A long-time-series human shadow operation effect quantitative evaluation method based on numerical simulation

Abstract

The application discloses a long-time-series human shadow operation effect quantitative evaluation method based on numerical simulation, which comprises the following steps: image processing is performed to determine a sub-evaluation period, a grid representative area and an area weight; driving data, real-time precipitation field data and artificial influence operation data of each sub-evaluation period of an area to be evaluated are obtained; numerical simulation of each sub-evaluation period is performed; the numerical simulation results of each sub-evaluation period are aligned with the real-time precipitation field data; accuracy is tested and the grid rain enhancement rate in each sub-evaluation period is calculated; the period rain enhancement amount of each sub-evaluation period of the area to be evaluated is calculated according to the grid rain enhancement rate and the area weight; and the period rain enhancement amount of each sub-evaluation period is accumulated to obtain a long-time-series rain enhancement amount evaluation result of the area to be evaluated. The method can more comprehensively and accurately evaluate the long-time-series human shadow operation effect, and has important significance for promoting the scientific management and reasonable distribution of water resources.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rainfall assessment, and in particular to a long-time series artificial influence operation effect quantitative assessment method based on numerical simulation. BACKGROUND

[0002] Numerical simulation test is one of the main technical methods for artificial influence operation effect assessment, and establishing a numerical simulation system capable of simulating the actual artificial influence operation process is an important basis for the practical application of such methods.

[0003] However, the current numerical simulation effect assessment method has great limitations: first, the simulation results are based on the precipitation field obtained by numerical model simulation, and there is a difference between the model-simulated precipitation field and the actual precipitation, so the assessment of the increased rainfall amount is more of a relative reference, and cannot give quantitative results on the actual rainfall background field; in addition, for artificial influence operations, it is often necessary to carry out long-time series artificial influence operation effect assessment for several days, months, seasons, years, etc. Due to the increasing error of the numerical model as the simulation period is extended, the model simulation is not suitable for using a single long-time simulation method for assessment. Therefore, the present application proposes a long-time series artificial influence operation effect quantitative assessment method based on numerical simulation, which combines simulation catalysis simulation with actual artificial influence operation information and real-time rainfall data to obtain the quantitative numerical simulation assessment results of the artificial influence operation effect based on the actual precipitation background field in this stage, thereby improving the scientificity and accuracy of artificial rainfall operation effect assessment, optimizing the operation scheme, and promoting the scientific management and rational allocation of water resources. SUMMARY

[0004] The purpose of the present application is to provide a long-time series artificial influence operation effect quantitative assessment method based on numerical simulation.

[0005] To achieve the above-mentioned purpose, the present application is implemented according to the following technical solutions:

[0006] The present application comprises the following steps:

[0007] Obtain a long-time series weather process graph of the region to be assessed, perform image processing on the weather process graph, and determine the sub-assessment period, grid representative region, and region weight according to the image processing results;

[0008] Obtain the driving data, real-time precipitation field data, and artificial influence operation data of each sub-assessment period of the region to be assessed, and perform numerical simulation of the first sub-assessment period; the numerical simulation includes natural cloud simulation and catalytic cloud simulation, and the numerical simulation results include background field data, assessment data, and catalytic influence data;

[0009] writing the catalytic effect data of the previous sub-evaluation period into the driving data of the next sub-evaluation period for numerical simulation, aligning the numerical simulation results of each sub-evaluation period with the real-time precipitation field data, performing accuracy verification, and calculating the grid rain enhancement rate in each sub-evaluation period; the grid rain enhancement rate includes a hit grid point rain enhancement rate and a false alarm grid point rain enhancement rate;

[0010] According to the grid rain enhancement rate and the area weight, the period rain enhancement amount of each sub-evaluation period in the to-be-evaluated area is calculated, and the period rain enhancement amounts of each sub-evaluation period are accumulated to obtain a long-time-series rain enhancement amount evaluation result of the to-be-evaluated area.

[0011] Further, the method for determining the sub-evaluation period, the grid representative area, and the area weight according to the image processing result comprises:

[0012] Taking the long-time-series initial time as the initial time of the first sub-evaluation period, a weather process map corresponding to the to-be-evaluated area is obtained as an initial map, and a weather process map is obtained every 1 h as a comparison map, and the initial map and the comparison map are processed by using a weather process map model to obtain an image change rate;

[0013] If the image change rate is less than a change rate threshold, the time interval of the comparison map is continuously increased for image processing, until the image change rate is not less than the change rate threshold, the current time is selected as the end time of the first sub-evaluation period, and the first sub-evaluation period is output;

[0014] The current time is selected as the initial time of the next sub-evaluation period, and the above operation is repeated by taking the weather process map at the current time as the initial map, until the end time of the last sub-evaluation period coincides with the long-time-series end time, and a plurality of sub-evaluation periods are divided;

