A prediction method for extreme rainfall based on future climate models

Abstract

The invention discloses a prediction method for extreme rainfall based on a future climate model, comprising the following steps: using a Delta downscaling method to correct the deviation of historical rainfall data and future rainfall data of a climate model, and verifying the error of extreme precipitation indexes observed in historical periods and downscaling simulations by drawing a Taylor diagram, selecting a climate model suitable for extreme precipitation index analysis in a study area, extracting historical and future extreme precipitation index sequences of the climate model, and using them to estimate parameters using a non-negative function of hydrometeorology and its improved form as an alternative function, selecting a probability distribution model for the long-term distribution of extreme precipitation indexes in a sub-region of the study area, and using a rainstorm formula in combination with a Chicago rain type method to obtain a design rainfall process under a specific return period under various future scenarios in the sub-region of the study area. The method improves the accuracy of rainfall prediction by designing extreme rainfall prediction of future climate models by considering extreme precipitation alone.

Description

Technical Field

[0001] The present invention relates to the technical field of rainfall prediction, and in particular to a method for predicting extreme rainfall based on future climate patterns. Background Art

[0002] The Sixth Assessment Report of the Intergovernmental Panel on Climate Change pointed out that from 2015 to 2100, the global warming trend will intensify, regional climate change will be more obvious, and its impact on the climate system is irreversible. Results from observations and climate model simulations show that climate warming will lead to frequent extreme climate events such as heavy precipitation, high temperature, and drought. For extreme precipitation, it mainly affects its frequency, intensity, duration, spatial distribution and other characteristics. The assessment and investigation of extreme precipitation events will help to formulate adaptive strategies and reduce the risk of such events. The Global Climate Model and Coupled Model Intercomparison Project developed by the World Climate Change Research Program is the most powerful tool for predicting future climate change, and has been used by many studies to reproduce current and predict future extreme precipitation events. Dong (2021) used CMIP6 to simulate 7 extreme precipitation indices in Asia and compared them with 28 models in CMIP5. The results showed that CMIP6's ability to reproduce the spatiotemporal distribution of extreme precipitation indices has been significantly improved compared with CMIP5. Faye and Akinsanola (2022) explored the simulation performance of 16 CMIP6 models for the extreme precipitation index in West Africa. The results showed that the models have different simulation capabilities for different indices, and also show significant simulation differences in different seasons and regions.

[0003] Extreme precipitation events refer to precipitation events in which the difference between the measured value and the normal value reaches or exceeds a specific threshold. At present, the most commonly used assessment of extreme precipitation events is the extreme precipitation index proposed by the Climate Expert Group, which can comprehensively evaluate precipitation characteristics in terms of precipitation, intensity, and number of days. Extreme climate events are ultimately about climate extremes. From the perspective of probability theory, the distribution law of extreme values ​​must be considered first when studying extreme climate events and predictability. Choosing the correct probability distribution function (PDF) to simulate the occurrence of high-frequency extreme precipitation is a requirement for estimating the probability of high-impact events and studying climate extremes within a wider range of climate change and variation. Different countries and regions recommend different distributions. For example, Pearson Type 3 (P-III) distribution is recommended in China; Log-Pearson 3 is used in the United States; and Generalized Extreme Value (GEV) distribution is usually used in Europe. The Sixth Assessment Report (AR6) of the Intergovernmental Panel on Climate Change (IPCC) pointed out that extreme hydrological events are more susceptible to climate change than ordinary events. However, it is not clear which theoretical distributions are consistent with the probability distribution characteristics of extreme hydrological events. In the past, the choice of distribution type mostly relied on empirical selection, which may be affected by region and variables.

[0004] Design rainfall is not only an essential element for ensuring flood safety in many infrastructure development projects, but also a fundamental input to hydrological models that quantify the impact of precipitation on other major hydrological components (evapotranspiration, soil water storage, snowmelt runoff, runoff). Design rainfall is defined as the calculated precipitation with a statistically determined return period "T", and the design rainfall for a given area is usually estimated based on a large amount of historical measured rainfall data.

[0005] Although the impact of climate change on future design rainfall on daily or longer time scales has been noted, the impact of climate change differs between total precipitation and extreme precipitation, thus ignoring the impact of potential changes in design extreme precipitation and reducing the accuracy of rainfall prediction. Summary of the invention

[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides a prediction method for extreme rainfall based on future climate models. By considering extreme precipitation and utilizing the Delta downscaling method combined with the Chicago rain pattern method, the future design extreme precipitation process is obtained, thereby providing data input for exploring the impact of climate change on hydrological processes at a fine scale, so as to improve the accuracy of rainfall prediction.

