JP4480630B2 - Lightning position prediction method and lightning position prediction system - Google Patents

Description

  The present invention relates to a lightning position prediction method and a lightning position prediction system that estimate a lightning position by predicting cloud movement.

  Lightning strikes are a type of disaster. When a disaster strikes, not only human lives, but also electronic devices, power transmission lines, communication lines, and the like are sometimes seriously damaged. For this reason, lightning protection technology has been studied by various organizations, but at the same time, there is a high demand for lightning prediction technology from the viewpoint of measures immediately before lightning strikes and speeding up of dissemination work. In particular, from the viewpoint of maintenance and operation of expensive electronic devices, there is a demand for highly accurate information about 30 minutes before a lightning strike, and more recently 10 minutes before a lightning strike.

  In response to such demands, conventionally, lightning prediction has been mainly performed by three methods.

(1) As the first method, the stability based on the temperature difference between the upper and lower layers called the Schwarter stability index (SSI) and the numerical prediction of the lightning strike probability based on the evaluation of the effective convective energy are non-patent literature. Reported as 1.

  This is a commonly used technique, and when numerical weather information is used as input data, typically it can be predicted up to about 40 hours ahead.

(2) As a second method, without using numerical weather information, real-time observations of lightning strikes are used as input data, the amount of movement is estimated from time series data, and a short-term prediction of about 10 to 60 minutes is performed. A technique to be performed is reported as Non-Patent Document 2.

  Since this is to see the movement of thunderclouds during the occurrence of lightning, once the lightning has started, the subsequent movement can be predicted.

(3) As a third method, a method using a rule of thumb or statistical processing has been reported.

When thunderclouds are predicted at a specific location, the next is to apply an empirical algorithm where lightning is likely to occur. This technique may be used in the vicinity of mountains where specific wind conditions are prominent and lightning strikes frequently.
Sugawara, “Development Research on Next-Day Lightning Prediction Method”, R & D NEWS KANSAI, 1999.10. Ohara et al., “Lightning Prediction Method Verification Experiment”, Chugoku Electric Power Engineering Co., Ltd. Time Report No. 91 1998 http: //www.golf-gtpa.orjp/xreport/pdf/kaminari_2.pdf http://csdl.computer.org/comp/proceedings/wacv/2002/1858/00/18580296abs.htm

  However, the conventional lightning prediction method has the following problems.

(1) In the first method, many observation points relating to the atmospheric temperature are installed at intervals of 20 to 40 km in the current state (2005). On the other hand, since the lightning generation radius is from 500 m to a maximum of about 5 km, when lightning prediction is performed from the atmospheric temperature obtained at observation points installed at intervals of 20 to 40 km, the hit rate, miss rate, miss rate, Alternatively, when evaluation is performed using an index used in the meteorological field such as a threat score, there is a problem that prediction with high accuracy cannot be performed. There is also a method of adding statistical analysis processing to improve the accuracy.

  In addition, cumulonimbus clouds develop rapidly in about 1 hour and may generate lightning, but the numerical weather information of RSM (Regional Scale Model) and MSM (Mesoscale Model) has a prediction time interval of 1 hour. , The time interval may not be in line with your prediction needs.

(2) In the second method, there is an error factor that cannot always accurately grasp the position of the lightning, and it is difficult to predict the position of the lightning in advance immediately before the lightning. there were.

(3) In the third method, the empirical algorithm has a drawback that it is difficult to apply only to a specific area.

  The present invention has been made in view of the above, and as its purpose, a lightning position prediction method capable of predicting a lightning strike position or a lightning strike time by calculating a cloud growth rate before the lightning strike. And providing a lightning position prediction system.

  In order to solve the above-mentioned problem, the invention according to claim 1 is a lightning position prediction method for predicting a lightning position by analyzing information on weather regularly distributed from a server provided in a weather information distribution center. And receiving the cloud mesh data K (t) and cloud top altitude mesh data U (t) distributed from the server, and cloud mesh data obtained up to time t every time the observation time interval Δt elapses. Based on the history of K (t), a step of calculating a cloud motion vector between time (t−Δt) and time t for each mesh, and the cloud motion vector is calculated as cloud top height mesh data U (t−Δt ) And calculating the estimated value Uex (t) of the cloud top height mesh data obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the cloud top height, and the cloud top height mesh data U ( t) From cloud top height mesh The result of dividing the difference value by subtracting the over data estimates Uex (t) at time intervals Δt and summarized in that comprises a step of calculating a cloud growth rate mesh data G (t), the.

  In order to solve the above problem, the invention according to claim 2 extracts a mesh cell in which the cloud growth rate mesh data G (t) has a value equal to or greater than a predetermined growth rate threshold value α, and a value of the growth rate. A step of calculating an estimated value Gex (t) of the growth rate mesh data G (t−Δt) obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the growth rate mesh data G (t− (T) projecting the motion vector to calculate the expected growth rate value Gex (t), and a mesh cell in which the expected growth rate value Gex (t) is equal to or greater than a growth rate threshold value α indicating cloud growth. And a step of extracting.

