Method for forecasting strong wind intensity grade of intelligent grid by using convection potential index

Abstract

The invention relates to the technical field of weather forecast, and particularly discloses a method for forecasting the strong wind intensity grade of an intelligent grid by using a convection potential index, which comprises the following steps of: firstly, acquiring a numerical mode product with a specified range and resolution, extracting five elements, namely a lifting index, 1000m vertical wind shear, a temperature difference between 500hPa and 850hPa, a multilayer vertical speed and convection effective potential energy, and calculating the wind intensity grade of the intelligent grid by using the convection potential index; the quartile value, the abnormal value and the like of each element are respectively calculated, and a standardized value is obtained by combining the correlation with the convection strong wind. The standardized values of all the elements are summed to obtain a convection gale potential index, then the gale intensity is calculated by fusing the maximum wind speed and the maximum wind speed, and after potential index correction is conducted, the grade of the gale intensity is determined by adopting a dual-threshold grading method in combination with rainfall elimination and control. According to the method, the instantaneous and continuous strength of the strong wind is considered, the convection potential effect is fused, strong wind forecasting skills of different time periods and the accurate forecasting advance of the local strong wind are improved, and the TS score of the forecasting result is obviously improved compared with an EC numerical mode on the whole, especially for high-grade strong wind forecasting.

Description

Technical Field

[0001] This invention relates to the field of meteorological forecasting technology, specifically to a method for forecasting the intensity level of strong winds using a smart grid based on the convection potential index. Background Technology

[0002] Severe convective winds are a common type of severe convective weather, characterized by their sudden onset, short duration, and high destructive power. They can pose a serious threat to transportation, agricultural production, power and communication facilities, and the safety of people's lives and property. Accurate forecasts of the intensity of convective winds can provide crucial support for disaster prevention and mitigation decisions, thereby reducing disaster losses.

[0003] Currently, convective wind forecasting largely relies on numerical model products and meteorological observation data, using statistical methods or empirical formulas for extrapolation. However, existing methods have several shortcomings: Firstly, the analysis of the impact of key meteorological factors on wind intensity is not comprehensive enough, neglecting the synergistic effects of multiple factors such as lifting index, vertical wind shear, temperature difference between upper and lower levels, vertical velocity, and convective available potential energy, thus limiting forecast accuracy. Secondly, traditional statistical forecasting methods seek patterns from historical data and assume that these patterns will continue to be valid in the future, often resulting in "conservative" forecasts. They lack a refined classification and consideration of the correlation between factors and convective winds, especially failing to adequately consider the impact of outliers, making it difficult to forecast "unusual" and extreme wind weather, and thus unable to meet the needs for accurate wind intensity forecasting and defense against highly destructive wind weather.

[0004] Furthermore, existing methods for correcting and classifying gale intensity are relatively simple, often employing a single threshold method without considering the dynamic impact of the convective potential index. They also lack effective mechanisms for controlling thunderstorm gale suppression, leading to false alarms and missed warnings. Therefore, there is an urgent need for a method for forecasting convective gale intensity levels that integrates multiple key meteorological factors based on intelligent grids, achieving refined classification and standardization, to overcome the shortcomings of existing technologies. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a method for intelligent grid-based wind intensity level forecasting using the convective potential index. By integrating multiple key convective meteorological factors, performing refined classification, standardization processing and weight allocation, a convective wind potential index is constructed. Combined with wind speed data correction and a dual threshold grading method, a high-resolution, high-precision intelligent grid-based convective wind intensity level forecast is achieved. At the same time, a precipitation control mechanism is introduced to reduce the false alarm rate of thunderstorms and strong winds.

[0006] To solve the above-mentioned technical problems, the technical solution provided by this invention is: a method for intelligent grid-based wind intensity level forecasting using the convection potential index, comprising the following steps:

[0007] S1, Optimal elevation index B of numerical model product x Obtain the latitude and longitude range and resolution, and calculate the 3rd quartile value B. Q3 1st quartile B Q1 and average value B aver IQR1, Positive deviation from outliers B PO Negative deviation outlier B NO The standardized value B of the smart grid lifting index is calculated based on its correlation with convective winds. Norm ;

[0008] S2, The numerical model product 1000m vertical wind shear VVS x Obtain the latitude and longitude range and resolution, and calculate its third quartile value (VVS). Q3 1st quartile VVS Q1 and average VVS aver IQR2, Positive Deviation from Outliers (VVS) PO Negative deviation outlier VVS NO The standardized value VVS of the vertical wind shear of the smart grid is calculated based on its correlation with convective winds. Norm ;

[0009] S3. Obtain the 500hPa and 850hPa temperatures from the numerical model product based on their latitude and longitude range and resolution, and calculate the temperature difference ΔT between the two levels for each grid point. x Calculate its third quartile value ΔT Q3 1st quartile value ΔT Q1 and average value ΔT aver IQR3, positive deviation from outlier ΔT PO Negative deviation outlier ΔT NO Based on its correlation with convective winds, the standardized value ΔT of the temperature difference between the upper and lower levels of the smart grid is calculated. Norm ;

