CN114818464A - Grid rainfall calculation method based on survey station - Google Patents

Grid rainfall calculation method based on survey station Download PDF

Info

Publication number
CN114818464A
CN114818464A CN202210254597.7A CN202210254597A CN114818464A CN 114818464 A CN114818464 A CN 114818464A CN 202210254597 A CN202210254597 A CN 202210254597A CN 114818464 A CN114818464 A CN 114818464A
Authority
CN
China
Prior art keywords
grid
rainfall
data
spatial
drainage basin
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210254597.7A
Other languages
Chinese (zh)
Inventor
王坤
万定生
余宇峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hohai University HHU
Original Assignee
Hohai University HHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hohai University HHU filed Critical Hohai University HHU
Priority to CN202210254597.7A priority Critical patent/CN114818464A/en
Publication of CN114818464A publication Critical patent/CN114818464A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Mathematical Analysis (AREA)
  • Data Mining & Analysis (AREA)
  • Pure & Applied Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Optimization (AREA)
  • Computational Mathematics (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Geometry (AREA)
  • Algebra (AREA)
  • Computer Hardware Design (AREA)
  • Databases & Information Systems (AREA)
  • Testing Or Calibration Of Command Recording Devices (AREA)

Abstract

The invention discloses a grid rainfall calculation method based on a measuring station, and aims to provide a method for acquiring more accurate spatial rainfall distribution. Aiming at emergency disposal of natural disasters caused by extreme rainfall, an effective method is to quickly and accurately acquire the spatial rainfall distribution condition with smaller granularity in a drainage basin and make a targeted preventive measure. The grid rainfall calculation method comprises the following steps: carrying out feature extraction on the grid meteorological discrete data; carrying out normalization processing on the structured data of the ground rainfall measurement station; constructing a multiple spatial feature topological graph; establishing a regularization grid; and constructing a grid rainfall mining model on the basis of the data to obtain the distribution condition of the grid rainfall in the medium and small drainage basin. According to the method, based on rainfall measurement station data and grid meteorological data, a multiple spatial feature topological relation is constructed, the multiple spatial feature topological relation is used as an adjacent matrix of a graph convolution neural network (GCN), a grid rainfall mining model MS-GCN based on multiple spatial features is constructed, and the rainfall distribution condition of the grids of medium and small watersheds can be acquired more accurately.

