Radar satellite precipitation rapid fusion algorithm based on radar quality index

Abstract

The application discloses a radar satellite precipitation rapid fusion algorithm based on a radar quality index, and comprises the following steps: a climate proportion correction factor RFC is utilized to design a radar sheltering area index RSA as a segmented function varying with the climate proportion correction factor RFC; a radar terrain sheltering index analysis field is utilized to optimize a radar quality index analysis field weight; and then, the radar quality index fusion weight is utilized to rapidly fuse radar and satellite precipitation data, so as to generate a radar-satellite fusion background field. Through utilization of the radar quality index and the terrain sheltering coefficient, the spatial reliability of the radar data is dynamically quantified, and the fusion weight of the radar and the satellite is optimized according to the spatial reliability, so that the fusion background field can more accurately reflect the actual precipitation condition, and in particular, in a radar coverage blind area and a sparse area of stations, problems such as local distortion, underestimation of precipitation intensity or structural ambiguity are effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of precipitation fusion technology. Specifically, it is a rapid precipitation fusion algorithm based on radar quality index. Background Technology

[0002] In multi-source precipitation fusion and real-time analysis, a high-spatiotemporal resolution and complete coverage fused background field is the core foundation supporting short-term forecasting, disaster early warning, and model assimilation. However, constrained by factors such as limited radar detection range, terrain obstruction effects, and uncertainties in satellite precipitation retrieval, existing fused background fields face severe challenges in balancing timeliness, spatial continuity, and accuracy. Especially in radar coverage blind spots (such as plateaus and offshore areas) and sparsely populated areas, traditional fusion methods often rely on static weight allocation or simple interpolation, making it difficult to dynamically adapt to the quality differences between radar and satellite data. This leads to problems such as local distortion, underestimated precipitation intensity, or structural ambiguity in the background field. Research shows that if the spatial heterogeneity of radar detection performance is not optimized, the ability of fused products to characterize precipitation gradient distribution and typhoon heavy precipitation centers in complex terrain areas will significantly decrease, directly affecting the accuracy and timeliness of disaster early warning.

[0003] The quality difference between radar and satellite precipitation data is mainly determined by their detection mechanisms and external environmental interference. While radar precipitation data boasts high spatiotemporal resolution, its detection efficiency is affected by multiple factors: 1) Spatial attenuation effect: as the distance from the radar center increases, beam widening and terrain obstruction (such as mountains and buildings) lead to echo signal attenuation, gradually reducing detection accuracy; 2) Coverage unevenness: the effective detection range of a single radar is limited, and multi-radar mosaics are prone to introducing fusion noise due to differences in overlapping and blind areas; 3) Climate-related terrain obstruction: long-term statistics show that the systematic bias of radar precipitation in areas with fixed terrain obstacles (such as basins and canyons) is closely related to the distribution of seasonal heavy precipitation. Although satellite precipitation data can compensate for insufficient radar coverage, its inversion algorithm is affected by cloud microphysical characteristics and surface radiation interference, resulting in significant uncertainties in the quantitative estimation of strong convective precipitation and the identification of weak precipitation. Therefore, dynamically quantifying the spatial reliability of radar data and optimizing the fusion weights of radar and satellite data accordingly is crucial for improving background field quality. CN201910114665.8 discloses a satellite-based method for monitoring severe convection and its application. It utilizes a daytime convective storm identification algorithm radar to compensate for blind spots in network monitoring. However, it primarily targets nearshore and offshore areas, and its effectiveness is limited for areas with fixed terrain obstacles (such as basins and canyons). CN202510174805.6 discloses an intelligent monitoring method and system for landslide surface deformation based on precipitation forecasts. This method considers the terrain shading index, but compensation is only triggered when the index is <0.3. There are no corresponding compensation measures for areas with long-term strong shading, which can easily lead to missing monitoring data. Furthermore, it does not address the complex terrain of areas with climatic shading.

[0004] Existing fusion technologies suffer from two main shortcomings: First, rigid weight allocation strategies. Traditional methods often use fixed thresholds (such as the effective detection radius of radar) or empirical coefficients to assign contribution weights to radar and satellites, failing to fully consider the dynamic changes in radar quality with distance, terrain obstruction, and coverage density. For example, in radar edge detection areas, where data quality has significantly degraded, it is still assigned a weight similar to that of the central area, leading to "pseudo-precipitation" or gradient distortion in the fusion results in transitional regions. Second, simplistic blind spot filling techniques. For areas not covered by radar, most schemes directly use satellite data or climatological mean values ​​for filling, without establishing a correlation model between radar quality attenuation and satellite error characteristics. This results in the loss of structural information from key systems such as typhoon outer rainbands and orographic lifting heavy precipitation. Furthermore, existing methods lack a collaborative optimization mechanism for the timeliness differences of multi-source data, making it difficult to meet the needs of rapidly updating operations. Especially during extreme weather events, the fused background field is prone to lag behind the actual evolution.