[0007] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0008] A prediction method for extreme rainfall based on future climate models includes the following steps:

[0009] S1. Delta downscaling method is used to correct the deviation of historical rainfall data and future rainfall data of climate model, and the error of extreme precipitation index between historical observation and downscaling simulation is verified by drawing Taylor diagram, so as to select the climate model suitable for extreme precipitation index analysis in the study area;

[0010] S2, extracting the historical and future extreme precipitation index sequences of the climate model selected in step S1 that are suitable for the extreme precipitation index analysis of the study area, using the non-negative function of hydrometeorology and its improved form as the candidate function, using the extreme precipitation index sequence to estimate the parameters of the candidate function, and selecting a probability distribution model for the long-term distribution of the extreme precipitation index of the sub-region of the study area;

[0011] S3. Based on the climate model selected in step S1 for the analysis of the extreme precipitation index in the study area and the probability model selected in step S2 for the long-term distribution of the extreme precipitation index in the sub-region of the study area, the rainstorm formula is used in combination with the Chicago rain pattern method to obtain the design rainfall process under a specific return period under various future scenarios in the sub-region of the study area.

[0012] Furthermore, step S1 specifically includes:

[0013] S11. Use the bilinear interpolation method to interpolate the historical precipitation data of the climate model and the ensemble results of the future precipitation data to a 0.25°×0.25° horizontal grid, and calculate the change factor of the climate model based on the precipitation data using the change ratio, that is:

[0014]

[0015] Among them, CFs represents the change factor of the climate model, GCMf and GCMh represent the climate variable values ​​of the global climate model base period and projection period at a certain time scale, respectively. They represent the average values ​​of climate variables in the global climate model base period and projection period at the specified time.

[0016] S12, according to the change factor of the climate model calculated in step S11, the change factor of the climate model is superimposed on the precipitation observation data using the Delta downscaling method to obtain the estimated value of the regional scale climate variable, that is:

[0017] GCMf i =CFs×OBS i

[0018] Among them, GCMf i represents the estimated value of the regional scale climate variable with i observations, OBS i Represents the observed value of a climate variable with i observations at a specified time scale, where i represents the number of observations

[0019] S13, plotting the simulation relationship between the climate model and the climate model data into a Taylor diagram, and plotting the Taylor diagram into a scatter plot based on polar coordinates, and using evaluation method indicators to evaluate the simulation results of the climate model and the climate model data;

[0020] S14. Use the temporal variability skill scoring index to evaluate the ability of climate models and climate model data to simulate interannual variability;

[0021] S15, ranking the climate models according to the size of the evaluation method index in step S13 and the size of the time variability skill scoring index in step S14, and using a comprehensive ranking index to rank the overall simulation performance of the climate models on the extreme precipitation index;

[0022] S16. According to the climate model ranking result of the comprehensive ranking index in step S15, a climate model suitable for analyzing the extreme precipitation index of the study area is selected.

[0023] Furthermore, in step S13, the simulation results of evaluating the climate model and the climate model data using the evaluation method index are:

[0024]

[0025] Among them, S represents the evaluation method index of climate model and climate model data, R represents the correlation coefficient between climate model and climate model data, and R 0 represents the maximum value of the correlation coefficient that can be achieved among all climate models, σ f Represents the ratio between the climate model and the standard deviation of the climate model data.

[0026] Furthermore, in step S14, the temporal variability skill scoring index is used to evaluate the simulation capability of the climate model and the climate model data for the interannual variability:

[0027]

[0028] Among them, IVS represents the temporal variability skill score index of the simulation ability of climate models and climate model data on interannual variability, and STD m 、STD o represent the standard deviation of the climate model and the time series of the climate model data, respectively.

[0029] Furthermore, in step S15, the overall simulation performance of the climate model on the extreme precipitation index is ranked using a comprehensive ranking index as follows:

[0030]

[0031] Among them, MR represents the comprehensive ranking index of the overall simulation performance of the climate model on the extreme precipitation index, m represents the total number of climate models, n represents the number of extreme precipitation indices, and rank i It indicates that the i-th extreme precipitation index is ranked according to the size of the evaluation method index S and the size of the time variability skill scoring index IVS.

[0032] Furthermore, step S2 specifically includes:

[0033] S21, extracting the historical and future extreme precipitation index sequences of the climate model selected in step S1 that are suitable for extreme precipitation index analysis in the study area;

[0034] S22, using the maximum likelihood method to construct a maximum likelihood function based on the extreme precipitation index sequence extracted in step S21 to estimate the parameters of the candidate function using the non-negative function of hydrometeorology and its improved form as the candidate function, and obtain the parameters of the appropriate candidate function;

[0035] S23, according to the parameters of the suitable candidate function obtained in step S22, the Kolmogorov-Smilov test method is used to determine whether each extreme precipitation index sequence conforms to the theoretical distribution of the candidate function, and the test statistics and graphical analysis method are used to judge the fitting quality of the candidate function, so as to obtain the candidate function with the best fitting effect and the extreme value distribution model;

[0036] S24. Using the extreme value distribution model of the candidate function with the best fitting effect obtained in step S23 as the probability distribution model of the long-term distribution characteristics of the extreme precipitation index in the sub-region of the study area, and using the probability distribution model to predict the design values ​​of extreme precipitation events under different recurrence periods.

