Post-flood season discrimination method based on flood frequency analysis-trend division

Abstract

A post-flood season discrimination method based on flood frequency analysis-trend division comprises the following steps: 1, collecting day-by-day hydrological factor data of historical flood seasons of stations, and fitting day-by-day probability distribution functions of the hydrological factors of the flood seasons of the stations; 2, estimating a parameter initial value in the day-by-day probability distribution function; 3, calculating a fitting degree and adjusting parameters; 4, calculating a hydrological factor design flood threshold value according to the day-by-day probability distribution function under a certain occurrence frequency, and then fitting to obtain a design flood hydrograph; 5, calculating a single-day magnitude index; 6, calculating data characteristic indexes, and drawing a change curve of the data characteristic indexes; 7, judging the most significant change point; 8, judging post-flood season staging points; and 9, judging the starting time of the post-flood season in the region or the drainage basin. According to the method, the potential of historical hydrological data is fully excavated, the probability statistical analysis and trend analysis technologies are combined, and compared with an existing fixed flood season judgment method, the method has higher adaptability, accuracy and practicability and has certain innovativeness and application value in the field of flood season staging.

Description

Technical Field

[0001] The present invention relates to the technical field of flood season staging, and in particular to a method for distinguishing a post-flood season based on flood frequency analysis-trend division. Background Art

[0002] The timing, sources, and composition of floods in many river basins across my country are extremely complex, posing severe challenges to flood control and water resources management. Therefore, developing a basin-scale method for identifying the post-flood period is crucial for improving flood control capabilities and optimizing water resource scheduling.

[0003] Faced with complex hydrological systems with multiple cross-sections and multiple variables, the comprehensiveness and systematic nature of traditional post-flood season identification methods often fail to meet practical needs. Furthermore, traditional technologies are also insufficient in mining historical hydrological factor information. Existing post-flood season identification methods often rely on real-time monitoring of a single hydrological factor, neglecting the accumulation of historical data and the coordinated analysis of multi-site data. This compromises the accuracy and reliability of the identification results.

[0004] Therefore, it is necessary to design a post-flood season identification method based on flood frequency analysis-trend division to overcome the above problems. Summary of the Invention

[0005] To overcome these issues, a post-flood season determination method based on flood frequency analysis and trend classification is provided. This method addresses the problems of existing post-flood season determination technologies, which fail to fully exploit historical hydrological factor information and lack comprehensiveness and systematicity when addressing complex hydrological systems with multiple sections and variables. The present invention proposes a post-flood season determination method that fully considers historical hydrological information and hydrological information from multiple control sections. The core of the present invention lies in coupling flood frequency analysis with design flood trend classification. Through flood frequency analysis, historical hydrological information is fully exploited, and design flood trend classification is performed on hydrological factors at multiple stations, systematizing the complex hydrological situation. The proposed method process includes flood frequency analysis, calculation of design flood hydrological factors, construction of data characteristic index curves, extraction of mutation points, and determination of the post-flood season in the basin. Furthermore, the present invention constructs a data characteristic index calculation formula, improving the original non-parametric test method to a parametric test method that considers relative data relationships. This method can more comprehensively consider the changes in hydrological factors, determine dynamic judgment criteria, and determine the start time of the post-flood season in the basin, providing a scientific basis for reservoir scheduling and flood control decisions.

[0006] The present invention provides a method for distinguishing the post-flood season based on flood frequency analysis-trend division, comprising the following steps:

[0007] Step 1: Collect the daily hydrological factor data of the historical flood season of the site and fit the daily probability distribution function of the hydrological factor of each site during the flood season;

[0008] Step 2, estimate the initial values of the parameters in the daily probability distribution function;

[0009] Step 3, calculate the fit and adjust the parameters;

[0010] Step 4: Under a certain occurrence frequency, the design flood threshold of the hydrological factor is calculated based on the daily probability distribution function, and then the design flood process line is obtained by fitting;

[0011] Step 5: Calculate the daily level index;

[0012] Step 6: Calculate the data characteristic index and draw its change curve;

[0013] Step 7, determine the most significant change point;

[0014] Step 8, determine the post-flood season stage point;

[0015] Step 9: Determine the start time of the post-flood season in the region or basin.