[0005] The construction of a high-timeliness fused background field has dual significance in meteorological operations. On the one hand, it provides a benchmark precipitation field for real-time analysis products, supporting precipitation phase identification, areal rainfall estimation, and flood risk assessment. On the other hand, as a rapidly updated assimilation field for numerical weather prediction models, its accuracy and timeliness directly affect key forecasting performance such as short-duration heavy precipitation and typhoon tracks. However, existing technologies are insufficient in characterizing the spatial heterogeneity of radar quality and have failed to effectively coordinate the complementary advantages of radar and satellite data, resulting in the reliability and resolution of the fused background field in key areas (such as topographic transition zones) failing to meet operational requirements. Therefore, it is urgent to innovate fusion algorithms, dynamically quantify radar quality indices, and optimize multi-source data fusion mechanisms to overcome the limitations of static weights and single-filling strategies, achieving the construction of a high-timeliness and high-consistency fused background field, providing technical support for disaster prevention and mitigation decision-making and refined meteorological services. Summary of the Invention

[0006] Therefore, the technical problem to be solved by the present invention is to provide a radar satellite precipitation fast fusion algorithm based on radar quality index, which can effectively solve the problems of rigid weight allocation, single blind zone information filling and lack of multi-source data timeliness collaborative optimization in existing fusion technologies.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] A rapid radar-satellite precipitation fusion algorithm based on radar quality index is proposed. The algorithm utilizes the climate proportionality correction factor (RFC) and designs the radar obstruction area index (RSA) as a piecewise function that varies with the RFC. The algorithm optimizes the radar quality index analysis field weights using the radar terrain obstruction index analysis field and then rapidly fuses radar and satellite precipitation data using the radar quality index fusion weights to generate a radar-satellite fused background field.

[0009] Preferably, the above includes the following steps:

[0010] S1. The radar quality index (RQI) is calculated by combining the detection distance from the radar center, the number of coverage points, and the terrain height at the radar network precipitation analysis grid points.

[0011] S2. Correction is performed by calculating the climate proportion correction factor (RFC) of radar precipitation using the INCA algorithm;

[0012] S3. The radar obstruction zone index RSA is designed as a piecewise function that varies with the climate proportional correction factor RFC.

[0013] S4. The weights of the radar quality index analysis field are further optimized and corrected using the radar obstruction zone index analysis field. The radar obstruction zone index RSA is applied to the radar quality index RQI using a hyperbolic function to obtain the optimized RQI, as shown in equation (6):

[0014] RQI opt (x,y)=RQI(x,y)·[1-α·tanh(β·RSA(x,y))] (6)

[0015] Wherein, RQI(x,y): the original radar quality index (value range [0,1]);

[0016] RSA(x,y): Radar Obstruction Index (value range [0,1], 1 indicates complete obstruction);

[0017] α: Intensity coefficient of topographic influence (0≤α≤1);

[0018] β: Steepness parameter of hyperbolic function (β>0);

[0019] tanh: Hyperbolic tangent function, used to smoothly transition occlusion effects;

[0020] S5. Utilize radar quality index fusion weights to quickly fuse radar and satellite precipitation data to generate a radar-satellite fused background field.

[0021] Preferably, the piecewise function in step S3 above is as shown in equation (5):

[0022]

[0023] When RFC is less than 1.0, let RSA = 0.0; when RFC is greater than or equal to 1.0, let RSA = 1 - e 0.05(rfc-1) When rain_rad = 0, that is, the grid point where the long-term cumulative radar precipitation is zero, let rsa = 1.0.

[0024] Preferably, in step S4 above, the radar obstruction index RSA is applied to the radar quality index RQI using a hyperbolic function to obtain the optimized RQI, as shown in equation (6):

[0025] RQI opt (x,y)=RQI(x,y)·[1-α·tanh(β·RSA(x,y))] (6)

[0026] Wherein, RQI(x,y): the original radar quality index (value range [0,1]);

[0027] RSA(x,y): Radar Obstruction Index (value range [0,1], 1 indicates complete obstruction);

[0028] α: Intensity coefficient of topographic influence (0≤α≤1);

[0029] β: Steepness parameter of hyperbolic function (β>0);

[0030] tanh: Hyperbolic tangent function, used to smoothly transition occlusion effects.

