JP2006242747A - Air temperature prediction correction device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an air temperature prediction correction device capable of determining an air temperature prediction correction value based on the actual situation by correcting an air temperature predicted value without using past time-series data. <P>SOLUTION: The air temperature prediction correction device 10 is equipped with an air temperature predicted value input part 11 for inputting the air temperature predicted value, an air temperature evaluation position input part 12, and the first air temperature correction means 13 for determining the air temperature prediction correction value at an air temperature evaluation position based on the air temperature predicted value inputted from the air temperature predicted value input part 11. The first air temperature correction means 13 determines a prediction correction value by correcting the air temperature predicted value by using the linear expression T<SP>1</SP><SB>adjust</SB>=a×T<SB>predict</SB>+b... (1). In the expression (1), T<SP>1</SP><SB>adjust</SB>is the air temperature prediction correction value, T<SB>predict</SB>is the air temperature predicted value at the evaluation position, a is a constant and b is a constant. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a weather prediction correction device that corrects a weather prediction value obtained by a weather prediction code.

Conventionally, when correcting a predicted temperature value predicted from a weather prediction code, an error between the predicted temperature value predicted by the weather prediction code and an actual measurement value at the position is accumulated as a database over a period of time, and correction is performed based on the data. Is doing. In these correction methods, in order to enhance the correction effect, correction is performed based on time series data of a past fixed period accumulated from the prediction time, or correction is performed using time series data updated every fixed period. (For example, refer to Patent Document 1).
JP 2004-86896 A

  By the way, in the correction method as described above, time series data accumulated in the past at the position is necessary for correction. Therefore, the position that can be corrected in the analysis region is only a limited position where the accumulated time series data exists, and the temperature predicted value cannot be corrected at the position where the time series data does not exist.

  The present invention has been made in consideration of such points, and provides a temperature prediction correction device that can obtain a temperature prediction correction value according to the actual situation without using past time-series data. Objective.

を用いて補正して気温の評価位置の気温予測補正値を求める気温補正手段と、気温補正手段で求めた気温予測補正値を出力する出力部とを備え、

であることを特徴とする気温予測補正装置である。 The present invention relates to a temperature prediction correction apparatus that corrects a temperature prediction value predicted by a temperature prediction means having a weather prediction code for obtaining a temperature prediction value corresponding to a temperature evaluation position, and predicts with a temperature evaluation position and a weather prediction code. Input unit to input the predicted temperature value and a linear expression for the predicted temperature value input from the input unit

A temperature correction unit that calculates the temperature prediction correction value of the temperature evaluation position by correcting using, and an output unit that outputs the temperature prediction correction value calculated by the temperature correction unit,

It is the temperature prediction correction device characterized by being.

  According to the present invention, in the temperature correction means, the constant b in the linear expression is the lowest value among the time series data of the temperature prediction values predicted by the temperature prediction means at the evaluation position. It is.

  In the temperature correction means, the constant b in the linear expression is an average value of time series data of the temperature prediction values predicted by the temperature prediction means at the evaluation position. is there.

  The present invention relates to a temperature prediction correction apparatus for correcting a temperature prediction value predicted by a temperature prediction means having a weather prediction code for obtaining a temperature prediction value corresponding to a temperature evaluation position based on mesh data. An input unit for inputting the mesh data used in the weather prediction code, the terrain data measured with respect to the mesh more detailed than the mesh data used in the weather prediction code, and the temperature predicted value predicted by the weather prediction code; The difference between the altitude value at the temperature evaluation position calculated from the mesh data and the altitude value at the temperature evaluation position calculated from the terrain data is obtained as an altitude error, and the predicted temperature value input from the input unit is the altitude error. Temperature correction means for calculating the temperature prediction correction value at the temperature evaluation position by correcting using the temperature, and the temperature prediction correction value calculated by the temperature correction means A temperature prediction correction apparatus characterized by comprising an output section for outputting.

  The present invention is characterized in that, in the temperature correction means, the temperature prediction value input from the input unit is corrected using a linear function of an altitude error to obtain a temperature prediction correction value at the evaluation position of the temperature. It is a correction device.