[0016] Preferably, the daily probability distribution function in step 1 includes a probability density function and a cumulative distribution function:

[0017] The single-day frequency fitting adopts the Pearson III frequency curve, and its probability density function is:

[0018]

[0019] Where x is the hydrological factor value of a hydrological station; Γ(α) is the gamma function of α; α, β, α0 are the shape parameters, scale parameters, and location parameters of the Pearson III frequency curve; α>0, β>0;

[0020] α is a shape parameter that determines the shape of the frequency curve, including skewness and peak degree. If the data points are generally close to a straight line, the value of α is generally large; if the data points are generally more curved, the value of α is generally small.

[0021] β is a scale parameter that determines the scale of the frequency curve. If the ratio between the maximum and minimum data points is generally small, the value of β is generally large; if the ratio between the maximum and minimum data points is generally large, the value of β is generally small.

[0022] α0 is the position parameter, which determines the position of the frequency curve. If the data value is small, the value of α0 is generally large; if the data value is large, the value of α0 is generally large.

[0023] Cumulative distribution function:

[0024] Where x is the hydrological factor value at a hydrological station, and F(x) is the probability that the daily average of the hydrological factor at that hydrological station is greater than the hydrological factor value. This function is obtained by integrating the probability density function.

[0025] Preferably, the estimation formula for the initial values of parameters α, β, and α0 in step 2 is:

[0026]

[0027] in, n is the total number of data points in the sample; x i is the hydrological factor value of the i-th data point; K i is the relative value of the i-th data point to the sample mean; is the average value of the hydrological factor series; C V is the coefficient of variation, which indicates the degree of dispersion of the sample; C S is the coefficient of deviation, which indicates the degree of asymmetry of the sample. The initial values of the parameters α, β, and α0 can be estimated based on the collected hydrological data.

[0028] Preferably, in step 3, the goodness of fit criterion is the residual sum of squares of the frequency curve, and the calculation formula is: y i is the observed value of the hydrological factor at the i-th data point; is the estimated value of the hydrological factor for the i-th data point. The smaller the residual sum of squares, the better the fitting of the daily probability distribution function. The parameters are estimated using the gradient descent method. The corresponding interval range is selected based on the initial estimated value of the parameter.

[0029] Preferably, in step 4, the occurrence frequency is known, and the inverse function of the cumulative distribution function is used to solve the hydrological factor design flood threshold, which is expressed as: x = F -1 (p); then the design flood process line of the hydrological factors in the flood season is obtained by fitting.

[0030] When obtaining the design flood process line under a certain occurrence frequency, a representative occurrence frequency is selected, and the daily hydrological factor value of the frequency is calculated through the daily probability distribution function, and finally the design flood process line of the hydrological factor during the flood season is obtained.

[0031] Preferably, the single-day magnitude index O in step 5 t The calculation formula is:

[0032]

[0033] Where, F t -1 (p) is the design value of the hydrological factor on day t under the occurrence frequency p; is the design value of the hydrological factor on the i-th day under the occurrence frequency p; T is the number of flood season periods; i is any data between 1 and T.

[0034] By extracting the hydrological data from the design flood process line and comparing the hydrological factors of the target date with those during the flood season on a daily basis, we can fully explore and utilize historical hydrological information; daily data comparison can provide more detailed information on water regime changes, making the division of the post-flood season more accurate.

[0035] Preferably, in step 6, for a time series of length T, a data characteristic index U is defined for any time t. t , and its calculation formula is:

[0036]

[0037] Where, F t -1 (p) is the design value of the hydrological factor on day t at frequency p; is the design value of the hydrological factor on the jth day under frequency p; T is the number of flood season periods; i is any number between 1 and t, and j is any number between 1 and T;

[0038] Then, according to each U t The corresponding time t is plotted as its changing curve.

[0039] Based on the single-day level index, the data characteristic index curve is calculated, and the trend of the hydrological factor data at the daily scale of the design flood of the representative station is divided. Before and after the continuous data show significant differences, the characteristic value will change significantly. By finding the change point, the change pattern of the continuous data can be discovered, and then the flood season phase points can be selected. Combining historical hydrological data with trend division helps to identify the pattern of hydrological changes. Historical data provides a rich pattern of hydrological changes. By analyzing the characteristics of past water conditions, specific trends and abnormal behaviors of the flood season can be identified, thereby improving the accuracy of post-flood season judgment. Calculate the characteristic index U of a single hydrological factor t After that, according to each U t The corresponding time t is plotted as its changing curve.