  In order to solve the above-described problem, the cloud top height mesh data U (t) is a mesh cell in which the cloud top height mesh data U (t) is a value equal to or greater than the initial growth height threshold β indicating the initial height of cloud growth. The present invention includes a step of extracting a mesh cell having a value equal to or higher than the developmental altitude threshold γ indicating the altitude of the developmental period.

  In order to solve the above problem, the invention according to claim 4 is a mesh cell in which the cloud top height mesh data U (t) has a value larger than the initial growth height threshold β or the developmental height threshold γ, and Extracting a mesh cell whose cloud growth rate mesh data G (t) has a value greater than or equal to the growth rate threshold value α, and recognizing the position of the mesh cell as a position at which there is a risk of lightning. The gist.

  In order to solve the above-mentioned problem, the invention according to claim 5 is obtained on the assumption that the cloud has moved after time xΔt while maintaining the cloud top altitude as a result value xΔt obtained by multiplying the natural number x by the time interval Δt. A step of calculating a future cloud top height predicted value UF (t + xΔt) and interpolating in time series from the history data of each time interval Δt of the growth rate mesh data U (t), so that the time of each future time interval Δt is excluded. A step of correcting the position of the extrapolated result based on the predicted value of the motion vector of the cloud and calculating a predicted growth rate GF (t + xΔt) at time (t + xΔt), and a predicted cloud top height UF (t + xΔt) Calculating the cloud top height {F (t + xΔt) + ΣΔt · GF (t + xΔt)} after the time xΔt from the predicted growth rate GF (t + xΔt), and the growth rate threshold value α for the predicted growth rate GF (t + xΔt) More than A step of extracting a mesh cell, high cloud top after time xΔt is, to the step of extracting the mesh cell to be a value larger than the initial growth altitude threshold β or developmental altitude threshold gamma, summarized in that with.

  In order to solve the above-mentioned problem, the invention according to claim 6 provides the cloud top height mesh data U (t), cloud growth speed mesh data G (t), cloud top height predicted value UF (t), growth speed predicted value GF. When calculating each (t), the gist is to have a step of performing an averaging process in the vicinity of each mesh.

  In order to solve the above-mentioned problem, the invention according to claim 7, when a cloud cell whose cloud top height after the time xΔt is larger than the initial growth height threshold β or the developmental height threshold γ is extracted. And a step of displaying a message to the effect that lightning is likely to occur at the position of the mesh cell after time xΔt.

  In order to solve the above-mentioned problems, the invention according to claim 8 receives and analyzes information on weather regularly distributed from a server provided in the weather information distribution center by a lightning position prediction device, and determines the position of lightning. A lightning position prediction system for prediction, wherein the lightning position prediction device receives cloud mesh data K (t) and cloud top height mesh data U (t) distributed from the server, and observation time interval Δt. Means for calculating a cloud motion vector between time (t−Δt) and time t for each mesh based on the history of cloud mesh data K (t) obtained until time t every time elapses, This cloud motion vector is projected onto the cloud top height mesh data U (t−Δt), and the estimated value Uex of the cloud top height mesh data obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the cloud top height. Hand to calculate (t) The difference obtained by subtracting the estimated value Uex (t) of the cloud top altitude mesh data from the stage top and the cloud top altitude mesh data U (t) and dividing the result by the time interval Δt is calculated as the cloud growth rate mesh data G (t). And a means.

  According to the first or eighth aspect of the invention, the cloud mesh data K (t) and the cloud top altitude mesh data U (t) distributed from a server provided in the weather information distribution center are received, and the observation time interval Δt The cloud motion vector between time (t−Δt) and time t is calculated for each mesh based on the history of cloud mesh data K (t) obtained up to time t every time elapses. Is projected onto the cloud top height mesh data U (t−Δt), and the estimated value Uex (t of the cloud top height mesh data obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the cloud top height. ) And the difference obtained by subtracting the estimated value Uex (t) of the cloud top height mesh data U (t) from the cloud top height mesh data U (t) is divided by the time interval Δt as the cloud growth rate mesh data G (t). By calculating, clouds are formed before the lightning strike. Can be calculated the speed, it is possible to predict the lightning strike position or lightning time.

  The best mode for carrying out the lightning position prediction system of the present invention will be described below with reference to the drawings.

[algorithm]
First, a prediction algorithm employed in the lightning position prediction system of the present invention will be described. To summarize the present invention, a cumulonimbus growth process is evaluated on the basis of time-series cloud top height mesh data U (t), and a position where it continuously grows up to a certain altitude is predicted and specified. It provides a method for predicting the position of lightning strike with high accuracy even in the previous stage.

