A simulation method for setting up virtual rain gauges based on satellite precipitation data

Abstract

The present invention discloses a method for simulating the installation of virtual rain gauges based on satellite precipitation data. The method comprises the following steps: obtaining potential installation points based on existing initial observation sites, calculating the probability of continuous precipitation variables taking values ​​in different intervals between the initial observation sites and the potential installation points; calculating the weighted cosine similarity of the distribution function of the value probability; calculating the information entropy between the initial observation points and the potential installation points; and selecting potential installation points that meet the requirements as virtual rain gauges based on the distribution characteristics and weighted cosine similarity of the satellite precipitation information entropy at different site locations. The present invention addresses the problems of the inconsistency between accuracy and coverage, as well as insufficient data processing efficiency, in existing precipitation information acquisition technologies, thereby providing more accurate, comprehensive, and timely precipitation information support.

Description

Technical Field

[0001] The present application relates to the field of precipitation measurement, and in particular to a method for simulating the setting of virtual rain gauges based on satellite precipitation data. Background Art

[0002] Precipitation, as a core component of the natural water cycle, is directly related to many aspects of the economy and society through its intensity variations and distribution characteristics. In agricultural production, timely and appropriate precipitation is an indispensable natural condition for crop growth; in water resource management, accurate precipitation information provides a scientific basis for reservoir scheduling and optimal water resource allocation; in disaster prevention and mitigation, timely monitoring of precipitation dynamics is crucial for predicting and responding to natural disasters such as floods and mudslides. Therefore, establishing an efficient and accurate precipitation monitoring network is of immeasurable value for ensuring agricultural safety, promoting the rational use of water resources, and enhancing disaster early warning capabilities. The rain gauge network is the cornerstone of this monitoring system. By setting up rain gauges in widely distributed geographical locations, it enables continuous observation and recording of precipitation, providing basic data support for meteorological services, hydrological research, and environmental management.

[0003] Although the rain gauge network plays an irreplaceable role in precipitation monitoring, its development and application also face many challenges. First, due to the high cost of construction and maintenance, especially in remote areas with poor economic conditions, the layout of rain gauges is often difficult to achieve the ideal density, resulting in scarce or even blank precipitation data in some areas. Secondly, factors such as complex terrain and inconvenient transportation also limit the coverage and data quality of rain gauges. Especially in hard-to-reach areas such as mountains and forests, the effective operation and maintenance of rain gauges face many difficulties. In addition, existing rain gauges often rely on manual maintenance, and the efficiency of data collection and processing is low. They are also easily affected by human factors, which affects the accuracy and timeliness of the data. The existence of these problems makes the existing rain gauge network have certain limitations in comprehensively and accurately reflecting the regional precipitation conditions.

[0004] Traditional statistical methods often struggle to accurately capture the spatial distribution of precipitation when dealing with this type of complex terrain, resulting in large errors in the estimated results. Existing precipitation information acquisition technologies often face a trade-off between accuracy and coverage. While statistical methods based on ground station networks can provide highly accurate precipitation data, their coverage is limited and data gaps are significant. While satellite remote sensing technology can provide large-scale, continuous precipitation observations, its data accuracy is relatively low, making it difficult to meet the needs of high-precision precipitation research or engineering planning. Summary of the Invention

[0005] In view of the above-mentioned deficiencies in the prior art, the present invention provides a virtual rain gauge setting simulation method based on satellite precipitation data, which solves the problems existing in the existing precipitation information acquisition technology such as the contradiction between accuracy and coverage and insufficient data processing efficiency.

[0006] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is: a virtual rain gauge setting simulation method based on satellite precipitation data, comprising:

[0007] S1. Based on the existing initial observation sites, obtain potential setting points and calculate the probability of the continuous precipitation variables at the initial observation sites and potential setting points taking values ​​in different intervals;

[0008] S2. Calculate the weighted cosine similarity of the distribution function of the value probability;

[0009] S3, calculating the information entropy between the initial observation point and the potential setting point;

[0010] S4. Based on the distribution characteristics and weighted cosine similarity of satellite precipitation information entropy at different station locations, potential setting points that meet the conditions are screened out as virtual rain gauges.

