CN114609698A - Short-time strong precipitation probability forecasting method based on proportioning method idea and dichotomy - Google Patents
Short-time strong precipitation probability forecasting method based on proportioning method idea and dichotomy Download PDFInfo
- Publication number
- CN114609698A CN114609698A CN202210084321.9A CN202210084321A CN114609698A CN 114609698 A CN114609698 A CN 114609698A CN 202210084321 A CN202210084321 A CN 202210084321A CN 114609698 A CN114609698 A CN 114609698A
- Authority
- CN
- China
- Prior art keywords
- short
- precipitation
- time
- time strong
- probability
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
Landscapes
- Environmental & Geological Engineering (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Atmospheric Sciences (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a short-time strong precipitation probability forecasting method based on a batching method idea and a dichotomy, which comprises the following steps of: step one, analyzing precipitation data; step two, constructing a forecasting model; step three, checking and forecasting effects; in the step one 2), when the time-space characteristics of short-time strong precipitation are analyzed, the precipitation amount of any one station of the 13 observation stations at a certain moment is more than or equal to 20 mm/h, or the precipitation amount is more than or equal to 20 mm/h for multiple times within 2 hours, and all the sites are regarded as the same short-time strong precipitation event; in the second step 2), screening out the short-time strong precipitation potential forecasting factors of 850hPa specific humidity, 850hPa pseudo equivalent temperature, K index, convection effective potential, 700hPa wind direction and 700hPa vertical speed, and basically covering the water vapor condition, stability condition and lifting condition required by the generation of short-time strong precipitation; in the third step 5), the reference probability threshold of the short-time strong precipitation forecast is set to be 0.98, so that the high forecast accuracy can be kept.
Description
Technical Field
The invention relates to the technical field of weather forecasting, in particular to a short-time strong precipitation probability forecasting method based on a batching method idea and a dichotomy.
Background
Defining short-time strong precipitation as a precipitation process with precipitation amount of more than 20mm in one hour on meteorological service; the process has the characteristics of strong burst property, high rainfall intensity and high forecasting difficulty, has great influence on urban traffic transportation and people production and life, and is more likely to cause geological disasters such as torrential flood, debris flow and the like, so that the research on the forecasting method of the short-time strong rainfall has important practical significance;
the adult city is located in the west of the Sichuan basin, the east is provided with a Longquan mountain, the west is provided with an even and wide mountain, and the short-time strong rainfall has obvious regional characteristics; at present, the short-time strong precipitation feature analysis of metropolis and even Sichuan basins has been studied, but the research on the short-time strong precipitation potential forecasting method of metropolis is not frequent.
Disclosure of Invention
The invention aims to provide a short-time strong precipitation probability forecasting method based on a batching method idea and a dichotomy, and aims to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a short-time strong precipitation probability forecasting method based on a batching method idea and a dichotomy comprises the following steps: step one, analyzing precipitation data; step two, constructing a forecasting model; step three, checking and forecasting effects;
wherein in the first step, analyzing precipitation data comprises the following steps:
1) obtaining data: the precipitation data adopts hourly precipitation records of 13 meteorological stations 2010-2019 in 4-9 months in metropolis, and the specific observation time is 20:00 to 20:00 of the day before; in the potential forecast, physical quantity analysis adopts ERA-5 hour-by-hour data analysis issued by European mesoscale weather forecast center, and grid resolution is 0.25 degree multiplied by 0.25 degree;
2) and (3) analysis of data: analyzing the time-space characteristics of the short-time strong precipitation according to the data obtained in the step 1) to obtain the time-space distribution characteristics of the Chengdu short-time strong precipitation, namely that urban short-time strong precipitation events are mostly concentrated in warm seasons, particularly in 7 and 8 months; carrying out bilinear interpolation on the ERA-5 reanalysis data according to the positions of 13 meteorological stations to obtain physical quantity data of the corresponding stations; in the second step, the construction of the forecasting model comprises the following steps:
