A method for ranking point rainstorm rainfall and a method for estimating the return period of point rainstorm

Abstract

The present invention discloses a method for sorting point rainstorms and a method for estimating the return period of point rainstorms: step 1, selecting rainstorm observation records with more than 20 consecutive years in the study area; step 2, grouping the effective equivalent rainfall samples; step 3, calculating the average effective equivalent rainfall of each group of samples; step 4, calculating the average return period of each sample group; step 5, constructing a regression model of the relationship between equivalent rainfall and return period; step 6, calculating the return period corresponding to the measured or designed extreme heavy rainstorm rainfall. The present invention eliminates the influence of the duration of rainfall on the size judgment of rainstorm rainfall by introducing equivalent rainfall, thereby making the sizes of various types of rainstorms comparable. On this basis, the rainstorm return period model constructed based on the return period law of historical rainstorms has a high credibility in the estimation result of the return period of extreme rainstorms.

Description

Technical Field

[0001] The present invention belongs to the field of water conservancy engineering and soil and water conservation engineering, and relates to water conservancy engineering construction, including reservoir storage capacity design, urban flood control engineering and determination of flood control standards for various soil and water conservation engineering construction, and in particular to a method for sorting point rainstorm rainfall and a method for estimating the rainstorm recurrence period and the closely related secondary disaster assessment caused by rainstorm. Background Art

[0002] Rainfall is the phenomenon that the hot and humid air in the atmosphere encounters a cold air current, causing the temperature to drop significantly, and then the water vapor is supersaturated. The excess water will condense and then fall to the ground in liquid or solid form. Rainfall is an important part of the water cycle on the earth's surface and a natural phenomenon in the climate process. For the earth's land surface, during the rainfall process, when the rainfall intensity exceeds the infiltration and water storage capacity of the underlying surface such as the surface soil in the corresponding period, surface runoff will be generated. Heavy rain is a rainfall with relatively large rainfall intensity and rainfall. It is the main cause of surface runoff and soil erosion, especially large-scale continuous heavy rain and concentrated heavy rain. Heavy rain not only produces large and rapid surface runoff, but also causes mountain torrents, reservoir dam collapse, river overflow, house collapse, farmland inundation, and then a series of serious natural disasters such as urban waterlogging, traffic and telecommunications interruption, which bring serious harm to the national economy and people's lives and property.

[0003] Rainfall is a form of water vapor cycle on the earth and a product of regional climate change. Under different climate and environmental conditions, the causes of rainfall are slightly different. It is usually divided into convective rainfall, orographic rainfall, frontal rainfall and typhoon rainfall, etc. The manifestations of rainfall from different causes are not the same. The performance characteristics of rainfall in different seasons and different time periods are also different, mainly manifested in: different rainfall intensities, different durations of rainfall, different final rainfall amounts, and different rainfall coverage. The rainfall amounts at different locations during the same rainfall process are also different, and sometimes the differences are very large. Therefore, the following five main parameters are usually used to describe the rainstorm process: rainfall intensity, rainfall duration, rainfall (usually refers to the point rainfall measured by the rain gauge, that is, the rainfall depth per unit area), rainfall area and rainfall center (location).

[0004] Among the five parameters describing rainfall characteristics, only rainfall amount and rainfall duration are measured values, while rainfall intensity is calculated values ​​(based on rainfall amount and rainfall duration), and rainfall area and rainfall center position are statistical analysis values. Since rainfall amount can relatively intuitively reflect the size of rainfall, it is usually used to classify and evaluate rainfall according to the size of rainfall in the same period. For a specific rainfall, the maximum rainfall (point rainfall) collected in the rainfall coverage area is generally used as the rainfall center, and then on this basis, a comprehensive analysis and evaluation is carried out in combination with other characteristic parameters of rainfall. In the past, due to the limitations of observation means and methods, the observation interval of meteorological data was mostly 12 hours or 24 hours. Therefore, in terms of rainfall classification, for consistency, my country's meteorological department stipulates that rainfall with a 24-hour rainfall of less than 10mm is light rain; rainfall with a 24-hour rainfall of between 10-25mm is moderate rain, rainfall with a 24-hour rainfall of between 25-50mm is heavy rain, and rainfall with a 24-hour rainfall of more than 50mm is torrential rain. Since rainstorms often cause serious natural disasters, and the range of more than 50 mm is too large, in order to facilitate accurate analysis and evaluation of the disasters that may be caused by rainstorms, rainstorms with different rainfall amounts are divided, among which: rainfall between 50-100 mm in 24 hours is rainstorm (ordinary rainstorm), rainfall between 100-250 mm in 24 hours is heavy rainstorm, and rainfall greater than 250 mm in 24 hours is extremely heavy rainstorm. Since rainstorms not only refer to large rainfall, but also include the attribute of large rainfall intensity, the meteorological department also includes some short-duration, heavy rainfall, such as heavy rainfall with more than 16 mm per hour and rainfall equal to or greater than 30 mm in 12 hours, into the scope of "rainstorms".

[0005] In addition, in terms of rainstorm classification, rainstorms are divided into short-duration rainstorms and long-duration rainstorms according to the duration of the rainstorms; according to the size of the rainstorms and the scope of their occurrence and impact, rainstorms are divided into local rainstorms, regional rainstorms, large-scale rainstorms, and extremely large-scale rainstorms. Short-duration rainstorms generally last only a few minutes to tens of minutes; while long-duration rainstorms may last from several hours to dozens of hours. In general, local rainstorms have a smaller impact range and a shorter duration, while regional rainstorms generally have a larger impact range and a longer duration.