[0011] Furthermore: S1 includes:

[0012] S11. Based on the existing initial observation sites, obtain the center point positions of each triangulated network structure as potential setting points;

[0013] S12, extracting satellite precipitation data at the initial observation point location and satellite precipitation data at the potential setting point location;

[0014] S13. Compare the probability density function characteristics of the satellite precipitation data at the initial observation point and the potential setting point, and calculate the probability of the continuous precipitation variables at the initial observation site and the potential setting point taking values ​​in different intervals.

[0015] Further: In S13, the coefficient of variation is used CV , kurtosis K and skewness S Describe the probability density function, its expression is:

[0016]

[0017]

[0018]

[0019] in, and are the mean and standard deviation of the random variable, n Indicates the amount of data, represents the data mean,x i Indicates the i Satellite precipitation data.

[0020] Furthermore: In S2, the weighted cosine similarity of the distribution function of the value probability WCS The expression is:

[0021]

[0022]

[0023] in, represents the probability of the continuous precipitation variable at the initial observation station taking values ​​in different intervals, The coefficient of variation in the probability density function of the continuous precipitation variable corresponding to the initial observation station CV , kurtosis K and skewness S , represents the probability of the continuous precipitation variable at the potential setting point taking values ​​in different intervals, The coefficient of variation in the probability density function of the continuous precipitation variable corresponding to the potential set point CV , kurtosis K and skewness S , WDP express A and B The weighted dot product of represents the diagonal weight matrix, ‖ AW ‖and‖ BW ‖ respectively represent vectors AW and BW The model, superscript T Indicates transpose.

[0024] Further: In S3, the information entropy between the initial observation point and the potential setting point The expression is:

[0025]

[0026] in, X and Y denote the precipitation random variables at the initial observation point and the potential setting point, express X and Y The joint probability distribution of represents mutual information, which is used to measure the information transfer between the initial observation point and the potential setting point. m and r are the total number of initial observation points and potential setting points, j and k Respectively refer toj The initial observation point and k potential setting points.

[0027] Furthermore: S4 includes:

[0028] S41, based on the distribution characteristics and weighted cosine similarity of satellite precipitation information entropy at different station locations, screening similar initial observation points of each potential setting point and assigning calculation weights;

[0029] S42, checking the information transmission strength of the potential setting points relative to the initial observation points, and further selecting points that meet the accuracy conditions from the screened potential setting points as the original virtual rain gauges;

[0030] S43, estimating the daily precipitation value of the original virtual rain gauge according to the assigned weight;

[0031] S44. The accuracy of the daily precipitation values ​​of the original virtual rain gauge is evaluated based on the Taylor score, Nash efficiency coefficient, and Klinger-Gupta efficiency coefficient. The preliminary virtual rain gauge that meets the accuracy evaluation requirements is used as the virtual rain gauge.

[0032] Further: S41 includes:

[0033] S411, calculating the satellite precipitation information entropy of each station;

[0034] S412, analyzing the distribution characteristics of satellite precipitation information entropy at different station locations, substituting the information entropy into a weighted cosine similarity formula, and sequentially calculating the similarity between each potential setting point and the initial observation point;

[0035] S413, screening out the initial observation points of each potential setting point that meet the similarity threshold according to the similarity between the initial observation points corresponding to each potential setting point;

[0036] S414, calculating the sum of similarities of each potential setting point with the initial observation points that meet the similarity threshold;

[0037] S415 , calculating weights according to the proportion of the similarity of each potential setting point to the initial observation point that meets the similarity threshold to the total similarity.

[0038] The beneficial effects of the present invention are:

[0039] 1. This invention not only considers direct comparison of precipitation amounts but also constructs potential observation points through a spatial triangulation network structure and introduces probability density function features to comprehensively describe the distribution characteristics of precipitation data from three dimensions: coefficient of variation, kurtosis, and skewness. This method better reflects the complexity and spatial variability of precipitation data than single-point observations or simple statistics alone.

[0040] 2. A weighted cosine similarity calculation is proposed, and weight assignment is introduced to make the comparison of precipitation data characteristics between different stations more precise and flexible. This allows for more accurate capture of subtle differences in precipitation characteristics between stations, providing a scientific basis for selecting similar stations and calculating weights.