1) and (3) correlation analysis: point-biserial correlation analysis is carried out on a dichotomized sequence of short-time strong precipitation days and an average physical quantity sequence of each station day, and common physical quantity elements in the research and forecast of the short-time strong precipitation are divided according to a water vapor condition, a thermodynamic stability condition, a dynamic stability condition and a lifting condition to obtain main physical quantity factors of the urban short-time strong precipitation, wherein each physical quantity factor is calculated by dividing the physical quantity factors into 500, 700 and 850hPa in different levels; performing correlation analysis on the four basic elements and the 1, 3 and 6 hour variable quantities thereof and the binary sequence of the short-time strong precipitation to obtain factors with the absolute value of the relation number greater than 0.3 in each element, thereby obtaining the main relevant physical quantity of the short-time strong precipitation in the metropolis;
2) screening a forecasting factor: screening main relevant physical quantities of the urban short-time heavy rainfall acquired in the step 1), and screening out physical quantity factors which are better in relevance to the short-time heavy rainfall;
3) establishing a short-time strong precipitation probability forecasting equation: calculating the short-time strong precipitation occurrence probability of each grid point by adopting a dichotomy;
4) calculating the weight coefficient of each parameter of the probability forecasting equation: based on the physical quantity factor box line graphs of the short-time strong precipitation day and the no short-time strong precipitation day, the upper quartile value and the lower quartile value when the short-time strong precipitation occurs can be set as a judgment threshold value of the parameter, the proportion of the crossed part of the two box bodies occupying the short-time strong precipitation box body is used for calculating a weight coefficient of the short-time strong precipitation box body, and a main parameter threshold value and the weight coefficient of the urban summer short-time strong precipitation probability forecast are obtained;
in the third step, the step of testing the forecasting effect comprises the following steps:
1) acquiring data to be checked: acquiring a forecast parameter value corresponding to each grid point in 2019 in 6-8 months by using ERA-5 data in 2019 of a metropolis city;
2) calculating a probability value: calculating the probability value of each grid point at each moment of the day by using the calculation method of the short-time strong precipitation occurrence probability of the single grid point mentioned in the step two 3), and interpolating the probability value to 13 sites;
3) judging and forecasting results: setting a probability threshold, considering that a short-time strong precipitation event exists in a certain day when the station probability is greater than the threshold, comparing the short-time strong precipitation event with station precipitation data, and judging whether the station precipitation event is correct or incorrect;
4) calculating a TS forecast score: calculating a TS forecast score according to the judgment result of the step 3) to obtain forecast quality under different threshold settings;
5) determining a reference probability threshold: selecting a proper threshold value as a reference probability threshold value of the urban short-time strong precipitation forecast according to the forecast quality of different threshold values obtained in the step 4).
Preferably, in the step one 2), when the temporal-spatial characteristics of the short-time strong precipitation are analyzed, the precipitation amount of any one of the 13 observation stations at a certain moment is more than or equal to 20 mm/h, or the precipitation amount of any one of the 13 observation stations is more than or equal to 20 mm/h within 2 hours, and all the sites are regarded as the same short-time strong precipitation event.
Preferably, in the step two 1), when performing point double row correlation analysis, if one group of variables is natural binary variables, point double row correlation coefficients need to be calculated; if the continuous variable is artificially divided into two variables, calculating a double-row correlation coefficient;
the calculation formula of the point biserial correlation coefficient is as follows:
wherein the content of the first and second substances,
and
respectively are the average values of continuous variable sequences corresponding to 0 and 1 in the binary sequence;
unbiased estimation of standard deviation for continuous variable sequence; p is a radical of0、p1Respectively the proportion of 0 and 1 in the binary sequence;
the calculation formula of the biserial correlation coefficient is as follows:
wherein h is the partition p0And p1The ordinate value of (c) is given by the above two formulae:
preferably, in the step two 2), the specific operation of predictor screening is as follows: dividing the short-time strong precipitation of 2010-2018 months into two sets according to existence and nonexistence, making various physical quantity factor box graphs of a short-time strong precipitation day and a non-short-time strong precipitation day, and selecting six physical quantity factors with good correlation with the short-time strong precipitation as the potentiality forecast equation independent variables according to the principle that the ratio of the overlapped part of each factor box body is small and comprehensively considering the water vapor condition, the thermodynamic stability condition, the dynamic stability condition and the lifting condition on the basis of the correlation analysis of various physical quantity factors and the short-time strong precipitation.
