A daily hydrological series reconstruction method based on change detection and deviation correction

Abstract

The present invention discloses a method for reconstructing daily hydrological series based on change detection and deviation correction, and the main steps are as follows: first, convert the daily hydrological series into a monthly series, detect the mutation point of the monthly series and its significance; second, test and analyze the trend before and after the mutation point to ensure that there is no significant trend; then, perform preliminary correction based on the cumulative distribution function to make the distribution characteristics of the monthly series before and after the mutation point consistent; then, distribute the cumulative correction total deviation of all mutation months to the daily hydrological series of non-mutation months in proportion; finally, reconstruct the daily hydrological series of non-mutation months and mutation months respectively, so as to obtain a complete daily hydrological series. The method can realize the restoration and reconstruction of the daily hydrological series, can effectively deal with the situation of multiple mutation points and ensure the smooth transition of data, and greatly improve the accuracy and adaptability of the reconstruction of the daily hydrological series.

Description

Technical Field

[0001] The present invention relates to the field of hydrological data analysis, and in particular to a daily hydrological sequence reconstruction method based on change detection and deviation correction. Background Art

[0002] Daily hydrological series are basic data in the fields of water resources management, climate change research, flood warning, etc. Their accuracy and consistency are of great significance for scientific research and practical applications. However, due to factors such as climate change, human activities and changes in observation conditions, daily hydrological series are often affected by mutation points or even multiple mutation points, making the hydrological series non-stationary. This non-stationarity has a significant impact on the analysis and prediction accuracy of hydrological series. Existing technologies for reconstructing daily hydrological series with multiple mutation points mainly include correction methods based on linear interpolation, regression analysis and proportional distribution, but these methods have obvious deficiencies in daily accuracy and systematicness of mutation point processing. Specifically, the deviation distribution mostly stays on the monthly scale, ignoring the refined distribution of daily data, which is easy to cause data jumps or uneven distribution; mutation month correction mostly adopts a unified adjustment method, which is difficult to take into account both the consistency of monthly scale data and daily dynamic characteristics; there is a lack of systematic solutions for the processing of multiple mutation point scenarios, resulting in the problem of deviation superposition or imbalanced distribution between multiple mutation points, which seriously affects the stability and consistency of the correction sequence. For example, when faced with hydrological sequences with multiple mutation points, most existing methods can only deal with a single mutation point, or process it through successive adjustments. However, this method can easily lead to accumulation or uneven distribution of deviations between mutation points, thereby destroying the continuity of the corrected data.

[0003] Therefore, the present invention proposes a daily hydrological series reconstruction method based on change detection and deviation correction, aiming to effectively deal with the situation of multiple mutation points and ensure the smooth transition of data. Through the systematic mutation point detection and deviation balance allocation mechanism, this method can not only accurately identify multiple mutation points, but also realize the effective alignment and dynamic transition of data before and after the mutation, which greatly improves the accuracy, adaptability and application value of daily hydrological series reconstruction. Summary of the invention

[0004] The present invention proposes a daily hydrological series reconstruction method based on change detection and deviation correction. This method is applicable to hydrological series with multiple mutation points, and is mainly used for the reconstruction of hydrological series at the monthly and daily scales, including the two scenarios of the reappearance and restoration of hydrological series. Through the systematic mutation point detection and deviation balance allocation mechanism, a smooth transition is achieved in the multi-mutation point scenario, avoiding distribution deviation caused by cumulative errors, and greatly improving the accuracy, adaptability and application value of daily hydrological series reconstruction.

[0005] To achieve the above object, the present invention adopts the following technical solution: a daily hydrological series reconstruction method based on change detection and deviation correction, comprising the following steps:

[0006] Step S1, mutation point diagnosis: the daily hydrological series is processed into a monthly hydrological series, and the mutation points in the monthly hydrological series are diagnosed by the Pettitt test and the Wilcoxon rank sum test method to obtain the mutation point diagnosis result;

[0007] Step S2, trend analysis: according to the mutation point diagnosis result obtained in step S1, the trend characteristics of the monthly hydrological series before and after the mutation point are analyzed by using the Mann-Kendall trend test method to ensure that there is no significant trend in the monthly hydrological series before and after the mutation point;

[0008] Step S3, preliminary correction based on cumulative distribution function: for the monthly hydrological series of the mutation month without significant trend, the monthly hydrological series with mutation points are corrected based on the cumulative distribution function, so that the distribution characteristics of the monthly hydrological series before and after the mutation point are consistent, and the monthly hydrological series of the mutation month after preliminary correction is obtained;

[0009] Step S4, deviation balance allocation: according to the cumulative corrected total deviation value between the monthly-scale hydrological sequence of the mutation month and the original monthly-scale hydrological sequence after preliminary correction in step S3, the cumulative corrected total deviation value of the mutation month is allocated to the daily-scale hydrological sequence of the non-mutation month according to the proportion of the original daily-scale hydrological sequence value, so as to obtain the deviation balance allocation value of the daily hydrological sequence of the non-mutation month;

[0010] Step S5, daily sequence reconstruction: apply the deviation balance distribution value of the daily hydrological sequence of the non-mutation month in step S4 to the daily hydrological sequence to obtain the reconstructed daily hydrological sequence of the non-mutation month; based on the ratio of the monthly hydrological sequence of the mutation month before and after the preliminary correction in step S3, scale the original daily hydrological sequence of the mutation month by the same proportion to obtain the reconstructed daily hydrological sequence of the mutation month.

[0011] Furthermore, the mutation point diagnosis in step S1 is specifically as follows:

[0012] Step S11, Pettitt test: identify the specific time of the monthly hydrological series mutation and determine the month and year when the mutation occurs;

[0013] Step S12, Wilcoxon rank sum test: verify the significance of the mutation points of the monthly hydrological series to ensure the accuracy and scientificity of the identification.

[0014] Furthermore, the trend analysis in step S2 is specifically as follows:

[0015] Step S21, data extraction: extracting the monthly hydrological series of the mutation month into monthly hydrological series before and after the mutation point respectively;

[0016] Step S22, trend test: Use the Mann-Kendall trend test method to analyze the trend characteristics of the monthly hydrological series before and after the mutation point respectively; if there is no significant trend, enter the next step of correction process; otherwise, further processing or exclusion is required to ensure that there is no significant trend in the monthly hydrological series before and after the mutation point.

