Ecological restorability calculation method based on time sequence

Abstract

The invention provides an ecological restoring force calculation method based on a time sequence, which is based on a system multistable state and critical deceleration phenomenon theory, carries out ecological system restoring force calculation and variation trend analysis by using time sequence data, and comprises the following steps: reading normalized difference vegetation index (NDVI) time sequence data of vegetation, generating a time sequence matrix, and calculating the ecological system restoring force according to the time sequence matrix; decomposing the NDVI time series data through an STL decomposition method to obtain a stable time series matrix of an NDVI residual error, and calculating a lagging autocorrelation coefficient by using a sliding window method; calculating a restoring force index by using a hysteresis autocorrelation coefficient change mode; and performing variation trend analysis on the restoring force index by using a trend analysis method, and checking the significance of the variation trend. And evaluating the ecological system restoring force and the state change potential according to the restoring force index change trend.

Description

Technical Field

[0001] The present invention relates to the field of ecological assessment, and in particular to a method for calculating ecosystem resilience and determining a change trend by using vegetation DNVI time series data based on system multi-stable states and critical deceleration theory. Background Art

[0002] Ecosystem resilience is the ability of an ecosystem to restore its original structure, function and state after natural or human disturbance. The concept was first proposed by C.S.Holling in 1973. It was later refined into engineering resilience (describing the speed at which a system recovers to a single equilibrium state) and ecological resilience (describing the ability of an ecosystem to resist disturbance before it undergoes disturbance and transforms to other stable states). With the deepening of research, a variety of theories and models have been formed to evaluate ecosystem resilience, such as the "cup and ball" model, the adaptation cycle model, the stability domain model, multi-stability, state transition threshold, attraction domain scaling, critical deceleration and turning point theory, as well as "self-organization" and dynamic system theory. The most important thing is that the rate at which an ecosystem recovers from disturbance near the critical point of change is significantly slowed down, the variance increases, and the autocorrelation of system fluctuations is enhanced. On this basis, a variety of indicators of changes in ecosystem structure and attribute characteristics are proposed as early warning signals of ecosystem state transitions to predict changes in ecosystem state. The early warning of system state changes characterized by changes in ecosystem resilience is of great significance to ecosystem management and protection, involving many aspects such as nature conservation, ecological restoration, and the formulation of ecosystem management strategies to cope with climate change and human disturbance. Despite the progress made, the assessment and application of theories related to ecosystem resilience still face many challenges, such as the identification of effective indicators and data acquisition, the interaction of multiple factors and the uncertainty of evaluation results, and the understanding of assessment and cross-scale effects.

[0003] At present, there are two main methods for studying changes in ecosystem resilience. The first is to quantify the critical deceleration phenomenon based on observational data and judge changes in the resilience of the system. The second is to estimate the transition probability of the ecosystem between different alternative states in the form of statistical probability based on long-term series data, and judge whether it is close to the critical point and the risk of state transition. Both methods contain a variety of specific indicators, including recovery rate / recovery time, lag correlation, variance / coefficient of variation, etc. for critical deceleration, kurtosis, spatial skewness, threshold autoregression, etc. for wandering effect.

[0004] In specific applications, methods such as multi-index comprehensive evaluation, time series analysis, and disturbance identification are involved. Among them, the multi-index comprehensive evaluation method mainly uses the entropy weight method and the hierarchical analysis method to integrate multiple related indicators to form a comprehensive indicator system; disturbance identification is based on time series or data that can reflect the natural fluctuations and patterns of the system, and evaluates the system's recovery capacity by identifying ecological disturbances and their impact on the ecosystem. The evaluation of regional ecological resilience mainly relies on a multi-index comprehensive evaluation method, which constructs an indicator system from different perspectives (such as biodiversity, soil quality, water resource recovery, etc.) and evaluates the state of the ecosystem from multiple dimensions.

