Multi-source ecological factor-based comprehensive evaluation method and system for biodiversity influence

Abstract

The invention provides a biodiversity influence comprehensive evaluation method and system based on multi-source ecological factors, and relates to the technical field of biodiversity analysis, and the method comprises the steps: obtaining multi-source ecological factor data of a geological disaster control influence region, and carrying out the maximum likelihood estimation and kernel function learning processing of the multi-source ecological factor data to obtain a space-time intensity field; performing adaptive time-frequency decomposition processing based on the time-space intensity field and the time sequence of the community functional traits to obtain a multi-scale modal set; performing low-rank dynamic recognition processing on the multi-scale modal set to obtain an approximate operator and control gain pair; performing sparse regression modeling processing to obtain a nonlinear state equation set; and carrying out distribution displacement measurement processing based on the nonlinear state equation set and preset quadrat baseline distribution to obtain a comprehensive evaluation result of the biodiversity influence. According to the method, the problems of single index, rough modeling and lack of multi-scene prediction capability in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the field of biodiversity analysis technology, and in particular to a method and system for comprehensive assessment of biodiversity impacts based on multi-source ecological factors. Background Art

[0002] At present, the disturbance of regional ecosystems in geological disaster management is inevitable, and its impact on biodiversity is receiving increasing attention. Existing technologies usually rely on single or limited ecological factors in ecological impact assessment, such as changes in vegetation coverage, species population statistics, or local hydrological monitoring indicators, and mainly conduct qualitative or semi-quantitative analysis through static monitoring or empirical models. Such methods can reflect the trend of ecosystems after disturbance to a certain extent, but often ignore the dynamic coupling relationship and spatiotemporal evolution characteristics between multi-source ecological factors. In addition, traditional methods mostly use modeling methods based on experience or linear assumptions, lack the ability to accurately characterize the disturbance triggering effect, and it is difficult to reveal the complex nonlinear response process of community functional traits after the disaster.

[0003] Therefore, in complex disturbance scenarios, existing technologies often suffer from deficiencies such as insufficient precision, poor adaptability, and weak interpretability, making them unable to meet the current practical needs of ecological restoration and biodiversity conservation. A comprehensive biodiversity impact assessment method and system based on multi-source ecological factors is urgently needed to address these technical issues. Summary of the Invention

[0004] The present invention aims to provide a comprehensive biodiversity impact assessment method and system based on multiple ecological factors to address the aforementioned issues. To achieve this objective, the present invention employs the following technical solutions: First, this application provides a comprehensive biodiversity impact assessment method based on multi-source ecological factors, including: Obtain multi-source ecological factor data for the area affected by geological disaster management, including the spatiotemporal event sequence of construction and natural disturbance, the time series of community functional traits, and the hydrological observation sequence; The multi-source ecological factor data are processed by maximum likelihood estimation and kernel function learning to obtain a spatiotemporal intensity field that describes the disturbance triggering effect; Performing adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits to obtain a multi-scale modal set of post-disaster ecological responses; Performing low-rank dynamic identification processing on the multi-scale mode set to obtain an approximate operator and a control gain pair of the ecological response; Based on the multi-scale modal set, the approximate operator of the ecological response and the control gain pair, a sparse regression modeling process is performed to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios; Based on the set of nonlinear state equations and the preset sample baseline distribution, distribution displacement measurement processing is performed to obtain a comprehensive assessment result of the biodiversity impact.

[0005] Secondly, this application also provides a comprehensive biodiversity impact assessment system based on multi-source ecological factors, including: An acquisition unit is used to obtain multi-source ecological factor data in the area affected by geological disaster control, wherein the multi-source ecological factor data includes a spatiotemporal event sequence of construction and natural disturbance, a time series of community functional traits, and a hydrological observation sequence; A processing unit, configured to perform maximum likelihood estimation and kernel function learning processing on the multi-source ecological factor data to obtain a spatiotemporal intensity field that describes the disturbance triggering effect; A decomposition unit is used to perform adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits to obtain a multi-scale mode set of post-disaster ecological response; an identification unit, configured to perform low-rank dynamics identification processing on the multi-scale modal set to obtain an approximate operator and a control gain pair of the ecological response; A modeling unit is used to perform sparse regression modeling based on the multi-scale modal set, the approximate operator of the ecological response, and the control gain pair to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios; An evaluation unit is used to perform distribution displacement measurement processing based on the set of nonlinear state equations and a preset sample baseline distribution to obtain a comprehensive evaluation result of the biodiversity impact.