を用いて気温の評価位置の気温予測補正値を求める第１気温補正手段と、気温の評価位置、気象予測コードで使用されたメッシュデータ、気象予測コードで使用されたメッシュよりも詳細なメッシュに対して測定された地形データ、および気象予測コードで予測された気温予測値を入力する第２入力部と、メッシュデータから算出された気温の評価位置における標高値と、地形データから算出された気温の評価位置における標高値との差を標高誤差として求め、第２入力部から入力された気温予測値を標高誤差を用いて補正して気温の評価位置の気温予測補正値を求める第２気温補正手段と、第１気温補正手段または第２気温補正手段のいずれかにより求めた気温予測補正値を出力する出力部と、を備えたことを特徴とする気温予測補正装置である。 The present invention relates to a temperature prediction correction apparatus for correcting a temperature prediction value predicted by a temperature prediction means having a weather prediction code for obtaining a temperature prediction value corresponding to a temperature evaluation position based on mesh data. A first input unit for inputting a predicted temperature value predicted by a weather prediction code, and a linear expression for the predicted temperature value input from the first input unit

The first temperature correction means for calculating the temperature prediction correction value of the temperature evaluation position using the temperature, the temperature evaluation position, the mesh data used in the weather prediction code, the mesh more detailed than the mesh used in the weather prediction code The second input unit for inputting the terrain data measured for the temperature and the temperature predicted value predicted by the weather prediction code, the altitude value at the temperature evaluation position calculated from the mesh data, and the temperature calculated from the terrain data Second temperature correction to obtain the difference between the altitude value at the evaluation position of the current as an altitude error, and to correct the temperature prediction value input from the second input unit using the altitude error to obtain the temperature prediction correction value at the temperature evaluation position And an output unit that outputs a temperature prediction correction value obtained by either the first temperature correction unit or the second temperature correction unit. A.

  As described above, according to the present invention, it is possible to correct a temperature predicted value without using past time-series data, and obtain a temperature predicted correction value according to the actual situation.

  Embodiments of the present invention will be described below with reference to the drawings. 1 to 8 are diagrams showing an embodiment of an air temperature prediction correction apparatus according to the present invention.

  As shown in FIG. 1, the temperature prediction correction device 10 corrects the weather prediction value predicted by the temperature prediction means having a weather prediction code for obtaining the temperature prediction value corresponding to the temperature evaluation position. In this case, a weather model RSM used by the Japan Meteorological Agency can be considered as the weather prediction code.

ここで

を用いて気温の評価位置の気温予測補正値を求める第１気温補正手段１３とを備えている。 The weather prediction correction device 10 includes a temperature prediction value input unit 11 to which a temperature prediction value predicted by a weather prediction code is input, and a temperature evaluation position input unit 12 to which an evaluation position (longitude and latitude) of the temperature is input. Based on the temperature prediction value input from the temperature prediction value input unit 11 and the evaluation position of the temperature evaluation position input from the input unit 12, a linear expression for the temperature prediction value

here

And a first temperature correction means 13 for obtaining a temperature prediction correction value at the temperature evaluation position.

  Among these, the temperature input predicted value input unit 11 and the temperature evaluation position input unit 12 constitute a first input unit.

  As the constant b of the linear function, the lowest value of the time series data of the temperature predicted value input from the temperature predicted value input unit 11 is used. In addition, you may use the average value of the time series data of temperature predicted value as the constant b.

  The constant a is determined based on AMeDAS data used in the Japan Meteorological Agency.

  Further, a mesh data input unit 15 for inputting mesh data used in the weather prediction code and a terrain data input unit 16 for inputting terrain data measured with respect to a mesh more detailed than the mesh data are provided.

  The mesh data input unit 15 and the terrain data input unit 16 are connected to the second temperature correction unit 17 and calculated from the altitude value at the temperature evaluation position calculated from the mesh data in the second temperature correction unit 17 and the terrain data. And the altitude value at the evaluation position of the measured temperature. The second temperature correction means 17 further obtains a difference between the altitude value calculated from the mesh data and the altitude value calculated from the topographic data as an altitude error, and corrects the temperature predicted value using the altitude error to evaluate the position. Find the temperature prediction correction value at.

  In this case, the temperature input predicted value input unit 11, the temperature evaluation position input unit 12, the mesh data input unit 15, and the terrain data input unit 16 constitute a second input unit.

の形で表される式を用いて気温予測値を補正する。ここで、

である。この場合、定数ｃとｄは、ＡＭｅＤＡＳ等の測定された過去のデータを用いて決定する。 As described above, the second temperature correction means 17 is information on the position (latitude / longitude) at which the temperature is to be evaluated, mesh data used in the weather prediction code, and detailed topographic data (for example, a numerical map issued by the Geographical Survey Institute). Based on the above, the temperature prediction correction value at the evaluation position is obtained. In this case, the second temperature correction means 17 corrects the temperature predicted value by using a linear function of an elevation difference between the elevation value of the evaluation position calculated from the mesh data and the elevation value calculated from the topographic data. In particular,

The predicted temperature value is corrected using an expression expressed in the form of. here,

It is. In this case, the constants c and d are determined using measured past data such as AMeDAS.