[0037] Furthermore, step S3 specifically includes:

[0038] S31. Calculate the rainstorm intensity in the study area using the rainstorm formula based on precipitation and rainfall duration;

[0039] S32, using the Chicago rain pattern method to calculate the rainfall intensity before and after the peak of a certain minute according to the rainstorm intensity in the study area in step S31, and obtain the rainfall intensity at any time, and connect them to obtain an instantaneous rainfall process line;

[0040] S33, according to the design values ​​of the extreme precipitation events predicted in step S23 at different return periods, using GIS technology to extract the design values ​​of the extreme precipitation index SDII, SDII (90), and SDII (95) to the design values ​​of the sub-areas of the study area;

[0041] S34. According to the rainstorm intensity formula of the sub-area of ​​the study area, the design return period is input to obtain the surface rainfall design value;

[0042] S35, taking the instantaneous rainfall process line obtained by the Chicago rain type method in step S32 as the prototype, taking the ratio of the surface rainfall design value obtained in step S34 to the total amount of heavy rainfall in different return periods obtained by the heavy rainfall formula of each province as the magnification ratio, adopting the principle of amplification with the same ratio in different time periods, and calculating the design rainfall process under a specific return period of extreme precipitation index SDII, SDII(90), SDII(95) in a certain region in the historical period;

[0043] S36. Based on the change factor of the climate model obtained in step S11, multiply the change factor by the design rainfall process under a specific return period of the extreme precipitation index SDII, SDII(90), and SDII(95) in a certain region in the historical period in step S35 to obtain the design rainfall process under a specific return period under various scenarios in the future for a certain region.

[0044] Furthermore, the calculation formula for calculating the rainstorm intensity in the study area using the rainstorm formula in step S31 is:

[0045]

[0046] Among them, i represents the intensity of heavy rain, q represents precipitation, t represents rainfall duration, a represents rain force parameter, c represents rain force variation parameter, LgT represents the logarithm of the return period T, T represents the return period, b represents the rainfall duration correction parameter, and n represents the heavy rain attenuation coefficient.

[0047] Furthermore, the calculation formula for calculating the rainfall intensity before and after the peak in a certain minute in step S32 is:

[0048]

[0049]

[0050] Among them, i b represents the rainfall intensity before the peak, r represents the rain peak coefficient, t b represents the rainfall duration before the peak, i a represents the post-peak rainfall intensity, t a represents the duration of rainfall after the peak.

[0051] The present invention has the following beneficial effects:

[0052] 1. The present invention proposes a prediction method for extreme rainfall based on future climate models, which explores the impact of climate change on extreme precipitation from the perspective of probability theory, and uses the Delta downscaling method combined with the Chicago rain pattern method to obtain the future design extreme precipitation process, providing data input for exploring the impact of climate change on hydrological processes at a fine scale, so as to improve the accuracy of rainfall prediction;

[0053] 2. The present invention uses the maximum likelihood method, Kolmogorov-Smilov test method, etc. to explore the evolution law of extreme precipitation events and establish a reasonable extreme distribution model, and provides a recommended function for extreme event assessment in the study area;

[0054] 3. The present invention uses a multi-index comprehensive evaluation of the climate model's simulation performance of the extreme precipitation index in the study area through deviation correction. This method can be used to explore the simulation performance of different models and provide assistance for the improvement of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic flow chart of a method for predicting extreme rainfall based on future climate models proposed by the present invention;

[0056] Figure 2 Schematic diagram of the relative positions of various research areas in the embodiment;

[0057] Figure 3It is a Taylor schematic diagram of the CMIP6 climate model's ability to simulate the extreme precipitation index climate state in the study area from 1980 to 2014 in the embodiment;

[0058] Figure 4 It is a schematic diagram of S and IVS values ​​of the CMIP6 climate model's ability to simulate the extreme precipitation index climate state in the study area from 1980 to 2014 in the embodiment;

[0059] Figure 5 Schematic diagram of the evaluation results of the extreme precipitation index in the study area from 1980 to 2014 in the embodiment;

[0060] Figure 6 It is a schematic diagram showing the variation of the multi-model ensemble data set's ability to simulate the climate state of the extreme precipitation index in the study area with the average number of model ensembles in the embodiment;

[0061] Figure 7 This is a schematic diagram of the multi-year mean spatial pattern of the index describing precipitation in the study area under various historical and future scenarios in the embodiment;

[0062] Figure 8 A schematic diagram of the interannual variation of the index describing precipitation in the historical and future scenario study areas in the embodiment;

[0063] Fig. 9 It is a schematic diagram of the coordination between the empirical data and the theoretical curve in the embodiment;

[0064] Fig.10 This is a schematic diagram of the spatial pattern of the superior probability distribution of the extreme precipitation index series under various scenarios in the history and future of the study area in the embodiment;

[0065] Fig.11 It is a schematic diagram of the probability frequency percentage of the extreme precipitation index sequence under various scenarios in the history and future of the study area in the embodiment;

[0066] Fig.12 This is a schematic diagram of the design value of the 20-year return period of the extreme precipitation index under various scenarios in the history and future of the study area in the embodiment;

[0067] Fig.13 Schematic diagram of the original rain pattern of the historical measured rainstorm in each province in the embodiment;

[0068] Fig.14 It is a schematic diagram of the rainstorm process designed for future extreme rainfall in various provinces in the embodiment. DETAILED DESCRIPTION

[0069] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0070] like Figure 1 As shown, a prediction method for extreme rainfall based on future climate models includes the following steps S1-S3:

[0071] S1. Delta downscaling method is used to correct the deviation of historical rainfall data and future rainfall data of climate model, and the error of extreme precipitation index between historical observation and downscaling simulation is verified by drawing Taylor diagram, so as to select climate model suitable for extreme precipitation index analysis in the study area.

