Method for effectively improving satellite rainfall data resolution

Abstract

The invention relates to a method for effectively improving the resolution of satellite rainfall data, which comprises the following steps of: acquiring normalized difference vegetation index (NDVI) data and digital elevation model (DEM) data of a meso-and small-scale watershed, processing the NDVI data and the DEM data, constructing a GWR model based on the processed NDVI data and the DEM data, and acquiring a coefficient, a constant term and a residual error under a target resolution; acquiring global rainfall measurement GPM data, and performing inverse reasoning on the global rainfall measurement GPM data based on the coefficient, the constant term and the residual error under the target resolution to acquire global rainfall measurement GPM data under the target resolution; and according to global rainfall measurement GPM data under the target resolution, obtaining the daily runoff volume of the small and medium-scale watershed. According to the method, firstly, a geographically weighted regression (GWR) model is adopted to carry out spatial downscaling on GPM data so as to improve the precision of input data GPM;

Description

Technical Field

[0001] The present invention relates to the technical field of satellite precipitation data analysis, and in particular to a method for effectively improving the resolution of satellite precipitation data. Background Art

[0002] Mountain station data is the main source of precipitation data, and the precipitation station network provides a guarantee for the simulation and prediction of the precipitation-runoff process in the basin. However, mountainous areas that are difficult for humans to reach often have few or even no stations. The scarcity or even lack of precipitation data is a universal problem, which brings great difficulties to the simulation of basin hydrology. Satellite monitoring precipitation data is a potential alternative data source for ground station monitoring precipitation data and a possible way to solve the problem. However, the highest resolution precipitation product currently available internationally is satellite precipitation GPM data, with a resolution of 0.1°×0.1° (10km×10km). Although this product is available for large-scale basins, it is not available for small and medium-scale basins, and its resolution is far from meeting the accuracy requirements. Summary of the invention

[0003] In order to make GPM data applicable to small and medium-sized watersheds, the present invention first uses a geographically weighted regression model (GWR) to spatially downscale the GPM data to improve the accuracy of the input data GPM.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A method for effectively improving the resolution of satellite precipitation data, comprising:

[0006] Obtaining normalized difference vegetation index NDVI data and digital elevation model DEM data of small and medium-scale watersheds, processing the NDVI data and DEM data, building a GWR model based on the processed NDVI data and DEM data, and obtaining coefficients, constant terms and residuals at a target resolution;

[0007] Acquire global precipitation measurement GPM data, perform reverse reasoning on the global precipitation measurement GPM data based on the coefficient, constant term and residual at the target resolution, and acquire global precipitation measurement GPM data at the target resolution;

[0008] The daily runoff of the small and medium-scale watersheds is obtained according to the global precipitation measurement (GPM) data at the target resolution.

[0009] Optionally, processing the normalized difference vegetation index NDVI data and the digital elevation model DEM data includes:

[0010] The normalized difference vegetation index NDVI data and the digital elevation model DEM data are resampled by using a bilinear interpolation method, and both the normalized difference vegetation index NDVI data and the digital elevation model DEM data are resampled to a preset resolution.

[0011] Optionally, the preset resolution includes: 10km×10km resolution.

[0012] Optionally, the method for obtaining the coefficient, constant term and residual at the target resolution is:

[0013] P i =β NDVI (u i ,v i )NDVI i +β DEM (u i ,v i )DEM i +β0(u i ,v i )+ε i

[0014] Among them, P i is the precipitation at the i-th point; (u i ,v i ) is the longitude and latitude coordinates of the i-th position point; β0(u i ,v i ) is the corresponding coefficient term; ε i is the residual, β NDVI is the vegetation coverage coefficient, NDVI i is the vegetation coverage of the hydrological unit, β DEM is the elevation coefficient, DEM i is the elevation of the hydrological unit.

[0015] Optionally, the global precipitation measurement GPM data at the target resolution includes: global precipitation measurement GPM data at a resolution of 1 km×1 km.

