JP2006220444A - Weather prediction system and its assimilation processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To appropriately correct the assimilation of data on a quantity of rain water in an unobservable region in which radar waves are shielded. <P>SOLUTION: Data on the quantity of rain water of a plurality of strata divided by altitude is acquired from results of radar observation in advance. Data on the quantity of rain water of lower strata is retrieved by the start of assimilation (S31) to determine the presence or absence of rain for every point of observation (S32). When the presence of rain is determined, water vapor assimilation is performed for all the lower strata at the points (S33). When the water vapor assimilation processing is completed or the absence of rain is determined in S32, data on the quantity of rain water of upper strata is retrieved (S34), and the quantity of rain water is compared with a threshold value again to determine the presence or absence of rain (S35). When the presence of rain is determined in S35, water vapor assimilation is performed for points at which water vapor assimilation has never been performed among lower strata at points at which rain has been observed (S36) to return to S34, and data on the quantity of rain water of the next strata is retrieved. When the absence of rain is determined in S35, data assimilation processing is assumed to be completed. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a weather prediction system that inputs various sensor observation data and weather radar data into a weather prediction model and simulates and predicts weather phenomena, and an assimilation processing method thereof.

  Currently, in a weather prediction model with a horizontal region of 500 km, calculation is performed using prediction data of a wide-area weather prediction model with a horizontal region of several thousand km as an initial value. In addition, an option for assimilating data on the amount of rainwater and wind speed vector observed by the weather radar and performing future prediction calculation is added (for example, see Patent Document 1).

  By the way, as a result of assimilating the rainwater data of the weather radar into a weather prediction model and conducting a variety of prediction calculation experiments, the rainwater forecast within the radar data assimilation period is AMeDAS (Automated Meteorological Data Acquisition). System regional meteorological observation system) often shows good agreement with rainfall data. However, if the prediction calculation is continued as it is after the radar data assimilation is completed, the rain that appears in the model immediately stops, and a phenomenon in which the coincidence with the observation cannot be seen occurs.

  There are several reasons for this, but it is said that the main factor that stops rain immediately after the data assimilation is completed is that data is only assimilated for the amount of rainwater and not data for water vapor that generates rain. Conceivable. That is, even if only the amount of rainwater is assimilated, it is considered that rainwater has a significant fall speed and immediately falls to the surface of the ground, so that it has almost no effect on the model after the fall. In order to keep raining, it is thought that the substance that makes rainwater, that is, water vapor is indispensable. Therefore, for the purpose of assimilating the amount of water vapor, “a water vapor bocus using rainwater data by a weather radar” has been proposed (see Patent Document 2).

The application of this water vapor assimilation can dramatically improve the accuracy of weather prediction. However, in the observation of weather radar, since it is impossible to observe in the shielded area of the radar radio wave entering the Sanin etc. at low altitude, assimilation of the radar data is performed except for that area. For this reason, even within the same radar coverage, the prediction calculation accuracy in the shielded area of the radar radio wave is extremely reduced as compared with the surrounding area.
JP 2003-090888 A Japanese Patent Application No. 2004-217608

  As described above, in the conventional weather forecasting system based on the weather forecast model using rainwater data of the weather radar, even if the prediction accuracy is improved by adopting the water vapor bocus, the influence of the unobservable area due to the radar wave shielding is a model. And the observation accuracy varies depending on the region.

  The present invention has been made to solve the above problem, and can appropriately correct assimilation of rainwater data in an unobservable region where radar waves are shielded, thereby improving prediction accuracy in the entire model area. An object of the present invention is to provide a weather forecasting system and an assimilation processing method.