[0006] The above rainstorm classification standard is to limit the duration of rainfall to 24 hours (or 12 hours). Since the duration of rainfall is fixed, the attribute of rainfall intensity is weakened. In fact, under the condition of the same rainfall, the disasters caused by short-duration heavy rainfall are usually more serious. Since the rainfall process is a natural phenomenon, the duration of each rainfall (rainfall duration) and rainfall amount are very random and uncertain. It is difficult to compare the size of rainstorms between the same rainfall duration and different rainfall amounts or different rainfall durations and the same rainfall amounts, and it is also inconvenient to evaluate the possible disasters. For example: if 20 mm of rainfall is evenly distributed within 24 hours, it will give people the feeling of drizzling rain, which can only be regarded as moderate rain, but if the rainfall in 1 minute reaches 20 mm (the current global land rainfall record for 1 minute is 38.1 mm), it must feel like the sky is falling and the earth is sinking, which cannot be described by heavy rain. The destructive power it produces is no less than a small rainstorm, but due to the small rainfall, it does not meet the rainstorm level standard. Similarly, the surface runoff and subsequent disasters generated by rainfall of 100 mm in 1 hour, 100 mm in 12 hours, and 100 mm in 24 hours are completely different, but there is a lack of relevant evaluation methods. In addition, if the rainfall is 120 mm in 24 hours and 150 mm in 48 hours, how big the difference in natural disasters is between the two cannot be assessed theoretically. Since it is impossible to directly compare the size of heavy rains with different rainfall durations, it is difficult to reliably estimate the frequency of heavy rain recurrence, and of course it is difficult to reasonably design the standards of various water conservancy and water conservation projects and road and urban flood control and drainage projects.

[0007] In terms of rainstorm recurrence period or frequency estimation, the estimation method currently used is mainly the fitting method. Its empirical frequency formula adopts the expected value formula, and the line type adopts the P-Ⅲ type. The specific operation method is to first select the typical rainstorm in the study area (referring to the point rainstorm rainfall with relatively large rainfall in the study area, and the point rainstorm refers to the rainfall measured by the meteorological station or rainfall observation station within a certain period of time that reaches or exceeds the rainstorm standard), and count the largest point rainstorm in the actual rainfall measured by the local meteorological station or rainfall observation station over the years (the result of this method is that the maximum sample number is equal to the number of observation years), and sort them by size, analyze and calculate the frequency of each rainstorm according to the number of observation years, and then plot them on the rainstorm frequency grid paper to draw the rainstorm frequency trend line. This method has a certain degree of credibility only when the number of samples is quite large (the size is related to the size of the rainstorm being assessed), but in most cases, due to the small number of samples, the recurrence period of the observed extreme rainstorm cannot be correctly assessed, so there will be a phenomenon of deviation from the trend line or a large deviation in the results. In order to include it in the trend line, only the unreasonable phenomenon (not on the trend line) caused by some extreme rainstorms is combined with the relationship between statistical parameters such as the mean value, deviation coefficient, and skewness coefficient of local rainfall, and then combined with the investigation data and experience of surrounding rainstorms to make appropriate corrections to the trend line. For example, in a certain place in southern my country, there was a 24-hour rainfall of 338mm. Combined with the typical rainstorm data in the early stage, the rainstorm frequency trend line was drawn on the rainstorm frequency grid paper. Through calculation, it was shown that the rainfall was a rainstorm that occurred once in ten thousand years, but after investigating the rainfall data in other nearby areas, it was found that the level of the rainstorm was not that high. Later, combined with the investigation data and experience of surrounding rainstorms, the trend line was appropriately corrected, and finally it was concluded that the rainfall was a rainstorm that occurred once in 200 years. Therefore, the near-line method (also known as the modified frequency curve) in the traditional method of determining the recurrence frequency of heavy rain is somewhat subjective, and it is unknown how credible the basis for the correction is. In addition, this method cannot compare the rainfall of heavy rains with different rainfall durations (especially heavy rains with a rainfall duration of more than 24 hours), thereby reducing the recurrence frequency of some heavy rains. The calculated recurrence period or recurrence frequency of heavy rains will produce a large deviation in the prevention and control of heavy rain disasters, resulting in various unnecessary losses. Summary of the invention

[0008] In view of the fact that there is currently a lack of methods for directly comparing and judging the rainfall amounts of various types of point rainstorms (short-duration rainstorms and long-duration rainstorms) in engineering hydrology at home and abroad, and of course it is impossible to accurately and reliably evaluate the recurrence period, frequency and disaster level of point rainstorms, the purpose of the present invention is to propose a universal point rainstorm size comparison method so that the sizes of point rainstorms with different rainfall durations are comparable; on this basis, a method for estimating the recurrence period or frequency of point rainstorms is proposed, thereby providing reliable technical support for rainstorm size classification, disaster evaluation and rainstorm frequency design.

[0009] In order to achieve the above objectives, the present invention adopts the following technical solutions:

[0010] On the one hand, the present invention discloses a method for sorting point rainstorm rainfall, which specifically comprises the following steps:

[0011] Step 1: Determine the rainstorm rainfall thresholds at different rainfall durations based on the historical rainfall characteristics in the study area;

[0012] Step 2, counting rainfall information in which the total rainfall of a continuous rainfall process in the historical rainfall data in the study area is greater than or equal to the rainstorm rainfall threshold corresponding to the rainfall duration, and the rainfall information includes rainfall amount and rainfall duration;

[0013] Step 3, dividing the rainfall of each rainstorm by the rainfall duration to obtain the average rainfall intensity of each rainstorm;

[0014] Step 4, deleting the rainstorm records with smaller rainfall intensity from the average rainfall intensity obtained in step 3 to obtain valid rainstorm data;

[0015] Step 5, calculating the equivalent rainfall of each rainstorm based on the effective rainstorm data obtained in step 4;

[0016] Step 6, determining the effective equivalent rainfall of the rainstorm according to the equivalent rainfall of each rainstorm obtained in step 5;

[0017] Step 7, arranging the effective equivalent rainfall corresponding to each rainstorm obtained in step 6 in descending order.

[0018] Furthermore, the step 1 specifically includes:

[0019] (1) If the rainfall duration is no more than 24 hours, the rainstorm rainfall thresholds for semi-arid areas, arid areas, and humid areas are 30 mm / 24 hours, 20 mm / 24 hours, and 50 mm / 24 hours, respectively;

[0020] (2) If the rainfall duration is greater than 24 hours, for every 24 hours increase in rainfall duration, the corresponding rainfall threshold is increased by 2 / 3 of the rainfall threshold for a heavy rain with a rainfall duration of no more than 24 hours.

