A method and system for attribution identification of sediment flux evolution in high-altitude and cold mountainous areas

Abstract

This invention proposes a method and system for attributing and identifying sediment flux evolution in high-altitude and cold mountainous areas, belonging to the field of sediment evolution attribution and identification technology. The method includes: using a joint mutation identification method to detect mutation points from multiple perspectives based on a driving factor dataset; using the Pettitt test to identify grade changes under distribution mutations to determine the year of distribution mutation; using the Mann-Kendall trend test to identify slope changes under trend mutations to determine the year of trend mutation; detecting mutation points based on the difference between the year of distribution mutation and the year of trend mutation, and dividing the evolution stages; and performing attribution analysis on all driving factors based on a three-stage progressive attribution analysis framework. This invention establishes a sediment attribution and identification framework for high-altitude and cold mountainous areas by integrating multivariate statistical learning and path modeling techniques, improving the stability and accuracy of attribution results, and providing decision support for soil and water conservation and regional ecological restoration.

Description

Technical Field

[0001] This invention belongs to the field of sediment evolution attribution identification technology, and particularly relates to a method and system for attribution identification of sediment flux evolution in high-altitude and cold mountainous areas. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Attribution identification of sediment flux evolution is a core technology for soil and water conservation and ecological restoration, especially in high-altitude mountainous areas where special climate and topographic conditions (such as freeze-thaw cycles, seasonal snowmelt, and vegetation vulnerability) lead to highly complex sediment transport mechanisms.

[0004] However, existing methods for attribution identification of sediment flux evolution in high-altitude and cold mountainous areas have some intractable technical problems:

[0005] (1) Existing methods often consider the influence of each driving factor individually, failing to effectively identify the complex interaction effects between driving factors, resulting in an incomplete analysis of the influence mechanism of sediment evolution. Sediment evolution is a complex process involving the combined effects of various anthropogenic factors (such as water conservancy projects, land use, etc.) and natural factors (such as climate, vegetation, topography, etc.). A single analytical method cannot accurately reveal the interaction and mutual influence between these driving factors.

[0006] (2) In the prior art, the identification of abrupt change points and the division of evolution stages in sediment concentration (SSC) data lack accuracy, resulting in unreliable analysis results. For example, when dealing with abrupt changes in complex sediment evolution processes, the prior art lacks efficient and accurate statistical methods, making it impossible to accurately identify abrupt change points or divide the evolution process into different stages.

[0007] (3) Existing methods for analyzing sediment evolution often neglect the nonlinear interactions between multiple driving factors, resulting in an incomplete analysis of the complex causal relationships in the sediment evolution process and affecting the accuracy of the attribution results. Because multiple driving factors in the sediment evolution process have strong interactions, a single model cannot accurately capture the complex relationships between these driving factors, often leading to instability and errors in the attribution results.

[0008] (4) Existing technologies often lack sufficient verification of attribution results in sediment evolution, especially in complex eco-hydrological systems such as high-altitude and cold mountainous areas, where the stability and interpretability of attribution results are poor. Existing methods may not be able to provide results with strong physical consistency when dealing with attribution analysis of multiple factors, which affects their application value. Summary of the Invention

[0009] To overcome the shortcomings of the prior art, this invention provides a method and system for attributing and identifying sediment flux evolution in high-altitude and cold mountainous areas. By integrating statistical learning and multi-model analysis techniques, it can accurately identify the key driving factors and their interactions in the sediment evolution process.

[0010] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0011] The first aspect of this invention provides a method for attributing and identifying the evolution of sediment flux in high-altitude and cold mountainous areas.