[0031] Preferably, the rapid fusion calculation formula in step S5 above is as shown in equation (7):

[0032] P fusion (x,y,t)=w radar (x,y,t)·P radar (x,y,t)+w sat (x,y,t)·P sat (x,y,t) (7)

[0033] Among them, P fusion : The merged precipitation background field;

[0034] P radar P sat Radar-estimated precipitation and satellite-retrieved precipitation estimates;

[0035] w radar (x,y,t),w sat (x,y,t): Radar quality index fusion weights, w radar (x,y,t)+w sat (x,y,t)=1.

[0036] Preferably, in step S1 above, the radar quality index (RQI) is calculated using the formula (1):

[0037]

[0038] Wherein, γ is the radar coverage factor, H is the radar beam height factor, H1 is the critical height characterizing the correlation between radar detection information and ground precipitation, with a value of 1000 meters, and H2 is the proportional factor of radar beam height change, with a value of 3500 meters. The closer to the center point, the lower the radar detection height, and the larger the RQI index.

[0039] Preferably, the formula for calculating the radar beam height factor H is as shown in equation (2):

[0040] H = H rad +D*tanα+β*E (2)

[0041] Among them, H rad α is the radar station height, D is the distance from the radar center to the radar network precipitation analysis grid point, α is the elevation angle of the lowest layer detected by the radar, β is the terrain effect factor, and E is the grid elevation (unit: m). The higher the elevation, the smaller the ROI index.

[0042] Preferably, the calculation formula for the radar coverage factor is as shown in equation (3):

[0043]

[0044] Where n is the number of radars covering the grid, and n0 takes the value 3, which means that when the number of radars covering the grid is less than 3, it will affect the RQI of that point. The fewer the number, the smaller the ROI index.

[0045] Preferably, in step S2 above, the climate proportion correction factor RFC is calculated using the ratio of historical ground observations and radar-estimated cumulative precipitation, and the calculation formula is as shown in equation (4):

[0046]

[0047] Among them, P i,j For ground-observed precipitation values ​​at specified grid locations, P Radar,i,j Estimated precipitation values ​​for radar at specified grid locations.

[0048] Preferably, the radar obstruction zone index RSA analysis field is an analysis field generated by processing the climate proportion correction factor RFC of radar precipitation calculated using the ratio of historical ground observations and radar-estimated cumulative precipitation over the past three months.

[0049] Preferably, the aforementioned radar and satellite precipitation data are used to generate corrected radar and satellite precipitation data using a bias correction method, including:

[0050] (1) Collection and preprocessing of ground station observation data, radar precipitation data, and satellite precipitation data:

[0051] Precipitation data from ground meteorological observation stations is collected and processed under quality control to obtain the quality-controlled ground observation precipitation data. The quality control is achieved using a single data source control module, a multi-source data collaborative quality control module, and a dynamic blacklist module. The single data source control module includes metadata checks, feature value checks, boundary value checks, and dead value checks, which can eliminate gross errors and long-term unchanging dead values ​​in real-time precipitation data. The multi-source data collaborative quality control module uses a QC algorithm developed based on the consistency of meteorological observation data related to radar and various weather phenomena, which can accurately identify false clear-sky precipitation and false zero-value data in typical precipitation areas that are difficult to identify with rapid quality control. The data, after being quality controlled by the single data source control module and the multi-source data collaborative quality control module, is further processed by the dynamic blacklist module to remove observation stations with high error rates and long error durations. A flexible dynamic evaluation mechanism is also used; data is removed from the blacklist only after its quality is restored. Data with quality control codes of 0, 1, 3, and 4 are selected for subsequent fusion analysis.

[0052] (2) Using the optimal interpolation method, the quality-controlled ground observation precipitation data are processed to generate a 1km ground grid analysis precipitation product, i.e., the ground precipitation grid analysis field:

[0053] 21) Establish a grid background field for precipitation climate values;

[0054] 22) Calculate the precipitation ratio data of each station and interpolate to generate the corresponding grid field. The ratio data is a new element defined with the help of the climate background field: Precipitation ratio = Station observed precipitation / Corresponding grid precipitation climate value;

[0055] 23) The precipitation grid field is generated by multiplying the precipitation ratio grid field with the corresponding climate background field. The interpolation method used in step 22) to generate the grid field is the optimal interpolation method, and the calculation formula (8) is as follows:

[0056]

[0057] That is, the analysis value A of the grid. k The initial estimate F at that point k Including the deviation between the observed value and the initial estimate at that point, the deviation is calculated from n known initial estimates F within the specified analysis range. i With the observed value O i The bias weighted estimate is obtained.