  The cloud top altitude mesh data U (t) indicates the height of the cloud, which is high when a cumulonimbus that causes thermal thunder occurs, but sometimes has a similar value depending on the high clouds. . For this reason, it only satisfies the requirements, and it is difficult to evaluate as a cumulonimbus cloud only with the numerical value of the cloud top height mesh data. Moreover, it is not sufficient as a method to perform lightning prediction by such evaluation.

  In cumulonimbus clouds, lightning is not generated due to the high altitude of the clouds, but is generated by friction due to the rising airflow that occurs inside the clouds and causing charge separation.

  Therefore, the present invention calculates the cloud growth speed from the time-series cloud top height mesh data U (t), and determines that the lightning strike probability is high when the speed is equal to or greater than a certain growth speed threshold value α. To do. The growth rate threshold value α may be based on the normal growth rate of cumulonimbus clouds as an index. If the growth rate threshold value α is a typical value of about 10 m / s or more, the predicted swing rate is low. However, the growth rate in the vertical direction may gradually decrease with the growth of cumulonimbus clouds, and not always accompanied by a fast ascending air flow of about 10 m / s or more, so the miss rate increases reciprocally. . Considering that lightning prediction is originally performed for disaster countermeasures, it is more important to reduce the miss rate, so in reality, the growth rate threshold value α is preferably about 2 m / s to 5 m / s. .

  However, the clouds advect from time to time by wind, and in some cases, they cause diffusion, divergence, convergence, etc., so the difference between the time series cloud top height mesh data U (t) is simply calculated without correcting the movement of the clouds. Therefore, the cloud growth rate cannot be fully evaluated.

  For example, as shown in FIGS. 1A and 1B, when the cloud C1 grows up to the cloud C2 while advancing, as shown in FIG. 1D, simply U (t) −U (t−Δt. ) Divided by the time interval Δt, that is, U (t) −U (t−Δt) / Δt, a point A having a high growth rate occurs, but a negative portion also occurs in B.

  Also, as shown in FIGS. 1B and 1C, when the cloud C2 has developed to the cloud C3, as shown in FIG. 1E, U (t + Δt) −U (t)) is simply expressed as time. At a value divided by the interval Δt, that is, U (t + Δt) −U (t)) / Δt, a point with a high growth point appears at another position C, and the cloud C1 at point A further grows and the point C It is difficult to determine the relationship that the cloud C3 has been reached.

  Furthermore, as shown in FIGS. 2A, 2B, and 2C, when the clouds C1, C2, and C3 having no change in size are horizontally moved during the time interval of 2Δt, FIG. ) U (t) −U (t−Δt) / Δt shown in FIG. 2 and U (t + Δt) −U (t)) / Δt shown in FIG. In the difference, it is difficult to determine whether or not the cloud is really growing.

  Therefore, in the present invention, separately, as shown in FIGS. 3A and 3B, the movement of the cloud from time (t−Δt) to time t is evaluated by momentum analysis, and this is calculated using the cloud top height mesh data U ( (t−Δt), as shown in FIG. 3C, a predicted value Uex that becomes an expected value of the cloud top altitude at time t when it is assumed that the cloud moves while maintaining the cloud top altitude. By obtaining (t) and calculating the change between it and the cloud top altitude mesh data U (t) at the observation site, it is possible to obtain the growth rate data G (t) as shown in FIG. Features.

That is, the growth rate data G (t) is calculated from the cloud top height mesh data U (t) and the cloud top height predicted value Uex (t).
G (t) = (U (t) −Uex (t)) / Δt (1)
And

  With such a method, it is possible to evaluate the growth rate of the cloud after correcting the movement of the cloud without producing an apparent negative portion as seen in B shown in FIG.

  Note that a known method may be used as the cloud momentum analysis process, but in particular, the dynamics texture method (DT) is a kind of image processing technique that can analyze diffusion, divergence, convergence, etc. along with the movement of the center of gravity. (For the DT method, see “Sakaino et al., Shingaku Techniques, IE2002-8, pp.43-48, 2002.”) is also used for prediction in the near future based on past time-series data. Since it is possible, it is also useful in the present invention.

  Since the growth rate data G (t) is evaluated by the method of the present invention, the vertical growth rate can be accurately evaluated for the first time even when the cumulonimbus grows while moving.

  However, even if only the growth rate data G (t) is evaluated, the cumulonimbus growth can be determined, but its scale cannot be determined. Assume that the advanced mesh data U (t) is evaluated.

  For example, when a cumulonimbus generates lightning, the cloud often grows from an altitude of about 3000 m to about 10,000 m. Since cumulonimbus clouds that have grown to 10,000 m generate lightning with considerably high accuracy, the cloud top height mesh data U (t) is evaluated by setting an index up to this height. On the other hand, as the lower numerical value, the height of a normal cloud is as high as around 1000 m. Therefore, about 3000 m is suitable as an index for recognizing the beginning of a cumulonimbus cloud, and it is recognized as a further developed cloud. Is preferably set with a threshold of about 5000 m.