[0016] Optionally, obtaining the daily runoff of the small and medium-scale watershed includes:

[0017] The global precipitation measurement (GPM) data at the target resolution is input into the SRM model to obtain the daily runoff of the small and medium-scale watersheds.

[0018] The beneficial effects of the present invention are:

[0019] The present invention selects NDVI and DEM as independent variables to construct the GWR model, estimates the parameters of each location to establish a geographically weighted regression model (GWR), and uses the geographically weighted regression model to spatially downscale the GPM data to improve the accuracy of the input data GPM. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0021] Figure 1 A flow chart of a method for effectively improving the resolution of satellite precipitation data according to an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of high-resolution (spatial resolution is 1 km×1 km) GPM precipitation data after downscaling in the Jingou River Basin according to an embodiment of the present invention; wherein, Figure 2 (a) is a schematic diagram of the high-resolution (spatial resolution is 1 km × 1 km) GPM precipitation data in April. Figure 2 (b) is a schematic diagram of the high-resolution (spatial resolution is 1 km × 1 km) GPM precipitation data in May. Figure 2 (c) is a schematic diagram of high-resolution (spatial resolution is 1 km × 1 km) GPM precipitation data in June. Figure 2 (d) is a schematic diagram of high-resolution (spatial resolution is 1 km × 1 km) GPM precipitation data in July. Figure 2 (e) is a schematic diagram of the high-resolution (spatial resolution is 1 km × 1 km) GPM precipitation data in August. Figure 2 (f) is a schematic diagram of high-resolution (spatial resolution is 1 km × 1 km) GPM precipitation data in September;

[0023] Figure 3 It is a schematic diagram of simulation results of running the SRM model using ground measured data according to an embodiment of the present invention;

[0024] Figure 4 This is a schematic diagram of simulation results of running the SRM model using downscaled GPM data according to an embodiment of the present invention;

[0025] Figure 5 It is a comparison diagram of the results of data simulation of an embodiment of the present invention; wherein, Figure 5 (a) is the result of simulation with ground-truth data. Figure 5 (b) is the simulation result of GPM data after downscaling. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] like Figure 1 As shown, this embodiment discloses a method for effectively improving the resolution of satellite precipitation data, including: obtaining normalized difference vegetation index NDVI data and digital elevation model DEM data of small and medium-scale watersheds, processing the NDVI data and DEM data, constructing a GWR model based on the processed NDVI data and DEM data, and obtaining coefficients, constant terms and residuals at a target resolution; obtaining global precipitation measurement GPM data, performing reverse reasoning on the global precipitation measurement GPM data based on the coefficients, constant terms and residuals at the target resolution, and obtaining global precipitation measurement GPM data at the target resolution; obtaining daily runoff of small and medium-scale watersheds based on the global precipitation measurement GPM data at the target resolution.

[0029] The processing of NDVI data and DEM data includes: resampling NDVI data and DEM data by bilinear interpolation method, and resampling NDVI data and DEM data to a preset resolution, specifically:

[0030] Data preparation: Since the source data NDVI (1km×1km) and DEM (90m×90m) have different resolutions, the source data are first resampled using bilinear interpolation. The DEM data is resampled to 1km and 10km, and the NDVI data is resampled to 10km.

[0031] Constructing GWR model: Using 10km resolution DEM, NDVI and GPM data to construct GWR model, and then obtain the coefficients, constant terms and residuals at 10km resolution, as shown in formula (a). Then resample the coefficients and constant terms at 10km to 1km, and resample the residuals to 1km using IDW interpolation.

[0032] P i =β NDVI (u i ,v i )NDVI i +β DEM(u i ,v i )DEM i +β0(u i ,v i )+ε i ...(a)

[0033] Where: P i is the precipitation at the ith point (mm); (u i ,v i ) represents the latitude and longitude coordinates of the i-th location point; NDVI is the vegetation coverage of the hydrological unit (%); β NDVI (u i ,v i ) is the vegetation coverage coefficient; DEM is the elevation of the hydrological unit (m); β DEM (u i ,v i ) is the elevation coefficient; β0(u i ,v i ) is the corresponding coefficient term; ε i is the residual.

