A wind profiler radar wind field data quality control method for data assimilation
The proposed quality control method for wind profiler radar data, using monthly averages and hourly rates, addresses inefficiencies in existing methods by improving data retention and elimination, enhancing the accuracy of numerical weather models.
Patent Information
- Application Number
- CN202110196258.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-02-22
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2041-02-22
AI Technical Summary
Existing wind profiler radar wind field data quality control methods, such as the effective detection height, empirical orthogonal function (EOF) analysis, and high confidence interval methods, fail to effectively account for variations in radar performance, weather conditions, and data acquisition rates, leading to inefficient data utilization and negative impacts on numerical weather forecasts.
A quality control method that incorporates monthly average standard deviation, monthly average correlation coefficient, and hourly data acquisition rate to determine high confidence intervals for wind profiler radar data, enhancing data retention and elimination criteria.
Improves the efficiency of wind profiler radar data quality control by reducing unnecessary data loss and eliminating poor-quality data, thereby enhancing the accuracy and reliability of numerical weather models.
Smart Images

Figure CN112859083B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of atmospheric sounding data processing, and in particular relates to a wind profiler radar wind field data quality control method oriented to data assimilation. Background Art
[0002] Wind profiler radar uses the Doppler shift principle to measure the speed at which the turbulent atmosphere moves closer or farther from the radar by measuring the change in frequency between the electromagnetic waves reflected from the clear sky turbulent atmosphere and the transmitted electromagnetic waves. This speed is called radial velocity. Wind profiler radar usually needs to measure the radial velocity on one zenith vertical beam and two inclined beams (three beams) or the radial velocity of each distance bin on one zenith vertical beam and four inclined beams in the east, south, west and north (five beams) in sequence. Since the radial velocity is not the true vector velocity of the turbulent atmosphere, in order to facilitate application, the wind profiler radar provides users with inverted wind profiler radar horizontal wind products and vertical wind products through certain software algorithms: that is, under the assumption that the atmospheric wind field is horizontally uniform, the radial velocity measurements in the directions of three beams or five beams at the same height are used to jointly solve the atmospheric three-dimensional wind field. It can be seen that the horizontal wind field (i.e., u and v wind fields) data retrieved by wind profiler radar has high temporal and spatial resolution, which can show the continuous and detailed change process of weather systems such as atmospheric fluctuations, greatly compensating for the shortcomings of insufficient time and spatial density of conventional sounding observations.
[0003] However, the quality of horizontal wind field data retrieved by wind profiler radar will be seriously affected by many factors, such as the accuracy of radial velocity observations of wind profiler radar, electromagnetic interference, ground echo interference, uniform wind assumptions, and different weather systems. If it is directly applied to numerical weather models without quality control, it will have a serious negative impact on the results of numerical weather forecasts. In order to overcome this difficulty, domestic and foreign researchers have developed wind profiler radar wind field data quality control technology for data assimilation, thereby eliminating bad wind profiler radar wind field data. At present, there are three main quality control methods for data assimilation used at home and abroad. One is the effective detection height quality control method proposed by Zhu Lijuan et al. (see "Preliminary Analysis of the Quality of Horizontal Wind Products of Wind Profiler Radar for Data Assimilation", "Meteorology", 2015); one is the empirical orthogonal function (EOF) analysis quality control method proposed by Zhang Xubin et al. (see "Quality Control of Wind Profiler Radar Data and Its Assimilation Application", "Acta Meteorologica Sinica", 2015); and the other is the high confidence interval quality control method proposed by Wang Yehong et al. (see "Analysis of Wind Field Characteristics of Wind Profiler Radar for Data Assimilation and Its Quality Control Methods", "Atmospheric Science", 2021).
