A method for determining the availability of spatial characteristics of multi-source precipitation products

Abstract

This invention discloses a method for determining the availability of spatial characteristics of multi-source precipitation products. The method includes the following steps: Step 1, acquiring basic data of the study area and constructing an estimated field and a measured field; Step 2, constructing an evaluation index system for multi-source precipitation products; Step 3, quantifying the missing measurement range of spatial characteristics of different precipitation products; Step 4, determining the availability of spatial characteristics of each precipitation product. This method, by analyzing and quantifying the missing measurement range of spatial characteristics of ground rain gauges and "sky" and "air" precipitation products, clarifies under what conditions "sky" and "air" precipitation products can effectively supplement these missing spatial characteristic ranges, achieving the determination of the availability of spatial characteristics of precipitation products, thereby supporting the optimized integration of multi-source precipitation products and improving the monitoring accuracy of the spatial distribution of precipitation fields.

Description

Technical Field

[0001] This invention belongs to the field of precipitation data processing technology, and in particular relates to a method for determining the availability of spatial characteristics of multi-source precipitation products. Background Technology

[0002] In mountainous small and medium-sized watersheds (drainage area ≤ 2000 km²) 2 The complex terrain results in a significantly lower density of surface rain gauges compared to plains areas. According to data from the water resources industry standard "SL 767-2018 Technical Specification for Investigation and Evaluation of Mountain Flood Disasters," the typical station density in mountainous areas is only 1 station per 150-200 km². 2 In plains areas, this can reach 1 station / 50km. 2 Low station density leads to a missed detection rate exceeding 30% for localized heavy rainfall centers. Because the disaster-causing precipitation threshold is extremely low in mountainous areas, even short-duration heavy rainfall of only 20-40 mm / h can trigger flash floods, and the false alarm rate for floods is even higher in areas with sparse stations. Therefore, improving the accuracy of precipitation data is a core requirement for disaster early warning in small watersheds in mountainous areas.

[0003] For areas with sparse ground-based rain gauges, the most effective way to improve their accuracy is to integrate data from satellite / radar precipitation products. However, current methods for evaluating the accuracy of precipitation products rely too heavily on ground-based station data as a benchmark. Their evaluation results only reflect the matching error of point grid cells (such as RMSE and MAE), failing to quantify the two-dimensional spatial structure characteristics of the precipitation field (such as rainband morphology, gradient distribution, and spatial autocorrelation). This limitation results in a lack of effective evaluation benchmarks for the spatial continuity of satellite / radar precipitation product data, thus restricting accuracy optimization in sparsely populated areas.

[0004] Therefore, how to analyze the missing range of spatial characteristics of ground-based measured rainfall data, clarify under what conditions "sky" and "air" precipitation products can effectively supplement the missing range of spatial characteristics, and realize the usability discrimination of spatial characteristics of multi-source precipitation products are the technical problems that urgently need to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide a method for determining the availability of spatial characteristics of multi-source precipitation products, so as to solve the above-mentioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention discloses a method for determining the availability of spatial characteristics of multi-source precipitation products, the method comprising the following steps:

[0008] Step 1: Obtain basic data for the study area and construct the estimation field and the measured field: First, select a small watershed with dense rain gauges as the study area. Define the area with rain gauge density less than or equal to the set value D1 as a dense rain gauge area, and define the area with rain gauge density greater than the set value D1 as a sparse rain gauge area. The small watershed is a closed catchment unit divided by the watershed and the river outlet. Then, obtain basic data for the study area, including "sky," "air," and "ground" precipitation product data, namely satellite precipitation product data, radar precipitation product data, and measured rainfall data from ground rain gauges. Then, based on the obtained basic data, construct the measured field and the estimated fields for each precipitation product, namely the satellite precipitation product estimated field, the radar precipitation product estimated field, and the ground rain gauge estimated field.

[0009] Step 2: Construct a multi-source precipitation product evaluation index system: Using the measured field as a benchmark, the accuracy of the estimated rainfall center location and spatial gradient of each precipitation product's estimated field is evaluated using both categorical and continuous indices. Specifically, the categorical index uses the rainfall center location hit rate (HR) to assess the accuracy of the estimated rainfall center location for each precipitation product's estimated field; the continuous index uses the spatial structure similarity index (SSIM) to quantify the accuracy of the spatial gradient between the estimated and measured rainfall fields for each precipitation product. Given that single calculation results are affected by random errors, the evaluation index calculation results for multiple samples are averaged under the same operating conditions, i.e., the average rainfall center location hit rate is calculated separately. And the average spatial structure similarity index

[0010] Step 3: Quantify the spatial characteristic gaps of different precipitation products: Quantify the spatial characteristic gaps of surface rain gauges and "sky" and "air" precipitation products separately, including the following steps:

[0011] Step 31: Quantify the spatial feature gap range of ground rain gauges: Based on the constructed multi-source precipitation product evaluation index system, the influence of station network density and distribution on the ability to capture spatial features of precipitation fields is quantified by the hierarchical control variable method, thereby quantifying the spatial feature gap range of ground rain gauges;

[0012] Specifically, the process involves: first, constructing different distribution schemes for rainfall stations, including uniform distribution, topographic gradient distribution, and random distribution; then, based on these schemes, using a stratified controlled variable method, randomly sampling different numbers of stations to form three different density distribution schemes: high density, medium density, and low density, resulting in a total of nine distribution schemes; finally, calculating the average rainfall center location hit rate for the estimated field under each of the different distribution schemes. And the average spatial structure similarity index Finally, the spatial feature gaps in ground rain gauge measurements were determined: Rainfall field monitoring accuracy under different station network distribution schemes was plotted and fitted as a function of station network density and uniformity, and the spatial feature monitoring accuracy was set accordingly. or The threshold is denoted as Y. When the monitoring accuracy is lower than the threshold Y, it indicates that the ground rain gauge has missed measuring the spatial characteristics of the rainfall field under the conditions of the station network. Then, the range of missing measurements of the spatial characteristics of the rainfall field by the ground rain gauge can be summarized.

[0013] Step 32: Quantify the spatial feature gaps of "sky" and "air" precipitation products: Based on the constructed multi-source precipitation product evaluation index system, quantify the spatial monitoring accuracy of satellite precipitation products and radar precipitation products under different observation conditions, and then quantify the spatial feature gaps of satellite precipitation products and radar precipitation products.