[0072] like Figure 2 As shown, Figure 2 Schematic diagram of the relative locations of the research areas. Figure 2 (a) shows the relative position of the Yellow River Basin; Figure 2 Middle (b) shows the relative positions of 92 daily precipitation stations in the Yellow River Basin; Figure 2 The middle (c) is the nine provinces in the Yellow River Basin.

[0073] In this embodiment, the Delta downscaling method is used to correct the deviation of the historical rainfall data and future rainfall data of the climate model, that is, the change factor is obtained by comparing the difference between the historical climate model data and the measured data, and it is assumed that the change factor remains unchanged in the future scenario, so as to achieve the purpose of correcting the future data. Then, the historical period is used as the verification period, and the extreme precipitation index observed and downscaled simulated in the historical period is error-verified according to the drawn Taylor diagram, the evaluation method based on the Taylor diagram, and the time variability skill score, so as to select the climate model suitable for the extreme precipitation index analysis of the study area. The study area described in this embodiment is the Yellow River Basin. By optimizing and evaluating the sixth generation global climate model (CMIP6) of the Yellow River Basin, the spatiotemporal distribution characteristics of extreme climate under different climate scenarios are analyzed.

[0074] Specifically, step S1 specifically includes S11-S16:

[0075] S11. Use the bilinear interpolation method to interpolate the historical precipitation data of the climate model and the ensemble results of the future precipitation data to a 0.25°×0.25° horizontal grid, and calculate the change factor of the climate model based on the precipitation data using the change ratio, that is:

[0076]

[0077] Among them, CFs represents the change factor of the climate model, GCMf and GCMh represent the climate variable values ​​of the global climate model base period and projection period at a certain time scale, respectively. Represents the average values ​​of climate variables in the global climate model base period and projection period at the specified time.

[0078] In this embodiment, the purpose of using the bilinear interpolation method to interpolate the historical precipitation data of the climate model and the aggregate results of the future precipitation data onto a 0.25°×0.25° horizontal grid is to maintain the consistency of data resolution.

[0079] S12, according to the change factor of the climate model calculated in step S11, the change factor of the climate model is superimposed on the precipitation observation data using the Delta downscaling method to obtain the estimated value of the regional scale climate variable, that is:

[0080] GCMf i =CFs×OBS i

[0081] Among them, GCMf i represents the estimated value of the regional scale climate variable with i observations, OBS i It represents the observed value of the climate variable with i observations at a specified time scale, where i represents the number of observations.

[0082] like Figure 3 As shown, Figure 3 Taylor diagram showing the ability of CMIP6 climate models to simulate the extreme precipitation index climate state in the study area from 1980 to 2014. Figure 3 There are 12 Taylor diagrams in total. For a single Taylor diagram, each point in the diagram represents a climate mode. The upper right corner of the diagram is the corresponding extreme precipitation index. The angular axis in the diagram is the spatial correlation between the simulated field and the observed field. The radial axis in the diagram is the spatial standard deviation (RMS deviation). The study area in this embodiment is the Yellow River Basin.

[0083] like Figure 4 As shown, Figure 4 Schematic diagram of S and IVS values ​​of the CMIP6 climate model's ability to simulate the extreme precipitation index climate state in the study area from 1980 to 2014. Figure 4 Each point in the figure represents a climate mode, and the corresponding extreme precipitation index is shown above the figure. In this embodiment, the study area is the Yellow River Basin.

[0084] S13. Plot the simulation relationship between the climate model and the climate model data into a Taylor diagram, and plot the Taylor diagram into a scatter plot based on polar coordinates. Use evaluation method indicators to evaluate the simulation results of the climate model and the climate model data.

[0085] The simulation results of evaluating the climate model and the climate model data using the evaluation method index in step S13 are:

[0086]

[0087] Among them, S represents the evaluation method index of climate model and climate model data, R represents the correlation coefficient between climate model and climate model data, and R 0 represents the maximum value of the correlation coefficient that can be achieved among all climate models, σ f Represents the ratio between the climate model and the standard deviation of the climate model data.

[0088] In this embodiment, the evaluation method index S is used to evaluate the simulation results of the climate model and the climate model data. The closer the evaluation method index S is to 1, the closer the climate model simulation value is to the measured value, that is, the closer the climate model is to the climate model data, and the better its simulation ability is.

[0089] S14. Use the temporal variability skill scoring index to evaluate the ability of climate models and climate model data to simulate interannual variability.

[0090] In step S14, the temporal variability skill scoring index is used to evaluate the simulation ability of the climate model and the climate model data for the interannual variability:

[0091]

[0092] Among them, IVS represents the temporal variability skill score index of the simulation ability of climate models and climate model data on interannual variability, and STD m 、STD o represent the standard deviation of the climate model and the time series of the climate model data, respectively.