[0015] The ground of the to-be-evaluated area is divided into uniform grids, the longitudinal area is uniformly divided into longitudinal evaluation areas according to the cloud height, the wind field humidity map of each longitudinal evaluation area in the same sub-evaluation period is obtained, and different wind field areas and different humidity areas are divided; the wind speed deviation value of the different wind field areas is greater than 2 m / s; and the humidity deviation value of the different humidity areas is greater than 5%;

[0016] The specific steps of dividing the different wind field areas and the different humidity areas are as follows:

[0017] A same wind speed gradient and a same humidity gradient are calculated, points with a gradient greater than a gradient threshold are marked, high-gradient points are taken as clustering seed points for adaptive clustering, and adjacent points with an attribute difference less than a wind speed deviation value and a humidity deviation value are merged to divide the different wind field areas and the different humidity areas; the adaptive clustering expression is as follows:

[0018]

[0019] wherein is the region of the kth cluster, (x, y, z) is the three-dimensional coordinate, p i is a point in space, c k is the center point of the kth cluster, r k is the dynamic radius of the kth cluster, ∈1 is a smoothing factor, G(p) is the gradient amplitude at point p i , G(c k ) is the gradient amplitude at the center point c k ;

[0020] projecting the wind field humidity maps of each longitudinal evaluation area in the same sub-evaluation period to the ground to obtain a wind field humidity overlay map, and selecting the area with the highest humidity overlap degree of the same wind field as the grid representative area; the grid representative area has an area greater than 30% of the standard grid area;

[0021] determining the area weight according to the area of the grid representative area in the grid and the overlap degree, and the expression is:

[0022]

[0023] wherein w rep,i is the area weight of the grid representative area in the grid i, l ove is the overlap degree of the grid representative area, ξ w is an area-sensitive coefficient, A rep,i is the area of the grid representative area in the grid i, A sta is the standard grid area.

[0024] Further, the method for processing the initial map and the contrast map by using the weather process map model to obtain the image change rate comprises the following steps:

[0025] The weather process map model comprises a multi-modal feature extraction layer, an adaptive weight distribution layer and a weighted change rate calculation layer.

[0026] The multi-modal feature extraction layer extracts the color feature of the image through the HSV space, calculates the texture feature by using the wavelet energy, enhances the calculation of the line feature by Hough transformation, and extracts the shape feature by Zernike moment, and calculates the change rate of each modal feature according to the feature extraction result, and the expression is:

[0027]

[0028] ΔC is the color feature change rate, ΔΓ is the texture feature change rate, ΔL is the line feature change rate, ΔS is the shape feature change rate, N pix is the total number of image pixels, respectively, hue, saturation and lightness value of pixel i in the initial map, respectively are hue, saturation, lightness value of pixel i in the contrast image, w c is a weight vector of color feature, is the jth layer low frequency sub-band energy of the initial image, is the jth layer low frequency sub-band energy of the contrast image, is the kth layer high frequency sub-band energy of the initial image, is the kth layer high frequency sub-band energy of the contrast image, N LL is the number of low frequency sub-bands, N HH is the number of high frequency sub-bands, P s (θ), P e (θ) are the cumulative line segment projection intensity of the initial image and the contrast image in the direction of θ, respectively, N d is the shape dimension, are the Zernike moments of the initial image and the contrast image respectively, n is the order, m is the repetition number;

[0029] The adaptive weight distribution layer adjusts the weight of the change rate of each modality feature according to the output of the multi-modal feature extraction layer, and the expression is:

[0030]

[0031] where w C , w Γ , w L , w S are the color feature change rate weight, the texture feature change rate weight, the line feature change rate weight, and the shape feature change rate weight, respectively, M is an adaptive matrix, and b is a bias vector;

[0032] The weighted change rate calculation layer calculates the image change rate of the initial image and the contrast image according to the change rate of each modality feature and the change rate of each modality feature.

[0033] Further, the method for performing numerical simulation in the first sub-evaluation period comprises:

[0034] Obtain driving data, real-time precipitation field data and artificial influence operation data of the first sub-evaluation period of the region to be evaluated; the driving data is used to drive the model to run, and the sub-evaluation period is divided and not affected by the actual catalytic operation, including dynamic field and temperature and humidity field; the real-time precipitation field data is divided into sub-evaluation periods, and is the real-time precipitation data of each grid and the corresponding equal latitude and longitude grid data; the artificial influence operation data is divided into days and affected by the actual catalytic operation;

[0035] The driving data of the first sub-evaluation period is processed to obtain initial field and boundary data field, and the initial field and the boundary data field are naturally cloud simulated to obtain background field data of the first sub-evaluation period; the background field data is not affected by the actual catalytic operation and includes precipitation field and cloud field;

[0036] According to the initial field, the boundary data field, the background field data and the artificial influence operation data, catalytic cloud simulation is performed to obtain evaluation data and catalytic influence area data of the first sub-evaluation period; the evaluation data includes precipitation field and cloud field; the catalytic influence area data includes three-dimensional spatial distribution data and two-dimensional ground distribution data of catalytic influence; the catalytic influence is represented by artificial ice crystals after silver iodide catalysis and catalyst nucleation.

