A method for bias correction of satellite precipitation data in data-deficient areas

Through the combination of KNN method and mixed geographic weighted weight values, the problem of precipitation data deviation in areas with insufficient data is solved, the accuracy of satellite precipitation data is improved, and high-precision input data is provided for the hydrological model, which enhances the disaster prevention and mitigation ability of flood disasters is enhanced. .

CN118467939BActive Publication Date: 2024-11-12HOHAI UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410391992.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-02
Publication Date
2024-11-12
Estimated Expiration
2044-04-02

AI Technical Summary

Technical Problem

The inaccurate acquisition of precipitation data in areas with lack of data leads to large deviations in the application of hydrological models in these areas, affecting the disaster prevention and mitigation capabilities of flood disasters.

Method used

The KNN method is used to combine the mixed geographic weighted weight value, and a weight coefficient database is constructed through the deviation analysis of satellite precipitation data and measured precipitation data to realize the correction of precipitation data in areas where the measured precipitation data is missing.

Benefits of technology

The accuracy and spatiotemporal resolution of satellite precipitation data in areas with insufficient data are improved, providing more accurate input data for the hydrological model, and enhancing the understanding of flood disaster mechanisms and disaster prevention and mitigation capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118467939B_ABST
    Figure CN118467939B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for correcting the deviation of satellite precipitation data in areas with insufficient data, including: collecting precipitation influencing factor data at a national scale, calculating the deviation between satellite precipitation and measured site data for areas with measured precipitation data, obtaining weight coefficient groups under the corresponding deviations of different influencing factor groups based on a hybrid geographically weighted regression model, and then constructing a national weight coefficient group database; for areas with insufficient data, presetting preliminary screening conditions, and screening weight coefficient groups from the database; sequentially using the KNN method, the Kriging interpolation method, and the weighted summation method to obtain the precipitation deviation under the current precipitation conditions in the areas with insufficient data; and then obtaining the corrected precipitation data. The present invention uses the hybrid geographically weighted weight values ​​of areas with data, obtains the hybrid geographically weighted weight values ​​of areas with insufficient data through the KNN method, realizes the deviation correction of satellite precipitation data in areas with insufficient data, and provides data support for the input of hydrological models in areas with insufficient data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of hydrology and meteorology, and in particular to a method for correcting satellite precipitation data deviation in areas lacking data. Background Art

[0002] Hydrological models are an important means to solve flood disasters. Precipitation is the most critical input source of hydrological models. The accuracy and timeliness of precipitation directly affect the accuracy and reliability of simulation results.

[0003] At present, the main ways to obtain precipitation data are ground observation, satellite and radar quantitative precipitation estimation, and model quantitative precipitation forecast. For a long time, conventional precipitation observations have mainly relied on observation stations arranged on the surface, using limited observation results to represent the real precipitation within the surrounding area of ​​dozens or even hundreds of square kilometers. The size and type of actual precipitation have significant spatiotemporal variability. Ground stations have the problem of using points to represent the surface, especially in areas with sparse stations. Observation of precipitation cannot effectively reflect the spatial variability of spatial precipitation. The spatial limitation of precipitation observation has become a difficulty in hydrological research. Radar quantitative precipitation estimation has the advantages of high spatial resolution and strong real-time performance, but because it is easily affected by covering objects, its coverage is limited. With the development of satellite remote sensing technology at home and abroad, remote sensing precipitation observations based on weather radars and satellites have been continuously improved, making up for the shortcomings of the spatial distribution of ground stations and providing a new means for precipitation monitoring. At present, satellite remote sensing has unique advantages in obtaining global precipitation with spatiotemporal variations, and provides unprecedented satellite precipitation products such as TRMM, GPM, COMRPH, PERSIANN, FY-3B, FY-3C, etc. Satellite quantitative precipitation estimation has the advantages of wide coverage and relatively continuous observation time. However, due to limitations of remote sensing detection instruments and inversion algorithms, the accuracy of satellite precipitation products is relatively low, and the inversion capability of solid precipitation is limited.