  For this reason, in the present invention, the cloud top height mesh data U (t) is evaluated by providing a growth initial height threshold value β indicating the height of the initial growth of the cloud. In the case of this threshold value β, a value of around 5000 m is appropriate, but it is more preferable to set a plurality of threshold values to evaluate the cloud top height mesh data U (t), typically It is preferable to provide a developmental altitude threshold γ indicating the altitude of the cloud development period in addition to the initial growth altitude threshold β, and evaluate by evaluating the lightning accuracy in stages.

  As shown in FIG. 3D, a location where the mesh cell extracted by the evaluation of the cloud top height mesh data U (t) and the mesh cell extracted by the evaluation of the growth rate data G (t) are overlapped with each other. D can be recognized as a portion having a high probability of lightning.

  Although the above-described method can be regarded as a real-time state analysis at the observed time t, it is also possible to perform a future lightning prediction by performing a cloud momentum analysis process.

  That is, the motion vector obtained as a result of the momentum analysis process is projected onto the cloud top height mesh data U (t) and the growth rate data G (t), and the cloud top height predicted value UF (t + xΔt) at a certain time (t + xΔt) A growth rate prediction value GF (t + xΔt) is calculated.

  Here, it is a natural number of x = 1, 2, 3,. In the prediction process based on the dynamics texture method described above, it is possible to predict within about 10 times the time interval Δt as a guide based on the inertial properties of the data elements. Therefore, when the dynamic texture method is used, the maximum value of x may be 10.

  Thus, the cloud top altitude predicted value UF (t + xΔt) is calculated by correcting the position based on the cloud motion vector predicted value while preserving the cloud top altitude value at time t.

  The predicted growth rate GF (t + xΔt) is a future time (t + Δt) interpolated in time series from the growth rate data G (t−2Δt), G (t−Δt), G (t), etc. It is a value calculated by extrapolating to (t + 2Δt), and correcting the position based on the predicted value of the cloud motion vector, and is the growth rate at time (t + xΔt).

Eventually, the cloud top height at time (t + xΔt) is
UF (t + xΔt) + ΣΔtGF (t + Δt)
It becomes. This is judged and evaluated by the initial growth altitude threshold β or further the developmental altitude threshold γ to evaluate the altitude of the cloud and predict a position with a high lightning probability in the growth rate predicted value GF (t + xΔt). It will be possible.

  It should be noted that the cloud mesh data K (t) and the cloud top height data U (t), which are input data, include errors, so U (t), G (t), UF (t), GF calculated in the present invention. When evaluating (t), there is a possibility that a slight misalignment may be caused to reduce the accuracy of prediction.

  Therefore, in order to round off the evaluation error of such positions, in the present invention, prior to evaluating U (t), G (t), UF (t), GF (t), their individual meshes are evaluated. More preferably, the averaging process is performed in the vicinity of the cell.

  That is, a preliminary calculation step is performed in which the value of the mesh cell at the coordinates (i, j) is the average value of the total of (2a + 1) ^ 2 cells from (i−a, j−a) to (i + a, j + a). To be inserted. Although a is a natural number, a numerical value of about 1 to 3 is preferable.

  In the present invention, the above algorithm is coded as software and displayed together with a map on a computer, thereby clearly indicating lightning position information and prediction information. As a better mode, it is preferable to acquire the latest weather information for each Δt, perform the calculation process inside, and update the screen display.

[Example 1]
FIG. 4 is a diagram showing the configuration of the cloud position prediction system according to the first embodiment of the present invention.

  The cloud position prediction system includes, for example, a weather information distribution center 3 that receives radio waves transmitted from the weather satellite “Himawari” 1 to generate weather information and periodically distribute the weather information, and the weather information distribution center 3. A first personal computer that receives weather information from the Internet 7 to predict and calculate the position of lightning, and receives calculation result information output from the first personal computer via the HUB 15 and displays it on a monitor. 2 personal computers.

  The host computer provided in the weather information distribution center 3 is provided with a server 5 having a database DB. For example, the weather information transmitted from the weather satellite “Himawari” 1 is analyzed and periodically meteorological information. Delivered or sent upon request.

  Next, the first personal computer 11 includes an arithmetic control unit 21, a communication control unit 23, an arithmetic memory unit 25, a hard disk unit 27, a display control unit 29, a monitor 31 and a LAN control unit 33.

  The arithmetic control unit 21 reads out the arithmetic control program stored in the hard disk unit 27 to the internal RAM, and executes arithmetic processing and control processing according to the program. The communication control unit 23 is connected to a server provided in the weather information distribution center 3 via the Internet 7, receives the weather information, and outputs the received weather information to the calculation memory unit 25. The calculation memory unit 25 is a memory space for expanding and calculating received weather information.