[0037] Further, the method for performing accuracy test and calculating grid rain enhancement rate in each sub-evaluation period comprises the following steps:

[0038] The driving data of each sub-evaluation period is processed to obtain initial field and boundary data field, and natural cloud simulation is performed to obtain background field data of the corresponding sub-evaluation period;

[0039] The three-dimensional spatial distribution data of the catalytic influence of the last time of the last sub-evaluation period is extracted, and the initial field, the boundary data field, the background field data and the artificial influence operation data of the sub-evaluation period are subjected to catalytic cloud simulation to obtain evaluation data and catalytic influence area data of the corresponding sub-evaluation period;

[0040] Data conversion and interpolation processing are performed to align the background field data, the evaluation data and the catalytic influence area data of each evaluation period with the real precipitation field data; the dimensions of the data alignment include grid precipitation area and grid resolution;

[0041] The regional rain enhancement rate is calculated according to the background field data and the evaluation data of the precipitation field of each grid representative area in the same evaluation period;

[0042] The ground influence area range of the artificial operation in the sub-evaluation period is determined according to the two-dimensional ground distribution data of the catalytic influence, and the evaluation data and the real precipitation field data are compared and TS scored in the corresponding ground influence area range; the hit grid and the false alarm grid are determined according to the TS score result and the comparison result;

[0043] The search radius range is determined, the adjacent grid weight in the hit grid radius range is determined according to the evaluation data and the real precipitation field data, and the hit grid rain enhancement rate is calculated by weighting according to the adjacent grid weight and the corresponding regional rain enhancement rate;

[0044] The mean value of the regional rain enhancement rate of the hit grid in the false alarm grid radius range is taken as the false alarm grid rain enhancement rate.

[0045] Further, the method for obtaining the long-time series rain enhancement amount evaluation result of the to-be-evaluated area comprises the following steps.

[0046] The grid rain enhancement amount is calculated according to the grid rain enhancement rate, the corresponding area weight, the grid real-time rainfall and the grid area; the grid rain enhancement amount comprises a hit grid rain enhancement amount and a missed grid rain enhancement amount.

[0047] The grid rain enhancement amounts in the operation influence area of the same sub-evaluation period in the to-be-evaluated area are accumulated to obtain the sub-evaluation period rain enhancement amount, and the rain enhancement amounts in each sub-evaluation period in the long-time series are accumulated to obtain the long-time series rain enhancement amount evaluation result of the to-be-evaluated area.

[0048] The present application has the following beneficial effects:

[0049] Compared with the prior art, the present application has the following technical effects:

[0050] The present application can improve the data preprocessing capability in the long-time series human shadow operation effect quantitative evaluation, can improve the speed of time and longitudinal space evaluation area division, thereby improving the efficiency and precision of the long-time series human shadow operation effect quantitative evaluation, optimizing the long-time series human shadow operation effect quantitative evaluation technology, can greatly save resources, improve work efficiency, realize the dynamic evaluation of the long-time series human shadow operation effect quantitative evaluation, and has important significance for improving the scientific nature and accuracy of the artificial rain enhancement operation effect evaluation, optimizing the operation scheme, and promoting the scientific management and reasonable allocation of water resources. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 The present application is a long-time series human shadow operation effect quantitative evaluation method based on numerical simulation. DETAILED DESCRIPTION

[0052] The present application will be further described below through specific embodiments, and the illustrative embodiments of the present application and the description are used to explain the present application, but not as a limitation of the present application.

[0053] The present application is a long-time series human shadow operation effect quantitative evaluation method based on numerical simulation, comprising the following steps:

[0054] As shown in the figure, Figure 1 In the present embodiment, the following steps are included:

[0055] A long-time series weather process graph of a to-be-evaluated area is obtained, the weather process graph is image-processed, and a sub-evaluation period, a grid representative area and an area weight are determined according to the image processing result;

[0056] acquire driving data, live precipitation field data and artificial influence operation data of each sub-evaluation period of the region to be evaluated, and perform numerical simulation of the first sub-evaluation period; the numerical simulation includes natural cloud simulation and catalytic cloud simulation, and the numerical simulation result includes background field data, evaluation data and catalytic influence data;

[0057] write the catalytic influence data of the previous sub-evaluation period into the driving data of the next sub-evaluation period for numerical simulation, align the numerical simulation results of each sub-evaluation period with the live precipitation field data, perform accuracy test and calculate the grid rain enhancement rate in each sub-evaluation period; the grid rain enhancement rate includes hit grid point rain enhancement rate and missed grid point rain enhancement rate;

[0058] calculate the period rain enhancement amount of each sub-evaluation period of the region to be evaluated according to the grid rain enhancement rate and the region weight, and accumulate the period rain enhancement amount of each sub-evaluation period to obtain the long-time series rain enhancement amount evaluation result of the region to be evaluated.