[0093] like Figure 5 As shown, Figure 5 Schematic diagram of the extreme precipitation index assessment results in the study area from 1980 to 2014. Figure 5 (a) is a scatter plot of the comprehensive ranking index MR of the climate model on 12 extreme precipitation indices given by the S graph (X axis, spatial pattern) and the IVS graph (Y axis, interannual variation), that is, the comprehensive ranking index MR score of the climate model on the temporal variability skill scoring index IVS and the evaluation method index S; Figure 5(b) is the comprehensive score and the final ranking based on the average of the comprehensive ranking index MR scores on the evaluation method index S and the time variation skill scoring index IVS. The research area in this embodiment is the Yellow River Basin.

[0094] S15. Rank the climate models according to the size of the evaluation method index in step S13 and the size of the time variability skill scoring index in step S14, and use a comprehensive ranking index to rank the overall simulation performance of the climate models on the extreme precipitation index.

[0095] In step S15, the overall simulation performance of the climate model on the extreme precipitation index is ranked using a comprehensive ranking index as follows:

[0096]

[0097] Among them, MR represents the comprehensive ranking index of the overall simulation performance of the climate model on the extreme precipitation index, m represents the total number of climate models, n represents the number of extreme precipitation indices, and rank i It indicates that the i-th extreme precipitation index is ranked according to the size of the evaluation method index S and the size of the time variability skill scoring index IVS.

[0098] In this embodiment, the climate models are ranked according to the size of the evaluation method index in step S13 and the size of the time variability skill scoring index in step S14. When the evaluation method index S is closer to 1, the better the ranking is, and when the time variability skill scoring index IVS is closer to 0, the higher the ranking is.

[0099] S16. According to the climate model ranking result of the comprehensive ranking index in step S15, a climate model suitable for analyzing the extreme precipitation index in the study area is selected.

[0100] like Figure 6 As shown, Figure 6 This is a schematic diagram of the change of the multi-model ensemble dataset's ability to simulate the extreme precipitation index climate state in the study area with the average number of model ensembles. The study area in this embodiment is the Yellow River Basin.

[0101] like Figure 7 As shown, Figure 7 Schematic diagram of the multi-year mean spatial pattern of the index describing precipitation in the study area under historical and future scenarios. Figure 7 From left to right, they are the multi-year spatial means of the same extreme precipitation index under the scenarios of HIS, SSP126, SSP245, and SSP585; Figure 7 From top to bottom, they respectively describe the multi-year spatial means of extreme precipitation indices RX1day, RX5day, PRCPTOT, R90P, and R95P under the same scenario. The study area in this embodiment is the Yellow River Basin.

[0102] like Figure 8 As shown, Figure 8 Schematic diagram of the interannual variation of the index describing precipitation in the historical and future study areas for each scenario. The study area in this embodiment is the Yellow River Basin.

[0103] S2. Extract the historical and future extreme precipitation index sequences of the climate model selected in step S1 that are suitable for the extreme precipitation index analysis of the study area, use the non-negative function of hydrometeorology and its improved form as alternative functions, use the extreme precipitation index sequence to estimate the parameters of the alternative function, and select the probability distribution model for the long-term distribution of the extreme precipitation index of the sub-region of the study area.

[0104] Specifically, step S2 specifically includes S21-S24:

[0105] S21. Extract the historical and future extreme precipitation index sequences of the climate model selected in step S1 that are suitable for extreme precipitation index analysis in the study area.

[0106] S22, using the maximum likelihood method to construct a maximum likelihood function according to the extreme precipitation index sequence extracted in step S21 to estimate the parameters of the alternative function using the non-negative function of hydrometeorology and its improved form as the alternative function, and obtain appropriate parameters of the alternative function.

[0107] In step S22, the maximum likelihood method is used to construct a maximum likelihood function based on the extreme precipitation index sequence extracted in step 821 to estimate the parameters of the non-negative function of hydrological meteorology and its improved formula as the candidate function. The calculation process is:

[0108]

[0109] Among them, L(θ) represents the likelihood function, f(x i θ i ) indicates that the unknown parameter is θ i The extreme precipitation index series x 1 , x 2 , x 3 , ..., x n The corresponding x i The probability distribution function of .

[0110] In this embodiment, seventeen non-negative functions widely used in hydrology and meteorology and their improved expressions are used as candidate functions, as shown in Table 1. In this embodiment, the extreme precipitation index sequence is used as sample data, the candidate function is used as the overall data, and the sample data is used to estimate the parameters of the overall data, so as to determine which distribution should be selected for the overall data to obtain the extreme precipitation design values ​​under different return periods in the study area. Table 1 is as follows:

[0111] Table 1 Information on candidate distribution functions considered in this study

[0112]

[0113]

[0114] S23. According to the parameters of the suitable candidate function obtained in step S22, the Kolmogorov-Smilov test method is used to determine whether each extreme precipitation index sequence conforms to the theoretical distribution of the candidate function, and the test statistics and graphical analysis method are used to judge the fitting quality of the candidate function, so as to obtain the candidate function with the best fitting effect and the extreme value distribution model.