[0010] S4. Obtain the vertical velocity (VVEL) of the numerical model product based on its latitude and longitude range and resolution, and take the sum of the four layers (500hPa, 600hPa, 700hPa, and 850hPa) for each grid point. x Calculate its third quartile value VVEL Q3 1st quartile value VVEL Q1 and average VVEL aver IQR4, Positive Deviation from Outliers VVEL PO Negative deviation from outlier VVEL NO Then, based on its correlation with convective winds, the standardized value VVEL of the vertical velocity of the smart grid is calculated. Norm ;

[0011] S5. The convective effective potential energy (CAPE) of the numerical model product x Obtain the latitude and longitude range and resolution, and calculate its third quartile value (CAPE). Q3 1st quartile CAPE Q1 and average CAPE aver IQR5, positive deviation from outliers CAPE PO Negative deviation from outlier CAPE NO The standardized value of the convective effective potential energy (CAPE) of the smart grid is calculated based on its correlation with convective winds. Norm ;

[0012] S6. Sum the standardized values ​​of the five elements—lift index, vertical velocity, upper-level temperature difference, vertical wind shear, and convective available potential energy—at each grid point to calculate the convective wind potential index P for that grid point:

[0013] P=B Norm +VVS Norm +ΔT Norm +VVEL Norm +CAPE Norm ;

[0014] S7. The average wind speed f of the numerical model product at 10 meters. a Maximum wind speed f mo Maximum wind speed of 10 meters f ma Obtain the latitude and longitude range and resolution; if the numerical model does not have a maximum wind speed product, but has a maximum wind speed or 10-meter average wind speed product, then select one of the following regression equations to calculate the maximum wind speed:

[0015] f ma =-0.8786+0.6634×f mo (R) 2 (Adjusted) = 94.5% or

[0016] f ma =0.6029294379+0.9548969974×f a +0.01003192183×f a 2 (R 2 =95.0%);

[0017] S8. Calculate the gale intensity value H using the maximum wind speed and extreme wind speed at that time. f The formula is as follows: H f =a8*f ma +b8*f moWhere a8 and b8 are weight values, obtained according to the grey relational analysis method, a8 is 0.53 and b8 is 0.47.

[0018] The wind intensity value H is then corrected using the convective wind potential index. f The wind intensity is obtained as: H = H f +H f *P;

[0019] S9. The dual threshold classification method is used to calculate the gale intensity level H. L Thunderstorm and strong wind suppression is carried out using precipitation R. If no precipitation occurs, it is assumed that the convective potential has not triggered the formation of strong winds.

[0020] Furthermore, data acquisition is performed before step S1, and the specific method is as follows:

[0021] Set the latitude and longitude range and resolution of the smart grid; the required upper-level layer for the numerical model products is between 200 hPa and 850 hPa, and the specific meteorological element layer selection is different. In terms of time selection, the mesoscale product is an hourly product, and the global model product is a 3-hour interval within 72 hours and a 6-hour interval within 72-120 hours.

[0022] Further, in step S1, the standardized value B of the smart grid lifting index... Norm The specific calculation method is as follows:

[0023] B Norm =-a1*b 11 *(B x -B aver ) / (B max -B min (B) x >(5.4 and B PO ))

[0024] B Norm =-a1*b 12 *(B x -B aver ) / (B max -B min (B) x ≤-1.1 and B x >(-8.8 and B NO ))

[0025] B Norm =-a1*b 13 *(B x -B aver ) / (B max -B min (B) x≤(-8.8 and B NO ))

[0026] B Norm =0(B x >=-1.1 and B x ≤(5.4 and B PO ));

[0027] Where, IQR1=B Q3 -B Q1 B PO =B Q3 +1.5×IQR1, B NO =B Q1 -1.5×IQR1; B aver The historical average is taken as -1.1, and a1 is the factor weighting coefficient; a1 is initially assigned an equal weight of 0.2, and then the empirical value of fluctuation is obtained based on the historical data, which is taken as 0.075 here; b 11 b 12 b 13 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 11 For a particularly negative correlation, take 2, b 12 For a typical positive correlation, take 1, b 13 For a particularly positive correlation, take 2.

[0028] Further, in step S2, the normalized value VVS of the vertical wind shear of the smart grid... Norm The specific calculation method is as follows:

[0029] VVS Norm =a2*b 21 *(VVS x -VVS aver ) / (VVS max -VVS min (VVS) x >(12.1 and VVS PO ))

[0030] VVS Norm =a2*b 22 *(VVS x -VVS aver ) / (VVS max -VVS min (VVS) x ≤(12.1 and VVS PO and VVS x >5.9)

[0031] VVSNorm=0 (VVSx≤5.9)

[0032] Where IQR2=VVS Q3 -VVS Q1 VVS PO =VVS Q3 +1.5×IQR2, VVS NO =VVS Q1 -1.5×IQR2, VVS aver a1 is the historical average, taken as 5.9; a2 is the factor weighting coefficient, initially assigned an equal weight of 0.2, then adjusted to an empirical value of 0.075 based on historical case statistics; b 21 b 22 , where b is the classification coefficient, representing the positive effect coefficient of convective winds. 21 For a particularly positive correlation, take 2, b 22 As a typical positive correlation, we take the value of 1.

