Rainfall runoff simulation method for performing session flood data cleaning based on dynamic threshold

Abstract

The invention discloses a rainfall runoff simulation method for performing session flood data cleaning based on a dynamic threshold value. The rainfall runoff simulation method comprises the following steps: collecting actually measured rainfall and section flow data of a target drainage basin in a multi-year hour scale; performing flood event segmentation on the continuous rainfall runoff data sequence of the target drainage basin based on the flow hydrograph morphological characteristics, and extracting rainfall runoff data corresponding to each flood session; through preliminary rainfall screening, rainfall runoff total amount comparison screening and fine screening based on runoff coefficient truncation analysis, abnormal flood sessions with unreasonable magnitude relationships in the rainfall-runoff data are identified and eliminated; by screening key time points in the flood process, flood session data with an abnormal time sequence is eliminated; and taking the cleaned flood data as the input of the session hydrological model, and carrying out normal simulation. According to the method, the flood data cleaning process is standardized, unreasonable flood session data are eliminated in advance, input data are ensured to be clean, and thus the flood forecasting precision is improved.

Description

Technical Field

[0001] The present invention belongs to the method for processing flood model data, and particularly relates to a rainfall-runoff simulation method for cleaning flood event data based on a dynamic threshold. Background Art

[0002] Flood forecasting, as the main non-engineering measure for reducing flood risks, is of great significance for reasonably predicting the formation of flood disasters and risk characterization, and scientifically defending against flood disasters. During the formation process of floods, affected by multiple factors such as the topographic features of the basin, the underlying surface vegetation, and human activities, the high spatiotemporal variability and nonlinearity of the runoff process also pose a huge challenge to the accuracy of flood forecasting. How to improve the accuracy of flood forecasting has become a key issue in the fields of hydrometeorology and disaster prevention and control.

[0003] To improve the accuracy of flood forecasting, the current main research directions include three major parts: input data, forecasting model structure, and parameter calibration. Combining the actual operation of the forecasting project and the experience of project managers, abnormal or incorrect flood data in the samples often greatly restricts the forecasting effect. Unreasonable data input cannot be fundamentally improved through the process real-time correction technology, and even affects the forecasting results and error correction directions of the remaining correct flood events.

[0004] For data input, error correction technology can be used to perform real-time correction on both rainfall and runoff data. Among them, multi-source data fusion is one of the effective means to improve the performance of hydrological forecasting simulation. Currently, the correction methods mainly focus on correcting according to the error relationship between the already forecast results and the measured runoff during or after the forecasting process, and try to attribute the error contribution rate in order to improve in the next forecast. Although these methods have good effects, most of them require complex mathematical statistics knowledge for operation, with a cumbersome process and low efficiency. In addition to multi-source data fusion, there are few methods to improve the data quality through technical means at the beginning of the model input to improve the forecasting accuracy. Summary of the Invention

[0005] The purpose of the present invention is to propose a rainfall-runoff simulation method for cleaning flood event data based on a dynamic threshold, by creating a scientific and feasible method for cleaning flood event data, standardizing the flood data cleaning process, and eliminating unreasonable flood event data in advance, so as to solve the above technical problems.

[0006] To solve the above technical problems, the present invention is realized through the following solutions: A rainfall-runoff simulation method for cleaning flood event data based on a dynamic threshold, comprising the following steps: Step S1: Data collection and compilation: Collect the measured rainfall and cross-section runoff data of the target basin at an hourly scale for many years.

[0007] Step S2: Flood event segmentation: Based on the morphological characteristics of the flow process line, the continuous rainfall and runoff data sequence of the target basin is segmented into flood events, and the rainfall and runoff data corresponding to each flood are extracted to form an independent flood event data sequence. The morphological characteristics of the flow process line (such as the rising point, flood peak, water retreat point, etc.) can reflect the occurrence, development and retreat process of the flood. By segmenting the flood event, the rainfall and runoff information of each flood can be accurately obtained, providing a basis for subsequent cleaning and simulation.

[0008] Step S3: Magnitude cleaning and correction: Through preliminary rainfall screening, rainfall-runoff total volume comparison screening, and fine screening based on runoff coefficient truncation analysis, abnormal flood events with unreasonable magnitude relationships in rainfall-runoff data are identified and eliminated.

[0009] Step S4: Process cleaning and correction: By screening the key time points (rising point, peak time) in the flood process, the flood event data with abnormal time sequence after screening in step S3 are eliminated; ensuring that the flood process conforms to the natural hydrological laws and providing data with reasonable physical meaning for subsequent simulations.