[0014] Specifically, for radar precipitation products, the average accuracy of the rainfall center location in the estimated field of the radar precipitation product is calculated based on the measured field. And the average spatial structure similarity index Then, the accuracy of radar precipitation products under different radar detection radii is obtained. The accuracy points corresponding to each radar detection radius are fitted into a radar detection radius-accuracy relationship curve to characterize the variation law of radar precipitation product accuracy with radar detection radius, thereby clarifying its spatial supplementation capability and applicable range. According to the monitoring accuracy threshold Y in step 31, the radar detection radius corresponding to the accuracy threshold is determined in the above curve and recorded as the radar accuracy threshold point. Using the radar monitoring capability zoning method, a three-level zoning is constructed based on the radar accuracy threshold point:

[0015] Zone I, or Independent Effective Zone: Within 10% of the threshold radius, radar data meets the preset accuracy threshold and can independently support the monitoring of spatial characteristics of the rainfall field;

[0016] Zone II is the Coordination and Supplementation Zone: the reliability of radar data in this zone is attenuated, and it is necessary to integrate satellite or ground station data to enhance it;

[0017] Zone III is the invalid zone: the radar data accuracy in this zone drops below the threshold and no longer has the ability to characterize spatial features;

[0018] The specific partitioning method is as follows:

[0019] The simplified expression for the relationship between radar accuracy and radius is:

[0020] P R =f R (r)(9)

[0021] When P R When =Y, the corresponding r obtained is r0;

[0022] The expressions for each region can also be obtained:

[0023]

[0024] In the formula, P R Indicates radar accuracy; f R This represents the radar accuracy function; r0 is the threshold radius, which refers to the effective detection radius of the radar when the accuracy threshold requirement of the ground rain gauge is met; μ is 10% of the threshold radius, i.e., μ = 0.1r0;

[0025] For satellite precipitation products, the average accuracy of the precipitation center location is calculated based on the measured field. And the average spatial structure similarity index The accuracy of the corresponding satellite precipitation product is calculated, which is a fixed value. By comparing this accuracy with the monitoring accuracy of radar and ground rain gauges, the missing measurement range of the satellite precipitation product is determined. That is, if its value is lower than the accuracy of radar precipitation products or ground rain gauges, it is determined that there is a missing measurement of satellite precipitation products in that area.

[0026] Step 4: Determine the availability of spatial features of each precipitation product: Based on the spatial missing measurement range of each precipitation product identified in Step 3, a coordinate transformation method is used to determine the availability of spatial features of each precipitation product by fixing points, lines, and zones.

[0027] Furthermore, the specific process of constructing the measured field and the estimated field of each precipitation product based on the acquired basic data in step 1 is as follows:

[0028] For the construction of estimation fields for each precipitation product:

[0029] Ground rain gauge estimation field: To simulate sparse areas of ground rain gauges, a small number of ground rain gauges in the study area are selected according to the principle of stratified random sampling to construct sparse rain gauge networks with different distributions. For each network distribution scheme, the Kriging interpolation spatial interpolation method is selected to construct the rainfall estimation field, which is the ground rain gauge estimation field.

[0030] Radar precipitation product estimation field: The measurement data under different radar detection radii are recorded as the radar precipitation product estimation field;

[0031] Satellite precipitation product estimation field: The satellite precipitation product estimation field is generated by using a grid resampling method to obtain satellite data of the study area;

[0032] For the construction of the test field:

[0033] Using the measured rainfall data from the remaining ground rain gauges after extraction, a near-real rainfall field, i.e., the measured field, is constructed. The specific construction process is as follows:

[0034] Let the location coordinates of each rain gauge be (xa y a (a = 1, 2, 3…), the actual measured rainfall is Z. a A near-realistic rainfall field is constructed using the inverse distance weighted interpolation method, that is, for any grid point (x... g y g The estimated rainfall was calculated using the inverse distance weighted interpolation method. The formula is as follows:

[0035]

[0036] In the formula, Z a ω represents the measured rainfall at the a-th rain gauge station; a For the a-th rain gauge station, pair it with grid point (x) g y g The weight of ); d a For the a-th rain gauge station to the grid point (x g y g The distance is m, which is the number of rain gauges involved in the calculation, determined by the search radius or the number of nearest neighbors; p is a power parameter, which is a constant.

[0037] Furthermore, the set value D1 for the rain gauge density in step 1 is 25 km². 2 / stand.

[0038] Furthermore, the formula for calculating the rainfall center location hit rate (HR) in step 2 is as follows:

[0039]

[0040] Among them, I i The function that indicates the location of the rainfall center has the following value:

[0041]

[0042] In the formula, d i δ represents the straight-line distance between the measured and estimated rainfall centers in time period i; δ represents the allowable error distance; and N represents the total number of time periods selected.

[0043] The formula for calculating the Spatial Structure Similarity Index (SSIM) is as follows:

[0044]

[0045] In the formula, x is the normalized measured field; y is the normalized estimated field; μ x and μ y These are the means of two normalized rainfall fields, respectively; σ x and σ y σ represents the variance of the two normalized rainfall fields. xyLet C1 be the covariance of the two normalized precipitation fields; where the constant is C1 = (0.01L). 2 C2 = (0.03L) 2 L is the dynamic range of the normalized rainfall field. When the rainfall data is normalized to the interval [0,1], L = 1.

[0046] The average rainfall center location hit rate The calculation formula is:

[0047]

[0048] The average spatial structure similarity index The calculation formula is:

[0049]

[0050] In the formula, M is the sample size under the same working conditions; HR j and SSIM j These are the rainfall center location hit rate and spatial structure similarity index value for the j-th sample, respectively.

[0051] Furthermore, the specific method for constructing different distribution schemes of rain gauge networks, including uniform distribution, topographic gradient distribution, and random distribution, as described in step 31, is as follows: The Moran index is used to quantify the uniformity of rain gauge station distribution, converting station locations into spatial density indicators to characterize the clustering and dispersion of stations in different schemes. The Moran index formula is as follows:

[0052]

[0053] In the formula, n is the total number of stations; x i x j These are the precipitation amounts at the i-th and j-th stations, respectively. This represents the average precipitation; w ij Let be the spatial weight matrix, representing the spatial relationship between the i-th and j-th stations; the closer the Moran index is to 0, the more uniform the station distribution; when the Moran index is in the range of [-0.1, 0.1], it is considered a uniform distribution; when the absolute value of the Moran index is in the range of (0.1, 0.7], it is considered a terrain gradient distribution; when the absolute value of the Moran index is in the range of (0.7, 1], it is considered a random distribution.

[0054] Furthermore, the high density mentioned in step 31 is <20km². 2 / station; medium density is 20-30km 2 / station; low density is >30km 2 / stand.

[0055] Furthermore, the specific process described in step 4, which uses a coordinate transformation method to determine the availability of spatial characteristics of each precipitation product by fixing points, lines, and zones, is as follows:

[0056] First, a new coordinate system is constructed, with the station control area ρ as the abscissa and the radar detection radius r as the ordinate. Based on this new coordinate system, to uniformly incorporate the accuracy influencing factors of rain gauges and radar products obtained in the previous steps into the analysis, spatial transformations are performed on the monitoring performance maps of the two types of products respectively. Among them, the rain gauge monitoring performance map is obtained in step 31, with the abscissa being the station control area ρ and the ordinate being the average rainfall center location hit rate. Surface rain gauges The relationship curve between the control area ρ of the site and the curve is used It indicates that radar precipitation products The relationship curve between the control area ρ of the site and the curve is used This indicates that, since the performance of radar precipitation products is not affected by the station's controlled area ρ, it is displayed as a horizontal straight line; the radar monitoring performance map is obtained from step 32, and the horizontal axis of the radar monitoring performance map represents the average rainfall center location hit rate. The vertical axis represents the radar detection radius r; radar precipitation products The curve showing the relationship between radar detection range r and the radar range is used. It indicates that satellite precipitation products The relationship curve between the radar detection radius r and the radar detection radius r is used This means that since the performance of satellite precipitation products is not affected by the radar detection radius r, it is displayed as a vertical straight line. Then, with the origin of the new coordinate system as a reference, the rain gauge monitoring performance map is shifted in the negative direction of the vertical axis so that its origin falls below the vertical axis of the new coordinate system. At the same time, the radar monitoring performance map is shifted in the positive direction of the horizontal axis so that its origin is located to the right of the horizontal axis of the new coordinate system. The shift distance is set reasonably to ensure that the three do not overlap in the coordinate space and to avoid graphic intersection.