[0040] Preferably, in step 7, the method for determining the most significant change point is: determining the significant change point according to the trend of the change curve of the data characteristic index, if |U t |Continuously increasing or decreasing indicates that there is no significant change point in this time series; if |U t |When a maximum value appears, it means that this point is the most significant change point. The calculation formula is:

[0041] Preferably, in step 8, at time point K TAs the dividing point, the flood season is divided into 1~K T With K T The flood process line is designed based on the hydrological factors to determine the period in which the post-flood season dividing point is located.

[0042] If the post-flood season boundary point is between 1 and K T The calculation formula for the second-stage single-day index is:

[0043]

[0044] The calculation formula for the characteristic index of the second-stage data is:

[0045]

[0046] Find U′ t After that, 1~K can be calculated according to the data characteristic index curve T The most significant change point is calculated as follows:

[0047] If the post-flood season boundary point is at K T During the T-period, the calculation formula for the second-stage daily index is:

[0048]

[0049] The calculation formula for the characteristic index of the second-stage data is:

[0050]

[0051] Find U′ t After that, 1~K can be calculated according to the data characteristic index curve T The most significant change point is calculated as follows:

[0052] Repeat the above steps, combine the hydrological factors to design the flood process line, and determine whether the most significant change point in each stage is the post-flood season staging point. After the post-flood season staging point is determined, the post-flood season threshold corresponding to the hydrological factor of the station is determined based on this point. After the main flood season, the post-flood season time calculated based on the hydrological factor of the hydrological station can be determined based on the measured data, which is calculated as D ij , represents the jth hydrological factor of the i-th hydrological station in the basin.

[0053] Preferably, step 9 specifically includes:

[0054] 9.1 To determine when a hydrological station enters the post-flood season, the formula is as follows:

[0055] D i =max(D i1 ,D i2,...,D ij ...,D im );

[0056] Where D i is the time when the i-th hydrological station enters the post-flood season, D ij is the jth hydrological factor of the i-th hydrological station in the basin, j is any data between 1 and m, m is the number of hydrological factors of this station, and the latest time among all hydrological factors is taken as the time when this station enters the post-flood season;

[0057] 9.2 The determination formula for when a region or river basin enters the post-flood season is as follows:

[0058] D=max(D1,D2,...,D i ...,D n );

[0059] Where D is the time when the region or basin enters the post-flood season, D i The time when the i-th hydrological station enters the post-flood season, i is any data between 1 and n, and n is the number of hydrological stations in the basin or region. The latest time among all hydrological stations is taken as the time when this station enters the post-flood season.

[0060] The method of the present invention can coordinate hydrological data from multiple control sections based on comprehensive consideration of historical hydrological information, and has important theoretical significance and practical application value. It also has dynamic adaptability to cope with changing hydrological conditions and ensure timely response under different hydrological scenarios. The present invention combines methods such as flood frequency analysis and trend classification to fully utilize historical hydrological data and systematically process hydrological factors at multiple sites. This method can not only reveal the long-term trend of hydrological factors, but also identify the mutation points of the daily hydrological situation at multiple stations during the flood season, thereby providing a more comprehensive basis for judging the post-flood season.

[0061] In flood frequency analysis, the Pearson Type III curve is widely used to analyze the frequency of extreme hydrological events. This method is suitable for asymmetric distributed data and can effectively capture the characteristics of flood events, thereby providing a more accurate estimate of flood frequency.

[0062] To better identify turning points in time series, it's crucial to collect historical hydrological data for flood frequency analysis and then employ parameter testing to identify trends. This approach systematically analyzes the changing patterns of historical hydrological factors, revealing potential hydrological trends and providing a more reliable basis for identifying post-flood season trends. Trend identification and comprehensive analysis of hydrological factors across multiple stations helps overcome the limitations of existing methods for complex, multi-section, and multi-variable hydrological systems, improving the accuracy and scientific nature of the identification results.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] This method fully exploits the potential information in flood season data from historical basin stations. By fitting the probability density function of these data, detailed frequency curves are obtained. Furthermore, the plotting of hydrological factor-based design flood process lines effectively depicts the dynamic changes in hydrological processes, providing an important basis for accurately identifying flood seasons. In contrast, existing fixed flood season identification methods often overlook the rich information contained in historical data, which can easily lead to misjudgments of flood seasons.

