Risk evaluation system based on rainstorm flood

Abstract

The invention relates to the technical field of meteorological disaster assessment, in particular to a rainstorm flood-based risk assessment system, which comprises an acquisition terminal, a law analysis terminal, a prediction terminal and an assessment terminal which are in progressive signal connection, the prediction terminal is used for predicting future extreme rainfall and flood influence through climate change conditions and in combination with historical rainstorm flood evolution rules analyzed by the rule analysis terminal, the prediction terminal comprises a rainfall prediction module and a flood influence module, and the evaluation terminal is used for comprehensive prediction through the rainfall prediction module and the flood influence module. And setting an evaluation framework for the flood risk, and estimating a future flood risk change trend. According to the method, multiple factors are comprehensively considered, the prediction precision is improved, more comprehensive and accurate support is provided for prediction, evaluation and response of rainstorm flood disasters, and therefore comprehensive and multi-level evaluation of disaster loss is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological disaster assessment, and particularly relates to a risk assessment system based on rainstorm floods. Background Art

[0002] Under the background of global climate change, extreme weather events occur frequently. Among them, rainstorm flood disasters, with their characteristics of strong suddenness, wide influence range, and great destructive power, have become important factors threatening the safety of human life and property. Rainstorm floods not only cause river floods and urban waterlogging, but also have a serious impact on multiple fields such as transportation, agriculture, and ecological environment. Therefore, it is particularly important to conduct risk assessment on rainstorm floods. Through scientific risk assessment, the occurrence probability and impact degree of rainstorm floods can be accurately predicted, providing decision-making support for relevant departments, thereby effectively reducing disaster losses and ensuring the safety of people's lives and property.

[0003] Although there are currently various rainstorm flood risk assessment methods and systems, there are still some obvious defects in actual applications. First of all, many existing assessment systems mainly rely on historical data for analysis, lacking in-depth exploration of the evolution laws of extreme rainfall and floods, resulting in inaccurate prediction results. Secondly, when processing data, some systems fail to fully consider the impact of climate change on rainstorm flood disasters, leading to deviations between the assessment results and the actual situation. In addition, when constructing the index system, some assessment systems often only focus on a single disaster-causing factor, ignoring the comprehensive effects of various factors such as the sensitivity of the disaster-forming environment and the vulnerability of the disaster-bearing body, making the assessment results incomplete. In view of these defects, a risk assessment system based on rainstorm floods is proposed. Summary of the Invention

[0004] To solve the above problems, the present invention provides a risk assessment system based on rainstorm floods, which provides more comprehensive and accurate support for the prediction, assessment, and response of rainstorm flood disasters by comprehensively considering various factors and improving the prediction accuracy, thereby effectively reducing all-round and multi-level assessments of disaster losses.

[0005] To achieve the above object, the technical solution of the present invention is as follows: A risk assessment system based on rainstorm floods, including a collection terminal, a law analysis terminal, a prediction terminal, and an assessment terminal that are connected by progressive signals;

[0006] The collection terminal is used to collect historical literature, geographic information data, meteorological and hydrological data, and CMIP6 model data;

[0007] The law analysis terminal is used to analyze the historical evolution law of rainstorm floods in a changing environment under the assumption of stationarity. The law analysis terminal includes a rainfall analysis module for analyzing the spatio-temporal variation characteristics of extreme rainfall and a flood analysis module for analyzing the variation of extreme floods;

[0008] The prediction terminal is used to estimate the impact of future extreme rainfall and floods by combining the climate change situation with the historical heavy rainfall and flood evolution laws analyzed by the terminal. The prediction terminal includes a rainfall prediction module and a flood impact module. The rainfall prediction module is used to estimate the adaptability evaluation and future extreme rainfall based on the CMIP6 model through the change characteristics analyzed by the rainfall analysis module. The flood impact module is used to perform hydrological simulation on the reservoir regulation through the flood analysis module and the rainfall prediction module and analyze the future flood characteristic changes;

[0009] The evaluation terminal is used to set an evaluation framework for flood risk through the comprehensive prediction of the rainfall prediction module and the flood impact module, and estimate the future flood risk change trend.

[0010] Furthermore, the law analysis terminal is built-in with a detrended pre-whitening MK test method for the change trend of extreme rainfall and flood extreme value sequences in river basins;

[0011] The flood analysis module is built-in with the Pettitt method for detecting mutation points in the maximum flood peak sequence of the river basin meteorological and hydrological basin collected by the collection terminal and the moving t-test method for testing each point of the maximum flood peak sequence in the river basin.

[0012] Furthermore, the law analysis terminal also includes a non-consistency calculation module. When the stationarity assumption is violated and the overall distribution of flood extremes in any year is different and the original method cannot be directly used to calculate the design flood value, the non-consistency calculation module is used to derive the hydrological design value equivalent to the T-year recurrence interval standard according to the "equal reliability".

[0013] Furthermore, the indicators of the extreme rainfall change characteristics analyzed by the rainfall analysis module include absolute indicators, threshold indicators, and percentile indicators.

[0014] Furthermore, the rainfall prediction module includes a CMIP6 evaluation module and a future rainfall prediction module;

[0015] The CMIP6 evaluation module is used to perform relative error evaluation, spatial structure evaluation, interannual variability evaluation, and comprehensive evaluation on the adaptability according to the extreme rainfall spatio-temporal change characteristics analyzed by the rainfall analysis module in combination with the CMIP6 model, and transmit the multi-climate model ensemble calculations to the future rainfall prediction module and the evaluation terminal respectively. Among them, three multi-model ensemble schemes are selected for the multi-model ensemble, including the arithmetic mean ensemble scheme, the improved reliability ensemble scheme considering the simulation performance of climate models, and the random forest ensemble scheme, and the simulation performance of the three multi-model ensemble schemes for the extreme rainfall indicators in the Huaihe River Basin is comprehensively compared;

[0016] The future rainfall prediction module is used to calculate the future rainfall spatio-temporal change characteristics according to the shared socioeconomic pathways and global warming scenarios collected by the collection terminal.