[0021] Furthermore, in step 2, the rainfall information obtained by statistics should also satisfy the following requirement: the rainfall in each 24-hour period should not be less than 1 / 3 of the rainfall threshold of a rainstorm with a rainfall duration of no more than 24 hours.

[0022] Furthermore, in step 4, the average rainfall intensity corresponding to the rainstorm rainfall threshold of three consecutive days of rainfall is used as the minimum rainfall intensity of an effective rainstorm, and the rainstorm records with a rainfall intensity less than the minimum rainfall intensity are deleted from the result of the average rainfall intensity obtained in step 3 to obtain effective rainstorm data.

[0023] Furthermore, in step 5, the equivalent rainfall P of the rainstorm is calculated using formulas (1) and (2):

[0024] P=η×R (1)

[0025]

[0026] Where: P--equivalent rainfall; R--measured rainfall in mm; η--conversion characteristic parameter; --The average rainfall intensity of a rainstorm, α—rainfall intensity characteristic index; for semi-arid areas, humid areas and arid areas, the values ​​of α are 0.20, 0.15 and 0.25 respectively.

[0027] On the other hand, the present invention discloses a method for estimating a point rainstorm return period, which specifically comprises the following steps:

[0028] Step 1, select the rainstorm observation records with more than 20 consecutive years in the study area; use the above-mentioned point rainstorm rainfall sorting method to calculate the equivalent rainfall of each rainstorm, determine the effective equivalent rainfall of the rainstorm and sort it; if the number of effective rainstorm equivalent rainfall is greater than 40, execute step 2, otherwise return to step 1;

[0029] Step 2, grouping the effective equivalent rainfall samples obtained in step 1;

[0030] Step 3, calculating the average effective equivalent rainfall of each group of samples obtained in step 2;

[0031] Step 4, calculating the average recurrence period of each sample group based on the results of step 3;

[0032] Step 5, constructing a regression model of the relationship between equivalent rainfall and return period based on the average effective equivalent rainfall of each sample group obtained in step 3 and the average return period of each sample group obtained in step 4;

[0033] Step 6: Calculate the return period of the measured or designed extreme heavy rain rainfall according to the regression model of step 5.

[0034] Furthermore, the step 1 includes the following sub-steps:

[0035] Step 1.1: Check whether the rainstorm data in the study area has continuous observation records for more than 20 years. If so, and there are several rainfall observation stations, select the rainstorm data with a longer rainfall data sequence and execute step 1.2. Otherwise, select new rainstorm data in areas with similar climate and environment around the study area and execute step 1.1.

[0036] Step 1.2, according to the point rainstorm rainfall sorting method of the present invention, obtaining the effective rainstorm equivalent rainfall in descending order;

[0037] Step 1.3: The number of effective rainstorm equivalent rainfall obtained in step 1.2 must be greater than 40, that is, more than 40 rainfall data are required. If so, use them as effective equivalent rainfall samples and execute step 2; otherwise, select new rainstorm data in areas with similar climate and environment around the study area and execute step 1.1.

[0038] Further, the step 2 is performed as follows:

[0039] The effective equivalent rainfall samples obtained in step 1 are grouped in descending order. The grouping rule adopts arithmetic grouping or geometric grouping, that is, the number of samples in the latter group increases in an arithmetic or geometric relationship with the number of samples in the previous group, and the grouping meets the following conditions:

[0040] (1) Grouping starts from the first sample in the sample group;

[0041] (2) The samples in each group are continuous in size order;

[0042] (3) The number of samples n1 in the first group is 3 if and only if 3 samples are selected, and the deviation between the value of each sample in the first group and the average value of the samples in the group cannot exceed 5%. If it exceeds, the sample with the larger value in the sample group is discarded, and the sample group is supplemented in order of size until the requirement is met;

[0043] (4) The number of groups the sample is grouped into is greater than or equal to 4;

[0044] (5) The second group selects samples from the same sequence as the first group. If the second and subsequent groups are grouped by arithmetic difference method, the difference is 2 or 3; if the geometric ratio method is used, the common ratio is 1.5-2;

[0045] (6) The number of samples in the last group must be sufficient and valid.

[0046] Furthermore, the step 3 is calculated using the following formula:

[0047]

[0048] In the formula, ---The average effective equivalent rainfall of the mth group of samples; P r ---The effective equivalent rainfall of the r[1, j]th sample in the mth group; n m ---Number of samples in the mth group, n≥3.

[0049] Furthermore, the step 4 is calculated using the following formula:

[0050]

[0051] In the formula, is the average return period of the mth group of samples, N is the number of years of historical rainfall data, n m is the number of elements in the mth group of samples.

[0052] Compared with the current fitting method used for rainstorm return period estimation, the point rainstorm return period estimation method of the present invention has the following advantages:

[0053] (1) The method of the present invention has innovative features. The method of the present invention converts the rainstorm amount at the measured point into a dimensionless comprehensive parameter, equivalent rainfall, which can characterize the properties of the rainstorm (including the size of rainfall and the size of rainfall intensity), and eliminates the influence of the duration of rainfall on the rainstorm, so that different types of rainstorms (long-duration rainstorms and short-duration rainstorms) are not restricted by the duration of rainfall, and the size of rainfall can be compared. In this way, when the rainstorm samples are counted, not only the possibility of sample omission is avoided, but also the number of rainstorm samples is expanded, making the analysis results more credible.

[0054] (2) The method of the present invention is scientific. On the basis of sorting the equivalent rainfall corresponding to the rainfall of the measured rainstorm, according to the law of previous rainstorms: the recurrence frequency of heavy rainstorms is relatively low, and the recurrence frequency of small rainstorms is relatively high, the equivalent rainfall of the measured rainstorm is used as a sample, and a "pyramid" sample grouping method is adopted to construct a historical rainstorm recurrence model, and the design rainstorm recurrence period is calculated using the historical rainstorm recurrence model. In particular, the first group uses 3 samples, and the deviation of the samples is no more than 5%, which is not only more representative, but also eliminates the small probability random errors that may be caused by a few heavy rainstorms in the calculation of the recurrence period.