[0017] Furthermore, the preliminary correction based on the cumulative distribution function in step S3 is specifically:

[0018] The monthly hydrological series after preliminary correction are divided into two scenarios: reappearance and restoration;

[0019] Step S31, restore the situation: make the monthly hydrological series distribution characteristics before the mutation point consistent with the monthly hydrological series distribution characteristics after the mutation point, so as to achieve the restoration effect of the monthly hydrological series in the mutation month; the specific steps are:

[0020] Step S311, cumulative distribution function definition: Let the monthly hydrological sequence before the mutation point be X, the monthly hydrological sequence after the mutation point be Z, and the cumulative distribution functions of the monthly hydrological sequence before the mutation point and the monthly hydrological sequence after the mutation point be F respectively. X and F Z , defined as follows:

[0021] F X (x) = P (x ≤ x);

[0022] F Z (z) = P (z ≤ z);

[0023] In the formula, F X (x) represents the cumulative distribution function of the monthly hydrological series X before the mutation point; F Z (z) represents the cumulative distribution function of the monthly hydrological sequence Z after the mutation point; P(X≤x) represents the probability that the data point of the monthly hydrological sequence X before the mutation point is less than or equal to x; P(Z≤z) represents the probability that the data point of the monthly hydrological sequence Z after the mutation point is less than or equal to z;

[0024] Step S312, cumulative distribution function transformation: Based on the transformation of the cumulative distribution function, the monthly hydrological sequence X before the mutation point is adjusted so that the distribution characteristics of the monthly hydrological sequence X before the mutation point are consistent with the monthly hydrological sequence Z after the mutation point. The correction formula is as follows:

[0025] ;

[0026] In the formula, F X (X s ) represents the monthly hydrological sequence X before the mutation point s The cumulative distribution value of F Z -1 is the inverse cumulative distribution function, which converts the monthly hydrological series X before the mutation point into s The cumulative distribution value F X (X s ) Distribution characteristics of the monthly hydrological series Z after being mapped to the mutation point; Y 还现s represents the monthly hydrological series X before the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series Z after the mutation;

[0027] Through steps S311 and S312, the cumulative distribution function is used to correct and eliminate the distribution difference between the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point, and the monthly hydrological series Y after preliminary correction is obtained. 还现 ;

[0028] Step S32, restoring the situation: making the monthly hydrological series distribution characteristics after the mutation point consistent with the monthly hydrological series distribution characteristics before the mutation point, so as to restore the monthly hydrological series in the mutation month; the specific steps are:

[0029] Step S321, cumulative distribution function definition: Let the monthly hydrological sequence before the mutation point be X, the monthly hydrological sequence after the mutation point be Z, and the cumulative distribution function of the monthly hydrological sequence before the mutation point and the monthly hydrological sequence after the mutation point be F X and F Z , defined as follows:

[0030] F X (x) = P(X≤x);

[0031] F Z (z) = P(Z ≤ z);

[0032] In the formula, F X (x) represents the cumulative distribution function of the monthly hydrological series X before the mutation point; F Z (z) represents the cumulative distribution function of the monthly hydrological sequence Z after the mutation point; P(X≤x) represents the probability that the data point of the monthly hydrological sequence X before the mutation point is less than or equal to x; P(Z≤z) represents the probability that the data point of the monthly hydrological sequence Z after the mutation point is less than or equal to z;

[0033] Step S322, cumulative distribution function inverse transformation: Based on the cumulative distribution function inverse transformation, the monthly hydrological series Z after the mutation point is adjusted to the distribution characteristics of the monthly hydrological series X before the mutation. The correction formula is as follows:

[0034] ;

[0035] In the formula, F Z (Z s ) represents the monthly hydrological sequence Z after the mutation point s The cumulative distribution value of F X -1 is the inverse cumulative distribution function, which converts the Z in the monthly hydrological series Z after the mutation point into s The cumulative distribution value F Z (Z s ) is mapped to the distribution characteristics of the monthly hydrological series X before the mutation point; Y 还原s Z represents the monthly hydrological series Z after the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series X before the mutation;

[0036] Through steps S321 and S322, the cumulative distribution function is used to correct and eliminate the distribution difference between the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point, and the monthly hydrological series Y after preliminary restoration and correction is obtained. 还原 .

[0037] Furthermore, the deviation balance allocation in step S4 is specifically as follows:

[0038] The deviation balance distribution values ​​of the daily hydrological series in non-mutation months are divided into two scenarios: the present scenario and the restoration scenario;

[0039] Step S41, the current situation: taking the most recent mutation year as the base year, and distributing the accumulated corrected total deviation values ​​of the monthly hydrological series of multiple mutation months to the non-mutation months according to the proportion of the original daily hydrological series values;

[0040] Step S411, calculate the preliminary correction deviation of the monthly hydrological series of each mutation month and accumulate them, see the following formula:

[0041] ;

[0042] In the formula, It represents the cumulative corrected total deviation value of the monthly hydrological series in all mutation months; Represents the index of the monthly hydrological series in the mutation month; Indicates the month of mutation; Indicates the mutation month The preliminary correction deviation value of k represents the total number of monthly hydrological series in the mutation months without significant trend;

[0043] Step S412, daily deviation value allocation:

[0044] Taking the most recent mutation year as the base year, the cumulative corrected total deviation value is allocated to the daily hydrological series of the non-mutation month base year and the non-mutation month base year before the base year according to the ratio of the original daily hydrological series value of the non-mutation month before the base year to the sum of the original daily hydrological series values ​​of all non-mutation months, to ensure a smooth transition of the data; the allocation formula is as follows:

[0045] ;

[0046] Among them, t represents the year, j represents the non-mutation month, and d represents the day; Indicates allocation to the non-mutation month in year t moon The correction deviation value of the day; Indicates the most recent mutation benchmark year; represents the starting year of the original daily-scale hydrological series of non-mutation month j; Indicates the most recent mutation base year and the annual sum of the original daily-scale hydrological series values ​​of non-mutation month j in year t before the most recent mutation benchmark year; The most recent mutation benchmark year and the original daily-scale hydrological series value of the non-mutation month j day d in the tth year before the most recent mutation benchmark year;

[0047] Step S42, restoring the situation: taking the earliest mutation year as the base year, and distributing the accumulated corrected total deviation values ​​of the monthly hydrological series of multiple mutation months to the non-mutation months according to the proportion of the original daily hydrological series values;

[0048] Step S411, calculate the preliminary correction deviation of the monthly hydrological series of each mutation month and accumulate them, see the following formula:

[0049] ;

[0050] Step S422, daily deviation value allocation:

[0051] Taking the earliest mutation year as the base year, the accumulated corrected total deviation value is allocated to the daily hydrological series of the non-mutation month base year and the non-mutation month base year according to the ratio of the original daily hydrological series value of the non-mutation month after the base year to the sum of the original daily hydrological series value of all non-mutation months, to ensure a smooth transition of the data; the specific allocation formula is as follows:

[0052] ;

[0053] in, Indicates the earliest mutation benchmark year; represents the end year of the original daily-scale hydrological series of non-mutation month j; The earliest mutation benchmark year and the annual sum of the original daily-scale hydrological series values ​​of non-mutation month j in the tth year after the earliest mutation benchmark year; The earliest mutation benchmark year and the original daily-scale hydrological series value of the dth day of the non-mutation month j in the tth year after the earliest mutation benchmark year.