  In order to solve the above problem, the weather prediction system according to the present invention assimilates the rainwater amount data observed by the weather radar into a weather prediction model, and periodically compares the prediction value obtained by the model with the observation value. A weather prediction system for interpolating the comparison result into the model, wherein the meteorological radar is scanned in a plurality of hierarchies previously classified by altitude to obtain storm water data for each hierarchy to obtain weather forecasts, respectively. Assimilation to the model, when assimilation of rainwater data is obtained in the upper layer of the point that was not observed on the lower hierarchy side, assimilation means adopting assimilation in the upper layer, and weather forecast model assimilated by the assimilation means The determination means for determining a point where the amount of rainwater exceeds a predetermined value, and steam bogus gas for the point where the amount of rainwater exceeds the predetermined value is created and assimilated in the weather prediction model. Bogas assimilation means, and the steam Bogas assimilation means is a saturated steam obtained by the temperature and the atmospheric pressure at the same point as the amount of water vapor represented in the model before data assimilation at a point where the amount of rainwater exceeds a predetermined value. The amount is compared to the amount of saturated water vapor by carrying out a nodding treatment.

  Further, the assimilation processing method of the weather prediction system according to the present invention assimilates the rainwater amount data observed by the weather radar into the weather prediction model, and periodically compares the predicted value obtained by the model with the observed value. A hierarchical data acquisition step that is applied to a weather prediction system that interpolates a comparison result into the model, and that acquires rainwater data by hierarchical level obtained by scanning the weather radar in a plurality of hierarchical levels that have been classified according to altitude in advance. , An associative step by step that assimilates the rainwater data obtained at each of the plurality of hierarchies into a weather prediction model, and an amount of rainwater data at the upper layer of the point that was not observed on the lower level from the processing result of the assimilation by step If an assimilation is obtained, there is an assimilation synthesis step that employs assimilation in the upper layer, and a weather forecast model assimilated in the assimilation synthesis step, and the amount of rainwater exceeds a predetermined value A determination step of determining a point to be detected, and a steam bogus assimilation step by creating and assimilating the steam bogus for the point where the amount of rainwater exceeds a predetermined value in the weather prediction model, the steam bogas assimilation step Compares the water vapor amount expressed in the model before data assimilation with the saturated water vapor amount obtained from the air temperature and the atmospheric pressure at a point where the rainwater amount is greater than or equal to the predetermined value, and performs nosing processing to obtain the saturated water vapor amount. It is characterized by being close.

  According to the present invention, even in a region where radar waves are shielded, since it is supplemented and assimilated with rainwater amount data in the sky, a weather prediction system capable of highly accurate prediction regarding rainwater amount and assimilation processing thereof A method can be provided.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a schematic configuration of a weather prediction system to which the present invention is applied. In FIG. 1, a communication processing unit 11 is connected to a network NT through a communication interface 12 and is distributed nationwide weather observation data (hereinafter referred to as GPV (Grid Point Value)) from a weather data server DS0 such as the Japan Meteorological Agency on the network. And local weather observation data provided from data servers DS1, DS2,... Such as the prediction target area and surrounding radar sites. The meteorological observation data obtained by the communication processing unit 11 is stored in the observation data storage unit 13 and selectively sent to the calculation unit 14 in response to a request from the weather prediction model calculation unit 14. Further, the weather prediction data obtained by the calculation unit 14 is accumulated in the prediction data storage unit 15.

  The weather prediction model calculation unit 14 first takes GPV data from the observation data storage unit 13 as an initial value, calculates a weather prediction model, and then calculates a local weather observation value such as a radar site from the observation value storage unit 13. The data is captured and interpolated into the spatial grid points of the weather forecast model, and data assimilation is performed to correct the predicted value. The corrected predicted value (predicted data) is stored in the predicted data storage unit 15. Next, the previously calculated prediction data is fetched from the prediction data storage unit 15, and a weather prediction model is calculated using this as an initial value. Thereafter, the weather prediction model is updated using the corrected predicted value as an initial value until new GPV data is obtained.

  FIG. 2 shows a flow of data assimilation processing of the weather forecast model calculation unit 14. In FIG. 2, first, an observed value and a predicted value are taken in and quality control processing is performed (S11). Here, poor quality data is excluded from assimilation by comparing observed values with predicted values.