[0012] A method for attributing and identifying the evolution of sediment flux in high-altitude and cold mountainous areas includes:

[0013] Based on the sediment transport mechanism in high-altitude and cold mountainous areas, the main driving factors were identified;

[0014] Obtain the basic data corresponding to the main driving factors, preprocess them, and construct the driving factor dataset;

[0015] A joint mutation identification method was adopted to detect mutation points from multiple perspectives based on the established driving factor dataset: the Pettitt test was used to identify the grade changes under the distribution mutation to determine the year of the distribution mutation; the Mann-Kendall trend test was used to identify the slope changes under the trend mutation to determine the year of the trend mutation; mutation points were detected based on the difference between the obtained distribution mutation years and trend mutation years, and the evolution stages were divided.

[0016] Attribution analysis is performed on all driving factors based on a three-stage progressive attribution analysis framework to quantify the attribution contribution of each driving factor at different evolution stages and in different spaces.

[0017] Furthermore, mutation points are detected based on the difference between the years of distribution mutation and the years of trend mutation, and evolution stages are divided, including: if the years of distribution mutation and the years of trend mutation are the same or within 3 years, the determined years of distribution mutation and trend mutation are defined as stable mutation points; if the difference between the years of distribution mutation and the years of trend mutation is greater than 3 years, the results of the Petitt test are used.

[0018] Furthermore, based on the difference between the years of distributional abrupt change and the years of trend abrupt change, the mutation points are detected and the evolution stages are divided. This also includes: if multiple trend change segments are detected through the Mann-Kendall trend test, the main mutation segment and the secondary stage of the evolution stages are divided according to the significance statistic of the Pettitt test.

[0019] Furthermore, the attribution analysis framework includes the random forest model, the partial least squares structural equation model, and the redundancy analysis model; among them, the random forest model is used for variable selection, the partial least squares structural equation model is used for path identification, and the redundancy analysis model is used for spatial attribution.

[0020] Furthermore, variable selection is performed based on the random forest model, including: calculating the relative importance of each driving factor to sediment concentration changes based on the random forest model, and selecting core driving factors to form a subset of variables.

[0021] Furthermore, path identification is performed based on the partial least squares structural equation model, including: using a subset of variables output by the random forest model as input for analysis to identify direct causal paths and indirect causal paths.

[0022] Furthermore, spatial attribution is performed based on a redundancy analysis model, including: interpreting the spatial distribution of sediment variation based on the redundancy analysis model, extracting the explanatory power of each driving factor for sediment variation in different regions and forming a unified weighted expression to quantify the attribution contribution of each driving factor in different spaces.

[0023] The second aspect of this invention provides an attribution and identification system for sediment flux evolution in high-altitude and cold mountainous areas.

[0024] An attribution and identification system for sediment flux evolution in high-altitude cold mountainous areas includes:

[0025] The main driving factor identification module is configured to identify the main driving factors based on the sediment transport mechanism in high-altitude and cold mountainous areas.

[0026] The preprocessing module is configured to: acquire the basic data corresponding to the main driving factors, preprocess them, and construct the driving factor dataset;

[0027] The mutation point detection module is configured to: use a joint mutation identification method to detect mutation points from multiple angles based on the established driving factor dataset; use the Pettitt test to identify the grade changes under distribution mutations to determine the year of distribution mutation; use the Mann-Kendall trend test to identify the slope changes under trend mutations to determine the year of trend mutation; and detect mutation points based on the difference between the obtained distribution mutation years and trend mutation years, and divide the evolution stages.

[0028] The attribution identification module is configured to perform attribution analysis on all driving factors based on a three-stage progressive attribution analysis framework, in order to quantify the attribution contribution of each driving factor at different evolution stages and in different spaces.

[0029] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of the attribution identification method for sediment flux evolution in high-altitude and cold mountainous areas as described in the first aspect of the present invention.

[0030] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the attribution identification method for sediment flux evolution in high-altitude cold mountainous areas as described in the first aspect of the present invention.