[0054] Furthermore, the reconstruction of the daily hydrological sequence in step S5 includes the reconstruction of the daily hydrological sequence in the non-mutation month and the reconstruction of the daily hydrological sequence in the mutation month, specifically:

[0055] Step S51, reconstruction of the daily hydrological series in non-mutation months:

[0056] The deviation balance distribution value of the daily hydrological sequence of the non-mutation month in step S4 is applied to the original daily scale hydrological sequence to obtain the reconstructed daily hydrological sequence of the non-mutation month. The specific calculation formula is as follows:

[0057] ;

[0058] in, represents the reconstructed hydrological sequence value of the non-mutation month j month d day in year t, It represents the original daily scale hydrological series value of the non-mutation month j month d day in year t;

[0059] Step S52, reconstructing the daily sequence of the mutation month:

[0060] Based on the ratio of the monthly hydrological series before and after preliminary correction in step S3, the original daily hydrological series of the mutation month is scaled to the corresponding daily hydrological series of the mutation month in the same proportion. The specific calculation formula is as follows:

[0061]

[0062] in: Indicates Year mutation month Indicates Year mutation month ; Indicates the mutation month in year t The reconstructed hydrological series value on the d day of the month; Indicates the mutation month in year t The original daily scale hydrological series value of day d of the month;

[0063] Through steps S51 and S52, the reconstruction of the daily hydrological sequence of the non-mutation month and the mutation month is completed respectively, and finally a complete daily hydrological sequence reconstruction result is formed, realizing the restoration and recovery of the daily hydrological sequence.

[0064] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0065] The first is the daily refined processing of deviation balance allocation, which allocates deviations based on the proportion of daily data, effectively avoiding data jumps and achieving a smooth transition of daily hydrological series. The second is the high-precision reconstruction of daily hydrological series, which ensures that the daily series not only accurately restores the original characteristics but also remains consistent with the monthly scale correction results through the same proportional scaling processing of mutation months and differentiated deviation allocation of non-mutation months. The third is the smooth processing of multiple mutation points, which uniformly processes all mutation points to ensure the stability and consistency of hydrological series under multiple mutation scenarios.

[0066] It is worth pointing out that the method of the present invention assumes that there is no significant trend change before and after the mutation, requires that the base year data is complete and of high quality, and assumes that the distribution characteristics of the data before and after the mutation have a certain similarity to ensure the effectiveness of the cumulative distribution function transformation; if the mutation point fluctuates frequently, there is a significant trend change, or it is a short-term fluctuation caused by random events, this method may be limited and should be used with caution. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0068] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with specific embodiments. 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.

[0069] like Figure 1 As shown, the present invention provides a daily hydrological series reconstruction method based on change detection and deviation correction, comprising the following steps:

[0070] Step S1, mutation point diagnosis: the daily hydrological series is processed into a monthly hydrological series, and the mutation points in the monthly hydrological series are diagnosed by the Pettitt Test and the Wilcoxon Rank-Sum Test methods to obtain the mutation point diagnosis results;

[0071] Step S2, trend analysis: according to the mutation point diagnosis result obtained in step S1, the Mann-Kendall Trend Test method is used to analyze the trend characteristics of the monthly hydrological series before and after the mutation point, to ensure that there is no significant trend in the monthly hydrological series before and after the mutation point;

[0072] Step S3, preliminary correction based on cumulative distribution function: for the monthly hydrological series of the mutation month without significant trend, the monthly hydrological series with mutation points are corrected based on the cumulative distribution function, so that the distribution characteristics of the monthly hydrological series before and after the mutation point are consistent, and the monthly hydrological series of the mutation month after preliminary correction is obtained;

[0073] Step S4, deviation balance allocation: according to the cumulative corrected total deviation value between the monthly-scale hydrological sequence of the mutation month and the original monthly-scale hydrological sequence after preliminary correction in step S3, the cumulative corrected total deviation value of the mutation month is allocated to the daily-scale hydrological sequence of the non-mutation month according to the proportion of the original daily-scale hydrological sequence value, so as to obtain the deviation balance allocation value of the daily hydrological sequence of the non-mutation month;

[0074] Step S5, daily sequence reconstruction: apply the deviation balance distribution value of the daily hydrological sequence of the non-mutation month in step S4 to the daily hydrological sequence to obtain the reconstructed daily hydrological sequence of the non-mutation month; based on the ratio of the monthly hydrological sequence of the mutation month before and after the preliminary correction in step S3, scale the original daily hydrological sequence of the mutation month by the same proportion to obtain the reconstructed daily hydrological sequence of the mutation month.

[0075] Furthermore, the mutation point diagnosis in step S1 is specifically as follows:

[0076] Step S11, Pettitt test: identify the specific time of the monthly hydrological series mutation and determine the month and year when the mutation occurs;

[0077] Step S12, Wilcoxon rank sum test: verify the significance of the mutation points of the monthly hydrological series to ensure the accuracy and scientificity of the identification.

[0078] Furthermore, the trend analysis in step S2 is specifically as follows:

[0079] Step S21, data extraction: extracting the monthly hydrological series of the mutation month into monthly hydrological series before and after the mutation point respectively;

[0080] Step S22, trend test: Use the Mann-Kendall trend test method to analyze the trend characteristics of the monthly hydrological series before and after the mutation point respectively; if there is no significant trend, enter the next step of correction process; otherwise, further processing or exclusion is required to ensure that there is no significant trend in the monthly hydrological series before and after the mutation point.