  Next, using the calculated interpolation weight value calculated separately based on the statistical error characteristics of the observed value and the predicted value, the observed value after quality control is spatially interpolated into the lattice points of the weather prediction model. (S12). At this time, in the first model calculation, GPV data is used as the initial value, and the corrected prediction data is used in the next model calculation.

  After the interpolation weight calculation process, a data assimilation process is performed (S13). In this data assimilation process, the interpolated result is taken into the numerical model continuously in time. The predicted value is compared with the observed observation value, and the predicted value is corrected.

  In the above system, when the rainwater amount data of the weather radar is synchronized with the data, the rainwater amount prediction within the data assimilation period is almost the same as the actual observation data, but the data synchronization is terminated. If the prediction calculation is continued as it is, the rain that appeared in the model stops immediately, and the observation does not match.

  There are several reasons for this, but it is said that the main factor that stops rain immediately after the data assimilation is completed is that data is only assimilated for the amount of rainwater and not data for water vapor that generates rain. Conceivable. That is, even if only the amount of rainwater is assimilated, it is considered that rainwater has a significant fall speed and immediately falls to the surface of the ground, so that it has almost no effect on the model after the fall. In order to keep raining, it is thought that the substance that makes rainwater, that is, water vapor is indispensable.

  Therefore, various attempts have been made to assimilate the amount of water vapor, but there are many problems in obtaining the raw observation data of water vapor and assimilating it into a model. In particular, the water vapor data assimilation method using a GPS (Global Positioning System) satellite or the like is accumulating data. However, since routine data for actual operation cannot be obtained, it is very difficult at present. In view of this, the present inventors have proposed “a water vapor bocus using meteorological radar data” in Patent Document 2 described above. This “water vapor bocus” means that the water vapor amount is estimated from the radar data and the data is assimilated into the model. Hereinafter, FIG. 3 shows a flow chart of water vapor amount data assimilation processing using a water vapor bocus, and the processing content will be described.

  First, assimilation of rainwater data in the weather radar is performed to calculate the amount of rainwater to be assimilated (S21). After assimilation of the rain water amount data, the water vapor amount is estimated (S22). This “estimation calculation of water vapor amount” is based on the assumption that the amount of rainwater assimilated around a defined point in the model calculation area exceeds a predetermined value, and at that point, the water vapor that generated the surrounding rain water limits the saturated water vapor amount to the upper limit. Based on the idea that it exists abundantly.

  As a result of the estimation calculation of the water vapor amount, it is determined whether or not there is rainwater assimilated by the radar in the vicinity (S23). If the determination does not exist, the rainwater volume data assimilation process is terminated. If it is determined that it exists, a steam bogus is created and assimilated in the model (S24).

  Specifically, at the definition point, the amount of water vapor expressed in the model before data assimilation is compared with the saturated water vapor amount obtained from the air temperature and atmospheric pressure at the same point, and saturation is achieved by using the following formula to perform nothing. An estimation method is adopted that approaches the amount of water vapor, that is, facilitates the generation of cloud water and rainwater.

qvassim = qvmod * (1-alpha) + qvsat * alpha
qvassim: amount of water vapor after assimilation (unknown amount)
qvmod: Water vapor volume (calculated by model) before assimilation (known volume)
qvsat: Saturated water vapor at the defined point (known amount)
alpha: Weighting parameter (0 <alpha <1)
However, in the above method, the following three points need to be determined in advance by a parameter experiment.
・ Search range of rainwater assimilated by surrounding radar when performing water vapor assimilation
・ Rainwater threshold to search
・ "Weighting parameters" for water vapor assimilation
By executing the above processing, the rainwater assimilated with the radar data immediately falls and does not leave an influence in the model area, but can leave water vapor drifting in the air in the model area. The place where it is raining is turbulent, including the surroundings, and by assimilating the amount of water vapor, it is expected that the water vapor will condense and generate clouds and rain. This would make it possible to “continue rain” in the model.