[0004] Among them, first, the limitations of the effective detection height quality control method. Using the effective detection height quality control method, the effective detection heights of the PA, PB model tropospheric wind profiler radar and the LC model boundary layer wind profiler radar are set to 14km, 8km and 3km respectively, and the horizontal wind field data of the wind profiler radar above the effective detection height are eliminated. This method has three main disadvantages:
[0005] 1) Since the effective detection height of wind profiler radar is affected by many factors including wind profiler radar performance parameters, transmission power, electromagnetic interference, weather system, etc., the effective detection heights of wind profiler radars of the same model located at different sites are not exactly the same, and may even differ greatly. The effective detection heights of wind profiler radars of different models such as PA, PB and LC obtained by empirical or averaging methods are used to control the quality of horizontal wind field data of wind profiler radars, and the data cannot be effectively retained or eliminated, which directly affects the efficiency of data use;
[0006] 2) With the further improvement of my country's wind profiler radar network, the assimilation and application of wind profiler radar data in numerical weather models has become more and more urgent. Using the fixed effective detection height quality control method corresponding to a specific model of radar, the actual effective detection height of some radars is higher or lower than the given effective detection height, which will inevitably cause some good data to be eliminated and some bad data to be retained. The former will cause data waste, and the latter will have a negative impact in the numerical weather model, which to some extent affects the effectiveness of the construction of the wind profiler radar observation network;
[0007] 3) Precipitation is an important weather system factor that affects the quality of the horizontal wind field of the wind profiler radar. This method does not consider the quality control scheme when there is precipitation, which directly affects the effect of quality control.
[0008] 2. Limitations of the empirical orthogonal function (EOF) analysis quality control method. The empirical orthogonal function (EOF) analysis quality control method is used. First, the real-time sampling u and v wind field data at intervals of 6 minutes before and after the analysis time are used to form a time series, and the time series is decomposed by EOF to obtain the main modes of the vertical distribution of the wind field during this period; then, the main modes obtained by EOF decomposition are used to reconstruct the time series within this period to obtain the vertical distribution of the wind field at the analysis time. This method has three main disadvantages:
[0009] 1) Theoretically, the EOF analysis method can be used to remove outliers and pulsation noise in observation data. However, when the observation data is always abnormal at a certain altitude and within the analysis time window, the EOF analysis will regard it as the main mode and retain it, which directly affects the use of the data in the numerical weather model.
[0010] 2) In this method, the u and v wind field data that constitute the time series must have values at the same vertical level. Therefore, if the wind field data at a certain height layer at a certain moment is not inverted, the data at that height layer in the time series will be eliminated. This method will greatly reduce the efficiency of data use;
[0011] 3) Precipitation is an important weather system factor that affects the quality of the horizontal wind field of the wind profiler radar. This method also does not consider the quality control plan when there is precipitation, which directly affects the effect of quality control.
[0012] 3. Limitations of the high confidence interval quality control method. In response to the above five limitations of the effective detection height quality control method and the empirical orthogonal function (EOF) analysis quality control method, Wang Yehong et al. (2021) proposed a "high confidence interval quality control method". Different wind profiler radars use the height layer where the annual average standard deviation of their respective u and v wind fields is less than 4.5m / s as the high confidence interval of the wind profiler radar. The data outside the high confidence interval is eliminated, and the quality control plan when there is precipitation is considered at the same time, thereby avoiding the limitations of the effective detection height quality control method and the empirical orthogonal function (EOF) analysis quality control method. However, in reality, the standard deviation of the u and v wind fields of the wind profiler radar has obvious seasonal and monthly variation characteristics. The high confidence interval determined based on the annual average standard deviation of the u and v wind fields has a certain deviation from the actual situation, so the wind profiler radar data cannot achieve the best utilization efficiency. Secondly, the wind field in the lower layer of the boundary layer detected by the wind profiler radar is often within the high confidence interval determined by the annual average standard deviation of the u and v wind fields. In fact, the wind field in the lower layer of the boundary layer detected by the wind profiler radar may be inconsistent with the actual wind direction or even in the opposite direction. However, because the wind speed in the boundary layer is relatively small, the standard deviation has not reached the threshold of 4.5m / s. Therefore, the wind field in the lower layer of the boundary layer cannot be well controlled, which will further affect the effect of its assimilation application in the numerical weather model. In addition, it is found in statistical studies that the hourly data acquisition rate of the wind profiler radar has great volatility over time. When the hourly data acquisition rate of the wind profiler radar is low, the quality of the u and v wind fields is often relatively poor, but it cannot be eliminated by the standard deviation or correlation coefficient method. Therefore, the hourly data acquisition rate is also an important factor in judging the quality of the u and v wind fields of the wind profiler radar. Therefore, in actual business applications, it is necessary to consider the wind profiler radar horizontal wind field quality control method that combines the monthly average standard deviation of the u and v winds, the monthly average correlation coefficient and the hourly data acquisition rate. Summary of the invention
[0013] In view of the shortcomings in the prior art, the purpose of the present invention is to provide a wind profiler radar wind field data quality control method for data assimilation, to improve and perfect the high confidence interval quality control method, to increase the consideration of the monthly average standard deviation of the u and v wind fields, the monthly average correlation coefficient and the hourly data acquisition rate, thereby improving the efficiency of the wind profiler radar horizontal wind data quality control.
