainfall variation points at different time scales. The sliding T test method determines whether there is a jump variation point by comparing the observed value of the sequence time point with the mean and standard deviation of the surrounding time points, and calculates the significance level of the variation point; the sliding T test method is calculated as follows:

[0038] Define the statistics:

[0039] (6)

[0040] In the formula, The mean and variance of the two subsequences are and , and ,in

[0041] (7)

[0042] In the formula, for Sample size time series , artificially set a certain moment as the reference point, and the two subsequences before and after the reference point , The samples are and , Formula (7) complies with the degree of freedom t distribution;

[0043] S303. Use the hydrological series jump variation classification method based on correlation coefficient to quantitatively describe the variation degree of the time series variation point. The calculation formula is as follows:

[0044] (8)

[0045] (9)

[0046] (10)

[0047] For a sequence of length n , split the entire sequence into and The mean of the whole sequence is The means of the two subsequences are and ; Construct a sequence consisting of jump components ,Right now ; , then the mean of the jump component sequence is and Equal, that is, they are both the means of the original sequences, expressed as:

[0048] (11)

[0049] When a sequence When a jump mutation occurs, its jump component The higher the proportion, by The greater the impact, The more it reflects the jumping variation characteristics, the stronger the correlation between the two, that is, the larger the correlation coefficient; the calculation formula is as follows:

[0050] (12)

[0051] right = For hypothesis testing, the commonly used significance levels are 0.01, 0.02, 0.05, and 0.1. When conducting hydrological analysis, the data length is required to be greater than or equal to 20. ;like Failed to pass the significant level The test classifies it as having no variation; When , it is classified as a giant variation, and the correlation coefficient is between 0 and 1;

[0052] S304. Based on the groundwater level data at different time scales, the groundwater level data and rainfall data of each monitoring point are divided into two subsequences with the same behavior through the groundwater level variation points with significant levels.

[0053] Specifically, the same behavior specifically refers to decomposing the sequence into two subsequences with the mutation point as a node, and the two subsequences have the same trend, either rising or falling.

[0054] Preferably, the harmonic analysis method is used in S4 to remove the periodicity of the sequence, as follows:

[0055] After removing the trend component, the periodic term analysis is performed and the periodic component is extracted by harmonic analysis. ,use The estimated value is expressed in the form of superposition of harmonics:

[0056] (13)

[0057] Where: For sequence An estimated value of is the number of harmonics, take The integer part of ; is the harmonic number, ; , is The period of a harmonic, n For sequence The length of the sequence;

[0058] , are the Fourier coefficients: ; ;like , then it is considered that The wave is significant, otherwise it is not significant; for The variance of is the significance level of the test; substitute the significant wave into the formula In this way, we get the estimated equation of the periodic term.

[0059] Preferably, the S5 is specifically as follows:

[0060] The aquifer is regarded as a filter, with rainfall data as the input signal and groundwater level data as the output signal. The cross-correlation function is used to quantitatively describe the response time of rainfall and groundwater level in the time domain. If rainfall is regarded as a random process, the cross-correlation function is the transfer function. The calculation process of rainfall-groundwater level transfer function is as follows:

[0061] Length is sequence The cross-correlation between is defined as:

[0062] (14)

[0063] In the cross-correlation function graph, the lag time corresponding to the maximum peak is defined as the response time of the system;

[0064] (15).

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] (1) The method of the present invention uses the coefficient of variation to characterize the long-term evolution characteristics of the regional hydrological sequence. Previous groundwater level change analysis methods usually only use trend detection to characterize the evolution trend of the groundwater level. The present invention calculates the interannual variability of the long-term groundwater level and rainfall based on the coefficient of variation, thereby preliminarily determining the overall variability of the regional hydrological sequence.

[0067] (2) The method of the present invention combines multiple variation point detection methods and introduces correlation coefficients to quantitatively identify jump variation points. In previous variation analysis methods, there are differences in the accuracy and test results of variation points obtained using different detection methods, and it is impossible to quantitatively describe the degree of variation of variation points. The present invention selects multiple variation point identification methods and conducts comprehensive evaluation based on actual conditions. Subsequently, correlation coefficients are introduced to grade and quantify the identified variation points, thereby further accurately identifying variation points in the hydrological sequence. Specifically, by combining the correlation coefficient-based hydrological sequence jump variation classification method and the Pettitt and sliding T tests, the variation points of the hydrological sequence are quantitatively identified, thereby improving the accuracy and precision of variation point identification.