[0005] In general, the two main evaluation methods currently have their own shortcomings:

[0006] Multi-indicator comprehensive evaluation method: Data collection is cumbersome, weight setting is subjective, comparability between different regions is poor, and there is little exploration of changes in resilience.

[0007] Disturbance analysis method: It is difficult to conduct long-term and continuous monitoring of the temporal changes of the restoring force, and the method is sensitive to the intensity of the disturbance, which affects the objectivity of the evaluation results. Summary of the invention

[0008] The technical problem to be solved by the present invention is to provide a method for calculating ecological resilience based on time series to quantitatively evaluate the characteristics and trends of regional ecological resilience changes in view of the above-mentioned defects in the prior art. This method not only evaluates resilience, but also monitors its changes over time, thereby providing a more accurate and objective spatiotemporal dynamic analysis of changes in regional ecosystem resilience.

[0009] According to the present invention, a method for calculating ecological resilience based on time series is provided, comprising:

[0010] Read the normalized difference vegetation index NDVI time series data, generate a time series matrix, decompose the NDVI time series data using the STL decomposition (Seasonal and Trend decomposition using Loess) method, and obtain the NDVI residual time series matrix;

[0011] For the generated NDVI residual time series matrix, the sliding window method is used to calculate the lagged autocorrelation coefficient; the lagged autocorrelation coefficient is used to calculate the resilience index;

[0012] Perform ecological resilience trend analysis using resilience indicators.

[0013] Preferably, the residual sequence obtained by decomposing the DNVI time series is used as the time series data basis for calculating the resilience index, and the sliding window method is used to calculate the lagged autocorrelation coefficient.

[0014] Preferably, during data processing, the lagged autocorrelation coefficient is calculated for each pixel of the time series matrix using a three-year sliding window as follows:

[0015]

[0016] Where ACF is the lagged autocorrelation coefficient, Xt is the observed value in the time series, is the mean of the observations, N is the total number of observations, and k = 1 indicates a lag of 1 period.

[0017] Preferably, when calculating the index, the multi-year mean of the time series of the lagged autocorrelation coefficient of each pixel is calculated by using the lagged autocorrelation coefficient, and then the negative value of the multi-year mean is taken as the resilience index.

[0018] Preferably, the resilience index is calculated as follows:

[0019]

[0020] Among them, R is the resilience index of the pixel, Represents the average of the lagged autocorrelation coefficients of the region.

[0021] Preferably, the least squares method and significance test are used in the trend analysis to analyze the changing trend of the resilience index over time.

[0022] Preferably, during the trend analysis, a trend that passes the significance test is determined as a decrease or increase in resilience, and a trend that fails the significance test is determined as a trend with no significant change.

[0023] Preferably, the least square method finds the best fitting line for the data by minimizing the square of the error and the residual sum of squares; wherein the straight line function y=mx+b is determined so that the sum of the vertical distances of all observed data points to the straight line is minimized.

[0024] Preferably, in trend analysis, least squares method, the residual sum of squares is expressed as:

[0025] RSS=∑(y i -(mx i +b)) 2

[0026] Where n is the number of data points, xi and yi are the independent and dependent variable values ​​of the ith data point, and are the time series number and the negative value of the lagged autocorrelation, respectively;

[0027] The calculation formula of slope m is:

[0028]

[0029] The calculation formula for intercept b is:

[0030]

[0031] Preferably, a t-statistic is used to perform a significance test to determine whether the slope m in the model is statistically significant, and the formula is as follows:

[0032]

[0033] Where m is the slope and SE is the standard error of the slope.