  Next, the temperature prediction correction value obtained by either the first temperature correction means 13 or the second temperature correction means 17 is output from the output unit 18.

  In the present embodiment, the predicted temperature correction value obtained by either the first temperature correction means 13 or the second temperature correction means 17 is output from the output unit 18, but the first temperature correction means 13 or the first temperature correction means 13 Only one of the two temperature correction means 17 may be provided, and only the temperature prediction correction value obtained by these temperature correction means 13 or 17 may be output from the output unit 18.

  Next, the operation of the present embodiment having such a configuration will be described.

  The weather model used in the weather prediction code has a habit of predicting higher temperatures or lower temperatures. The first air temperature correction unit 13 extracts the characteristics of the cocoon from the past temperature predicted value data and corrects it using the above-described linear expression to obtain a highly accurate temperature predicted correction value.

  The second temperature correction means 17 performs correction by paying attention to an altitude error of the temperature evaluation position caused by the mesh data including the analysis mesh. Generally, when the altitude increases by 100 m, there is a temperature drop of 0.5 to 1 ° C. In meteorological analysis, the size of the analysis mesh in the horizontal direction is 5 to 20 km or more due to the region to be analyzed and the limitation of the calculation amount. Therefore, the altitude value of the evaluation position calculated from the mesh data including these meshes is significantly different from the actual altitude in the Chubu region and regions with large undulations such as Hakone.

  In the present invention, such an altitude error is calculated from detailed topographic data, and a temperature difference due to the altitude difference is corrected, thereby generating a highly accurate temperature prediction correction value.

  The first air temperature correction means 13 and the second air temperature correction means 17 correct the prediction error caused by the weather prediction code and the prediction error caused by the size of the analysis mesh of the mesh data, respectively. It depends only on the model of the prediction code and the analysis mesh. Therefore, once the constant of the linear function used when correcting the predicted temperature value is determined once, it can be used thereafter. In addition, after the determination, it is possible to correct the predicted temperature value at an arbitrary position regardless of the presence or absence of past measurement data.

  Next, the function and effect of the present invention will be described with reference to specific examples.

  First, the effect of the 1st temperature correction means 13 is demonstrated. FIG. 2 shows an average error between the temperature prediction value and the actual measurement value obtained at the AMeDAS point of 265 by the weather prediction code, the temperature prediction value and the actual measurement for the Chubu / Kanto region including 265 points (location) of the Japan Meteorological Agency AMeDAS. It is time series data which calculated the root mean square error of a value for every time.

  As shown in FIG. 2, both the average error and the square root mean square error are small at night, and both are large during the day. Therefore, the accuracy of temperature prediction is improved by correcting the rate of temperature rise during the day with a coefficient based on a low estimated temperature value at night when the average error can be expected to be close to zero.

  FIG. 3 shows time series data of the average error between the temperature prediction value and the actual measurement value at 256 points of AMeDAS, and the average error between the temperature prediction correction value and the actual measurement value by the first temperature correction means 13. FIG. 4 shows the square root mean square error of the value and the square root mean square error of the temperature prediction correction value and the actual measurement value. As shown in FIGS. 3 and 4, it can be seen that the error is reduced by the correction by the first temperature correction means 13.

  Next, the function and effect of the second temperature correction means 17 will be described. Fig. 5 shows the AMeDAS locations calculated from the elevation values of each AMeDAS point calculated from mesh data at intervals of 5 km in the horizontal direction and 250 m topographic data published by the Geospatial Information Authority of Japan for the Chubu and Kanto regions including 265 points of the Meteorological Agency AMeDAS. The difference from the elevation value (true value) of a point is represented as a histogram. In FIG. 5, there is also a point having an altitude error of 500 m or more. An altitude difference of 500 m causes a large error in the predicted temperature value.

に現れる定数は、
ｃ＝−０．０４９
ｄ＝０．１８７４
となる。この一次式は、過去の事例を用いて求めることができる。この一次式は、標高誤差に基づいて算出されるものであり、原理的にはメッシュやデータ測定の場所には依存しない。したがって、一次式の算出は一度行なえば、その後も使用することができる。 Fig. 6 shows the result of meteorological analysis in the same area using a mesh of 5km in the horizontal direction, comparing the predicted temperature value with the measured value at the AMeDAS256 point, and calculating the altitude error and the average error between the predicted temperature value and the measured value. The relationship is plotted. Here, the temperature prediction value was calculated within the analysis mesh. There is a correlation between the altitude error and the average error between the predicted temperature value and the actual measurement value. In this case, a linear expression representing the correlation

The constant that appears in
c = −0.049
d = 0.1874
It becomes. This linear expression can be obtained using past cases. This linear expression is calculated based on the altitude error, and in principle does not depend on the location of the mesh or data measurement. Therefore, once the primary expression is calculated, it can be used thereafter.