[0014] In order to achieve the above object, the technical solution of the present invention is as follows:
[0015] A method for controlling wind field data quality of wind profiler radar for data assimilation comprises the following steps:
[0016] S1, monthly average standard deviation of the u and v wind fields of the statistical wind profiler radar relative to the background wind field of the large-scale model;
[0017] S2, monthly average correlation coefficient of the u and v wind fields of the statistical wind profiler radar relative to the background wind field of the large-scale model;
[0018] S3. Statistical annual average standard deviation of wind profiler radar u and v wind fields relative to large-scale model background wind fields under precipitation conditions;
[0019] S4. Determination of high confidence interval of wind profile radar based on monthly average standard deviation and monthly average correlation coefficient;
[0020] S5. Determination of high confidence intervals of wind profiler radar u and v wind fields under precipitation conditions;
[0021] S6. Calculate the hourly data acquisition rate of the wind profiler radar;
[0022] S7. Wind profile radar u, v wind field quality control.
[0023] Furthermore, in step S1, it is necessary to statistically calculate the monthly average standard deviation of each wind profile radar u, v wind field relative to the large-scale model background wind field at each vertical level. The calculation method of the monthly average standard deviation is specifically as follows:
[0024] Assume that there are i = 1, 2, ..., m wind profiler radars; j = 1, 2, ..., l represents the vertical level; t = 1, 2, ..., n represents the observation time; k = 1, 2, ..., 12 represents the month. The calculation formula is as follows:
[0025]
[0026] In the formula, represents the monthly mean standard deviation of u and v winds of wind profiler radar i at vertical level j in month k;
[0027] represents the hourly averaged u and v winds sampled by wind profiler radar i at vertical level j at time t in month k;
[0028] Represents the u and v winds of the large-scale model background interpolated to the u and v winds at the geographical location of wind profiler i and the vertical level j at time t in month k.
[0029] Furthermore, in step S2, it is necessary to statistically calculate the monthly average correlation coefficient of each wind profile radar u, v wind field relative to the large-scale model background wind field at each vertical level. The calculation method of the monthly average correlation coefficient is specifically as follows:
[0030]
[0031] In the formula, represents the average monthly correlation coefficient of wind u and v for wind profiler radar i at vertical level j in month k;
[0032] represents the average value of u and v winds of wind profiler radar i at vertical level j in month k;
[0033] Represents the large-scale model background u, v wind interpolation to the geographical location of wind profiler i and the average u, v wind value of month k at vertical level j.
[0034] Furthermore, in step S3, it is necessary to statistically calculate the annual average standard deviation of the wind field of each wind profile radar u and v at each vertical level relative to the large-scale model background wind field under precipitation conditions. The specific calculation method of the annual average standard deviation is:
[0035]
[0036] In the formula, represents the annual average standard deviation of wind u and v at vertical level j of wind profiler radar i in the case of precipitation;
[0037] represents the hourly average sampled u and v winds at time t when there is precipitation at vertical level j of wind profiler radar i;
[0038] Represents the large-scale model background u, v winds interpolated to the geographical location of wind profiler i and vertical level j at time t when there is precipitation.