[0115] like Fig. 9 As shown, Fig. 9 Schematic diagram of the matching of empirical data and theoretical curve. In this embodiment, test statistics and graphical analysis are used to judge the fitting quality of the candidate function. Fig. 9 It can be concluded that the smaller the statistical value is, the more concentrated the scatter plot of theoretical and empirical data is. ° The closer the extreme value distribution model is to the two sides of the line, the better the fitting effect. One of the ultimate purposes of establishing the extreme value distribution model in this embodiment is to predict the design value under a fixed return period of extreme precipitation events.

[0116] S24. Using the extreme value distribution model of the candidate function with the best fitting effect obtained in step S23 as the probability distribution model of the long-term distribution characteristics of the extreme precipitation index in the sub-region of the study area, and using the probability distribution model to predict the design values ​​of extreme precipitation events under different recurrence periods.

[0117] The calculation process of using the probability distribution model to predict the design value of extreme precipitation events under different return periods in step S24 is:

[0118]

[0119]

[0120] Where p represents the critical probability under a given return period T, X represents the value of the random variable, and x p represents the design value of extreme precipitation events under a given return period T, P(X>x p ) represents the probability that the value of the random variable is greater than the design value of the extreme precipitation event under a given return period T, μ T represents the sequence expectation, T represents the return period, It represents the inverse function of the candidate function F(x) with the best fitting effect.

[0121] like Fig.10 As shown, Fig.10It is a schematic diagram of the spatial pattern of the superior probability distribution of extreme precipitation index series under various scenarios in the history and future of the study area. The study area in this embodiment is the Yellow River Basin. Fig.10 From left to right, the superior probability distribution spatial patterns of the same extreme precipitation index under the his, SSP126, SSP245, and SSP585 scenarios; Fig.10 From top to bottom, the superior probability distribution spatial patterns of extreme precipitation indices PRCPTOT, SDII, R90P, SDII(90), R95P, and SDII(95) under the same scenario.

[0122] like Fig.11 As shown, Fig.11 Schematic diagram of the probability frequency percentage of extreme precipitation index series under various historical and future scenarios in the study area. Fig.11 The upper left corner of each pie chart shows the corresponding extreme precipitation index name.

[0123] like Fig.12 As shown, Fig.12 This is a schematic diagram of the design value of the extreme precipitation index 20-year return period under various scenarios in the history and future of the study area. The study area in this embodiment is the Yellow River Basin. Fig.12 From left to right, they are the 20-year return period design values ​​of the same extreme precipitation index under the his, SSP126, SSP245, and SSP585 scenarios; Fig.12 From top to bottom, they are the 20-year return period design values ​​of the extreme precipitation indices PRCPTOT, SDII, R90P, SDII(90), R95P, and SDII(95) under the same scenario.

[0124] In this embodiment, the design value under the 20-year return period is taken as an example to show the design extreme precipitation process under various scenarios in the future of the sub-region of the study area, namely the nine provinces in the Yellow River Basin. Based on the latest rainstorm formula for the nine provinces in the Yellow River Basin, combined with the Chicago rain type, the rain peak coefficient is selected as 0.4 and the rainfall duration is 2h. The actual precipitation process under the 20-year return period of the nine provinces in the Yellow River Basin is obtained as the original rain type, and the downscaling delta method is superimposed to obtain the change factors of the future scenarios of the nine provinces in the Yellow River Basin relative to the historical period, so as to obtain the design precipitation process under the 20-year return period precipitation intensity index, namely SDII, SDII (90), and SDII (95). This is of great significance for the medium- and long-term hydrological forecast, water resources planning and management, and regional flood and drought prevention and control in the Yellow River Basin.

[0125] S3. Based on the climate model selected in step S1 for the analysis of the extreme precipitation index in the study area and the probability model selected in step S2 for the long-term distribution of the extreme precipitation index in the sub-region of the study area, the rainstorm formula is used in combination with the Chicago rain pattern method to obtain the design rainfall process under a specific return period under various future scenarios in the sub-region of the study area.

[0126] Step S3 specifically includes S31-S36:

[0127] S31. Calculate the rainstorm intensity in the study area using the rainstorm formula based on precipitation and rainfall duration;

[0128] The calculation formula for calculating the rainstorm intensity in the study area using the rainstorm formula in step S31 is:

[0129]

[0130] Among them, i represents the intensity of rainstorm, q represents the precipitation, t represents the duration of rainfall, a represents the rain force parameter, c represents the rain force change parameter, LgT represents the logarithm of the return period T, T represents the return period, b represents the rainfall duration correction parameter, and n represents the rainstorm attenuation coefficient. Among them, the unit of rainstorm intensity is mm / min, the unit of precipitation is mm, the unit of rainfall duration is min, and the value is 5-180min, the unit of rain force parameter is mm, the unit of return period is a, and the value is from 1 to 100a, and the unit of rainfall duration correction parameter is min.