[0006] The beneficial effects of the present invention are: The present invention systematically constructs a nonlinear coupling relationship between the disturbance triggering effect and the community dynamic response by introducing a variety of technical means such as maximum likelihood estimation, kernel function learning, adaptive time-frequency decomposition, low-rank dynamics identification and sparse regression modeling. This method can not only extract the multi-scale response mode of the post-disaster community based on the multi-source spatiotemporal events of construction and natural disturbances, but also simulate the community evolution trajectory under different disturbance scenarios through a set of nonlinear state equations. Finally, combined with the optimal transmission model, a quantitative assessment of the impact on biodiversity is achieved, thereby overcoming the shortcomings of the existing technology of single indicators, rough modeling and lack of multi-scenario prediction capabilities, and can provide more scientific, accurate and operational decision support for ecosystem governance and biodiversity conservation.

[0007] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the embodiments of the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0009] Figure 1 A schematic flow chart of a comprehensive biodiversity impact assessment method based on multi-source ecological factors according to an embodiment of the present invention; Figure 2 Schematic diagram of the structure of the biodiversity impact comprehensive assessment system based on multi-source ecological factors described in an embodiment of the present invention.

[0010] In the figure: 701, acquisition unit; 702, processing unit; 703, decomposition unit; 704, identification unit; 705, modeling unit; 706, evaluation unit; 7021, first processing subunit; 7022, second processing subunit; 7023, third processing subunit; 7031, first decomposition subunit; 7032, second decomposition subunit; 7033, third decomposition subunit; 7041, first identification subunit; 7042, second identification subunit; 7043, third identification subunit; 7051, first modeling subunit; 7052, second modeling subunit; 7053, third modeling subunit. DETAILED DESCRIPTION

[0011] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0012] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are used only to distinguish the description and should not be understood as indicating or implying relative importance.

[0013] Example 1

[0014] This embodiment provides a comprehensive biodiversity impact assessment method based on multi-source ecological factors.

[0015] See the instructions attached Figure 1 , the figure shows that the method includes step S1, step S2, step S3, step S4, step S5 and step S6.

[0016] Step S1, obtaining multi-source ecological factor data of the geological disaster control impact area, wherein the multi-source ecological factor data includes a spatiotemporal event sequence of construction and natural disturbance, a time series of community functional traits, and a hydrological observation sequence; It can be understood that this step builds a comprehensive data foundation by integrating multi-dimensional dynamic monitoring technology. In terms of spatial dimension, high-resolution drone images (accuracy 0.1m) are used to 2 A grid-based system (likely a grid) combined with satellite remote sensing precisely locates construction disturbances (such as waste dump excavation and road construction) and natural geological hazard sites (such as landslide boundaries). RTK mapping equipment is used to georeference elevation stratification (10-meter contour lines) with the microtopography of the travertine beach dam, ensuring the spatial accuracy of spatiotemporal events. In terms of time, a network of hydrological sensors (pressure gauges and turbidity meters) is deployed to collect minute-by-minute data on runoff and sediment concentration in real time. Furthermore, seasonal dynamics of community functional traits (such as leaf nitrogen content, specific leaf area, and canopy height) are collected through fixed sampling (quarterly) and remotely sensed vegetation indices.

[0017] Step S2: performing maximum likelihood estimation and kernel function learning processing on the multi-source ecological factor data to obtain a spatiotemporal intensity field that describes the disturbance triggering effect; It can be understood that this step converts multi-source heterogeneous ecological data into a spatiotemporal intensity field that can quantify disturbance effects. Essentially, it constructs a three-dimensional probability density field describing the "probability of a disturbance event occurring, spatial location, and time" through statistical learning. In this step, step S2 includes steps S21, S22, and S23.