[0004] Data-deficient areas are usually located in remote or underdeveloped areas and lack observation data. Climate change in these areas is more significantly affected by global climate change. Hydrological models usually require a large amount of observation data for calibration and verification. Research on precipitation bias correction in data-deficient areas can provide a scientific basis for disaster reduction and prevention and better understand the occurrence mechanism of these disasters. In order to better describe the spatial variation of precipitation in data-deficient areas, the KNN method is used to perform bias correction on satellite quantitative precipitation with high temporal and spatial resolution, improve the quality of precipitation products, and construct a high-precision, high-temporal and spatial resolution hydrological model driving field to provide strong support for hydrological simulation in data-deficient areas. Summary of the invention

[0005] The purpose of the present invention is to provide a method for correcting the deviation of satellite precipitation data in areas with insufficient data. The method utilizes the mixed geographical weighted weight values ​​of areas with sufficient data and obtains the mixed geographical weighted weight values ​​of areas with insufficient data through the KNN method, thereby realizing the deviation correction of satellite precipitation data in areas with insufficient data and providing data support for the input of hydrological models in areas with insufficient data.

[0006] In order to achieve the above functions, the present invention designs a satellite precipitation data deviation correction method for areas with insufficient data, and performs the following steps S1 to S5 to complete the correction of precipitation data in areas lacking measured precipitation data:

[0007] Step S1: Download national-scale satellite precipitation data, vegetation index data NDVI, wind speed data, and elevation data DEM, collect measured precipitation data at each rainfall station, and obtain national-scale slope, slope aspect, surface roughness, and distance to the coastline based on elevation data DEM;

[0008] Step S2: For areas with measured precipitation data, the deviation between satellite precipitation data and measured precipitation data at rain gauges is calculated, and the relationship between the deviation and the influencing factor is established based on the hybrid geographical weighted regression model to obtain the weight coefficient group corresponding to the precipitation deviation of different influencing factor groups in areas with measured precipitation data nationwide, and then construct a national weight coefficient group database;

[0009] Step S3: For the area without measured precipitation data, based on the precipitation conditions and influencing factors of the area, a similar area between the area without measured precipitation data and the area with measured precipitation data is established, and the similar area is used as a preliminary screening condition to select each weight coefficient group that meets the preliminary screening condition from the national weight coefficient group database constructed in step S2;

[0010] Step S4: using the KNN method to select s preferred weight coefficient groups with the highest matching degree with the current precipitation conditions in the area where measured precipitation data are missing from the weight coefficient groups screened in step S3, using the Kriging interpolation method to obtain the optimal weight coefficient group under the current precipitation conditions in the area where measured precipitation data are missing, and performing weighted summation of the optimal weight coefficient group and the influencing factors to obtain the precipitation deviation under the current precipitation conditions in the area where measured precipitation data are missing;

[0011] Step S5: Add the deviation of the current precipitation conditions in the area lacking measured precipitation data to the corresponding satellite precipitation data to obtain corrected precipitation data, thereby completing the correction of precipitation data in the area lacking measured precipitation data.

[0012] Beneficial effects: Compared with the prior art, the advantages of the present invention include:

[0013] The method designed in the present invention utilizes the mixed geographical weighted weight values ​​of areas with measured precipitation data, and obtains the mixed geographical weighted weight values ​​of areas without measured precipitation data through the KNN method, thereby realizing the deviation correction of satellite precipitation data in areas with insufficient data, providing data support for the input of hydrological models in areas with insufficient data, and solving the problem of lack of measured precipitation data in remote or underdeveloped areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flow chart of a method for correcting satellite precipitation data deviation in a data-deficient area provided according to an embodiment of the present invention;

[0015] Figure 2 is a schematic diagram of DEM elevation data of the Han River Basin at a uniform resolution provided by an embodiment of the present invention;

[0016] Figure 3 Schematic diagram of satellite precipitation data in areas lacking measured precipitation data provided by an embodiment of the present invention;

[0017] Figure 4 It is a schematic diagram of precipitation data after correction for a region lacking measured precipitation data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0019] A method for correcting satellite precipitation data deviation in a data-deficient area is provided in an embodiment of the present invention. Figure 1 , perform the following steps S1 to S5 to complete the precipitation data correction in the area where the measured precipitation data is missing:

[0020] Step S1: Download national-scale satellite precipitation data (25 km × 25 km), vegetation index data NDVI (25 km × 25 km), wind speed data (1 km × 1 km), and elevation data DEM (5 km × 5 km), collect measured precipitation data at each rainfall station, and obtain national-scale slope, slope aspect, surface roughness, and distance to the coastline based on the elevation data DEM;

