Plateau mountainous area rainfall erosivity estimation optimization method based on satellite rainfall product

By constructing an interpolation model and accuracy evaluation indicators, and combining satellite precipitation products and ground station data, the problem of accuracy in estimating rainfall erosivity in plateau and mountainous areas was solved, achieving rapid and accurate estimation of rainfall erosivity, which is applicable to soil erosion research in different regions.

CN119026453APending Publication Date: 2024-11-26KUNMING UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410932287.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately obtain continuous high spatiotemporal resolution precipitation data across high-altitude and mountainous regions, leading to inaccurate estimations of rainfall erosivity. Furthermore, satellite precipitation products exhibit inconsistent performance under different geographical conditions, lacking adaptability studies.

Method used

By acquiring multi-source satellite precipitation products, ground meteorological station data, and DEM data, an interpolation model was constructed using ANSUPLIN software. Accuracy was evaluated by combining various statistical and classification indicators, and a precipitation erosivity estimation model for plateau mountainous areas was established, thus optimizing the precipitation erosivity estimation process.

Benefits of technology

It enables rapid and accurate estimation of rainfall erosivity in plateau and mountainous areas, provides valuable soil erosion index information, saves time and resources, and has portability and versatility.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119026453A_ABST
    Figure CN119026453A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of soil erosion, and discloses a plateau mountain area rainfall erosivity estimation optimization method based on satellite rainfall products. According to the method, the special geographical environment characteristics of the plateau mountain area are considered, the performance of the satellite precipitation product is comprehensively evaluated from multiple spatial and temporal scales, reliable support can be provided for selecting the optimal satellite precipitation product in a region, and valuable soil erosion index information is provided for the plateau mountain area; and meanwhile, the rainfall erosivity of the region can be quickly and accurately estimated by acquiring the continuous high-temporal-spatial-resolution satellite rainfall product of the target region, so that time, manpower, material resources and financial resources are saved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of soil erosion, and in particular to an optimization method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products. Background Art

[0002] Soil erosion is the main cause of land degradation worldwide and the number one ecological and environmental problem in plateau mountainous areas. As the indicator factor with the strongest temporal and spatial heterogeneity and variability in soil erosion, large-scale, high-intensity precipitation not only causes a series of flood disasters such as urban inundation, landslides, and mudslides, but also causes raindrop splashing, which directly provides the basic driving force for the separation of soil particles, surface runoff scouring and transportation. The rainfall erosivity factor characterizes the potential impact of rainfall on soil erosion. It works together with factors such as soil, topography, and surface vegetation cover to determine the actual intensity of soil and water loss in the region. Therefore, accurately calculating rainfall erosivity is the key to preventing soil and water loss in plateau mountainous areas, building regional ecological environment, and protecting food security. However, the plateau mountainous areas have fragmented terrain and complex and changeable climate. How to obtain high-precision spatial and temporal distribution information of precipitation is a prerequisite for accurate estimation of rainfall erosivity;

[0003] Traditional regional precipitation data are usually based on discrete point-like rain gauge observation data and calculated with the help of spatial interpolation methods. Although the regional precipitation data obtained by this method is most consistent with the actual situation, due to the lack of dense and continuous observation points, and the problems of short recording time and missing information caused by the irregular monitoring process of local stations, the observation data are poorly representative. The limited station data can only reflect the precipitation situation within a certain radius of a single point location, and it is difficult to obtain a reliable global continuous high-temporal and spatial resolution precipitation data set in mountainous areas. As an alternative, ground-based radar can provide real-time and accurate precipitation observations, but its actual detection performance is often affected by factors such as beam blocking, electromagnetic wave attenuation and other measurement errors, and large-scale precipitation observations cannot be achieved. Although some influencing factors can be solved by network radar, it is not suitable for mountainous and marine areas. With the gradual improvement of satellite remote sensing technology and sensor resolution, multi-satellite joint inversion precipitation technology has overcome the limitations of station rain gauges and ground radars, and has become the primary way to quickly capture high-temporal and spatial resolution precipitation data. Since the 20th century, a series of satellite high-temporal and spatial resolution precipitation products such as TRMM, CHIRPS, ERA5, IMERG, and PERSIANN have been derived, which have made it possible to observe regional precipitation efficiently and accurately. Related studies have shown that satellite precipitation datasets have good applicability in precipitation observations in most regions, and their applications have involved research fields such as extreme precipitation analysis, hydrological modeling, and rainfall erosivity characteristics assessment.