Description

Grid rainfall calculation method based on survey station
Technical Field
The invention belongs to the technical field of space rainfall excavation, and particularly relates to a grid rainfall calculation method based on a measuring station.
Background
Precipitation is an important characteristic forming meteorological change, is a main source of surface runoff, and is a main reason causing flood disasters and landslide due to instability and nonuniformity in space-time distribution. The small and medium rivers in China are numerous, and how to quickly and effectively react to flood disasters caused by extreme rainfall becomes a problem which needs to be solved urgently. Aiming at emergency disposal of natural disasters caused by extreme rainfall, an effective method is to quickly and accurately acquire the spatial rainfall distribution condition with smaller granularity in a drainage basin and make a targeted preventive measure. However, the rainfall distribution of the middle and lower watershed has the characteristics of nonlinearity, high complexity and unstable space-time distribution, and the spatial distribution condition of the rainfall in the watershed cannot be accurately obtained only by relying on a small number of ground rainfall measurement stations which are not distributed uniformly.
At present, research institutions at home and abroad mainly adopt a traditional spatial interpolation method and satellite radar multi-source fusion to generate a spatial rainfall field. The conventional spatial interpolation method includes: such as an inverse distance weight method, a kriging method and the like, the structure is simple, the calculation is convenient, but the interpolation precision is low due to neglect of the complex space-time characteristics of nonlinear rainfall data, and the method is mostly adopted for spatial interpolation in domestic research institutes. The meteorological departments at home and abroad mainly adopt satellite radar multi-source fusion products, such as TMPA in the United states and GsMAP in Japan, and adopt various data sources, including ground survey station data, satellite radar maps and the like, so that the precision is high, but the satellite radar fusion products excessively depend on hardware facilities, have certain hysteresis and cannot effectively react to sudden extreme rainfall.
In recent years, deep learning theory is rapidly developed, a new development direction is provided for the research of the rainfall distribution in the space of medium and small watersheds, and a Convolutional Neural Network (CNN) obtains a large amount of research results in the fields of image processing, time series prediction, natural language processing and the like by virtue of the strong modeling capability of the CNN. The mesh rainfall and the station rainfall are used as discrete non-European space data and are more suitable for being described by adopting a graph structure, and a graph convolution neural network (GCN) has strong topological structure feature extraction capability and can be better used for excavating abundant space features of rainfall distribution when being applied to mesh space rainfall excavation.
Disclosure of Invention
The purpose of the invention is as follows: the invention aims to solve the defects in the prior art, and provides a grid rainfall calculation method based on a measuring station.
The technical scheme is as follows: the invention relates to a grid rainfall calculation method based on a measuring station, which comprises the following steps:
step S1, establishing a 1km multiplied by 1km regularized grid in an experimental drainage basin so as to extract multiple spatial features of the grid in a target drainage basin and carry out rainfall excavation on the multiple spatial features;
step S2, comprehensively influencing grid rainfall distribution in the drainage basin by various factors, including rainfall data of a ground station, meteorological factors such as air pressure, temperature and wind speed in the drainage basin, and geographic attribute factors such as terrain, elevation and underlying surface; geographic attribute factors such as terrain, elevation, underlying surface, and the like. Firstly, selecting and collecting rainfall data, meteorological data and geographic attribute data of a ground station of a target drainage basin, and filling the rainfall data, the meteorological data and the geographic attribute data into a regularized grid;
step S3, on the basis of the data, researching complex relationships contained in the spatial geographic features of the medium and small watersheds, constructing a multi-spatial-feature topological relationship between the grid to be researched and the ground measuring station, and fully mining the multi-spatial features of the medium and small watersheds;
and S4, according to the constructed multiple-feature topological graph, taking the grid to be researched and the ground survey station in the drainage basin as topological nodes, taking meteorological factors such as rainfall, air pressure, temperature and wind speed in a time period, geographic attribute factors such as terrain, elevation and underlying surface as node features, taking the multiple-space feature topological relation as an adjacency matrix, constructing a grid rainfall mining model MS-GCN based on multiple-space features, and outputting a model mining result.