[0021] In step S1, the satellite precipitation data, vegetation index data NDVI, wind speed data, and elevation data DEM are unified to a resolution of 1km×1km using bilinear interpolation, and the slope, aspect, surface roughness, and distance to the coastline data of the national scale are obtained based on the elevation data DEM at this resolution;

[0022] Step S2: For areas with measured precipitation data, in this embodiment, the areas with measured precipitation data in the Han River Basin are used to calculate the deviation between the satellite precipitation data and the measured precipitation data of the rain gauge. According to the hybrid geographical weighted regression model, the relationship between the deviation and the influencing factors (including vegetation index data NDVI, wind speed, elevation data DEM, slope, aspect, surface roughness, and distance to the coastline) is established to obtain the weight coefficient group under the precipitation deviation corresponding to different influencing factor groups in the areas with measured precipitation data nationwide, and then construct the Han River Basin weight coefficient group database; Figure 2 This is a schematic diagram of DEM elevation data of the Han River Basin at a uniform resolution.

[0023] The specific steps of step S2 are as follows:

[0024] Step S2.1: Determine the location of the rainfall station in the satellite precipitation data grid based on the longitude and latitude of the rainfall station:

[0025]

[0026]

[0027] In the formula, row is the row where the measured precipitation data of the rain gauge station is located, col is the column where the measured precipitation data of the rain gauge station is located, delta is the grid resolution of the satellite precipitation data, lat u is the maximum latitude of satellite precipitation data, lat s is the latitude of the rainfall station, lon l is the minimum longitude of satellite precipitation data, lon s is the longitude of the rainfall station;

[0028] Step S2.2: Calculate the deviation between the satellite precipitation data and the measured precipitation data at the rain gauge station as follows:

[0029] Δpre=pre obs -pre sat

[0030] In the formula, pre obs is the measured precipitation data at the rainfall station, sat is the satellite precipitation data, Δpre is the deviation between the satellite precipitation data and the measured precipitation data at the rain gauge station;

[0031] Step S2.3: Based on the deviation and impact factor calculated in step S2.2, a hybrid geographically weighted regression model is established as follows:

[0032]

[0033] Where (i, j) is the coordinate of the sampling point, Δpre(i, j) is the deviation between the satellite precipitation data and the measured precipitation data at the sampling point; x k is the kth time-invariant influence factor of the sampling point, is the lth time-varying influencing factor at sampling point t, where the time-invariant factors include: elevation data DEM, slope, aspect, surface roughness, and distance to the coastline; the time-varying factors include vegetation index data NDVI and wind speed; β0(i,j) is the regression coefficient of the sampling point; β k (i, j) is the regression coefficient of the kth time-invariant influencing factor at the sampling point; is the regression coefficient of the lth time-varying influencing factor at time t at the sampling point; p and q are the total number of time-invariant influencing factors and time-varying influencing factors respectively; ε(i,j) is the residual term of the sampling point;

[0034] Among them, the regression coefficient β k The calculation formula for (i,j) is as follows:

[0035]

[0036]

[0037] Where I is the unit matrix, w ij Represents the weight coefficient group at the sampling point (i, j);

[0038] Among them, w ij The Gauss function is selected as the spatial weight function, and the function form is as follows:

[0039] w ij =exp(-(d ij / b) 2 )

[0040] In the formula, w ij Represents the weight coefficient group at the sampling point (i, j), d ij is the distance between the sampling point and the measured rainfall site, b is a non-negative attenuation parameter describing the functional relationship between weight and distance;

[0041] Step S2.4: Traverse different influencing factor groups in the areas with historical measured precipitation data in the Hanjiang River Basin on spatial and temporal scales, obtain weight coefficient groups corresponding to different influencing factor groups, and then construct a weight coefficient group database for the Hanjiang River Basin.