[0065] This method uses parameter testing to analyze the design flood process lines of watershed hydrological factors. It accurately identifies significant change points in continuous data and dynamically determines the start and end of the flood season, avoiding the limitations of traditional fixed flood season divisions. This method better captures the dynamic characteristics of hydrological processes and reflects the fluctuations of water regime factors. Compared with existing methods that rely solely on statistical eigenvalues, it offers greater adaptability and flexibility.

[0066] By combining flood frequency analysis with design flood trend classification, this method can more deeply explore the inherent patterns of watershed hydrological data. Probability density analysis of water regime factors can identify hydrological characteristics under different conditions. Simultaneously, the study of design flood hydrographs helps us understand historical patterns of flood seasons, while trend classification can promptly capture significant changes in hydrological data. Compared with flood season classification methods that rely solely on hydrological simulation, this statistical analysis-based approach is more closely aligned with actual hydrological processes and can more accurately capture flood season characteristics.

[0067] This dynamic flood season identification method is simple to implement within the basin, requiring no specialized hydrological knowledge or complex computational procedures. Relying solely on historical water level or flow data from each monitoring station, it can rapidly identify flood seasons, making it applicable to basins of all sizes within the region. Compared to traditional methods that require numerous parameter inputs and complex post-flood season identification models, this statistical analysis-based approach is more convenient and practical.

[0068] In summary, this hydrological discrimination method for dynamic discrimination of the post-flood season fully taps the potential of historical hydrological data and combines the techniques of probability statistical analysis and trend analysis. Compared with the existing fixed flood season discrimination method, it has stronger adaptability, accuracy and practicality, and has certain innovation and application value in the field of flood season classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a flow chart of a method for distinguishing the post-flood season by flood frequency analysis-trend division according to a preferred embodiment of the present invention;

[0070] Figure 2This is a flow frequency curve diagram of three stations in the middle reaches of the Yangtze River according to a preferred embodiment of the present invention;

[0071] Figure 3 This is a graph showing the flow process lines and one-stage data characteristic index curves of three stations in the middle reaches of the Yangtze River in a preferred embodiment of the present invention.

[0072] Figure 4 This is a graph showing the flow process lines and two-stage data characteristic index curves of three stations in the middle reaches of the Yangtze River in a preferred embodiment of the present invention.

[0073] Figure 5 This is a graph showing the flow process lines and three-stage data characteristic index curves of three stations in the middle reaches of the Yangtze River in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to illustrate the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0075] like Figure 1 As shown, this embodiment provides a method for distinguishing the post-flood season based on flood frequency analysis-trend division, comprising the following steps:

[0076] Step 1: First, collect the daily hydrological factor data of the historical flood season of the site and determine the daily probability distribution function of the hydrological factor of each site during the flood season.

[0077] The single-day frequency fitting adopts the Pearson III frequency curve, and its probability density function is:

[0078]

[0079] Where x is the hydrological factor value of a hydrological station; Γ(α) is the gamma function of α; α, β, α0 are the shape parameters, scale parameters, and location parameters of the Pearson III frequency curve; α>0, β>0;

[0080] α is a shape parameter that determines the shape of the frequency curve, including skewness and peak degree. If the data points are generally close to a straight line, the value of α is generally large; if the data points are generally more curved, the value of α is generally small.

[0081] β is a scale parameter that determines the scale of the frequency curve. If the ratio between the maximum and minimum data points is generally small, the value of β is generally large; if the ratio between the maximum and minimum data points is generally large, the value of β is generally small.

[0082] α0 is the position parameter, which determines the position of the frequency curve. If the data value is small, the value of α0 is generally large; if the data value is large, the value of α0 is generally large.

[0083] Cumulative distribution function:

[0084] Where x is the hydrological factor value of a hydrological station, and F(x) is the probability that the daily average value of the hydrological factor at the hydrological station is greater than the hydrological factor value.