[0033] Furthermore, in step S3, the standardized value ΔT of the temperature difference between the upper and lower layers of the smart grid is... Norm The specific calculation method is as follows:

[0034] ΔT Norm =a3*b 31 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x ≤(18 and ΔT NO ))

[0035] ΔT Norm =a3*b 32 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x >(30 and ΔT PO ))

[0036] ΔT Norm =a3*b 33 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x >24.7 and ΔT x ≤(30 and ΔT PO ))

[0037] ΔT Norm =0(ΔT x≤24.7 and ΔT x >(18 and ΔT NO ))

[0038] Where, IQR3=ΔT Q3 -ΔT Q1 ;ΔT PO =ΔT Q3 +1.5×IQR3, ΔT NO =ΔT Q1 -1.5×IQR3; ΔT x =T 850 -T 500 ΔT aver a3 is the historical average, taken as 24.7; a3 is the factor weighting coefficient, initially given an equal weight of 0.2, then adjusted to an empirical value of 0.25 based on historical case statistics; b 31 b 32 b 33 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 31 For a negative correlation, take 2, b 32 For a particularly positive correlation, take 3, b 33 As a typical positive correlation, we take the value of 1.

[0039] Further, in step S4, the normalized value VVEL of the vertical velocity of the smart grid... Norm The specific calculation method is as follows:

[0040] VVEL Norm =-a4*b 41 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min (VVEL) x >(26 and VVEL PO ))

[0041] VVEL Norm =-a4*b 42 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min (VVEL) x ≤(-655 and VVEL) NO ))

[0042] VVEL Norm =-a4*b 43 *(VVEL x -VVELaver ) / (VVEL max -VVEL min (VVEL) x >

[0043] (-655 and VVEL NO ) and VVEL x ≤-84.8))

[0044] VVEL Norm =0(VVEL x >-84.8andVVEL x ≤(26 and VVEL PO ))

[0045] Where IQR4=VVEL Q3 -VVEL Q1 VVEL PO =VVEL Q3 +1.5×IQR4, VVEL NO =VVEL Q1 -1.5×IQR4; VVEL aver The historical average is -84.8; a4 is the factor weighting coefficient, initially weighted equally at 0.2, then adjusted to 0.35 based on historical statistical data to determine the fluctuation range; b 41 b 42 b 43 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 41 For a particularly negative correlation, take 2, b 42 For a particularly positive correlation, take 2, b 43 As a typical positive correlation, we take the value of 1.

[0046] Further, in step S5, the normalized value of the convective effective potential energy (CAPE) of the smart grid is... Norm The specific calculation method is as follows:

[0047] CAPE Norm =a5*b 52 *(CAPE x -CAPE aver ) / (CAPE max -CAPE min (CAPE) x >(780 and CAPE PO ))

[0048] CAPE Norm =a5*b 53 *(CAPE x -CAPEaver ) / (CAPE max -CAPE min (CAPE) x >117 and CAPE x ≤(780and CAPE PO ))

[0049] CAPE Norm =0(CAPE x ≤117)

[0050] Where IQR5 = CAPE Q3 -CAPE Q1 CAPE PO =CAPE Q3 +1.5×IQR5, CAPE NO =CAPE Q1 -1.5×IQR5; CAPE aver The historical average is set to 117. a5 is the factor weighting coefficient, initially weighted equally at 0.2, then adjusted to 0.25 based on historical statistical data to determine fluctuations. 52 b 53 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 52 For a particularly positive correlation, take 3, b 53 As a typical positive correlation, we take the value of 1.

[0051] Further, in step S9, the gale intensity level H L The specific calculation method is as follows:

[0052] Based on historical cases, the percentiles c7, c9, and c of the initial thresholds for classifying gusts at levels 7, 9, 11, and 13 are used. 11 c 13 Find historical examples in c7, c9, and c 11 c 13 The percentile corrected wind intensity forecast values ​​are used as empirical thresholds d7, d9, and d for classifying wind intensity. 11 d 13 And at a certain time, c7, c9, c 11 c 13 The percentile wind intensity forecast values ​​are the percentile thresholds e7, e9, and e1. 11 e 13 The specific divisions are as follows:

[0053] H>(d 13 and e 13 ) and R>0.01, H L =Level 1, higher;

[0054] (d 13 and e 13 )≥H>(d 11 and e 11 ) and R>0.01, H L =Level 2, high;

[0055] (d 11 and e 11 ) ≥H>(d9and e9) and R>0.01, H L =Level 3, Medium;

[0056] (d9and e9) ≥H>(d7and e7) and R>0.01,H L =Level 4, low.