[0068] (3) The method of the present invention constructs a rainfall-groundwater level transfer function based on variation points. Traditional physical models for simulating groundwater are often affected by the complexity of parameters and subjective assumptions, which limits their applicability in different regions. In particular, in areas where groundwater is severely exploited, directly constructing a transfer function may result in large errors due to human interference. Based on variation point analysis, the present invention removes the impact of artificial groundwater exploitation, simulates the process of rainfall to groundwater level changes at the water cycle level, and constructs a transfer function of rainfall and groundwater level to calculate the response time of groundwater level to rainfall, thus overcoming the problem of parameter complexity of traditional physical models. The present invention quantitatively identifies the response time of regional long-term series groundwater level to rainfall by combining variation analysis and transfer function, which helps to clearly distinguish the impact of rainfall and groundwater exploitation on groundwater level, and provides an important basis for the scientific management of groundwater resources. Moreover, through this method, the degree of influence of rainfall and groundwater exploitation on groundwater level fluctuations in different subsequences can be finally determined, thereby extracting the main driving factors affecting groundwater level changes in different subsequences.

[0069] (4) The transfer function method based on variation analysis in the method of the present invention has a wider range of applications. Due to the complex stratigraphic structure, the application area of ​​the traditional physical model is limited. In contrast, this method can better characterize the random dependence between groundwater level and rainfall, so that the groundwater level can be analyzed over a long span and a large area. Therefore, it has a wider applicability in time and space.

[0070] (5) The method in this invention is based on public data platforms, such as the Water Resources Bulletin and GEE, and establishes an effective technical method for extracting the main driving factors affecting groundwater levels, especially in groundwater overexploitation areas. This invention reduces the cost of data acquisition, can accurately quantify the long-term groundwater level evolution characteristics, and extract the main driving factors of different subsequences, providing strong technical support for the healthy development of regional groundwater. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a flow chart of a method for extracting groundwater level driving factors by using a variation analysis combined with a transfer function in the present invention;

[0072] Figure 2 This is a result diagram of the rainfall random process involved in the transfer function in Example 1 of the present invention. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] Embodiment 1:

[0075] The groundwater level driving factor extraction method based on the variability analysis and transfer function is used to extract the main driving factors affecting the groundwater level change. This method can obtain the groundwater level change information over a long period of time and a large area, and extract the main influencing factors.

[0076] See also Figure 1 Taking a certain experimental area as an example, a long-term series study of groundwater level was conducted. The groundwater depth data is 37 years long (1986-2021), and the rainfall and groundwater level data cover 22 years (2000-2021). The specific implementation of the method is as follows:

[0077] Step 1: Calculate and compare the interannual variability of regional long-term series rainfall and groundwater levels using the coefficient of variation.

[0078] S1.1, groundwater level fluctuation amplitude and interannual variability;

[0079] Obtain regional long-term series groundwater depth data, calculate the difference between groundwater levels in two consecutive years, and thus evaluate the groundwater fluctuation amplitude as a whole. This step can accurately reflect the fluctuation of groundwater levels, thus providing important information for studying the interannual variability of groundwater levels. If the groundwater level itself does not fluctuate strongly, it means that the groundwater level in the area is very stable; if the fluctuation is very violent, it means that the groundwater level fluctuation in the area is greatly affected by external factors, mainly natural factors (rainfall) and human factors (groundwater exploitation), and then further analysis is carried out on each monitoring well.

[0080] S1.2, coefficient of variation ( ) can reflect the interannual changes in the maximum fluctuation amplitude of groundwater levels to solve the problem that groundwater level variation cannot be detected by the average value of a single year.

[0081] Interannual variability of groundwater level:

[0082] The monthly average groundwater level data is obtained for each monitoring well in the area, and the interannual coefficient of variation of the groundwater level (CV_G) is calculated. The details are as follows:

[0083] For each monitoring well, first calculate the difference between the extreme values ​​of the monthly average groundwater level (GWL) each year to obtain the maximum annual fluctuation amplitude sequence of the GWL. In this embodiment, the groundwater level data is monthly data, and this monthly data is composed of the average value of each month (the groundwater level data also includes the monthly maximum and monthly minimum values, and the present invention uses the monthly average data). Calculate the annual groundwater level fluctuation amplitude (obtained by subtracting the average data of every two months), find the maximum and minimum values ​​of the annual groundwater level fluctuation amplitude, and then make a difference between the two. This data is the maximum fluctuation amplitude of the year. Calculate year by year to obtain the maximum annual fluctuation amplitude sequence of the groundwater level, and then calculate it using the coefficient of variation calculation formula.