[0081] Furthermore, the preliminary correction based on the cumulative distribution function in step S3 is specifically:

[0082] The monthly hydrological series after preliminary correction are divided into two scenarios: reappearance and restoration;

[0083] Step S31, restore the situation: make the monthly hydrological series distribution characteristics before the mutation point consistent with the monthly hydrological series distribution characteristics after the mutation point, so as to achieve the restoration effect of the monthly hydrological series in the mutation month; the specific steps are:

[0084] Step S311, cumulative distribution function definition: Let the monthly hydrological sequence before the mutation point be X, the monthly hydrological sequence after the mutation point be Z, and the cumulative distribution functions of the monthly hydrological sequence before the mutation point and the monthly hydrological sequence after the mutation point be F respectively. X and F Z , defined as follows:

[0085] F X (x) = P (x ≤ x);

[0086] F Z (z) = P (z ≤ z);

[0087] In the formula, F X (x) represents the cumulative distribution function of the monthly hydrological series X before the mutation point; F Z (z) represents the cumulative distribution function of the monthly hydrological sequence Z after the mutation point; P(X≤x) represents the probability that the data point of the monthly hydrological sequence X before the mutation point is less than or equal to x; P(Z≤z) represents the probability that the data point of the monthly hydrological sequence Z after the mutation point is less than or equal to z;

[0088] Step S312, cumulative distribution function transformation: Based on the transformation of the cumulative distribution function, the monthly hydrological sequence X before the mutation point is adjusted so that the distribution characteristics of the monthly hydrological sequence X before the mutation point are consistent with the monthly hydrological sequence Z after the mutation point. The correction formula is as follows:

[0089] ;

[0090] In the formula, F X (X s ) represents the monthly hydrological sequence X before the mutation point s The cumulative distribution value of FZ -1 is the inverse cumulative distribution function, which converts the monthly hydrological series X before the mutation point into i The cumulative distribution value F X (X s ) Distribution characteristics of the monthly hydrological series Z after being mapped to the mutation point; Y 还现s represents the monthly hydrological series X before the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series Z after the mutation;

[0091] Through steps S311 and S312, the cumulative distribution function is used to correct and eliminate the distribution difference between the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point, and the monthly hydrological series Y after preliminary correction is obtained. 还现 ;

[0092] Step S32, restoring the situation: making the monthly hydrological series distribution characteristics after the mutation point consistent with the monthly hydrological series distribution characteristics before the mutation point, so as to restore the monthly hydrological series in the mutation month; the specific steps are:

[0093] Step S321, cumulative distribution function definition: Let the monthly hydrological sequence before the mutation point be X, the monthly hydrological sequence after the mutation point be Z, and the cumulative distribution function of the monthly hydrological sequence before the mutation point and the monthly hydrological sequence after the mutation point be F X and F Z , defined as follows:

[0094] F X (x) = P(X≤x);

[0095] F Z (z) = P(Z ≤ z);

[0096] In the formula, F X (x) represents the cumulative distribution function of the monthly hydrological series X before the mutation point; F Z (z) represents the cumulative distribution function of the monthly hydrological sequence Z after the mutation point; P(X≤x) represents the probability that the data point of the monthly hydrological sequence X before the mutation point is less than or equal to x; P(Z≤z) represents the probability that the data point of the monthly hydrological sequence Z after the mutation point is less than or equal to z;

[0097] Step S322, cumulative distribution function inverse transformation: Based on the cumulative distribution function inverse transformation, the monthly hydrological series Z after the mutation point is adjusted to the distribution characteristics of the monthly hydrological series X before the mutation. The correction formula is as follows:

[0098] ;

[0099] In the formula, F Z (Z s ) represents the monthly hydrological sequence Z after the mutation point s The cumulative distribution value of F X -1 is the inverse cumulative distribution function, which converts the Z in the monthly hydrological series Z after the mutation point into s The cumulative distribution value F Z (Z s ) is mapped to the distribution characteristics of the monthly hydrological series X before the mutation point; Y 还原s Z represents the monthly hydrological series Z after the mutation point i The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series X before the mutation;

[0100] Through steps S321 and S322, the cumulative distribution function is used to correct and eliminate the distribution difference between the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point, and the monthly hydrological series Y after preliminary restoration and correction is obtained. 还原 .

[0101] Furthermore, the deviation balance allocation in step S4 is specifically as follows:

[0102] The deviation balance distribution values ​​of the daily hydrological series in non-mutation months are divided into two scenarios: the present scenario and the restoration scenario;

[0103] Step S41, the current situation: taking the most recent mutation year as the base year, and distributing the accumulated corrected total deviation values ​​of the monthly hydrological series of multiple mutation months to the non-mutation months according to the proportion of the original daily hydrological series values;

[0104] Step S411, calculate the preliminary correction deviation of the monthly hydrological series of each mutation month and accumulate them, see the following formula:

[0105] ;

[0106] In the formula, It represents the cumulative corrected total deviation value of the monthly hydrological series in all mutation months; Represents the index of the monthly hydrological series in the mutation month; Indicates the month of mutation; Indicates the mutation month The preliminary correction deviation value of k represents the total number of monthly hydrological series in the mutation months without significant trend;

[0107] Step S412, daily deviation value allocation:

[0108] Taking the most recent mutation year as the base year, the cumulative corrected total deviation value is allocated to the daily hydrological series of the non-mutation month base year and the non-mutation month base year before the base year according to the ratio of the original daily hydrological series value of the non-mutation month before the base year to the sum of the original daily hydrological series values ​​of all non-mutation months, to ensure a smooth transition of the data; the allocation formula is as follows:

[0109] ;

[0110] Among them, t represents the year, j represents the non-mutation month, and d represents the day; Indicates allocation to the non-mutation month in year t moon The correction deviation value of the day; Indicates the most recent mutation benchmark year; represents the starting year of the original daily-scale hydrological series of non-mutation month j; Indicates the most recent mutation base year and the annual sum of the original daily-scale hydrological series values ​​of non-mutation month j in year t before the most recent mutation benchmark year; The most recent mutation benchmark year and the original daily-scale hydrological series value of the non-mutation month j day d in the tth year before the most recent mutation benchmark year;

[0111] Step S42, restoring the situation: taking the earliest mutation year as the base year, and distributing the accumulated corrected total deviation values ​​of the monthly hydrological series of multiple mutation months to the non-mutation months according to the proportion of the original daily hydrological series values;

[0112] Step S411, calculate the preliminary correction deviation of the monthly hydrological series of each mutation month and accumulate them, see the following formula:

[0113] ;

[0114] Step S422, daily deviation value allocation:

[0115] Taking the earliest mutation year as the base year, the accumulated corrected total deviation value is allocated to the daily hydrological series of the non-mutation month base year and the non-mutation month base year according to the ratio of the original daily hydrological series value of the non-mutation month after the base year to the sum of the original daily hydrological series value of all non-mutation months, to ensure a smooth transition of the data; the specific allocation formula is as follows:

[0116] ;

[0117] in, Indicates the earliest mutation benchmark year; represents the end year of the original daily-scale hydrological series of non-mutation month j; The earliest mutation benchmark year and the annual sum of the original daily-scale hydrological series values ​​of non-mutation month j in the tth year after the earliest mutation benchmark year; The earliest mutation benchmark year and the original daily-scale hydrological series value of the dth day of the non-mutation month j in the tth year after the earliest mutation benchmark year.