[0017] Further, the rainfall prediction module combines the spatiotemporal variation characteristics of extreme rainfall analyzed by the rainfall analysis module with the adaptability evaluation of the CMIP6 evaluation module, and transmits them to the flood impact module and the evaluation terminal respectively.

[0018] Further, the flood impact module combines the flood analysis module and the rainfall prediction module to calculate the impact of climate change on future floods in river basins and transmits it to the evaluation terminal. The flood impact module includes a hydrological simulation module and a future flood prediction module;

[0019] The hydrological simulation module is used to perform hydrological simulations through the comprehensive reservoir regulation obtained by the acquisition module and evaluate through the daily runoff simulation results;

[0020] The future flood prediction module is used to calculate the changes in future flood characteristics based on the shared socioeconomic pathways and global warming scenarios collected by the acquisition terminal.

[0021] Further, the evaluation terminal includes a flood risk assessment module and a flood risk change module;

[0022] The flood risk assessment module is used to construct an index system and calculate weights for the change characteristics of future extreme rainfall through the rainfall prediction module;

[0023] The flood risk change module is used to obtain the future risk change trend of the river basin based on the evaluation results of the flood risk assessment module and the evaluation results of the flood impact module.

[0024] Further, the flood risk assessment module, based on the flood disaster system theory and risk assessment principles, screens indicators according to the three major driving factors of hazard factor hazard, disaster-bearing environment sensitivity, and disaster-affected body vulnerability, and constructs an index system to evaluate the flood risk.

[0025] Further, the flood risk change module, based on the flood risk assessment model of the river basin, calculates the flood risk index in the historical period of the Huaihe River Basin and the future flood risk index of the future basin under four SSPs scenarios.

[0026] The above scheme has the following beneficial effects:

[0027] 1. In this solution, the risk assessment system can comprehensively consider various climate models and data to accurately predict extreme rainstorm and flood events. By applying the latest CMIP6 climate model and multi-model ensemble scheme, the system can predict in advance the spatio-temporal evolution characteristics of future extreme rainfall in the Huaihe River Basin, thereby improving the accuracy and timeliness of rainstorm and flood warnings. This helps relevant departments take preventive measures in advance and reduce disaster losses. The system constructs a prediction framework for future flood changes in the basin and can quantitatively evaluate the impact of climate change on future flood changes in the Huaihe River Basin. This helps the water conservancy department optimize flood control strategies, rationally allocate water resources, ensure that water conservancy facilities such as reservoirs and flood storage areas can play the maximum utility under extreme weather conditions, and safeguard the safety of people's lives and property.

[0028] 2. In this solution, by comprehensively considering various factors such as the danger of disaster-causing factors, the sensitivity of disaster-bearing environments, and the vulnerability of disaster-affected bodies, the system can comprehensively evaluate the flood risk zoning in historical and future periods. This provides a scientific basis for the government and relevant departments to formulate disaster prevention and mitigation plans, emergency response plans, and post-disaster reconstruction plans, and helps reduce the impact of flood disasters on social economy.

[0029] 3. In this solution, the system not only focuses on the physical impact of flood disasters but also considers the adaptability and resilience of the social and economic systems. By revealing the evolution law of flood risk zoning over a large time span, the system can help all sectors of society better understand the risks and challenges of flood disasters, thereby enhancing the overall adaptability and resilience of society. Against the background of global warming, the frequent occurrence of extreme rainstorm and flood events poses a serious threat to the ecological environment and the sustainable development of social economy. The risk assessment system helps guide all sectors of society to take more proactive and effective measures to address the challenges brought by climate change and promotes the sustainable development process of the Huaihe River Basin and even the whole country by providing scientific prediction and assessment results.

[0030] The additional aspects and advantages of the present invention will be partly given in the following description, partly become apparent from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the system framework of the embodiment of the risk assessment system based on rainstorm and flood of the present invention;

[0032] Figure 2 It is a schematic diagram of the Huaihe River Basin zoning and Thiessen polygons of the embodiment of the risk assessment system based on rainstorm and flood of the present invention;

[0033] Figure 3 It is a schematic diagram of the interannual change trend of the extreme rainfall Rx1day index and the TFPW-MK test results in the Huaihe River Basin from 1961 to 2020 of the embodiment of the risk assessment system based on rainstorm and flood of the present invention;

[0034] Figure 4 Schematic diagram of the risk assessment system for the inter-annual variation trend of the extreme rainfall Rx3day index and the TFPW-MK test results in the Huaihe River Basin from 1961 to 2020 based on rainstorm floods according to an embodiment of the present invention;

[0035] Figure 5 Schematic diagram of the inter-annual variation trend of the extreme rainfall R20mm index and the TFPW-MK test results in the Huaihe River Basin from 1961 to 2020 based on rainstorm floods according to an embodiment of the risk assessment system of the present invention;

[0036] Figure 6 Schematic diagram of the inter-annual variation trend of the extreme rainfall R50mm index and the TFPW-MK test results in the Huaihe River Basin from 1961 to 2020 based on rainstorm floods according to an embodiment of the risk assessment system of the present invention;

[0037] Figure 7 Schematic diagram of the inter-annual variation trend of the extreme rainfall R90p index and the TFPW-MK test results in the Huaihe River Basin from 1961 to 2020 based on rainstorm floods according to an embodiment of the risk assessment system of the present invention;

[0038] Figure 8 Schematic diagram of the inter-annual variation trend of the extreme rainfall R95p index and the TFPW-MK test results in the Huaihe River Basin from 1961 to 2020 based on rainstorm floods according to an embodiment of the risk assessment system of the present invention;

[0039] Figure 9 Schematic diagram of the inter-annual variation trend of the annual maximum flood peak flow series and the TFPW-MK test results at the main and tributary hydrological stations in the Huaihe River Basin based on rainstorm floods according to an embodiment of the risk assessment system of the present invention;

[0040] Figure 10 Schematic diagram of the Taylor diagram according to an embodiment of the risk assessment system of the present invention based on rainstorm floods;

[0041] Figure 11 Schematic diagram of the random forest structure using the Bootstrap method for random sampling according to an embodiment of the risk assessment system of the present invention based on rainstorm floods. Detailed implementation manners

[0042] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0044] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "installed", "connected", and "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0045] The following is a further detailed description through specific embodiments:

[0046] Embodiment 1:

[0047] As shown in the Figure 1 accompanying drawings: A risk assessment system based on rainstorm floods includes a collection terminal, a pattern analysis terminal, a prediction terminal, and an evaluation terminal that are progressively signal-connected.