[0057] Then, based on the coordinate transformation method, the availability of spatial characteristics of each precipitation product can be determined by fixing points, lines, and zones. The specific steps are as follows:

[0058] Step 41, Point Determination: This involves identifying the key dividing points for spatial monitoring accuracy of various precipitation products; first, determine the dividing points of the radar's applicable range, and then, in the radar monitoring performance map, through... and The intersection point determines the radius that ensures consistent accuracy between radar and satellite precipitation products; the intersection point is denoted as B. r The radius corresponding to this point is called the critical radius, denoted as r2. Based on the precision corresponding to this critical radius, in The control area of ​​the station at this level of accuracy is determined by the following formula:

[0059]

[0060] in,

[0061]

[0062] P O,ρ ×f O (ρ) (13)

[0063] In the formula, P represents the accuracy value corresponding to a radar radius of r2. O,ρ f represents the accuracy value corresponding to the control area of ​​a ground rain gauge station being ρ. R g O (ρ) represent the accuracy function of the ground rain gauge and the accuracy function of the radar, respectively;

[0064] The point obtained at this point is denoted as B. ρ , past B ρ Draw a vertical line through B r Draw a horizontal line. The intersection of the two straight lines in the new coordinate system is a key dividing point of the radar's applicable range, denoted as point B.

[0065] The next key point starts with the accuracy threshold of the ground rain gauge, which corresponds to the monitoring accuracy threshold Y in step 31. The intersection of the horizontal line drawn from this threshold and the accuracy curve of the ground rain gauge is the critical point A. ρ The radius corresponding to this radar accuracy is obtained by the following formula:

[0066] P R =f R (r)=Y (14)

[0067] The radius obtained from the above formula is denoted as the effective radius r1, which corresponds to the radar monitoring performance diagram. A of the curve r Point; similarly, through A ρ Draw a vertical line through A r Draw a horizontal line. The intersection of the two straight lines in the new coordinate system is point A, which represents a key dividing point of the applicable range of the ground rain gauge.

[0068] The third key point C is determined by the density of ground rain gauges. When the controlled area of ​​a station is less than or equal to the set value D2, it is considered that the ground rain gauge does not need to integrate other data sources. Therefore, in the new coordinate system, this point is initially set at the position with the horizontal coordinate D2. Furthermore, if the key point C is located to the right of the key point B after initial determination, in order to ensure the consistency of the judgment logic and the rationality of the diagram structure, the position of the key point C is adjusted to be directly below the key point B.

[0069] Step 42, Line Determination: This involves constructing the boundary lines for the spatial availability of precipitation products. Based on the two boundary points A and B obtained in the point determination step, horizontal lines L1 and L2 are drawn through these two points respectively. In the area above the horizontal line L1 passing through point A, only satellite data is available. In the area between the two horizontal lines, both satellite and radar data are available. In the area below the horizontal line L2 passing through point B, only radar data is available.

[0070] Then, based on the spatial characteristics of the missing measurement range of the ground rain gauge, a boundary line for supplementary availability is drawn; a vertical line L3 is drawn through the boundary point of the applicable range of the ground rain gauge, i.e., point A. The area to the right of this line needs to utilize data from satellite and radar precipitation products to improve the accuracy of the ground rain gauge; the accuracy of the ground rain gauge in the area to the left of this line has reached the threshold requirement.

[0071] In respectively A in the curve ρ B ρ Take several points from the segment, and then... A in the curve r B r Find the point in the segment that corresponds to its accuracy. The specific method is the same as the method for finding points A and B above. Draw a vertical line on the point in the rain gauge monitoring performance map and a horizontal line on the point in the radar monitoring performance map. Fit all their intersections into a curve, which is called the AB curve. The left side of this curve indicates that the ground rain gauge has the ability to monitor spatial characteristics, and the right side indicates that satellite and radar precipitation products are needed to make up for the lack of measurement of the spatial characteristics of the precipitation field by the ground rain gauge.

[0072] Step 43, Partitioning: After defining the points and lines, the newly constructed coordinate system is divided into four regions:

[0073] "Ground" area: refers to the area located to the left of curve ABC and straight line L3, where ground rain gauges do not need to rely on data from satellite or radar precipitation products for supplementation;

[0074] "Ground + Air" area: refers to the area to the right of line BC and below line L2. Within this area, the ability to monitor the spatial characteristics of rainfall needs to be improved by combining ground rain gauge data and radar rainfall data.

[0075] The "ground + sky + air" region is the area to the right of curve AB, between straight lines L1 and L2. In this region, it is necessary to combine three rainfall data sources: ground rain gauges, radar, and satellites, in order to monitor the spatial distribution characteristics of rainfall.

[0076] "Ground + Sky" region: refers to the area located to the right of line L3 and above line L1. In this region, ground rain gauges and satellite data are used to reflect the spatial characteristics of rainfall.

[0077] Step 44: Dynamic Adjustment: Adjust the data utilization strategy in real time according to different actual application needs and environmental changes.

[0078] Furthermore, the set value D2 in step 41 is 20 km. 2 / stand.

[0079] The beneficial effects of the present invention are as follows: The method described in the present invention analyzes and quantifies the missing measurement range of spatial characteristics of ground rain gauges and "sky" and "air" precipitation products, clarifies under what conditions "sky" and "air" precipitation products can effectively supplement these missing measurement ranges of spatial characteristics, realizes the usability judgment of the spatial characteristics of precipitation products, thereby supporting the optimized fusion of multi-source precipitation products and improving the monitoring accuracy of the spatial distribution of precipitation fields.

[0080] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0081] Figure 1 This is a flowchart of the method described in this invention;

[0082] Figure 2 A schematic diagram illustrating the quantification of the missing measurement range of spatial characteristics of rainfall fields from ground rain gauges;

[0083] Figure 3 A graph showing the radius-accuracy relationship of radar precipitation products;

[0084] Figure 4 This is a schematic diagram illustrating the process of determining the availability of spatial characteristics of multi-source precipitation products. Detailed Implementation

[0085] This invention discloses a method for determining the availability of spatial characteristics of multi-source precipitation products, such as... Figure 1 As shown, the method includes the following steps:

[0086] Step 1: Obtain basic data of the study area and construct the estimated field and the measured field.

[0087] First, a small watershed with a high density of rain gauge stations was selected as the study area. Rain gauge station density was defined as 25 km² or less. 2 The area defined by a single station is a densely populated area of ​​rain gauges; greater than 25km.2 The area defined by a rain gauge station is a sparsely populated area; a small watershed is a closed catchment unit defined by the watershed and the outlet of the river channel.

[0088] Then, basic data of the study area are obtained, including precipitation product data of the study area in the "sky", "air", and "ground", namely satellite precipitation product data, radar precipitation product data, and measured rainfall data of ground rain gauges. Among them, the measured rainfall data of ground rain gauges includes the geographical coordinates of each rain gauge station and its corresponding measured rainfall value.

