TWM608757U - Device for numerical weather prediction - Google Patents

Abstract

A device for numerical weather prediction is applied to a numerical weather prediction program using observation data from a plurality of real stations. The device includes: a wireless network unit, a measuring unit measuring the surrounding weather data by one or more sensing methods to store them in the storage unit, a computing unit connecting the wireless network unit with the storage unit and a power supply unit providing the driving power for all the units. The wireless network unit downloads observation data of the real stations adjacent to the device during an observation period to the calculation unit when the network is found. The meaning of the calculation unit includes: reading the weather data and the geographic location in the storage unit during the observation period; calculating the cross-correlation value of the weather data and the observation data; and reporting the cross-correlation value and the geographic location to the numerical weather prediction program.

Description

Numerical weather forecasting device

This new creation is about a numerical weather forecasting device; in particular, it is about a numerical weather forecasting device. The method can correct the false correlation errors in the numerical weather forecast, especially the correction based on actual empirical values.

Numerical Weather Predication (NWP) is a method of predicting weather changes, which is based on atmospheric dynamics and thermal principles, applying momentum equations, continuity equations, ideal gas equations, equations of the first law of thermodynamics, and changes in moisture content Equations to establish a basic atmospheric model; finally, a large high-speed computer is used to substitute the initial conditions of the observations to replace the unknowns in these equations: solar radiation, evapotranspiration, and air temperature. ), relative humidity, precipitation, air pressure, wind speed and wind direction, changes over time are predicted.

The differential equation is a non-linear equation, which represents that the numerical weather prediction (NWP) is a complete chaos system, which relies heavily on the accuracy of the initial parameters, especially for short-term weather forecasts with a forecast time limit of only 1-3 days. -term weather forecast) or even very short-term weather forecast (very short-term weather forecast). Whether it is conventional observation or radar reflectivity, both Doppler velocity or small errors in the conversion of meteorological satellite signals to initial parameters can cause this chaotic system to produce more than ten times the error growth.

The initial parameter error does not mainly come from the instrument itself, but from the position between the actual observation point and the grid and the instrument calibration gap. The grid can be regarded as a virtual observation point in the NWP, and its position is roughly uniformly distributed and different from the distribution of the actual observation station. Therefore, the actual observations must be assimilated before they can be analogized to the initial parameters on the grid and then numerically calculated. When the gap between the grid position and the actual observation point is greater, the instrument difference is more or the calibration procedure is more complicated, the initial parameters on the grid will have a greater error. Although in order to reduce the sensitivity of NWP to the initial parameters, forecasters began to use ensemble forecasts that examine the probability densities of different initial parameters instead of single forecasts with the most likely weather: It is impossible to avoid forecast gaps caused by spatially correlated initial parameter errors. The research results show that as long as the initial parameters have a kilometer scale (kilometer scale) horizontal spatial correlation (horizontally spatial correlation), or a sub-kilometer scale (sub-kilometer scale) vertical spatial correlation (vertically spatial correlation) error, even if it is Weak intensity can also cause severe weather forecast results above the kilometer level, which are significantly different from actual observation results.

Known numerical weather prediction methods can only add white noise to weaken the influence of spatial correlation errors on NWP false alarms, but ignore that the blank areas of the observation data have strong spatial correlation during data assimilation. The spatial correlation of this blank area will be as great as the sky has a large area of white clouds the same size as this blank area. It is the same reason that it will have a very strong impact on NWP.

Therefore, it is necessary to develop a numerical weather forecasting device that can provide reduced spatial correlation. In addition to overcoming the false correlation error caused by the blank area of the observation data, it can further correct the correlation with the actual experience value to present a more accurate weather forecast.

A numerical weather forecasting device created by the present invention is applied to a numerical weather forecasting program using observation data from a plurality of real observatories. It includes: a geographic location detecting unit capable of providing the geographic location of the device and storing it in a storage unit , A wireless network unit, a measurement unit measure the surrounding weather data by one or more sensing methods and store it in the storage unit, a computing unit connects the wireless network unit with the storage unit, and a power supply unit to supply the The driving power of other units. The wireless network unit downloads the observation data of the real observation station adjacent to the device during an observation period to the calculation unit when searching for the network. The meaning of the calculation unit includes: reading the weather data and the geographic location during the observation period in the storage unit; calculating the cross-correlation value (cross-correlation) of the weather data and the observation data; and through the wireless The network unit reports the cross-correlation value and the geographic location on the network.