[0085] Step 2: Estimation of the initial values of the parameters. The initial values of the three parameters α, β, and α0 of the daily probability distribution function are usually obtained through the following three overall parameters: C V 、C S Make an estimate:

[0086]

[0087] Where, C V 、C S The calculation formula for the initial value is as follows:

[0088]

[0089] The initial values of the parameters α, β, and α0 can be estimated based on the collected hydrological data; n is the total number of data points in the sample; x i is the hydrological factor value of the i-th data point; K i is the relative value of the i-th data point to the sample mean; is the average value of the hydrological factor series; C V is the coefficient of variation, which indicates the degree of dispersion of the sample; C S is the coefficient of deviation, which indicates the degree of asymmetry of the sample.

[0090] Step 3: Calculate the goodness of fit and adjust the parameters, and select the corresponding interval range based on the initial estimated values of the parameters.

[0091] The fitting criterion is the residual sum of squares of the frequency curve, and the calculation formula is as follows:

[0092]

[0093] y i is the observed value of the hydrological factor at the i-th data point; is the estimated value of the hydrological factor for the i-th data point; the smaller the residual sum of squares, the better the model fitting effect.

[0094] The gradient descent method is used to estimate the curve parameters in order to obtain a probability distribution model that is as accurate as possible.

[0095] Step 4: Obtain the design flood process line under a certain occurrence frequency, select a representative occurrence frequency, calculate the daily hydrological factor value of the frequency through the daily probability distribution function of the flood season hydrological factor obtained in the above steps, and finally obtain the design flood process line of the flood season hydrological factor.

[0096] Given the occurrence frequency P, the inverse function of the cumulative distribution function can be used to solve the design flood threshold of the hydrological factor, the expression is: x = F -1 (p); then the design flood process line of the hydrological factors in the flood season is obtained by fitting.

[0097] Step 5: Calculate the daily level index O t , extract the hydrological data from the flood process line designed in the above steps, and compare the hydrological factors of the target date with the hydrological factors during the flood season on a daily basis, so as to fully explore and utilize historical hydrological information. Daily data comparison can provide more detailed information on water regime changes, making the classification of the post-flood season more accurate. Its expression is as follows:

[0098]

[0099] Where, F t -1 (p) is the design value of the hydrological factor on day t under the occurrence frequency p; F i -1 (p) is the design value of the hydrological factor on the i-th day under the occurrence frequency p; T is the number of flood season periods; i is any data between 1 and T.

[0100] Step 6: Based on the daily level indicators, calculate the data characteristic index curve and divide the trend of the hydrological factor data of the design flood daily scale of the representative station. Before and after the continuous data show significant differences, the characteristic values will change significantly. By finding the change points, the change pattern of the continuous data can be discovered, and then the flood season stages can be selected.

[0101] Combining historical hydrological data with trend classification helps to identify the pattern of hydrological changes. Historical data provides a rich set of hydrological change patterns. By analyzing the characteristics of past water conditions, we can identify specific trends and abnormal behaviors during the flood season, thereby improving the accuracy of post-flood season discrimination. For a time series of length T, define a data indicator feature U for any time t. t , which is expressed as follows:

[0102]

[0103] Where, F t -1(p) is the design value of the hydrological factor on day t at frequency p; is the design value of the hydrological factor on the jth day under frequency p; T is the number of flood season periods; i is any number between 1 and t, and j is any number between 1 and T;

[0104] Calculate the characteristic index U of a single hydrological factor t After that, according to each U t The corresponding time t is plotted as its changing curve.

[0105] Step 7: Determine the most significant change point. Determine the significant change point based on the curve trend. If |U t |Continuously increasing or decreasing indicates that there is no significant change point in this time series; if |U t |When a maximum value appears, it means that this point is the most significant change point. The calculation formula is:

[0106] Step 8, at time point K T As the dividing point, the flood season is divided into 1~K T With K T The flood process line is designed based on the hydrological factors to determine the period in which the post-flood season dividing point is located.