[0041] 3. Information entropy is used to evaluate the uncertainty of precipitation data. The amount of information transmitted between stations is quantified by calculating mutual information, thereby evaluating the impact of a station on the surrounding area. Based on information entropy, virtual rain gauges that meet the accuracy requirements are selected, which can optimize the layout of the precipitation observation network and improve the simulation accuracy of precipitation data. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Flowchart of the simulation method for setting up a virtual rain gauge based on satellite precipitation data.

[0043] Figure 2 Construct a triangulated network structure for the initial observation points and a schematic diagram of the spatial distribution of potential setting points.

[0044] Figure 3 Calculation diagram of the probability density distribution and similarity of satellite precipitation corresponding to the initial observation point and the potential setting point.

[0045] Figure 4 Schematic diagram of Taylor score results for screening and validation evaluation of potential set points based on information entropy. DETAILED DESCRIPTION

[0046] The specific embodiments of the present invention are described below to facilitate understanding of the present invention by those skilled in the art. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations utilizing the concepts of the present invention are protected.

[0047] like Figure 1 As shown, in one embodiment of the present invention, a method for simulating the setting of a virtual rain gauge station based on satellite precipitation data is provided, comprising:

[0048] S1. Based on the existing initial observation sites, obtain potential setting points and calculate the probability of the continuous precipitation variables at the initial observation sites and potential setting points taking values ​​in different intervals;

[0049] S2. Calculate the weighted cosine similarity of the distribution function of the value probability;

[0050] S3, calculating the information entropy between the initial observation point and the potential setting point;

[0051] S4. Based on the distribution characteristics and weighted cosine similarity of satellite precipitation information entropy at different station locations, potential setting points that meet the conditions are screened out as virtual rain gauges.

[0052] Furthermore, S1 includes:

[0053] S11. Based on the existing initial observation sites, obtain the center point positions of each triangulated network structure as potential setting points;

[0054] Schematic diagram of the spatial distribution of the triangulated network structure and potential setting points constructed from the initial observation points, as shown in Figure 2 As shown;

[0055] S12, extracting satellite precipitation data at the initial observation point location and satellite precipitation data at the potential setting point location;

[0056] S13. Compare the probability density function characteristics of the satellite precipitation data at the initial observation point and the potential setting point, and calculate the probability of the continuous precipitation variables at the initial observation site and the potential setting point taking values ​​in different intervals.

[0057] Specifically, in S13, the coefficient of variation is used CV , kurtosis K and skewness S Describe the probability density function, its expression is:

[0058]

[0059]

[0060]

[0061] in, and are the mean and standard deviation of the random variable, n Indicates the amount of data, represents the data mean, x i Indicates the i Satellite precipitation data.

[0062] The probability density distribution results calculated using the above three key parameters are as follows Figure 3 shown.

[0063] In this embodiment, an improved calculation scheme for cosine similarity is proposed for the three key parameters of the probability distribution function, that is, the original cosine similarity is extended from two-dimensional space to three-dimensional space. At the same time, considering the weight change distribution of the parameters, a weighted cosine similarity calculation method is proposed.

[0064] Specifically, in S2, the weighted cosine similarity of the distribution function of the value probability WCSThe expression is:

[0065]

[0066]

[0067] in, represents the probability of the continuous precipitation variable at the initial observation station taking values ​​in different intervals, The coefficient of variation in the probability density function of the continuous precipitation variable corresponding to the initial observation station CV , kurtosis K and skewness S , represents the probability of the continuous precipitation variable at the potential setting point taking values ​​in different intervals, The coefficient of variation in the probability density function of the continuous precipitation variable corresponding to the potential set point CV , kurtosis K and skewness S , WDP express A and B The weighted dot product of represents the diagonal weight matrix, ‖ AW ‖and‖ BW ‖ respectively represent vectors AW and BW The model, superscript T Indicates transpose.

[0068] Information entropy is used to measure the uncertainty of information and can be used to describe the complexity and information content of data. By comparing the information entropy between the initial observation point and the potential setting point, the loss or gain of information in the transmission process can be evaluated.