The step S1 is to establish a regularized grid, and the step S1 is further to: when the regularization grids are established, the grid establishment range is slightly larger than the area of the experimental drainage basin, all ground rainfall points of the experimental drainage basin are covered, and the discrete measurement stations are distributed in different grids as much as possible. The regularization mesh contains the following attributes: num _ Column is the number of Grid columns, Num _ Row is the number of Grid lines, Origin _ Longitude and Origin _ Latitude are the Longitude and Latitude of the initial Grid, Grid _ Size is the Size of Grid pixels, and invaid _ Value is the identification Value of the out-of-watershed Grid.
The step S2 is data preprocessing, and the step S2 is further:
step S2.1, the preprocessing of the ground rainfall measurement station data in the step S2 comprises missing value processing, processing of the rainfall at the moment into integral point rainfall, extreme value elimination and normalization;
the normalization formula is as follows:
Figure BSA0000268682150000021
wherein the content of the first and second substances,
Figure BSA0000268682150000022
for normalized values, X i Is an original value, X min Is the minimum value in the original sequence, X max Is the maximum value in the original sequence;
step S2.2, the preprocessing of the meteorological and geographic attribute data in the step S2 comprises non-numerical feature conversion and shp layer data extraction to grids;
and S2.3, taking the first 80% of the data preprocessed in the step S2 as a model training set L, and taking the remaining 20% of the data as a test set T.
The spatial geographic feature relationship of the medium and small watersheds is complex, and the medium and small watersheds need to be researched from multiple angles to construct multiple element relationship matrixes and fully mine multiple spatial features of the medium and small watersheds, wherein the step S3 includes:
and S3.1, constructing a site-grid rainfall correlation coefficient matrix. The Correlation Coefficient (Correlation Coefficient) is a statistical index for describing the linear Correlation degree between different variables, and the Correlation Coefficient is used for calculating the rainfall linear relation between the grid to be researched and the ground observation station, wherein the higher the Correlation degree of the two is, the larger the influence of the rainfall of the ground observation station on the rainfall of the grid to be researched is, and the station-grid rainfall Correlation Coefficient matrix is as follows:
Figure BSA0000268682150000031
wherein, c m，k Is the correlation coefficient between node m and node k;
and S3.2, constructing a site-grid spatial distance relation matrix. The closer the spatial distance, the more similar the climate conditions, the lower the spatial distribution dissimilarity of rainfall, and therefore, the spatial distance is also a large factor affecting the spatial distribution of rainfall. Describing the spatial position relationship between nodes by using Euclidean distance (eugler distance), wherein the Euclidean distance is the actual distance between two points in a three-dimensional space, and the formula is as follows:
Figure BSA0000268682150000032
Figure BSA0000268682150000033
wherein atan2(x, y) is a function of calculating the arctan value for a given location (x, y), and R is the earth's equatorial radius; delta theta and Delta a are respectively the latitude difference and longitude difference between two nodes, alpha i And alpha j Is the longitude of two nodes, it should be noted that the longitude and latitude in the trigonometric function need to be converted to a unit value of radian. The spatial distance matrix constructed accordingly is as follows:
Figure BSA0000268682150000034
wherein, de m，k Is the spatial distance between node m and node k;
and S3.3, constructing a site-grid geographic attribute distance matrix. In addition to considering the influence of spatial distance on rainfall distribution, geographic attributes are also important factors influencing spatial rainfall distribution. The more similar the geographic attributes are, the more similar the meteorological conditions are, and the similar rainfall process is possessed among the nodes. The geographic attributes include elevation, whether the slope is a windward slope, and underlying surface conditions such as vegetation coverage and soil type. Calculating the geographic attribute distance between two nodes, firstly calculating the absolute difference value of a single geographic attribute between the nodes, and then carrying out weighted summation on the absolute difference values of all the geographic attributes, wherein the formula is as follows:
Figure BSA0000268682150000035
wherein dl is a，b Is the distance of the geographic attributes between the node a and the node b, n is the number of the node attributes,
Figure BSA0000268682150000036
is the mth attribute value in the node a, and η m is the weighting coefficient of the mth attribute value in the node. From this, a geographical attribute distance matrix between nodes is constructed as follows.
Figure BSA0000268682150000041
Wherein dl is m.k Representing the geographical attribute distance between the node m and the node k, wherein when m is equal to k, the distance is 0;
according to the constructed multiple feature topological relation, constructing a grid rainfall mining model MS-GCN of multiple spatial features, wherein the step S4 comprises the following steps:
s4.1, firstly, constructing a single relation topological graph based on different relation matrixes, then converting the topological relation into an adjacency matrix, explicitly describing the relation among nodes, and finally performing weighted summation on the adjacency matrixes of all the relation topological graphs to obtain a multiple spatial characteristic topological graph and an adjacency matrix thereof;