[0039] Furthermore, in step S4, the monthly average standard deviation and monthly average correlation coefficient As the threshold, the high confidence interval of the wind farm is determined, which is specifically expressed as:
[0040]
[0041] In the formula, Represents the high confidence interval of the u and v wind fields of wind profiler radar i in month k.
[0042] Furthermore, in step S5, the annual average standard deviation under the condition of precipitation is As the threshold, the high confidence interval of the wind farm is determined, which is specifically expressed as:
[0043]
[0044] In the formula, Represents the high confidence interval of the u and v wind fields of wind profiler radar i in the presence of precipitation.
[0045] Furthermore, the step S6 is specifically as follows: for the hourly average sampled u and v winds of the wind profile radar i at time t, the data acquisition rate at that time is characterized by the ratio of the amount of wind field data actually acquired to the total amount of data that should be acquired.
[0046] Furthermore, the step S7 is specifically:
[0047] (1) For the hourly averaged wind u and v sampled by wind profiler radar i at vertical level j at time t in month k, if vertical level j is If they are within the range, they are retained, otherwise they are discarded;
[0048] (2) If there is precipitation at wind profiler radar i 3 hours before time t, then if vertical level j is If they are within the range, they are retained, otherwise they are discarded;
[0049] (3) If the data acquisition rate of the hourly average sampling of u and v winds by wind profiler radar i at time t is less than 0.3, all data at that time will be discarded.
[0050] Beneficial effects: Under the differentiated quality control scheme, the present invention adopts the monthly average standard deviation of the wind field of wind profile radar u and v as the quality control criterion, and adds the monthly average correlation coefficient and hourly data acquisition rate as the quality control criterion, thereby improving the quality control effect of wind profile radar data. It is convenient for wind field data of different wind profile radar stations to be more fully and effectively identified, which not only reduces unnecessary loss of radar data, but also further eliminates poor quality data. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] Figure 1 A flow chart of a method for quality control of wind field data of wind profiler radar for data assimilation provided by the present invention;
[0052] Figure 2 Schematic diagram of the distribution of the monthly average standard deviation of the wind field u(2a) and v(2b) of the Jianning wind profile radar relative to the background wind field of the large-scale model with respect to the height in an embodiment of the present invention;
[0053] Figure 3 Schematic diagram of the distribution of the monthly average correlation coefficients of the wind fields u(3a) and v(3b) of the Jianning wind profile radar relative to the background wind field of the large-scale model with respect to the height in an embodiment of the present invention;
[0054] Figure 4 Schematic diagram of the annual average standard deviation distribution of the wind fields u(4a) and v(4b) of the Jianning wind profile radar relative to the background wind field of the large-scale model in an embodiment of the present invention;
[0055] Figure 5 The figure is a distribution evolution diagram of the acquisition rate of Jianning wind profile radar data calculated over time taking 2017 as an example in an embodiment of the present invention;
[0056] Figure 6 This is a standard deviation distribution diagram of the u wind and v wind of the Jianning wind profile radar before and after step-by-step quality control in an embodiment of the present invention. DETAILED DESCRIPTION
[0057] The present invention is described below with reference to specific examples. It will be appreciated by those skilled in the art that these examples are only used to illustrate the present invention and are not intended to limit the scope of the present invention in any way.
[0058] Taking the LC wind profiler radar in Jianning, Fujian Province from January to December 2017 as an example, Figure 1 A wind profiler radar wind field data quality control method for data assimilation is shown and an example is given to illustrate it.