[0059] In the embodiment, the method for determining the sub-evaluation period, the grid representative region and the region weight according to the image processing result comprises:

[0060] take the long-time series initial time as the initial time of the first sub-evaluation period, acquire the weather process graph corresponding to the initial time as an initial graph, acquire a weather process graph every 1h as a comparison graph, and process the initial graph and the comparison graph by using a weather process graph model to obtain an image change rate;

[0061] if the image change rate is less than a change rate threshold, continue to increase the time interval of the comparison graph for image processing, until the image change rate is not less than the change rate threshold, select the current time as the end time of the first sub-evaluation period, and output the first sub-evaluation period;

[0062] select the current time as the initial time of the next sub-evaluation period, continue to perform the above operation by taking the weather process graph of the current time as the initial graph, until the end time of the last sub-evaluation period coincides with the long-time series end time, and the multiple sub-evaluation periods are divided;

[0063] divide the ground of the region to be evaluated into uniform grids, divide the longitudinal region into longitudinal evaluation zones according to the cloud height, acquire the wind field humidity graph of each longitudinal evaluation zone in the same sub-evaluation period, and divide different wind field regions and different humidity regions; the wind speed deviation value of the different wind field regions is greater than 2m / s; the humidity deviation value of the different humidity regions is greater than 5%;

[0064] The specific steps of dividing the different wind field regions and the different humidity regions are as follows:

[0065] The same wind speed gradient and humidity gradient are calculated, points with a gradient greater than a gradient threshold value are marked, high gradient points are taken as cluster seed points for adaptive clustering, and points with adjacent and attribute differences less than a wind speed deviation value and a humidity deviation value are combined to divide different wind field regions and different humidity regions; the adaptive clustering expression is:

[0066]

[0067] wherein represents a region of the kth cluster, (x, y, z) is a three-dimensional space coordinate, p i is a point in space, c k is a center point of the kth cluster, r k is a dynamic radius of the kth cluster, ∈1 is a smoothing factor, G(p) is a gradient amplitude at the point p i , G(c k ) is a gradient amplitude at the center point c k ;

[0068] The wind field humidity map of each longitudinal evaluation area in the same sub-evaluation period is projected to the ground to obtain a wind field humidity overlay map, and the area with the highest overlap degree of the same wind field and the same humidity is selected as a grid representative area; the grid representative area has an area greater than 30% of a standard grid area;

[0069] The area weight of the grid representative area in the grid is determined according to the area and the overlap degree of the grid representative area in the grid, and the expression is:

[0070]

[0071] wherein w rep,i is the area weight of the grid representative area in the grid i, l ove is the overlap degree of the grid representative area, ξ w is an area-sensitive coefficient, A rep,i is the area of the grid representative area in the grid i, and A sta is a standard grid area;

[0072] In actual evaluation, the change rate threshold value is taken as 20%, if the image change rate of the comparison map after 1h compared with the initial map is less than 20%, the image change rate of the comparison map after 2 / 3 / 4h compared with the initial map is continuously calculated at an interval of 1h, until the image change rate of the comparison map after 4h compared with the initial map is 22%, the first sub-evaluation period length is taken as 4h, 4h is selected as the initial time of the next sub-evaluation period, and the above operation is continuously performed with the weather process map at the present moment as the initial map until the end time of the last sub-evaluation period coincides with the end time of the long sequence, to divide a plurality of sub-evaluation periods;

[0073] The longitudinal region is evenly divided into 5 longitudinal evaluation zones according to the cloud height, the wind field gradient and the humidity gradient of each longitudinal evaluation zone are calculated according to the wind field humidity chart, the points with the gradient greater than the gradient threshold value (the wind field gradient is 0.5 m / s, and the humidity gradient is 1%) are marked, the high gradient points are taken as the clustering seed points for adaptive clustering by taking the smoothing factor ∈1 as 1, the points adjacent to each other and having an attribute difference less than a wind speed deviation value 2 m / s and a humidity deviation value 5% are combined to divide different wind field regions and different humidity regions;

[0074] In the same grid, the wind field deviation of different longitudinal evaluation zones at the same horizontal projection is within ±2 m / s, and the humidity is within ±5% to be recorded as a first overlap degree, for example, the a region in the grid (2.3), the longitudinal wind field projection is 14.6, 13.7, 15.4, 16, 16.1 m / s, and the wind field deviation is recorded as 4 with 14.6 m / s as the reference, the longitudinal humidity projection is 88%, 85%, 90%, 80%, 81%, and the humidity deviation is recorded as 2 with 88% as the reference, and the overlap degree of the a region in the grid (2.3) is 6, if the area and the overlap degree of two overlapping regions in the grid (2.3) are respectively: the a region (0.21A sta / 6), and the b region (0.55A sta / 5), since 0.21A sta <0.3A sta , the b region is selected as the grid representative region, the area sensitivity coefficient ξ w is 0.6, and the area weight of the b region is calculated as 0.741.