[0034] Preferably, when using the least squares method and t-statistic test to analyze the changes in ecosystem resilience indicators over time, the judgment criteria are as follows: if the t-value is less than 0.05 and the slope is positive, the ecological resilience is judged to be significantly improved; if the t-value is less than 0.05 and the slope m is negative, the ecological resilience is judged to be significantly reduced over time; if the t-value of the slope m fails the significance test, that is, the t-value is greater than or equal to 0.05, the change is judged to be not statistically significant. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] A more complete understanding of the present invention and its attendant advantages and features will be more readily appreciated by reference to the following detailed description taken in conjunction with the accompanying drawings, in which:

[0036] Figure 1 The overall flow chart of the ecological resilience calculation method based on time series according to a preferred embodiment of the present invention is schematically shown.

[0037] It should be noted that the drawings are used to illustrate the present invention, rather than to limit the present invention. Note that the drawings showing the structures may not be drawn to scale. In addition, in the drawings, the same or similar elements are marked with the same or similar reference numerals. DETAILED DESCRIPTION

[0038] In order to make the contents of the present invention clearer and easier to understand, the contents of the present invention are described in detail below in conjunction with specific embodiments and drawings.

[0039] The inventors of the present invention have made full use of the existing methods and indicators related to early warning of ecosystem state transitions, and clarified that the autocorrelation coefficient measures the degree of correlation between a time series and its own data at a certain lag period, which can reflect the changes in ecosystem resilience and warn of ecosystem state transitions; in the context of ecosystems accepting disturbances, if ecosystem fluctuations show a high degree of time correlation, it indicates that the rate at which the system recovers to its original state will slow down, which is consistent with the theory that when the system approaches the critical point, the speed of recovery from interference slows down significantly, indicating that the system's resilience is reduced and the possibility of transition is high; on the contrary, low autocorrelation indicates that the ecosystem is far from the critical point, the resilience changes little, and the system is highly stable. The lagged autocorrelation coefficient provides an effective means to quantitatively warn of changes in the resilience of ecosystems. The risk of critical transitions in regional ecosystem states can be judged by analyzing changes and trends.

[0040] Specifically, Figure 1 The overall flow chart of the ecological resilience calculation method based on time series according to a preferred embodiment of the present invention is schematically shown.

[0041] like Figure 1 As shown, the ecological resilience calculation method based on time series according to a preferred embodiment of the present invention includes:

[0042] Step S1: Read the normalized difference vegetation index NDVI time series data of vegetation and generate a time series matrix.

[0043] Specifically, for example, the data acquisition step reads the Normalized Difference Vegetation Index (NDVI) data of the MODIS satellite. MODIS NDVI provides global vegetation coverage monitoring with high temporal resolution (8 days to 1 month) and long time span (more than 5 years).

[0044] Preferably, the data acquisition step also performs data preprocessing on the time series matrix, including pixel alignment, clipping according to geographic boundaries, and data format conversion, thereby making the data suitable for further analysis.

[0045] Step S2: For the read NDVI time series data, a detrended stable time series is formed.

[0046] Specifically, the STL (Seasonal and Trend decomposition using Loess) decomposition method is used to decompose the NDVI time series data into three parts: trend, season, and residual. The residual part is the change in the NDVI series without the influence of trend and seasonal changes, which can reduce the random fluctuations caused by seasonal and trend effects in the time series and form a stable time series. The residual time series of NDVI is used to generate a stable time series matrix;

[0047] Step S3: For the generated NDVI residual time series matrix, the lagged autocorrelation coefficient is calculated using the sliding window method;

[0048] Specifically, for example, the residual sequence obtained by decomposing the DNVI time series is used as the time series data basis for calculating the resilience index, and the sliding window method is used to calculate the lagged autocorrelation coefficient;

[0049] The lagged autocorrelation coefficient is calculated for each pixel of the time series matrix using a three-year sliding window as follows:

[0050]

[0051] Where ACF is the lagged autocorrelation coefficient, Xt is the observed value in the time series, is the mean of the observations, N is the total number of observations, and k = 1 indicates a lag of 1 period.