[0004] Although the above studies can obtain spatially continuous precipitation information, the quality of satellite observations is essentially affected by systematic or random errors caused by differences in sensor parameters, observation methods, and inversion algorithms. In addition, due to the heterogeneity of terrain and climate conditions in different regions, satellite precipitation products will show unique performance levels in different spatiotemporal environments. Therefore, it is necessary to conduct an adaptability assessment of satellite precipitation products before applying them to specific regional studies. In addition, there are several problems in the current research on satellite precipitation assessment: (1) The research scale is single, and the regional and grid-scale precipitation observation accuracy cannot be analyzed from the annual, seasonal, and monthly scales; (2) The regional geographical environment characteristics are not well considered, and there is a lack of in-depth discussion on the causes of satellite precipitation estimation errors caused by unique terrain in special regions; (3) There is little research on the adaptability of satellite precipitation products in plateau and mountainous areas.

[0005] Therefore, how to evaluate the observation accuracy of satellite precipitation products in plateau and mountainous areas and then optimize the estimation of rainfall erosivity is an urgent problem that technicians in this field need to solve. Summary of the invention

[0006] 1. Technical issues to be resolved

[0007] In view of the shortcomings of the existing technology, the present invention provides an optimization method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products, which has the advantages of quickly and accurately estimating the rainfall erosivity in a region, and solves the above-mentioned technical problems.

[0008] (II) Technical solution

[0009] To achieve the above object, the present invention provides the following technical solution: a method for estimating and optimizing rainfall erosivity in plateau mountainous areas based on satellite precipitation products, comprising the following steps:

[0010] S1. Obtain multi-source satellite precipitation products, ground meteorological station data and DEM data in the target area, and pre-process the multi-source satellite precipitation products, ground meteorological station data and DEM data;

[0011] S2. Based on the preprocessed ground meteorological station data and DEM data, the SPLINA and LAPGRD modules in the ANSUPLIN software were used to set different spline times, and interpolation models were constructed under DEM data of different spatial resolutions. The optimal interpolation model of the precipitation data of the ground meteorological station was determined by diagnosing the error sources and evaluating the interpolation quality parameters generated by different models. The spatialized layer of the precipitation information observed at the ground meteorological station was obtained based on the optimal interpolation model, and the spatial point comparison of the measured precipitation values ​​and the interpolation prediction values ​​of the ground meteorological station was verified at the annual and monthly scales.

[0012] S3. Using the interpolated ground meteorological station precipitation observation raster layer as a reference, multiple statistical indicators and multiple classification indicators are used to evaluate the accuracy of the pre-processed satellite precipitation products at different temporal and spatial scales and vertical zonality;

[0013] S4. Establish a rainfall erosivity estimation model for plateau mountainous areas to obtain continuous rainfall erosivity with high temporal and spatial resolution over the entire plateau mountainous area.

[0014] As a preferred technical solution of the present invention, the preprocessing in step S1 includes summarizing, resampling, converting units and unifying the observation period of precipitation information on annual, seasonal, monthly and daily scales for multi-source satellite precipitation products; screening for missing values, data classification and unit conversion for ground meteorological station data; and resampling, hierarchical statistics and format conversion for DEM data.

[0015] As a preferred technical solution of the present invention, the parameters for diagnosing the source of error and evaluating the interpolation quality in step S2 include mean (Mean), standard deviation (SD), signal degrees of freedom (Signal), residual degrees of freedom (Error), signal-to-noise ratio (SNR), generalized cross validation (GCV), expected true mean square error (MSE), mean square residual (MSR), variance (VAR) estimation and its square root;

[0016] The criteria for determining the optimal interpolation model in step S2 are: minimum generalized cross validation, minimum signal-to-noise ratio, signal degrees of freedom less than half the number of ground meteorological stations, and no * symbol displayed in the error file generated by SPLINA.

[0017] As a preferred technical solution of the present invention, the statistical indicators used in step S3 include Pearson correlation coefficient CC, root mean square error RMSE, central root mean square error RMSD, mean error ME, relative deviation Bias, standard deviation STD and consistency index KGE, and the classification indicators used in step S3 include key success index CSI, hit rate POD and false alarm rate FAR.