  The hard disk unit 27 records the history data K (t), K (t−Δt), K (t−2Δt), etc. of the cloud mesh data as the received weather information, and at the same time, the history of the cloud top altitude mesh data. Data U (t), U (t−Δt), U (t−2Δt), and so on are recorded.

  The display control unit 29 displays the operation screen of the geographic information system GIS on the monitor 31. The LAN control unit 33 transmits the calculation processing result to the second personal computer via the HUB 15.

  Next, the second personal computer 13 includes an arithmetic control unit 41, a LAN control unit 43, a hard disk unit 45, a display control unit 47, a display memory unit 49, and a monitor 51.

  The arithmetic control unit 41 reads out an arithmetic control program stored in the hard disk unit 45 to the internal RAM, and executes arithmetic processing and control processing according to the program. The LAN control unit 43 receives the calculation processing result of the first personal computer via the HUB 15. The hard disk unit 45 records the latest data of cloud mesh data K (t), cloud top altitude mesh data U (t), and growth rate mesh data G (t) that are calculation results received from the first computer 11. The display control unit 47 displays the result display screen of the geographic information system GIS on the monitor 31, and the cloud mesh data K (t) and the cloud top altitude mesh data U (t) that are the calculation results received from the first computer 11. The growth rate mesh data G (t) is expanded in the display memory 49, updated to the latest data, and displayed on the monitor 51 by GIS.

  Next, the operation of the lightning position prediction system of the present invention will be described with reference to the flowchart shown in FIG. Note that the processing related to step Sa is executed in the first personal computer 11 and the processing related to step Sb is executed in the second personal computer 13.

  First, in step Sa10, the cloud control data 23 (t) and the cloud top altitude mesh data U (t) obtained as a function of time t are connected from the communication control unit 23 to the server 5 of the weather information distribution center 3 via the Internet 7. ) Is received via the communication control unit 23, stored in the arithmetic memory unit 25, and recorded in a predetermined holder of the hard disk unit 27.

  That is, as one of the weather information distribution centers 3, for example, a meteorological service support center installed in the Japan Meteorological Agency is connected via the Internet 7, and 1 km square range (mesh) of the weather information released by this center is taken as one unit National Synthetic Radar GPV (Grid Point Value) to be handled is received and input as cloud mesh data K (t), and at the same time, the National Synthetic Radar Echo Altitude GPV which treats 2.5km square range (mesh) as one unit is cloud top. It is received and input as altitude mesh data U (t), and each data is stored in the arithmetic memory unit 25 and recorded in a predetermined holder of the hard disk unit 27.

  The cloud mesh data K (t) and cloud top altitude mesh data U (t) for each mesh distributed from the weather information distribution center are for wind pressure data of 925, 850, 700, 500 hPa on the air pressure surface, Calculation prediction value of weather information such as wind speed data, temperature data, dew point difference data, grid point spacing of east-west 30 ', north-south 24' (cloud top altitude mesh data), east-west 15 ', north-south 12' (cloud mesh data), respectively It is comprised from the positional information on the mesh.

  Since these weather data are information distributed every 10 minutes, Δt = 10 minutes. Note that when inputting these weather data, the grid point data defines a mesh coordinate system so that the grid point is in the center of the mesh, and is expanded in the memory space on the arithmetic memory unit 25. In addition, since the above-described weather data distributed from the meteorological service support center is composed of binary data, the arithmetic control unit 21 converts data from binary data to mesh data when expanding to addresses on the arithmetic memory unit 25. To do.

  As a result, cloud mesh data K (t) and cloud top height mesh data U (t) are newly stored in the arithmetic memory unit 25, and K (t) is stored in the hard disk unit 27 as history data of cloud mesh data. , K (t−Δt), K (t−2Δt),... Are recorded, and at the same time, U (t), U (t−Δt), U ( t-2Δt),... are recorded.

  Next, in step Sa20, momentum analysis processing (DT method) is performed on the cloud mesh data K (t). That is, m = 3, and the calculation control unit 21 performs t−3Δt (that is, 30 minutes before) t at a certain time t with respect to the cloud mesh data K (t) stored in the calculation memory unit 25, t The cloud mesh data K (t) of −2Δt (that is, 20 minutes before) and t−Δt (that is, 10 minutes before) is analyzed by the dynamic texture method.

  Next, in step Sa30, the calculation control unit 21 performs the motion vector (a, cloud motion) of each mesh at the time (t−Δt) until the first order with respect to the result obtained by the momentum analysis processing in step S20. b) is calculated as at time t.

  That is, based on cloud mesh data K (t), K (t−Δt), K (t−2Δt), K (t−mΔt) obtained until time t with Δt as the observation time interval, Cloud motion vectors (a, b) between time (t−Δt) and time t are calculated for each mesh.