step S4.2, taking the grid to be researched and the ground survey station in the drainage basin as topological nodes, taking meteorological factors such as rainfall, air pressure, temperature and wind speed in a time period, geographic attribute factors such as terrain, elevation and underlying surface as node characteristics, taking a multiple spatial characteristic topological relation as an adjacent matrix, and excavating by using a graph convolution neural network (GCN), wherein the GCN has strong topological structure characteristic extraction capability, and can better excavate rich spatial characteristics of rainfall distribution when being applied to spatial rainfall excavation, and the formula of a GCN hidden layer is as follows:
X (l+1) ＝f(X (l) ，A)
wherein, l is the current layer number, X (l+1) The model is characterized in that the model is l +1 layer node input characteristics, A is an adjacent matrix, and in the model, the model input characteristics comprise characteristics of nodes, including meteorological conditions and time-interval rainfall; the adjacency matrix is used as the relation between nodes, namely edges, and is generated by the constructed multiple spatial characteristic topological graph;
and S4.3, outputting a model mining result.
Has the advantages that: compared with the prior art, the invention has the advantages that:
the invention comprehensively considers a plurality of data sources including rainfall data, meteorological data such as air pressure and wind speed, and geographic attribute data such as DEM (digital elevation model), underlying surface and the like based on the established regularized grid, establishes a multiple spatial feature topological graph, fully excavates the complex relation of multiple spatial features in the drainage basin, uses GCN (grid traffic control network) to excavate the rainfall of the grid, can more accurately acquire the spatial distribution condition of the rainfall in the drainage basin,
compared with the traditional interpolation algorithm, the MS-GCN solves the problems that the consideration factor is single and the research on the space rainfall change is too shallow, so that the precision of the excavated grid rainfall is greatly improved; compared with a multi-source fusion rainfall product, the method avoids dependence on hardware facilities such as a satellite radar map and the like, has better effectiveness, and provides a decision basis for prevention and emergency disposal of sudden extreme rainfall.
Drawings
FIG. 1 is an overall process diagram of an embodiment of the present invention;
fig. 2 is a view of a grid map at the tunny river basin in a specific example of the invention;
FIG. 3 is a diagram of a model MS-GCN network architecture proposed in the present invention;
FIG. 4 is a comparison graph of experimental results of the experimental grid in the embodiment in the flood season and other conventional interpolation methods;
fig. 5 is a comparison graph of the experimental results of the experimental grid in the embodiment in the flood season and other conventional interpolation methods.
Detailed Description
The technical solution of the present invention is described in detail below, but the scope of the present invention is not limited to the embodiments.
As shown in fig. 1, a method for calculating a grid rainfall based on a station survey of the present embodiment includes the following steps:
step S1, a 1km × 1km regularized grid is established in the tunxi territory so as to extract multiple spatial features of the grid in the target territory and perform rainfall mining on the multiple spatial features.
When the regularized grids are established, the grid establishment range is slightly larger than the area of the experimental drainage basin, all ground rainfall points of the experimental drainage basin are covered, the discrete measuring stations are distributed in different grids as much as possible, the attribute of the regularized grids of the tunxi drainage basin is shown in table 1, and the distribution of the grids in the tunxi drainage basin is shown in fig. 2.
TABLE 1
Grid_Attribute Value
Num_Column 106
Num_Row 81
Origin_Longitude 117.627
Origin_Latitude 29.447
Grid_Size 8.333334E-03
Invalid_Value -9999
Step S2, selecting and collecting rainfall data, meteorological data and geographic attribute data of a ground station of the Tunxi river basin, and preprocessing the rainfall data, the meteorological data and the geographic attribute data;
step S2.1, the preprocessing of the ground rainfall measurement station data of the tunxi watershed in step S2 includes missing value processing, processing of the time rainfall into an integral point rainfall, extreme value rejection, and normalization;
the normalization formula is as follows:
Figure BSA0000268682150000051
wherein the content of the first and second substances,
Figure BSA0000268682150000052
for normalized values, X i Is an original value, X min Is the minimum value in the original sequence, X max Is the maximum value in the original sequence; the specifications of the rainfall data of the Tunxi ground station are as follows: 42 ground measuring stations, wherein the time interval is the hourly rainfall of 1 hour and 0 hour of 2016 to 30 hours and 23 hours of 9 months and 30 months of 2020;
step S2.2, in step S2, the preprocessing of the meteorological and geographic attribute data of the tunny river basin includes non-numerical feature conversion and shp map layer data extraction to a grid; and extracting the geographic attribute data in the layer by using desktop GIS drawing software ArcMap. Firstly, converting the Raster file into an ASCII file through a Raster to ASCII tool, extracting corresponding geographic attribute Data in the ASCII file from grid to grid through a method in a grid Data extraction and storage operation tool class according to a regulated grid established in the preamble, wherein the geographic attribute Data of vegetation coverage, soil type and the like are subjected to one-hot extraction of characteristics and finally stored in a NetCDF (network Common Data form) file;