[0010] Step S5: Simulation and effect evaluation: The cleaned flood data is used as the input of the hydrological model for simulation, and the simulation effect is evaluated using evaluation indicators. The occurrence and development of floods can be predicted through simulation, providing a basis for decision-making for flood prevention and disaster reduction; the quality of data cleaning and the applicability of the model can be tested by using accuracy evaluation indicators to evaluate the simulation effect, providing a reference for further improving the model and data processing methods.

[0011] Further optimization, in step S1, the collected data is required to be continuous in time and meet the consistency requirements. For data sequences affected by human activities, runoff restoration calculation is performed and compiled. Human activities (such as water conservancy project construction, water resources development and utilization, etc.) will change the natural runoff process of the basin. These effects can be eliminated through runoff restoration calculation, so that the data can better reflect the natural hydrological characteristics of the basin.

[0012] The Thiessen polygon method is used to calculate the target basin surface rainfall to form a continuous rainfall runoff data sequence for the target basin. The Thiessen polygon method can reasonably distribute the rainfall data of each rain gauge in the basin to the entire basin, and obtain more accurate basin surface rainfall data. The Thiessen polygon method is an existing technology and will not be described in detail.

[0013] During the collection and compilation process, understanding the basin profile helps to understand the basin's hydrological characteristics and possible influencing factors. Checking the consistency of the data can ensure data quality. Large basins, due to their large area and complex impacts of human activities, require runoff restoration calculations to ensure data reliability.

[0014] Further optimization, the step S3 specifically includes: Step S3.1: Preliminary rainfall screening: Determine whether there is rainfall in the time series of each flood event. If there is no rainfall, it is determined that the rainfall magnitude is unqualified and excluded; if there is rainfall, proceed to Step S3.2 for the next magnitude cleaning. Rainfall is an important factor in forming floods. If there is no rainfall in the flood time series, it indicates that there may be data anomalies or the basic conditions for flood formation are not met in this event, so it is excluded.

[0015] Step S3.2: Comparison and screening of total rainfall runoff. By comparing the total rainfall and total runoff, further screen out the flood events with unreasonable rainfall-runoff relationships. Specifically, it includes: Step S3.2.1: Calculate the total rainfall P for each event 总 : ; where n is the number of time intervals, that is, the number of time periods into which a rainfall process is divided; is the time interval, p i represents the average rainfall intensity in the i th time period.

[0016] Step S3.2.2: Calculate the total flood runoff R for each event 总 ; ; where Q t represents the cross-sectional flow at time t .

[0017] Step S3.2.3: Calculate the net flood runoff R for each event net : R net =R 总 -R b -R pre ; ; ; where Q0 is the initial recession flow of the corresponding event, Q b is the base flow, Q p is the previous recession flow; λ is the recession coefficient, t 0 is the time corresponding to the flood rise point of this event, t e is the time corresponding to the end point of the previous recession, t f is the time corresponding to the end point of this flood event; Rb represents base flow; R pre Indicates the previous period of runoff.

[0018] Step S3.2.4: Introduce the sgn() function to perform R net and P 总 Compare, ; δ = R net -P 总 ; The corresponding flood events with sgn(δ) value of 1 are retained, and the rest of the flood event data are discarded, and then the next step of magnitude cleaning is entered.

[0019] Step S3.3: Based on the magnitude cleaning of runoff coefficient truncation analysis, further remove the events with unreasonable rainfall-runoff magnitude relationship to improve data quality. Specifically include: Step S3.3.1: Calculate the flood runoff coefficient: Use the formula C = R net / P 总 Calculate the runoff coefficients of all floods screened in step S3.2; then sort the runoff coefficients in ascending order to obtain an ordered array {C j | 1≤j≤N}, where N is the total number of flood events screened in step S3.2.

[0020] Step S3.3.2: Set the percentage ρ 1% and ρ 2%, and then calculate the characteristic quantile m k , and the corresponding runoff coefficient sequence quantile G ρk% ; m k =( N +1)× ρ k% ; ; Where K = 1, 2; Indicates floor operation.

[0021] Step S3.3.3: Determine reasonable value boundaries: Keep C j ∈[G ρ 1% , G ρ 2% ] corresponding flood event data, eliminating the cases where C is less than G ρ 1% and greater than G ρ 2%The data of abnormal flood events are thus completed with magnitude cleaning and calibration. By calculating the quantiles and defining the boundaries of reasonable values, the normal range of runoff coefficients can be determined to exclude outliers.