Preferably, in the second step 2), the short-term strong precipitation potential forecasting factors are selected to be 850hPa specific humidity, 850hPa pseudo equivalent temperature, K index, convective effective potential, 700hPa headwind and 700hPa vertical speed, and basically cover the moisture condition, stability condition and lifting condition required by the generation of short-term strong precipitation.
Preferably, in the step two 3), the calculation formula of the short-time strong precipitation occurrence probability of a single lattice point is as follows:
wherein N is the number of forecast parameters; when the value of the ith parameter falls within the threshold range, set AiIs 1, otherwise is 0; w is aiFor the weighting coefficients of parameter i, the sum of the weighting coefficients of all parameters is equal to 1.
Preferably, in the step two 4), the weight coefficient calculation formula is:
wherein QiThe i-th parameter represents the proportion (in%) of the intersection of the two boxes in the short-term strong precipitation box.
Preferably, in the step three 4), the calculation formula of the TS forecast score is as follows:
wherein NA is the number of times of the short-term strong precipitation and the correct forecast, NB is the number of times of the empty forecast, NC is the number of times of the missed forecast, and ND is the number of times of the no short-term strong precipitation and the no forecast.
Preferably, in the step three 5), the reference probability threshold of the short-time strong precipitation forecast is set to 0.98, so that a high forecast accuracy can be maintained.
Compared with the prior art, the invention has the beneficial effects that: based on the idea of a batching method, the urban short-time strong rainfall potentiality prediction factor is screened out, and the urban short-time strong rainfall probability prediction equation is constructed, so that more scientific bases are provided for urban short-time strong rainfall prediction services, the method can effectively perform potentiality prediction on urban short-time strong rainfall days, and the method has important practical significance.
Drawings
FIG. 1 is a schematic diagram of site and grid distribution of metropolis weather station;
FIG. 2 is an annual distribution diagram of a 2010-2019 annual urban short-term heavy precipitation event;
FIG. 3 is a monthly distribution graph of 2010-2019 annual urban short-term heavy precipitation events;
FIG. 4 is a characteristic diagram of daily change of occurrence frequency of short-term strong precipitation in 2010-2019 adult cities;
FIG. 5 is a distribution diagram of the average precipitation in the annual urban city of 2010-2019;
FIG. 6 is an annual average distribution diagram of 2010-2019 annual short-term heavy precipitation events in adult cities;
FIG. 7 is a diagram of a box of wet short-term heavy precipitation days and non-short-term heavy precipitation days with a physical quantity of 850 hPa;
FIG. 8 is a diagram of a physical quantity of 700hPa versus a wet short-time heavy precipitation day and a non-short-time heavy precipitation day box;
FIG. 9 is a line graph of a short-time heavy precipitation day and a non-short-time heavy precipitation day at a relative humidity of 850hPa as a physical quantity;
FIG. 10 is a box diagram of 850hPa false equivalent temperature short time heavy precipitation days and non-short time heavy precipitation days;
FIG. 11 is a diagram of a case of a short-time heavy precipitation day and a non-short-time heavy precipitation day with a physical quantity of K index;
FIG. 12 is a box diagram of short-duration heavy precipitation days and non-short-duration heavy precipitation days with a physical quantity of 700hPa false equivalent temperature;
FIG. 13 is a diagram of a case of convection effective potential energy short-time heavy precipitation days and non-short-time heavy precipitation days with physical quantities;
FIG. 14 is a box diagram of short-time heavy precipitation days and non-short-time heavy precipitation days with physical quantities of a thermodynamic total index;
FIG. 15 is a diagram of the short-term strong precipitation day and the non-short-term strong precipitation day box with the physical quantity of 850-500hPa pseudo equivalent temperature difference;
FIG. 16 is a diagram of a case of short-term strong precipitation days and non-short-term strong precipitation days with physical quantities of the Sauter's index;
FIG. 17 is a diagram of a radial wind-time short-time heavy precipitation day and non-short-time heavy precipitation day box with a physical quantity of 700 hPa;
FIG. 18 is a line graph of the short term heavy precipitation days and non-short term heavy precipitation days at a vertical speed of 700hPa for a physical quantity;