[0018] Step S21: The multi-source ecological factor data is processed for spatiotemporal event construction. By identifying the mutation time point for each data stream, the preset altitude layer and travertine beach bar location mutation is used as a single triggering event to obtain a spatiotemporal event sequence, where the time and position accuracy of each event is corrected using a preset drone image grid; It can be understood that this step constructs a high-precision spatiotemporal event sequence by identifying the mutation time points of multi-source data streams and constraining geographic space rules: first, the mutation time points of three types of data streams, construction disturbances (such as mechanical vibration waveforms), natural changes (such as hydrological parameter drift) and travertine beach dam morphology (based on drone images), are identified respectively. Among them, the construction disturbance adopts the dynamic threshold method to capture the amplitude surge (such as greater than 30dB), the hydrological parameters are detected by sliding window variance analysis of statistical characteristics (such as standard deviation > 3 times the baseline within 24 hours), and the travertine beach dam uses morphological contour analysis to identify the dam boundary displacement (preset displacement The system uses pre-set elevation stratification rules (e.g., 1500-1600 meters as a shrub-grassland transition zone) to forcibly merge all concurrent disturbances within this zone (e.g., a sudden increase in vibration and a sharp drop in plant area) into a single event. Furthermore, the system uses a triggering mechanism based on sudden changes in the position of travertine dams (e.g., a porosity drop >10% triggers dam instability) to correlate ecological responses. Finally, using drone image grid correction technology, the event locations are projected onto a pre-set geographic grid (e.g., 1m×1m) and the timestamps are calibrated to eliminate terrain distortion and clock asynchrony errors. This outputs event sequences with a temporal and spatial accuracy of ±0.8m and ±2 seconds. This approach successfully decouples construction blasting from aftershock disturbances during post-earthquake restoration. Furthermore, the elevation stratification rules reveal the specific propagation patterns of disturbances along the 1600m contour, providing reliable event input for the spatiotemporal intensity field.

[0019] Step S22: performing a preliminary fitting process based on maximum likelihood estimation according to the spatiotemporal event sequence to obtain preliminary parameterized kernel coefficients and spatial background fields for non-parametric kernel learning; It can be understood that in this step, by constructing a spatiotemporal point process model and implementing maximum likelihood estimation (MLE) optimization, the initial parameters required for non-parametric kernel learning (preliminary parameterized kernel coefficients and spatial background field) are output. Its mathematical model is defined as: , in, λ (x, y, t) is the space-time intensity field function, describing the disturbance trigger intensity at position (x, y) and time 𝑡, μ(x, y) is the spatial background field function, describing the basic disturbance level at position (x, y) and time t, k is the parameterized kernel function, t j is the time when the jth disturbance event occurs, x j and y jare the spatial coordinates of the j-th disturbance event, x and y are the geographical locations of the last disturbance event, j is the j-th disturbance event, and t is the time of occurrence of the last disturbance event.

[0020] Log-likelihood function: , Among them, log L is the objective function (log likelihood function), is the intensity function, describing the event at position x j， y j and time t i The instantaneous probability density of the disturbance occurs at , i represents the i-th disturbance event, n represents the total number of disturbance events, represents the triple integral of the spacetime element dxdydt, λ It is the temporal and spatial impact intensity of construction blasting.

[0021] This step uses the gradient ascent method to iteratively optimize the log-likelihood function to solve the grid distribution of the spatial background field μ(x,y), and then simultaneously fit the coefficients of the parameterized kernel function κ to obtain the initial parameters. The kernel function in this step is an exponential decay kernel, and its formula is as follows: ; in, is the kernel function, Δt is the time difference, Δr is the spatial distance, α is the basic intensity coefficient, β is the time attenuation coefficient, and γ is the spatial attenuation coefficient.

[0022] Step S23: Based on the preliminary parameterized kernel coefficients and the spatial background field, non-parametric learning and uncertainty quantification processing of the kernel function are performed, and by constructing a space-time basis function and introducing time monotonicity and non-negativity constraints, a space-time intensity field that characterizes the disturbance triggering effect is obtained.

[0023] It can be understood that this step is based on the preliminary parameterized kernel coefficients and the spatial background field. By constructing the space-time basis function and introducing the dual constraints of time monotonicity and non-negativity, high-fidelity modeling of the disturbance triggering effect is achieved: first, the kernel function is decomposed into the tensor product form of the time basis function (discrete trigonometric function describes the attenuation characteristics) and the spatial basis function (radial basis function describes the diffusion characteristics), and the negative log-likelihood function is minimized, and the time monotonicity constraint and non-negativity constraint are imposed.