[0022] Further optimization, ρ 1% is 5%, ρ 2% is 95%.

[0023] Further optimization, step S4 specifically includes: Step S4.1: Screening of the rising point: Identify the time point t0 corresponding to the rising point of each flood and the rainfall start time point t p , compare their sequence. If t0 < t p , then consider this flood event as an abnormal flood and eliminate it; if t0 ≥ t p , then enter the next process of cleaning. By screening the rising point time, abnormal flood events with illogical rising point times are excluded, ensuring the rationality of flood data in terms of time sequence, improving the quality and reliability of the data, and providing a more accurate data basis for subsequent simulations and analyses.

[0024] Step S4.2: Screening of the peak occurrence time: Identify the peak occurrence time t Qmax of each flood and the peak rainfall occurrence time t Pmax , compare their sequence. If t Qmax < t Pmax , then consider this flood event as an abnormal flood and eliminate it; if t Qmax ≥ t Pmax , then retain this flood event, thus completing the cleaning of flood event data.

[0025] Generally, the flood peak should occur after the rainfall peak. If the flood peak occurrence time is earlier than the rainfall peak occurrence time, it does not conform to the natural process of rainfall runoff, so this event is excluded. By screening the peak occurrence time, abnormal flood events with illogical peak occurrence time relationships are excluded, ensuring the rationality of flood data in terms of peak occurrence time, improving the quality and reliability of the data, and further providing a more accurate data basis for subsequent simulations and analyses.

[0026] Further optimization, in step S5, the simulation effect is evaluated using the relative error of flood peak Rep, the relative error of peak occurrence time Ret, the relative error of flood volume Rew, and the Nash efficiency coefficient NSE as indicators.

[0027] Compared with the prior art, the beneficial effects of the present invention are: 1. The present invention forms a dual screening mechanism through magnitude cleaning and process cleaning, which not only pays attention to the magnitude relationship between rainfall and runoff, but also ensures the rationality of the time sequence of the flood process, standardizes the flood data cleaning process, eliminates unreasonable flood event data in advance, ensures that the input data is clean, and thus improves the accuracy of flood forecasting.

[0028] 2. The method described in the present invention does not perform complex and arduous error attribution or numerical contribution rate calculation. The method is simple, the physical mechanism is clear, and it does not rely on subsequent real-time simulation (forecast) feedback. It is independent, stable, and has strong anti-interference ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a flow chart of the rainfall-runoff simulation method for cleaning flood data based on dynamic thresholds according to the present invention; Figure 2 This is the distribution map of flood runoff coefficients for the screened events; Figure 3 This is a schematic diagram of the flood process on 2020070413; Figure 4 This is a schematic diagram of the flood process No. 2021072108. DETAILED DESCRIPTION

[0030] The technical solution of the present invention is described in detail below in conjunction with embodiments, but the protection scope of the present invention is not limited to the embodiments.

[0031] In this example, a small and medium-sized river basin in a key flood-prone area in southwest China was selected as a case study, and the Xin'anjiang model (the Xin'anjiang model is a hydrological model proposed by Professor Zhao Renjun's team at East China University of Water Resources (now Hohai University)) was used to simulate the rainfall-runoff process.

[0032] In this embodiment, if Figure 1 As shown in FIG. 1 , a rainfall runoff simulation method for cleaning flood data based on a dynamic threshold comprises the following steps: Step S1: First, a hydrological survey is conducted on the small and medium-sized watersheds. Through investigation and data review, it is determined that there are no new water conservancy projects and human activities in the basin after 2015. Therefore, the collected continuous rainfall and runoff data sequences meet the representativeness, reliability and consistency review and can be used for the subsequent hydrological simulation.

[0033] In this embodiment, the Thiessen polygon method is used to compile rainfall data, and finally a continuous time series of data from 2017.6 to 2022.10 is obtained.

[0034] Step S2: Select 48 flood events from the continuous time series data according to flood characteristics, extract the corresponding rainfall and runoff flow data, and form a separate excel file for each flood event, including three columns of data: time, rainfall, and runoff. The flood is named after its starting time.

[0035] Step S3: Magnitude cleaning and calibration, specifically including: Step S3.1: Preliminary rainfall screening: Conduct the magnitude cleaning section for flood events. Check whether there is rainfall in the extracted flood events, and calculate the total rainfall P for each flood event 总 . It is found that the total rainfall of the two flood events on July 6, 2017, 18:00 and May 24, 2019, 11:00 is 0, which is determined to have a false rainfall-runoff relationship and is removed. The remaining 46 reasonable flood events enter the next magnitude cleaning through this cleaning.