[0084] The coefficient of variation is calculated using the following formula :

[0085] (1)

[0086] in is the mean of the series, is the standard deviation, The units and magnitude are independent.

[0087] Interannual variability of rainfall:

[0088] First, the rainfall data set in the Google Search Engine (GEE) platform was used to collect the rainfall of each monitoring well in the region. Multi-value extraction was performed through the rainfall data set of each monitoring point, and the daily data was summarized monthly to obtain the monthly rainfall data of each monitoring point.

[0089] Then, the coefficient of variation of the multi-year rainfall series is calculated. The same principle as that of obtaining the interannual maximum fluctuation amplitude series of groundwater level is used to calculate using the above formula (1) to obtain the interannual coefficient of variation of rainfall (CV_P).

[0090] S1.3. The interannual coefficient of variation of rainfall (CV_P) was compared with the interannual coefficient of variation of groundwater level (CV_G) to investigate the extent to which rainfall and anthropogenic groundwater extraction affect the interannual variability of GWL.

[0091] For example, if the coefficient of variation of groundwater levels in each monitoring well in the region is greater than 0.3, it indicates that the interannual variation of groundwater levels in the region is large. By comparing it with the coefficient of variation of rainfall, we can preliminarily determine the driving factors that have a greater impact on groundwater level fluctuations in the region. If the interannual variation of rainfall is less than 0.3 and is relatively evenly distributed in the region, it means that rainfall has little impact on groundwater level variation in the region.

[0092] Step 2: Combine the jump variation classification method of hydrological series based on correlation coefficient with Pettitt and sliding T tests to identify the variation points of groundwater level and rainfall at different time scales, quantitatively describe the degree of variation, and divide the subsequences.

[0093] From the calculation results of S1, it can be seen that the coefficient of variation determines the interannual variability of groundwater level and rainfall in the region. By identifying the variation points of long-term series of groundwater level and rainfall at different time scales, the dependence of groundwater level on rainfall is further determined.

[0094] S2.1, analysis of variation points;

[0095] The Pettitt method is used to detect variation points in hydrological time series. The calculation process of the Pettitt test is:

[0096] (2)

[0097] (3)

[0098] in, is a statistical variable; n is the sample size; time series ; is a sign function whose value is -1, 0 or 1;

[0099] The mutation point is defined as :

[0100] (4)

[0101] The significance level of the variation point is determined by:

[0102] (5)

[0103] When the significance level P When the value is <0.05, the detected variation point is considered to be statistically significant.

[0104] The calculation accuracy and test results of the variation points obtained by different detection methods may be inconsistent. In order to obtain more accurate variation point detection results, the present invention uses the sliding T test method and the Pettitt method to identify variation points. The sliding T test compares the observed values ​​of the sequence time point with the mean and standard deviation of the surrounding time points to determine whether there is a jump variation point, and calculates the significance level of the variation point. Define statistics:

[0105] (6)

[0106] In the formula, The mean and variance of the two subsequences are and , and ,in

[0107] (7)

[0108] In the formula, for Sample size time series , artificially set a certain moment as the reference point, and the two subsequences before and after the reference point , The samples are and , Formula (7) complies with the degree of freedom The t-distribution of .

[0109] S2.2, classification method of jump variation of hydrological series based on correlation coefficient;

[0110] For the variation points detected by Pettitt and sliding T, considering the complexity and high variability of the hydrological process, the correlation coefficient is introduced as an indicator, and the hydrological series jump variation classification method based on the correlation coefficient is used to quantitatively describe the variation degree of the time series variation points. The calculation formula is as follows:

[0111] (8)

[0112] (9)

[0113] (10)

[0114] For a sequence of length n , split the entire sequence into and The mean of the entire sequence is The means of the two subsequences are and ; Construct a sequence consisting of jump components ,Right now ; , then the mean of the jump component sequence is and Equal, that is, they are both the means of the original sequences, which can be expressed as:

[0115] (11)