[0075] In the embodiment, the method for processing the initial image and the contrast image by using the weather process model to obtain the image change rate comprises the following steps:

[0076] The weather process model comprises a multi-modal feature extraction layer, an adaptive weight distribution layer, and a weighted change rate calculation layer.

[0077] The multi-modal feature extraction layer extracts the color feature of the image through the HSV space, calculates the texture feature by using the wavelet energy, enhances the line feature by performing the Hough transformation, and extracts the shape feature by using the Zernike moment, and calculates the change rate of each modal feature according to the feature extraction result, and the expression is as follows:

[0078]

[0079] ΔC is the color feature change rate, ΔΓ is the texture feature change rate, ΔL is the line feature change rate, ΔS is the shape feature change rate, N pix is the total number of image pixels, respectively, is the hue, saturation, and brightness value of pixel i in the initial image, respectively are hue, saturation, lightness value of pixel i in the contrast image, w c is a weight vector of color feature, is the jth layer low frequency subband energy of the initial image, is the jth layer low frequency subband energy of the contrast image, is the kth layer high frequency subband energy of the initial image, is the kth layer high frequency subband energy of the contrast image, N LL is the number of low frequency subbands, N HH is the number of high frequency subbands, P s (θ), P e (θ) are the cumulative line segment projection intensity of the initial image and the contrast image in the direction of θ, respectively, N d is the shape dimension, are the Zernike moments of the initial image and the contrast image respectively, n is the order, m is the repetition number;

[0080] The adaptive weight distribution layer adjusts the weight of the change rate of each modality feature according to the output of the multi-modal feature extraction layer, and the expression is:

[0081]

[0082] where w C , w Γ , w L , w S are the color feature change rate weight, the texture feature change rate weight, the line feature change rate weight, and the shape feature change rate weight, respectively, M is an adaptive matrix, and b is a bias vector;

[0083] The weighted change rate calculation layer calculates the image change rate of the initial image and the contrast image according to the change rate weight of each modality feature and the change rate of each modality feature;

[0084] In actual evaluation, the expression for calculating the image change rate of the initial image and the contrast image according to the change rate weight of each modality feature and the change rate of each modality feature is:

[0085]

[0086] where R is the image change rate, the color feature change rate weight, the texture feature change rate weight, the line feature change rate weight, and the shape feature change rate weight are 0.4, 0.2, 0.35, and 0.25, respectively.

[0087] In this embodiment, the method for performing numerical simulation in the first sub-evaluation period comprises:

[0088] obtain driving data, live precipitation field data and artificial influence operation data of the first sub-evaluation period of the region to be evaluated; the driving data is used to drive the model to run, divided by sub-evaluation period and not affected by actual catalytic operation, including dynamic field and temperature and humidity field; the live precipitation field data is divided by sub-evaluation period, live precipitation data of each grid and corresponding equal latitude and longitude grid data; the artificial influence operation data is divided by day and affected by actual catalytic operation;

[0089] process the driving data of the first sub-evaluation period to obtain initial field and boundary data field, and perform natural cloud simulation on the initial field and boundary data field to obtain background field data of the first sub-evaluation period; the background field data is not affected by actual catalytic operation, including precipitation field and cloud field;

[0090] perform catalytic cloud simulation according to the initial field, boundary data field, background field data and artificial influence operation data to obtain evaluation data and catalytic influence area data of the first sub-evaluation period; the evaluation data includes precipitation field and cloud field; the catalytic influence area data includes three-dimensional spatial distribution data and two-dimensional ground distribution data of catalytic influence; the catalytic influence is represented by artificial ice crystals after silver iodide catalysis and catalyst nucleation;

[0091] In actual evaluation, reanalysis field data (dynamic field and temperature and humidity field) of the sub-evaluation period is extracted from the global model of the official meteorological agency as driving data, which is not affected by actual catalytic operation; grid-based multi-source fusion live analysis hourly precipitation product data and corresponding equal latitude and longitude grid data are obtained from the official meteorological agency to form live precipitation field data, and are combined according to the sub-evaluation period; actual human shadow operation information (aircraft / ground operation position, time, dose) is processed by the catalytic data comprehensive processing system of the CMA-CPEFS model to obtain artificial influence operation data;

[0092] CRA global atmospheric reanalysis field data of the first sub-evaluation period is processed into initial field data file and boundary condition data file that can drive the CMA-CPEFS model, natural cloud simulation (simulation resolution is 1km) is performed using the initial field data file and boundary condition data file to obtain background field data without reading catalytic information, and catalytic cloud simulation (simulation resolution is 1km) is performed using the background field data, artificial influence operation data, initial field data file and boundary condition data file to obtain evaluation data and catalytic influence area data reading catalytic information (matching artificial influence operation data corresponding to the first sub-evaluation period); the catalytic influence area data is represented by three-dimensional and two-dimensional concentration distribution of artificial ice crystals after silver iodide catalysis and catalyst nucleation.