[0055] (3) The results are highly credible. The equivalent rainfall samples used to construct the point rainstorm return period model are all converted based on the historically measured rainstorm rainfall. It is statistically reasonable to establish the correlation between historical rainfall and the corresponding return period. From the estimated results of the return period of the larger rainstorms that have occurred and been recorded in history, it can be seen that the method is credible and feasible for estimating the return period of rainstorms.

[0056] (4) The method of the present invention is universal. On the one hand, the method of the present invention can not only be used to evaluate and calculate the return period of measured extreme heavy rainstorms or future designed heavy rainstorms, but can also be used to estimate the rainstorm rainfall corresponding to the design return period during engineering design. On the other hand, when the method is applied, it is not affected by statistical parameters such as the mean value, coefficient of deviation, and coefficient of skewness of the region and rainfall. If the rainfall and duration of the rainstorm are known and the corresponding equivalent rainfall is calculated, the rainstorm can be estimated. DETAILED DESCRIPTION

[0057] The present invention is further described in detail below based on the above technical solutions and embodiments.

[0058] The method for estimating the point rainstorm recurrence period (or recurrence frequency) of the present invention includes two parts: a method for ranking the relative sizes of rainfall of various types of point rainstorms and a method for constructing a recurrence period (or recurrence frequency) model (regression equation) based on historical point rainstorms on this basis to further estimate the recurrence period (or recurrence frequency) of the target point rainstorm.

[0059] The first part is the point rainstorm rainfall ranking method, which is aimed at various types of rainstorms in the study area, including rainstorms of different durations, rainstorms of different intensities, and rainstorms of different rainfall. By introducing characteristic parameters related to rainfall intensity attributes, the actual rainstorm rainfall is converted into equivalent rainfall, so that the rainfall of each rainstorm has the same invisible physical meaning, making their sizes comparable. It mainly includes the following steps:

[0060] Step 1: Determine the rainstorm rainfall threshold.

[0061] According to the historical rainfall characteristics in the study area, the minimum rainfall that belongs to the category of heavy rain is determined, that is, the heavy rain rainfall threshold (the heavy rain rainfall threshold may be different in different regions). (1) If the rainfall lasts no more than 24 hours, for semi-arid areas with low frequency of heavy rain and small rainfall, the heavy rain rainfall threshold is recommended to be 30 mm / 24 hours; for arid areas with extremely low frequency of heavy rain and small rainfall, the heavy rain rainfall threshold is recommended to be 20 mm / 24 hours; for humid areas with abundant rainfall, high frequency of heavy rain and large rainfall, the heavy rain rainfall threshold is recommended to be 50 mm / 24 hours. (It should be noted here that although my country's meteorological department uses 50mm / 24h as the classification standard for heavy rain, in most parts of my country, especially in the arid and semi-arid areas in the north, on the one hand, the frequency of rainfall above 50mm / 24h is too low, and even does not occur for many years. At the same time, due to the short rainfall data series in many places, it is impossible to analyze the possible recurrence period of extreme heavy rain; on the other hand, considering the rainfall disaster, the rainfall in many places in the north mostly appears in the form of short-duration heavy rainfall, and rainfall below 50mm / 24h can still cause serious rain and flood disasters. Therefore, the present invention The rainstorm threshold is determined according to the characteristics of different regions); (2) If the rainfall duration is greater than 24 hours (including intermittent periods of cessation of rainfall during the rainfall process), for every additional 24 hours of rainfall duration (considering that the rainfall data in the past were all based on a 24-hour statistical period, every additional 1-24 hours of rainfall duration is calculated as 24 hours), the corresponding rainfall threshold is recommended to be determined based on the rainstorm threshold for rainfall duration not exceeding 24 hours. Specifically, the rainstorm threshold is determined based on 2 / 3 of the rainstorm threshold for rainfall duration not exceeding 24 hours as the increase gradient. Taking semi-arid areas as an example, the 24-hour rainstorm threshold is 30 mm, the rainstorm threshold for continuous 48 hours of rainfall is 50 mm, the rainstorm threshold for continuous 72 hours of rainfall is 70 mm, and so on.

[0062] Step 2: Collect the rainfall information of heavy rains in previous years.

[0063] Select rainfall observation stations with continuous rainfall observation records for more than 20 years (including 20 years) in the study area, and count the rainstorm information of previous years. Specifically: according to the rainstorm rainfall thresholds of different rainfall durations determined in step 1, the rainfall information of a continuous rainfall process (not limited by the number of rainy days) in the historical rainfall data is counted, and the total rainfall is greater than or equal to the rainstorm rainfall threshold corresponding to the rainfall duration, mainly including rainfall (mm) and rainfall duration (24h or min). However, the rainfall in each 24h period cannot be less than 1 / 3 of the rainstorm rainfall threshold for a rainfall duration of no more than 24h, that is, not less than 7mm / 24h in arid areas, not less than 10mm / 24h in semi-arid areas, and not less than 17mm / 24h in humid areas. For example, in semi-arid areas, if continuous rainfall lasts for 48 hours, only rainfall events with a total rainfall of 50 mm or more in 48 hours will be counted, and the rainfall in each 24-hour period within the 48 hours must also be greater than or equal to 10 mm; if the rainfall lasts for 72 hours, only rainfall events with a total rainfall of 72 hours or more in 70 mm will be counted, and the rainfall in each 24-hour period must also be greater than or equal to 10 mm.

[0064] Step 3, calculate the average rainfall intensity (mm / min) of each rainstorm.