[0118] Furthermore, the reconstruction of the daily hydrological sequence in step S5 includes the reconstruction of the daily hydrological sequence in the non-mutation month and the reconstruction of the daily hydrological sequence in the mutation month, specifically:

[0119] Step S51, reconstruction of the daily hydrological series in non-mutation months:

[0120] The deviation balance distribution value of the daily hydrological sequence of the non-mutation month in step S4 is applied to the original daily scale hydrological sequence to obtain the reconstructed daily hydrological sequence of the non-mutation month. The specific calculation formula is as follows:

[0121] ;

[0122] in, represents the reconstructed hydrological sequence value of the non-mutation month j month d day in year t, It represents the original daily scale hydrological series value of the non-mutation month j month d day in year t;

[0123] Step S52, reconstructing the daily sequence of the mutation month:

[0124] Based on the ratio of the monthly hydrological series before and after preliminary correction in step S3, the original daily hydrological series of the mutation month is scaled to the corresponding daily hydrological series of the mutation month in the same proportion. The specific calculation formula is as follows:

[0125]

[0126] in: Indicates Year mutation month Indicates Year mutation month ; Indicates the mutation month in year t The reconstructed hydrological series value on the d day of the month; Indicates the mutation month in year t The original daily scale hydrological series value of day d of the month;

[0127] Through steps S51 and S52, the reconstruction of the daily hydrological sequence of the non-mutation month and the mutation month is completed respectively, and finally a complete daily hydrological sequence reconstruction result is formed, realizing the restoration and recovery of the daily hydrological sequence.

[0128] Taking the daily-scale hydrological series of a hydrological station (May 1, 1952 to December 31, 2022) as an example, this paper demonstrates how to achieve daily-scale hydrological series reconstruction based on change detection and deviation correction methods, including two implementation examples: daily-scale hydrological series presentation and daily-scale hydrological series restoration.

[0129] Example 1: Daily hydrological series still exists

[0130] Based on the daily-scale hydrological series of the above hydrological stations (May 1, 1952 to December 31, 2022), the implementation steps of how to use the method of the present invention to realize the restoration of the daily-scale hydrological series are shown as follows.

[0131] Step S1 mutation point diagnosis: The above daily hydrological series are processed into monthly series data, and the mutation points in the monthly hydrological series are diagnosed by Pettitt test and Wilcoxon rank sum test. The analysis shows that the mutation time of the flow series in January, August, November and December of the hydrological station is 1988, 1993, 1980 and 1993 respectively. The year, month and mutation position corresponding to each mutation point are clearly marked and passed the Wilcoxon rank sum test. The specific results are shown in Table 1.

[0132] Table 1 Results of mutation test for monthly hydrological series at a hydrological station

[0133]

[0134] Step S2 Trend analysis: According to the diagnosis results of the mutation point in step S1, the monthly runoff data of January, August, November and December are extracted as monthly hydrological sequences before and after the mutation point according to the position of the mutation point; the Mann-Kendall trend test is used to analyze the trend characteristics of the monthly hydrological sequences before and after the mutation point in the monthly hydrological sequences of January, August, November and December, and it is concluded that there is no significant trend in the monthly hydrological sequences before and after the mutation in the above four mutation months.

[0135] Step S3 Preliminary correction based on cumulative distribution function: Based on the detected mutation point information, preliminary corrections are performed on the monthly hydrological series before and after the mutation points (1988, 1993, 1980 and 1993, respectively) in the monthly hydrological series of January, August, November and December. Specifically, the monthly hydrological series of the above four mutation months are corrected based on the cumulative distribution function to keep the distribution characteristics of the monthly hydrological series before and after the mutation point consistent, and the monthly hydrological series of the four mutation months after preliminary correction are obtained.

[0136] Further, the preliminary correction based on the cumulative distribution function in step S3 is specifically as follows:

[0137] Step S3.1: Perform a preliminary correction on the monthly hydrological series in January: Let the monthly hydrological series from January 1953 to 1988 before the mutation point be X, and the monthly hydrological series from January 1989 to 2022 after the mutation point be Z. The preliminary corrected series is obtained by the following formula:

[0138] ;

[0139] In the formula, F X (X s ) represents the monthly hydrological sequence X before the mutation point s The cumulative distribution value of X in the monthly hydrological series X from 1953 to January 1988 before the mutation point s The cumulative distribution value of F Z -1 is the inverse cumulative distribution function, which converts the monthly hydrological series X before the mutation point into s The cumulative distribution value F X (X s ) Distribution characteristics of monthly hydrological series Z from 1989 to January 2022 after being mapped to the mutation point; Y 还现s represents the monthly hydrological series X before the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series Z after the mutation.

[0140] Step S3.2 Repeat the above process for the monthly hydrological series in August, November and December to ensure that the distribution characteristics of the monthly hydrological series before and after the mutation point remain consistent.

[0141] Step S4 Secondary correction based on deviation value: After the preliminary correction, the most recent mutation year (1993) is taken as the base year, and the cumulative corrected total deviation values ​​of the monthly hydrological series of multiple mutation months are distributed to the non-mutation months according to the proportion of the original daily hydrological series values.

[0142] Step S4.1 Calculate the preliminary correction deviations of the monthly hydrological series in January, August, November and December and accumulate them:

[0143] ;

[0144] Step S4.2 Daily deviation value allocation:

[0145] Taking 1993 as the time point, according to the ratio of the original daily hydrological series values ​​of the non-mutation months in 1993 and before to the sum of the original daily hydrological series values ​​of all non-mutation months, the cumulative corrected total deviation values ​​of the monthly hydrological series of the above four mutation months are allocated to the daily hydrological series of the remaining eight non-mutation months in 1993 and before. The specific allocation formula is as follows:

[0146] ;

[0147] Wherein, t represents the year, which is the year between 1952 and 1993; j represents the non-mutation month; Indicates allocation to the non-mutation month in year t moon The correction deviation value of the day; It represents the cumulative corrected total deviation value of the monthly hydrological series of all mutation months. Here it represents the cumulative corrected total deviation value of the monthly hydrological series of the four mutation months of January, August, November and December. It represents the annual sum of the original daily-scale hydrological series values ​​of non-mutation month j in year t; It represents the original daily scale hydrological series value of day d in non-mutation month j in year t.