[0031] The above one or more technical solutions have the following beneficial effects:

[0032] (1) This invention integrates three methods and models: random forest, partial least squares structural equation model (PLS-SEM) and redundancy analysis (RDA) to construct a three-stage progressive attribution analysis framework to comprehensively analyze the interaction between various driving factors. It can accurately identify the different degrees of influence of each driving factor on sediment evolution and make up for the shortcomings of existing technologies in analyzing the interaction of driving factors.

[0033] (2) This invention employs a joint mutation identification method for multi-angle mutation point detection, namely: using the Pettitt test to identify grade changes under distribution mutations to determine the year of distribution mutation; using the Mann-Kendall trend test to identify slope changes under trend mutations to determine the year of trend mutation; and detecting mutation points and dividing evolution stages based on the difference between the year of distribution mutation and the year of trend mutation. Therefore, compared with the prior art, this invention can effectively divide evolution stages, providing more reliable basic data for subsequent attribution analysis.

[0034] (3) This invention provides a comprehensive analytical framework by combining three different analytical methods: random forest, partial least squares and redundancy analysis. It can not only effectively evaluate the independent influence of each driving factor, but also reveal the interrelationship and interaction effect between the driving factors, thereby improving the accuracy and stability of attribution analysis.

[0035] (4) By proposing a multi-model analysis framework (attribution analysis framework), this invention solves the defects of existing technologies in the analysis of synergistic effects of multiple driving factors, identification of mutation points, and stability of attribution results. It effectively improves the accuracy, stability and physical consistency of attribution identification of sediment evolution process in high-altitude and cold mountainous areas, and provides a scientific basis for soil and water conservation and regional ecological restoration.

[0036] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0037] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0038] Figure 1 This is a flowchart of a method for attribution identification of sediment flux evolution in high-altitude cold mountainous areas according to Embodiment 1 of the present invention. Detailed Implementation

[0039] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0040] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0041] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0042] This invention provides an attribution identification method for sediment flux evolution in high-altitude cold mountainous areas. The overall approach of this method can be summarized as follows: S1. Analyze the sediment transport mechanism in high-altitude cold mountainous areas and identify the main driving factors, which generally include precipitation, snowmelt, temperature, vegetation index, land use, etc.; S2. Collect and preprocess the basic data of driving factors to construct a complete dataset; S3. Identify abrupt change points in the sediment transport process and divide the evolution stages; S4. Construct an attribution framework that integrates random forest, PLS-SEM, and redundancy analysis to assess the importance, path dependence, and spatial pattern of driving factors; S5. Combine the results of the three methods to quantitatively identify the degree of influence and action path of driving factors on sediment evolution.

[0043] Example 1

[0044] This embodiment discloses a method for attribution and identification of sediment flux evolution in high-altitude and cold mountainous areas.

[0045] like Figure 1 As shown, a method for attribution and identification of sediment flux evolution in high-altitude and cold mountainous areas includes:

[0046] Step S1: Identify the main driving factors based on the sediment transport mechanism in high-altitude and cold mountainous areas;

[0047] Step S2: Obtain the basic data corresponding to the main driving factors, perform preprocessing, and construct the driving factor dataset;

[0048] Step S3: Using a joint mutation identification method, multi-angle mutation point detection is performed based on the established driving factor dataset: the Pettitt test is used to identify the grade change under the distribution mutation to determine the year of the distribution mutation; the Mann-Kendall trend test is used to identify the slope change under the trend mutation to determine the year of the trend mutation; mutation points are detected based on the difference between the obtained distribution mutation years and trend mutation years, and the evolution stages are divided.

[0049] Step S4: Perform attribution analysis on all driving factors based on the three-stage progressive attribution analysis framework to quantify the attribution contribution of each driving factor in different evolution stages and different spaces.

[0050] Based on the above process, this invention establishes a sediment attribution identification framework for high-altitude and cold mountainous areas by integrating multivariate statistical learning and path modeling techniques. This improves the stability and accuracy of the attribution results and can provide decision support for soil and water conservation and regional ecological restoration. To facilitate understanding of the technical solution of this invention, the specific implementation methods are further explained and described below.