[0036] Step S3.2: Comparison and screening of total rainfall and runoff: First, calculate the total flood runoff R for each rainfall 总 , according to the formula R net = R 总 - R b - R pre Remove the base flow and the part where the previous flood has not receded completely.

[0037] Find the dry period without rain from the long time series, and obtain the base flow value of 2.35 m 3 / s. The total rainfall of all flood events has been calculated in Step S3.1. Calculate the sgn value of the difference between the total rainfall and runoff of each flood event. It is found that the sgn values of 4 flood events are less than or equal to 0, which are determined to have an incorrect rainfall-runoff relationship and are excluded. The remaining 42 flood events are reasonable and then enter the next magnitude cleaning.

[0038] Step S3.3: Use the runoff coefficient to remove the flood events with unreasonable rainfall-runoff magnitude relationships. Since the underlying surface of all flood generation and concentration is unified, the runoff coefficient should not show too large a difference in the frequency distribution. Specifically including: Step S3.3.1: Calculation of flood runoff coefficient: Use the formula C = R net / P 总 to calculate the runoff coefficients of all 42 selected flood events. Since the name is too long, the flood number is used instead of the flood name. The runoff coefficient distribution is as Figure 2 shown.

[0039] Step S3.3.2: Calculation of quantiles: Sort the runoff coefficients in ascending order, and calculate the corresponding characteristic quantiles after calculating the 5% and 95% m k , as well as the quantile G ρk% of the corresponding runoff coefficient sequence.

[0040] m k =( N +1)× ρ k% ; ; Calculate the characteristic quantile G of the flood sequence 5% =2.353, G 95% =5.287, K = 1, 2.

[0041] Step S3.3.3: Determine the reasonable value boundary. Use the truncation method to determine the reasonable value boundary: ; Identify and eliminate 19 flood events that do not meet the boundary constraint conditions, and output 23 reasonable flood events, completing the magnitude cleaning and correction.

[0042] Step S4: Process cleaning and correction. By screening key time points in the flood process, eliminate the flood event data with abnormal time sequence after screening in step S3. Specifically include: Step S4.1: Rising point screening: Initially traverse 23 flood events for rising point screening, identify the time point t0 corresponding to the rising point of each flood event and the rainfall start time point t p , compare their sequence. If t0 < t p , then regard this flood event as an abnormal flood and eliminate it; if t0 ≥ t p , then proceed to the next process cleaning.

[0043] In this embodiment, a total of 1 flood event is eliminated in this process cleaning, and the remaining 22 flood events pass the cleaning. As Figure 3 shown: The 2020070413 flood is observed as a single-peak flood. From the part marked by the green dot in the figure, it can be seen that the flood peak rising point t0 is 18:00:00 on July 5, 2020, and the corresponding rainfall during this period is 0. The rainfall starting point t p is 06:00:00 on July 6, 2020, which belongs to the situation where the flood rising point is before the rainfall occurs, and the rainfall-runoff relationship is incorrect, so it is removed.

[0044] Step S4.2: Peak appearance time screening: Identify the flood peak appearance time t Qmax and the rainfall peak appearance time t Pmax , compare their sequence. If t Qmax < t Pmax , then regard this flood event as an abnormal flood and eliminate it; if t Qmax ≥ t Pmax , then retain this flood event, and thus complete the cleaning of the flood event data.

[0045] In this embodiment, this cleaning process removes a total of 5 floods. Take flood No. 2021072108 as an example. Figure 4 As shown, it can be clearly seen that the peak time of the rain peak is 14:00:00 on July 22, 2021, while the peak time of the flood peak is 05:00:00 on July 22, 2021. The flood peak appears before the rain peak, which is inconsistent with the runoff generation mechanism and the rainfall-runoff relationship is unreasonable, so it is removed. At this point, the flood scene cleaning is completed, and a total of 17 high-quality flood scenes are output.

[0046] Step S5: Select the Sanshuiyuan Xin'anjiang model for hydrological simulation, select the flood peak relative error Rep, peak time relative error Ret, flood volume relative error Rew and Nash efficiency coefficient NSE as accuracy assessment indicators, and use the 48 floods before dynamic threshold data cleaning and the 17 floods after cleaning as input for hydrological process simulation and accuracy assessment.