[0089] Then, based on the acquired basic data, the estimated field and the measured field are constructed:

[0090] 1) Construction of the estimated field:

[0091] Ground Rain Gauge Estimation Field: To simulate sparse areas of ground rain gauges, a small number of ground rain gauges were selected from the study area using stratified random sampling to construct sparse rain gauge networks with varying distributions. For each network distribution scheme, the Kriging interpolation spatial interpolation method was selected to construct the rainfall estimation field, which is the ground rain gauge estimation field.

[0092] Radar precipitation product estimation field: The measurement data under different radar detection radii are recorded as the radar precipitation product estimation field.

[0093] Satellite precipitation product estimation field: The satellite precipitation product estimation field is generated by using a grid resampling method to obtain satellite data of the study area.

[0094] 2) Construction of the measured field: A near-real rainfall field, i.e., the measured field, is constructed using the measured rainfall data from the remaining rain gauges after extraction. The specific construction process is as follows:

[0095] Let the location coordinates of each rain gauge be (x a y a (a = 1, 2, 3…), the actual measured rainfall is Z. a A near-realistic rainfall field is constructed using the inverse distance weighted interpolation method, that is, for any grid point (x... g y g The estimated rainfall was calculated using the inverse distance weighted interpolation method. The formula is as follows:

[0096]

[0097]

[0098] In the formula, Z a ω represents the measured rainfall at the a-th rain gauge station; a For the a-th rain gauge station, pair it with grid point (x) g y g The weight of ); da For the a-th rain gauge station to the grid point (x g y g The distance is m, which is the number of rain gauges involved in the calculation, determined by the search radius or the number of nearest neighbors; p is a power parameter, which is a constant.

[0099] Step 2: Construct an evaluation index system for multi-source precipitation products.

[0100] Given the differences in rainfall monitoring characteristics between "sky," "air," and "ground" precipitation products, the main supplementary point of "sky" and "air" products to "ground" products lies in capturing the spatial distribution characteristics of rainfall, including key information such as the location of the rainfall center and spatial gradient. Therefore, a quantitative evaluation index system is constructed based on the above characteristics. Using the measured field as a benchmark, both categorical and continuous indices are used to evaluate the accuracy of the rainfall center location and spatial gradient of the rainfall field for each estimated field.

[0101] The classification metric uses the hit rate (HR) of the rainfall center location (with the grid point of maximum rainfall as the rainfall center location) to evaluate the accuracy of the estimated rainfall center location. A hit is counted when the straight-line distance between the measured and estimated rainfall centers is less than a certain allowable error distance. The formula is as follows:

[0102]

[0103] Among them, I i The function that indicates the location of the rainfall center has the following value:

[0104]

[0105] In the formula, d i denoted by HR, we represent the straight-line distance between the measured and estimated rainfall centers in time period i; δ represents the allowable error distance; and N represents the total number of time periods selected. A larger HR value indicates a more accurate location of the rainfall center in the estimated field.

[0106] The continuous index uses the Spatial Structure Similarity Index (SSIM) to quantify the accuracy of the estimated field and the measured field of rainfall spatial gradient. Its calculation formula is as follows:

[0107]

[0108] In the formula, x is the normalized measured field; y is the normalized estimated field; μ x and μ y These are the means of two normalized rainfall fields, respectively; σ x and σ y σ represents the variance of the two normalized rainfall fields. xy Let C1 be the covariance of the two normalized precipitation fields; where the constant is C1 = (0.01L). 2 C2 = (0.03L0)2 L represents the dynamic range of the normalized precipitation field. When the precipitation data is normalized to the interval [0,1], L = 1. The larger the SSIM, the higher the similarity between the measured field and the estimated field, and the higher the accuracy.

[0109] Given that single calculation results may be affected by random errors, leading to some bias in the evaluation results, and considering the large amount of sample data, to improve the accuracy of the evaluation results, the calculation results of the evaluation indicators for multiple samples under the same working conditions are averaged, i.e., the average rainfall center location hit rate is calculated separately. And the average spatial structure similarity index This leads to more reliable assessment results. Average rainfall center location accuracy. The calculation formula is:

[0110]

[0111] Average spatial structure similarity index The calculation formula is:

[0112]

[0113] In the formula, M is the sample size under the same working conditions; HR j and SSIM j These are the rainfall center location hit rate and spatial structure similarity index value for the j-th sample, respectively.

[0114] Step 3: Quantify the missing measurement range of spatial characteristics of different precipitation products: Quantify the missing measurement range of spatial characteristics of ground rain gauges and "sky" and "air" precipitation products respectively.

[0115] Step 31: Quantifying the Spatial Feature Missing Range of Ground Rain Gauge Stations: Based on the constructed precipitation product evaluation index system, the impact of station network density and distribution on the ability to capture spatial features of precipitation fields is quantified using the stratified control variable method. Specifically:

[0116] First, three representative rain gauge network distribution schemes are constructed: uniform distribution, topographic gradient distribution, and random distribution. Specifically, Moran's I, a spatial autocorrelation analysis tool, is used to quantify the uniformity of rain gauge station distribution, transforming station locations into spatial density indicators to characterize the clustering and dispersion of stations in different schemes. The Moran's I formula is as follows:

[0117]

[0118] In the formula, n is the total number of stations; x i x j These are the precipitation amounts at the i-th and j-th stations, respectively. This represents the average precipitation; w ij Let be the spatial weight matrix, representing the spatial relationship between the i-th and j-th stations; the closer the Moran index is to 0, the more uniform the station distribution. Specifically, a Moran index in the range [-0.1, 0.1] is considered a uniform distribution; an absolute value in the range (0.1, 0.7) is considered a topographic gradient distribution; and an absolute value in the range (0.7, 1) is considered a random distribution.

[0119] Then, based on the above three spatial distribution schemes, a high-density (<20km) network was formed by using a stratified controlled variable method to randomly sample different numbers of stations in a stratified manner. 2 / station), medium density (20-30km) 2 / station) and low density (>30km) 2 There are three types of station network distribution schemes ( / station). Therefore, nine station network distribution schemes are formed: high-density, medium-density, and low-density station networks constructed based on the uniform distribution scheme; high-density, medium-density, and low-density station networks constructed based on the terrain gradient distribution scheme; and high-density, medium-density, and low-density station networks constructed based on the random distribution scheme.

[0120] For the nine station network distribution schemes mentioned above, a spatial characteristic monitoring accuracy assessment was conducted. Specifically, based on the measured fields of the corresponding regions, the average accuracy of the estimated rainfall center location for each scheme was calculated. And the average spatial structure similarity index

[0121] Finally, the spatial characteristics of missing measurements at ground rain gauges were determined. Specifically, this involved plotting and fitting curves showing the variation of rainfall field monitoring accuracy with network density and uniformity for different station network distribution schemes. Figure 2 As shown, N1 is the curve fitted under a uniform distribution, N2 is the curve fitted under a terrain gradient distribution, and N3 is the curve fitted under a random distribution. The spatial feature monitoring accuracy is set as follows ( or The threshold value for this is denoted as Y. Given that accuracy requirements vary across different regions, the quantile value satisfying the accuracy requirements of 70%–90% of samples is taken as this threshold, typically the 80th quantile. When the monitoring accuracy is lower than this threshold Y, it indicates that the surface rain gauges under this network condition have significantly missed measuring the spatial characteristics of the rainfall field, thus allowing for the summarization of the missing measurement range of the spatial characteristics of the rainfall field by the surface rain gauges. For example... Figure 2 The gray shaded area shows the missing measurements from the ground rain gauges.