  Next, in step S40a, by projecting the cloud motion vector (a, b) onto the cloud top height mesh data U (t-Δt), the horizontal movement amount of the cloud top height data accompanying the cloud movement is corrected, The cloud top height mesh data estimated value Uex (t) is calculated.

  That is, the cloud motion vector (a, b) is projected onto the cloud top altitude mesh data U (t−Δt) and calculated, and obtained by assuming that the cloud has moved between Δt while maintaining the cloud top altitude. The estimated value Uex (t) of the calculated cloud top height mesh data is calculated.

  Next, in step Sa50, a cloud growth rate vector is calculated for each mesh from the cloud growth rate mesh data G (t). The growth rate threshold value α was set to 3 m / s.

G (t) = (U (t) −Uex (t)) / Δt
On the other hand, the display control unit 47 provided in the second computer 13 uses the well-known geographic information system GIS (Geographic Information System) to set the layers of Japan map, K (t), U (t), G (t). These are prepared in the order from the lower layer to the upper layer, and these are linked to the basic map data using the position as a key, and are displayed in a superimposed manner. This makes it possible to grasp the mutual positional relationship, search and display data, analyze the relationship between data, and the like.

  Here, the first computer 11 reads out cloud mesh data K (t), cloud top altitude mesh data U (t), and growth rate mesh data G (t) newly stored in the arithmetic memory unit 25, and performs LAN control. The data is transmitted from the unit 33 to the second personal computer 13 via the HUB 15. Next, in the second personal computer 13, the LAN control unit 43 receives this calculation result, and the cloud mesh data K (t), cloud top altitude mesh data U (t), and the growth rate, which are the calculation results received by the hard disk unit 45. Record the latest data of the mesh data G (t). In the display control unit 47 of the second computer 13, K (t), U (t), and G (t) are updated to the latest data, aligned by GIS, and displayed on the monitor 51.

  In step Sb60, an estimated value Gex (t) of the growth rate mesh data obtained on the assumption that the cloud has moved during the time interval Δt while storing the growth rate value is calculated.

  That is, in the layer of the growth rate mesh data G (t), the cloud growth point is displayed by extracting a mesh having a value larger than the growth rate threshold value α and appropriately assigning and displaying the symbol to the mesh. The display control unit 47 is controlled.

  Next, in step Sb70, a mesh cell having a value greater than or equal to the growth rate threshold value α is extracted from the estimated value Gex (t) of the growth rate mesh data.

  That is, by projecting the motion vector (a, b) onto the growth rate mesh data G (t−Δt), the growth rate expected value Gex (t) is calculated, and a mesh cell satisfying Gex (t)> α is extracted. To do.

  In step Sb110, an initial growth altitude threshold β or a developmental altitude threshold γ is provided for the cloud top altitude mesh data U (t), and mesh cells having values higher than these thresholds are extracted and displayed on the monitor 51. The display control unit 47 is controlled as described above.

  That is, in the second computer 13, in the cloud top altitude mesh data U (t), an initial growth altitude threshold β representing the altitude when the cloud is growing is set to, for example, 3000 (the range of β is 1500 to 3000). M), and the developmental altitude threshold γ representing the altitude of the cloud development period is, for example, 5000 (the range of γ is preferably 5000 to 7000) m. A cloud mesh with a high altitude is selected, and the display color is changed and highlighted, and the position of the cumulonimbus cloud is displayed on the monitor 51.

  In step Sb210, in step Sb110, the cell selected by the cloud top height mesh data U (t) with the growth initial height threshold β and the developmental height threshold γ, and in step Sb60, the growth rate mesh data G (t) A logical OR with a mesh cell having a value equal to or higher than the growth rate threshold value α was taken. When this logical sum is 1, the display control unit 47 is controlled so that the cell is blinked and highlighted, and the mesh cell indicating the position where the risk of lightning is high is aligned and displayed on the monitor 51. To do.

  In step Sa310, time t + Δt, t + 2Δt,..., T + 6Δt, and UF (t) and GF (t) up to one hour ahead are predicted in a sub-window different from the above. UF (t + xΔt) is a cloud top height predicted value at time (t + xΔt) calculated by correcting the position based on the predicted value of the cloud motion vector while holding the value of the cloud top height at time t, and GF (t + xΔt ) Is a predicted growth rate at time (t + xΔt) calculated by correcting the position based on the predicted value of the cloud motion vector.

  In step Sb320, based on the cloud top altitude predicted value based on the cloud top altitude {UF (t + xΔt) + ΣΔtGF (t + xΔt)} and GF (t), a cell having a threshold value or more is extracted in the same manner as described above, and lightning is generated by logical sum. A lightning strike position was predicted for a short time by displaying cells with a high risk level.

  That is, the time value ΔΔt is a result value obtained by multiplying the natural number x by the time interval Δt, and the future cloud top height predicted value UF (t + xΔt) obtained on the assumption that the cloud has moved after the time xΔt while calculating the cloud top height is calculated. .