[0116] When a sequence When a jump mutation occurs, its jump component The higher the proportion, by The greater the impact, The more it reflects the jump variation characteristics, the stronger the correlation between the two, that is, the larger the correlation coefficient. The calculation formula is as follows:

[0117] (12)

[0118] right = For hypothesis testing, the commonly used significance levels are 0.01, 0.02, 0.05, and 0.1. When conducting hydrological analysis, it is generally required that the data length is greater than or equal to 20. .like Failed to pass the significant level The test classifies it as having no variation; When , it is classified as a large variation, and the correlation coefficient is between 0 and 1. The details are shown in Table 1:

[0119] Table 1 Classification of hydrological sequence variation degree

[0120]

[0121] S2.3. Based on the variation points with significant levels in the groundwater level series at different time scales, the long-term series of groundwater levels are divided into subsequences with the same behavior, and the dependence of groundwater levels on rainfall is discovered.

[0122] Specifically, the same behavior means having the same trend. After the mutation point is identified, the sequence is decomposed into two with the mutation point as the node, and the two sequences have the same trend.

[0123] Step 3: Based on the subsequence division, the rainfall and groundwater level data of each monitoring point are stabilized.

[0124] In order to further determine the response time of groundwater to rainfall, the rainfall-groundwater level transfer function can be constructed based on the variation point. As the input and output signals of the transfer function, the rainfall and groundwater level sequences are often composed of a deterministic component superimposed on a random component, and this deterministic component often forms a strong autocorrelation. Therefore, the input and output signals need to be stabilized before constructing the transfer function. Based on the subsequences divided in S2, the rainfall and groundwater level data are stabilized respectively.

[0125] Usually, a linear regression equation combined with a power spectrum is used to remove the trend and seasonality of a sequence. Subsequently, a periodic term analysis is performed on the portion after the trend component is removed. The present invention uses a harmonic analysis method to extract the periodic component. , you can use The estimated value is expressed in the form of superposition of harmonics:

[0126] (13)

[0127] Where: For sequence An estimated value of is the number of harmonics, take The integer part of ; is the harmonic number, ; , is The period of a harmonic, n For sequence The length of the sequence;

[0128] , are the Fourier coefficients: ; .like , then it is considered that The wave is significant, otherwise it is not significant. for The variance of is the significance level of the test. Substituting the significant wave into the formula In this way, we get the estimated equation of the periodic term.

[0129] Step 4: Construct the rainfall-groundwater level transfer function based on the variation points to characterize the response of groundwater levels in different subsequences to rainfall.

[0130] Groundwater level fluctuations are mainly affected by the combined effects of rainfall and groundwater extraction. At present, due to the limitations of relevant conditions, the amount of groundwater extraction in regional monitoring wells cannot be officially measured. How to separate the impact of groundwater extraction on groundwater levels, so as to more accurately and quantitatively calculate the response time of rainfall and groundwater levels, so as to compare the impact of groundwater extraction on groundwater levels, is an urgent problem to be solved in extracting the driving factors of groundwater levels in different periods. Cross-correlation analysis can reveal the random dependence and response relationship between groundwater levels and rainfall. Based on the data after smoothing processing of the variation points, a transfer function is constructed, and the aquifer is regarded as a filter, with rainfall as the input signal and groundwater level as the output signal. The cross-correlation function is used to quantitatively describe the response time of rainfall and groundwater levels in the time domain. If rainfall is regarded as a random process, such as Figure 2 As shown, the cross-correlation function is the transfer function.

[0131] The transfer function is calculated as follows:

[0132] Length is sequence The cross-correlation between is defined as:

[0133] (14)

[0134] In the cross-correlation function diagram, the lag time corresponding to the maximum peak is generally defined as the response time of the system.

[0135] (15)

[0136] Based on this method, by comparing the hysteresis coefficients of different monitoring wells in the same subsequence, the spatiotemporal response relationship of the groundwater system to rainfall can be determined. At the same time, according to the hysteresis time of each monitoring well in different subsequences and combined with the overall groundwater extraction data of the region, the main driving factors affecting the groundwater level changes in different subsequences can be extracted.

[0137] By introducing the cross-correlation method based on the variation point analysis, the response time of the groundwater level to rainfall can be calculated more accurately. Therefore, the method of the present invention provides a more suitable research tool for analyzing the evolution of groundwater levels under different degrees of human disturbance and extracting the main driving factors affecting groundwater level changes.