[0058] (3) PDF deviation correction:

[0059] 31) Perform consistency matching analysis between ground-observed precipitation and radar-estimated precipitation, including: (1) data preparation and time alignment, (2) consistency index calculation, and (3) determination of optimal lag time; analyze the difference between the cumulative precipitation observed by the ground automatic station in the first 10 minutes and the radar QPE at the lag time of 0 minutes during the precipitation process, and determine the optimal lag time using correlation coefficient, root mean square error, and relative deviation index;

[0060] 32) Construct a correction model and adjust the spatiotemporal matching window of the ground and radar precipitation PDF samples. The spatiotemporal matching window parameters are set to 1 hour and 35 km, and the minimum number of valid sample pairs participating in PDF matching is 120.

[0061] 33) Use the ground precipitation grid analysis field to perform PDF bias correction on radar network-estimated precipitation data.

[0062] The technical solution of the present invention achieves the following beneficial technical effects:

[0063] 1. This invention utilizes radar quality index and terrain obstruction coefficient to dynamically quantify the spatial reliability of radar data and optimize the fusion weights of radar and satellite data accordingly, overcoming the limitations of rigid weight allocation strategies in traditional methods. This innovation enables the fused background field to more accurately reflect actual precipitation conditions, especially in radar coverage blind spots and sparsely populated areas, effectively reducing problems such as local distortion, underestimation of precipitation intensity, or structural ambiguity, and significantly improving the reliability and operational applicability of the fusion product.

[0064] 2. The high-timeliness fused background field constructed by this invention not only provides a high-precision benchmark precipitation field for real-time analysis products, supporting precipitation phase identification, areal rainfall estimation, and flood risk assessment, but also serves as a rapid update assimilation field for numerical weather prediction models, improving key forecasting performance such as short-duration heavy precipitation and typhoon tracks. This technical solution effectively coordinates the complementary advantages of radar and satellite data, meeting the meteorological operational needs for a rapidly updated, highly consistent fused background field, and providing a solid technical guarantee for disaster prevention and mitigation decision-making and refined meteorological services. Attached Figure Description

[0065] Figure 1 This is the effect curve of the radar precipitation terrain obstruction factor (RSA) on the radar quality index (RQI) of this invention. Detailed Implementation

[0066] This embodiment presents a rapid radar-satellite precipitation fusion algorithm based on radar quality index, including:

[0067] S1. The radar quality index (RQI) is calculated by comprehensively considering the detection distance from the radar center, the number of coverage points, and the terrain height at the radar network precipitation analysis grid points. The specific formulas are shown in equations (1)-(3):

[0068]

[0069] H = H rad +D*tanα+β*E(2)

[0070]

[0071] Where γ is the radar coverage factor, H is the radar beam height factor, H1 is the critical height characterizing the correlation between radar detection information and ground precipitation (valued at 1000 meters), and H2 is the proportionality factor for radar beam height variation (valued at 3500 meters). The closer to the center point, the lower the radar detection height, and the larger the RQI index. radD is the radar station height, α is the distance from the radar center to the radar network precipitation analysis grid point, β is the elevation angle of the lowest layer detected by the radar, E is the grid elevation (unit: m), the higher the elevation, the smaller the ROI index; n is the number of radars covering the grid, n0 takes the value 3, which means that when the number of radars covering the grid is less than 3, it will affect the RQI of that point, the fewer the number, the smaller the ROI index;

[0072] S2. The climate scaling factor (RFC) of radar precipitation is calculated using the INCA algorithm. The RFC of the analysis grid is calculated by the ratio of the cumulative precipitation of historical ground observations and radar estimates over the past three months. The calculation formula is as shown in equation (4):

[0073]

[0074] Among them, P i,j For ground-observed precipitation values ​​at specified grid locations, P Radar,i,j Estimating precipitation values ​​for radar at specified grid locations;

[0075] S3. The Radar Shaded Area (RSA) index is designed as a piecewise function that varies with the climate proportional correction factor (RFC), as shown in equation (5):

[0076]

[0077] For RFC values ​​less than 1.0, assuming no occlusion, RSA is set to 0.0. For RFC values ​​greater than 1.0, assuming some degree of occlusion exists, and the higher the RFC, the more severe the occlusion, RSA is set to 1 - e. 0.05(rfc-1) For example, when RFC is greater than 15, RSA is approximately greater than 0.5, corresponding to severe obstruction. Additionally, grid points with zero long-term cumulative radar precipitation are considered to have absolute obstruction, and their RSA is set to 1.0.