[0047] According to GB / T22482-2008 "Specifications for Hydrological Information Forecasting", the permissible error of flood peak forecast is 20% of the measured flood peak flow, and the peak time is 30% of the time interval between the forecast time and the measured flood peak time as the permissible error. When the permissible error is less than 3h or a calculation period is longer, 3h or a calculation period is used as the permissible error. Here, since the calculated permissible error is generally less than 3h, 3h is taken as the permissible error of the peak time in this embodiment. The runoff forecast uses 20% of the measured value as the permissible error, and calculates each qualified rate separately.

[0048] Based on the above analysis, the flood data before and after data cleaning were input into the model for simulation, and the results were output and the accuracy was evaluated. The accuracy evaluation tables of the Xin'anjiang model before flood cleaning are shown in Tables 1, 2, 3, and 4, and the Xin'anjiang model after flood cleaning is shown in Tables 5, 6, 7, and 8.

[0049] Table 1 Evaluation of the simulation accuracy of the peak flow of the Xin'anjiang model before cleaning

[0050] Table 2 Evaluation of the simulation accuracy of the peak time of the Xin'anjiang model before cleaning

[0051] Table 3 Evaluation of flood simulation accuracy of Xinanjiang model before cleaning

[0052] Table 4 Evaluation of the certainty coefficient of the Xinanjiang model before cleaning

[0053] Table 5 Evaluation of the simulation accuracy of the peak flow of the Xin'anjiang model after cleaning

[0054] Table 6 Evaluation of the simulation accuracy of the peak appearance time of the Xin'anjiang model after cleaning

[0055] Table 7 Evaluation of the simulation accuracy of the flood volume of the Xin'anjiang model after cleaning

[0056] Table 8 Evaluation of the coefficient of determination of the Xin'anjiang model after cleaning

[0057] From Tables 1-4, it can be seen that rainfall-runoff simulations were carried out on a total of 48 floods before cleaning for individual floods. From the simulation results of peak flow runoff, the qualified rate was only 39.58%. The maximum simulation error was for the flood on August 24, 2017, No. 08, with a relative error of 315.2%. The qualified rate of peak appearance time simulation was only 29.17%. The flood event with the largest error was on August 3, 2017, No. 03, and the peak appearance time was advanced by 41 h. The qualified rate of flood volume simulation was 35.42%. The flood event with the largest relative error was the flood on August 24, 2017, No. 08, reaching -173.9%. Generally speaking, the simulation effect of individual floods before cleaning was poor, and none of them met the minimum Class C accuracy requirements in the specifications and had no reference significance. Only the simulation accuracies of three floods, namely on July 24, 2020, No. 00, on September 3, 2019, No. 00, and on July 10, 2020, No. 15, were good and reached Class A accuracy. The rest of the flood events were unqualified and had large errors.

[0058] From Tables 5-8, it can be seen that rainfall-runoff simulations were carried out on a total of 17 floods after cleaning based on the method described in the present invention. The qualified rate of peak flow runoff simulation increased to 88.24%. The flood event with the largest error was on September 15, 2018, No. 00, with a relative error of -46.08%. The qualified rate of peak appearance time increased to 52.94%. The flood event with the largest error was on October 16, 2019, No. 00, and it was delayed by 12 h. The qualified rate of flood volume increased to 94.12%. The only unqualified flood event was on September 15, 2018, No. 00, with a relative error of 35.1%. Generally, the simulation effect of individual floods was good, and the simulation result of peak appearance time was poor. Overall, it could reach above Class B accuracy, and the improvement effect of the simulation accuracy compared with that before cleaning was significant.

Claims

1. A rainfall-runoff simulation method for cleaning flood data of each flood event based on a dynamic threshold, characterized in that The steps include: Step S1: Data collection and compilation: Collect the measured rainfall and cross-sectional runoff data of the target basin at the hourly scale for many years; Step S2: Flood event segmentation: Based on the flow process line morphological characteristics, the continuous rainfall runoff data sequence of the target basin is segmented into flood events, and the rainfall runoff data corresponding to each flood event is extracted to form an independent flood event data sequence; Step S3: Magnitude cleaning and correction: Through preliminary rainfall screening, rainfall-runoff total volume comparison screening, and fine screening based on runoff coefficient truncation analysis, abnormal flood events with unreasonable magnitude relationships in rainfall-runoff data are identified and eliminated; Step S4: Process cleaning and correction: by screening the key time points in the flood process, the flood event data with abnormal time sequence after screening in step S3 are eliminated; Step S5: Simulation and effect evaluation: The cleaned flood data is used as the input of the hydrological model for simulation, and the simulation effect is evaluated using evaluation indicators.