[0122] Step 32: Quantifying the Spatial Feature Missing Range of "Airborne" and "Ground-Based" Precipitation Products: To address the missing spatial features of ground-based rain gauges, "airborne" and "ground-based" precipitation products are used for supplementation. However, due to limitations imposed by observation conditions, different detection technologies (satellites, radar) exhibit significant differences in their spatial monitoring capabilities. The accuracy degradation of each product in specific regions can lead to secondary missing data risks. Directly using data from such low-precision areas to supplement the ground station network may actually reduce the reliability of the fusion results. Therefore, it is also necessary to quantify the spatial feature missing range of "airborne" and "ground-based" precipitation products.

[0123] When quantifying the spatial characteristics of missing measurements for "sky" and "air" precipitation products, the same precipitation product evaluation index system as that used for surface rain gauges is adopted, namely, the average rainfall center location hit rate. And the average spatial structure similarity index As a quantitative evaluation indicator, it quantifies the spatial monitoring accuracy of aquatic product decline under different observation conditions. The difference lies in that the estimated field needs to be adjusted according to the object of calculation during the calculation.

[0124] For radar precipitation products, using the measured field as a benchmark, calculate the average accuracy of the rainfall center location in the estimated field of the radar precipitation product. And the average spatial structure similarity index This allows us to obtain the accuracy of radar precipitation products under different radar detection radii. The accuracy points corresponding to each radar detection radius are then fitted to a radar detection radius-accuracy relationship curve, as shown below. Figure 3 As shown, the variation of radar precipitation product accuracy with radar detection radius is characterized, thereby clarifying its spatial supplementation capability and applicable range. Based on the monitoring accuracy threshold Y in step 31, the radar detection radius corresponding to this accuracy threshold is determined in the above curve and denoted as the radar accuracy threshold point.

[0125] A radar monitoring capability zoning method is adopted. Although radar detection has wide coverage, its spatial accuracy decreases with increasing distance, such as... Figure 3 As shown, a boundary needs to be defined within the radar coverage area. Beyond this boundary, data reliability may decrease, requiring supplementation with satellite data or ground-based rain gauges. Beyond a certain boundary, radar data alone is insufficient. To quantify the effective monitoring range, a three-level partitioning system is constructed based on accuracy threshold points:

[0126] Zone I (Independent Effective Zone): Within 10% of the threshold radius, radar data meets the preset accuracy threshold and can independently support the monitoring of spatial characteristics of rainfall fields;

[0127] Zone II (Cooperative Supplement Zone): The reliability of radar data in this zone is reduced, but it can still be enhanced by fusing satellite or ground station data;

[0128] Zone III (Invalid Zone): The radar data accuracy in this zone drops below the threshold and no longer has the ability to characterize spatial features.

[0129] The partitioning method is as follows: The expression relating radar accuracy and radius can be simplified as:

[0130] P R =f R (r)(9)

[0131] When P R When =Y, the corresponding r obtained is r0.

[0132] The expressions for each region can also be obtained:

[0133]

[0134] In the formula, P R Indicates radar accuracy; f R The function represents the radar accuracy; r0 is the threshold radius, which refers to the effective detection radius of the radar when the accuracy threshold requirement of the ground rain gauge is met; μ is 10% of the threshold radius, i.e., μ = 0.1r0.

[0135] The accuracy of satellite precipitation products is primarily influenced by a combination of sensor characteristics, precipitation type and intensity, and surface environment, exhibiting significant regional and seasonal variations. Therefore, the impact of radar detection radius and ground rain gauge density on the accuracy of satellite precipitation products does not need to be considered during implementation. Using the measured field as a benchmark, the average rainfall center location accuracy of the estimated field of satellite precipitation products is calculated. And the average spatial structure similarity index The accuracy of the corresponding satellite precipitation product is calculated and given a fixed value. By comparing this accuracy with the monitoring accuracy of radar and ground rain gauges, the area of ​​missing satellite precipitation data can be determined. Since this value is fixed, if it is lower than the accuracy of radar or ground rain gauges, it is determined that there is a missing satellite precipitation data in that area.

[0136] Step 4: Determine the availability of spatial characteristics of each precipitation product.

[0137] Based on the spatial missing measurement ranges of various precipitation products identified in step 3, a coordinate transformation-based method is adopted to achieve unified integration of key parameter conditions (including station density constraints and effective detection radius) for various precipitation products. This method first constructs a new coordinate system, where the station control area ρ is the abscissa and the radar detection radius r is the ordinate. Based on this new coordinate system, to uniformly incorporate the accuracy influencing factors of rain gauges and radar products obtained in the preceding steps into the analysis, spatial transformations are performed on the monitoring performance maps of the two types of products respectively.

[0138] Among them, the performance map of the rain gauge monitoring station is obtained from step 31, such as... Figure 4 As shown in (c), the horizontal axis of the rain gauge monitoring performance graph represents the station's controlled area ρ, and the vertical axis represents the average accuracy of the rainfall center location. in For ground rain gauges The curve showing the relationship between the site's controlled area ρ and its variation. For radar precipitation products The curve showing the relationship between radar precipitation products and the station control area ρ. Since the performance of radar precipitation products is not affected by the station control area ρ, therefore... Figure 4 (c) is displayed as a horizontal straight line. The radar monitoring performance diagram is obtained from step 32, as shown below. Figure 4 As shown in (b), the horizontal axis of the radar monitoring performance graph represents the average hit rate of the rainfall center location. The vertical axis represents the radar detection radius r; where Representative radar precipitation products The curve showing the relationship between radar detection range r and the radar range. Representative satellite precipitation products The curve showing the relationship between the radar detection radius r and the performance of satellite precipitation products. Since the performance of satellite precipitation products is not affected by the radar detection radius r, therefore... exist Figure 4 (b) is displayed as a vertical straight line. Then, using the origin of the new coordinate system as a reference, the rain gauge monitoring performance map is shifted in the negative direction of the vertical axis so that its origin falls below the vertical axis of the new coordinate system; at the same time, the radar monitoring performance map is shifted in the positive direction of the horizontal axis so that its origin is located to the right of the horizontal axis of the new coordinate system. The shift distance is set reasonably to ensure that the three do not overlap in the coordinate space and to avoid the graphs intersecting.

[0139] Then, based on coordinate transformation, by fixing points, lines, and zones, the availability of spatial characteristics of each precipitation product can be determined. The specific steps are as follows:

[0140] Step 41, Point Determination: This involves identifying the key dividing points for spatial monitoring accuracy of various precipitation products. First, determine the dividing points of the radar's applicable range. This is done by... (The sentence is incomplete and requires further context to be fully translated.) and The intersection point determines the radius that ensures consistent accuracy between radar and satellite precipitation products; the intersection point is denoted as B. r (like Figure 4 (as shown in (b)), the radius at this point is called the critical radius, denoted as r2. Based on the precision corresponding to this critical radius, it can be determined that... Above (e.g.) Figure 4 (c) shows the determination of the station control area corresponding to this level of accuracy, which satisfies the following formula:

[0141]

[0142] in,

[0143] P O,ρ =f O (ρ)(13)

[0144] In the formula, P represents the accuracy value corresponding to a radar radius of r2. O,ρ f represents the accuracy value corresponding to the control area of ​​a ground rain gauge station being ρ. R g O (ρ) represents the accuracy function of the ground rain gauge and the accuracy function of the radar, respectively.