[0148] Step S5: Daily sequence reconstruction (reconstruction): including the reconstruction of the daily-scale hydrological sequences of 8 non-mutation months and the reconstruction of the daily-scale hydrological sequences of 4 mutation months.

[0149] Step S51, the daily hydrological series of non-mutation months is restored:

[0150] The deviation balance distribution value of the daily sequence of the non-mutation month in step S4 is applied to the original daily-scale hydrological sequence to obtain the reconstructed daily-scale hydrological sequence of the non-mutation month. The specific calculation formula is as follows:

[0151] ;

[0152] in, represents the reconstructed hydrological sequence value of the non-mutation month j month d day in year t, It represents the original daily scale hydrological series value of the non-mutation month j month d day in year t; is the non-mutation number assigned to year t in step S4 moon The corrected deviation value for the day.

[0153] Step S52, reconstructing the daily sequence of the mutation month:

[0154] Based on the ratios of the monthly hydrological series of the four mutation months of January, August, November and December before and after the preliminary correction in step S3, the original daily hydrological series of the mutation months are scaled to the corresponding daily hydrological series of the mutation months in the same proportion. The specific calculation formula is as follows:

[0155]

[0156] in: Indicates Year mutation month Indicates Year mutation month ; Indicates the mutation month in year t The reconstructed hydrological series value on the d day of the month; Indicates the mutation month in year t The original daily scale hydrological series value of day d of the month;

[0157] Through the above steps, the current sequence value of the daily hydrological series of the hydrological station (May 1, 1952 to December 31, 2022) is obtained.

[0158] Example 2: Daily-scale hydrological sequence restoration

[0159] Similarly, based on the daily-scale hydrological series of the above hydrological stations (May 1, 1952 to December 31, 2022), the implementation steps of how to use the method of the present invention to restore the daily-scale hydrological series are demonstrated.

[0160] Step S1 mutation point diagnosis and step S2 trend analysis are consistent with implementation case 1. The following focuses on step S3, step S4 and step S5.

[0161] Step S3 Preliminary correction based on cumulative distribution function: Based on the detected mutation point information, preliminary corrections are performed on the monthly hydrological series before and after the mutation points (1988, 1993, 1980 and 1993, respectively) in the monthly hydrological series of January, August, November and December. Specifically, the monthly hydrological series of the above four mutation months are corrected based on the cumulative distribution function to keep the distribution characteristics of the monthly hydrological series before and after the mutation point consistent, and the monthly hydrological series of the four mutation months after preliminary correction are obtained.

[0162] Further, the preliminary correction based on the cumulative distribution function in step S3 is specifically as follows:

[0163] Step S3.1: Perform a preliminary correction on the monthly hydrological series in January: Let the monthly hydrological series from January 1953 to 1988 before the mutation point be X, and the monthly hydrological series from January 1989 to 2022 after the mutation point be Z. The preliminary corrected monthly hydrological series is obtained by the following formula:

[0164] ;

[0165] In the formula, F Z (Z s ) represents the monthly hydrological sequence Z after the mutation point s The cumulative distribution value of Z in the monthly hydrological series Z from 1989 to January 2022 after the mutation point s The cumulative distribution value of F X -1 is the inverse cumulative distribution function, which converts the Z in the monthly hydrological series Z after the mutation point into s The cumulative distribution value F Z (Z s ) is mapped to the distribution characteristics of the monthly hydrological series X from January 1953 to 1988 before the mutation point; Y 还原s Z represents the monthly hydrological series Z after the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series X before the mutation.

[0166] Step S3.2 Repeat the above process for the monthly hydrological series in August, November and December to ensure that the distribution characteristics of the monthly hydrological series before and after the mutation point remain consistent.

[0167] Step S4 Secondary correction based on deviation value: After the initial correction, the earliest mutation year (1980) is taken as the base year, and the cumulative corrected total deviation values ​​of the monthly hydrological series of multiple mutation months are distributed to the non-mutation months according to the proportion of the original daily hydrological series values.

[0168] Step S4.1 Calculate the preliminary correction deviations of the monthly hydrological series in January, August, November and December and accumulate them:

[0169] ;

[0170] Step S4.2 Daily deviation value allocation:

[0171] Taking 1980 as the base year, according to the proportion of the original daily hydrological series values ​​of non-mutation months in 1980 and thereafter to the sum of the original daily hydrological series values ​​of all non-mutation months, the cumulative corrected total deviation values ​​of the monthly hydrological series of the above four mutation months are allocated to the daily hydrological series of the remaining eight non-mutation months in 1980 and thereafter. The specific allocation formula is as follows:

[0172] ;

[0173] Among them, t represents the year, which is the year between 1980 and 2022; j represents the non-mutation month; Indicates allocation to the non-mutation month in year t moon The correction deviation value of the day; It represents the annual sum of the original daily-scale hydrological series values ​​of non-mutation month j in year t; It represents the original daily scale hydrological series value of day d in non-mutation month j in year t.

[0174] Step S5: Daily sequence reconstruction (restoration): including restoration of daily-scale hydrological sequences of 8 non-mutation months and restoration of daily-scale hydrological sequences of 4 mutation months.

[0175] Step S51, restoring the daily hydrological series of non-mutation months:

[0176] The deviation balance distribution value of the daily sequence of the non-mutation month in step S4 is applied to the original daily-scale hydrological sequence to obtain the reconstructed daily-scale hydrological sequence of the non-mutation month. The specific calculation formula is as follows:

[0177] ;

[0178] in, represents the reconstructed runoff sequence value of the non-mutation month j on day d in year t, represents the original daily scale hydrological series value of the non-mutation month j month d day in year t, is the non-mutation month assigned to the tth year in step S4 moon The corrected deviation value for the day.