[0018] As a preferred technical solution of the present invention, the accuracy assessment of different spatiotemporal scales in step S3 is based on the Voronoi partitioning idea and K-Means clustering method to describe the spatial distribution law of grid scale-related indicators, specifically including:

[0019] S3a.1. Based on the ground meteorological stations in the target area, construct Thiessen polygons through the spatial analysis tools in ArcGIS and obtain the mean value of the accuracy evaluation index in each Thiessen polygon;

[0020] S3a.2, using the K-Means clustering method in GeoDa software to analyze the accuracy evaluation index in each Thiessen polygon;

[0021] The vertical zonality accuracy assessment in step S3 specifically includes: using the DEM data after hierarchical statistics in step S1, obtaining precipitation information of satellite observations and ground station observations in different elevation intervals through overlay analysis in ArcGIS.

[0022] As a preferred technical solution of the present invention, the rainfall erosivity estimation model expression in the plateau mountainous area in step S4 is as follows:

[0023]

[0024] R j =5.249WF j

[0025]

[0026] β=1.205

[0027] Among them, R is the annual rainfall erosivity, R j is the rainfall erosivity of the jth month, P is the annual rainfall, P j is the rainfall in month j, F j is the weighted rainfall in the jth month, W is the weighted average annual rainfall, β is the model parameter, Indicates the sum of 12 months of data.

[0028] Compared with the prior art, the present invention provides an optimization method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products, which has the following beneficial effects:

[0029] 1. The present invention takes into account the special geographical environment characteristics of plateau mountainous areas, comprehensively evaluates the performance of satellite precipitation products from multiple temporal and spatial scales, can provide reliable support for the selection of the optimal satellite precipitation products for the region, and provide valuable soil erosion indicator information for plateau mountainous areas. At the same time, by obtaining continuous high-temporal and spatial resolution satellite precipitation products in the target area, the rainfall erosion force of the region can be estimated quickly and accurately, saving time, manpower, material resources and financial resources.

[0030] 2. The present invention provides a satellite precipitation data-ground station combined rainfall erosivity estimation optimization method, which has transferability and versatility in the study of rainfall erosivity factors of soil erosion in different regions. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the process of the present invention;

[0032] Figure 2 It is a schematic diagram of the framework of the present invention. DETAILED DESCRIPTION

[0033] 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.

[0034] See also Figure 1 - Figure 2 , a method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products, comprising the following steps:

[0035] S1. Obtain multi-source satellite precipitation products, ground meteorological station data and DEM data in the target area, and pre-process the multi-source satellite precipitation products, ground meteorological station data and DEM data;

[0036] Multi-source satellite precipitation data (TRMM, IMERG, CHIRPS, ERA5, PERSIANN-CDR) and DEM data were obtained based on Google Earth Engine; ground meteorological station data were obtained based on the "China Ground Climate Data Log Dataset (V3.0)" of the National Meteorological Data Center; specific data information is shown in Table 1; the preprocessing of multi-source satellite precipitation products, ground meteorological station data and DEM data includes: summarizing and arranging precipitation information on annual, seasonal, monthly and daily scales for multi-source satellite precipitation products, resampling, unit conversion and unifying observation periods; screening for missing values, data classification and unit conversion for ground meteorological station data; resampling, hierarchical statistics and format conversion for DEM data;

[0037]

[0038] Table 1

[0039] S2. Based on the preprocessed ground meteorological station data and DEM data, the SPLINA and LAPGRD modules in the ANSUPLIN software were used to set different spline times, and interpolation models were constructed under DEM data of different spatial resolutions. The optimal interpolation model of the precipitation data of the ground meteorological station was determined by diagnosing the error sources and evaluating the interpolation quality parameters generated by different models. The spatialized layer of the precipitation information observed at the ground meteorological station was obtained based on the optimal interpolation model, and the spatial point comparison of the measured precipitation values ​​and the interpolation prediction values ​​of the ground meteorological station was verified at the annual and monthly scales.