[0057] The advantages of this invention compared to existing technologies are as follows: When calculating wind intensity, this invention considers both the maximum wind speed and the sustained wind speed, taking into account both the instantaneous and sustained wind intensity. This provides a more comprehensive evaluation of wind intensity, overcoming the shortcomings of relying solely on the sustained wind speed. Furthermore, this invention considers the effect of convective potential, further improving the forecasting skills for wind intensities at different time points, especially for high-level winds, and increasing the lead time for accurate local wind forecasts. Attached Figure Description

[0058] Figure 1 This is a flowchart of the method for intelligent grid-based wind intensity level forecasting using the convection potential index, as per the present invention.

[0059] Figure 2 This is the 48-hour forecast result of the gale potential index and gale intensity level for the gale process on April 11, 20xx. The forecast of the potential index and intensity level is in good agreement with the actual gale level. The accuracy rates of the TS scores for level 4, level 3 and level 2 in 48 hours are 95.8%, 70.4% and 6.5% respectively (level 1 did not occur).

[0060] Figure 3 This invention relates to a method for forecasting wind intensity levels using a potential forecast index. The forecasts were conducted for six major convective wind events in 20xx, with different forecast lead times of 24h, 48h, and 72h. Compared to the skill score of the EC numerical model for wind intensity forecasting, the skill score of this method was generally positive. The 24h wind intensity products (levels 1, 2, and 4) were all higher than the EC numerical model products' skill scores, by 33.3%, 2.7%, and 31.3%, respectively. The 48h and 72h wind intensity products (levels 1, 2, 3, and 4) had skill scores higher than (only the 72h level 1 was equal to) the EC numerical model products' skill scores. Detailed Implementation

[0061] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present invention.

[0062] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0063] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0064] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0065] The following detailed description, in conjunction with the accompanying drawings, provides a further explanation of the method for intelligent grid-based wind intensity level forecasting using the convection potential index.

[0066] Combined with appendix Figure 1-3 The specific implementation process of the method for intelligent grid-based wind intensity level forecasting using the convection potential index in this invention is as follows:

[0067] A method for intelligent grid-based wind intensity forecasting using the convection potential index includes the following steps:

[0068] S0. Data Acquisition. The smart grid's latitude and longitude range is set to (108.25E, 116.25E), (28.90N, 33.40N) and resolution is 0.05°. The required upper-level altitude for numerical model products is between 200hPa and 850hPa. The specific altitude selection varies for different meteorological elements. For time intervals, mesoscale products are hourly products, while global model products use 3-hour intervals within 72 hours and 6-hour intervals within 72-120 hours. Different factor weight coefficients 'a' are determined using a factor importance ranking method. Factor classification uses positive outliers (PO) and negative outliers (NO) for classification. Classification weight coefficients 'b' are used. i The expert experience-based weighting method was adopted.

[0069] S1, Optimal elevation index B of numerical model product x (Unit: K) Obtain the latitude and longitude range and resolution, and calculate the 3rd quartile value B. Q31st quartile B Q1 and average value B aver IQR1, Positive deviation from outliers B PO Negative deviation outlier B NO The standardized value B of the smart grid lifting index is calculated based on its correlation with convective winds. Norm :

[0070] B Norm =-a1*b 11 *(B x -B aver ) / (B max -B min (B) x >(5.4 and B PO ))

[0071] B Norm =-a1*b 12 *(B x -B aver ) / (B max -B min (B) x ≤-1.1 and B x >(-8.8 and B NO ))

[0072] B Norm =-a1*b 13 *(B x -B aver ) / (B max -B min (B) x ≤(-8.8 and B NO ))

[0073] B Norm =0(B x >-1.1andB x ≤(5.4 and B PO ));

[0074] Where, IQR1=B Q3 -B Q1 B PO =B Q3 +1.5×IQR1, B NO =B Q1 -1.5×IQR1; B aver The historical average is taken as -1.1, and a1 is the factor weighting coefficient; a1 is initially assigned an equal weight of 0.2, and then the empirical value of fluctuation is obtained based on the historical data, which is taken as 0.075 here; b 11 b12 b 13 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 11 For a particularly negative correlation, take 2, b 12 For a typical positive correlation, take 1, b 13 For a particularly positive correlation, take 2.