[0145] The point obtained at this point is denoted as B. ρ (like Figure 4 (as shown in (c)). Through B ρ Draw a vertical line through B r Draw a horizontal line. The intersection of the two straight lines in the new coordinate system is a key dividing point of the radar's applicable range, denoted as point B (e.g., ...). Figure 4 (as shown in (a)).

[0146] The next key point starts with the accuracy threshold of the ground rain gauge, which corresponds to the monitoring accuracy threshold Y in step 31. The intersection of the horizontal line drawn from this threshold and the accuracy curve of the ground rain gauge is the critical point A. ρ The radius corresponding to this radar accuracy can be obtained by the following formula:

[0147] P R =f R (r)=Y(14)

[0148] The radius obtained from the above formula is denoted as the effective radius r1, which corresponds to the radar monitoring performance diagram. A of the curve r Point. Similarly, through A. ρ Draw a vertical line through A r Draw a horizontal line; the intersection of the two lines in the new coordinate system is point A. (e.g.) Figure 4 (as shown in (a)) This point represents a key dividing point in the applicable range of the ground rain gauge.

[0149] The third key point C is determined by the density of surface rain gauges. When the rain gauges are distributed sufficiently densely, i.e., the area controlled by each station is less than or equal to 20 km², the rain gauges will be densely distributed. 2Since the ground-based rain gauge station has sufficient observation accuracy, there is no need to integrate other data sources. Therefore, in the new coordinate system, key point C is initially set at an x-coordinate of 20. Furthermore, based on the method used to determine key point B, areas with x-coordinates greater than the value corresponding to point B indicate that the observation accuracy of the ground-based rain gauge station is lower than that of radar and satellites, and usually require the introduction of other data sources for integration. Therefore, if key point C is initially located to the right of key point B, to ensure consistency of the judgment logic and the rationality of the diagram structure, the position of key point C is adjusted to directly below key point B.

[0150] Step 42, Line Determination: This involves constructing the boundary lines for the spatial availability of precipitation products. The two boundary points obtained from the point determination step are... Figure 4 (a) Points A and B are connected by horizontal lines L1 and L2. In the area above the horizontal line L1 passing through point A, only satellite data is available. In the area between the two horizontal lines, both satellite and radar data are available. In the area below the horizontal line L2 passing through point B, only radar data is available.

[0151] Next, to address the issue of missing spatial characteristics due to low accuracy of surface rain gauge data, a supplementary usability boundary line is established. A plumb line L3 is drawn through the boundary point (point A) of the applicable range of the surface rain gauges. The area to the right of this line, due to the low accuracy of the surface rain gauges, requires the use of satellite and radar precipitation data to improve the accuracy of the surface rain gauges. In the area to the left of this line, the accuracy of the surface rain gauges has already met the threshold requirements.

[0152] In respectively Figure 4 (c) A in the curve ρ B ρ Take several points from the segment, and then... Figure 4 (b) Curve A r B r Find the point in the segment that corresponds to its precision. The specific method is the same as the previous method for finding points A and B, respectively... Figure 4 Draw a vertical line from the point in (c). Figure 4 (b) Draw a horizontal line from the points, and fit all their intersections to form a curve, which can also be considered the AB curve. The left side of this curve indicates that ground-based rain gauges have strong spatial characteristic monitoring capabilities, while the right side indicates the need to utilize satellite and radar precipitation products to compensate for the lack of spatial characteristic measurement of rainfall fields by ground-based rain gauges. This invention schematically illustrates three possible cases for curve AB, see [link to diagram]. Figure 4 (a) The three dashed lines between A and B.

[0153] Step 43, Partitioning: After defining the points and lines, divide... Figure 4 (a) Divided into four areas (the four color blocks shown in the figure):

[0154] "Ground" area: The area located to the left of curve ABC and straight line L3. The spatial density and monitoring accuracy of ground rain gauges in this area are high enough to independently and completely reflect the spatial characteristics of the precipitation field without relying on data from satellite or radar precipitation products.

[0155] The "ground + air" area refers to the region to the right of line BC and below line L2. This area has a relatively low spatial density of ground-based rain gauges, but good radar coverage and accuracy. Therefore, it is necessary to combine ground-based rain gauge data with radar rainfall data to improve the monitoring capability of spatial rainfall characteristics.

[0156] The "Ground + Sky + Air" region refers to the area to the right of curve AB, between straight lines L1 and L2. In this region, the density of ground-based rain gauges is low and radar accuracy is limited. Therefore, it is necessary to combine three rainfall data sources—ground-based rain gauges, radar, and satellite—to achieve comprehensive capture and high-precision monitoring of the spatial distribution characteristics of rainfall.

[0157] The "Ground + Sky" region: Located to the right of line L3 and above line L1. In this region, the density of ground rain gauges is low, and the radar monitoring range is limited. It mainly relies on ground rain gauge and satellite data to reflect the spatial characteristics of rainfall.

[0158] Step 44, Dynamic Adjustment: Adjust the data utilization strategy in real time according to different application needs and environmental changes. In some cases, such as extreme weather or areas with complex terrain, it is necessary to recalculate the applicable thresholds. Radar data may be more representative in humid climates, while satellite data may be more important in arid or mountainous terrains. Dynamic adjustment ensures that the data utilization strategy is always optimal to cope with constantly changing environmental and application needs.

[0159] Finally, it should be noted that the above description is only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention.