[0042] Step S3: For the area without measured precipitation data, based on the precipitation conditions and influencing factors of the area (including vegetation index data NDVI, wind speed, elevation data DEM, slope, aspect, surface roughness, and distance to the coastline), a similar area between the area without measured precipitation data and the area with measured precipitation data is established, and the similar area is used as a preliminary screening condition to select each weight coefficient group that meets the preliminary screening condition from the weight coefficient group database constructed in step S2;

[0043] The specific steps of step S3 are as follows:

[0044] Step S3.1: Taking the area with measured precipitation data as the reference basin and the area without measured precipitation data as the target basin, based on the precipitation conditions and influencing factors of the target basin, establish the similarity area between the reference basin and the target basin based on precipitation conditions and influencing factors, and screen the weight coefficient group;

[0045] The precipitation conditions include: meteorological conditions, cooling conditions, and power conditions:

[0046] Meteorological conditions: The atmosphere must contain sufficient water vapor;

[0047] Cooling conditions: water vapor in the atmosphere condenses into liquid water or solid water;

[0048] Dynamic conditions: Condensed liquid precipitation or solid ice and snow particles are large and small. Only liquid water droplets and solid ice and snow particles that can overcome air resistance and reach the ground from the air under the action of gravity can become precipitation.

[0049] In one embodiment, the similarity zone between the reference basin and the target basin can be determined by analyzing the spatial distribution patterns of meteorological factors to extract spatial statistical indicators of the climate of small and medium-sized rivers, extracting physical characteristic factors (topography, soil, vegetation, basin shape, etc.) based on sub-basins as basic units, extracting basin physical similarity indicators based on principal component analysis, and establishing a mapping relationship between climate spatial statistical indicators and basin physical similarity indicators. The similarity degree of the basins is then evaluated to complete the establishment of the similarity zone between the reference basin and the target basin.

[0050] Step S3.2: For the influencing factors of the sampling points at the current moment in the area where measured precipitation data is missing, the upper and lower fluctuation ranges of the influencing factor benchmark data are preset, and then the weight coefficient groups that meet the fluctuation range are screened out;

[0051] Step S3.3: For areas without measured precipitation data, the cumulative distribution probability of each grid day scale 10×365 time series is calculated based on the historical satellite precipitation data of the past 10 years. The satellite precipitation data of a certain day is selected as the benchmark precipitation, and its corresponding probability in the cumulative distribution probability is found as the benchmark precipitation probability. The heavier or lighter the rainfall, that is, in extreme cases, the fluctuation range is small, and in the middle, the fluctuation range is large; the fluctuation range of precipitation probability is set as follows:

[0052] If the baseline precipitation probability The fluctuation range of precipitation probability is

[0053] If the baseline precipitation probability The fluctuation range of precipitation probability is

[0054] Step S3.4: Under the premise of satisfying the similarity area and the fluctuation range of the influencing factors, the baseline precipitation probability As a judgment benchmark, according to the corresponding relationship between the influencing factors and the weight coefficient groups, the weight coefficient groups that meet the conditions are screened out from the weight coefficient group database constructed in step S2 to achieve preliminary screening of the weight coefficient groups.

[0055] Step S4: using the KNN method to select s preferred weight coefficient groups with the highest matching degree with the current precipitation conditions in the area where measured precipitation data are missing from the weight coefficient groups screened in step S3, using the Kriging interpolation method to obtain the optimal weight coefficient group under the current precipitation conditions in the area where measured precipitation data are missing, and performing weighted summation of the optimal weight coefficient group and the influencing factors to obtain the precipitation deviation under the current precipitation conditions in the area where measured precipitation data are missing;

[0056] The specific steps of step S4 are as follows:

[0057] Step S4.1: Based on each weight coefficient group screened in step S3, construct a feature space according to its corresponding influencing factor vector;

[0058] Step S4.2: Taking the influencing factor vector of the current precipitation conditions in the area without measured precipitation data as the item to be classified, calculate the Manhattan distance between the item to be classified and each vector in the feature space. The Manhattan distance formula is as follows:

[0059] d m =|a1-b1|+|a2-b2|+…+|a n -b n |

[0060] In the formula, (a1, a2, … a n ) is the feature vector of the item to be classified, (b1,b2,…b n ) is the mth vector in the feature space, d m is the Manhattan distance between the two;

[0061] Step S4.3: according to the calculated Manhattan distances between the item to be classified and each vector in the feature space, find s vectors with the smallest Manhattan distances to the item to be classified, thereby determining the corresponding s preferred weight coefficient groups;

[0062] Step S4.4: Use the Kriging interpolation method to perform weighted averaging on the s optimal weight coefficient groups to obtain the optimal weight coefficient group under the current precipitation conditions in the area where measured precipitation data is missing;

[0063] Step S4.5: By performing weighted summation on the optimal weight coefficient group and the influencing factors, the precipitation deviation under the current precipitation conditions in the area where measured precipitation data is missing is obtained.