[0052] Step S4: Calculate the resilience index using the lagged autocorrelation coefficient of the residual;

[0053] Specifically, in the indicator calculation step, the multi-year mean of the lagged autocorrelation coefficient time series of each pixel is calculated by using the lagged autocorrelation coefficient, and then the negative value of the multi-year mean is taken as the resilience indicator corresponding to this time period. The resilience indicator reflects the difference between the system and the stable state. The resilience indicator can be used to judge the trend of ecological resilience changes, and the resilience indicator can also be used to judge the significance of the trend.

[0054] Specifically, the resilience index is calculated as follows:

[0055]

[0056] Among them, R is the resilience index of the pixel, Represents the average of the lagged autocorrelation coefficients of the region.

[0057] Step S5: Use the resilience index to perform ecological resilience change trend analysis and test the significance of the change trend;

[0058] Preferably, the trend analysis step uses the least square method and significance test to analyze the change trend of the resilience index over time. For example, a trend that passes the significance test is determined as a decrease or increase in resilience, and a trend that fails the significance test is determined as no significant change trend.

[0059] Specifically, the least squares method finds the best fitting line for the data by minimizing the square of the error and the residual sum of squares; wherein the straight line function y=mx+b is determined so that the sum of the vertical distances of all observed data points to the straight line is minimized.

[0060] For example, in the trend analysis step, in the least squares method, the residual sum of squares is expressed as:

[0061]

[0062] Where n is the number of data points, xi and yi are the independent and dependent variable values ​​of the ith data point, and are the time series number and the negative value of the lagged autocorrelation, respectively;

[0063] To find the best estimates of the slope m and intercept b, you can use the following formulas:

[0064] The calculation formula of slope m is:

[0065]

[0066] The calculation formula for intercept b is:

[0067]

[0068] Use the t-statistic to perform a significance test to determine whether the slope m in the model is statistically significant. The formula is as follows:

[0069]

[0070] Where m is the slope and SE is the standard error of the slope.

[0071] When using the least squares method and t-statistic test to analyze the changes in ecosystem resilience indicators over time, the judgment criteria are as follows: if the t-value is less than 0.05 and the slope is positive, the ecological resilience is judged to have significantly increased; if the t-value is less than 0.05 and the slope m is negative, the ecological resilience is judged to have significantly decreased over time. If the t-value of the slope m does not pass the significance test, that is, the t-value is greater than or equal to 0.05, then the change is judged to be not statistically significant, that is, no significant change has occurred.

[0072] Preferably, step S6 may also be performed: the resilience index of each pixel is graded according to the value of the resilience index to evaluate the change in ecosystem resilience and the potential for state transition.

[0073] For example, based on the values ​​of the resilience indicators, the resilience of the ecosystem can be divided into five levels: excellent, good, medium, low, and poor.

[0074] For example, in a specific example, the grading criteria may be as shown in the following table.

[0075] Grading Grading Standards excellent x>μ+1.5σ good μ+0.5σ <x<μ+1.5σ middle μ-0.5σ <x<μ+0.5σ Low μ-1.5σ <x<μ-0.5σ Difference x>μ-1.5σ

[0076] where x is the value of the resilience index, and μ and σ are preset parameters used to determine the range.

[0077] The present invention has at least the following beneficial technical effects:

[0078] 1. Simplify operations and data acquisition: By using remote sensing data combined with data detrending processing, the effective data collection process is simplified and the operational complexity is reduced.

[0079] 2. Improve the timeliness and accuracy of evaluation: It can continuously monitor the changes in regional ecological resilience with the help of remote sensing data and provide spatiotemporal dynamic evaluation results.

[0080] 3. Enhance comparability and objectivity: Standardized methods facilitate comparison and evaluation between different regions.

[0081] In addition, it should be noted that, unless otherwise specified, the terms "first", "second", "third", etc. in the specification are only used to distinguish the various components, elements, steps, etc. in the specification, and are not used to indicate the logical relationship or sequential relationship between the various components, elements, steps, etc.