[0065] The average rainfall intensity (mm / min) of each rainstorm is obtained by dividing the rainfall amount (mm) of each rainstorm by the rainfall duration (min). Specifically, for rainfall data with rainfall duration in days (24 hours) as the monitoring and statistical unit, the rainfall duration corresponding to 1 day (24 hours) is 1440 minutes, and the rainfall duration corresponding to n days is n×1440 minutes; for rainfall data with rainfall duration in hours (h) as the statistical unit, it is converted into min to calculate the rainfall intensity, and the calculation result represents the average rainfall intensity in the corresponding period.

[0066] Step 4: Screen the effective rainstorm intensity.

[0067] Considering that heavy rain refers to rainfall with high rainfall intensity or rainfall, in order not to miss the heavy rain produced for many consecutive days and to exclude the influence of the light rainfall intensity for many consecutive days on the calculation results, the average rainfall intensity corresponding to the heavy rain rainfall threshold of three consecutive days is used as the minimum rainfall intensity of effective heavy rain. Therefore, the heavy rain records with small rainfall intensity are deleted from the result of the average rainfall intensity obtained in step 3 to obtain the effective heavy rain data. For example, for semi-arid areas, if the 72h rainfall threshold of 70mm is used as the standard, the rainfall records with average rainfall intensity i≤0.016mm / min can be deleted, and the rainfall records with average rainfall intensity i>0.016mm / min can be retained.

[0068] Step 5, calculate the equivalent rainfall P of the rainstorm.

[0069] The pure rainfall in physical sense is converted into a dimensionless comprehensive parameter, equivalent rainfall P, which has both rainfall attribute and rainfall intensity attribute. The calculation formula is shown in (1). The rainstorm rainfall R in the effective rainstorm data selected in step 4 (i.e., the measured rainfall corresponding to the statistics in step 2) is multiplied by a dimensionless conversion characteristic parameter η which is closely related to the rainstorm rainfall intensity attribute. The calculation of η is shown in formula (2). The problem that it was previously impossible to compare the rainfall of different rainfall durations (including any rainfall with a rainfall duration of less than 24 hours and more than 24 hours) is now comparable through attribute normalization.

[0070] P=η×R (1)

[0071] Where: P--equivalent rainfall; R--measured rainfall in mm; η--conversion characteristic parameter;

[0072]

[0073] Where: --The average rainfall intensity of a rainstorm, α—rainfall intensity characteristic index. For semi-arid areas, α is 0.20; for humid areas, α can be 0.15; for arid areas, α can be 0.25.

[0074] Step 6, determine the effective equivalent rainfall of the rainstorm.

[0075] Principle for determining effective equivalent rainfall: the equivalent rainfall corresponding to the local 24-hour rainfall threshold in 12 hours is used as the minimum equivalent rainfall of rainfall data required for rainstorm frequency analysis. From the equivalent rainfall P of all previous rainstorms obtained in step 5, the rainfall records with equivalent rainfall P less than the minimum equivalent rainfall are discarded, and the remaining equivalent rainfall is used as the effective equivalent rainfall of the corresponding rainstorm.

[0076] For example, for semi-arid areas, the equivalent rainfall corresponding to 30 mm rainfall in 12 hours (P = 16) is used as the minimum equivalent rainfall of the rainfall data required for the rainstorm frequency analysis. From the equivalent rainfall P of all previous rainstorms obtained in step 5, the rainfall records with equivalent rainfall P < 16 are discarded, and the equivalent rainfall with P ≥ 16 is retained as the effective equivalent rainfall. For humid areas with more annual rainfall and greater rainfall intensity, the minimum value of the effective equivalent rainfall can be relatively large.

[0077] Step 7, sort the equivalent rainfall.

[0078] The effective equivalent rainfall corresponding to each rainstorm obtained in step 6 is arranged in descending order according to the numerical value. If the equivalent rainfall (or rain intensity) value is large, it indicates that the actual rainfall of the rainstorm is either large, or the rainfall intensity is large, or both are large. In this case, either the runoff of the rainstorm is large, or the runoff intensity is large, and the overall destructiveness is greater, and the secondary disasters are more serious; otherwise, the opposite is true. Therefore, the sorting result reflects the relative size of the destructiveness that the actual rainstorm corresponding to different equivalent rainfall in the region has or may have.

[0079] The second part is to construct a calculation method for the return period (or recurrence frequency) of a point rainstorm. According to the research results of statistical hydrology, there is a certain recurrence period for the occurrence of rainstorms, and the recurrence periods of different levels of rainstorms are different. This method is based on the first part above (i.e., based on the ranking of the equivalent rainfall of historical measured point rainstorms). By constructing a correlation regression equation (recurrence period model) between the equivalent rainfall and its corresponding recurrence period, the recurrence period of measured (known) or designed extreme rainstorms can be calculated. It mainly includes the following steps:

[0080] Step 1: Selection of rainstorm data series.

[0081] Step 1.1: Does the rainstorm data in the study area have continuous observation records for more than 20 years (including 20 years)? If so, and there are several rainfall observation stations, then the rainstorm data with a longer rainfall data sequence is preferred, and step 1.2 is executed. Otherwise, new rainstorm data is selected from areas with similar climate and environment around the study area, and step 1.1 is executed;

[0082] Step 1.2, sort the rainstorm data according to the sorting method of the relative size of the rainstorm rainfall in the first part above, and obtain the effective rainstorm equivalent rainfall in descending order;

[0083] Step 1.3: The number of effective rainstorm equivalent rainfall obtained in step 1.2 must be greater than 40, that is, more than 40 rainfall data are required. If so, take them as effective equivalent rainfall samples and execute step 2; otherwise, select new rainstorm data in areas with similar climate and environment around the study area and execute step 1.1;

[0084] Step 2: Group the effective equivalent rainfall samples.