[0051] In step S1, the main driving factors are identified based on the sediment transport mechanism in high-altitude and cold mountainous areas.

[0052] First, an in-depth analysis of the sediment transport mechanism in high-altitude plateau regions is conducted to identify the main driving factors that may affect sediment evolution, including both natural and anthropogenic factors, such as precipitation, snowmelt, temperature, vegetation index (NDVI), land use change, and disturbances from water conservancy projects. It should be noted that identifying the main driving factors based on the sediment transport mechanism in high-altitude mountainous areas is not the primary technical point of this invention. Therefore, in this embodiment, the main driving factors are determined by integrating the identification results of existing technologies based on the sediment transport mechanism in high-altitude mountainous areas.

[0053] In step S2, the basic data corresponding to the main driving factors are obtained and preprocessed to construct the driving factor dataset.

[0054] Based on the identified main driving factors, corresponding basic data are collected, such as meteorological, surface condition, topographic, and hydrological data. These data undergo preprocessing, including missing value imputation, standardization, and normalization, to ensure data quality and construct a complete driving factor dataset.

[0055] In step S3, a joint mutation identification method is used to detect mutation points from multiple angles based on the established driving factor dataset.

[0056] By combining the Pettitt test and the Mann-Kendall trend test (MK), multi-angle abrupt change points can be detected in sediment concentration time series (SSC). This can be achieved through the following methods:

[0057] 1) The Pettitt test (suitable for detecting abrupt changes in the mean or "structural jumps" in time series) is used to identify changes in the range under abrupt changes in distribution, in order to determine the year of the abrupt change in distribution, i.e.:

[0058] ;

[0059] In the formula, Indicates the preceding Year and after The difference between the rank sums of annual observations This represents the statistical significance of a mutation. A significant mutation is considered to have occurred when P < 0.05; the P-value is obtained by looking up a table or calculation, and is calculated under the assumption of no mutation. The probability; when P < 0.05, it indicates that the mutation is significant, and it is considered that the time series has undergone a non-random structural jump.

[0060] Based on this, the years of abrupt changes in distribution can be identified, i.e.:

[0061] ;

[0062] in, Indicates the year of a sudden change in distribution. Indicates that The largest time point .

[0063] 2) The Mann-Kendall trend test (suitable for identifying abrupt changes in long-term trends, such as a shift from increase to decrease) is used to identify the slope change under the trend abrupt change in order to determine the year of the trend abrupt change.

[0064] ;

[0065] in, Represents the first in a time series Annual observations (e.g., sediment concentration SSC at a certain station); Represents the first in a time series Observations for the year ( ); This is a trend statistic used to represent the direction of monotonic change throughout the time series. This indicates that the overall sequence shows an upward trend; if This indicates that the overall sequence shows a downward trend. The sign function, used to determine the direction of a trend, is defined as follows:

[0066] .

[0067] The Mann-Kendall trend test generates a forward-looking (UF) and backward-looking (UB) curve. The intersection of the UF and UB curves represents the year of the trend abrupt change, which can be used to help determine the "direction" of the change (e.g., increase). Subtract or reduce increase).

[0068] 3) Detect mutation points based on the difference between the years of distribution mutation and the years of trend mutation, and divide the evolution stages.

[0069] If the years of mutation identified by the Pettitt test and the MK test are the same or similar (difference) If the difference between the distributional mutation year and the trend mutation year is greater than 3 years, the Pettt test result (applicable to jump mutations) is preferred, and MK is used as an auxiliary interpretation of the trend direction.