step S2.3, the first 80% of the tunxi basin data preprocessed in step S2 is used as a model training set L, and the remaining 20% is used as a test set T.
Step S3, based on the data, using the grid G where the semi-source station in the drainage basin is located BY As an experimental object, researching a complex relation between the experimental object and spatial geographic features of a rainfall station, constructing a multiple spatial feature topological relation, and fully excavating multiple spatial features of the Tunxi river basin;
step S3.1, constructing Tunxi basin site-grid G BY A rainfall correlation coefficient matrix. Computing grid G using correlation coefficients BY The higher the correlation degree between the station rainfall and the ground observation station is, the greater the influence of the ground observation station rainfall on the grid rainfall to be researched is, and the station-grid rainfall correlation coefficient matrix is as follows:
Figure BSA0000268682150000061
wherein, c m，k Is the correlation coefficient between node m and node k;
step S3.2, constructing Tunxi basin site-grid G BY A spatial distance relationship matrix. The closer the spatial distance, the more similar the climate conditions, the lower the spatial distribution dissimilarity of rainfall, and therefore, the spatial distance is also a large factor affecting the spatial distribution of rainfall. Describing the spatial position relationship between nodes by using Euclidean distance (eucladometric), and constructing a spatial distance matrix according to the eucladometric distance as follows:
Figure BSA0000268682150000062
wherein, de m，k Is the spatial distance between node m and node k;
step S3.3, constructing Tunxi basin site-grid G BY A geographic attribute distance matrix. In addition to considering the influence of spatial distance on rainfall distribution, geographic attributes are also important factors influencing spatial rainfall distribution. The more similar the geographic attributes are, the more similar the meteorological conditions are, and the similar rainfall process is possessed among the nodes. From this, a geographical attribute distance matrix between nodes is constructed as follows.
Figure BSA0000268682150000063
Wherein dl is m.k Represents the geographical attribute distance between node m and node k, and when m ═ k, the distance is 0.
Step S4, according to the constructed Tunxi drainage basin multiple feature topological graph, taking the grid to be researched and the ground survey station in the drainage basin as topological nodes, taking the meteorological factors such as rainfall, air pressure, temperature and wind speed in a time period, the geographic attribute factors such as terrain, elevation and underlying surface as node features, taking the multiple spatial feature topological relation as an adjacency matrix, and constructing a grid rainfall mining model MS-GCN with multiple spatial features.
S4.1, firstly, constructing a single relation topological graph based on different relation matrixes, then converting the topological relation into an adjacent matrix, explicitly describing the relation between nodes, and finally carrying out weighted summation on the adjacent matrixes of all relation topological graphs to obtain a multiple spatial characteristic topological graph and an adjacent matrix thereof;
step S4.2, grid G BY The method is characterized in that a ground survey station in a drainage basin is used as a topological node, meteorological factors such as rainfall, air pressure, temperature and wind speed in a time period, geographic attribute factors such as terrain, elevation and underlying surface are used as node characteristics, a multiple spatial characteristic topological relation is used as an adjacency matrix, and a graph convolution neural network (GCN) is used for mining, the GCN has strong topological structure characteristic extraction capability, and can be used for mining the spatial rainfall to better mine rich spatial characteristics of rainfall distribution, and the MS-GCN network structure is shown in figure 3, and the hidden layer formula is as follows:
X (l+1) ＝f(X (l) ，A)
wherein, l is the current layer number, X (l+1) The model is characterized in that the model is l +1 layer node input characteristics, A is an adjacent matrix, and in the model, the model input characteristics comprise characteristics of nodes, including meteorological conditions and time-interval rainfall; the adjacency matrix is used as the relation between nodes, namely edges, and is generated by the constructed multiple spatial characteristic topological graph.
Step S4.2, outputting grid G BY Rainfall excavating knotAnd (5) fruit. In this embodiment, the root mean square error RMSE and the average absolute error MAE are used to evaluate the model mining effect. FIG. 4 and FIG. 5 are grids G BY And (3) comparing the model mining results in the flood season and the non-flood season with other traditional interpolation methods, wherein the table 2 shows the comparison of the MS-GCN and the predicted values of the other traditional interpolation methods.
TABLE 2
Figure BSA0000268682150000071
Figure BSA0000268682150000081
Table 3 shows evaluation indexes of the respective models.
TABLE 3
Figure BSA0000268682150000082
Through the graph, each index of the MS-GCN rainfall mining effect in the embodiment is superior to that of a traditional interpolation algorithm, and the accuracy rate of the method is higher.