[0040] In the above step S2, based on the two program modules SPLINA and LAPGRD in the ANUSPLIN software, different spline times are set to build interpolation models under DEM data of different spatial resolutions. The optimal interpolation model is selected by sorting out the diagnostic error sources and parameters for evaluating the interpolation quality generated by different models. The measured precipitation values ​​of the target area are compared with the predicted values ​​of the optimal interpolation model at the annual and monthly scales, where:

[0041] The parameters for diagnosing the source of error and evaluating the interpolation quality include mean, standard deviation, signal degrees of freedom, residual degrees of freedom, signal-to-noise ratio, generalized cross validation, expected true mean square error, mean square residual, variance (VAR) estimate and its square root. The optimal model judgment criteria are as follows: ① minimum generalized cross validation; ② minimum signal-to-noise ratio; ③ signal degrees of freedom are less than half of the number of sites; ④ no * symbol is displayed in the error file generated by SPLINA.

[0042] Generally speaking, all four criteria must be met at the same time. First, the model can be roughly filtered according to items ③ and ④, and then items ① and ② are judged, but item ① has the highest priority.

[0043] S3. Using the interpolated ground meteorological station precipitation observation raster layer as a reference, multiple statistical indicators and multiple classification indicators are used to evaluate the accuracy of the pre-processed satellite precipitation products at different temporal and spatial scales and vertical zonality;

[0044] The various statistical indicators include Pearson correlation coefficient (CC), root mean square error (RMSE), central root mean square error (RMSD), mean error (ME), relative deviation (Bias), standard deviation (STD) and Kling-Gupta efficiency (KGE); the various classification indicators include key success index (CSI), hit rate (POD) and false alarm rate (FAR). The above accuracy evaluation indicators are shown in Table 2; spatial statistical analysis includes Voronoi subdivision, K-Means clustering and vertical zonality characteristics; among them:

[0045] In order to more clearly characterize the error characteristics of the precipitation dataset, a Taylor diagram for precipitation intensity assessment is introduced. This diagram takes into account the arithmetic triangular conversion relationship between CC, STD and RMSD between the two sets of data series and is presented in the form of a polar coordinate diagram; each point in the diagram represents three statistics at the same time, thereby more effectively evaluating the degree of match between each satellite precipitation product and the observed data.

[0046] The calculation of CSI requires a threshold to determine whether there is rain or no rain. Percentiles (25th, 50th, 75th and 90th percentiles) are used to divide the thresholds of precipitation events observed by ground stations in different periods, and then the precipitation detection of multi-source satellite precipitation products under the corresponding thresholds is evaluated;

[0047]

[0048] Table 2

[0049] Where n is the total number of samples; S i and G i represents the precipitation observation data samples of the ith satellite and ground station, respectively, S and G are the corresponding average values; H is the number of precipitation events hit by the satellite, that is, the number of precipitation events detected by both the satellite dataset and the observation dataset; M is the number of precipitation events missed by the satellite, that is, the number of missed precipitation events detected by the observation dataset but not by the satellite dataset; F is the number of false precipitation events reported by the satellite, that is, the number of false precipitation events detected by the satellite dataset but not by the observation dataset; r is the Pearson correlation coefficient between S and G; β is the offset rate; γ is the coefficient of variation; μ is the average precipitation; σ is the standard deviation;

[0050] Based on the Voronoi subdivision idea and K-Means clustering method, the spatial distribution law of grid scale related indicators is characterized, including: based on the ground meteorological stations in the target area, the Thiessen polygons are constructed through the spatial analysis tools in ArcGIS to obtain the mean of the accuracy assessment indicators in each Thiessen polygon; the K-Means clustering method in GeoDa software is used to analyze the accuracy indicators in each Thiessen polygon to obtain the spatial distribution law of grid scale accuracy indicators.

[0051] Based on the vertical zonality characteristics, the accuracy of precipitation observations is evaluated, and satellite precipitation products that are suitable for the special geographical environment of plateau and mountainous areas are preferred. Specifically, the DEM data after hierarchical statistics are used to obtain precipitation information from satellite observations and ground station observations in different elevation intervals through overlay analysis in ArcGIS, and then relevant accuracy indicators are used to characterize the response characteristics of satellite precipitation estimates to terrain undulations.