[0075] S2, The numerical model product 1000m vertical wind shear VVS x (Unit: 1 / s) Obtain the latitude and longitude range and resolution, and calculate its 3rd quartile value VVS. Q3 1st quartile VVS Q1 and average VVS aver IQR2, Positive Deviation from Outliers (VVS) PO Negative deviation outlier VVS NO The standardized value VVS of the vertical wind shear of the smart grid is calculated based on its correlation with convective winds. Norm :

[0076] VVS Norm =a2*b 21 *(VVS x -VVS aver ) / (VVS max -VVS min (VVS) x >(12.1 and VVS PO ))

[0077] VVS Norm =a2*b 22 *(VVS x -VVS aver ) / (VVS max -VVS min (VVS) x ≤(12.1 and VVS PO and VVS x >5.9)

[0078] VVSNorm=0 (VVSx≤5.9)

[0079] Where IQR2=VVS Q3 -VVS Q1 VVS PO =VVS Q3 +1.5×IQR2, VVS NO =VVS Q1 -1.5×IQR2, VVS avera1 is the historical average, taken as 5.9; a2 is the factor weighting coefficient, initially assigned an equal weight of 0.2, then adjusted to an empirical value of 0.075 based on historical case statistics; b 21 b 22 , where b is the classification coefficient, representing the positive effect coefficient of convective winds. 21 For a particularly positive correlation, take 2, b 22 As a typical positive correlation, we take the value of 1.

[0080] S3. Obtain the 500hPa and 850hPa temperatures from the numerical model product based on their latitude and longitude range and resolution, and calculate the temperature difference ΔT between the two levels for each grid point. x (Unit: °C), calculate its 3rd quartile value ΔT Q3 1st quartile value ΔT Q1 and average value ΔT aver IQR3, positive deviation from outlier ΔT PO Negative deviation outlier ΔT NO Based on its correlation with convective winds, the standardized value ΔT of the temperature difference between the upper and lower levels of the smart grid is calculated. Norm :

[0081] ΔT Norm =a3*b 31 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x ≤(18 and ΔT NO ))

[0082] ΔT Norm =a3*b 32 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x >(30 and ΔT PO ))

[0083] ΔT Norm =a3*b 33 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x >24.7 and ΔT x ≤(30 and ΔT PO ))

[0084] ΔT Norm =0(ΔTx ≤24.7 and ΔT x >(18 and ΔT NO ))

[0085] Where, IQR3=ΔT Q3 -ΔT Q1 ;ΔT PO =ΔT Q3 +1.5×IQR3, ΔT NO =ΔT Q1 -1.5×IQR3; ΔT x =T 850 -T 500 ΔT aver a3 is the historical average, taken as 24.7; a3 is the factor weighting coefficient, initially given an equal weight of 0.2, then adjusted to an empirical value of 0.25 based on historical case statistics; b 31 b 32 b 33 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 31 For a negative correlation, take 2, b 32 For a particularly positive correlation, take 3, b 33 As a typical positive correlation, we take the value of 1.

[0086] S4. Measure the vertical velocity VVEL of the numerical model product (unit: 10e). -2 .Pa.s -1 The latitude and longitude range and resolution are obtained, and the sum of the four layers (500hPa, 600hPa, 700hPa, and 850hPa) for each grid point is taken as VVEL. x Calculate its third quartile value VVEL Q3 1st quartile value VVEL Q1 and average VVEL aver IQR4, Positive Deviation from Outliers VVEL PO Negative deviation from outlier VVEL NO Then, based on its correlation with convective winds, the standardized value VVEL of the vertical velocity of the smart grid is calculated. Norm :

[0087] VVEL Norm =-a4*b 41 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min (VVEL) x >(26 and VVEL PO ))

[0088] VVEL Norm =-a4*b 42 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min (VVEL) x ≤(-655 and VVEL) NO ))

[0089] VVEL Norm =-a4*b 43 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min (VVEL) x >

[0090] (-655 and VVEL NO ) and VVEL x ≤-84.8))

[0091] VVEL Norm =0(VVEL x >-84.8 and VVEL x ≤(26 and VVEL PO ))

[0092] Where IQR4=VVEL Q3 -VVEL Q1 VVEL PO =VVEL Q3 +1.5×IQR4, VVEL NO =VVEL Q1 -1.5×IQR4; VVEL aver The historical average is -84.8; a4 is the factor weighting coefficient, initially weighted equally at 0.2, then adjusted to 0.35 based on historical statistical data to determine the fluctuation range; b 41 b 42 b 43 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 41 For a particularly negative correlation, take 2, b 42 For a particularly positive correlation, take 2, b 43 As a typical positive correlation, we take the value of 1.

[0093] S5. The convective effective potential energy (CAPE) of the numerical model product x(Unit: J / kg) Obtain the latitude and longitude range and resolution, and calculate its 3rd quartile value CAPE. Q3 1st quartile CAPE Q1 and average CAPE aver IQR5, positive deviation from outliers CAPE PO Negative deviation from outlier CAPE NO The standardized value of the convective effective potential energy (CAPE) of the smart grid is calculated based on its correlation with convective winds. Norm :

[0094] CAPE Norm = a5 * b 51 *(CAPE x -CAPE aver ) / (CAPE max -CAPE min (CAPE) x ≤0) (due to CAPE) x In reality, the value will not be less than 0, so this situation will not occur and the formula can be ignored.