  Next, FIG. 7 shows the time series data of the average error between the temperature predicted value and the actual measured value and the average error between the temperature predicted correction value and the actual measured value by the second temperature correction means 17. FIG. 8 shows time-series data of the square root mean square error, and the square root mean square error between the temperature prediction correction value by the second temperature correction means 17 and the actual measurement value.

  As shown in FIGS. 7 and 8, it can be seen that the error is reduced by the correction by the second temperature correction means 17.

The block diagram which shows the whole structure of the temperature prediction correction | amendment apparatus by this invention. The figure which shows the average error (ME) and square root mean square error (RMSE) of an estimated temperature value and an actual measurement value. The figure which shows the average error before and behind the correction | amendment of a 1st temperature correction means. The figure which shows the square root mean square error before and behind the correction | amendment of a 1st temperature correction means. The figure which shows distribution of the altitude error of an AMeDAS observation point. The figure which shows the relationship between an altitude error and the average error of temperature predicted value. The figure which shows the average error before and behind the correction | amendment of a 2nd temperature correction means. The figure which shows the square root mean square error before and behind the correction | amendment of a 2nd temperature correction means.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Temperature prediction correction | amendment apparatus 11 Temperature prediction value input part 12 Temperature evaluation position input part 13 1st temperature correction means 15 Mesh data input part 16 Terrain data input part 17 2nd temperature correction means 18 Output part

Claims (6)

In the temperature prediction correction device for correcting the temperature prediction value predicted by the temperature prediction means having the weather prediction code for obtaining the temperature prediction value corresponding to the temperature evaluation position,
An input unit for inputting an estimated temperature value and an estimated temperature value predicted by a weather prediction code;
Linear equation for the predicted temperature value input from the input unit

A temperature correction means for calculating the temperature prediction correction value of the temperature evaluation position by correcting using
An output unit for outputting a temperature prediction correction value obtained by the temperature correction means,

The temperature prediction correction device characterized by being.   2. The temperature prediction according to claim 1, wherein the constant b in the linear expression is the lowest value of the time series data of the temperature predicted values predicted by the temperature prediction means at the evaluation position. Correction device.   2. The temperature prediction correction according to claim 1, wherein the constant b in the linear expression is an average value of time series data of the temperature prediction values predicted by the temperature prediction means at the evaluation position. apparatus. In the temperature prediction correction device for correcting the temperature prediction value predicted by the temperature prediction means having the weather prediction code for obtaining the temperature prediction value corresponding to the temperature evaluation position based on the mesh data,
The temperature evaluation position, the mesh data used in the weather prediction code, the terrain data measured for the mesh more detailed than the mesh data used in the weather prediction code, and the temperature prediction value predicted by the weather prediction code An input section to input,
The difference between the altitude value at the temperature evaluation position calculated from the mesh data and the altitude value at the temperature evaluation position calculated from the topographic data is obtained as an altitude error, and the estimated temperature value input from the input unit is calculated as the altitude error. Temperature correction means for correcting and calculating the temperature prediction correction value of the temperature evaluation position,
An temperature prediction correction apparatus comprising: an output unit that outputs a temperature prediction correction value obtained by the temperature correction means.   5. The temperature according to claim 4, wherein the temperature correction means corrects the temperature prediction value input from the input unit using a linear function of an altitude error to obtain a temperature prediction correction value at the temperature evaluation position. Predictive correction device. In the temperature prediction correction device for correcting the temperature prediction value predicted by the temperature prediction means having the weather prediction code for obtaining the temperature prediction value corresponding to the temperature evaluation position based on the mesh data,
A first input unit that inputs an estimated temperature position and an estimated temperature value predicted by a weather prediction code;
A linear expression for the predicted temperature value input from the first input section

First temperature correction means for obtaining a temperature prediction correction value of the temperature evaluation position using
Enter the temperature assessment location, mesh data used in the weather forecast code, terrain data measured on a more detailed mesh than the mesh used in the weather forecast code, and the temperature forecast predicted by the weather forecast code A second input unit to
The difference between the altitude value at the temperature evaluation position calculated from the mesh data and the altitude value at the temperature evaluation position calculated from the terrain data is obtained as an altitude error, and the temperature prediction value input from the second input unit is calculated as the altitude. A second temperature correction means for correcting the error using an error to obtain a temperature prediction correction value at the temperature evaluation position;
An output unit for outputting a temperature prediction correction value obtained by either the first temperature correction means or the second temperature correction means;
A temperature prediction correction apparatus characterized by comprising:

Images