Claims

1. A method for determining the availability of spatial characteristics of multi-source precipitation products, characterized in that, The method includes the following steps: Step 1: Obtain basic data for the study area and construct the estimation field and the measured field: First, select a small watershed with dense rain gauges as the study area. Define the area with rain gauge density less than or equal to the set value D1 as the dense rain gauge area, and define the area with rain gauge density greater than the set value D1 as the sparse rain gauge area. The small watershed is a closed catchment unit divided by the watershed and the river outlet. Then, obtain basic data for the study area, including "sky," "air," and "ground" precipitation product data, namely satellite precipitation product data, radar precipitation product data, and measured rainfall data from ground rain gauges. Then, based on the obtained basic data, construct the measured field and the estimated fields for each precipitation product, namely the satellite precipitation product estimated field, the radar precipitation product estimated field, and the ground rain gauge estimated field. Step 2: Construct a multi-source precipitation product evaluation index system: Using the measured field as a benchmark, evaluate the accuracy of the estimated precipitation center location and spatial gradient of the precipitation field for each precipitation product using both categorical and continuous indices; among them, the categorical index uses the precipitation center location hit rate. The accuracy of the estimated rainfall center location for each precipitation product field is evaluated. The Spatial Structure Similarity Index (SSIM) is used to quantify the accuracy of the spatial gradient between the estimated and measured rainfall fields for each precipitation product. Given that single calculations are susceptible to random errors, the evaluation index calculations for multiple samples are averaged under the same operating conditions; that is, the average rainfall center location hit rate is calculated separately. And the average spatial structure similarity index ; Step 3: Quantify the spatial characteristic gaps of different precipitation products: Quantify the spatial characteristic gaps of surface rain gauges and "sky" and "air" precipitation products respectively, specifically including the following steps: Step 31: Quantify the spatial feature gap range of ground rain gauges: Based on the constructed multi-source precipitation product evaluation index system, the influence of station network density and distribution on the ability to capture spatial features of precipitation fields is quantified by the hierarchical control variable method, thereby quantifying the spatial feature gap range of ground rain gauges; Specifically, the process involves: first, constructing different distribution schemes for rainfall monitoring stations, including uniform distribution, topographic gradient distribution, and random distribution; then, based on these schemes, using a stratified controlled variable method, randomly sampling different numbers of stations to form three different density distribution schemes: high density, medium density, and low density, resulting in a total of nine distribution schemes; the high density scheme is defined as <20km². 2 / station; medium density is 20~30km 2 / station; low density is >30km 2 / station; calculate the average rainfall center location hit rate of the estimated field under different station network distribution schemes. And the average spatial structure similarity index Finally, the spatial characteristic measurement gaps of ground rain gauges were determined: Rainfall field monitoring accuracy was plotted and fitted as a function of station network density and uniformity under different station network distribution schemes, and the spatial characteristic monitoring accuracy was set. or The threshold is denoted as Y. When the monitoring accuracy is lower than the threshold Y, it indicates that the ground rain gauge has missed measuring the spatial characteristics of the rainfall field under the conditions of the station network. Then, the range of missing measurements of the spatial characteristics of the rainfall field by the ground rain gauge can be summarized. Step 32: Quantify the missing range of spatial characteristics of "sky" and "air" precipitation products: Based on the constructed multi-source precipitation product evaluation index system, quantify the spatial monitoring accuracy of satellite precipitation products and radar precipitation products under different observation conditions, and then quantify the missing range of spatial characteristics of satellite precipitation products and radar precipitation products. Specifically, for radar precipitation products, the average accuracy of the rainfall center location in the estimated field of the radar precipitation product is calculated based on the measured field. And the average spatial structure similarity index This process yields the accuracy of radar precipitation products under different radar detection radii. The accuracy points corresponding to each radar detection radius are then fitted to a radar detection radius-accuracy relationship curve, characterizing the variation of radar precipitation product accuracy with radar detection radius, thereby clarifying its spatial supplementation capability and applicable range. Based on the monitoring accuracy threshold Y in step 31, the radar detection radius corresponding to this accuracy threshold is determined from the above curve and denoted as the radar accuracy threshold point. A radar monitoring capability zoning method is adopted, and a three-level zoning is constructed based on the radar accuracy threshold points. Zone I, or Independent Effective Zone: Within 10% of the threshold radius, radar data meets the preset accuracy threshold and can independently support the monitoring of spatial characteristics of the rainfall field; Zone II is the Coordination and Supplementation Zone: the reliability of radar data in this zone is attenuated, and it is necessary to integrate satellite or ground station data to enhance it; Zone III is the invalid zone: the radar data accuracy in this zone drops below the threshold and no longer has the ability to characterize spatial features; The specific partitioning method is as follows: The simplified expression for the relationship between radar accuracy and radius is: （9） when The corresponding value of r is . ; The expressions for each region can also be obtained: （10） In the formula, Indicates radar accuracy; Represents the radar accuracy function; Threshold radius refers to the effective radar detection radius that meets the accuracy threshold requirements of ground rain gauges. It is 10% of the threshold radius, that is ; For satellite precipitation products, the average accuracy of the precipitation center location is calculated based on the measured field. And the average spatial structure similarity index The accuracy of the corresponding satellite precipitation product is calculated, which is a fixed value. By comparing this accuracy with the monitoring accuracy of radar and ground rain gauges, the missing measurement range of satellite precipitation products is determined. That is, if its value is lower than the accuracy of radar precipitation products or ground rain gauges, it is determined that there is a missing measurement of satellite precipitation products in that area. Step 4: Determine the availability of spatial features of each precipitation product: Based on the spatial missing measurement range of each precipitation product identified in Step 3, a coordinate transformation method is used to determine the availability of spatial features of each precipitation product by fixing points, lines, and zones.

2. The method for determining the availability of spatial characteristics of multi-source precipitation products according to claim 1, characterized in that, The specific process of constructing the measured field and the estimated field of each precipitation product based on the acquired basic data in step 1 is as follows: For the construction of estimation fields for each precipitation product: Ground rain gauge estimation field: To simulate sparse areas of ground rain gauges, a small number of ground rain gauges in the study area are selected according to the principle of stratified random sampling to construct sparse rain gauge networks with different distributions. For each network distribution scheme, the Kriging interpolation spatial interpolation method is selected to construct the rainfall estimation field, which is the ground rain gauge estimation field. Radar precipitation product estimation field: The measurement data under different radar detection radii are recorded as the radar precipitation product estimation field; Satellite precipitation product estimation field: The satellite precipitation product estimation field is generated by using a grid resampling method to obtain satellite data of the study area; For the construction of the actual test field: Using the measured rainfall data from the remaining ground rain gauges after extraction, a near-real rainfall field, i.e., the measured field, is constructed. The specific construction process is as follows: Let the coordinates of each rain gauge be... ( The actual measured rainfall was... A near-realistic rainfall field is constructed using the inverse distance weighted interpolation method, that is, for any grid point... The estimated rainfall was calculated using the inverse distance weighted interpolation method. The formula is as follows: （1） （2） In the formula, For the first The measured rainfall at each rain gauge station; For the first Each rain gauge point is a grid point The weights; For the first From each rain gauge station to the grid point The distance; The number of rain gauges participating in the calculation is determined by the search radius or the number of nearest neighbor points; is an exponential parameter, and is a constant.

3. The method for determining the availability of spatial characteristics of multi-source precipitation products according to claim 1, characterized in that, The set value D1 for the density of rain gauges in step 1 is 25 km. 2 / stand.

4. The method for determining the availability of spatial characteristics of multi-source precipitation products according to claim 1, characterized in that, The accuracy rate of the rainfall center location mentioned in step 2 The calculation formula is: （3） in, The function that indicates the location of the rainfall center has the following value: （4） In the formula, For the first The straight-line distance between the centers of rainfall in the measured and estimated fields during the same period; Allowable error distance; This represents the total number of time periods selected. The formula for calculating the Spatial Structure Similarity Index (SSIM) is as follows: （5） In the formula, For normalized measured fields; For the normalized estimation field; and These are the means of two normalized rainfall fields, respectively; and These represent the variances of the two normalized rainfall fields; Let be the covariance of two normalized precipitation fields; where is a constant. , L is the dynamic range of the normalized rainfall field. When the rainfall data is normalized to the interval [0,1], L=1. The average rainfall center location hit rate The calculation formula is: （6） The average spatial structure similarity index The calculation formula is: （7） In the formula, M is the number of samples collected under the same working conditions; and These are the rainfall center location hit rate and spatial structure similarity index value for the j-th sample, respectively.