[0069] Specifically, in S3, the information entropy between the initial observation point and the potential setting point is The expression is:

[0070]

[0071] in, X and Y denote the precipitation random variables at the initial observation point and the potential setting point, express X and Y The joint probability distribution of Represents mutual information, which is used to measure the information transmission between the initial observation point and the potential setting point. The value of information transmission ranges from 0 to 1. The larger the value, the stronger the correlation between the variables. If the two sites are independent of each other, then If it is 0, there is no information transmission; if the two sites are completely related, then If it is 1, it means that all information is transmitted between the two. m and r are the total number of initial observation points and potential setting points, j and k Respectively refer to j The initial observation point and k potential setting points.

[0072] Furthermore, S4 includes:

[0073] S41, based on the distribution characteristics and weighted cosine similarity of satellite precipitation information entropy at different station locations, screening similar initial observation points of each potential setting point and assigning calculation weights;

[0074] S42, checking the information transmission strength of the potential setting points relative to the initial observation points, and further selecting points that meet the accuracy conditions from the screened potential setting points as the original virtual rain gauges;

[0075] S43, estimating the daily precipitation value of the original virtual rain gauge according to the assigned weight;

[0076] S44. The accuracy of the daily precipitation values ​​of the original virtual rain gauge is evaluated based on the Taylor score, Nash efficiency coefficient, and Klinger-Gupta efficiency coefficient. The preliminary virtual rain gauge that meets the accuracy evaluation requirements is used as the virtual rain gauge.

[0077] Specifically, S41 includes:

[0078] S411, calculating the satellite precipitation information entropy of each station;

[0079] S412, analyzing the distribution characteristics of satellite precipitation information entropy at different station locations, substituting the information entropy into a weighted cosine similarity formula, and sequentially calculating the similarity between each potential setting point and the initial observation point;

[0080] S413, screening out the initial observation points of each potential setting point that meet the similarity threshold according to the similarity between the initial observation points corresponding to each potential setting point;

[0081] S414, calculating the sum of similarities of each potential setting point with the initial observation points that meet the similarity threshold;

[0082] S415 , calculating weights according to the proportion of the similarity of each potential setting point to the initial observation point that meets the similarity threshold to the total similarity.

[0083] Specifically, in S42, the method for verifying the information transmission strength of the potential setting point relative to the initial observation point includes:

[0084] S421, calculating the information entropy of the potential setting points after screening and the initial observation point at the nearest spatial position;

[0085] S422. Compare the relative errors of the information entropy of the potential setting points after screening and the information entropy of the initial observation point at the nearest position to obtain the information transmission strength value.

[0086] Specifically, in S43, Taylor score TSS The spatial correlation coefficient, root mean square error, and standard deviation are combined to evaluate the accuracy of the precipitation expression ability of the potential setting point. The calculation formula is:

[0087]

[0088] in, R is the correlation coefficient between the simulated and observed values, σ is the ratio of the simulated and observed standard deviations, R 0 represents the ideal optimal correlation, and the value is generally 1.

[0089] Nash efficiency coefficient NSE Used to effectively calculate the relative error between data, and the Klinger-Gupta efficiency coefficient KGE Taking into account variability and bias, the following is published:

[0090]

[0091]

[0092] in, O t is the observed value, P t is the model prediction value, is the mean of the observations, B is the ratio of the simulated and observed means;

[0093] In order to verify the information transmission strength of the potential setting point relative to the initial observation point, gridded precipitation data can be used to verify the precipitation performance of the station. Gridded precipitation data is spatial gridded data formed by interpolation of dense rain gauge sites. The storage format is consistent with satellite precipitation data, and it records the actual observed precipitation information. It can be used to verify whether the accuracy of the virtual rain gauge set up by the present invention meets the requirements.

[0094] According to the number and location of rain gauge stations used in gridded precipitation data, this application can be divided into verification data 1, verification data 2, and verification data 3. The Taylor score results of potential setting point screening and verification evaluation based on information entropy are as follows: Figure 4 shown.

[0095] The technology in this application has been applied and verified in plain areas, enabling virtual rainfall station encryption and precipitation simulation based on satellite precipitation data. This technology can generate more effective precipitation information, promoting widespread coverage and precise monitoring of regional precipitation information.