[0085] The effective equivalent rainfall samples obtained in step 1 are grouped in descending order. The grouping rule adopts arithmetic grouping or geometric grouping, that is, the number of samples in the latter group increases in arithmetic or geometric relationship with the number of samples in the former group (i.e. "pyramid" sample grouping), and the grouping meets the following conditions:

[0086] (1) Grouping starts from the first sample in the sample group (the sample with the largest value);

[0087] (2) The samples in each group must be continuous in size order;

[0088] (3) The number of samples n1 in the first group is 3 only, and the deviation (i.e., deviation) of the value of each sample in the first group from the average value of the samples in the group cannot exceed 5% (this is to improve the consistency of the sample function in the first group). If it exceeds, the sample with the larger value in the sample group is discarded, and the sample group is supplemented in order of size until the requirement is met;

[0089] (4) The number of groups the sample is grouped into must be greater than or equal to 4 (i.e., if the number of sample groups is less than 4, return to step 1 in the second part);

[0090] (5) The second group selects samples following the sample sequence of the first group. If the arithmetic difference method is used for the second and subsequent sample groupings, the preferred value of the difference is 2 or 3, that is, the number of samples in the latter group is 2-3 more than that in the former group. If the geometric ratio method is used, the preferred value of the common ratio is 1.5-2, that is, the number of samples in the latter group is 1.5-2 times that of the former group.

[0091] (6) The number of samples in the last group must be sufficient and must be valid samples (i.e., if the number of remaining samples is insufficient for further grouping, they will be discarded and not grouped);

[0092] Step 3: Calculate the average effective equivalent rainfall of each group of samples obtained in step 2

[0093] Specifically, the sum of the effective equivalent rainfall of each group of rainstorms is divided by the corresponding number of samples, that is:

[0094]

[0095] Where: ---The average effective equivalent rainfall of the mth group of samples; P r ---The effective equivalent rainfall of the r[1, j]th sample in the mth group; n m ---Number of samples in group m (n≥3)

[0096] Step 4: Calculate the average return period for each sample group Specifically, the number of years of historical rainfall data series N (N ≥ 20) is divided by the number of elements in each group of samples n m ,Right now:

[0097]

[0098] Step 5: Construct a regression model for the relationship between equivalent rainfall and return period.

[0099] According to the average effective equivalent rainfall of each sample group obtained in step 3 and the average return period of each sample group obtained in step 4, the average equivalent rainfall The horizontal axis is the average return period As the vertical coordinate, the correlation between equivalent rainfall and return period is constructed, and it is recommended to use polynomials to establish a regression equation. The equation based on polynomial regression is shown in formula (5):

[0100] T=a·P 2 +b·P+c (5)

[0101] In the formula: T---corresponding to the recurrence period (year) of the rainstorm with equivalent rainfall of P, that is, the magnitude of this rainstorm is a rainstorm that occurs once in T years; P---equivalent rainfall of the rainstorm; a, b, c---coefficients and constant terms of the polynomial regression equation.

[0102] Step 6: Determine the recurrence period or frequency corresponding to the measured (or designed) extreme heavy rain rainfall.

[0103] According to the known extreme heavy rain rainfall and duration (or design heavy rain rainfall and duration), firstly, use step 3 of the first part to obtain the average rainfall intensity i of the heavy rain, and use formulas (1) and (2) in step 5 to calculate the equivalent rainfall corresponding to the known heavy rain (or design heavy rain). Secondly, substitute the equivalent rainfall P into formula (5) obtained in step 5 to calculate the recurrence period T (in years) corresponding to the heavy rain. The value T means that under similar climate and environmental conditions, a heavy rain of this magnitude may occur again in the next T years. The reciprocal of the recurrence period T is the annual recurrence probability of the heavy rain. The annual recurrence probability is expressed as a percentage, thousandth or ten thousandth, which is the recurrence period frequency of the target heavy rain.

[0104] Example:

[0105] In early August 2020, a place in the north received 214.6 mm of rainfall for two consecutive days, exceeding the maximum rainfall in history. The rainfall on the first day was 152.4 mm, and the rainfall on the second day was 62.2 mm, causing serious soil and water loss such as collapse of terraces and roads and slope scour. In order to assess the magnitude of the rainstorm and the flood control capacity of water conservancy projects, it is necessary to know how many years this rainstorm is a rainstorm, that is, the possible recurrence frequency or recurrence period.

[0106] According to the technical solution of the present invention, firstly, the rainfall of the local point rainstorms that have occurred is sorted according to the size of the corresponding amount of rainfall.

[0107] Step 1: Obtain rainfall data. Select a local rainfall observation station with more than 20 consecutive years of rainfall observation records and collect detailed rainfall data over the years. The meteorological station that collected this extreme rainfall has 21 years (2000-2020) of complete meteorological data, including rainfall data, so it meets the requirements of rainfall data sequence.

[0108] Step 2: Determine the minimum rainstorm rainfall threshold. According to the historical rainfall characteristics in the study area, determine the minimum rainfall that belongs to the category of rainstorm (rainfall with disastrous effects), that is, the rainstorm rainfall threshold. The local area belongs to a semi-arid area with relatively small rainfall and low frequency, so the 24h (1 day) rainfall of 30mm is taken as the rainstorm rainfall threshold. For two consecutive days or more of rainfall, the corresponding rainfall threshold is based on 30mm per day. If the rainfall duration increases by 1 day, the rainfall threshold will increase by 20mm. And the actual rainfall in the corresponding time period is not less than 10mm.

[0109] Step 3: Statistical data of “rainstorms” over the years. Statistics are collected for the rainfall information of a continuous rainfall process (not limited by the number of rainy days) in the historical rainfall data with rainfall greater than the threshold, mainly including the rainfall amount (mm) and rainfall duration (d). According to the historical time sequence of the rainstorms, the statistical results are shown in columns (1)-(3) and (6)-(8) in Table 1 below.

[0110] Table 1. Statistical calculation results of rainstorm data in previous years

[0111]

[0112]

[0113] Step 4: Calculate the average rainfall intensity (mm / min) of each rainstorm and delete the rainfall information with an average rainfall intensity less than or equal to 0.016 mm / min. The results are shown in columns (4) and (9) of Table 1.