[0052] S4. Establish a rainfall erosivity estimation model for plateau mountainous areas to obtain continuous rainfall erosivity with high temporal and spatial resolution over the entire plateau mountainous area;

[0053] By collecting a large number of literature, the commonly used rainfall erosivity estimation models are summarized and concluded, and then the rainfall erosivity estimation model suitable for plateau mountainous areas is determined; based on the optimal satellite precipitation products, the rainfall erosivity estimation model suitable for plateau mountainous areas is adopted to obtain the annual and monthly scales of the whole area continuous high temporal and spatial resolution rainfall erosivity information in plateau mountainous areas; among them:

[0054] In southern China, commonly used rainfall erosivity estimation models include: the monthly rainfall estimation model established by Wischmeier et al.; the modified daily rainfall model established by Shi Zhihua et al.; the monthly rainfall model established by Shi Dongmei et al. using artificial rainfall simulation experiments; the CREAMS model; the model established by Zhang Wenbo et al. using daily rainfall; and the daily rainfall estimation model for cold and warm seasons established by Yin Shuiqing et al. by distinguishing between dry and rainy seasons. According to the characteristics of satellite precipitation products and the results of model adaptability research, the rainfall erosivity model based on monthly average rainfall proposed by Shi Dongmei et al. is more suitable for the mountainous areas of Yunnan Plateau. The specific calculation formula is:

[0055]

[0056] R j =5.249WF j

[0057]

[0058] β=1.205

[0059] Among them, R is the annual rainfall erosivity, R j is the rainfall erosivity of the jth month, P is the annual rainfall, P j is the rainfall in month j, F j is the weighted rainfall in the jth month, W is the weighted average annual rainfall, β is the model parameter, Indicates the sum of 12 months of data;

[0060] Next, the Jinsha River dry-hot valley is used as a research area to illustrate the optimization method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products provided by the present invention. Figure 2 shown.

[0061] The dry-hot river valley of Jinsha River is the most widely distributed river valley belt in my country. Due to its location in a complex area with transverse mountains and deep river valleys, the distribution of rain gauge monitoring stations is uneven, the density is low, and most of them are located in low-altitude areas in the river valley. It has the disadvantages of poor representativeness and discontinuity, making it difficult to accurately capture the precipitation information with high spatiotemporal resolution in the whole region, which brings difficulties to the effective implementation of regional soil and water conservation projects. In the example of the present invention, the special geographical environment characteristics of the plateau mountainous area are comprehensively considered, and a variety of accuracy evaluation indicators, spatial interpolation methods and spatial analysis methods are fully utilized to achieve satellite observation-ground station combined rainfall erosivity estimation optimization in the plateau mountainous area, solve the data gap of continuous high spatiotemporal resolution precipitation information in the whole region of the plateau mountainous area, and provide valuable soil erosion indicator information for the region.

[0062] The specific contents are as follows:

[0063] Step 1: Obtain multi-source satellite precipitation products, ground meteorological station data and DEM data in the target area, and pre-process the multi-source satellite precipitation products, ground meteorological station data and DEM data; specifically including:

[0064] (1) Multi-source satellite precipitation data (TRMM, IMERG, CHIRPS, ERA5, PERSIANN-CDR) and DEM data of the study area obtained based on the Google Earth Engine platform; ground meteorological station data of the study area obtained based on the "China Ground Climate Data Log Dataset (V3.0)" of the National Meteorological Data Center; the satellite precipitation data and ground station data span the period from 2001 to 2019;

[0065] (2) The preprocessing of the above data includes: summarizing, resampling, unit conversion and unifying the observation period of precipitation information on annual, seasonal, monthly and daily scales for multi-source satellite precipitation products; screening for missing values, data classification and unit conversion of ground meteorological station data, and finally selecting a total of 32 meteorological stations in the study area and its surrounding areas; resampling, hierarchical statistics and format conversion of DEM data.

[0066] Step 2: Based on the pre-processed ground meteorological station data and DEM data, the ANSUPLIN software is used to determine the optimal interpolation model of the station precipitation data to obtain the spatialized layer of the station observation precipitation information; multi-scale spatial point comparison verification is performed based on the station precipitation measured value and interpolation prediction value; specifically including:

[0067] (1) Based on the two program modules SPLINA and LAPGRD in ANUSPLIN software, the longitude and latitude of the ground station were used as independent variables, and the 1km and 90m resolution SRTM DEM data were used as covariates, respectively. Different spline times were set, and a total of three interpolation models were constructed (see Table 3). The diagnostic error sources and parameters for evaluating the interpolation quality generated by different models were sorted out (see Table 4) to select the optimal interpolation model. After conducting experiments based on the actual situation in the study area, it was found that when the spline times were set to 2, the GCV reached the lowest value. Therefore, Model 1 was selected as the optimal standby model. In order to match the resolution of the precipitation data, 1km resolution SRTM DEM was finally selected as the covariate and input into Model 1 for interpolation.