[0048] The collection terminal is used to collect historical literature, geographical information data, meteorological and hydrological data, and CMIP6 model data. In this embodiment and the following embodiments, the Huaihe River Basin collected by the collection terminal is taken as the implementation object. Since the Huaihe River Basin spans five provinces of Henan, Hubei, Anhui, Jiangsu, and Shandong, with a catchment area of about 270,000 km 2 , and the main stream of the Huaihe River originates from the northern foot of Tongbai Mountain in Henan Province, flows through the southern part of Henan Province and the central part of Anhui Province, and injects into Hongze Lake in the central part of Jiangsu Province. After being regulated by Hongze Lake, the main stream flows through the Jiangshuidao to Sanjiangying in Yangzhou and injects into the Yangtze River. The total length of the main stream of the Huaihe River is about 1000 km, and the total elevation drop is about 200 m. The elevation drop of the upper reaches is much greater than that of the middle and lower reaches. Therefore, once rainstorm floods occur, the flood discharge in the middle reaches is slow, which is extremely likely to cause serious flood disasters. Therefore, the upper and middle reaches of the Huaihe River Basin (above Wujia Ferry Station) are selected as the research object, including the daily precipitation and daily runoff data of the Huaihe River Basin from 1961 to 2020.

[0049] The law analysis terminal is used to analyze the historical evolution law of rainstorm floods in a changing environment under the assumption of stationarity. The law analysis terminal includes a rainfall analysis module for analyzing the spatio-temporal variation characteristics of extreme rainfall and a flood analysis module for analyzing extreme flood changes.

[0050] The law analysis terminal is built-in with a detrending pre-whitening MK test method for the changing trends of extreme rainfall and flood extreme value sequences in river basins.

[0051] The flood analysis module is built-in with the Pettitt method for detecting mutation points in the maximum flood peak sequence of the meteorological and hydrological basins of river basins collected by the acquisition terminal and the moving t-test method for testing each point of the maximum flood peak sequence of river basins.

[0052] The indicators of the extreme rainfall variation characteristics analyzed by the rainfall analysis module include absolute indicators, threshold indicators, and percentile indicators.

[0053] The law analysis terminal also includes a non-consistency calculation module. When the stationarity assumption is violated and the overall distribution of flood extreme values in any year is different and the original method cannot be directly used to calculate the design flood value, the non-consistency calculation module is used to derive the hydrological design value equivalent to the T-year recurrence interval standard according to the "equal reliability".

[0054] The prediction terminal is used to estimate the future impacts of extreme rainfall and floods by combining the climate change situation with the historical evolution law of rainstorm floods analyzed by the law analysis terminal. The prediction terminal includes a rainfall prediction module and a flood impact module. The rainfall prediction module is used to estimate the adaptability evaluation and future extreme rainfall based on the CMIP6 model through the variation characteristics analyzed by the rainfall analysis module. The flood impact module is used to perform hydrological simulation on the reservoir regulation through the flood analysis module and the rainfall prediction module and analyze the changes in future flood characteristics.

[0055] The rainfall prediction module includes a CMIP6 evaluation module and a future rainfall prediction module.

[0056] The CMIP6 evaluation module is used to conduct relative error assessment, spatial structure assessment, inter-annual variability assessment, and comprehensive assessment of adaptability according to the spatio-temporal variation characteristics of extreme rainfall analyzed by the rainfall analysis module in combination with the CMIP6 model, and transmit the results calculated by the multi-climate model ensemble to the future rainfall prediction module and the evaluation terminal respectively. Among them, three multi-model ensemble schemes are selected for the multi-model ensemble, including the arithmetic mean ensemble scheme, the improved reliability ensemble scheme considering the simulation performance of climate models, and the random forest ensemble scheme, and the simulation performances of the three multi-model ensemble schemes for extreme rainfall indicators in the Huaihe River Basin are comprehensively compared.

[0057] The future rainfall prediction module is used to calculate the spatio-temporal variation characteristics of future rainfall according to the shared socioeconomic pathways and global warming scenarios collected by the acquisition terminal.

[0058] The flood impact module combines the flood analysis module and the rainfall prediction module to calculate the impact of climate change on future floods in river basins and transmit it to the evaluation terminal. The flood impact module includes a hydrological simulation module and a future flood prediction module;

[0059] The hydrological simulation module is used to perform hydrological simulation through the comprehensive reservoir regulation obtained by the acquisition module and evaluate it through the daily runoff simulation results;

[0060] The future flood prediction module is used to calculate the future flood characteristic changes according to the shared socioeconomic pathways and global warming scenarios collected by the acquisition terminal.

[0061] The rainfall prediction module combines the spatio-temporal variation characteristics of extreme rainfall analyzed by the rainfall analysis module with the adaptability evaluation of the CMIP6 evaluation module and transmits them to the flood impact module and the evaluation terminal respectively.

[0062] The evaluation terminal is used to set an evaluation framework for flood risk and estimate the future flood risk change trend through the comprehensive prediction of the rainfall prediction module and the flood impact module.

[0063] The evaluation terminal includes a flood risk assessment module and a flood risk change module.