[0059] The method comprises the following steps:
[0060] S1, monthly average standard deviation of the u and v wind fields of the statistical wind profiler radar relative to the background wind field of the large-scale model;
[0061] In this embodiment, there are i=1 wind profiler radars; j=1, 2, ..., 59 represents vertical levels; t=1, 2, ..., n represents observation time (when k=1, 3, 5, 7, 8, 10, 12 months, n=124; when k=2 months, n=112; when k=4, 6, 9, 11 months, n=120); k=1, 2, ..., 12 represents months, and is substituted into the following formula group (1) to calculate the monthly average standard deviation of the u and v wind fields of the Jianning LC wind profiler radar relative to the large-scale model background wind field at each vertical level from January to December 2017:
[0062]
[0063] In the formula, represents the monthly mean standard deviation of u and v winds of wind profiler radar i at vertical level j in month k; represents the hourly averaged u and v winds sampled by wind profiler radar i at vertical level j at time t in month k; represents the large-scale model background u, v wind interpolated to the u, v wind at the kth month t time on the vertical level j at the geographical location of the wind profiler i. The calculation results are expressed as Figure 2 a and 2b show the distribution of the monthly average standard deviation of the u wind and v wind of the Jianning LC wind profile radar relative to the large-scale model background wind field. In the figure, the hollow circle represents a standard deviation ≤4.5m / s; the solid circle represents a standard deviation >4.5m / s; the larger the circle, the larger the monthly average standard deviation; the shaded area represents the interval determined by the annual average standard deviation ≤4.5m / s.
[0064] S2, monthly average correlation coefficient of the u and v wind fields of the statistical wind profiler radar relative to the background wind field of the large-scale model;
[0065] In this embodiment, there are i=1 wind profiler radars; j=1, 2, ..., 59 represents vertical levels; t=1, 2, ..., n represents observation time (when k=1, 3, 5, 7, 8, 10, 12 months, n=124; when k=2 months, n=112; when k=4, 6, 9, 11 months, n=120); k=1, 2, ..., 12 represents months, and is substituted into the following formula group (2) to calculate the monthly average correlation coefficients of the u and v wind fields of the Jianning LC wind profiler radar relative to the large-scale model background wind field at each vertical level from January to December 2017;
[0066]
[0067]
[0068] In the formula, represents the average monthly correlation coefficient of wind u and v for wind profiler radar i at vertical level j in month k; represents the average value of u and v winds of wind profiler radar i at vertical level j in month k; represents the average value of the u and v winds of the large-scale model background interpolated to the geographical location of the wind profiler radar i and the vertical level j for month k. The calculation results are expressed as Figure 3 a and 3b show the distribution of the monthly average correlation coefficients of the u wind and v wind of the Jianning LC wind profile radar relative to the large-scale model background wind field with height; at the same time, in the figure, the hollow circle represents a correlation coefficient ≥ 0.7; the solid circle represents a correlation coefficient < 0.7; the larger the circle, the smaller the correlation coefficient.
[0069] S3. Statistical annual average standard deviation of wind profiler radar u and v wind fields relative to large-scale model background wind fields under precipitation conditions;
[0070] In this embodiment, there are i=1 wind profiler radars; j=1,2,…,59 represents the vertical level; t=1,2,…,1460 represents the observation time, which is substituted into the following formula group (3) to calculate the annual average standard deviation of the u and v wind fields of the Jianning LC wind profiler radar relative to the large-scale model background wind field at each vertical level in 2017 under the precipitation condition;
[0071]
[0072] In the formula, represents the annual average standard deviation of wind u and v at vertical level j of wind profiler radar i in the case of precipitation; represents the hourly average sampled u and v winds at time t when there is precipitation at vertical level j of wind profiler radar i; represents the large-scale model background u, v wind interpolated to the geographical location of wind profiler i and vertical level j at time t when there is precipitation. The calculation results are expressed as Figure 4 a and 4b show the annual average standard deviation of the u and v wind fields of the Jianning LC wind profiler radar relative to the large-scale model background wind field. In the figure, the solid line represents the case with precipitation; the short dashed line represents the case without precipitation; and the long dashed line represents all samples.
[0073] S4. Determination of high confidence interval of wind profile radar based on monthly average standard deviation and monthly average correlation coefficient; Figure 2 and Figure 3 The height layers shown are all hollow circles, which are based on the monthly average standard deviation threshold of 4.5m / s ( Figure 2 ) and the monthly average correlation coefficient threshold is 0.7 ( Figure 3 ) The high confidence intervals of the u and v wind fields of the Jianning LC wind profiler radar were determined by the LC wind profiler, and the specific height layers are shown in Table 1.