[0068] (2) The measured values ​​and interpolated predicted values ​​of 32 meteorological stations in the study area were compared at the spatial point positions on the annual and monthly scales, and the statistical indicators used for accuracy evaluation were CC, |ME| and RMSE. In terms of CC, the CC of each station on the monthly scale was above 0.90; in terms of |ME| and RMSE, |ME| was concentrated in the range of 0.210-19.753 mm / month and 0.034-19.422 mm / year, with average values ​​of 11.454 mm / month and 11.443 mm / year respectively; RMSE was mostly distributed in the range of 9.482-32.443 mm / month and 0.117-67.278 mm / year, with average values ​​of 20.925 mm / month and 39.639 mm / year respectively.

[0069]

[0070] Table 3 Three candidate thin disk smooth spline function models

[0071]

[0072] Table 4 Error statistical parameters of each model

[0073] Step 3. Take the interpolated station precipitation observation raster layer as a reference, combine 7 statistical indicators and 3 classification indicators, supplemented by Voronoi subdivision and K-Means clustering method, and conduct multi-spatiotemporal scale and vertical zonality error analysis on multi-source satellite precipitation products. In terms of space, starting from the overall distribution and grid scale, using precipitation spatial distribution map, scatter density map and Thiessen polygon K-means clustering distribution map, etc., explore the error distribution characteristics of satellite data in space; in terms of time, considering the annual, seasonal and monthly scales, combining time series curves with visual expressions, clarify the performance of satellite precipitation products at different time scales;

[0074] Step 4. Based on the observational accuracy performance of multi-source satellite precipitation products (TRMM, IMERG, CHIRPS, ERA5, PERSIANN-CDR) at multiple temporal and spatial scales, select the optimal satellite precipitation product suitable for the study area. On the annual scale, the performance of IMERG long-term series is significantly better than other products, and PCDR and ERA5 need to be considered with caution. On the seasonal scale, the dry season is the best period for estimating precipitation, and overestimation occurs frequently in the rainy season with abundant precipitation. On the monthly scale, the performance of CHIRPS, PCDR and ERA5 is stable from November to April, while the performance decreases and is greatly unbalanced from May to October; in addition, compared with the area above 1600m, the 1200-1600m elevation zone performs better, and IMERG and T3B42 have excellent performance in each elevation range, followed by CHIRPS, while PCDR and ERA5 are relatively more affected by elevation disturbances; in general, IMERG satellite precipitation products can be used as a substitute for regional precipitation observation data gaps;

[0075] Based on the IMERG satellite precipitation products from 2001 to 2019, the rainfall erosivity in the study area was estimated on an annual and monthly scale using the monthly average rainfall estimation rainfall erosivity model, and characteristic analysis was carried out. On the annual scale, the regional R factor value is generally low, with an average value of 2929.98MJ·mm / (hm2·h·a), and the distribution pattern decreases from northwest to southeast with the northeast corner and southwest corner as the boundary; on the monthly scale, the R value from January to May and October to December is low, showing a distribution pattern of low northwest and high southeast. Due to the gradual strengthening of the monsoon after May, the R value from June to September increases and shows a distribution pattern of high northwest and low southeast. During this period, the rainfall erosivity accounts for 90.94% of the annual rainfall erosivity, which is easy to cause serious soil erosion.

[0076] In summary, the example of the present invention provides an optimization method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products. The method adopts a combination of global continuous high-precision satellite precipitation products, multiple precision evaluation indicators, spatial interpolation methods and spatial analysis. Compared with the method of observing precipitation at traditional rain gauge sites, the invention comprehensively considers the special geographical environment characteristics of plateau mountainous areas, and makes full use of multiple precision evaluation indicators, spatial interpolation methods and spatial analysis methods. It can provide reliable support for the selection of optimal satellite precipitation products in the region, and provide valuable soil erosion indicator information for plateau mountainous areas. The invention can quickly and accurately estimate the rainfall erosivity of a region, saving time, manpower, material resources and financial resources, and the research on rainfall erosivity factors of soil erosion in different regions is transferable and universal, which is of great significance for preventing regional soil erosion, building a regional ecological environment, and protecting food security.