[0138] The core of the technical solution of the present invention is as follows:

[0139] 1) A variety of variation point detection methods are used, and the correlation coefficient is introduced to quantitatively identify the degree of variation of variation points in long-term hydrological series, and the variation degree is graded and quantified, which improves the accuracy of variation point identification.

[0140] 2) Based on the variation points, a transfer function is constructed to simulate the process of rainfall to groundwater change, which overcomes the influence of parameter complexity and subjective factors, and more effectively quantifies the response time of groundwater level and rainfall, thereby improving the accuracy of obtaining the driving factors that affect groundwater level changes in different periods.

[0141] 3) Taking into account the synergistic effects of climate change and anthropogenic exploitation, the impact of groundwater exploitation on groundwater levels is stripped away, especially in areas where groundwater is overexploited and the amount of exploitation cannot be officially measured. This invention provides a more effective tool for quantifying the temporal and spatial response relationship between rainfall and groundwater levels and extracting the main driving factors of groundwater level changes.

[0142] The above description is only used to help understand the method of the present invention and its core essence, but the protection scope of the present invention is not limited thereto. For those skilled in the art in the art, equivalent replacement or change according to the technical solution and inventive concept of the present invention within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for extracting groundwater level driving factors by using variation analysis combined with transfer function, characterized in that: The steps include: S1. Obtain regional long-term series hydrological data, including groundwater depth data, groundwater level data at each monitoring point, and rainfall data at each monitoring point; S2. Based on the hydrological data, the interannual maximum fluctuation amplitude sequence of groundwater level and the interannual maximum fluctuation amplitude sequence of rainfall at each monitoring point are calculated, and the interannual variation coefficient of groundwater level and the interannual variation coefficient of rainfall at each monitoring point are calculated, and the interannual variation coefficient of groundwater level and the interannual variation coefficient of rainfall are compared to determine the interannual variability of regional long-term series rainfall and groundwater level; S3. The Pettitt method and sliding T test method were used to identify groundwater level variation points and rainfall variation points at different time scales, respectively. The hydrological series jump variation classification method based on correlation coefficient was used to quantitatively describe the degree of variation of the time series variation points. At the same time, the groundwater level data and rainfall data of each monitoring point were divided into subsequences with the same behavior based on the groundwater level variation points. S4, performing stabilization processing on the rainfall data and groundwater level data of each subsequence of each monitoring point, wherein the stabilization processing includes using a linear regression equation combined with a power spectrum to remove the trend and seasonality of the subsequence, and using a harmonic analysis method to remove the periodicity of the subsequence; S5. Construct a rainfall-groundwater level transfer function based on the smoothed rainfall data and groundwater level data, wherein the rainfall-groundwater level transfer function is used to describe the response time of rainfall and groundwater level in the time domain; extract the main driving factors affecting the groundwater level change in different subsequences according to the response time of different subsequences of each monitoring point and in combination with the overall groundwater extraction data of the region.

2. The method for extracting groundwater level driving factors by using variation analysis combined with transfer function according to claim 1 is characterized in that: The S2 specifically includes the following steps: S201. Based on the regional long-term series groundwater depth data, the difference between the groundwater depths of two consecutive years is calculated to evaluate the fluctuation range of the groundwater level and determine whether the groundwater level in the region is affected by rainfall and groundwater exploitation factors. S202, if the groundwater level calculated in S201 fluctuates violently, it is determined that the groundwater level in the region is affected by rainfall factors and groundwater exploitation factors; based on the groundwater level data of each monitoring point in the regional long-term series, the interannual maximum fluctuation amplitude sequence of the groundwater level is calculated, and the interannual variation coefficient of the groundwater level at each monitoring point is calculated using the variation coefficient calculation formula; Based on the rainfall data of each monitoring point in the regional long-term series, the maximum annual fluctuation amplitude sequence of rainfall is calculated, and the annual variation coefficient of rainfall at each monitoring point is calculated using the variation coefficient calculation formula; S203. Compare the interannual coefficient of variation of groundwater level and the interannual coefficient of variation of rainfall; The interannual coefficient of variation of groundwater level at monitoring points is greater than 0.3, indicating that the interannual variation of groundwater level in the region is large; correspondingly, the interannual coefficient of variation of rainfall at monitoring points is less than 0.3 and is evenly distributed in the region, indicating that rainfall has little impact on the variation of groundwater level in the region.