[0064] Step S5: Add the deviation of the current precipitation conditions in the area lacking measured precipitation data to the corresponding satellite precipitation data to obtain the corrected precipitation data, and complete the precipitation data correction in the area lacking measured precipitation data. Figure 3 , the schematic diagram of the corrected precipitation data in the area without measured precipitation data is shown in Figure 4 .

[0065] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A method for correcting satellite precipitation data bias in areas with insufficient data, characterized in that: Execute the following steps S1 to S5 to complete the precipitation data correction in the area where the measured precipitation data is missing: Step S1: Download national-scale satellite precipitation data, vegetation index data NDVI, wind speed data, and elevation data DEM, collect measured precipitation data at each rainfall station, and obtain national-scale slope, slope aspect, surface roughness, and distance to the coastline based on elevation data DEM; Step S2: For areas with measured precipitation data, the deviation between satellite precipitation data and measured precipitation data at rain gauges is calculated, and the relationship between the deviation and the influencing factor is established based on the hybrid geographical weighted regression model to obtain the weight coefficient group corresponding to the precipitation deviation of different influencing factor groups in areas with measured precipitation data nationwide, and then construct a national weight coefficient group database; The specific steps of step S2 are as follows: Step S2.1: Determine the location of the rainfall station in the satellite precipitation data grid based on the longitude and latitude of the rainfall station: ; In the formula, row is the row where the measured precipitation data of the rainfall station is located. col This is the column where the measured precipitation data of the rain gauge station is located. delta is the grid resolution of satellite precipitation data, lat u is the maximum latitude of satellite precipitation data, lat s is the latitude of the rainfall station, lon l is the minimum longitude of satellite precipitation data, lon s is the longitude of the rainfall station; Step S2.2: Calculate the deviation between the satellite precipitation data and the measured precipitation data at the rain gauge station as follows: ; In the formula, pre obs The measured precipitation data of the rain gauge station. pre sat is the satellite precipitation data, △ pre It is the deviation between satellite precipitation data and measured precipitation data at rain gauges; Step S2.3: Based on the deviation and impact factor calculated in step S2.2, a hybrid geographically weighted regression model is established as follows: ; In the formula, ( i,j ) is the sampling point coordinate, △ pre ( i,j ) is the deviation between the satellite precipitation data and the measured precipitation data at the sampling point; x k The sampling point k time-invariant impact factor, x t l For sampling point t Moment l time-varying factors, of which the time-invariant factors include: elevation data DEM, slope, aspect, surface roughness, and distance to the coastline; the time-varying factors include vegetation index data NDVI and wind speed; β 0 ( i , j ) is the regression coefficient of the sampling point; β k ( i , j ) is the sampling point k The regression coefficient of each time-invariant influencing factor; β t l ( i , j ) is the sampling point t Moment l The regression coefficients of the time-varying influencing factors; p and q are the total number of time-invariant impact factors and time-varying impact factors respectively; ε( i , j ) is the residual term of the sampling point; Among them, the regression coefficient β k ( i , j ) is calculated as follows: ; In the formula, I is the identity matrix, w ij Indicates the sampling point ( i , j ) at the weight coefficient group; in, w ij The spatial weight function is used to select the Gauss function for calculation, and the function form is as follows: ; In the formula, w ij Indicates the sampling point ( i , j ), d ij is the distance between the sampling point and the measured rainfall station, b is a non-negative attenuation parameter that describes the functional relationship between weight and distance; Step S2.4: Traverse different influencing factor groups in regions with historical measured precipitation data across the country on spatial and temporal scales, obtain weight coefficient groups corresponding to different influencing factor groups, and then construct a national weight coefficient group database; Step S3: For the area without measured precipitation data, based on the precipitation conditions and influencing factors of the area, a similar area between the area without measured precipitation data and the area with measured precipitation data is established, and the similar area is used as a preliminary screening condition to select each weight coefficient group that meets the preliminary screening condition from the national weight coefficient group database constructed in step S2; The specific steps of step S3 are as follows: Step S3.1: Taking the area with measured precipitation data as the reference basin and the area without measured precipitation