Claims (5)

1. The grid rainfall calculation method based on the measuring station is characterized by comprising the following steps: the method comprises the following steps:
step S1, establishing a 1km multiplied by 1km regularized grid in an experimental drainage basin so as to extract multiple spatial features of the grid in a target drainage basin and carry out rainfall excavation on the multiple spatial features;
step S2, comprehensively influencing grid rainfall distribution in the drainage basin by various factors, including rainfall data of a ground station, meteorological factors such as air pressure, temperature and wind speed in the drainage basin, and geographic attribute factors such as terrain, elevation and underlying surface; geographic attribute factors such as terrain, elevation, underlying surface, and the like. Firstly, selecting and collecting rainfall data, meteorological data and geographic attribute data of a ground station of a target drainage basin, and filling the rainfall data, the meteorological data and the geographic attribute data into a regularized grid;
step S3, on the basis of the data, researching complex relationships contained in the spatial geographic features of the medium and small watersheds, constructing a multi-spatial-feature topological relationship between the grid to be researched and the ground measuring station, and fully mining the multi-spatial features of the medium and small watersheds;
and S4, according to the constructed multiple-feature topological graph, taking the grid to be researched and the ground survey station in the drainage basin as topological nodes, taking meteorological factors such as rainfall, air pressure, temperature and wind speed in a time period, geographic attribute factors such as terrain, elevation and underlying surface as node features, taking the multiple-space feature topological relation as an adjacency matrix, constructing a grid rainfall mining model MS-GCN based on multiple-space features, and outputting a model mining result.
2. The station-based grid rainfall calculation method of claim 1, wherein: when the regularization grids are established, the grid establishment range is slightly larger than the area of the experimental drainage basin, all ground rainfall points of the experimental drainage basin are covered, and the discrete measurement sites are distributed in different grids as much as possible. The regularization mesh contains the following attributes: num _ Column is the number of Grid columns, Num _ Row is the number of Grid lines, Origin _ Longitude and Origin _ Latitude are the Longitude and Latitude of the initial Grid, Grid _ Size is the Size of Grid pixels, and invaid _ Value is the identification Value of the out-of-watershed Grid.
3. The method of claim 1, wherein the grid rainfall calculation based on the stations comprises:
step S2.1, the preprocessing of the ground rainfall measurement station data in the step S2 comprises missing value processing, processing of the rainfall at the moment into integral point rainfall, extreme value elimination and normalization;
step S2.2, the preprocessing of the meteorological and geographic attribute data in the step S2 comprises non-numerical feature conversion and shp layer data extraction to grids;
and S2.3, taking the first 80% of the data preprocessed in the step S2 as a model training set L, and taking the rest 20% of the data as a test set T.
4. The station-based grid rainfall calculation method of claim 1, wherein:
s3.1, constructing a site-grid rainfall correlation coefficient matrix;
s3.2, constructing a site-grid spatial distance relation matrix;
and S3.3, constructing a site-grid geographic attribute distance matrix.
5. The method of claim 1 wherein the method comprises:
s4.1, firstly, constructing a single relation topological graph based on different relation matrixes, then converting the topological relation into an adjacent matrix, explicitly describing the relation between nodes, and finally carrying out weighted summation on the adjacent matrixes of all relation topological graphs to obtain a multiple spatial characteristic topological graph and an adjacent matrix thereof;
s4.2, taking the grid to be researched and a ground survey station in the drainage basin as topological nodes, taking meteorological factors such as rainfall, air pressure, temperature and wind speed in a time period and geographical attribute factors such as terrain, elevation and underlying surface as node characteristics, taking a topological relation of multiple spatial characteristics as an adjacent matrix, and excavating by using a graph convolution neural network (GCN);
and S4.3, outputting a model mining result.
CN202210254597.7A 2022-03-15 2022-03-15 Grid rainfall calculation method based on survey station Pending CN114818464A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210254597.7A CN114818464A (en) 2022-03-15 2022-03-15 Grid rainfall calculation method based on survey station