  Therefore, in the weather forecasting system having the above configuration, even if the prediction calculation is continued after the assimilation of the rainwater amount radar data is completed, it is possible to continue to simulate the rainfall situation in the model. Prediction becomes possible.

  In the weather prediction system having the above-described configuration, it is estimated that the rainfall is observed by the radar = the water vapor is abundant on the spot, and the method of correcting the water vapor data in addition to the rainfall data is adopted. As a result, the effective correction period after completion of data assimilation has been extended to several hours, and the prediction accuracy has been improved. However, in spite of the radar observation range, as shown in FIG. 4, the radio wave is shielded in the area existing behind mountains B1 and B2 with respect to the direction of radio wave emission of weather radar A. Can't catch the area. For this reason, in the conventional radar assimilation technique, correction is performed by assimilation only in an area where radar observation is possible.

  However, as shown in FIG. 4, even if it is an area where radar observation is impossible, if there is rain in the upper layer, it can be estimated that the lower non-assimilable area is also rainy. If the correction by assimilation is performed with the amount of rainwater, it is possible to suppress partial accuracy degradation due to the inability to observe the lower layer.

  The present invention pays attention to this point, and in order to solve the shielding problem, when raindrops exist further above, it is assumed that water vapor / rain also exists in the lower layer, and the water vapor amount is given to the lower layer by a nudge method. A method is proposed.

  The processing flow according to the present invention will be described below with reference to FIG.

  First, at the start of data assimilation, the rain water quantity data (spatial three-dimensional data) obtained by scanning the weather radars in a plurality of hierarchies classified in advance by altitude from the radar site servers DS1 and DS2 shown in FIG. Is stored in the observation data storage unit 13. The number of hierarchies is appropriately selected based on the altitudes of the mountain parts of the radar. Here, three layers are assumed for the sake of simplicity.

  When the start of assimilation is instructed, first, the rainwater amount data obtained in the lower layer scan is fetched (step S31), and the rainwater amount is compared with a threshold value for each observation point to determine whether there is rain (step S32). If it is determined that there is rain, the water vapor assimilation shown in FIG. 3 is performed for all lower layers at that point (step S33).

  If the water vapor assimilation process in step S33 is completed, or if it is determined that there is no rain in step S32, the rainwater volume data of the upper layer is taken in (step S34), and the rainwater volume is compared with the threshold value again. It is determined whether there is any (step S35). If it is determined in step S35 that there is rain, after performing water vapor assimilation on a point below the point where rain has been observed, where water vapor assimilation has never been performed (step S36), step Returning to S34, the rainwater amount data of the next hierarchy is taken in. If it is determined in step S35 that there is no rain, the data assimilation process is completed.

  According to the above process, water vapor assimilation is determined from rainwater volume data in order from the lower layer side, and water vapor assimilation is determined from rainwater volume data of the upper layer at a point where there is no rain. Therefore, water vapor assimilation is performed if there is rain in the upper layer, regardless of whether or not the radar radio wave is shielded at each point. Thus, by virtually providing raindrops that could not be seen due to shielding, raindrops and water vapor can be given in the weather model, prediction of data can be corrected by data assimilation, and improvement in prediction accuracy can be expected.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

1 is a block diagram showing a schematic configuration of a weather prediction system to which the present invention is applied. The flowchart for demonstrating the flow of the data assimilation process of the weather prediction model calculating part of the system of FIG. The flowchart for demonstrating the data assimilation process of the water vapor | steam amount applied to the system of FIG. The figure which shows a mode that a radar radio wave is shielded by the mountain in the system of FIG. 1, and an unobservable area | region is made. The flowchart which shows the process sequence at the time of applying the hierarchical assimilation process which concerns on this invention to the system of FIG.

Explanation of symbols

11: Communication processing unit,
12 ... Communication interface,
13: Observation data storage unit,
14 ... Weather forecast model calculation unit,
15 ... prediction data storage unit,
NT ... Network,
DS0 ... Weather data server,
DS1, DS2 ... Radar site data server.