[0064] The flood risk assessment module is used to construct an index system and calculate weights for the change characteristics of future extreme rainfall through the rainfall prediction module. Based on the flood disaster system theory and risk assessment principles, the flood risk assessment module screens indicators according to three major driving factors: the hazard of the disaster-causing factor, the sensitivity of the disaster-bearing environment, and the vulnerability of the disaster-affecting body, and forms an index system to evaluate the flood risk.

[0065] The flood risk change module is used to obtain the future river basin risk change trend according to the evaluation results of the flood risk assessment module and the evaluation results of the flood impact module. Based on the river basin flood risk assessment model, the flood risk change module calculates the flood risk index in the historical period of the Huaihe River Basin and the future flood risk index trend under four future SSP scenarios.

[0066] Example 2:

[0067] The difference from Example 1 is that specifically, in the trend analysis terminal, the method steps of the detrended pre-whitening MK test method are as follows:

[0068] S1. First, divide the original sequence X t by the mean value of the sequence respectively to obtain a new sequence X t , and then calculate the Sen's slope β of the new sequence X t .

[0069]

[0070] S2. Assume the sequence X t is linearly varying. Use the following formula (2-4) to remove the trend term of the sequence, forming the detrended residual term Y t :

[0071] Y t = X t —β.t (2-4)

[0072] S3. Calculate the first-order autocorrelation coefficient r of the sequence Y t . If the value of r is small, it is considered that the sequence Y t is independent, and the sequence X t can be directly used for the MK test. Otherwise, it is considered that the sequence has autocorrelation, and pre-whitening processing must be performed to remove the autocorrelation terms in the sequence to obtain the white noise term N t , and then recombine the detrended term Y t and the white noise term N t to form the recombined sequence M t :

[0073] N t = Y t -r.Y t -1

[0074] M t = N t +β.t

[0075] S4. Use the MK trend test method to perform a detrended significance test on the recombined sequence.

[0076] Specifically, the Pettitt method in the flood analysis module uses this test method to detect mutation points in the annual maximum flood peak sequences of 8 hydrological stations in the Huaihe River Basin from 1961 to 2020. The test statistic U t,n is defined as follows:

[0077]

[0078] In the formula, x i , x j are the variables in the sequence to be tested; n is the sequence length; U t,n is a new sequence composed of the number of times the first sample sequence exceeds the second sample sequence. If the τ moment satisfies:

[0079] K τ = |U τ，n | = max|U t，n |

[0080] Then the τ point is the mutation point. At the same time, the statistic can be calculated:

[0081]

[0082] If p ≤ 0.05, the mutation point τ is statistically significant and is identified as the mutation point of the sequence.

[0083] Specifically, the moving t-test method first selects two subsequence samples according to the positions of the moving points in the data time series, with data lengths of n1 and n2 respectively. Calculate the means and variances of the two subsequence samples, and construct a test statistic to test whether there is a difference in the means of the two subsequence samples, so as to determine whether there is a mutation. The calculation formula of the statistic is as follows:

[0084]

[0085]

[0086] In the formula, Z is the test statistic, s is the sequence variance. Here, Z follows the distribution of t(n1 + n2 - 2), and α is the significance level. If Z > t α / 2 It indicates that there is a significant difference between the two samples and there is a mutation. Otherwise, there is no significant difference and no mutation.

[0087] Specifically, under the consistency condition, the hydrological reliability R T,S of the design value (X S ) with a return period of T years is calculated as follows:

[0088]

[0089] Under the non-consistency condition, the non-consistency extrapolation module can still derive the hydrological design value (X T,NS ) equivalent to the standard of once in T years according to the concept of "equal reliability". Let R NS represent the hydrological reliability of the design value under the non-consistency condition, then

[0090]

[0091] F t (X T，NS ) = P(X t > X T，NS )

[0092] In the formula, F t (X) represents the probability distribution function of the hydrological extreme value in the t-th year, where t = 1, 2,..., L. Obviously, due to non-consistency, F t (X) is different from year to year. Let R NS = R S , then X T,NS can be obtained, thus solving the problem of estimating the hydrological design value under the non-consistency condition.

[0093] Example 3:

[0094] As shown in the attached Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8 , different from Example 2, specifically, the rainfall analysis module assigns a certain area weight coefficient to each rainfall station by using the Thiessen polygon method based on the daily rainfall data of 71 rainfall stations in the Huaihe River Basin from 1961 to 2020, and derives the area-averaged extreme rainfall index of the Huaihe River Basin from the extreme rainfall indices of the stations. Considering that there is a strong spatial variability in the changing trend of extreme rainfall, in this example, the Huaihe River Basin is further divided into 7 sub-regions to comprehensively explore the interannual variation characteristics of the extreme rainfall index in the Huaihe River Basin in the past 60 years, including the north bank above Wangjiaba, the south bank above Wangjiaba, above Zhoukou, the north bank between Wangjiaba and Luxi, the south bank between Wangjiaba and Luxi, the north bank between Luxi and Wusong, and the south bank between Luxi and Wusong, as shown in Figure 2 .

[0095] Among them, taking the interannual variation of the extreme rainfall Rx1day index in the Huaihe River Basin and each sub-region from 1961 to 2020 as an example of the absolute index of the extreme rainfall variation characteristics analyzed by the rainfall analysis module, as shown in Figure 3 , the multi-year average value of the Rx1day index in the Huaihe River Basin is 97.2 mm, and the climate tendency rate is 0.98 mm / 10a, showing an upward trend. From the perspective of each sub-region, the Rx1day index of the south bank above Wangjiaba is the highest, with a multi-year average value of 114.04 mm, and the Rx1day index above Zhoukou is the lowest, with a multi-year average value of 87.07 mm. In addition, except for the north bank above Wangjiaba and above Zhoukou, the Rx1day indices of the other sub-regions show an upward trend. Among them, the climate tendency rate of Rx1day in the south bank between Wangjiaba and Luxi is the highest, at 4.78 mm / 10a, and the climate tendency rate of the Rx1day index above Zhoukou is the lowest, at -0.57 mm / 10a. The results of the TFPW-MK test are as shown in Figure 3 . The results show that the increasing trend of the Rx1day index in the south bank between Wangjiaba and Luxi is significant, passing the 0.05 significance level test, and the Rx1day index in the south bank between Luxi and Wusong shows a relatively significant upward trend, passing the 0.1 significance level test, while the changing trends of the area-averaged Rx1day index in the Huaihe River Basin and the other sub-regions are not significant.