[0114] Step 5: Calculate the equivalent rainfall P of the rainstorm. Multiply the actual rainfall of the rainstorm by the dimensionless conversion parameter The corresponding equivalent rainfall results are shown in columns (5) and (10) of Table 1.

[0115] Step 6, determine the effective equivalent rainfall P of the rainstorm. Delete the rainfall records with equivalent rainfall P < 16, and keep the equivalent rainfall with P ≥ 16 as the effective equivalent rainfall. The results are shown in columns (3) and (7) of Table 2.

[0116] Table 2. Calculation and ranking of rainstorm equivalent rainfall

[0117]

[0118]

[0119] Step 7: Sort the equivalent rainfall. Sort the effective equivalent rainfall in descending order according to the numerical value. The sorting results are shown in columns (4) and (8) of Table 2.

[0120] Secondly, on the basis of completing the above steps, the correlation between the equivalent rainfall corresponding to the heavy rain at the historical measured point and the return period (return period model) is constructed. It mainly includes the following steps:

[0121] Step 8, selection of rainstorm data series. According to the previous sorting of effective equivalent rainfall, the number of rainstorm samples (effective equivalent rainfall) recorded by the meteorological station in 21 years is 53, which meets the requirements.

[0122] Step 9, group the measured rainstorm samples. According to the grouping rules, select 3 samples n1 in the first group, and start grouping from the first sample. After calculation, due to the large difference between the values ​​of the first three samples in the sample ranking (column 4 of Table 2), the sample combinations [1,2,3], [2,3,4], and [3,4,5] do not meet the requirement that the deviation of the sample value is less than or equal to 5%. Only the combination of samples [4,5,6] has a maximum upper and lower deviation of 3%, which meets the requirement that the deviation of the sample value is less than or equal to 5%. Therefore, samples No. 4, 5, and 6 are selected as the first group; the second group starts from sample No. 7 in the sample ranking of column 4 of Table 2, and adopts the geometric series grouping method. The number of samples n2 is 1.5 times that of the first group, that is, 5 samples, and the five samples [7, 8, 9, 10, 11] in the sample ranking are selected; the number of samples n3 in the third group is 1.5 times that of the second group n2, that is, 8 samples; and grouping is performed by analogy. The grouping results are shown in columns (3) and (8) of Table 3 below.

[0123] Step 10: Calculate the average equivalent rainfall for each group of rainstorms Divide the sum of the equivalent rainfall of each group of rainstorms by the corresponding number of samples n m , the average equivalent rainfall corresponding to each group of measured rainstorms See columns (4) and (9) in Table 3.

[0124] Step 11: Calculate the average return period of each group of measured rainstorm samples Divide the number of years of the measured rainfall data series N (N = 21) by the number of samples in each group n m , the average return period of each group of measured rainstorms See columns (5) and (10) in Table 3.

[0125] Step 12: Construct a regression model for the relationship between the return period and the equivalent rainfall. The average equivalent rainfall and the corresponding average return period in Table 3 are statistically analyzed, and the results are shown in Table 4.

[0126] Table 3. Sample grouping results

[0127]

[0128]

[0129] Taking the average equivalent rainfall as the horizontal axis and the average return period as the vertical axis, the correlation between the two is regressed using polynomial regression. The correlation between the two is shown in the following formula (9).

[0130] Table 4. Statistics of mean values ​​of each group

[0131] Group number Average equivalent rainfall Average return period (years) 5 20.1 1.17 4 27.8 1.75 3 33.8 2.6 2 41 4.2 1 49 7

[0132] T=0.0065P 2 -0.2485P+3.5816 (9)

[0133] R 2 =0.9993

[0134] In the formula: T---the return period (year) of a rainstorm with an equivalent rainfall of P; P---the equivalent rainfall of a rainstorm. From the correlation coefficient, it can be seen that the equivalent rainfall of a rainstorm that has occurred has a high correlation with the return period.

[0135] Step 13, the recurrence period or frequency corresponding to the target rainstorm rainfall is determined. For the heavy rainstorm that occurred in a certain place in the north in early August 2020, the rainfall for two consecutive days was 214.6 mm, that is, the rainfall duration was 2880 min, and the corresponding average rainfall intensity was 0.0745 mm / min. The equivalent rainfall of this rainfall was P = 127.7.

[0136] Substituting the equivalent rainfall of 127.7 corresponding to the rainfall of 214.6 mm into equation (9),

[0137] So: T = 78 (years),

[0138] That is, the rainstorm recurrence period corresponding to the local rainfall of 214.6 mm in two days is 78 years (scientifically speaking, it should be about 78 years), and the corresponding recurrence frequency is 1.23%. That is, under the condition of unchanged climate and environment, rainstorms of this magnitude may occur again in the next 80 years, or the probability of rainstorms of this magnitude occurring again in the local area is 1.23%, that is, the probability of occurrence in the next 100 years is 1.23 times.

[0139] For a daily rainfall of 152.4 mm, the corresponding equivalent rainfall is 97. Substituting it into formula (9), we get: T = 41 (years)

[0140] The same method can be used to derive the return periods of the two days of rainfall of 133.9 mm and 121.3 mm (statistical rainfall data) that occurred in the local area, which are 20 years and 15 years respectively. The above two types of rainstorms have only occurred once in the 21-year rainfall record, which also shows that the estimation method of the return period of this rainstorm is credible.

[0141] According to the investigation, in the mid-1970s, another place in the same landform type area about 100 km away from the study area had a heavy rainstorm with a 12-hour rainfall of 320 mm and a 24-hour rainfall of 400 mm, which caused extremely serious flood disasters. According to the method proposed in this patent, the equivalent rainfall corresponding to the rainstorm rainfall is 272 and 310 respectively. Substituting them into the return period equation (9), the corresponding return periods can be obtained as 417 years and 551 years (the recurrence frequency is about 2‰). Combined with the local climate characteristics, this result has a high degree of credibility.