[0077] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for estimating rainfall erosivity in plateau mountainous areas based on satellite precipitation products, characterized by: The following steps are involved: S1. Obtain multi-source satellite precipitation products, ground meteorological station data and DEM data in the target area, and pre-process the multi-source satellite precipitation products, ground meteorological station data and DEM data; S2. Based on the preprocessed ground meteorological station data and DEM data, the SPLINA and LAPGRD modules in the ANSUPLIN software were used to set different spline times, and interpolation models were constructed under DEM data of different spatial resolutions. The optimal interpolation model of the precipitation data of the ground meteorological station was determined by diagnosing the error sources and evaluating the interpolation quality parameters generated by different models. The spatialized layer of the precipitation information observed at the ground meteorological station was obtained based on the optimal interpolation model, and the spatial point comparison of the measured precipitation values ​​and the interpolation prediction values ​​of the ground meteorological station was verified at the annual and monthly scales. S3. Using the interpolated ground meteorological station precipitation observation raster layer as a reference, multiple statistical indicators and multiple classification indicators are used to evaluate the accuracy of the pre-processed satellite precipitation products at different temporal and spatial scales and vertical zonality; S4. Establish a rainfall erosivity estimation model for plateau mountainous areas to obtain continuous rainfall erosivity with high temporal and spatial resolution over the entire plateau mountainous area.

2. The method for estimating and optimizing rainfall erosivity in plateau mountainous areas based on satellite precipitation products according to claim 1 is characterized by: The preprocessing in step S1 includes summarizing, resampling, unit conversion and unifying the observation period of precipitation information on annual, seasonal, monthly and daily scales for multi-source satellite precipitation products; screening for missing values, data classification and unit conversion for ground meteorological station data; and resampling, hierarchical statistics and format conversion for DEM data.

3. The method for estimating and optimizing rainfall erosivity in plateau and mountainous areas based on satellite precipitation products according to claim 1 is characterized by: The parameters for diagnosing the error sources and evaluating the interpolation quality in step S2 include mean, standard deviation, signal degrees of freedom, residual degrees of freedom, signal-to-noise ratio, generalized cross validation, expected true mean square error, mean square residual, variance estimate and its square root; The criteria for determining the optimal interpolation model in step S2 are: minimum generalized cross validation, minimum signal-to-noise ratio, signal degrees of freedom less than half the number of ground meteorological stations, and no * symbol displayed in the error file generated by SPLINA.

4. The method for estimating and optimizing rainfall erosivity in plateau mountainous areas based on satellite precipitation products according to claim 2 is characterized by: The statistical indicators used in step S3 include Pearson correlation coefficient CC, root mean square error RMSE, central root mean square error RMSD, mean error ME, relative deviation Bias, standard deviation STD and KGE, and the classification indicators used in step S3 include key success index CSI, hit rate POD and false alarm rate FAR.

5. The method for optimizing rainfall erosivity estimation in plateau mountainous areas based on satellite precipitation products according to claim 4 is characterized by: The accuracy assessment of different spatiotemporal scales in step S3 is based on the Voronoi partitioning idea and K-Means clustering method to describe the spatial distribution law of grid point scale-related indicators, specifically including: S3a.

1. Based on the ground meteorological stations in the target area, construct Thiessen polygons through the spatial analysis tools in ArcGIS and obtain the mean value of the accuracy evaluation index in each Thiessen polygon; S3a.2, using the K-Means clustering method in GeoDa software to analyze the accuracy evaluation index in each Thiessen polygon; The vertical zonality accuracy assessment in step S3 specifically includes: using the DEM data after hierarchical statistics in step S1, obtaining precipitation information of satellite observations and ground station observations in different elevation intervals through overlay analysis in ArcGIS.

6. The method for estimating and optimizing rainfall erosivity in plateau mountainous areas based on satellite precipitation products according to claim 1 is characterized by: The rainfall erosivity estimation model in the plateau mountainous area in step S4 is expressed as follows: R j =5.249WF j β=1.205 Among them, R is the annual rainfall erosivity, R j is the rainfall erosivity of the jth month, P is the annual rainfall, P j is the rainfall in month j, F j is the weighted rainfall in the jth month, W is the weighted average annual rainfall, β is the model parameter, Indicates the sum of 12 months of data.