[0070] Additionally, if MK detects multiple trend change segments, then the significance level of Pettitt (i.e., significance statistic) is used to help distinguish between the primary abrupt change segment and the secondary stage. Specifically:

[0071] First, obtain the set of trend inflection points detected by Mann-Kendall. ;in, Represents the first in the set of trend inflection points There are several trend inflection points. Then, for each trend inflection point, its relationship with the Petittt inflection point is calculated. The difference, i.e. Based on this, if a certain trend abruptly changes... satisfy ,and Significant (i.e.) If a point is identified as a primary mutation point, then that point is the primary mutation point; other mutation points serve as secondary stage division nodes, dividing the stage segments according to time sequence (e.g., P1, P2, P3).

[0072] Furthermore, the stage division rule of the present invention is as follows: the sediment evolution process is divided into different time periods (such as P1: 1981–2000, P2: 2001–2020) with the main mutation point as the boundary; the sediment trend remains relatively consistent within each stage.

[0073] This invention uses methods such as the Pettt test and the Mann-Kendall test to detect abrupt change points in sediment concentration (SSC) time series, identify key change nodes in the sediment evolution process, and divide the evolution into different stages, which can provide an accurate time frame for subsequent attribution analysis.

[0074] In step S4, attribution analysis is performed on all driving factors based on a three-stage progressive attribution analysis framework to quantify the attribution contribution of each driving factor at different evolution stages and in different spaces.

[0075] To systematically identify the driving mechanisms of sediment flux changes in high-altitude and cold mountainous areas, this invention constructs a three-stage progressive fusion modeling method, sequentially integrating Random Forest (RF), Partial Least Squares Structural Equation Modeling (PLS-SEM), and Redundancy Analysis (RDA). These three methods are used for variable selection, path identification, and spatial attribution, respectively, ultimately achieving a unified expression output. Specifically, this can be implemented through the following methods:

[0076] 1) Variable selection based on random forest model.

[0077] A random forest model was used to train all driving factors, and their relative importance to changes in sediment concentration (SSC) was calculated, i.e.:

[0078] ;

[0079] in, Indicates the first Several driving factors, such as temperature (T), precipitation (P), NDVI, snowmelt, land use index, etc. Representing variables in the RF model The contribution score to the target variable. From this, the top [values] with the highest importance can be output. These variables constitute a subset of variables. The next step is path modeling and analysis. In this invention, a Random Forest (RF) model is used to train all candidate driving factors to identify the factors that have the most significant impact on sediment concentration (SSC) changes. This algorithm belongs to the nonparametric supervised classification / regression algorithm in ensemble learning and has advantages such as strong ability to model nonlinear relationships and stable identification of feature importance.

[0080] 2) Identifying direct and indirect causal paths based on partial least squares structural equation modeling, i.e.:

[0081] ;

[0082] in, This represents mediating variables (such as MDVI, snowmelt amount, and other hydrological factors). express For the direct path coefficients of SSC, Indicates the mediation path Path coefficients for SSC; This indicates the error term. and The model was automatically fitted from the PLS-SEM model, and the modeling process is as follows:

[0083] First, construct the structural path model graph (T, P). NDVI SSC) is then imported into tools such as SmartPLS for modeling; subsequently, the regression coefficients and statistical significance (e.g., p-value) of each path are output. Based on this, the total effect is... ( (For intermediate path coefficients).

[0084] 3) Spatial attribution based on redundancy analysis model.

[0085] The key variables identified by PLS-SEM are input into the redundancy analysis model to interpret the spatial distribution of sediment variation. The explanatory power (contribution rate) of each factor to sediment variation in different regions is extracted and a unified weighted expression is formed.

[0086] The Redundancy Analysis (RDA) model can be expressed as:

[0087] ;

[0088] in, This represents the sediment flux response variables for each region (such as the SSC of multiple watersheds or sub-regions). The real number field indicates that the variable takes the value of a continuous real number. The number of response variables typically corresponds to multiple sub-basins or cross-sections in space (e.g., SSC_1, SSC_2, ..., SSC_p). This represents the key driver matrix confirmed by RF and PLS. This represents the contribution coefficient of the variable to each response item (similar to the coefficient of multiple regression). This represents the residual matrix.