[0034] Inversely infer GPM data: Based on the coefficients, auxiliary variables, constant terms and residuals obtained above at 1 km, the downscaling results of the GPM data at 1 km are obtained according to formula a, as shown in Figure 2 As shown in (a)-(f).

[0035] Obtaining the daily runoff of small and medium-scale basins includes: inputting the global precipitation measurement GPM data at the target resolution into the SRM model to obtain the daily runoff of small and medium-scale basins, specifically:

[0036] The ground-based measured data and downscaled high-resolution remote sensing data were input into the SRM model to simulate the multi-year daily snowmelt runoff in the Jingou River Basin on the northern slope of the Tianshan Mountains. Figure 3 , Figure 4 ) and verification( Figure 5 The results of (a)-(b)) are as follows:

[0037] Effect evaluation of this embodiment: The R2 of the ground measured data and the downscaled GPM data are 0.85 and 0.84 respectively, and the results are very different. The downscaled GPM precipitation data (1km×1km) was substituted into the SRM model to simulate the daily runoff of the Jingou River Basin, and compared with the results simulated using the measured precipitation data. As a result, the accuracy of simulating runoff using downscaled GPM remote sensing precipitation data is 2.5% lower than that of simulating runoff using measured data, and a very good effect was achieved. It can be seen that this product after downscaling can be used for snowmelt runoff simulation in high-cold mountainous areas where observation data is lacking. It has a good application prospect as a substitute data for mountainous areas with scarce data, solves the problem of scarcity of precipitation data in high-altitude mountainous areas without measurement stations, and provides an alternative high-resolution precipitation data source for hydrological models with low accuracy.

[0038] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for effectively improving the resolution of satellite precipitation data, characterized in that: include: Obtaining normalized difference vegetation index NDVI data and digital elevation model DEM data of small and medium-scale watersheds, processing the NDVI data and DEM data, building a GWR model based on the processed NDVI data and DEM data, and obtaining coefficients, constant terms and residuals at a target resolution; Acquire global precipitation measurement GPM data, perform reverse reasoning on the global precipitation measurement GPM data based on the coefficient, constant term and residual at the target resolution, and acquire global precipitation measurement GPM data at the target resolution; The daily runoff of the small and medium-scale watersheds is obtained according to the Global Precipitation Measurement (GPM) data at the target resolution.

2. The method for effectively improving the resolution of satellite precipitation data according to claim 1, characterized in that: Processing the normalized difference vegetation index NDVI data and the digital elevation model DEM data includes: The normalized difference vegetation index NDVI data and the digital elevation model DEM data are resampled by using a bilinear interpolation method, and both the normalized difference vegetation index NDVI data and the digital elevation model DEM data are resampled to a preset resolution.

3. The method for effectively improving the resolution of satellite precipitation data according to claim 2, characterized in that: The preset resolution includes: 10km×10km resolution.

4. The method for effectively improving the resolution of satellite precipitation data according to claim 1, characterized in that: The method for obtaining the coefficients, constant terms and residuals at the target resolution is: P i =β NDVI (u i ,v i )NDVI i +β DEM (u i ,v i )DEM i +β0(u i ,v i )+ε i Among them, P i is the precipitation at the i-th point; (u i ,v i ) is the longitude and latitude coordinates of the i-th position point; β0(u i ,v i ) is the corresponding coefficient term; ε i is the residual, β NDVI is the vegetation coverage coefficient, NDVI i is the vegetation coverage of the hydrological unit, β DEM is the elevation coefficient, DEM i is the elevation of the hydrological unit.

5. The method for effectively improving the resolution of satellite precipitation data according to claim 1, characterized in that: The global precipitation measurement GPM data at the target resolution includes: global precipitation measurement GPM data at a resolution of 1km×1km.

6. The method for effectively improving the resolution of satellite precipitation data according to claim 1, characterized in that: Obtaining the daily runoff of the small and medium-scale watershed includes: The global precipitation measurement (GPM) data at the target resolution is input into the SRM model to obtain the daily runoff of the small and medium-scale watersheds.

Images