[0093] In this embodiment, the method for performing accuracy test and calculating grid rain enhancement rate in each sub-evaluation period comprises the following steps:

[0094] processing the driving data of each sub-evaluation period to obtain an initial field and a boundary data field and performing natural cloud simulation to obtain the background field data of the corresponding sub-evaluation period;

[0095] extracting the three-dimensional spatial distribution data of catalytic influence of the last time of the previous sub-evaluation period, and performing catalytic cloud simulation on the initial field, the boundary data field, the background field data and the artificial influence operation data of the sub-evaluation period to obtain the evaluation data and the catalytic influence area data of the corresponding sub-evaluation period;

[0096] performing data conversion and interpolation processing to align the background field data, the evaluation data and the catalytic influence area data of each evaluation period with the real-time precipitation field data; the dimensions of the data alignment include the grid precipitation area and the grid resolution;

[0097] According to the background field data and the evaluation data of the grid representative area in each grid of the same evaluation period, the regional rain enhancement rate is calculated and obtained;

[0098] According to the two-dimensional ground distribution data of catalytic influence, the ground influence area range of the artificial operation in the sub-evaluation period is determined, and the evaluation data is compared with the real-time precipitation field data in the corresponding ground influence area range and is scored by TS, and the hit grid and the false alarm grid are determined according to the TS score result and the comparison result;

[0099] determine the search radius range, according to the evaluation data and the real-time precipitation field data, determine the weight of the adjacent grid in the hit grid radius range, and according to the adjacent grid weight and the corresponding regional rain enhancement rate, the hit grid rain enhancement rate is obtained by weighted calculation;

[0100] Take the average of the regional rain enhancement rate of the hit grid in the false alarm grid radius range as the false alarm grid rain enhancement rate;

[0101] In actual evaluation, the grid rain enhancement rate of the grid representative area in each grid of the same evaluation period is calculated according to "(evaluation rainfall-background field precipitation) / background field precipitation";

[0102] According to the position of the two-dimensional ground distribution data (silver iodide catalysis and catalyst nucleated ice crystal), the ground influence area range of the artificial operation in the sub-evaluation period is determined, the evaluation precipitation field data (forecast value) is compared with the real-time precipitation field data, the grid with TS score greater than 0.6 and forecast hit rate greater than 50% is taken as the hit grid, and the other grid is taken as the false alarm grid;

[0103] The search radius is determined as 5km, the inverse of the difference between the adjacent grid evaluation precipitation field data (forecast value) and the real-time precipitation field data is calculated as the adjacent forecast difference coefficient, the ratio of the adjacent grid forecast difference coefficient and the cumulative value of the adjacent grid forecast difference coefficient in the radius range is taken as the adjacent grid weight, and the hit grid rain enhancement rate is obtained by weighted calculation according to the adjacent grid weight and the corresponding grid rain enhancement rate.

[0104] In the embodiment, the method for obtaining the long-time-series rain enhancement evaluation result of the to-be-evaluated area comprises the following steps:

[0105] The grid rain enhancement is calculated according to the grid rain enhancement rate, the corresponding area weight, the grid real-time rainfall and the grid area; the grid rain enhancement comprises the hit grid rain enhancement and the missed grid rain enhancement;

[0106] The grid rain enhancement in the operation influence area of the same sub-evaluation period of the to-be-evaluated area is accumulated to obtain the sub-evaluation period rain enhancement, and the rain enhancement of each sub-evaluation period in the long-time-series is accumulated to obtain the long-time-series rain enhancement evaluation result of the to-be-evaluated area;

[0107] In the actual evaluation, the expression for calculating the grid rain enhancement is:

[0108]

[0109] wherein dR i is the rain enhancement of the grid i, OR i is the grid real-time rainfall of the grid i, A sta is the standard grid area, A rep,i is the grid representative area of the grid i, dRP i is the grid rain enhancement rate of the grid i, and w rep,i is the area weight of the grid representative area in the grid i.