FIG. 19 is a line graph of the short-term heavy precipitation days and non-short-term heavy precipitation days with the physical quantity of 850hPa divergence;
FIG. 20 is a diagram of a case of short-term heavy precipitation days and non-short-term heavy precipitation days at a vertical speed of 500hPa as a physical quantity;
FIG. 21 is a graph of TS variation for different probability thresholds;
FIG. 22 is a flow chart of a method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
Referring to fig. 1-22 and tables 1-3, an embodiment of the present invention is provided: a short-time strong precipitation probability forecasting method based on a batching method idea and a dichotomy comprises the following steps: step one, analyzing precipitation data; step two, constructing a forecasting model; step three, checking and forecasting effects;
wherein in the first step, analyzing precipitation data comprises the following steps:
1) obtaining data: the precipitation data adopts hourly precipitation records of 13 meteorological stations 2010-2019 in 4-9 months in metropolis, and the specific observation time is 20:00 to 20:00 of the day before; in the potential forecast, physical quantity analysis adopts ERA-5 hourly reanalysis data issued by a European mesoscale weather forecast center, the grid resolution is 0.25 degrees multiplied by 0.25 degrees, and the spatial distribution of site positions and grid data is shown in figure 1;
2) and (3) analysis of data: analyzing the space-time characteristics of the short-time strong precipitation according to the information obtained in the step 1) to obtain the space-time distribution characteristics of the short-time strong precipitation of the city, namely that urban short-time strong precipitation events are mostly concentrated in warm seasons, especially in 7 and 8 months (as shown in figures 2-6), and when analyzing the space-time characteristics of the short-time strong precipitation, regarding the precipitation amount of any one station in 13 observation stations at a certain moment as more than or equal to 20 mm/h, or regarding the precipitation amount as more than or equal to 20 mm/h for multiple times within 2 hours, and regarding the precipitation events as the same short-time strong precipitation event; carrying out bilinear interpolation on the ERA-5 reanalysis data according to the positions of 13 meteorological stations to obtain physical quantity data of the corresponding stations;
in the second step, the construction of the forecasting model comprises the following steps:
1) and (3) correlation analysis: point-biserial correlation analysis is carried out on the binary sequence of the short-time strong precipitation day and the daily average physical quantity sequence of each station, and the physical quantity elements commonly used in the research and forecast of the short-time strong precipitation are divided according to the water vapor condition, the thermodynamic stability condition, the dynamic stability condition and the lifting condition to obtain the main physical quantity factors of the urban short-time strong precipitation, wherein each physical quantity factor is calculated by dividing the physical quantity factors into 500hPa, 700hPa and 850hPa at different levels (as shown in Table 1); performing correlation analysis on the four basic elements and the 1, 3 and 6 hour variable quantities thereof and the binary sequences of the short-time strong precipitation to obtain factors (shown in a table 2) with the absolute value of the relation number being greater than 0.3 in each element, thereby obtaining the main relevant physical quantity of the urban short-time strong precipitation; when point double-row correlation analysis is carried out, if one group of variables are natural binary variables, point double-row correlation coefficients need to be calculated; if the continuous variable is artificially divided into two variables, calculating a double-row correlation coefficient;
the calculation formula of the point biserial correlation coefficient is as follows:
wherein, the first and the second end of the pipe are connected with each other,
and
respectively are the average values of continuous variable sequences corresponding to 0 and 1 in the binary sequence;
unbiased estimation of standard deviation for continuous variable sequence; p is a radical of0、p1Respectively the proportion of 0 and 1 in the binary sequence;
the calculation formula of the biserial correlation coefficient is as follows:
wherein h is the partition p0And p1The ordinate value of (d) is given by the following two formulae:
2) screening a forecasting factor: screening main relevant physical quantities of the urban short-time heavy rainfall acquired in the step 1), and screening out physical quantity factors which are better in relevance to the short-time heavy rainfall; the specific operation of the forecast factor screening is as follows: dividing the short-time strong precipitation of 2010-2018 months into two sets according to existence and nonexistence, making various physical quantity factor box graphs (shown in figures 7-20) of a day with the short-time strong precipitation and a day without the short-time strong precipitation, and selecting six physical quantity factors with good correlation with the short-time strong precipitation as potential forecast equation independent variables according to the principle that the proportion of overlapped parts of various factor box bodies is small and comprehensively considering a water vapor condition, a thermal stability condition, a dynamic stability condition and a lifting condition on the basis of the correlation analysis of various physical quantity factors and the short-time strong precipitation: screening out the potential forecasting factors of the short-time strong precipitation, which are 850hPa specific humidity, 850hPa false equivalent potential temperature, K index, convection effective potential energy, 700hPa wind and 700hPa vertical speed, and basically covering the water vapor condition, stability condition and lifting condition required by the generation of the short-time strong precipitation;
3) establishing a short-time strong precipitation probability forecasting equation: calculating the short-time strong precipitation occurrence probability of each grid point by adopting a dichotomy; the short-time strong precipitation occurrence probability calculation formula of a single lattice point is as follows:
wherein N is the number of forecast parameters; when the value of the ith parameter falls within the threshold range, set AiIs 1, otherwise is 0; w is aiThe weighting coefficients of the parameter i are added to be equal to 1;
4) calculating the weight coefficient of each parameter of the probability forecasting equation: based on the physical quantity factor box line graphs of the short-time strong precipitation day and the no short-time strong precipitation day, the upper quartile value and the lower quartile value when the short-time strong precipitation occurs can be set as a judgment threshold value of the parameter, the proportion of the crossed part of the two box bodies occupying the short-time strong precipitation box body is used for calculating a weight coefficient of the short-time strong precipitation box body, and a main parameter threshold value and the weight coefficient of the urban summer short-time strong precipitation probability forecast are obtained; the weight coefficient calculation formula is as follows:
wherein QiThe proportion (unit is%) of the crossed part of the two boxes in the ith parameter in the short-time strong precipitation box is expressed, and the threshold value and the weight coefficient of each parameter of the probability forecast equation are obtained as shown in the table 3; in the third step, the step of testing the forecasting effect comprises the following steps:
1) acquiring data to be checked: acquiring a forecast parameter value corresponding to each grid point in 2019 in 6-8 months by using ERA-5 data in 2019 of a metropolis city;
2) calculating a probability value: calculating the probability value of each grid point at each moment of the day by using the calculation method of the short-time strong precipitation occurrence probability of the single grid point mentioned in the step two 3), and interpolating the probability value to 13 sites;
3) judging and forecasting results: setting a probability threshold, considering that a short-time strong precipitation event exists in a certain day when the station probability is greater than the threshold, comparing the short-time strong precipitation event with station precipitation data, and judging whether the station precipitation event is correct or incorrect;
4) calculating a TS forecast score: calculating a TS forecast score according to the judgment result of the step 3) to obtain forecast quality under different threshold settings (as shown in figure 21); the calculation formula of the TS forecast score is as follows:
wherein NA is the number of times of the short-time strong precipitation and the correct forecast, NB is the number of times of the empty forecast, NC is the number of times of the missed forecast, and ND is the number of times of the no short-time strong precipitation and the no forecast;
5) determining a reference probability threshold: selecting 0.98 as a reference probability threshold value of the urban short-time strong rainfall forecast according to the forecast quality of different threshold values obtained in the step 4).