The effect of the new creation can provide a numerical weather forecast device that can solve false correlation errors, and does not need to modify the source code of the numerical weather forecast program, and can also reduce the probability of false alarms in severe weather.

A: Numerical weather forecasting device

c ₆ -c ₇ : Supply station of _{H 2}

c ₈ -c ₉ : Supply station of _{H 1}

H ₁ : Geographical location during the first period

H ₂ : Geographical location during the second period

I: Open area

K ₁ : Observation data of S _{20 in the first period}

K ₂ : Observation data of S _{20 in the second period}

L: The route of the ship

S ₂₀ -S ₂₅ : Reference stations in open areas

V _H : Cross correlation value between _{S 20} and Z ₂

V _L : Cross correlation value between _{S 20} and Z ₁

Z ₁ : Temporary data during the first period

Z ₂ : Temporary data during the second period

10: Geographic location detection unit

20: Wireless network unit

30: Measuring unit

31: Anemometer

40: Computing unit

50: storage unit

60: power supply unit

70: Satellite communication antenna

Figure 1 is a schematic diagram of the numerical weather forecast method and device used in this creation

Figure 2 is a schematic diagram of the numerical weather forecast device created by this book

Since real observation stations cannot be distributed evenly like the grid in the numerical weather forecast (NWP) program, the observation data must be assimilated by data first. Only then can it become the initial condition of the grid in the program, and then start the numerical weather forecast program. If the relative distance of the actual observation station is greater than or close to the grid spacing, then some grids will share approximate initial values through data assimilation. These spatially relevant initial values are called in this content False correlation errors may affect the accuracy of NWP forecasts, especially for severe weather forecasts.

Please refer to "Figure 1", a numerical weather forecasting device A created by this new model, which is applied to a numerical weather forecasting program using observation data from multiple real observatories. Taking the numerical weather forecasting device A on a moving object in an open area I as an example, the plural real observation stations adjacent to the open area have six reference stations S ₂₀ -S ₂₅ , and the numerical weather forecasting device A is on the moving object _{On the walking path L, two temporary data Z 1} and Z ₂ will be generated in a first period and a second period, respectively, for calculating the cross-correlation value between the observation data of the six reference stations S ₂₀ -S _{25 (} cross-correlation) and recorded as twelve expected values. The twelve expected values will be compared with the cross-correlation values of the observation data between the six reference stations (a total of fifteen empirical values) to find one to several of the _{six reference stations S 20} -S ₂₅ The value that meets the expected value is the supply station c ₆ -c ₉ . The expected value of the reference station S ₂₀ and the temporary data Z _{1 is V L} , and the five empirical values (solid squares in the figure) between the reference station S ₂₀ and the other reference stations S ₂₁ -S _{25 can be compared with the reference} station S _21, S ₂₄ experience value and the desired value V _L of approximation, it will be S _21, S ₂₄ as a supply station c _8, c _9. The expected value of the reference station S ₂₀ and the temporary data Z _{2 is V H} , and the five empirical values (solid squares in the figure) between the reference station S ₂₀ and the other reference stations S ₂₁ -S _{25 can be compared with the reference} The empirical values of stations S ₂₂ and S _{23 are similar to the expected value V H} , so S ₂₂ and S _{23 are} listed as supply stations c ₆ and c ₇ .

Since the locations H ₁ and H _{2 of} the temporary data are close to the other reference station S ₂₄ , the smallest absolute value is selected from the cross-correlation values of _{S 24} and S ₂₄ and S ₂₄ and S _{21 , that is, S 21} Observation data is used as the virtual data of the virtual station H ₁ ; and the smallest absolute value is taken from the empirical values of _{S 24} and S ₂₂ and S ₂₄ and S ₂₃ , for example: make the observation data of _{S 23 do} the virtual station for a H ₂ of the dummy data. Finally, the two virtual stations are connected to the observation network formed by the plurality of real observation stations, so that the virtual data and the observation data are used together for the numerical weather prediction program to perform data assimilation (assimilation).

In this way, even if the _{time provided by the temporary data Z 1} and Z ₂ is very short and is not sufficient for the numerical weather prediction program, it is still possible to select approximate observation stations through spatial correlation data to provide empirical data and solve false correlations. The impact of sexual error on the numerical weather forecasting program. To make the empirical data provided more reliable, the frequency and location of the temporary data can be increased, and the number of supply stations can also be increased.