[0078] S4. The radar terrain obstruction index analysis field further optimizes and corrects the weights of the radar quality index analysis field. The radar obstruction index RSA is applied to the radar quality index RQI using a hyperbolic function to obtain the optimized radar quality index RQI, as shown in equation (6):

[0079] RQI opt (x,y)=RQI(x,y)·[1-α·tanh(β·RSA(x,y))] (6)

[0080] Wherein, RQI(x,y): the original radar quality index (value range [0,1]);

[0081] RSA(x,y): Radar Obstruction Index (value range [0,1], 1 indicates complete obstruction);

[0082] α: Intensity coefficient of topographic influence (0≤α≤1);

[0083] β: Steepness parameter of hyperbolic function (β>0);

[0084] tanh: Hyperbolic tangent function, used to smoothly transition occlusion effects.

[0085] S5. Rapid Fusion of Radar and Satellite Precipitation: Radar and satellite precipitation data are rapidly fused using the Radar Quality Index (RQI) fusion weight to generate a radar-satellite fused background field. The calculation formula is shown in equation (7).

[0086] P fusion (x,y,t)=w radar (x,y,t)·P radar (x,y,t)+w sat (x,y,t)·P sat (x,y,t) (7)

[0087] Among them, P fusion : The merged precipitation background field;

[0088] P radar P sat Radar-estimated precipitation and satellite-retrieved precipitation estimates;

[0089] w radar (x,y,t),w sat (x,y,t): Radar quality index fusion weights, w radar (x,y,t)+w sat (x,y,t)=1.

[0090] The present invention is provided in the following information for specific implementation.

[0091] In most areas east of 105°E and south of 42°N on land in China, the RQI value is basically 1. However, the RQI values ​​are relatively low in areas such as northern and western China, the Qinghai-Tibet Plateau and its surrounding southeastern regions, and the eastern and southern coastal areas, which are farther from the radar center, have fewer radars covering them, and have complex terrain.

[0092] In this embodiment, radar and satellite precipitation data can be corrected using a bias correction method to generate corrected radar and satellite precipitation data, specifically:

[0093] (1) Collection and preprocessing of ground station observation data, radar precipitation data, and satellite precipitation data:

[0094] Precipitation data from ground meteorological observation stations is collected and processed under quality control to obtain the quality-controlled ground observation precipitation data. The quality control is achieved using a single data source control module, a multi-source data collaborative quality control module, and a dynamic blacklist module. The single data source control module includes metadata checks, feature value checks, boundary value checks, and dead value checks, which can eliminate gross errors and long-term unchanging dead values ​​in real-time precipitation data. The multi-source data collaborative quality control module uses a QC algorithm developed based on the consistency of meteorological observation data related to radar and various weather phenomena, which can accurately identify false clear-sky precipitation and false zero-value data in typical precipitation areas that are difficult to identify with rapid quality control. The data, after being quality controlled by the single data source control module and the multi-source data collaborative quality control module, is further processed by the dynamic blacklist module to remove observation stations with high error rates and long error durations. A flexible dynamic evaluation mechanism is also used; data is removed from the blacklist only after its quality is restored. Data with quality control codes of 0, 1, 3, and 4 are selected for subsequent fusion analysis.

[0095] (2) Using the optimal interpolation method, the quality-controlled ground observation precipitation data are processed to generate a 1km ground grid analysis precipitation product, i.e., the ground precipitation grid analysis field:

[0096] 21) Establish a grid background field for precipitation climate values;

[0097] 22) Calculate the precipitation ratio data of each station and interpolate to generate the corresponding grid field. The ratio data is a new element defined with the help of the climate background field: Precipitation ratio = Station observed precipitation / Corresponding grid precipitation climate value;

[0098] 23) The precipitation grid field is generated by multiplying the precipitation ratio grid field with the corresponding climate background field. The interpolation method used in step 22) to generate the grid field is the optimal interpolation method, and the calculation formula (8) is as follows:

[0099]

[0100] That is, the analysis value A of the grid. k The initial estimate F at that point k Including the deviation between the observed value and the initial estimate at that point, the deviation is calculated from n known initial estimates F within the specified analysis range. i With the observed value O i The bias weighted estimate is obtained.