[0074] Table 1 High confidence intervals of Jianning LC wind profile radar based on the monthly average standard deviation (threshold of 4.5 m / s) and the monthly average correlation coefficient (threshold of 0.7)
[0075]
[0076] S5. Determination of high confidence intervals of wind profiler radar u and v wind fields under precipitation conditions;
[0077] In this embodiment, based on Figure 4The statistical distribution characteristics of the annual average standard deviation of the u and v wind fields at the Jianning wind profile radar station with precipitation are obtained with the annual average standard deviation ≤ 4.5m / s as the threshold to obtain the high confidence high interval of the u and v wind fields at the Jianning radar station, as shown in Table 2.
[0078] Table 2 High confidence intervals of u and v wind fields of Jianning wind profile radar under precipitation conditions
[0079] Wind profiler radar station u Wind height confidence interval (m) v Wind height confidence interval (m) Jianning 870—6750 870—9870
[0080] S6, calculate the hourly data acquisition rate of the wind profile radar; for the hourly average sampled u and v winds of wind profile radar i at time t, the data acquisition rate of this time is characterized by the ratio of the actual amount of wind field data obtained to the total amount of data that should be obtained. In this embodiment, Figure 5 The distribution evolution of Jianning wind profile radar data acquisition rate over time is calculated based on 2017. After investigation and research, 0.3 is selected as the threshold of data acquisition rate, that is, when the data acquisition rate at a certain moment is less than 0.3, all data at that moment are eliminated. Figure 5 Figure 2 shows the temporal evolution of the data acquisition rate of the Jianning wind profiler radar station in 2017.
[0081] S7. Wind profiler radar u and v wind field quality control. (1) For the hourly average sampled u and v winds of wind profiler radar i (i = 1) at vertical level j (j = 1, 2, ..., 59) at time t (t = 1, 2, ..., n, when k = 1, 3, 5, 7, 8, 10, 12 months, n = 124; when k = 2 months, n = 112; when k = 4, 6, 9, 11 months, n = 120), if vertical level j is If the wind profiler radar i (i = 1) has precipitation 3 hours before time t (t = 1, 2, ..., 1460), then if the vertical level j is within If the data is within the range, it is retained, otherwise it is discarded; (3) If the hourly average sampling u and v wind data acquisition rate of wind profile radar i (i = 1) at time t (t = 1, 2, ..., 1460) is less than 0.3, all data at that time are discarded. The result is as follows Figure 6The figure shows the standard deviation distribution of u wind (first row) and v wind (second row) of Jianning wind profile radar before and after step-by-step quality control in 2017. The solid line is before quality control; the thin short dash line is after quality control of monthly average standard deviation (threshold is 4.5m / s); the thin long dash line is after quality control of monthly average correlation coefficient (threshold is 0.7) on the basis of the thin short dash line; the thick short dash line is after quality control of annual average standard deviation (threshold is 4.5m / s) under precipitation on the basis of the thin long dash line; the thick long dash line is after quality control of hourly data acquisition rate (threshold is 0.3) on the basis of the thick short dash line.
Claims
1. A wind profiler radar wind field data quality control method for data assimilation, characterized in that: The following steps are involved: S1, monthly average standard deviation of the u and v wind fields of the statistical wind profiler radar relative to the background wind field of the large-scale model; S2, monthly average correlation coefficient of the u and v wind fields of the statistical wind profiler radar relative to the background wind field of the large-scale model; S3. Statistical annual average standard deviation of wind profiler radar u and v wind fields relative to large-scale model background wind fields under precipitation conditions; S4. Determination of high confidence interval of wind profile radar based on monthly average standard deviation and monthly average correlation coefficient; S5. Determination of high confidence intervals of wind profiler radar u and v wind fields under precipitation conditions; S6. Calculate the hourly data acquisition rate of the wind profiler radar; S7. Wind profiler radar u, v wind field quality control based on the high confidence interval of the wind profiler radar based on the monthly average standard deviation and the monthly average correlation coefficient, the high confidence interval of the wind profiler radar u, v wind field under precipitation conditions, and the wind profiler radar hourly data acquisition rate, specifically: (1) For the hourly averaged wind u and v sampled by wind profiler radar i at vertical level j at time t in month k, if vertical level j is If the value is within the range, it will be retained, otherwise it will be removed. represents the high confidence interval of the u and v wind fields of wind profiler radar i in month k; (2) If there is precipitation at wind profiler radar i 3 hours before time t, then if vertical level j is If the value is within the range, it will be retained, otherwise it will be removed. Represents the high confidence interval of the u and v wind fields of wind profiler radar i under precipitation conditions; (3) If the data acquisition rate of the hourly average sampling of u and v winds by wind profiler radar i at time t is less than 0.3, all data at that time will be discarded.