[0131] S32. Using the Chicago Rain Pattern Method, according to the rainstorm intensity in the study area in step S31, calculate the rainfall intensity before and after the peak of a certain minute, obtain the rainfall intensity at any time, and connect them to obtain the instantaneous rainfall process line.

[0132] The calculation formula for calculating the rainfall intensity before and after the peak in a certain minute in step S32 is:

[0133]

[0134]

[0135] Among them, i b represents the rainfall intensity before the peak, r represents the rain peak coefficient, and the extreme precipitation in the Yellow River Basin is mostly single-peak rainfall, with a comprehensive rain peak coefficient of 0.4, t b represents the rainfall duration before the peak, i a represents the post-peak rainfall intensity, t a represents the duration of rainfall after the peak.

[0136] The rainstorm formula in this embodiment directly reflects the relationship between rainstorm intensity, rainfall duration and recurrence period. Therefore, by calculating the rainstorm intensity in the study area and determining its parameters, the Chicago rain pattern method is used to calculate the rainfall intensity before and after the peak of a certain minute according to the rainstorm intensity in the study area, and the rainfall intensity at any time is obtained, and they are connected to obtain the instantaneous rainfall process line.

[0137] like Fig.13 As shown, Fig.13 A schematic diagram of the original rain pattern of heavy rain designed based on historical measurements for each province.

[0138] S33. According to the design values ​​of extreme precipitation events predicted in step S23 at different recurrence periods, GIS technology is used to extract the design values ​​of extreme precipitation indices SDII, SDII(90), and SDII(95) to the design values ​​of sub-areas in the study area.

[0139] S34. According to the rainstorm intensity formula of the sub-area of ​​the study area, input the design return period and obtain the surface rainfall design value.

[0140] like Fig.14 As shown, Fig.14 Design a schematic diagram of rainstorm processes for future extreme rainfall in each province.

[0141] S35. Taking the instantaneous rainfall process line obtained by the Chicago rain type method in step S32 as the prototype, the ratio of the surface rainfall design value obtained in step S34 to the total amount of heavy rainfall in different return periods obtained by the heavy rainfall formula of each province is used as the amplification ratio, and the principle of amplification with the same ratio in different time periods is adopted to calculate the design rainfall process under a specific return period of extreme precipitation index SDII, SDII(90), and SDII(95) in a certain area in the historical period.

[0142] S36. Based on the change factor of the climate model obtained in step S11, multiply the change factor by the design rainfall process under a specific return period of the extreme precipitation index SDII, SDII(90), and SDII(95) in a certain region in the historical period in step S35 to obtain the design rainfall process under a specific return period under various scenarios in the future for a certain region.

[0143] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0144] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A prediction method for extreme rainfall based on future climate patterns, It is characterized in that The following steps are involved: S1. Delta downscaling method is used to correct the deviation of historical rainfall data and future rainfall data of climate model to obtain the change factor of climate model, and the error of extreme precipitation index between historical observation and downscaling simulation is verified by drawing Taylor diagram, and the climate model suitable for extreme precipitation index analysis in the study area is selected; S2, extracting the historical and future extreme precipitation index sequences of the climate model selected in step S1 that is suitable for the extreme precipitation index analysis of the study area, using the non-negative function of hydrometeorology and its improved form as the candidate function, using the extreme precipitation index sequence to estimate the parameters of the candidate function, and selecting a probability distribution model for the long-term distribution of the extreme precipitation index of the sub-region of the study area, so as to predict the design values ​​of extreme precipitation events under different return periods; S3. Based on the climate model selected in step S1 for the extreme precipitation index analysis of the study area and the probability model selected in step S2 for the long-term distribution of the extreme precipitation index of the sub-region of the study area, the rainstorm formula is used in combination with the Chicago rain pattern method to obtain the design rainfall process under a specific return period under various scenarios in the future sub-region of the study area, specifically: S31. Calculate the rainstorm intensity in the study area using the rainstorm formula based on precipitation and rainfall duration; S32, using the Chicago rain pattern method to calculate the rainfall intensity before and after the peak of a certain minute according to the rainstorm intensity in the study area in step S31, and obtain the rainfall intensity at any time, and connect them to obtain an instantaneous rainfall process line; S33, according to the design values ​​of the extreme precipitation events predicted in step S2 at different return periods, using GIS technology to extract the design values ​​of the extreme precipitation index SDII, SDII (90), and SDII (95) to the design values ​​of the sub-areas of the study area; S34. According to the rainstorm intensity formula of the sub-area of ​​the study area, the design return period is input to obtain the surface rainfall design value; S35, taking the instantaneous rainfall process line obtained by the Chicago rain type method in step S32 as the prototype, taking the ratio of the surface rainfall design value obtained in step S34 to the total amount of heavy rainfall in different return periods obtained by the heavy rainfall formula of each province as the magnification ratio, adopting the principle of amplification with the same ratio in different time periods, and calculating the design rainfall process under a specific return period of extreme precipitation index SDII, SDII(90), SDII(95) in a certain region in the historical period; S36. According to the change factor of the climate model in step S1, multiply the change factor by the design rainfall process under a specific return period of the extreme precipitation index of a certain region in the historical period of SDII, SDII(90), and SDII(95) in step S35 to obtain the design rainfall process under a specific return period under various scenarios in the future for a certain region.