[0096] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A virtual rain gauge setting simulation method based on satellite precipitation data, characterized in that: include: S1. Based on the existing initial observation sites, obtain potential setting points and calculate the probability of the continuous precipitation variables at the initial observation sites and potential setting points taking values ​​in different intervals; S2. Calculate the weighted cosine similarity of the distribution function of the value probability; S3, calculating the information entropy between the initial observation site and the potential setting point; S4. Based on the distribution characteristics and weighted cosine similarity of satellite precipitation information entropy at different station locations, potential setting points that meet the conditions are selected as virtual rain gauges; In S2, the weighted cosine similarity of the distribution function of the value probability WCS The expression is: in, represents the probability of the continuous precipitation variable at the initial observation station taking values ​​in different intervals, The coefficient of variation in the probability density function of the continuous precipitation variable corresponding to the initial observation station CV , kurtosis K and skewness S , represents the probability of the continuous precipitation variable at the potential setting point taking values ​​in different intervals, The coefficient of variation in the probability density function of the continuous precipitation variable corresponding to the potential set point CV , kurtosis K and skewness S , WDP express A and B The weighted dot product of represents the diagonal weight matrix, ‖ AW ‖and‖ BW ‖ respectively represent vectors AW and BW The model, superscript T Indicates transpose.

2. The virtual rain gauge setting simulation method based on satellite precipitation data according to claim 1 is characterized in that: S1 includes: S11. Based on the existing initial observation sites, obtain the center point positions of each triangulated network structure as potential setting points; S12, extracting satellite precipitation data at the initial observation site location and satellite precipitation data at the potential setting point location; S13. Compare the probability density function characteristics of the satellite precipitation data at the initial observation site and the potential setting point, and calculate the probability of the continuous precipitation variables at the initial observation site and the potential setting point taking values ​​in different intervals.

3. The virtual rain gauge setting simulation method based on satellite precipitation data according to claim 2 is characterized in that: In S13, the coefficient of variation is used CV , kurtosis K and skewness S Describe the probability density function, its expression is: in, and are the mean and standard deviation of the random variable, n Indicates the amount of data, represents the data mean, x i Indicates the i Satellite precipitation data.

4. The virtual rain gauge setting simulation method based on satellite precipitation data according to claim 1 is characterized in that: In S3, the information entropy between the initial observation site and the potential setting point The expression is: in, X and Y denote the precipitation random variables at the initial observation site and the potential setting point, respectively, express X and Y The joint probability distribution of represents the mutual information, which is used to measure the information transfer between the initial observation site and the potential setting point. m and r are the total number of initial observation sites and potential setting points, j and k Respectively refer to j The initial observation site and k potential setting points.

5. The virtual rain gauge setting simulation method based on satellite precipitation data according to claim 1 is characterized in that S4 include: S41. Based on the distribution characteristics and weighted cosine similarity of satellite precipitation information entropy at different station locations, similar initial observation stations of each potential setting point are screened and calculation weights are assigned. S42, checking the information transmission strength of the potential setting points relative to the initial observation site, and further selecting points that meet the accuracy conditions from the screened potential setting points as the original virtual rain gauges; S43, estimating the daily precipitation value of the original virtual rain gauge according to the assigned weight; S44. The accuracy of the daily precipitation values ​​of the original virtual rain gauge is evaluated based on the Taylor score, Nash efficiency coefficient, and Klinger-Gupta efficiency coefficient. The original virtual rain gauge that meets the accuracy evaluation requirements is used as the virtual rain gauge.

6. The virtual rain gauge setting simulation method based on satellite precipitation data according to claim 5 is characterized in that: S41 includes: S411, calculating the satellite precipitation information entropy of each station; S412, analyzing the distribution characteristics of satellite precipitation information entropy at different station locations, substituting the information entropy into a weighted cosine similarity formula, and sequentially calculating the similarity between each potential setting point and the initial observation station; S413, screening out the initial observation sites that meet the similarity threshold for each potential setting point based on the similarity between each potential setting point and the initial observation site; S414, calculating the sum of similarities of the initial observation sites that meet the similarity threshold for each potential setting point; S415 , allocating calculation weights according to the ratio of the similarity of each potential setting point to the initial observation site that meets the similarity threshold to the total similarity.

Images