[0179] Step S52, restoring the daily sequence of the mutation month:

[0180] Based on the ratios of the monthly hydrological series of the four mutation months of January, August, November and December before and after the preliminary correction in step S3, the original daily hydrological series of the mutation months are scaled to the corresponding daily hydrological series of the mutation months in the same proportion. The specific calculation formula is as follows:

[0181]

[0182] in: Indicates Year mutation month Indicates Year mutation month ; Indicates the mutation month in year t The reconstructed hydrological series value on the d day of the month; Indicates the mutation month in year t The original daily scale hydrological series value of day d of the month;

[0183] Through the above steps, the restored sequence values ​​of the daily hydrological series of the hydrological station (May 1, 1952 to December 31, 2022) are obtained.

[0184] In summary, this method achieves a smooth transition in a multi-mutation point scenario through a systematic mutation point detection and deviation balance allocation mechanism, avoiding distribution deviations caused by cumulative errors. In particular, it can better handle the problem of reappearance or restoration of non-stationary daily-scale hydrological series, reduce the deviations caused by mutations, and greatly improve the accuracy, adaptability and application value of daily hydrological series reconstruction.

[0185] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0186] It should be noted that although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A daily hydrological series reconstruction method based on change detection and deviation correction, characterized by: The following steps are involved: Step S1, mutation point diagnosis: the daily hydrological series is processed into a monthly hydrological series, and the mutation points in the monthly hydrological series are diagnosed by the Pettitt test and the Wilcoxon rank sum test method to obtain the mutation point diagnosis result; Step S2, trend analysis: according to the mutation point diagnosis result obtained in step S1, the trend characteristics of the monthly hydrological series before and after the mutation point are analyzed by using the Mann-Kendall trend test method to ensure that there is no significant trend in the monthly hydrological series before and after the mutation point; Step S3, preliminary correction based on cumulative distribution function: for the monthly hydrological series of the mutation month without significant trend, the monthly hydrological series with mutation points are corrected based on the cumulative distribution function, so that the distribution characteristics of the monthly hydrological series before and after the mutation point are consistent, and the monthly hydrological series of the mutation month after preliminary correction is obtained; Step S4, deviation balance allocation: according to the cumulative corrected total deviation value between the monthly scale hydrological sequence of the mutation month and the original monthly scale hydrological sequence after preliminary correction in step S3, according to the ratio of the original daily scale hydrological sequence value of the non-mutation month to the sum of the original daily scale hydrological sequence values ​​of all non-mutation months, the cumulative corrected total deviation value of the mutation month is allocated to the daily scale hydrological sequence of the non-mutation month, so as to obtain the deviation balance allocation value of the daily hydrological sequence of the non-mutation month; Step S5, daily sequence reconstruction: apply the deviation balance distribution value of the daily hydrological sequence of the non-mutation month in step S4 to the daily hydrological sequence to obtain the reconstructed daily hydrological sequence of the non-mutation month; based on the ratio of the monthly hydrological sequence of the mutation month before and after the preliminary correction in step S3, scale the original daily hydrological sequence of the mutation month by the same proportion to obtain the reconstructed daily hydrological sequence of the mutation month.

2. The method for reconstructing daily hydrological series based on change detection and deviation correction according to claim 1, characterized in that: The mutation point diagnosis in step S1 is specifically as follows: Step S11, Pettitt test: identify the specific time of the monthly hydrological series mutation and determine the month and year when the mutation occurs; Step S12, Wilcoxon rank sum test: verify the significance of the mutation points of the monthly hydrological series to ensure the accuracy and scientificity of the identification.

3. The method for reconstructing daily hydrological series based on change detection and deviation correction according to claim 2, characterized in that: The trend analysis in step S2 is specifically as follows: Step S21, data extraction: extracting the monthly hydrological series of the mutation month into monthly hydrological series before and after the mutation point respectively; Step S22, trend test: using the Mann-Kendall trend test method to analyze the trend characteristics of the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point respectively; If there is no significant trend, the next step of correction will be entered; otherwise, further processing or exclusion is required to ensure that there is no significant trend in the monthly hydrological series before and after the mutation point.

4. The method for reconstructing daily hydrological series based on change detection and deviation correction according to claim 3, characterized in that: The preliminary correction based on the cumulative distribution function in step S3 is specifically: The monthly hydrological series after preliminary correction are divided into two scenarios: reappearance and restoration; Step S31, restore the situation: make the monthly hydrological series distribution characteristics before the mutation point consistent with the monthly hydrological series distribution characteristics after the mutation point, so as to achieve the restoration effect of the monthly hydrological series in the mutation month; the specific steps are: Step S311, cumulative distribution function definition: Let the monthly hydrological sequence before the mutation point be X, the monthly hydrological sequence after the mutation point be Z, and the cumulative distribution functions of the monthly hydrological sequence before the mutation point and the monthly hydrological sequence after the mutation point be F respectively. X and F Z , defined as follows: F X (x)=P(X≤x); F Z (z)=P(Z≤z); In the formula, F X (x) represents the cumulative distribution function of the monthly hydrological series X before the mutation point; F Z (z) represents the cumulative distribution function of the monthly hydrological sequence Z after the mutation point; P(X≤x) represents the probability that the data point of the monthly hydrological sequence X before the mutation point is less than or equal to x; P(Z≤z) represents the probability that the data point of the monthly hydrological sequence Z after the mutation point is less than or equal to z; Step S312, cumulative distribution function transformation: Based on the transformation of the cumulative distribution function, the monthly hydrological sequence X before the mutation point is adjusted so that the distribution characteristics of the monthly hydrological sequence X before the mutation point are consistent with the monthly hydrological sequence Z after the mutation point. The correction formula is as follows: ； In the formula, F X (X s ) represents the monthly hydrological sequence X before the mutation point s The cumulative distribution value of F Z -1 is the inverse cumulative distribution function, which converts the monthly hydrological series X before the mutation point into s The cumulative distribution value F X (X s ) Distribution characteristics of the monthly hydrological series Z after being mapped to the mutation point; Y 还现s represents the monthly hydrological series X before the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series Z after the mutation; Through steps S311 and S312, the cumulative distribution function is used to correct and eliminate the distribution difference between the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point, and the monthly hydrological series Y after preliminary correction is obtained. 还现 ; Step S32, restoring the situation: making the monthly hydrological series distribution characteristics after the mutation point consistent with the monthly hydrological series distribution characteristics before the mutation point, so as to restore the monthly hydrological series in the mutation month; the specific steps are: Step S321, cumulative distribution function definition: Let the monthly hydrological sequence before the mutation point be X, the monthly hydrological sequence after the mutation point be Z, and the cumulative distribution function of the monthly hydrological sequence before the mutation point and the monthly hydrological sequence after the mutation point be F X and F Z , defined as follows: F X (x)=P(X≤x); F Z (z)=P(Z≤z); In the formula, F X (x) represents the cumulative distribution function of the monthly hydrological series X before the mutation point; F Z (z) represents the cumulative distribution function of the monthly hydrological sequence Z after the mutation point; P(X≤x) represents the probability that the data point of the monthly hydrological sequence X before the mutation point is less than or equal to x; P(Z≤z) represents the probability that the data point of the monthly hydrological sequence Z after the mutation point is less than or equal to z; Step S322, cumulative distribution function inverse transformation: Based on the cumulative distribution function inverse transformation, the monthly hydrological series Z after the mutation point is adjusted to the distribution characteristics of the monthly hydrological series X before the mutation. The correction formula is as follows: ； In the formula, F Z (Z s ) represents the monthly hydrological sequence Z after the mutation point s The cumulative distribution value of F X -1 is the inverse cumulative distribution function, which converts the Z in the monthly hydrological series Z after the mutation point into s The cumulative distribution value F Z (Z s ) is mapped to the distribution characteristics of the monthly hydrological series X before the mutation point; Y 还原s Z represents the monthly hydrological series Z after the mutation point s The values ​​after preliminary restoration and correction are consistent with the distribution characteristics of the monthly hydrological series X before the mutation; Through steps S321 and S322, the cumulative distribution function is used to correct and eliminate the distribution difference between the monthly hydrological series before the mutation point and the monthly hydrological series after the mutation point, and the monthly hydrological series Y after preliminary restoration and correction is obtained. 还原 .