2. A method for predicting extreme rainfall based on future climate patterns according to claim 1, It is characterized in that Step S1 specifically includes: S11. Use the bilinear interpolation method to interpolate the historical precipitation data of the climate model and the ensemble results of future precipitation data to On the horizontal grid, the change factor of the climate model is calculated based on the precipitation data using the change ratio, that is: in, represents the factor of change in climate patterns, , Respectively represent the climate variable values ​​of the global climate model base period and projection period at a certain time scale, , They represent the average values ​​of climate variables in the global climate model base period and projection period at the specified time. S12, according to the change factor of the climate model calculated in step S11, the change factor of the climate model is superimposed on the precipitation observation data using the Delta downscaling method to obtain the estimated value of the regional scale climate variable, that is: in, Indicates the number of observations Projections of regional-scale climate variables, Indicates the number of observations The observed values ​​of the climate variables at a specified time scale, Indicates the number of observations; S13, plotting the simulation relationship between the climate model and the climate model data into a Taylor diagram, and plotting the Taylor diagram into a scatter plot based on polar coordinates, and using evaluation method indicators to evaluate the simulation results of the climate model and the climate model data; S14. Use the temporal variability skill scoring index to evaluate the ability of climate models and climate model data to simulate interannual variability; S15, ranking the climate models according to the size of the evaluation method index in step S13 and the size of the time variability skill scoring index in step S14, and using a comprehensive ranking index to rank the overall simulation performance of the climate models on the extreme precipitation index; S16. According to the climate model ranking result of the comprehensive ranking index in step S15, a climate model suitable for analyzing the extreme precipitation index in the study area is selected.

3. A method for predicting extreme rainfall based on future climate patterns according to claim 2, It is characterized in that The simulation results of evaluating the climate model and the climate model data using the evaluation method index in step S13 are: in, Indicators representing the evaluation methods of climate models and climate model data, represents the correlation coefficient between the climate model and the climate model data, represents the maximum correlation coefficient that can be achieved among all climate models. Represents the ratio between the climate model and the standard deviation of the climate model data.

4. The method for predicting extreme rainfall based on future climate patterns according to claim 2, It is characterized in that In step S14, the temporal variability skill scoring index is used to evaluate the simulation ability of the climate model and the climate model data for the interannual variability: in, The temporal variability skill score index indicates the ability of climate models and climate model data to simulate interannual variability. , represent the standard deviation of the climate model and the time series of the climate model data, respectively.

5. The method for predicting extreme rainfall based on future climate patterns according to claim 2, It is characterized in that In step S15, the overall simulation performance of the climate model on the extreme precipitation index is ranked using a comprehensive ranking index as follows: in, A comprehensive ranking index representing the overall simulation performance of climate models on the extreme precipitation index. represents the total number of climate modes, represents the number of extreme precipitation indices, Indicates The extreme precipitation index is based on the evaluation method indicators Size and time rate skill scoring indicators Ranked by size.

6. The method for predicting extreme rainfall based on future climate patterns according to claim 1, It is characterized in that Step S2 specifically includes: S21, extracting the historical and future extreme precipitation index sequences of the climate model selected in step S1 that are suitable for extreme precipitation index analysis in the study area; S22, using the maximum likelihood method to construct a maximum likelihood function based on the extreme precipitation index sequence extracted in step S21 to estimate the parameters of the candidate function using the non-negative function of hydrometeorology and its improved form as the candidate function, and obtain the parameters of the appropriate candidate function; S23, according to the parameters of the suitable candidate function obtained in step S22, the Kolmogorov-Smilov test method is used to determine whether each extreme precipitation index sequence conforms to the theoretical distribution of the candidate function, and the test statistics and graphical analysis method are used to judge the fitting quality of the candidate function, so as to obtain the candidate function with the best fitting effect and the extreme value distribution model; S24. Using the extreme value distribution model of the candidate function with the best fitting effect obtained in step S23 as the probability distribution model of the long-term distribution characteristics of the extreme precipitation index in the sub-region of the study area, and using the probability distribution model to predict the design values ​​of extreme precipitation events under different recurrence periods.

7. The method for predicting extreme rainfall based on future climate patterns according to claim 1, It is characterized in that The calculation formula for calculating the rainstorm intensity in the study area using the rainstorm formula in step S31 is: in, Indicates the intensity of rainstorm. Indicates the amount of precipitation, Indicates the duration of rainfall. represents the rain force parameter, represents the rainfall variation parameter, Recurrence period Take the logarithm, represents the return period, represents the rainfall duration correction parameter, Represents the rainstorm attenuation coefficient.

8. A method for predicting extreme rainfall based on future climate patterns according to claim 7, It is characterized in that The calculation formula for calculating the rainfall intensity before and after the peak in a certain minute in step S32 is: in, represents the pre-peak rainfall intensity, represents the rain peak coefficient, represents the rainfall duration before the peak, Indicates the post-peak rainfall intensity, represents the duration of rainfall after the peak.

Images