Please refer to "Figure 2", a numerical weather forecasting device A created by this new model is applied to a numerical weather forecasting program that uses observation data from multiple real observation stations. It includes: a geographic location detection unit 10, such as a GPS component, which can provide geographic locations H ₁ and H ₂ where the device is located and store them in a storage unit 50, a wireless network unit 20, and a measurement unit 30 The surrounding weather data is measured by one or more sensing methods and stored in the storage unit, a computing unit 40 is connected to the wireless network unit 20 and the storage unit 50, and a power supply unit 60 supplies the driving power (power line) of the units. Not shown). The wireless network unit 20 can be a WiFi network chip, a satellite communication chip or other wireless network standard chips. In order to enhance the communication strength, an external antenna created by the present invention, such as an external satellite communication antenna 70 for a satellite communication chip. The measuring unit 30 may include: a thermometer for sensing ambient temperature data, a barometer for measuring ambient atmospheric pressure data, and a hygrometer for measuring ambient relative humidity data. Or even an external anemometer 31 for measuring the surrounding wind data; an external rain gauge for measuring the accumulated rainfall; and an external solar radiometer for measuring the radiation intensity.

Take a ship sailing in the open area I as an example. The ship is equipped with the numerical weather forecasting device A, and travels along its route L via H ₁ , H during the first period and the second period, respectively. ₂ two points. When the wireless unit 20 to the network to search the web, download the reference station ₂₀ S K observations in the first period and observations ₁ K ₂ in the second period to the calculation unit 30 for calculating the The unit 30 respectively calculates the cross-correlation value of the weather data in the first period and the weather data in the second period with the observation data K ₁ in the first period and the observation data K _{2 in the second period.} The calculation unit 30 then reports the cross-correlation value and the geographic locations H ₁ , H ₂ on the network through the wireless network unit 20. If the cross-correlation value based on the complete network communication section, the cross-correlation value and the location of these H _1, H ₂ to obtain the web of the reference station ₂₀ S, S and the reference station ₂₀ of the data observed Report to the data processing center of the numerical weather forecast program together.

In one embodiment, the numerical weather forecasting device can be installed on a weather balloon and connected to a geostationary satellite through the wireless network unit such as a satellite communication chip to calculate the weather at different times and in different regions. The cross-correlation value between the data and the reference station is sent back to the numerical weather forecast program to create the virtual station and its virtual data. In this way, this creation can download neighboring observation data through the real-time or non-real-time network on the measurement side and directly perform the cross-correlation calculation on the measurement side, and upload the cross-correlation value when the network is connected again. It does not occupy network bandwidth and does not need to surf the Internet at any time. It can still provide precious spatial correlation data to the numerical weather forecast program to improve the accuracy of the numerical weather forecast.

To sum up, the numerical weather forecast method and device created by this author are to solve the false correlation error of the numerical weather forecast instantly and quickly at the lowest cost. There is no need to occupy the network bandwidth or change the original numerical weather forecast program. It has indeed met the requirements of the creation patent application, and the patent application is filed in accordance with the law. However, the above are only the preferred embodiments of this creation, and should not be used to limit the scope of implementation of this creation; therefore, all simple equivalent changes and modifications made in accordance with the scope of the patent application for this creation and the content of the creation specification, All should still fall within the scope of this creation patent.

10‧‧‧Geographic location detection unit

20‧‧‧Wireless Network Unit

30‧‧‧Measuring unit

31‧‧‧Anemometer

40‧‧‧Computer unit

50‧‧‧Storage unit

60‧‧‧Power Supply Unit

70‧‧‧Satellite communication antenna

Claims (3)

A numerical weather forecasting device, applied to a numerical weather forecasting program using observation data from a plurality of real observatories, comprising: a geographic location detecting unit capable of providing the geographic location of the numerical weather forecasting device and storing it in a storage Unit; a wireless network unit that can download the observation data of the real observation station adjacent to the device during an observation period to a calculation unit when searching for the network; a measurement unit for one or more sensing methods The surrounding weather data is measured and stored in the storage unit; and a power supply unit supplies the driving power of the units, wherein the meaning of the calculation unit includes: reading the weather in the storage unit during the observation period Data and the geographic location; calculate the cross-correlation value of the weather data and the observation data; and report the cross-correlation value and the geographic location on the network through the wireless network unit. The numerical weather forecast device according to claim 1, wherein the geographic location detection unit and the measurement unit respectively store the geographic location and the weather data according to the period during which the geographic location is located through the calculation unit. The numerical weather forecast device according to claim 1, further comprising: a waterproof and dustproof housing for accommodating the above-mentioned units, and a conductive head for electrically connecting a power socket and the power supply unit.

Images