[0089] Based on this, the multi-factor fusion expression (final output) can be expressed as:

[0090] ;

[0091] ;

[0092] in, Indicates the final attribution contribution. 、 and Indicates the weighting coefficients (default equal weights or set via cross-validation); This indicates the load value of the variable on the principal axis; This represents the fitted sediment concentration or flux value calculated by the model, with units consistent with the original SSC. This represents the path coefficient of the variable in PLS-SEM. Furthermore, in redundancy analysis (RDA), the principal axis load value represents the explanatory power or projected weight of each variable along the principal component direction. It is a variable The load value on the first principal axis, specifically: the larger the value (positive or negative), the stronger its explanatory power for the spatial variation of sediment; this value is usually derived from the ordination analysis output of RDA (such as a biplot). In this invention, using... As a weighting factor, it is incorporated into the multi-factor fusion expression to generate the final attribution contribution weight.

[0093] This three-step fusion framework integrates variable selection (RF), causal identification (PLS-SEM), and spatial interpretation (RDA) in an orderly manner, outputting an interpretable expression similar to multiple linear regression, and quantifying the attribution contribution of each driving factor at different stages and in different spatial units, which greatly improves the interpretability, stability, and practicality of the model.

[0094] Through the above technical solution, this invention has for the first time achieved a multi-dimensional and comprehensive attribution analysis of sediment evolution in high-altitude and cold mountainous areas, which greatly improves the stability and accuracy of the analysis results and provides a scientific basis for the formulation of strategies for soil and water conservation, sediment regulation and regional ecological restoration.

[0095] Example 2

[0096] This embodiment discloses an attribution and identification system for sediment flux evolution in high-altitude and cold mountainous areas.

[0097] An attribution and identification system for sediment flux evolution in high-altitude cold mountainous areas includes:

[0098] The main driving factor identification module is configured to identify the main driving factors based on the sediment transport mechanism in high-altitude and cold mountainous areas.

[0099] The preprocessing module is configured to: acquire the basic data corresponding to the main driving factors, preprocess them, and construct the driving factor dataset;

[0100] The mutation point detection module is configured to: use a joint mutation identification method to detect mutation points from multiple angles based on the established driving factor dataset; use the Pettitt test to identify the grade changes under distribution mutations to determine the year of distribution mutation; use the Mann-Kendall trend test to identify the slope changes under trend mutations to determine the year of trend mutation; and detect mutation points based on the difference between the obtained distribution mutation years and trend mutation years, and divide the evolution stages.

[0101] The attribution identification module is configured to perform attribution analysis on all driving factors based on a three-stage progressive attribution analysis framework, in order to quantify the attribution contribution of each driving factor at different evolution stages and in different spaces.

[0102] Example 3

[0103] The purpose of this embodiment is to provide a computer-readable storage medium.

[0104] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the attribution identification method for sediment flux evolution in high-altitude and cold mountainous areas as described in Embodiment 1 of this disclosure.

[0105] Example 4

[0106] The purpose of this embodiment is to provide an electronic device.

[0107] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the attribution identification method for sediment flux evolution in high-altitude cold mountainous areas as described in Embodiment 1 of this disclosure.

[0108] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0109] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0110] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for attributing and identifying the evolution of sediment flux in high-altitude and cold mountainous areas, characterized in that, include: Based on the sediment transport mechanism in high-altitude and cold mountainous areas, the main driving factors were identified; Obtain the basic data corresponding to the main driving factors, preprocess them, and construct the driving factor dataset; A joint mutation identification method was adopted to detect mutation points from multiple perspectives based on the established driving factor dataset: the Pettitt test was used to identify the grade changes under the distribution mutation to determine the year of the distribution mutation; the Mann-Kendall trend test was used to identify the slope changes under the trend mutation to determine the year of the trend mutation; mutation points were detected based on the difference between the obtained distribution mutation years and trend mutation years, and the evolution stages were divided. Attribution analysis is performed on all driving factors based on a three-stage progressive attribution analysis framework to quantify the attribution contribution of each driving factor at different stages of evolution and in different spaces. The attribution analysis framework includes a random forest model, a partial least squares structural equation model, and a redundancy analysis model; wherein, the random forest model is used for variable selection, the partial least squares structural equation model is used for path identification, and the redundancy analysis model is used for spatial attribution.