[0101] (3) PDF deviation correction:

[0102] 31) Perform consistency matching analysis between ground-observed precipitation and radar-estimated precipitation, including: (1) data preparation and time alignment, (2) consistency index calculation, and (3) determination of optimal lag time; analyze the difference between the cumulative precipitation observed by the ground automatic station in the first 10 minutes and the radar QPE at the lag time of 0 minutes during the precipitation process, and determine the optimal lag time using correlation coefficient, root mean square error, and relative deviation index;

[0103] 32) Construct a correction model and adjust the spatiotemporal matching window of the ground and radar precipitation PDF samples. The spatiotemporal matching window parameters are set to 1 hour and 35 km, and the minimum number of valid sample pairs participating in PDF matching is 120.

[0104] 33) Use the ground precipitation grid analysis field to perform PDF bias correction on radar network-estimated precipitation data.

[0105] The spatial distribution maps of 24-hour cumulative precipitation from radar (corrected for 10-minute intervals at 1 km) and satellite (corrected for 10-minute intervals at 1 km) on July 31, 2021, are presented. Using RQI as the fusion weighting coefficient for radar precipitation, FY4 satellite data is linearly weighted and incorporated, filling in areas not covered by radar. This optimizes the background field in sparsely populated areas, creating a complete radar-satellite fused background field covering China and effectively utilizing satellite data to improve the monitoring capability of typhoon precipitation systems in near-shore areas. Furthermore, the RQI-based weighted fusion method significantly reduces computation time compared to the BMA radar-satellite background field fusion method used in the three-source fusion scheme of smaller time-based products, making it more suitable for the timeliness requirements of minute-level real-time product development.

[0106] Based on the cumulative precipitation data of 10 minutes from July 20 to 31, 2021, observed by more than 2,400 national automatic weather stations, the optimization effect of introducing minute-level FY4 satellite precipitation data on the precipitation background field was independently tested and evaluated. The results are shown in Table 1.

[0107] Table 1. Error Statistics of Precipitation Test Products at 1 km Per 10 Minutes, July 20-31, 2021

[0108]

[0109] Nationwide, although the accuracy of FY4 precipitation data is much lower than that of radar precipitation products, the quality of the radar combined background field is still improved to some extent compared with single-source radar precipitation products, such as an increase in correlation coefficient (CC) and a decrease in root mean square error (RMSE).

[0110] The Radar Shaded Area (RSA) index is established based on the radar climate correction factor (RFC). The RSA index is designed as a piecewise function that varies with the climate correction factor (RFC). For RFC values ​​less than 1.0, it is assumed that there is no shading effect, and RSA is set to 0.0. For RFC values ​​greater than 1.0, it is assumed that there is some degree of shading, with the degree increasing as RFC increases. For example, when RFC is greater than 15, RSA is approximately greater than 0.5, corresponding to severe shading. Furthermore, grid points with zero long-term cumulative radar precipitation are considered to have absolute shading, and RSA is set to 1.0. The resulting radar precipitation topographic shading data shows that areas with severe radar topographic shading are mainly located in the western and northeastern regions, with Gansu, Tibet, Yunnan, Sichuan, and Xinjiang experiencing particularly severe shading. Areas with absolute shading calculated based on RSA are also primarily located in these regions.

[0111] By applying the Radar Obstruction Area Index (RSA) to the Radar Quality Index (RQI) using a hyperbolic function, an optimized RQI is obtained. Figure 1 When the obstruction factor is low, the radar quality index changes little. When the obstruction factor is greater than 0.7, the radar quality index drops rapidly, and when the obstruction factor approaches 1, the radar quality index approaches 0. It can be seen that after introducing the influence of terrain obstruction, the RQI index decreases significantly in northern Northeast China, Xinjiang, eastern and surrounding areas of the Qinghai-Tibet Plateau, and Southwest China, making the radar quality index more scientifically sound.

[0112] The quality of the background field before and after the introduction of the radar obstruction zone index (RSA) weighting into the radar quality index (RQI) was independently verified and evaluated using on-site precipitation observations. The results are shown in Table 2.

[0113] Table 2. Impact of Incorporating Radar Obstruction Area Index (RSA) into the Radar Quality Index (RQI) Weighting on the Background Field

[0114]

[0115] Overall, using radar quality index to fuse radar and satellite precipitation data can significantly improve the background field quality. Furthermore, the radar and satellite background field quality fused using the optimized radar quality index is even higher, with the relative deviation decreasing from -2.6% to -0.2%.