[0082] It is to be understood that, although the present invention has been disclosed as above with specific embodiments, the above embodiments are not intended to limit the present invention. For any person skilled in the art, without departing from the scope of the technical solution of the present invention, the technical contents disclosed above can be used to make many possible changes and modifications to the technical solution of the present invention, or modified into equivalent embodiments of equivalent changes. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.

Claims

1. A method for calculating ecological resilience based on time series, characterized in that include: Read the normalized difference vegetation index NDVI time series data of vegetation, generate a time series matrix, decompose the NDVI time series data by STL decomposition method, and obtain the residual time series matrix of NDVI; For the residual time series matrix of NDVI, the sliding window method is used to calculate the lagged autocorrelation coefficient of the residual, and the lagged autocorrelation coefficient of the residual is used to calculate the resilience index; The resilience indicators were used to analyze the trend of ecological resilience and test the significance of the trend.

2. The ecological resilience calculation method based on time series according to claim 1 is characterized in that: In data processing, the residual sequence obtained by decomposing the DNVI time series is used as the time series data basis for calculating the resilience index, and the sliding window method is used to calculate the lagged autocorrelation coefficient; The lagged autocorrelation coefficient is calculated for each pixel of the time series matrix using a three-year sliding window as follows: Where ACF is the lagged autocorrelation coefficient, Xt is the observed value in the time series, is the mean of the observations, N is the total number of observations, and k = 1 indicates a lag of 1 period.

3. The ecological resilience calculation method based on time series according to claim 1 or 2, characterized in that: When calculating the index, the multi-year mean of the lagged autocorrelation coefficient time series of each pixel is calculated by using the lagged autocorrelation coefficient, and then the negative value of the multi-year mean is taken as the resilience index.

4. The ecological resilience calculation method based on time series according to claim 1 or 2, characterized in that: Preferably, the resilience index is calculated as follows: Among them, R is the resilience index of the pixel, Represents the average of the lagged autocorrelation coefficients of the region.

5. The ecological resilience calculation method based on time series according to claim 1 or 2, characterized in that: In trend analysis, Mann-Kendal l trend analysis and least squares method are used to analyze the changing trend of resilience indicators over time.

6. The ecological resilience calculation method based on time series according to claim 5 is characterized in that: In the trend change significance test, the trend that passes the t-statistical characteristic significance test is judged as a decrease or increase in resilience, and the trend that fails the significance test is judged as no significant change trend.

7. The ecological resilience calculation method based on time series according to claim 5 is characterized in that: The least squares method finds the best fitting line for the data by minimizing the square of the error and the residual sum of squares; wherein the straight line function y=mx+b is determined so that the sum of the vertical distances of all observed data points to the straight line is minimized.

8. The ecological resilience calculation method based on time series according to claim 7 is characterized in that: In the least squares method, the residual sum of squares is expressed as: RSS=∑(y i -(mx i +b)) 2 Where n is the number of data points, xi and yi are the independent and dependent variable values ​​of the ith data point, and are the time series number and the negative value of the lagged autocorrelation, respectively; The calculation formula of slope m is: The calculation formula for intercept b is:

9. The method for calculating the significance test of the ecological resilience change trend based on time series according to claim 6 is characterized in that: The t-statistic is used to perform a significance test to determine whether the slope m in the model is statistically significant. The formula is as follows: Where m is the slope and SE is the standard error of the slope.

10. The ecological resilience calculation method based on time series according to claim 9 is characterized in that: When using the least squares method and t-statistic test to analyze the changes in ecosystem resilience indicators over time, the judgment criteria are as follows: if the t-value is less than 0.05 and the slope is positive, it is judged that the ecological resilience has significantly improved; If the t-value is less than 0.05 and the slope m is negative, it is judged that the ecological resilience has decreased significantly over time; If the t-value of the slope m fails the significance test, that is, the t-value is greater than or equal to 0.05, then the change is judged to be not statistically significant.

Images