[0107] If the post-flood season boundary point is between 1 and K T The calculation formula for the second-stage single-day index is:

[0108]

[0109] The calculation formula for the characteristic index of the second-stage data is:

[0110]

[0111] Find U′ t After that, 1~K can be calculated according to the data characteristic index curve T The most significant change point is calculated as follows:

[0112] If the post-flood season boundary point is at K T During the T-period, the calculation formula for the second-stage daily index is:

[0113]

[0114] The calculation formula for the characteristic index of the second-stage data is:

[0115]

[0116] Find U′ t After that, 1~K can be calculated according to the data characteristic index curveT The most significant change point is calculated as follows:

[0117] Repeat the above steps, combine the hydrological factors to design the flood process line, and determine whether the most significant change point in each stage is the post-flood season staging point. After the post-flood season staging point is determined, the post-flood season threshold corresponding to the hydrological factor of the station is determined based on this point. After the main flood season, the post-flood season time calculated based on the hydrological factor of the hydrological station can be determined based on the measured data, which is calculated as D ij , represents the jth hydrological factor of the i-th hydrological station in the basin.

[0118] Step 9: Determine the start time of the post-flood season in the region or basin.

[0119] First, determine when a hydrological station enters the post-flood season. The determination formula is as follows:

[0120] D i =max(D i1 ,D i2 ,...,D ij ...,D im );

[0121] Where D i is the time when the i-th hydrological station enters the post-flood season, D ij is the jth hydrological factor of the i-th hydrological station in the basin, j is any data between 1 and m, m is the number of hydrological factors of this station, and the latest time among all hydrological factors is taken as the time when this station enters the post-flood season.

[0122] Finally, the time when the region or basin enters the post-flood season is determined by the following formula:

[0123] D=max(D1,D2,...,D i ...,D n );

[0124] Where D is the time when the region or basin enters the post-flood season, D i The time when the i-th hydrological station enters the post-flood season, i is any data between 1 and n, and n is the number of hydrological stations in the basin or region. The latest time among all hydrological stations is taken as the time when this station enters the post-flood season.

[0125] Focusing on the middle Yangtze River's control basin, we selected the daily average flow rates for 92 days from June 1st to August 31st over the past several decades or centuries at three stations: Huangzhuang Station in the middle and lower reaches of the Yangtze River, the combined flow rate of the four Dongting Lakes, and the combined flow rate of the five Poyang Lakes. Using a probability distribution model to fit the frequency curves daily, we estimated the probability density function of the flow rate at each station over these 92 days, resulting in a total of 276 cumulative distribution function curves. The parameter selection intervals for each station are shown in Table 1:

[0126] Table 1 Parameter ranges for flood frequency analysis at three stations in the middle reaches of the Yangtze River

[0127] parameter Flow of the four rivers of Dongting Lake Flow of Poyang Five Lakes Traffic flow at Huangzhuang Station α 0.1～5 0.1～2 0.1～3 β 0～0.1 0～0.02 0～0.01 <![CDATA[a0]]> 10～5000 300～3000 10～2000

[0128] Figure 2 The flow frequency curve of Huangzhuang Station on June 16, the composite flow frequency curve of the four Dongting Lakes on July 16, and the composite flow frequency curve of the five Poyang Lakes on August 16 are displayed.

[0129] On this basis, the design flow process lines of 0.01% occurrence frequency, 0.1% occurrence frequency, 1% occurrence frequency, 5% occurrence frequency, 10% occurrence frequency and 20% occurrence frequency are drawn for each station.

[0130] A trend analysis was conducted on the daily scale flow data of the design flood at three representative stations, and then the flood season stages were selected as the post-flood season thresholds of the corresponding hydrological factors at this station. Figure 3 The flow process line of 10% occurrence frequency at three stations in the middle reaches of the Yangtze River and the characteristic index curve of the first stage data, Figure 4 The flow process line with 10% occurrence frequency and the characteristic index curve of the second stage data at three stations in the middle reaches of the Yangtze River are shown. Figure 3 The flow process lines of 10% occurrence frequency at three stations in the middle reaches of the Yangtze River and the curves of three-stage data characteristic indicators.

[0131] Combined with the flow process line, for the 10% frequency flow process at Huangzhuang Station, the first stage |U t |When the maximum value appears, it is the beginning of the pre-flood season. The second stage|U t |When the maximum value appears, it is the end time of the pre-flood season. t |The time when the maximum value appears is the beginning of the flood season. For the 10% occurrence frequency process of the composite flow of the four Dongting Lakes and the composite flow of the five Poyang Lakes, the first stage |U t |The end time of the post-flood season is when the maximum value appears. The second stage|U t |When the maximum value appears, it is the beginning of the pre-flood season. t The peak value indicates the start of the post-flood season. Table 2: Flow thresholds for Huangzhuang Station, the composite flow of the four Dongting Lakes, and the composite flow of the five Poyang Lakes under different design flood frequencies.