[0116] In summary, this invention utilizes the Radar Quality Index (RQI) to weightedly fuse radar bias-corrected precipitation and satellite bias-corrected precipitation, rapidly generating a higher-quality fused background field. Furthermore, it introduces the Radar Obstruction Area Index (RSA, Terrain Obstruction Factor) to optimize the RQI, and compares and evaluates the fusion effects before and after optimization. The overall national evaluation results show that both the RQI-based weighted fusion method and the method optimized with RSA can improve the quality of the fused background field to some extent, but the optimized method is more effective. While the RQI weighted fusion method can utilize satellite data to compensate for insufficient radar coverage and improve background field quality, certain biases still exist in areas with complex terrain and radar coverage edges. The RQI weighted fusion method with RSA optimization, by dynamically quantifying the spatial heterogeneity of radar quality and considering the impact of terrain obstruction on radar detection performance, significantly improves the accuracy of the fused background field in key areas (such as terrain transition zones and radar coverage blind spots). The fusion results were independently verified by retaining precipitation observation data from over 2,400 national-level automatic weather stations. Statistical analysis showed that the optimized fusion method demonstrated superior improvement in areas with low radar quality and complex terrain. In areas with low radar quality, the relative bias of the fused background field improved from -2.6% to -0.2%. From the spatial distribution and statistical analysis of the fused precipitation, the optimized Radar Quality Index (RQI) was used for rapid weighted fusion of radar bias-corrected and satellite bias-corrected precipitation, compensating for insufficient radar coverage and significantly improving background field quality. This resulted in better consistency with station observations, enhancing the reliability and operational applicability of the fused background field.

[0117] This invention applies radar quality index analysis fields and radar terrain obstruction index analysis fields. On the one hand, it dynamically quantifies the spatial heterogeneity of radar quality by introducing the radar obstruction area index (RSA) to optimize the radar quality index (RQI), thereby achieving dynamic adjustment of the radar and satellite fusion weights and overcoming the limitations of static weights and single filling strategies. On the other hand, the optimized fusion method is more suitable for the timeliness requirements of minute-level real-time product development, enabling rapid response to extreme weather processes. It provides a highly timely and consistent fusion background field for disaster prevention and mitigation decision-making and refined meteorological services, promoting the overall improvement of meteorological services in short-term forecasting, disaster early warning, and model assimilation.

[0118] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of the claims of this patent application.

Claims

1. A rapid fusion algorithm for radar satellite precipitation based on radar quality index, characterized in that, Using the climate proportionality correction factor RFC, the radar obstruction area index RSA is designed as a piecewise function that varies with the climate proportionality correction factor RFC. The radar terrain obstruction index analysis field is used to optimize the radar quality index analysis field weights. Then, the radar quality index fusion weights are used to quickly fuse radar and satellite precipitation data to generate a radar-satellite fused background field. Includes the following steps: S1. The radar quality index (RQI) is calculated by combining the detection distance from the radar center, the number of coverage points, and the terrain height at the radar network precipitation analysis grid points. S2. Correction is performed by calculating the climate proportion correction factor (RFC) of radar precipitation using the INCA algorithm; S3. The radar obstruction zone index RSA is designed as a piecewise function that varies with the climate proportional correction factor RFC. S4. The weights of the radar quality index analysis field are further optimized and corrected using the radar obstruction zone index analysis field. The radar obstruction zone index RSA is applied to the radar quality index RQI using a hyperbolic function to obtain the optimized RQI, as shown in equation (6): （6） Where RQI(x,y): the original radar quality index, with a value range of [0,1]; RSA(x,y): Radar Blockage Index, with a value range of [0,1], where 1 indicates complete blockage; α: Intensity coefficient of terrain influence, 0≤α≤1; β: kurtosis parameter of the hyperbolic function, β>0; tanh: Hyperbolic tangent function, used to smoothly transition occlusion effects; S5. Rapidly fuse radar and satellite precipitation data using radar quality index fusion weights to generate a radar-satellite fused background field; The piecewise function in step S3 is as shown in equation (5): （5） When RFC is less than 1.0, let RSA = 0.0; when RFC is greater than or equal to 1.0, ;when , that is, the grid point where the long-term cumulative radar precipitation is zero, let rsa=1.

0.

2. The radar satellite precipitation fast fusion algorithm based on radar quality index according to claim 1, characterized in that, The rapid fusion calculation formula in step S5 is as follows: (7) （7） in, : The merged precipitation background field; , Radar-estimated precipitation and satellite-retrieved precipitation estimates; , Radar quality index fusion weights .

3. The radar-satellite precipitation fast fusion algorithm based on radar quality index according to claim 1, characterized in that, In step S1, the radar quality index (RQI) is calculated using the formula (1): （1） Wherein, γ is the radar coverage factor, H is the radar beam height factor, H1 is the critical height characterizing the correlation between radar detection information and ground precipitation, with a value of 1000 meters, and H2 is the proportional factor of radar beam height change, with a value of 3500 meters. The closer to the center point, the lower the radar detection height, and the larger the RQI index.