[0096] Figure 4The figure shows the interannual variation trends of the Rx3day index in the Huaihe River Basin and its sub - regions in the past nearly 60 years. It can be seen from the figure that the Rx3day index in the Huaihe River Basin shows an upward trend, with a multi - year average value of 150.17 mm and a climate tendency rate of 0.1 mm / 10a. Among them, the Rx3day index of the south bank above Wangjiaba is the highest, with a multi - year average value of 154.61 mm, and the Rx3day index above Zhoukou is the lowest, with a multi - year average value of 117.18 mm. In addition, the Rx3day indices of the south and north banks above Wangjiaba and above Zhoukou all show a decreasing trend, while the Rx3day indices of the remaining sub - regions all show an increasing trend. According to the TFPW - MK test results ( Figure 4 ), the interannual variation trends of the Rx3day index in the Huaihe River Basin and its sub - regions all fail to pass the significance test.

[0097] In addition, the threshold indicators of the extreme rainfall change characteristics analyzed by the rainfall analysis module are as Figure 5 shown. It can be seen that the multi - year average value of the R20mm index in the Huaihe River Basin is 12.05 d, and the climate tendency rate is 0.12 d / 10a, showing an upward trend. The results of each sub - region show that the R20mm index of the south bank of Wang - Lu is the highest, with a multi - year average value of 17.44 d, while the R20mm index above Zhoukou is the smallest, with a multi - year average value of 9.89 d. In addition, except for the south bank above Wangjiaba, the R20mm indices of the remaining sub - regions all show an increasing trend. Among them, the climate tendency rate of the R20mm index of the north bank of Wang - Lu is the highest, which is 0.44 d / 10a. From the TFPW - MK test results ( Figure 5 ), only the upward trend of the R20mm index of the north bank of Wang - Lu is relatively significant and passes the 0.1 significance level test.

[0098] As Figure 6 shown, Figure 6 it reflects the interannual variation of the R50mm index in the Huaihe River Basin and its sub - regions. It can be seen from the figure that the R50mm index in the Huaihe River Basin generally shows an increasing trend, with a multi - year average value of 2.8 d and a climate tendency rate of 0.052 d / 10a. In addition, the R50mm index of the south bank above Wangjiaba is the highest, with a multi - year average value of 4.01 d and a climate tendency rate of 0.17 d / 10a, while the R50mm index above Zhoukou is the smallest, with a multi - year average value of 2.17 d and a climate tendency rate of - 0.007 d / 10a. As Figure 6 known, except above Zhoukou, the R50mm indices of the remaining sub - regions all show an increasing trend. Among them, the increasing trends of the R50mm indices of the south and north banks of Wang - Lu and the south bank of Lu - Wu are relatively significant and pass the 0.1 significance level test.

[0099] The percentile indicators of the extreme rainfall change characteristics analyzed by the rainfall analysis module are shown in the inter-annual change trend of the R90p index of extreme rainfall in the Huaihe River Basin and its sub-regions from 1961 to 2020 as Figure 7 shown. It can be seen from the figure that the R90p index in the Huaihe River Basin shows an upward trend, with its multi-year average value being 31.07 mm and the climate tendency rate being 0.43 mm / 10a. The results of each sub-region show that the R90p index of the south bank above Wangjiaba is the highest, with its multi-year average value being 33.02 mm, while the R90p index of the area above Zhoukou is the smallest, with its multi-year average value being 29.33 mm. From the perspective of the change trend, except for the south bank above Wangjiaba and the south bank of Lu-Wu, the R90p indices of the other sub-regions all show an upward trend, among which the climate tendency rate of the R90p index of the north bank of Wang-Lu is the highest, being 0.79 mm / 10a. The TFPW-MK test results show that there is a relatively significant increasing trend in the R90p index of the north bank of Wang-Lu in the Huaihe River Basin, passing the 0.1 significance level test.

[0100] As Figure 8 shown, Figure 8 shows the inter-annual change trend of the R95p index in the Huaihe River Basin and its sub-regions from 1961 to 2020. The results show that there is an increasing trend in the R95p index in the Huaihe River Basin, with its multi-year average value being 46.45 mm and the climate tendency rate being 0.48 mm / 10a. The R95p indices of each sub-region all show an increasing trend, among which the R95p index of the south bank above Wangjiaba is the highest, with its multi-year average value being 49.85 mm, while the R95p index of the area above Zhoukou is the smallest, with its multi-year average value being 43.5 mm. In addition, the climate tendency rate of the R95p index of the north bank of Wang-Lu is the highest, being 0.86 mm / 10a. The TFPW-MK test results are as Figure 8 shown, and it is found that the increasing trend of the R95p index of the north bank of Wang-Lu is relatively significant, passing the 0.1 significance level test.

[0101] Generally speaking, the extreme rainfall index in the Huaihe River Basin as a whole shows an insignificantly increasing trend. At the same time, the change trend directions of the extreme rainfall indices in each sub-region also basically show an increasing trend. However, there are differences in the magnitude and significance of the extreme rainfall changes among different sub-regions, which more intuitively reflects the spatial variability of the extreme rainfall index changes at the sub-region scale. Among them, the TFPW-MK test results show that there are relatively significant increasing trends in the R20mm, R50mm, R90p, and R95p indices of the north bank of Wang-Lu. There is a significant increasing trend in the Rx1day index of the south bank of Wang-Lu, a relatively significant increasing trend in the R50mm index, and a relatively significant increasing trend in the Rx1day index of the south bank of Lu-Wu.