5. A daily hydrological series reconstruction method based on change detection and deviation correction according to claim 4, characterized in that: The deviation balance allocation in step S4 is specifically: The deviation balance distribution values ​​of the daily hydrological series in non-mutation months are divided into two scenarios: the present scenario and the restoration scenario; Step S41, the current situation: taking the most recent mutation year as the base year, according to the ratio of the original daily scale hydrological series values ​​of the non-mutation months before the base year including the base year to the sum of the original daily scale hydrological series values ​​of all non-mutation months before the base year including the base year, the accumulated corrected total deviation values ​​of the monthly scale hydrological series of multiple mutation months are distributed to the non-mutation months; Step S411, calculate the preliminary correction deviation of the monthly hydrological series of each mutation month and accumulate them, see the following formula: ； In the formula, It represents the cumulative corrected total deviation value of the monthly hydrological series in all mutation months; Represents the index of the monthly hydrological series in the mutation month; Indicates the month of mutation; Indicates the mutation month The preliminary correction deviation value of k represents the total number of monthly hydrological series in the mutation months without significant trend; Step S412, daily deviation value allocation: Taking the most recent mutation year as the base year, the total cumulative corrected deviation value is allocated to the daily hydrological series of the non-mutation months including the base year before the base year according to the ratio of the original daily hydrological series values ​​of the non-mutation months including the base year before the base year to the sum of the original daily hydrological series values ​​of all non-mutation months including the base year before the base year, to ensure a smooth transition of the data; the allocation formula is as follows: ； Among them, t represents the year, j represents the non-mutation month, and d represents the day; Indicates allocation to the non-mutation month in year t moon The correction deviation value of the day; Indicates the most recent mutation benchmark year; represents the starting year of the original daily-scale hydrological series of non-mutation month j; Indicates the most recent mutation base year and the sum of the original daily-scale hydrological series values ​​of all non-mutation months before the most recent mutation benchmark year; The most recent mutation benchmark year and the original daily-scale hydrological series value of the dth day of the non-mutation month j in the tth year before the most recent mutation benchmark year; Step S42, restoring the situation: taking the earliest mutation year as the base year, according to the ratio of the original daily-scale hydrological series values ​​of the non-mutation months after the base year including the base year to the sum of the original daily-scale hydrological series values ​​of all non-mutation months after the base year including the base year, the accumulated corrected total deviation values ​​of the monthly-scale hydrological series of multiple mutation months are distributed to the non-mutation months; Step S411, calculate the preliminary correction deviation of the monthly hydrological series of each mutation month and accumulate them, see the following formula: ； Step S422, daily deviation value allocation: Taking the earliest mutation year as the base year, the accumulated corrected total deviation value is allocated to the daily hydrological series of the non-mutation months including the base year after the base year according to the ratio of the original daily hydrological series values ​​of the non-mutation months after the base year to the sum of the original daily hydrological series values ​​of all non-mutation months after the base year including the base year, to ensure a smooth transition of the data; the specific allocation formula is as follows: ； in, Indicates the earliest mutation benchmark year; represents the end year of the original daily-scale hydrological series of non-mutation month j; The earliest mutation benchmark year and the sum of the original daily-scale hydrological series values ​​of all non-mutation months after the earliest mutation benchmark year; The earliest mutation benchmark year and the original daily-scale hydrological series value of the dth day of the non-mutation month j in the tth year after the earliest mutation benchmark year.

6. A daily hydrological series reconstruction method based on change detection and deviation correction according to claim 5, characterized in that: The reconstruction of the daily hydrological sequence in step S5 includes the reconstruction of the daily hydrological sequence in the non-mutation month and the reconstruction of the daily hydrological sequence in the mutation month, specifically: Step S51, reconstruction of the daily hydrological series in non-mutation months: The deviation balance distribution value of the daily hydrological sequence of the non-mutation month in step S4 is applied to the original daily scale hydrological sequence to obtain the reconstructed daily hydrological sequence of the non-mutation month. The specific calculation formula is as follows: ； in, represents the reconstructed hydrological sequence value of the non-mutation month j month d day in year t, It represents the original daily scale hydrological series value of the non-mutation month j month d day in year t; Step S52, reconstructing the daily sequence of the mutation month: Based on the ratio of the monthly hydrological series before and after preliminary correction in step S3, the original daily hydrological series of the mutation month is scaled to the corresponding daily hydrological series of the mutation month in the same proportion. The specific calculation formula is as follows: ； in: Indicates Year mutation month Monthly scale hydrological series values ​​after monthly correction; Indicates Year mutation month The original monthly scale hydrological series value before monthly correction; Indicates the mutation month in year t The reconstructed hydrological series value on the d day of the month; Indicates the mutation month in year t The original daily scale hydrological series value of day d of the month; Through steps S51 and S52, the reconstruction of the daily hydrological sequence of the non-mutation month and the mutation month is completed respectively, and finally a complete daily hydrological sequence reconstruction result is formed, realizing the restoration and recovery of the daily hydrological sequence.

Images