4. The radar-satellite precipitation fast fusion algorithm based on radar quality index according to claim 3, characterized in that, The formula for calculating the radar beam height factor H is as follows: (2) （2） in, α is the radar station height, D is the distance from the radar center to the grid point of the radar network precipitation analysis, α is the elevation angle of the lowest layer detected by the radar, β is the terrain effect factor, and E is the grid elevation in meters. The higher the elevation, the smaller the ROI index.

5. The radar-satellite precipitation fast fusion algorithm based on radar quality index according to claim 3, characterized in that, The formula for calculating the radar coverage factor is as follows: (3) （3） Where n is the number of radars covering the grid, and n0 takes the value 3, which means that when the number of radars covering the grid is less than 3, it will affect the RQI of that point. The fewer the number, the smaller the ROI index.

6. The radar-satellite precipitation fast fusion algorithm based on radar quality index according to claim 1, characterized in that, In step S2, the climate proportion correction factor RFC is calculated using the ratio of historical ground observations and radar-estimated cumulative precipitation, as shown in formula (4): （4） in, For ground-observed precipitation values ​​at specified grid locations, Estimated precipitation values ​​for radar at specified grid locations.

7. The radar-satellite precipitation fast fusion algorithm based on radar quality index according to claim 1, characterized in that, The radar obstruction zone index (RSA) analysis field is generated by processing the climate proportion correction factor (RFC) of radar precipitation calculated using the ratio of historical ground observations and radar-estimated cumulative precipitation over the past three months.

8. The radar-satellite precipitation fast fusion algorithm based on radar quality index according to claim 1, characterized in that, The radar and satellite precipitation data are used to generate corrected radar and satellite precipitation data using a bias correction method, including: (1) Collect and preprocess ground station observation data, radar precipitation data, and satellite precipitation data: Precipitation data from ground meteorological observation stations is collected and processed under quality control to obtain the quality-controlled ground observation precipitation data. The quality control is achieved using a single data source control module, a multi-source data collaborative quality control module, and a dynamic blacklist module. The single data source control module includes metadata checks, feature value checks, boundary value checks, and dead value checks, which can eliminate gross errors and long-term unchanging dead values ​​in real-time precipitation data. The multi-source data collaborative quality control module uses a QC algorithm developed based on the consistency of meteorological observation data related to radar and various weather phenomena, which can accurately identify false clear-sky precipitation and false zero-value data in typical precipitation areas that are difficult to identify with rapid quality control. The data, after being quality controlled by the single data source control module and the multi-source data collaborative quality control module, is further processed by the dynamic blacklist module to remove observation stations with high error rates and long error durations. A flexible dynamic evaluation mechanism is also used; data is removed from the blacklist only after its quality is restored. Data with quality control codes of 0, 1, 3, and 4 are selected for subsequent fusion analysis. Using the optimal interpolation method, the quality-controlled ground-observed precipitation data is processed to generate a 1km ground-grid analysis precipitation product, i.e., a ground-grid analysis field: 21) Establish a grid background field for precipitation climate values; 22) Calculate the precipitation ratio data for each station and interpolate to generate the corresponding grid field. The ratio data is a new element defined with the help of the climate background field: Precipitation ratio = Station observed precipitation / Corresponding grid precipitation climate value; 23) The precipitation grid field is generated by multiplying the precipitation ratio grid field with the corresponding climate background field. The interpolation method used in step 22) to generate the grid field is the optimal interpolation method, and the calculation formula (8) is as follows: （8） That is, the analysis value A of the grid. k The initial estimate F at that point k Including the deviation between the observed value and the initial estimate at that point, the deviation is calculated from n known initial estimates F within the specified analysis range. i With the observed value O i The bias-weighted estimate is obtained; PDF deviation correction: 31) Perform consistency matching analysis between ground-observed precipitation and radar-estimated precipitation, including: (1) data preparation and time alignment, (2) consistency index calculation, and (3) determination of optimal lag time; analyze the difference between the cumulative precipitation observed by the ground automatic station in the first 10 minutes and the radar QPE at the lag time of 0 minutes during the precipitation process, and use the correlation coefficient, root mean square error, and relative deviation index to determine the optimal lag time; 32) Construct a correction model and adjust the spatiotemporal matching window of the ground and radar precipitation PDF samples. The spatiotemporal matching window parameters are set to 1 hour and 35 km, and the minimum number of valid sample pairs participating in PDF matching is 120. 33) Use the ground precipitation grid analysis field to perform PDF bias correction on radar network-estimated precipitation data.

Images