  Further, by interpolating in time series from the historical data for each time interval Δt of the growth rate mesh data U (t), extrapolating the time for each future time interval Δt, and using the extrapolated result as the predicted value of the motion vector of the cloud Based on this, the position is corrected and the growth rate predicted value GF (t + xΔt) at time (t + xΔt) is calculated.

  Further, the cloud top height {F (t + xΔt) + ΣΔt · GF (t + xΔt)} after time xΔt is calculated from the cloud top height predicted value UF (t + xΔt) and the growth rate predicted value GF (t + xΔt).

  Next, a mesh cell having a value equal to or higher than the growth rate threshold value α with respect to the growth rate predicted value GF (t + xΔt) is extracted. Next, a mesh cell is extracted in which the cloud top height after time xΔt is larger than the initial growth height threshold β or the developmental height threshold γ.

  For the growth rate predicted value GF (t + xΔt), a mesh cell having a value equal to or greater than the growth rate threshold value α is extracted and displayed on the monitor 51. In addition, for the cloud top height {UF (t + xΔt) + ΣΔt · GF (t + xΔt)} after time xΔt, a mesh cell that is higher than the initial growth height threshold β and further develops to a value higher than the developmental height threshold γ is extracted and monitored. 51.

  Here, in step Sb320, if there is a mesh cell in which the cloud top height {UF (t + xΔt) + ΣΔt · GF (t + xΔt)} after time xΔt is equal to or greater than the initial growth height threshold β or the developmental height threshold γ, the lightning occurrence probability Is extremely high, the calculation control unit 41 controls the display control unit 47 to blink the position of the mesh cell on the monitor. Furthermore, since there is a high possibility that lightning will strike at this position after time xΔt (xΔt = ○ minutes), the arithmetic control unit 41 will display a message “There is a lightning prediction after time OO minutes” in the vicinity of the blinking position. The display control unit 47 is controlled so as to display.

  In addition, in order to correct the position evaluation error, evaluation by averaging processing of U (t), G (t), UF (t), and GF (t) was also performed. That is, a = 1, and the value of the mesh cell at the coordinates (i, j) located at the center is a total of 9 values of the values of the surrounding mesh cells (i−1, j−1) to (i + 1, j + 1). The average value was used.

  As a result, the position evaluation error can be corrected, so that the noise component is reduced, and a relatively stable output can be obtained even when the wind speed is fast and the cloud moves quickly.

[Example 2]
With reference to FIG. 4, the structure of the cloud position prediction system which concerns on 2nd Embodiment of this invention is demonstrated. In the first embodiment, as the cloud mesh data K (t), for example, a weather satellite photograph is used as distribution data of the weather satellite “Himawari”.

  For this reason, since the input timing is different from the 10-minute value and the delivery interval from the server 5 is an hour interval, this is set to Δt ′. That is, Δt ′ = 6Δt.

  Therefore, in the processing of steps Sa10 to Sa30, Δt ′ is used for the momentum evaluation using the cloud mesh data K (t), and the cloud top height mesh data U (t) is evaluated from (t−5Δt) to (t + 5Δt). Is assumed to be applied by linear interpolation that the change in momentum by the cloud mesh data K (t) from (t−Δt ′) to (t + Δt ′) is uniform, and the processing after step Sa40 is the processing shown in FIG. The same shall apply.

  As described above, according to the first and second embodiments, it is possible to predict the lightning strike position or lightning time at the stage before the lightning by calculating the cloud growth speed at the stage before the lightning. In particular, in the short-term prediction of thermal lightning, it is possible to predict the lightning position and lightning time with high accuracy even before the lightning.

  In addition, by measuring the cumulonimbus growth process as time-series cloud top height mesh data and estimating the subsequent cloud growth process and position, it is possible to predict the location of lightning strikes even before the stage of lightning. it can.

  As a result, it is possible to effectively take proactive measures in accordance with lightning strike predictions, and contribute to speeding up recovery after a disaster by strengthening advance preparations.

It is a schematic diagram (a), (b), (c) which shows that the cloud moved while growing, and a schematic diagram (d), (e) of the growth rate by the cloud top height and a simple difference. They are schematic diagrams (a), (b), (c) showing that the cloud has moved while maintaining its size, and schematic diagrams (d), (e) of the growth rate based on the cloud top height and a simple difference. It is a schematic diagram in the case where the growth rate is evaluated by correcting the position of the cloud, and is a schematic diagram (a), (b) showing that the cloud has moved while becoming larger, and a schematic diagram showing the estimated value of the cloud top height ( c) and a schematic diagram (d) showing the growth rate. It is a figure which shows the structure of the lightning position prediction system of this invention. It is a flowchart which shows operation | movement of the lightning position prediction system of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Meteorological satellite 3 Weather information distribution center 5 Server 7 Internet 11 1st personal computer 13 2nd personal computer 15 HUB
21, 41 Operation control unit 23 Communication control unit 25 Operation memory unit 27, 45 Hard disk unit 29, 47 Display control unit 31, 51 Monitor 33, 43 LAN control unit 49 Display memory unit