Table 1 classification table of major physical parameters of short-term heavy rainfall in metropolis;
table 2 data table of main relevant physical quantities of short-time strong precipitation in metropolis;
TABLE 3 data sheet of main parameter threshold and weight coefficient for forecasting probability of strong rainfall in summer
Based on the above, the method has the advantages that based on the idea of a batching method, the method screens out the urban short-time strong rainfall potentiality prediction factor through correlation analysis and boxplot characteristic analysis, then establishes an urban short-time strong rainfall probability prediction equation through dichotomy, and utilizes TS prediction scoring to test the prediction effect, so that the optimal probability threshold value is determined to be 0.98, the number of times of missed reports can be ensured not to be too large, the number of times of correct prediction can not be obviously reduced, meanwhile, higher accuracy can be maintained, and more scientific bases are provided for urban short-time strong rainfall prediction services.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Claims (9)
1. A short-time strong precipitation probability forecasting method based on a batching method idea and a dichotomy comprises the following steps: step one, analyzing precipitation data; step two, constructing a forecasting model; step three, checking and forecasting effects; the method is characterized in that:
wherein in the first step, analyzing precipitation data comprises the following steps:
1) obtaining data: the precipitation data adopts hourly precipitation records of 13 meteorological stations 2010-2019 in 4-9 months in metropolis, and the specific observation time is 20:00 to 20:00 of the day before; in the potential forecast, physical quantity analysis adopts ERA-5 published by a European mesoscale weather forecast center to analyze data hour by hour, and the grid resolution is 0.25 degrees multiplied by 0.25 degrees;
2) and (3) analysis of data: analyzing the space-time characteristics of the short-time strong precipitation according to the data obtained in the step 1) to obtain the space-time distribution characteristics of the short-time strong precipitation of the city, namely that the short-time strong precipitation events of the city are mostly concentrated in warm seasons, particularly in 7 and 8 months; carrying out bilinear interpolation on the ERA-5 reanalysis data according to the positions of 13 meteorological stations to obtain physical quantity data of the corresponding stations;
in the second step, the construction of the forecasting model comprises the following steps:
1) and (3) correlation analysis: point-biserial correlation analysis is carried out on a dichotomized sequence of short-time strong precipitation days and an average physical quantity sequence of each station day, and common physical quantity elements in the research and forecast of the short-time strong precipitation are divided according to a water vapor condition, a thermodynamic stability condition, a dynamic stability condition and a lifting condition to obtain main physical quantity factors of the urban short-time strong precipitation, wherein each physical quantity factor is calculated by dividing the physical quantity factors into 500, 700 and 850hPa in different levels; performing correlation analysis on the four basic elements and the 1, 3 and 6 hour variable quantities thereof and the binary sequence of the short-time strong precipitation to obtain factors with the absolute value of the relation number greater than 0.3 in each element, thereby obtaining the main relevant physical quantity of the short-time strong precipitation in the metropolis;
2) screening a forecasting factor: screening main relevant physical quantities of the urban short-time heavy rainfall acquired in the step 1), and screening out physical quantity factors which are better in relevance to the short-time heavy rainfall;
3) establishing a short-time strong precipitation probability forecasting equation: calculating the short-time strong precipitation occurrence probability of each grid point by adopting a dichotomy;
4) calculating the weight coefficient of each parameter of the probability forecasting equation: based on the physical quantity factor box line graphs of the short-time strong precipitation day and the no short-time strong precipitation day, the upper quartile value and the lower quartile value when the short-time strong precipitation occurs can be set as a judgment threshold value of the parameter, the proportion of the crossed part of the two box bodies occupying the short-time strong precipitation box body is used for calculating a weight coefficient of the short-time strong precipitation box body, and a main parameter threshold value and the weight coefficient of the urban summer short-time strong precipitation probability forecast are obtained;
in the third step, the step of testing the forecasting effect comprises the following steps:
1) acquiring data to be checked: acquiring a forecast parameter value corresponding to each grid point in 2019 in 6-8 months by using ERA-5 data in 2019 of a metropolis city;
2) calculating a probability value: calculating the probability value of each grid point at each moment of the day by using the calculation method of the short-time strong precipitation occurrence probability of the single grid point mentioned in the step two 3), and interpolating the probability value to 13 sites;
3) judging and forecasting results: setting a probability threshold, considering that a short-time strong precipitation event exists in a certain day when the station probability is greater than the threshold, comparing the short-time strong precipitation event with station precipitation data, and judging whether the station precipitation event is correct or incorrect;
4) calculating a TS forecast score: calculating a TS forecast score according to the judgment result of the step 3) to obtain forecast quality under different threshold settings;
5) determining a reference probability threshold: selecting a proper threshold value as a reference probability threshold value of the urban short-time strong precipitation forecast according to the forecast quality of different threshold values obtained in the step 4).
2. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the step one 2), when the time-space characteristics of the short-time strong precipitation are analyzed, the precipitation amount of any one station of the 13 observation stations at a certain moment is more than or equal to 20 mm/h, or the precipitation amount is more than or equal to 20 mm/h for multiple times within 2 hours, and all the sites are regarded as the same short-time strong precipitation event.
3. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the second step 1), when point double-row correlation analysis is carried out, if one group of variables is natural binary variables, point double-row correlation coefficients need to be calculated; if the continuous variable is artificially divided into two variables, calculating a biserial correlation coefficient;
the calculation formula of the point double-row correlation coefficient is as follows:
wherein the content of the first and second substances,
and
respectively are the average values of continuous variable sequences corresponding to 0 and 1 in the binary sequence;
unbiased estimation of standard deviation for continuous variable sequence; p is a radical of0、p1Respectively the proportion of 0 and 1 in the binary sequence;
the calculation formula of the biserial correlation coefficient is as follows:
wherein h is the division p0And p1The ordinate value of (d) is given by the following two formulae:
4. the method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the step 2), the specific operations of screening the forecasting factors are as follows: dividing the short-time strong precipitation of 2010-2018 months into two sets according to existence and nonexistence, making various physical quantity factor box graphs of a short-time strong precipitation day and a non-short-time strong precipitation day, and selecting six physical quantity factors with good correlation with the short-time strong precipitation as the potentiality forecast equation independent variables according to the principle that the ratio of the overlapped part of each factor box body is small and comprehensively considering the water vapor condition, the thermodynamic stability condition, the dynamic stability condition and the lifting condition on the basis of the correlation analysis of various physical quantity factors and the short-time strong precipitation.
5. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the second step 2), the short-time strong precipitation potential forecasting factors are screened out to be 850hPa specific humidity, 850hPa pseudo equivalent temperature, K index, convection effective potential, 700hPa wind-passing direction and 700hPa vertical speed, and basically cover the water vapor condition, stability condition and lifting condition required by the generation of short-time strong precipitation.
6. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the second step 3), the calculation formula of the short-time strong precipitation occurrence probability of a single lattice point is as follows:
wherein N is the number of forecast parameters; when the value of the ith parameter falls within the threshold range, set AiIs 1, otherwise is 0; w is aiFor the weighting coefficients of parameter i, the sum of the weighting coefficients of all parameters is equal to 1.
7. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the second step 4), the weight coefficient calculation formula is as follows:
wherein QiThe i-th parameter represents the proportion (in%) of the intersection of the two boxes in the short-term heavy precipitation box.
8. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the step three 4), the calculation formula of the TS forecast score is as follows:
wherein NA is the number of times of the short-term strong precipitation and the correct forecast, NB is the number of times of the empty forecast, NC is the number of times of the missed forecast, and ND is the number of times of the no short-term strong precipitation and the no forecast.