data as the target basin, based on the precipitation conditions and influencing factors of the target basin, establish the similarity area between the reference basin and the target basin based on precipitation conditions and influencing factors, and screen the weight coefficient group; Step S3.2: For the influencing factors of the sampling points at the current moment in the area where measured precipitation data is missing, the upper and lower fluctuation ranges of the influencing factor benchmark data are preset, and then the weight coefficient groups that meet the fluctuation range are screened out; Step S3.3: For areas without measured precipitation data, the cumulative distribution probability of each grid day scale 10×365 time series is calculated based on the historical satellite precipitation data of the past 10 years. The satellite precipitation data of a certain day is selected as the benchmark precipitation, and its corresponding probability in the cumulative distribution probability is found as the benchmark precipitation probability Ф. The fluctuation range of precipitation probability is set as follows: If the baseline precipitation probability Ф>0.5, the precipitation probability fluctuation range is 1-Ф; If the baseline precipitation probability Ф≤0.5, the fluctuation range of precipitation probability is 0.5-Ф; Step S3.4: Under the premise of satisfying the similarity area and the fluctuation range of the influencing factors, the baseline precipitation probability As a judgment benchmark, according to the corresponding relationship between the impact factor and the weight coefficient group, the weight coefficient group that meets the conditions is screened out from the national weight coefficient group database to achieve preliminary screening of the weight coefficient group; Step S4: Using the KNN method, select the weight coefficient group that best matches the current precipitation conditions in the area where measured precipitation data is missing from the weight coefficient groups screened in step S3. s The optimal weight coefficient group is selected, and the optimal weight coefficient group under the current precipitation conditions in the area with missing measured precipitation data is obtained by using the Kriging interpolation method. The precipitation deviation under the current precipitation conditions in the area with missing measured precipitation data is obtained by weighted summation of the optimal weight coefficient group and the influencing factors. The specific steps of step S4 are as follows: Step S4.1: Based on each weight coefficient group screened in step S3, construct a feature space according to its corresponding influencing factor vector; Step S4.2: Taking the influencing factor vector of the current precipitation conditions in the area without measured precipitation data as the item to be classified, calculate the Manhattan distance between the item to be classified and each vector in the feature space. The Manhattan distance formula is as follows: ; In the formula, ( a 1 ,a 2 ,…a n ) is the feature vector of the item to be classified, ( b 1 ,b 2 ,…b n ) is the first m vector, d m is the Manhattan distance between the two; Step S4.3: Based on the calculated Manhattan distances between the item to be classified and each vector in the feature space, find the vector with the smallest Manhattan distance to the item to be classified. s vectors, thereby determining the corresponding s A group of optimal weight coefficients; Step S4.4: Use Kriging interpolation to s The optimal weight coefficient group under the current precipitation conditions in the area without measured precipitation data is obtained by weighted average of the optimal weight coefficient groups. Step S4.5: By performing weighted summation of the optimal weight coefficient group and the influencing factors, the precipitation deviation under the current precipitation conditions in the area where measured precipitation data is missing is obtained; Step S5: Add the deviation of the current precipitation conditions in the area lacking measured precipitation data to the corresponding satellite precipitation data to obtain corrected precipitation data, thereby completing the correction of precipitation data in the area lacking measured precipitation data.

2. The method for correcting satellite precipitation data deviation in data-deficient areas according to claim 1, characterized in that: In step S1, the satellite precipitation data, vegetation index data NDVI, wind speed data, and elevation data DEM are unified in resolution using bilinear interpolation, and the national-scale slope, slope aspect, surface roughness, and distance to the coastline data are obtained based on the elevation data DEM at this resolution.

3. The method for correcting satellite precipitation data deviation in data-deficient areas according to claim 1, characterized in that: The influencing factors in step S2, step S3 and step S4 include vegetation index data NDVI, wind speed data, elevation data DEM, slope, aspect, surface roughness, and distance to the coastline.

Citation Information

Cited By

  • Data-deficient region hydrological forecasting method based on multi-source rainfall fusion

    CN122595865A

  • A hydrological forecasting method for areas lacking data based on multi-source precipitation fusion

    CN122595865B