[0110] The above only describes the preferred embodiments of the present application and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A quantitative evaluation method for the effect of long-term artificial weathering operations based on numerical simulation, characterized in that, Includes the following steps: S1. Obtain a long-term weather process map of the area to be evaluated, perform image processing on the weather process map, and determine the sub-evaluation period, grid representative area, and area weight based on the image processing results. S2. Obtain driving data, actual precipitation field data, and artificial influence operation data for each sub-evaluation period of the area to be evaluated, and perform numerical simulation for the first sub-evaluation period; the numerical simulation includes natural cloud simulation and catalytic cloud simulation, and the numerical simulation results include background field data, evaluation data, and catalytic influence data; S3. Write the catalytic impact data of the previous sub-evaluation period into the driving data of the next sub-evaluation period for numerical simulation. Align the numerical simulation results of each sub-evaluation period with the actual precipitation field data, perform accuracy verification, and calculate the grid rain enhancement rate in each sub-evaluation period. The grid rain enhancement rate includes the rain enhancement rate of the hit grid points and the rain enhancement rate of the missed grid points. S4. Calculate the time-period rainfall increase of each sub-evaluation period in the area to be evaluated based on the grid rainfall increase rate and the regional weight, and sum the time-period rainfall increases of each sub-evaluation period to obtain the long-term rainfall increase evaluation result of the area to be evaluated.

2. The method for quantitative evaluation of the effect of long-term artificial weathering operations based on numerical simulation as described in claim 1, characterized in that, The method for determining the sub-evaluation period, the grid representative region, and the region weight based on the image processing results includes: Using the long-term initial time as the initial time of the first sub-evaluation period, the weather process map corresponding to the area to be evaluated is obtained as the initial map, and a weather process map is obtained every 1 hour as the comparison map. The initial map and the comparison map are processed using the weather process map model to obtain the image change rate. If the image change rate is less than the change rate threshold, continue to increase the time interval of the comparison image for image processing until the image change rate is not less than the change rate threshold. Then, select this moment as the end time of the first sub-evaluation period and output the first sub-evaluation period. Select this moment as the initial time of the next sub-evaluation period, and continue to repeat the above operation with the weather process map of this moment as the initial map until the end time of the last sub-evaluation period coincides with the end time of the long time series, thus dividing the time into multiple sub-evaluation periods; The ground area to be evaluated is divided into a uniform grid. The longitudinal area is then uniformly divided into longitudinal evaluation zones according to cloud height. Wind field humidity maps of each longitudinal evaluation zone are obtained for the same sub-evaluation period, and different wind field zones and different humidity zones are defined. The wind speed deviation value of the different wind field zones is greater than 2 m / s. The humidity deviation value of the different humidity zones is greater than 5%. The specific steps for dividing the wind field regions and humidity regions are as follows: Calculate the same wind speed gradient and humidity gradient, mark points with gradients greater than the gradient threshold, use high gradient points as clustering seed points for adaptive clustering, merge adjacent points with attribute differences less than the wind speed deviation and humidity deviation values ​​to divide them into different wind field regions and different humidity regions; the adaptive clustering expression is: in To represent the region of the k-th cluster, (x, y, z) are three-dimensional spatial coordinates, p i Let c be a point in space. k Let r be the center point of the k-th cluster. k Let G(p) be the dynamic radius of the k-th cluster, ∈1 be the smoothing factor, and G(p) be the dynamic radius of the k-th cluster. i Gradient magnitude at point G(c) k ( ) is the center point c k The gradient magnitude; The wind field humidity maps of each longitudinal assessment area in the same sub-assessment period are projected onto the ground to obtain a wind field humidity overlay map. The area with the highest overlap of the same wind field and the same humidity is selected as the grid representative area; the area of ​​the grid representative area is greater than 30% of the standard grid area. The region weight is determined based on the area and overlap of the regions represented by the grid cells. The expression is as follows: Where w rep,i Let l be the region weight representing the area within grid i. ove Let ξ represent the overlap of the grid regions. w A is the area sensitivity coefficient. rep,i Let A be the area of ​​the region represented by the grid within grid i. sta This represents the standard grid area.

3. The method for quantitative evaluation of the effect of long-term artificial weathering operations based on numerical simulation according to claim 2, characterized in that, The method for obtaining the rate of change of an image by processing an initial map and a comparison map using a weather process map model includes the following steps: The weather process map model includes a multimodal feature extraction layer, an adaptive weight allocation layer, and a weighted rate of change calculation layer; The multimodal feature extraction layer extracts image color features using HSV space, calculates texture features using wavelet energy, enhances line feature calculation using Hough transform, and extracts shape features using Zernike moments. Based on the feature extraction results, it calculates the rate of change of each modality's features, expressed as: ΔC is the rate of change of color features, ΔΓ is the rate of change of texture features, ΔL is the rate of change of line features, ΔS is the rate of change of shape features, and N is the rate of change of color features. pix The total number of pixels in the image. These represent the hue, saturation, and brightness values ​​of pixel i in the initial image. These represent the hue, saturation, and brightness values ​​of pixel i in the comparison image, respectively, and w c Let be the weight vector of the color features. Let the energy of the j-th low-frequency subband of the initial graph be . To compare the energy of the low-frequency subband of the j-th layer in the diagram, The energy of the high-frequency subband of the k-th layer of the initial graph. To compare the energy of the high-frequency subband of the k-th layer in the diagram, N LL N represents the number of low-frequency subbands. HH P represents the number of high-frequency subbands. s (θ), P e (θ) represents the cumulative line segment projection intensity in the θ direction of the initial and comparison images, respectively, N. d For shape dimension, These are the Zernike moments of order n with m repetitions for the initial and comparison graphs, respectively. The adaptive weight allocation layer adjusts the weights of the change rate of each modality feature based on the output of the multimodal feature extraction layer, as expressed by: Where w C w Γ w L w S These are the weights for the rate of change of color features, the rate of change of texture features, the rate of change of line features, and the rate of change of shape features, respectively. M is the adaptive matrix, and b is the bias vector. The weighted rate of change calculation layer calculates the image change rate by calculating the initial image and the comparison image based on the weights of the change rates of each modal feature and the change rates of each modal feature.