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210254597.7A CN114818464A (en) 2022-03-15 2022-03-15 Grid rainfall calculation method based on survey station

Publications (1)

Publication Number Publication Date
CN114818464A true CN114818464A (en) 2022-07-29

Family

ID=82529492

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210254597.7A Pending CN114818464A (en) 2022-03-15 2022-03-15 Grid rainfall calculation method based on survey station

Country Status (1)

Country Link
CN (1) CN114818464A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117574960A (en) * 2023-12-14 2024-02-20 江苏省气象台 Multi-factor plum rainfall prediction method based on self-adaptive graph structure and reinforcement integration

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117574960A (en) * 2023-12-14 2024-02-20 江苏省气象台 Multi-factor plum rainfall prediction method based on self-adaptive graph structure and reinforcement integration
CN117574960B (en) * 2023-12-14 2024-07-09 江苏省气象台 Multi-factor plum rainfall prediction method based on self-adaptive graph structure and reinforcement integration

Similar Documents

Publication Publication Date Title
Chao et al. Geographically weighted regression based methods for merging satellite and gauge precipitation
Sadler et al. Modeling urban coastal flood severity from crowd-sourced flood reports using Poisson regression and Random Forest
CN107316095B (en) Regional weather drought level prediction method coupled with multi-source data
CN111339711B (en) Small watershed design flood calculation method
CN112016052B (en) Near-surface daily maximum air temperature estimation method, system and terminal based on multi-source data
CN108268735A (en) Soil moisture NO emissions reduction method based on multi-source remote sensing satellite fused data
Mustafa et al. Evaluation of land development impact on a tropical watershed hydrology using remote sensing and GIS
CN113222283B (en) Mountain torrent forecasting and early warning method and system based on digital twinning
Pietro et al. Developing terrestrial-LIDAR-based digital elevation models for monitoring beach nourishment performance
CN111340668B (en) Typhoon disaster assessment system
Huang et al. Generating high-accuracy and cloud-free surface soil moisture at 1 km resolution by point-surface data fusion over the Southwestern US
Yılmaz et al. Determination of the appropriate zone on dam surface for floating photovoltaic system installation using RS and GISc technologies
CN115755221A (en) Multi-source precipitation data fusion method based on mathematical uncertainty analysis
CN114818464A (en) Grid rainfall calculation method based on survey station
CN114461983A (en) Satellite precipitation product space downscaling method based on water balance principle
CN114528672A (en) Urban hydrological station network layout method and system based on 3S technology
CN117494419A (en) Multi-model coupling drainage basin soil erosion remote sensing monitoring method
Shabani et al. An integrated approach for simulation and prediction of land use and land cover changes and urban growth (Case study: Sanandaj city in Iran)
Matkan et al. Flood early warning with integration of hydrologic and hydraulic models, RS and GIS (Case Study: Madarsoo Basin, Iran)
Rahman A study on determining land use/land cover changes in dhaka over the last 20 years and observing the impact of population growth on land use/land cover using remote sensing
Ning et al. Precipitation scale effect of the TRMM satellite in Tianshan, China
Shahabi et al. Application of satellite images and fuzzy set theory in landslide hazard mapping in central Zab basin
Jabar et al. Incorporating a GIS-Based Approach and SWAT Model to Estimate Sediment in the Western Desert of Iraq
Shan et al. Precipitation scale effect of the TRMM satellite in Tianshan, China
CN117556695B (en) Crop root soil water content simulation method based on deep learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
DD01 Delivery of document by public notice
DD01 Delivery of document by public notice

Addressee: Wang Kun

Document name: Notice of Publication of Patent Application for Invention

Addressee: Wang Kun

Document name: Notification on Qualification of Preliminary Examination of Invention Patent Application