2. The method for controlling wind field data quality of wind profiler radar for data assimilation according to claim 1, characterized in that: In step S1, it is necessary to statistically calculate the monthly average standard deviation of the wind field of each wind profiler radar u and v relative to the large-scale model background wind field at each vertical level. The calculation method of the monthly average standard deviation is specifically as follows: assuming that there are i=1, 2, ..., m wind profiler radars; j=1, 2, ..., l represents the vertical level; t=1, 2, ..., n represents the observation time; k=1, 2, ..., 12 represents the month, and the calculation formula is as follows: In the formula, represents the monthly mean standard deviation of u and v winds of wind profiler radar i at vertical level j in month k; represents the hourly averaged u and v winds sampled by wind profiler radar i at vertical level j at time t in month k; Represents the u and v winds of the large-scale model background interpolated to the u and v winds at the geographical location of wind profiler i and the vertical level j at time t in month k.
3. The method for controlling wind field data quality of wind profiler radar for data assimilation according to claim 1, characterized in that: In step S2, it is necessary to statistically calculate the monthly average correlation coefficient of each wind profile radar u, v wind field relative to the large-scale model background wind field at each vertical level. The calculation method of the monthly average correlation coefficient is specifically as follows: In the formula, represents the average monthly correlation coefficient of wind u and v for wind profiler radar i at vertical level j in month k; represents the hourly averaged u and v winds sampled by wind profiler radar i at vertical level j at time t in month k; represents the u and v winds of the large-scale model background interpolated to the u and v winds at the geographical location of the wind profiler radar i and the vertical level j at time t in month k; represents the average value of u and v winds of wind profiler radar i at vertical level j in month k; Represents the large-scale model background u, v wind interpolation to the geographical location of wind profiler i and the average u, v wind value of month k at vertical level j.
4. The method for controlling wind field data quality of wind profiler radar for data assimilation according to claim 1, characterized in that: In step S3, it is necessary to statistically calculate the annual average standard deviation of the wind field of each wind profile radar u and v at each vertical level relative to the large-scale model background wind field under precipitation conditions. The specific method for calculating the annual average standard deviation is: In the formula, represents the annual average standard deviation of wind u and v at vertical level j of wind profiler radar i in the case of precipitation; represents the hourly average sampled u and v winds at time t when there is precipitation at vertical level j of wind profiler radar i; Represents the large-scale model background u, v winds interpolated to the geographical location of wind profiler i and vertical level j at time t when there is precipitation.
5. The method for controlling wind field data quality of wind profiler radar for data assimilation according to claim 1, characterized in that: In step S4, the monthly average standard deviation and monthly average correlation coefficient As the threshold, the high confidence interval of the wind farm is determined, which is specifically expressed as: In the formula, Represents the high confidence interval of the u and v wind fields of wind profiler radar i in month k.
6. The method for controlling wind field data quality of wind profiler radar for data assimilation according to claim 1, characterized in that: In step S5, the annual average standard deviation under the condition of precipitation is As the threshold, the high confidence interval of the wind farm is determined, which is specifically expressed as: In the formula, Represents the high confidence interval of the u and v wind fields of wind profiler radar i in the presence of precipitation.
7. The method for controlling wind field data quality of wind profiler radar for data assimilation according to claim 1, characterized in that: The step S6 is specifically as follows: for the hourly average sampled u and v winds of the wind profile radar i at time t, the data acquisition rate at time t is characterized by the ratio of the amount of wind field data actually acquired to the total amount of data that should be acquired.