Claims (8)

A lightning position prediction method for predicting the position of lightning by analyzing information on weather regularly distributed from a server provided in a weather information distribution center,
Receiving cloud mesh data K (t) and cloud top height mesh data U (t) distributed from the server;
Based on the history of cloud mesh data K (t) obtained up to time t each time the observation time interval Δt elapses, the cloud motion vector from time (t−Δt) to time t is determined for each mesh. A process of calculating;
This cloud motion vector is projected onto the cloud top height mesh data U (t−Δt), and the estimated value Uex of the cloud top height mesh data obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the cloud top height. calculating (t);
Calculating a result obtained by dividing a difference value obtained by subtracting the estimated value Uex (t) of the cloud top altitude mesh data from the cloud top altitude mesh data U (t) by the time interval Δt as cloud growth rate mesh data G (t);
The lightning position prediction method characterized by having. Extracting a mesh cell in which the cloud growth rate mesh data G (t) is equal to or greater than a predetermined growth rate threshold value α;
Calculating an estimated value Gex (t) of the growth rate mesh data G (t−Δt) obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the value of the growth rate;
Projecting the motion vector onto growth rate mesh data G (t−Δt) to calculate a growth rate expected value Gex (t);
Extracting a mesh cell in which a growth rate expected value Gex (t) is a value equal to or greater than a growth rate threshold value α indicating cloud growth;
The lightning position prediction method according to claim 1, further comprising:   The cloud top height mesh data U (t) is a mesh cell having a value equal to or higher than the initial growth height threshold β indicating the initial height of cloud growth, or a value equal to or higher than the developmental height threshold γ indicating the height of cloud development. The lightning position prediction method according to claim 1, further comprising a step of extracting a mesh cell. The cloud top height mesh data U (t) is a mesh cell having a value larger than the initial growth height threshold β or the developmental height threshold γ, and the cloud growth speed mesh data G (t) is the growth speed threshold α. Extracting a mesh cell having the above value and recognizing the position of the mesh cell as a position with a lightning risk,
The lightning position prediction method according to claim 3, further comprising: Calculating a future cloud top height predicted value UF (t + xΔt) obtained by assuming that the cloud has moved after time xΔt while maintaining the cloud top height as time xΔt obtained by multiplying the natural number x by the time interval Δt; ,
By interpolating in time series from the history data for each time interval Δt of the growth rate mesh data U (t), extrapolating the time for each future time interval Δt, and using the extrapolated result as the predicted value of the motion vector of the cloud Correcting the position based on the calculated value, and calculating a growth rate predicted value GF (t + xΔt) at time (t + xΔt);
Calculating a cloud top height {F (t + xΔt) + ΣΔt · GF (t + xΔt)} after time xΔt from the cloud top height predicted value UF (t + xΔt) and the growth rate predicted value GF (t + xΔt);
Extracting a mesh cell having a value equal to or higher than the growth rate threshold value α with respect to the growth rate prediction value GF (t + xΔt);
Extracting a mesh cell in which the cloud top height after time xΔt is greater than the initial growth height threshold β or the developmental height threshold γ;
The lightning position prediction method according to claim 1, further comprising:   When calculating the cloud top altitude mesh data U (t), cloud growth rate mesh data G (t), cloud top altitude predicted value UF (t), and growth rate predicted value GF (t), respectively, The lightning position prediction method according to claim 4, further comprising a step of performing an averaging process in the vicinity.   When a mesh cell in which the cloud top height after the time xΔt is larger than the initial growth height threshold β or the developmental height threshold γ is extracted, there is a possibility that lightning will occur at the position of the mesh cell after the time xΔt. The lightning position prediction method according to claim 5, further comprising a step of displaying a message indicating that the power is high. A lightning position prediction system that receives and analyzes information on weather regularly distributed from a server provided in a weather information distribution center, analyzes the lightning position prediction device, and predicts the position of lightning,
The lightning position prediction device
Means for receiving cloud mesh data K (t) and cloud top height mesh data U (t) distributed from the server;
Based on the history of cloud mesh data K (t) obtained up to time t each time the observation time interval Δt elapses, the cloud motion vector from time (t−Δt) to time t is determined for each mesh. Means for computing;
The cloud motion vector is projected onto the cloud top height mesh data U (t−Δt), and the estimated value Uex of the cloud top height mesh data obtained on the assumption that the cloud has moved during the time interval Δt while maintaining the cloud top height. means for calculating (t);
Means for calculating, as cloud growth rate mesh data G (t), a result obtained by dividing a difference value obtained by subtracting the estimated value Uex (t) of the cloud top altitude mesh data U (t) from the cloud top altitude mesh data U (t);
A lightning position prediction system comprising:

Images