2. The method for attribution and identification of sediment flux evolution in high-altitude cold mountainous areas as described in claim 1, characterized in that, The mutation point is detected based on the difference between the year of the distribution mutation and the year of the trend mutation, and the evolution stages are divided, including: if the year of the distribution mutation is the same as or within 3 years of the year of the trend mutation, the determined year of the distribution mutation and the year of the trend mutation are defined as stable mutation points; if the difference between the year of the distribution mutation and the year of the trend mutation is greater than 3 years, the test result of the Pettitt test is used.

3. The method for attribution and identification of sediment flux evolution in high-altitude cold mountainous areas as described in claim 2, characterized in that, The method detects mutation points based on the difference between the years of distribution mutation and the years of trend mutation, and divides the evolution stages. It also includes: if multiple trend change segments are detected by the Mann-Kendall trend test, the main mutation segment and the secondary stage of the evolution stage are divided according to the significance statistic of the Pettitt test.

4. The method for attribution and identification of sediment flux evolution in high-altitude cold mountainous areas as described in claim 1, characterized in that, The variable selection based on the random forest model includes: calculating the relative importance of each driving factor to the change in sediment concentration based on the random forest model, and selecting the core driving factors to form a subset of variables.

5. The method for attribution and identification of sediment flux evolution in high-altitude cold mountainous areas as described in claim 1, characterized in that, Path identification based on partial least squares structural equation modeling includes: analyzing a subset of variables output by a random forest model as input to identify direct causal paths and indirect causal paths.

6. The method for attribution and identification of sediment flux evolution in high-altitude cold mountainous areas as described in claim 1, characterized in that, Spatial attribution based on redundancy analysis models includes: interpreting the spatial distribution of sediment variation based on redundancy analysis models, extracting the explanatory power of each driving factor for sediment variation in different regions and forming a unified weighted expression to quantify the attribution contribution of each driving factor in different spaces.

7. A system for attributing and identifying the evolution of sediment flux in high-altitude and cold mountainous areas, characterized in that, include: The main driving factor identification module is configured to identify the main driving factors based on the sediment transport mechanism in high-altitude and cold mountainous areas. The preprocessing module is configured to: acquire the basic data corresponding to the main driving factors, preprocess them, and construct the driving factor dataset; The mutation point detection module is configured to: use a joint mutation identification method to detect mutation points from multiple angles based on the established driving factor dataset; use the Pettitt test to identify the grade changes under distribution mutations to determine the year of distribution mutation; use the Mann-Kendall trend test to identify the slope changes under trend mutations to determine the year of trend mutation; and detect mutation points based on the difference between the obtained distribution mutation years and trend mutation years, and divide the evolution stages. The attribution identification module is configured to perform attribution analysis on all driving factors based on a three-stage progressive attribution analysis framework to quantify the attribution contribution of each driving factor at different stages of evolution and in different spaces. The attribution analysis framework includes a random forest model, a partial least squares structural equation model, and a redundancy analysis model; wherein, the random forest model is used for variable selection, the partial least squares structural equation model is used for path identification, and the redundancy analysis model is used for spatial attribution.

8. A computer-readable storage medium having a program stored thereon, characterized in that, When executed by the processor, the program implements the steps in the attribution identification method for sediment flux evolution in high-altitude and cold mountainous areas as described in any one of claims 1-6.

9. An electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the attribution identification method for sediment flux evolution in high-altitude cold mountainous areas as described in any one of claims 1-6.

Images