[0102] Example 4:

[0103] As shown in the appendix Figure 9As shown, the difference from Example 3 is that, specifically, the flood analysis module selects flood extreme events using the annual maximum method, that is, it selects the annual maximum flood peak flow to form a sequence of flood extreme events. First, the mean, coefficient of variation (Cv), and extreme ratio are used to reflect the inter-annual variation characteristics of the annual maximum flood peak sequence in the Huaihe River Basin. Among them, Cv is the ratio of the standard deviation to the arithmetic mean of the annual maximum flood peak sequence. The larger this value is, the greater the fluctuation degree of the annual maximum flood peak sequence between years, indicating the stability degree of the long-term flood change. The extreme ratio is the ratio of the maximum and minimum flood peak flows during the observation period, which reflects the unevenness degree of the annual maximum flood peak sequence. The larger this value is, the more uneven the inter-annual variation of the annual maximum flood peak sequence is. Table 2-5 shows the inter-annual variation results of the annual maximum flood peak sequence in the Huaihe River Basin from 1961 to 2020. Generally speaking, during this time period, the Cv values and extreme ratios of the 8 hydrological stations are all relatively large, indicating that the annual maximum flood peaks of the 8 hydrological stations fluctuated greatly as a whole from 1961 to 2020.

[0104] Table 1 Inter-annual variation characteristics of the annual maximum flood peak sequence in the Huaihe River Basin from 1961 to 2020

[0105]

[0106] As Figure 9 shown, except for the Bantai Station and the Wujiadu Station, the annual maximum flood peak sequences of the other hydrological stations all show a downward trend. The TFPW-MK test results show that only the annual maximum flood peak sequence of the Zhoukou Station has a significant downward trend and passes the 0.05 significance level test, while the change trends of the annual maximum flood peak sequences of the other hydrological stations are not significant.

[0107] The flood analysis module uses the Pettitt test method and the moving t-test method to detect the jump mutations of the annual maximum flood peak sequence in the Huaihe River Basin. By synthesizing the results of the two mutation detection methods, the mutation points of the annual maximum flood peak sequence of each hydrological station are diagnosed. The results are shown in Table 2. It is found that both mutation detection methods detect that the annual maximum flood peak sequence of the Xixian Station mutated around 2011, the annual maximum flood peak flow sequence of the Bantai Station was detected to have a significant mutation in 2009, the annual maximum flood peak sequence of the Zhoukou Station was detected to have a mutation in 1986, the moving t-test method detected that the annual maximum flood peak sequence of the Jiangjiaji Station mutated in 2008, while no significant mutation points were detected in the annual maximum flood peak sequences of the Miaowan Station, Wangjiaba Station, Lutai Station, and Wujiadu Station.

[0108] Table 2-6 Diagnosis results of mutation points of the annual maximum flood peak flow sequence in the Huaihe River Basin

[0109]

[0110] Note: * and ** in the table respectively indicate passing the significance tests at confidence levels of 0.1 and 0.05.

[0111] Example 5:

[0112] As shown in the Figure 10 appendix, the difference from Example 4 is that specifically, taking the observed daily precipitation data in the Huaihe River Basin and the historical scenario data (1961 - 2014) of the CMIP6 global climate model and the daily rainfall data (2015 - 2100) under four SSPs scenarios as examples, the CMIP6 evaluation module comprehensively evaluated the simulation ability of a single climate model for extreme rainfall indicators in the Huaihe River Basin from three dimensions: relative error, spatial structure, and interannual variability, and obtained a climate model suitable for simulating extreme rainfall in the Huaihe River Basin. In addition, the simulation abilities of the arithmetic mean multi - model ensemble scheme, the improved REA multi - model ensemble scheme, and the random forest multi - model ensemble scheme for extreme rainfall indicators in the Huaihe River Basin were compared. And based on the optimal multi - model ensemble scheme, the spatio - temporal variation characteristics of future extreme rainfall in the Huaihe River Basin in the medium - long term (2031 - 2060 and 2071 - 2100) and under the conditions of global warming of 1.5℃ and 2.0℃ were predicted.

[0113] Due to the low spatial resolution of global climate models and the existence of model biases, downscaling or bias correction is required. In this chapter, first, the daily rainfall data of 22 CMIP6 global climate models in historical and future periods were uniformly downscaled to 71 stations in the Huaihe River Basin using the inverse distance weighted interpolation method, and then the daily bias correction method (Daily bias correction, DBC) was used to perform site - by - site bias correction on the daily rainfall data. The DBC method is a distribution - based bias correction method proposed by Chen et al. (2013) by combining two methods: Daily translation (DT) (Mpelasoka and Chiew, 2009) and Local intensity scaling (LOCI)S (Schmidli, 2006). The DBC method first corrects the precipitation occurrence probability of climate model data. By making the precipitation occurrence probability of model data in the reference period in each month consistent with the observed data, the phenomenon of generally overestimating the precipitation occurrence probability in climate models is eliminated. The specific steps are as follows:

[0114] S1, Calculate the monthly precipitation occurrence probability of historical observed precipitation data;

[0115] S2, Calculate the occurrence threshold of the precipitation data output by the climate model so that the probability of the model monthly precipitation data being higher than this threshold is equal to the occurrence probability of the corresponding observed precipitation data, and the model precipitation data lower than this threshold is set to zero;

[0116] S3. Use the monthly precipitation thresholds obtained in the historical period to correct the future model precipitation data, and obtain the corrected future precipitation data.