4. The method for quantitative evaluation of the effect of long-term artificial weathering operations based on numerical simulation according to claim 1, characterized in that, The method for conducting numerical simulation of the first sub-evaluation period includes the following steps: Acquire driving data, actual precipitation field data, and artificial influence operation data for the first sub-evaluation period of the area to be evaluated; the driving data is used for driving mode operation, is divided by sub-evaluation period and is not affected by actual catalytic operations, and includes dynamic field and temperature and humidity field; the actual precipitation field data is divided by sub-evaluation period, consisting of actual precipitation data for each grid and corresponding isotropic grid data; the artificial influence operation data is divided by day and is affected by actual catalytic operations; The initial field and boundary data fields are obtained by processing the driving data of the first sub-evaluation period. The initial field and boundary data fields are then simulated using natural clouds to obtain the background field data of the first sub-evaluation period. The background field data is not affected by the actual catalytic operation and includes precipitation field and cloud field. The evaluation data and catalytic impact zone data for the first sub-evaluation period were obtained by catalytic cloud simulation based on the initial field, boundary data field, background field data and artificial influence operation data. The evaluation data includes precipitation field and cloud field. The catalytic impact zone data includes three-dimensional spatial distribution data and two-dimensional ground distribution data of catalytic impact. The catalytic impact is represented by artificial ice crystals after silver iodide catalysis and catalyst nucleation.

5. The method for quantitative evaluation of the effect of long-term artificial weathering operations based on numerical simulation according to claim 1, characterized in that, The method for performing accuracy verification and calculating the grid-based rainfall enhancement rate within each sub-evaluation period includes the following steps: The driving data of each sub-evaluation period is processed to obtain the initial field and boundary data field, and natural cloud simulation is performed to obtain the background field data of the corresponding sub-evaluation period; Extract the three-dimensional spatial distribution data of the catalytic impact from the last time period of the previous sub-evaluation period, and perform catalytic cloud simulation with the initial field, boundary data field, background field data and artificial influence operation data of the sub-evaluation period to obtain the evaluation data and catalytic impact area data of the corresponding sub-evaluation period; Data transformation and interpolation are performed to align the background field data, assessment data, and catalytic impact zone data with the actual precipitation field data for each assessment period; the dimensions of the data alignment include the gridded precipitation area and the grid resolution. The regional rainfall increase rate is calculated based on the precipitation field of the background field data and the assessment data of the representative areas of each grid within the same assessment period. Based on the two-dimensional ground distribution data of the catalytic effect, the ground impact area of ​​the artificial weather modification operation within the sub-assessment period is determined. Within the corresponding ground impact area, the assessment data is compared with the actual precipitation field data and TS score is performed. Based on the TS score results and comparison results, the hit grid points and missed grid points are determined. Determine the search radius range, determine the weight of neighboring grids within the radius range of the hit grid point based on the evaluation data and the actual precipitation field data, and calculate the rainfall increase rate of the hit grid by weighting the neighboring grid weights and the corresponding regional rainfall increase rate. The average rainfall increase rate of the area within the radius of the missed grid point that hit the grid point is taken as the missed grid rainfall increase rate.

6. The method for quantitative evaluation of the effect of long-term artificial weathering operations based on numerical simulation according to claim 1, characterized in that, The method for obtaining the long-term precipitation enhancement assessment results for the area to be assessed includes the following steps: The grid-based rainfall enhancement amount is calculated based on the grid-based rainfall enhancement rate and the corresponding regional weight, the actual rainfall in the grid, and the grid area; the grid-based rainfall enhancement amount includes the rainfall enhancement amount of the hit grid and the rainfall enhancement amount of the missed grid. The rainfall increase of the grid within the operational impact area of ​​the region to be evaluated during the same sub-evaluation period is accumulated to obtain the rainfall increase of the sub-evaluation period. The rainfall increase of each sub-evaluation period within the long time series is accumulated to obtain the long-term rainfall increase evaluation result of the region to be evaluated.

Images