[0095] CAPE Norm =a5*b 52 *(CAPE x -CAPE aver ) / (CAPE max -CAPE min (CAPE) x >(780 and CAPE PO ))

[0096] CAPE Norm =a5*b 53 *(CAPE x -CAPE aver ) / (CAPE max -CAPE min (CAPE) x >117 and CAPE x ≤(780and CAPE PO ))

[0097] CAPE Norm =0(CAPE x ≤117)

[0098] Where IQR5 = CAPE Q3 -CAPE Q1 CAPE PO =CAPE Q3 +1.5×IQR5, CAPE NO=CAPE Q1 -1.5×IQR5; CAPE aver The historical average is set to 117. a5 is the factor weighting coefficient, initially weighted equally at 0.2, then adjusted to 0.25 based on historical statistical data to determine fluctuations. 51 b 52 b 53 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 51 For a particularly negative correlation, we take 2, b. 52 For a particularly positive correlation, take 3, b 53 As a typical positive correlation, we take the value of 1.

[0099] S6. Sum the standardized values ​​of the five elements—lift index, vertical velocity, upper-level temperature difference, vertical wind shear, and convective available potential energy—at each grid point to calculate the convective wind potential index P for that grid point:

[0100] P=B Norm +VVS Norm +ΔT Norm +VVEL Norm +CAPE Norm ;

[0101] S7. The average wind speed f of the numerical model product at 10 meters. a Maximum wind speed f mo Maximum wind speed of 10 meters f ma Obtain the latitude and longitude range and resolution; if the numerical model does not have a maximum wind speed product, but has a maximum wind speed or 10-meter average wind speed product, then select one of the following regression equations to calculate the maximum wind speed:

[0102] f ma =-0.8786+0.6634×f mo (R) 2 (Adjusted) = 94.5%

[0103] or f ma =0.6029294379+0.9548969974×f a +0.01003192183×f a 2 (R 2 =95.0%);

[0104] S8. Calculate the gale intensity value H using the maximum wind speed and extreme wind speed at that time. f The formula is as follows: H f =a8*f ma +b8*f moWhere a8 and b8 are weight values, obtained according to the grey relational analysis method, a8 is 0.53 and b8 is 0.47.

[0105] The wind intensity value H is then corrected using the convective wind potential index. f The wind intensity is obtained as: H = H f +H f *P;

[0106] S9. The dual threshold classification method is used to calculate the gale intensity level H. L Based on the percentiles c7, c9, and c8 of the initial thresholds for gust classification at levels 7, 9, 11, and 13 in historical cases. 11 c 13 Find historical examples in c7, c9, and c 11 c 13 The percentile corrected wind intensity forecast values ​​are used as empirical thresholds d7, d9, and d for classifying wind intensity. 11 d 13 And at a certain time, c7, c9, c 11 c 13 The percentile wind intensity forecast values ​​are the percentile thresholds e7, e9, and e1. 11 e 13 Thunderstorm gale control is based on precipitation R (mm). If no precipitation occurs, it is assumed that the convective potential has not triggered gale generation.

[0107] The specific classification of gale intensity levels is as follows:

[0108] H>(d 13 and e 13 ) and R>0.01, H L =Level 1, higher;

[0109] (d 13 and e 13 )≥H>(d 11 and e 11 ) and R>0.01, H L =Level 2, high;

[0110] (d 11 and e 11 ) ≥H>(d9and e9) and R>0.01, H L =Level 3, Medium;

[0111] (d9and e9) ≥H>(d7and e7) and R>0.01,H L =Level 4, low.