[0117] After the above precipitation frequency correction, the empirical cumulative distribution function of the climate model output data is corrected by the empirical cumulative distribution function of the historical observation data. In the specific calculation, the cumulative distribution function is represented by 100 quantiles from the 0th to the 100th, and the specific formula is as follows:

[0118] P cor，fut，d = P sim，fut，d ×(P obs，per / P sim，ref，per )

[0119] In the formula, P represents precipitation, cor represents the corrected data, d represents daily, ref and fut represent the historical period and the future period respectively, obs and sim represent the observation data and the model simulation data respectively, and per represents a certain quantile.

[0120] To evaluate the simulation ability of the climate model output on the spatial structure of extreme rainfall in the Huaihe River Basin, the extreme rainfall index sequence of the climate model output in the historical reference period is calculated. The standard deviation, correlation coefficient, root mean square error index, and the Taylor diagram image distribution and skill score TS are used to conduct a comparative study on the simulation performance of the climate model. The Taylor diagram can effectively synthesize the information of the three indicators of the correlation coefficient, standard deviation, and root mean square error, and intuitively reflect the simulation ability of each model in the form of an image. The specific steps are as follows:

[0121] a n and b n are the basin-average extreme rainfall sequence simulated by climate model A and the basin-average measured extreme rainfall sequence respectively. Calculate each index according to the following formula:

[0122] a n and b n The standard deviations of are expressed as:

[0123]

[0124] a n and b n The correlation coefficient of is:

[0125]

[0126] a n and b n The standardized root mean square error is:

[0127]

[0128] The standard deviation, correlation coefficient, and normalized root mean square error between the measured extreme rainfall (bn) and the simulated value (an) of climate model A satisfy −2σ a σ b the relationship of R. Presenting these three evaluation indicators on the same graph gives the Taylor diagram, as shown in the schematic diagram Figure 10 as follows. The REF point in the figure represents the measured data. The arc axis in the figure represents the correlation coefficient R between the model simulation value and the measured value. The rectangular coordinate system represents the standard deviation ratio (the standard deviation of the climate model simulation value / the standard deviation of the measured value, i.e., σ a / σ b ). The distance from the point to the REF point is the normalized root mean square error E relative to the measured value. The correlation coefficient represents the similarity between the simulation result and the measured value. The root mean square error and the standard deviation ratio respectively characterize the differences in the intensity and uniformity of the climate model simulation result compared to the measured value. When the correlation coefficient between the model shown in the Taylor diagram and the measured value is larger, the standard deviation ratio is closer to 1, and the root mean square error is smaller, indicating that the model value is closer to the measured value and the simulation performance of the climate model is better.

[0129] To more intuitively reflect the quality of the model simulation performance, a quantifiable index TS proposed by Talyor based on these two parameters, the correlation coefficient and the standard deviation ratio, is introduced to quantify the simulation performance of the climate model:

[0130]

[0131] In the formula, TS is the skill score, and the closer its value is to 1, the better the model simulation performance; is the standard deviation ratio between the model and the observation; R0 is the maximum obtainable correlation, usually set to 0.999.

[0132] Example 6:

[0133] As shown in the appendix Figure 11As shown, the difference from Example 5 is that specifically, in the multi-model ensemble method, when improving the reliability ensemble scheme for future climate projection, due to the different internal mechanisms and applicable scopes among models, there is usually a large uncertainty in the projection results based on a single climate model only. The multi-model ensemble (MME) method has been proven to effectively reduce the uncertainty of the physical processes and parameterization schemes of climate models in future projections to a certain extent, so it is widely used in future climate projection research. The most commonly used current scheme is the equal-weight multi-model ensemble average. However, it should be noted that climate models with poor simulation performance will reduce the overall simulation ability of the multi-model ensemble, and using a simple arithmetic average ensemble scheme can no longer meet the research needs. Therefore, based on the reliability ensemble average framework, this paper comprehensively considers the simulation performance of climate models in aspects such as simulating the observed climate state of extreme rainfall, spatial structure, and interannual variability, and further proposes an improved REA scheme (UREA). The reliability coefficient R i of this scheme is calculated as follows:

[0134] R i = [f1(P bias )] m1 × [f2(P Talyor )] m2 × [f3(P IVS )] m3

[0135]

[0136] In the formula: f1 represents the ability of the model to reproduce the observed climate state, f2 represents the ability of the model to simulate the observed spatial structure, and f3 measures the performance of the model in reproducing the observed interannual variability; P represents the extreme rainfall index; P bias is the deviation term of the simulated observed climate state, P Talyor is the spatial structure term of the simulated observation; P IVS is the relative deviation term of the simulated observed interannual variability. ε P is the difference between the maximum and minimum values of the observed values of a certain variable in the historical period, P mod and P obs are the extreme rainfall indices of simulation and observation respectively, abs represents the absolute value; r is the spatial correlation coefficient between observation and simulation, r0 is the maximum correlation that can be obtained, and its value is 0.999; is the ratio of the spatial standard deviation between observation and simulation; STD obs and STD mod are the standard deviations of the interannual variations of observation and simulation; the value ranges of f1, f2, and f3 are 0 to 1. In this example, the parameters m1, m2, and m3 are all set to 1 to equally treat the contributions of f1, f2, and f3.

[0137] Once the reliability coefficient R of each pattern is calculated i , the weight of each pattern can be determined, and the calculation formula is as follows:

[0138]

[0139] In the formula: N is the number of climate models, and W i is the weight of the i-th model, and W i is actually the normalized value of R i .

[0140] In addition, the Random Forest (RF) method is based on a tree structure and has advantages such as good anti-noise ability and being unaffected by multicollinearity, enabling it to perform well in both classification and regression problems. By integrating multiple decision trees (classifiers), a forest (ensemble classifier) with better prediction results is constructed. Each sampling is sampling with replacement, that is, randomly extracting n training samples from the training set as a new training set; each time, a new sub-model is constructed using the newly formed training set; the output results of the final model are divided into two types for classification and regression. For classification problems, the result is determined by the classification result of the sub-model with the most votes, while for regression problems, the average value of the output results of each sub-model is calculated as the final prediction value. The random sampling with replacement of the samples here is the manifestation of the first "random" property of the random forest.