3. The method for extracting groundwater level driving factors by using variation analysis combined with transfer function according to claim 2 is characterized in that: In S202, the inter-annual maximum fluctuation amplitude sequence of the groundwater level is calculated, and the inter-annual variation coefficient of the groundwater level at each monitoring point is calculated using the variation coefficient calculation formula, which is as follows: Based on the groundwater level data of each monitoring point in the regional long-term series, the monthly groundwater level fluctuation amplitude of each monitoring point is calculated and its extreme value is subtracted, calculated year by year, and the annual maximum fluctuation amplitude sequence of the groundwater level of each monitoring point is obtained by merging. The interannual coefficient of variation of groundwater level at each monitoring point is calculated. The coefficient of variation calculation formula is as follows: in, is the coefficient of variation, is the average value of the values ​​in the sequence, is the standard deviation.

4. The method for extracting groundwater level driving factors by using variation analysis combined with transfer function according to claim 3 is characterized in that: In S202, the maximum annual fluctuation amplitude sequence of rainfall is calculated, and the annual variation coefficient of rainfall at each monitoring point is calculated using the variation coefficient calculation formula, which is obtained in the same way as the annual variation coefficient of groundwater level, as follows: Based on the rainfall data of each monitoring point in the regional long-term series, the monthly rainfall fluctuation amplitude of each monitoring point is calculated and its extreme value is subtracted, calculated year by year, and the interannual maximum fluctuation amplitude sequence of rainfall at each monitoring point is obtained; finally, the interannual variation coefficient of rainfall at each monitoring point is calculated using the coefficient of variation calculation formula.

5. The method for extracting groundwater level driving factors by using variation analysis and transfer function according to claim 4 is characterized in that: The S3 specifically includes the following steps: S301. Use the Pettitt method to identify groundwater level variation points and rainfall variation points at different time scales; S302, using a sliding T test method to identify groundwater level variation points and rainfall variation points at different time scales; S303. Use the hydrological series jump variation classification method based on correlation coefficient to quantitatively describe the variation degree of the time series variation point. The calculation formula is as follows: For a sequence of length n , split the entire sequence into and The mean of the whole sequence is The means of the two subsequences are and ; Construct a sequence consisting of jump components ,Right now ; , then the mean of the jump component sequence is and Equal, that is, they are both the means of the original sequences, expressed as: When a sequence When a jump mutation occurs, its jump component The higher the proportion, by The greater the impact, The more it reflects the jumping variation characteristics, the stronger the correlation between the two, that is, the larger the correlation coefficient; The calculation formula is as follows: right = For hypothesis testing, the commonly used significance levels are 0.01, 0.02, 0.05, and 0.

1. When conducting hydrological analysis, the data length is required to be greater than or equal to 20. ;like Failed to pass the significant level The test classifies it as having no variation; When , it is classified as a giant variation, and the correlation coefficient is between 0 and 1; S304. Based on the groundwater level data at different time scales, the groundwater level data and rainfall data of each monitoring point are divided into two subsequences with the same behavior through the groundwater level variation points with significant levels.

6. The method for extracting groundwater level driving factors by using variation analysis combined with transfer function according to claim 1 or 5, characterized in that: In S4, the harmonic analysis method is used to remove the periodicity of the subsequence, as follows: After removing the trend component, the periodic term analysis is performed and the periodic component is extracted by harmonic analysis. ,use The estimated value is expressed in the form of superposition of harmonics: Where: For sequence An estimated value of is the number of harmonics, take The integer part of ; is the harmonic number, ; , is the The period of a harmonic, n For sequence The length of the sequence; , are the Fourier coefficients: ; ;like , then it is considered that The wave is significant, otherwise it is not significant; for The variance of is the significance level of the test; substitute the significant wave into the formula In this way, we get the estimated equation of the periodic term.

7. The method for extracting groundwater level driving factors by using variation analysis combined with transfer function according to claim 6 is characterized in that: The S5 is specifically as follows: The aquifer is regarded as a filter, with rainfall data as the input signal and groundwater level data as the output signal. The cross-correlation function is used to quantitatively describe the response time of rainfall and groundwater level in the time domain. If rainfall is regarded as a random process, the cross-correlation function is the transfer function. The calculation process of rainfall-groundwater level transfer function is as follows: Length is sequence and The cross-correlation function between is: In the cross-correlation function graph, the lag time corresponding to the maximum peak is defined as the response time of the system; 。