[0132] Table 2 Discrimination thresholds based on flood frequency analysis-trend division

[0133]

[0134] Through the above steps, we can get the threshold values of post-flood season division with different occurrence frequencies. We select 10% occurrence frequency as the threshold value for water regime factor discrimination in post-flood season. When the flow rate at a station reaches the corresponding threshold, we can judge that the control basin of this station has entered the post-flood season. For example, when the flow rate at Huangzhuang station drops to 5280m after the main flood season, 3 / s, it can be determined that Huangzhuang Station has entered the post-flood season. Then, based on the comprehensive analysis of the three stations, the time of entering the post-flood season can be determined.

[0135] This method fully exploits the potential information in flood season data from historical basin stations. By fitting the probability density function of these data, detailed frequency curves are obtained. Furthermore, the plotting of hydrological factor-based design flood process lines effectively depicts the dynamic changes in hydrological processes, providing an important basis for accurately identifying flood seasons. In contrast, existing fixed flood season identification methods often overlook the rich information contained in historical data, which can easily lead to misjudgments of flood seasons.

[0136] This method uses parameter testing to analyze the design flood process lines of watershed hydrological factors. It accurately identifies significant change points in continuous data and dynamically determines the start and end of the flood season, avoiding the limitations of traditional fixed flood season divisions. This method better captures the dynamic characteristics of hydrological processes and reflects the fluctuations of water regime factors. Compared with existing methods that rely solely on statistical eigenvalues, it offers greater adaptability and flexibility.

[0137] By combining flood frequency analysis with design flood trend classification, this method can more deeply explore the inherent patterns of watershed hydrological data. Probability density analysis of water regime factors can identify hydrological characteristics under different conditions. Simultaneously, the study of design flood hydrographs helps us understand historical patterns of flood seasons, while trend classification can promptly capture significant changes in hydrological data. Compared with flood season classification methods that rely solely on hydrological simulation, this statistical analysis-based approach is more closely aligned with actual hydrological processes and can more accurately capture flood season characteristics.

[0138] This dynamic flood season identification method is simple to implement within the basin, requiring no specialized hydrological knowledge or complex computational procedures. Relying solely on historical water level or flow data from each monitoring station, it can rapidly identify flood seasons, making it applicable to basins of all sizes within the region. Compared to traditional methods that require numerous parameter inputs and complex post-flood season identification models, this statistical analysis-based approach is more convenient and practical.

[0139] In summary, this hydrological discrimination method for dynamic discrimination of the post-flood season fully taps the potential of historical hydrological data and combines the techniques of probability statistical analysis and trend analysis. Compared with the existing fixed flood season discrimination method, it has stronger adaptability, accuracy and practicality, and has certain innovation and application value in the field of flood season classification.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying the post-flood season based on flood frequency analysis and trend division, characterized in that: The steps include: Step 1: Collect the daily hydrological factor data of the historical flood season of the site and fit the daily probability distribution function of the hydrological factor of each site during the flood season; Step 2, estimate the initial values of the parameters in the daily probability distribution function; Step 3, calculate the fit and adjust the parameters; Step 4: Under a certain occurrence frequency, the design flood threshold of the hydrological factor is calculated based on the daily probability distribution function, and then the design flood process line is obtained by fitting; Step 5: Calculate the daily level index; Step 6: Calculate the data characteristic index and draw its change curve; Step 7, determine the most significant change point; Step 8, determine the post-flood season stage point; Step 9: Determine the start time of the post-flood season in the region or basin.

2. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 1, characterized in that: The daily probability distribution function in step 1 includes the probability density function and the cumulative distribution function: The single-day frequency fitting adopts the Pearson III frequency curve, and its probability density function is: Where x is the hydrological factor value of a hydrological station; Γ(α) is the gamma function of α; α, β, α0 are the shape parameters, scale parameters, and location parameters of the Pearson III frequency curve; α>0, β>0; Cumulative distribution function: Where x is the hydrological factor value of a hydrological station, and F(x) is the probability that the daily average value of the hydrological factor at the hydrological station is greater than the hydrological factor value.

3. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 2, characterized in that: The estimation formula for the initial values of parameters α, β, and α0 in step 2 is: in, n is the total number of data points in the sample; x i is the hydrological factor value of the i-th data point; K i is the relative value of the i-th data point to the sample mean; is the average value of the hydrological factor series; C V is the coefficient of variation, which indicates the degree of dispersion of the sample; C S is the coefficient of deviation, which indicates the degree of asymmetry of the sample.

4. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 3, characterized in that: In step 3, the goodness of fit criterion is the residual sum of squares of the frequency curve, and the calculation formula is: y i is the observed value of the hydrological factor at the i-th data point; is the estimated value of the hydrological factor for the i-th data point; the smaller the residual sum of squares, the better the fitting effect of the daily probability distribution function; the gradient descent method is used to estimate the parameters.

5. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 4, characterized in that: In step 4, given the occurrence frequency P, the inverse function of the cumulative distribution function is used to solve the design flood threshold of the hydrological factor, which is expressed as: x = F -1 (p); then the design flood process line of the hydrological factors in the flood season is obtained by fitting.

6. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 5, characterized in that: Single-day volume indicator O in step 5 t The calculation formula is: Where, F t -1 (p) is the design value of the hydrological factor on day t under the occurrence frequency p; F i -1 (p) is the design value of the hydrological factor on day i under the occurrence frequency p; T is the number of flood season periods; i is any data between 1 and T.

7. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 6, characterized in that: In step 6, for a time series of length T, a data characteristic index U is defined for any time t t , and its calculation formula is: Where, F t -1 (p) is the design value of the hydrological factor on day t at frequency p; is the design value of the hydrological factor on the jth day under frequency p; T is the number of flood season periods; i is any number between 1 and t, and j is any number between 1 and T; Then, according to each U t The corresponding time t is plotted as its changing curve.

8. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 7, characterized in that: In step 7, the method for determining the most significant change point is: determine the significant change point based on the trend of the change curve of the data characteristic index. If |U t |Continuously increasing or decreasing indicates that there is no significant change point in this time series; if |U t |When a maximum value appears, it means that this point is the most significant change point. The calculation formula is:

9. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 8, characterized in that: In step 8, at time point K T As the dividing point, the flood season is divided into 1~K T With K T The flood process line is designed based on the hydrological factors to determine the period in which the post-flood season dividing point is located. If the post-flood season boundary point is between 1 and K T The calculation formula for the second-stage single-day index is: The calculation formula for the characteristic index of the second-stage data is: Obtain U' t After that, 1~K can be calculated according to the data characteristic index curve T The most significant change point is calculated as follows: If the post-flood season boundary point is at K T During the T-period, the calculation formula for the second-stage daily index is: The calculation formula for the characteristic index of the second-stage data is: Obtain U' t After that, 1~K can be calculated according to the data characteristic index curve T The most significant change point is calculated as follows: Repeat the above steps, combine the hydrological factors to design the flood process line, and determine whether the most significant change point in each stage is the post-flood season staging point. After the post-flood season staging point is determined, the post-flood season threshold corresponding to the hydrological factor of the station is determined based on this point. After the main flood season, the post-flood season time calculated based on the hydrological factor of the hydrological station can be determined based on the measured data, which is calculated as D ij , represents the jth hydrological factor of the i-th hydrological station in the basin.

10. The method for identifying the post-flood season based on flood frequency analysis and trend division as claimed in claim 9, characterized in that: Step 9 specifically includes: 9.1 To determine when a hydrological station enters the post-flood season, the formula is as follows: D i =max(D i1 ,D i2 ,...,D ij ...,D im ); Where D i is the time when the i-th hydrological station enters the post-flood season, D ij is the jth hydrological factor of the i-th hydrological station in the basin, j is any data between 1 and m, m is the number of hydrological factors of this station, and the latest time among all hydrological factors is taken as the time when this station enters the post-flood season; 9.2 The determination formula for when a region or river basin enters the post-flood season is as follows: D=max(D1,D2,...,D i ...,D n ); Where D is the time when the region or basin enters the post-flood season, D i The time when the i-th hydrological station enters the post-flood season, i is any data between 1 and n, and n is the number of hydrological stations in the basin or region. The latest time among all hydrological stations is taken as the time when this station enters the post-flood season.

Images