2. The rainfall-runoff simulation method for cleaning flood data of each flood event based on a dynamic threshold according to claim 1, wherein, In step S1, the collected data is required to be continuous in time and meet the consistency requirements; for the data series affected by human activities, runoff restoration calculation and compilation are performed; the Thiessen polygon method is used to calculate the target basin surface rainfall to form a continuous rainfall runoff data series of the target basin.

3. The rainfall-runoff simulation method for cleaning flood data of each flood event based on a dynamic threshold according to claim 2, characterized in that, The step S3 specifically includes: Step S3.1: Preliminary rainfall screening: determine whether there is rainfall in the time series of each flood event. If not, the rainfall level is determined to be unqualified and eliminated; if there is rainfall, proceed to step S3.2 for the next level cleaning; Step S3.2: Comparison and screening of total rainfall runoff, including: Step S3.2.1: Calculate the total rainfall amount P for each rainfall event 总 : ； Among them, n is the number of time periods, that is, the number of multiple time periods into which a rainfall process is divided; is the time period interval, p i represents the i average rainfall intensity within the nth time period; Step S3.2.2: Calculate the total flood runoff volume R for each flood event 总 ; ； Among them, Q t represents t the cross-sectional flow rate at a certain moment, t 0 is the time corresponding to the rising point of the flood in this flood event, t f is the time corresponding to the termination point of the flood in this flood event; Step S3.2.3: Calculate the net runoff R of each flood event net : R net =R 总 -R b -R pre ; ； ； Among them, Q0 is the starting recession flow of the corresponding flood event, Q b is the base flow, Q p is the previous recession flow; λ is the recession coefficient, t 0 is the time corresponding to the start point of rising of the flood event, t e is the time corresponding to the end point of the previous recession, t f is the time corresponding to the end point of the flood event; R b represents the base flow runoff; R pre represents the previous recession runoff; Step S3.2.4: Introduce the sgn() function to perform the comparison between R net and P 总 comparison ； δ = R net -P 总 ; Keep the corresponding flood events with sgn(δ) value of 1, remove the rest of the flood event data, and then proceed to the next level cleaning; Step S3.3: Magnitude cleaning based on runoff coefficient truncation analysis, specifically including: Step S3.3.1: Calculate the flood runoff coefficient: Use the formula C = R net / P 总 to calculate the flood runoff coefficients of all flood events screened in Step S3.2; then sort the runoff coefficients in ascending order to obtain an ordered array {C j | 1 ≤ j ≤ N}, where, N is the total number of flood events screened in Step S3.2; Step S3.3.2: Set the percentage ρ 1% and ρ 2%, and then calculate the characteristic quantile m k , and the quantile G of the corresponding runoff coefficient sequence ρk% ; m k =( N +1)× ρ k% ； ； where K = 1, 2; represents the floor operation; Step S3.3.3: Determine the reasonable value boundary: Retain the flood event data corresponding to C j ∈[G ρ 1% , G ρ 2% , and eliminate the abnormal flood event data where C is less than G ρ 1% and greater than G ρ 2% . Thus, the magnitude cleaning and calibration are completed.

4. The rainfall runoff simulation method for cleaning flood data based on dynamic threshold according to claim 3 is characterized in that: ρ 1% is 5%, ρ 2% is 95%.

5. The rainfall-runoff simulation method for cleaning flood data of each event based on a dynamic threshold according to claim 4, wherein The step S4 specifically includes: Step S4.1: Identification of the rising point: Identify the time point t0 corresponding to the rising point of each flood and the rainfall start time point t p , compare their order. If t0 < t p , then regard this flood as an abnormal flood and eliminate it; if t0 ≥ t p , then proceed to the next step of process cleaning; Step S4.2: Peak appearance time screening: Identify the peak appearance time t of the flood peak for each flood Qmax and the peak appearance time t Pmax of the rainfall peak, compare their order. If t Qmax < t Pmax , then consider this flood as an abnormal flood and eliminate it; if t Qmax ≥ t Pmax , then retain this flood. Thus, the data cleaning of flood events is completed.

6. The rainfall runoff simulation method for cleaning flood data based on dynamic threshold according to claim 5 is characterized in that: In step S5, the simulation effect is evaluated using the flood peak relative error Rep, the peak appearance time relative error Ret, the flood volume relative error Rew and the Nash efficiency coefficient NSE as indicators.

Images