5. The method for determining the availability of spatial characteristics of multi-source precipitation products according to claim 1, characterized in that, The specific method for constructing different distribution schemes of rain gauge networks, including uniform distribution, topographic gradient distribution, and random distribution, as described in step 31, is as follows: The Moran index is used to quantify the uniformity of rain gauge station distribution, converting station locations into spatial density indicators to characterize the clustering and dispersion of stations in different schemes. The Moran index formula is as follows: （8） In the formula, Total number of sites; , The first The, the Rainfall at each station; This represents the average precipitation. Let be the spatial weight matrix, representing the first... The and the first The spatial relationship between stations; the closer the Moran index is to 0, the more uniform the station distribution; when the Moran index is in the range of [-0.1, 0.1], it is considered a uniform distribution; when the absolute value of the Moran index is in the range of (0.1, 0.7], it is considered a topographic gradient distribution; when the absolute value of the Moran index is in the range of (0.7, 1], it is considered a random distribution.

6. The method for determining the availability of spatial characteristics of multi-source precipitation products according to claim 1, characterized in that, The specific process described in step 4, which uses a coordinate transformation method to determine the availability of spatial characteristics of each precipitation product by fixing points, lines, and zones, is as follows: First, a new coordinate system is constructed, with the station control area as the coordinate system. The horizontal axis represents the radar detection radius. The vertical axis represents the area controlled by the rain gauge. Based on this new coordinate system, to integrate the accuracy influencing factors of rain gauges and radar products obtained in the preceding steps into the analysis, spatial transformations are performed on the monitoring performance maps of the two types of products. The rain gauge monitoring performance map is obtained in step 31, and its horizontal axis represents the area controlled by the station. The vertical axis represents the average accuracy of the rainfall center location. ; Surface rain gauges Controlled area of ​​the site The curve of change relationship is used It indicates that radar precipitation products Controlled area of ​​the site The curve of change relationship is used This indicates that the performance of radar precipitation products is not affected by the area controlled by the station. The impact is therefore represented as a horizontal straight line; the radar monitoring performance map is obtained from step 32, and the horizontal axis of the radar monitoring performance map is the average hit rate of the rainfall center location. The vertical axis represents the radar detection radius. Radar precipitation products With radar detection range The curve of change relationship is used It indicates that satellite precipitation products With radar detection radius The curve of change relationship is used This indicates that the performance of satellite precipitation products is not affected by radar detection radius. The influence of the radar is therefore displayed as a vertical straight line; then, with the origin of the new coordinate system as a reference, the rain gauge monitoring performance map is shifted in the negative direction of the vertical axis so that its origin falls below the vertical axis of the new coordinate system; at the same time, the radar monitoring performance map is shifted in the positive direction of the horizontal axis so that its origin is located to the right of the horizontal axis of the new coordinate system; the shift distance is reasonably set to ensure that the three do not overlap in the coordinate space and to avoid the intersection of the graphics. Then, based on the coordinate transformation method, the availability of spatial characteristics of each precipitation product can be determined by fixing points, lines, and zones. The specific steps are as follows: Step 41, Point Determination: This involves identifying the key dividing points for spatial monitoring accuracy of various precipitation products; first, determine the dividing points of the radar's applicable range, and then, in the radar monitoring performance map, through... and The intersection point determines the radius that ensures consistent accuracy between radar and satellite precipitation products; the intersection point is denoted as... The radius corresponding to this point is called the critical radius, denoted as . Based on the accuracy corresponding to this critical radius, in The control area of ​​the station at this level of accuracy is determined by the following formula: （11） in, （12） （13） In the formula, Indicates the radar radius as The precision value corresponding to the time. The area controlled by the ground rain gauge station is... The precision value corresponding to the time. These represent the accuracy functions of ground rain gauges and radar, respectively. The point obtained at this time is denoted as... ,Pass Draw a vertical line through Draw a horizontal line. The intersection of the two straight lines in the new coordinate system is a key dividing point of the radar's applicable range, denoted as point A. ; The next key point starts with the accuracy threshold of the ground rain gauge, which corresponds to the monitoring accuracy threshold Y in step 31. The intersection of the horizontal line drawn from this threshold and the accuracy curve of the ground rain gauge is the critical point. The radius corresponding to this radar accuracy is obtained by the following formula: （14） The radius obtained from the above formula is denoted as the effective radius. Corresponding radar monitoring performance diagram curve Point; similarly, pass Draw a vertical line through Draw a horizontal line. The intersection of the two straight lines in the new coordinate system is point A, which represents a key dividing point of the applicable range of the ground rain gauge. The third key point C is determined by the density of ground rain gauges. When the controlled area of ​​a station is less than or equal to the set value D2, it is considered that the ground rain gauge does not need to integrate other data sources. Therefore, in the new coordinate system, this point is initially set at the position with the horizontal coordinate D2. Furthermore, if the key point C is located to the right of the key point B after initial determination, in order to ensure the consistency of the judgment logic and the rationality of the diagram structure, the position of the key point C is adjusted to be directly below the key point B. Step 42, Line Determination: This involves constructing the boundary lines for the spatial availability of precipitation products; based on the two boundary points obtained in the point determination step, ... , Two points, draw horizontal lines through each point. , In the past Horizontal line of a point The area above is only available for satellite data. The area between the two horizontal lines is available for both satellite and radar data. Beyond the point... horizontal line The following areas have radar data available only; Then, based on the spatial characteristics of the missing measurement range of the ground rain gauges, a supplementary usability boundary line is drawn; the boundary line of the applicable range of the ground rain gauge is the point. Draw a plumb line If the area to the right of the line requires the use of satellite and radar precipitation data to improve the accuracy of ground rain gauges, then the area to the left of the line requires that the accuracy of ground rain gauges has already reached the threshold requirement. In respectively In the curve Take several points from the segment, and then... curve Find the point in the segment that corresponds to its accuracy. The specific method is the same as the method for finding points A and B above. Draw a vertical line on the point in the rain gauge monitoring performance map and a horizontal line on the point in the radar monitoring performance map. Fit all their intersections into a curve, which is called the AB curve. The left side of this curve indicates that the ground rain gauge has the ability to monitor spatial characteristics, and the right side indicates that satellite and radar precipitation products are needed to make up for the lack of measurement of the spatial characteristics of the precipitation field by the ground rain gauge. Step 43, Partitioning: After defining the points and lines, the newly constructed coordinate system is divided into four regions: "Earth" area: refers to the area located on curve ABC and the vertical line. In the area on the left, ground-based rain gauges do not need to rely on data from satellite or radar precipitation products for supplementation. "Ground + Air" area: refers to the area to the right of line BC and located on the horizontal line. In the lower area, the ability to monitor the spatial characteristics of rainfall needs to be improved by combining ground rain gauge data with radar rainfall data; The "Earth + Sky + Space" region: is the right side of curve AB, located between the horizontal line. and In the area between these two points, it is necessary to combine three rainfall data sources: ground rain gauges, radar, and satellite, in order to monitor the spatial distribution characteristics of rainfall. "Earth + Sky" region: refers to the area located on the vertical line Right side and on the horizontal line In the upper region, ground-based rain gauges and satellite data are used to reflect the spatial characteristics of rainfall. Step 44: Dynamic Adjustment: Adjust the data utilization strategy in real time according to different actual application needs and environmental changes.

7. The method for determining the availability of spatial characteristics of multi-source precipitation products according to claim 6, characterized in that, The set value D2 in step 41 is 20 km² / station.

Images