9. The method for forecasting the probability of strong rainfall in short time based on the idea of the ingredient method and the dichotomy as claimed in claim 1, wherein: in the third step 5), the reference probability threshold of the short-time strong rainfall forecast is set to be 0.98, so that the higher forecast accuracy can be kept.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210084321.9A CN114609698A (en) | 2022-01-20 | 2022-01-20 | Short-time strong precipitation probability forecasting method based on proportioning method idea and dichotomy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210084321.9A CN114609698A (en) | 2022-01-20 | 2022-01-20 | Short-time strong precipitation probability forecasting method based on proportioning method idea and dichotomy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114609698A true CN114609698A (en) | 2022-06-10 |
Family
ID=81857516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210084321.9A Withdrawn CN114609698A (en) | 2022-01-20 | 2022-01-20 | Short-time strong precipitation probability forecasting method based on proportioning method idea and dichotomy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114609698A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115356789A (en) * | 2022-10-08 | 2022-11-18 | 南京气象科技创新研究院 | Plum rain period short-time strong precipitation grading early warning method |
| CN117290810A (en) * | 2023-11-27 | 2023-12-26 | 南京气象科技创新研究院 | Short-time strong precipitation probability prediction fusion method based on cyclic convolutional neural network |
-
2022
- 2022-01-20 CN CN202210084321.9A patent/CN114609698A/en not_active Withdrawn
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115356789A (en) * | 2022-10-08 | 2022-11-18 | 南京气象科技创新研究院 | Plum rain period short-time strong precipitation grading early warning method |
| CN117290810A (en) * | 2023-11-27 | 2023-12-26 | 南京气象科技创新研究院 | Short-time strong precipitation probability prediction fusion method based on cyclic convolutional neural network |
| CN117290810B (en) * | 2023-11-27 | 2024-02-02 | 南京气象科技创新研究院 | Short-time strong precipitation probability prediction fusion method based on cyclic convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112070286B (en) | Precipitation forecast and early warning system for complex terrain river basin | |
| Lloyd‐Hughes et al. | Seasonal prediction of European spring precipitation from El Niño–Southern Oscillation and local sea‐surface temperatures | |
| Wong et al. | Variability of rainfall in Peninsular Malaysia | |
| Cacciamani et al. | Monitoring and forecasting drought on a regional scale: Emilia-Romagna region | |
| CN114609698A (en) | Short-time strong precipitation probability forecasting method based on proportioning method idea and dichotomy | |
| Brugnara et al. | High‐resolution analysis of daily precipitation trends in the central Alps over the last century | |
| Merino et al. | Synoptic environment, mesoscale configurations and forecast parameters for hailstorms in Southwestern Europe | |
| KR20170005553A (en) | Floods, drought assessment and forecasting techniques development for intelligent service | |
| CN111639865B (en) | High-speed rail line meteorological disaster occurrence risk analysis method | |
| CN112686426A (en) | Incoming water quantity early warning method and system based on hydropower station basin key points | |
| Zhu et al. | Predictability and skill of convection-permitting ensemble forecast systems in predicting the record-breaking “21· 7” extreme rainfall event in Henan Province, China | |
| Kuleshov et al. | Improving preparedness to natural hazards: Tropical cyclone seasonal prediction for the Southern Hemisphere | |
| Ahmadi et al. | Spatial modeling of seasonal precipitation–elevation in Iran based on aphrodite database | |
| Mazzarella et al. | Reflectivity and velocity radar data assimilation for two flash flood events in central Italy: A comparison between 3D and 4D variational methods | |
| CN114912359A (en) | Water level dynamic prediction method based on XGboost machine learning model | |
| Javadinejad et al. | Climate change simulation and impacts on extreme events of rainfall and storm water in the Zayandeh Rud Catchment | |
| CN114444020A (en) | Air temperature forecast correction system for power grid tower point | |
| Jia et al. | An application of the LTP_DSEF model to heavy precipitation forecasts of landfalling tropical cyclones over China in 2018 | |
| CN111680886A (en) | Waterlogging risk prediction method and system | |
| Tadić et al. | Mapping of maximum snow load values for the 50-year return period for Croatia | |
| Tiriolo et al. | Verification of a real time weather forecasting system in southern Italy | |
| CN115236770B (en) | Nonlinear short-time adjacent precipitation prediction method based on space-time stacking and sample reconstruction | |
| Kyselý et al. | Return periods of the August 2010 heavy precipitation in northern Bohemia (Czech Republic) in the present climate and under climate change | |
| CN115511192A (en) | Rainfall forecasting method and system based on lightning data assimilation | |
| CN114996943A (en) | Mesoscale numerical simulation method for reservoir storage climate effect evaluation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20220610 |