[0141] The structure of the random forest is as Figure 11 shown, and the model construction mainly includes the following steps:

[0142] The first step: Use the Bootstrap method to construct a new training set. Sampling with replacement is performed on the training samples to construct m new training samples;

[0143] The second step: Train the m training samples respectively to obtain m decision tree models.

[0144] Obviously, the above embodiments are only examples clearly described and not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or variations can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. And the obvious changes or variations derived therefrom are still within the protection scope of the present invention.

Claims

1. A risk assessment system based on rainstorm and flood, characterized in that: It includes a progressive signal connection acquisition terminal, a law analysis terminal, a prediction terminal and an evaluation terminal; The collection terminal is used to collect historical documents, geographic information data, meteorological and hydrological data, and CMIP6 model data; The law analysis terminal is used to analyze the evolution law of historical rainstorm floods in a changing environment under the assumption of stability. The law analysis terminal includes a rainfall analysis module for analyzing the spatiotemporal variation characteristics of extreme rainfall and a flood analysis module for analyzing the variation of extreme floods. The prediction terminal is used to predict the impact of future extreme rainfall and floods through climate change conditions and the evolution law of historical rainstorms and floods analyzed by the law analysis terminal. The prediction terminal includes a rainfall prediction module and a flood impact module. The rainfall prediction module is used to evaluate adaptability and predict future extreme rainfall based on the CMIP6 model through the change characteristics analyzed by the rainfall analysis module. The flood impact module is used to perform hydrological simulation of reservoir regulation and analyze future changes in flood characteristics through the flood analysis module and the rainfall prediction module. The assessment terminal is used to set up an assessment framework for flood risk and predict future trends in flood risk through comprehensive predictions of the rainfall prediction module and the flood impact module.

2. The risk assessment system based on rainstorm and flood according to claim 1 is characterized in that: The regularity analysis terminal has a built-in preset White MK test method for detrending the trend of extreme rainfall and flood extreme value series in river basins; The flood analysis module has built-in Pettitt method for detecting mutation points of the maximum flood peak sequence of the river basin meteorological and hydrological basin collected by the collection terminal and sliding t test method for testing the maximum flood peak sequence of the river basin point by point.

3. The risk assessment system based on rainstorm and flood according to claim 2 is characterized in that: The regularity analysis terminal also includes an inconsistency calculation module, which is used to deduce the hydrological design value equivalent to the T-year return standard based on "equal reliability" when the stability assumption is violated and the overall distribution of flood extremes in any year is different and the design flood value cannot be directly calculated using the original method.

4. The risk assessment system based on rainstorm and flood according to claim 3 is characterized in that: The indicators of extreme rainfall change characteristics analyzed by the rainfall analysis module include absolute indicators, threshold indicators and percentile indicators.

5. The risk assessment system based on rainstorm and flood according to claim 4 is characterized in that: The rainfall prediction module includes the CMIP6 evaluation module and the future rainfall prediction module; The CMIP6 evaluation module is used to conduct relative error evaluation, spatial structure evaluation, interannual variability evaluation and comprehensive evaluation of adaptability based on the spatiotemporal variation characteristics of extreme rainfall analyzed by the rainfall analysis module and the CMIP6 model, and transmit the multi-climate model ensemble to the future rainfall prediction module and the evaluation terminal respectively after calculation. Among them, the multi-model ensemble selects three multi-model ensemble schemes, including the arithmetic mean ensemble scheme, the improved reliability ensemble scheme considering the simulation performance of the climate model, and the random forest ensemble scheme, and comprehensively compares the simulation performance of the three multi-model ensemble schemes on the extreme rainfall indicators in the Huaihe River Basin; The future rainfall prediction module is used to calculate the spatiotemporal variation characteristics of future rainfall based on the shared socioeconomic pathways and global warming scenarios collected by the collection terminals.

6. The risk assessment system based on rainstorm and flood according to claim 5 is characterized in that: The rainfall prediction module combines the spatiotemporal variation characteristics of extreme rainfall analyzed by the rainfall analysis module with the adaptability evaluation of the CMIP6 evaluation module, and transmits them to the flood impact module and the evaluation terminal respectively.

7. The risk assessment system based on rainstorm and flood according to claim 6 is characterized in that: The flood impact module combines the flood analysis module and the rainfall prediction module to calculate the impact of climate change on future river basin floods and transmits it to the assessment terminal. The flood impact module includes the hydrological simulation module and the future flood prediction module; The hydrological simulation module is used to simulate the hydrological data of the comprehensive reservoir transfer obtained by the acquisition module and evaluate the results through daily runoff simulation; The future flood prediction module is used to calculate future changes in flood characteristics based on the shared socioeconomic pathways and global warming scenarios collected by the collection terminals.

8. The risk assessment system based on rainstorm and flood according to claim 7 is characterized in that: The assessment terminal includes a flood risk assessment module and a flood risk change module; The flood risk assessment module is used to construct an indicator system and calculate weights for the changing characteristics of future extreme rainfall through the rainfall prediction module; The flood risk change module is used to derive the future risk change trend of the river basin based on the evaluation results of the flood risk assessment module and the evaluation results of the flood impact module.

9. The risk assessment system based on rainstorm and flood according to claim 8 is characterized in that: The flood risk assessment module is based on the flood disaster system theory and risk assessment principles. It selects indicators based on the three driving factors: the danger of disaster-causing factors, the sensitivity of the disaster-prone environment, and the vulnerability of the disaster-bearing body, and constructs an indicator system to evaluate flood risks.

10. The risk assessment system based on rainstorm and flood according to claim 9 is characterized in that: The flood risk change module is based on the river basin flood risk assessment model to calculate the flood risk index of the Huaihe River Basin in the historical period and the future basin flood risk index under the four future SSPs scenarios.

Images