Claims

1. A method for sorting point rainstorm rainfall, characterized in that: The specific steps include: Step 1: Determine the rainstorm rainfall thresholds at different rainfall durations based on the historical rainfall characteristics in the study area; Step 2, counting rainfall information in which the total rainfall of a continuous rainfall process in the historical rainfall data in the study area is greater than or equal to the rainstorm rainfall threshold corresponding to the rainfall duration, and the rainfall information includes rainfall amount and rainfall duration; Step 3, dividing the rainfall of each rainstorm by the rainfall duration to obtain the average rainfall intensity of each rainstorm; Step 4, deleting the rainstorm records with smaller rainfall intensity from the average rainfall intensity obtained in step 3 to obtain valid rainstorm data; specifically, taking the average rainfall intensity corresponding to the rainstorm rainfall threshold of three consecutive days of rainfall as the minimum rainfall intensity of valid rainstorm, deleting the rainstorm records with less than the minimum rainfall intensity from the result of the average rainfall intensity obtained in step 3 to obtain valid rainstorm data; Step 5, calculating the equivalent rainfall of each rainstorm based on the effective rainstorm data obtained in step 4; The equivalent rainfall P of a rainstorm can be calculated using formulas (1) and (2): P = η × R （1） （2） Where: P -- equivalent rainfall; R --Measured rainfall in mm; η-- Conversion feature parameters; --Average rainfall intensity of rainstorm, α—rainfall intensity characteristic index; For semiarid, humid, and arid regions, the values ​​of α are 0.20, 0.15, and 0.25, respectively; Step 6, determining the effective equivalent rainfall of the rainstorm according to the equivalent rainfall of each rainstorm obtained in step 5; Step 7, arranging the effective equivalent rainfall corresponding to each rainstorm obtained in step 6 in descending order.

2. The method for sorting point rainstorm rainfall according to claim 1, characterized in that: The step 1 specifically includes: (1) If the rainfall duration is no more than 24 hours, the rainstorm rainfall thresholds for semi-arid areas, arid areas, and humid areas are 30 mm / 24 hours, 20 mm / 24 hours, and 50 mm / 24 hours, respectively; (2) If the rainfall duration is greater than 24 hours, the rainfall threshold corresponding to the rainfall duration increases by 24 hours, with a gradient of 2 / 3 of the rainfall threshold for a heavy rain with a rainfall duration of no more than 24 hours.

3. The method for sorting point rainstorm rainfall according to claim 2, characterized in that: In step 2, the rainfall information obtained by statistics should also meet the following requirements: the rainfall in each 24-hour period is not less than 1 / 3 of the rainfall threshold of a rainstorm with a rainfall duration of no more than 24 hours.

4. A method for estimating the return period of a point rainstorm, characterized in that: The specific steps include: Step 1, selecting the rainstorm observation records with more than 20 consecutive years in the study area; using the point rainstorm rainfall sorting method described in any one of claims 1-3 to calculate the equivalent rainfall of each rainstorm, determine the effective equivalent rainfall of the rainstorm and sort it; if the number of effective rainstorm equivalent rainfall is greater than 40, execute step 2, otherwise return to step 1; Step 2, grouping the effective equivalent rainfall samples obtained in step 1; Step 3, calculating the average effective equivalent rainfall of each group of samples obtained in step 2; Step 4, calculating the average recurrence period of each sample group based on the results of step 3; Step 5, constructing a regression model of the relationship between equivalent rainfall and return period based on the average effective equivalent rainfall of each sample group obtained in step 3 and the average return period of each sample group obtained in step 4; Step 6: Calculate the return period of the measured or designed extreme heavy rain rainfall according to the regression model of step 5.

5. The method for estimating the point rainstorm return period according to claim 4, characterized in that: The step 1 includes the following sub-steps: Step 1.1: Check whether the rainstorm data in the study area has continuous observation records for more than 20 years. If so, and there are several rainfall observation stations, select the rainstorm data with the longest rainfall data sequence and execute step 1.

2. Otherwise, select new rainstorm data in areas with similar climate and environment around the study area and execute step 1.

1. Step 1.2, according to the point rainstorm rainfall sorting method, obtain the effective rainstorm equivalent rainfall in descending order; Step 1.3: The number of effective rainstorm equivalent rainfall obtained in step 1.2 must be greater than 40, that is, more than 40 rainfall data are required. If so, use them as effective equivalent rainfall samples and execute step 2; otherwise, select new rainstorm data in areas with similar climate and environment around the study area and execute step 1.

1.

6. The method for estimating the point rainstorm return period according to claim 4, characterized in that: The step 2 is performed as follows: The effective equivalent rainfall samples obtained in step 1 are grouped in descending order. The grouping rule adopts arithmetic grouping or geometric grouping, that is, the number of samples in the latter group increases in an arithmetic or geometric relationship with the number of samples in the previous group, and the grouping meets the following conditions: (1) Grouping starts from the first sample in the sample group; (2) The samples in each group are continuous in size order; (3) The number of samples n1 in the first group is 3 if and only if 3 samples are selected, and the deviation between the value of each sample in the first group and the average value of the samples in the group cannot exceed 5%. If it exceeds, the sample with the largest value in the sample group is discarded, and the sample group is supplemented in order of size until the requirement is met; (4) The number of groups the sample is grouped into is greater than or equal to 4; (5) The second group selects samples from the same sequence as the first group. If the second and subsequent groups are grouped by arithmetic difference method, the difference is 2 or 3; if the geometric ratio method is used, the common ratio is 1.5-2; (6) The number of samples in the last group must be sufficient and valid.

7. The method for estimating the point rainstorm return period according to claim 4, characterized in that: The step 3 is calculated using the following formula: （r=1,2,3,4,5…j） In the formula, ---The average effective equivalent rainfall of the mth group of samples; P r ---Effective equivalent rainfall of the r[1, j]th sample in the mth group; n m ---Number of samples in the mth group, n≥3.

8. The method for estimating the point rainstorm return period according to claim 4, characterized in that: The step 4 is calculated using the following formula: = N / n m In the formula, is the average return period of the mth group of samples, N is the number of years of historical rainfall data, n m is the number of elements in the mth group of samples.

Images