[0112] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for intelligent grid-based wind intensity level forecasting using the convection potential index, characterized in that, Includes the following steps: S1, Optimal elevation index B of numerical model product x Obtain the latitude and longitude range and resolution, and calculate the 3rd quartile value B. Q3 1st quartile B Q1 and average value B aver IQR1, Positive deviation from outliers B PO Negative deviation outlier B NO The standardized value B of the smart grid lifting index is calculated based on its correlation with convective winds. Norm ; S2, The numerical model product 1000m vertical wind shear VVS x Obtain the latitude and longitude range and resolution, and calculate its third quartile value (VVS). Q3 1st quartile VVS Q1 and average VVS aver IQR2, Positive Deviation from Outliers (VVS) PO Negative deviation outlier VVS NO The standardized value VVS of the vertical wind shear of the smart grid is calculated based on its correlation with convective winds. Norm ; S3. Obtain the 500hPa and 850hPa temperatures from the numerical model product based on their latitude and longitude range and resolution, and calculate the temperature difference ΔT between the two levels for each grid point. x Calculate its third quartile value ΔT Q3 1st quartile value ΔT Q1 and average value ΔT aver IQR3, positive deviation from outlier ΔT PO Negative deviation outlier ΔT NO Based on its correlation with convective winds, the standardized value ΔT of the temperature difference between the upper and lower levels of the smart grid is calculated. Norm ; S4. Obtain the vertical velocity (VVEL) of the numerical model product based on its latitude and longitude range and resolution, and take the sum of the four layers (500hPa, 600hPa, 700hPa, and 850hPa) for each grid point. x Calculate its third quartile value VVEL Q3 1st quartile VVEL Q1 and average VVEL aver IQR4, Positive Deviation from Outliers VVEL PO Negative deviation from outlier VVEL NO Then, based on its correlation with convective winds, the standardized value VVEL of the vertical velocity of the smart grid is calculated. Norm ; S5. The convective effective potential energy (CAPE) of the numerical model product x Obtain the latitude and longitude range and resolution, and calculate its third quartile value (CAPE). Q3 1st quartile CAPE Q1 and average CAPE aver IQR5, positive deviation from outliers CAPE PO Negative deviation from outlier CAPE NO The standardized value of the convective effective potential energy (CAPE) of the smart grid is calculated based on its correlation with convective winds. Norm ; S6. Sum the standardized values ​​of the five elements—lift index, vertical velocity, upper-level temperature difference, vertical wind shear, and convective available potential energy—at each grid point to calculate the convective wind potential index P for that grid point: P=B Norm +VVS Norm +ΔT Norm +VVEL Norm +CAPE Norm ; S7. The average wind speed f of the numerical model product at 10 meters. a Maximum wind speed f mo Maximum wind speed of 10 meters f ma Obtain the latitude and longitude range and resolution; if the numerical model does not have a maximum wind speed product, but has a maximum wind speed or 10-meter average wind speed product, then select one of the following regression equations to calculate the maximum wind speed: f ma =-0.8786+0.6634×f mo (R) 2 (Adjusted) = 94.5% or f ma =0.6029294379+0.9548969974×f a +0.01003192183×f a 2 (R 2 =95.0%); S8. Calculate the gale intensity value H using the maximum wind speed and extreme wind speed at that time. f The formula is as follows: H f =a8*f ma +b8*f mo ; Where a8 and b8 are weight values, which are obtained according to the grey relational analysis method, with a8 being 0.53 and b8 being 0.

47. The wind intensity value H is then corrected using the convective wind potential index. f The wind intensity is obtained as: H = H f +H f *P; S9. The dual threshold classification method is used to calculate the gale intensity level H. L Thunderstorm and strong wind suppression is carried out using precipitation R. If no precipitation occurs, it is assumed that the convective potential has not triggered the formation of strong winds.

2. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 1, characterized in that: Data acquisition is performed before step S1, and the specific method is as follows: Set the latitude and longitude range and resolution of the smart grid; the required upper-level layer for the numerical model products is between 200 hPa and 850 hPa, and the specific meteorological element layer selection is different. In terms of time selection, the mesoscale product is an hourly product, and the global model product is a 3-hour interval within 72 hours and a 6-hour interval within 72-120 hours.

3. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 2, characterized in that: In step S1, the standardized value B of the smart grid lifting index... Norm The specific calculation method is as follows: B Norm =-a1*b 11 *(B x -B aver ) / (B max -B min )(B x >(5.4 and B PO )) B Norm =-a1*b 12 *(B x -B aver ) / (B max -B min )(B x ≤-1.1 and B x >(-8.8 and B NO )) B Norm =-a1*b 13 *(B x -B aver ) / (B max -B min )(B x ≤(-8.8 and B NO )) B Norm =0(B x >-1.1 and B x ≤(5.4 and B PO )); Where, IQR1=B Q3 -B Q1 B PO =B Q3 +1.5×IQR1, B NO =B Q1 -1.5×IQR1; B aver The historical average is taken as -1.1, and a1 is the factor weighting coefficient; a1 is initially assigned an equal weight of 0.2, and then the empirical value of fluctuation is obtained based on the historical data, which is taken as 0.075 here; b 11 b 12 b 13 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 11 For a particularly negative correlation, take 2, b 12 For a typical positive correlation, take 1, b 13 For a particularly positive correlation, take 2.

4. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 3, characterized in that: In step S2, the normalized value VVS of the vertical wind shear of the smart grid is... Norm The specific calculation method is as follows: Plumbing Norm =a2*b 21 *(Plumbing x -Plumbing aver ) / (PLUMBING max -Plumbing min )(Plumbing x >(12.1 and HVAC PO )) Plumbing Norm =a2*b 22 *(Plumbing x -Plumbing aver ) / (PLUMBING max -Plumbing min )(Plumbing x ≤(12.1 and HVAC PO ) and plumbing x >5.9) VVSNorm=0 (VVSx≤5.9) Where IQR2=VVS Q3 -VVS Q1 VVS PO =VVS Q3 +1.5×IQR2, VVS NO =VVS Q1 -1.5×IQR2, VVS aver a1 is the historical average, taken as 5.9; a2 is the factor weighting coefficient, initially assigned an equal weight of 0.2, then adjusted to an empirical value of 0.075 based on historical case statistics; b 21 b 22 , where b is the classification coefficient, representing the positive effect coefficient of convective winds. 21 For a particularly positive correlation, take 2, b 22 As a typical positive correlation, we take the value of 1.

5. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 4, characterized in that: In step S3, the standardized value ΔT of the temperature difference between the upper and lower layers of the smart grid is... Norm The specific calculation method is as follows: ΔT Norm =a3*b 31 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x ≤(18 and ΔT NO )) ΔT Norm =a3*b 32 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x >(30 and ΔT PO )) ΔT Norm =a3*b 33 *(ΔT x -ΔT aver ) / (ΔT max -ΔT min )(ΔT x >24.7 and ΔT x ≤(30 and ΔT PO )) ΔT Norm =0(ΔT x ≤24.7 and ΔT x > (18 and ΔT NO )) Where, IQR3=ΔT Q3 -ΔT Q1 ;ΔT PO =ΔT Q3 +1.5×IQR3, ΔT NO =ΔT Q1 -1.5×IQR3; ΔT x =T 850 -T 500 ΔT aver a3 is the historical average, taken as 24.7; a3 is the factor weighting coefficient, initially given an equal weight of 0.2, then adjusted to an empirical value of 0.25 based on historical case statistics; b 31 b 32 b 33 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 31 For a negative correlation, take 2, b 32 For a particularly positive correlation, take 3, b 33 As a typical positive correlation, we take the value of 1.

6. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 5, characterized in that: In step S4, the normalized value VVEL of the vertical velocity of the smart mesh is... Norm The specific calculation method is as follows: VVEL Norm =-a4*b 41 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min )(VVEL x >(26 and VVEL PO )) VVEL Norm =-a4*b 42 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min )(VVEL x ≤(-655 and VVEL NO )) VVEL Norm =-a4*b 43 *(VVEL x -VVEL aver ) / (VVEL max -VVEL min )(VVEL x > (-655 and VVEL NO ) and VVEL x ≤-84.8)) VVEL Norm =0(VVEL x >-84.8 and VVEL x ≤(26 and VVEL PO )) Where IQR4=VVEL Q3 -VVEL Q1 VVEL PO =VVEL Q3 +1.5×IQR4, VVEL NO =VVEL Q1 -1.5×IQR4; VVEL aver The historical average is -84.8; a4 is the factor weighting coefficient, initially weighted equally at 0.2, then adjusted to 0.35 based on historical statistical data to determine the fluctuation range; b 41 b 42 b 43 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 41 For a particularly negative correlation, take 2, b 42 For a particularly positive correlation, take 2, b 43 As a typical positive correlation, we take the value of 1.

7. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 6, characterized in that: In step S5, the normalized value of the convective effective potential energy (CAPE) of the smart grid is... Norm The specific calculation method is as follows: CAPE Norm =a5*b 52 *(CAPE x -CAPE aver ) / (CAPE max -CAPE min )(CAPE x >(780 and CAPE PO )) CAPE Norm =a5*b 53 *(CAPE x -CAPE aver ) / (CAPE max -CAPE min )(CAPE x >117andCAPE x ≤(780 andCAPE PO )) CAPE Norm =0(CAPE x ≤117) Where IQR5 = CAPE Q3 -CAPE Q1 CAPE PO =CAPE Q3 +1.5×IQR5, CAPE NO =CAPE Q1 -1.5×IQR5; CAPE aver The historical average is set to 117. a5 is the factor weighting coefficient, initially assigned an equal weight of 0.2, then adjusted to an empirical value of 0.25 based on historical case statistics to account for fluctuations. 52 b 53 The classification coefficient, representing the positive / negative interaction coefficient with convective winds, b 52 For a particularly positive correlation, take 3, b 53 As a typical positive correlation, we take the value of 1.

8. The method for intelligent grid-based wind intensity level forecasting using the convection potential index according to claim 7, characterized in that: In step S9, the gale intensity level H L The specific calculation method is as follows: Based on historical cases, the percentiles c7, c9, and c of the initial thresholds for classifying gusts at levels 7, 9, 11, and 13 are used. 11 c 13 Find historical examples in c7, c9, and c 11 c 13 The percentile corrected wind intensity forecast values ​​are used as empirical thresholds d7, d9, and d for classifying wind intensity. 11 d 13 And at a certain time, c7, c9, c 11 c 13 The percentile wind intensity forecast values ​​are the percentile thresholds e7, e9, and e1. 11 e 13 The specific divisions are as follows: H>(d 13 and e 13 ) and R>0.01, H L = Level 1, higher; (d 13 and e 13 )≥H>(d 11 and e 11 ) and R>0.01,H L =Level 2, high; (d 11 and e 11 ) ≥ H>(d9 and e9) and R>0.01, H L = Level 3, medium; (d9 and e9) ≥H>(d7and e7) and R>0.01,H L = Level 4, low.

Images