Claims (6)

A weather prediction system that assimilates rainwater data observed by a weather radar into a weather prediction model, periodically compares the prediction value obtained by the model with the observation value, and interpolates the comparison result into the model. ,
The rain radar data obtained by scanning the meteorological radar in a plurality of hierarchies that have been classified according to altitude in advance is obtained and assimilated into a weather prediction model. An assimilation means that employs assimilation in the upper layer when the assimilation of quantity data is obtained,
In the weather forecast model assimilated by the assimilation means, a determination means for determining a point where the amount of rainwater exceeds a predetermined value;
Steam vapor bogas assimilation means by creating a steam bogus for the point where the amount of rainwater is present above a predetermined value and assimilating in the weather prediction model,
The steam bogas assimilation means compares the water vapor amount expressed in the model before data assimilation with the saturated water vapor amount obtained from the air pressure and the atmospheric pressure at a point where the rainwater amount is greater than or equal to a predetermined value, and performs a nothing processing. A weather forecasting system that is close to saturated water vapor. The nothing processing
qvassim = qvmod * (1-alpha) + qvsat * alpha
qvassim: amount of water vapor after assimilation (unknown amount)
qvmod: Water vapor volume (calculated by model) before assimilation (known volume)
qvsat: Saturated water vapor at the defined point (known amount)
alpha: Weighting parameter (0 <alpha <1)
The meteorological prediction system according to claim 1, wherein qvassim is obtained from the amount of water vapor after assimilation.   In order to execute the nosing process, the search range of rainwater assimilated by radar when performing water vapor amount assimilation, the threshold value of the rain water amount to be searched, and the weighting parameters for water vapor amount assimilation are determined in advance. The weather prediction system according to claim 2, wherein the system is a weather prediction system. It is applied to a weather prediction system that assimilates rainwater data observed by a weather radar into a weather prediction model, periodically compares the prediction values obtained by the model with observation values, and interpolates the comparison results into the model. ,
Hierarchical data acquisition step for acquiring rainwater data by hierarchy obtained by scanning the weather radar in a plurality of hierarchies previously classified by altitude,
An assimilation step by layer for assimilating the rainwater data obtained in each of the plurality of layers into a weather prediction model;
If the assimilation of rainwater data is obtained in the upper layer of the point that was not observed on the lower hierarchy side from the processing result of the assimilation step by hierarchy, an assimilation synthesis step adopting assimilation in the upper layer,
In the weather prediction model assimilated in the assimilation synthesis step, a determination step for determining a point where the amount of rainwater is equal to or greater than a predetermined value;
A steam bogus assimilation step by creating a steam bogus for a point where the amount of rainwater exists above a predetermined value and assimilating it in the weather forecast model,
In the steam bogas assimilation step, the water vapor amount expressed in the model before data assimilation is compared with the saturated water vapor amount obtained from the air temperature and the atmospheric pressure at a point where the rainwater amount is greater than or equal to a predetermined value, and a nosing process is performed. An assimilation method for a weather forecasting system, characterized in that it approaches the saturated water vapor amount. The nothing processing
qvassim = qvmod * (1-alpha) + qvsat * alpha
qvassim: amount of water vapor after assimilation (unknown amount)
qvmod: Water vapor volume (calculated by model) before assimilation (known volume)
qvsat: Saturated water vapor at the defined point (known amount)
alpha: Weighting parameter (0 <alpha <1)
5. The assimilation method for a weather forecasting system according to claim 4, wherein qvassim is obtained from the amount of water vapor after assimilation.   In order to execute the nosing process, when searching for water vapor amount assimilation, the search range of rainwater assimilated by the surrounding radar, the threshold value of the rainwater amount to be searched, and the weighting parameter for water vapor amount assimilation are determined in advance. The assimilation method for a weather prediction system according to claim 5.

Images