[0024] Among them, the time monotonicity constraint is as follows: , Among them, k is the time basis function index, m is the space basis function index, C km represents the interaction strength between the time basis function and the space basis function, represents the speed of change of the time decay rate, Ψm represents the diffusion pattern of the disturbance effect with spatial distance (Gaussian diffusion), It means that the constraint condition holds for all time differences Δt and spatial distances Δr.

[0025] The non-negativity constraints are as follows: , Among them, C km represents the time basis function, k is the time basis function index, m is the space basis function index, The constraints are valid for all time basis functions and space basis functions.

[0026] At the same time, this step uses Bayesian posterior sampling to generate a parameter set and calculate the confidence interval of the spatiotemporal intensity field. This step reduces the error of the traditional Gaussian kernel and avoids redundant ecological compensation area.

[0027] Step S3, performing adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits to obtain a multi-scale modal set of post-disaster ecological responses; It can be understood that this step reveals the multiscale response mechanisms of post-disaster ecological disturbances through a joint analysis of the time and frequency domains. First, using the spatiotemporal intensity field as a guiding signal, a synchronously compressed wavelet transform is applied to the time series of community functional traits (such as canopy height and chlorophyll content). This involves multiresolution analysis using the complex Morlet wavelet kernel and calculating the instantaneous frequency compression energy diffusion through phase gradient. The filter bandwidth is dynamically adjusted based on the time-varying characteristics of the intensity field (for example, widening the high-frequency band to 0.1-1 Hz after a blast event). The energy entropy criterion (threshold He > 0.85) is then applied to filter dominant modes and eliminate noise components (such as random fluctuations caused by instrument errors). Finally, high-frequency oscillations that are strictly aligned with the disturbance event (such as instantaneous stomatal closure triggered by construction vibration), medium-frequency fluctuations (such as cyclical changes in biomass influenced by monsoon precipitation), and low-frequency trends (such as the direction of community succession caused by travertine degradation) are isolated, constructing a physically interpretable multiscale mode set. In this step, step S3 includes steps S31, S32, and S33.

[0028] Step S31: performing multi-resolution wavelet pre-decomposition processing on the spatiotemporal intensity field and the community functional trait time series, extracting disturbance response components of different frequency bands by continuous wavelet transform, and redistributing the time-frequency energy by synchronous compression method to obtain a preliminary time-frequency localization representation; It can be understood that in this step, a refined expression of the ecological response in the time-frequency domain is constructed through multi-resolution wavelet pre-decomposition and synchronous compression technology: first, the spatiotemporal intensity field is input as the disturbance driving signal, and a continuous wavelet transform is performed synchronously with the time series of community functional traits (such as photosynthetic rate and root biomass) - a complex Morlet wavelet basis (matching the plant physiological cycle) is used to perform multi-scale decomposition on the original signal to generate an initial time-frequency coefficient matrix, whose scale parameters are logarithmically distributed covering the 0.01-10Hz ecological response frequency band. In order to address energy diffusion defects (such as the fuzzification of the blast disturbance signal in the 2-3Hz frequency band), this step introduces synchronous compression transform (SST) to recalibrate the time-frequency plane: by calculating the instantaneous frequency of the analytical wavelet coefficients, the diffuse energy is recompressed to the true frequency ridge to generate an energy-focused time-frequency distribution. Among them, the instantaneous frequency of the wavelet coefficients is as follows: , Among them, ω f (a, b) is the local frequency characteristic of the signal at scale a and time shift b, q represents the imaginary unit, W f (a,b) are wavelet coefficients, is the symbol of partial derivative.

[0029] Step S32: performing empirical wavelet decomposition processing based on the time-frequency localization representation, separating the dominant mode of the community functional trait sequence from the background noise by adaptively constructing a filter function, and screening effective modes in combination with the energy entropy criterion to obtain a denoised multi-scale mode subset; It can be understood that this step is based on the time-frequency localization representation and adaptive empirical wavelet decomposition technology to separate the dominant modes of ecological response: first, an adaptive filter function group is constructed - the frequency band boundaries are automatically divided according to the energy distribution spectrum (for example, in the Jiuzhaigou case, the bandpass boundary is set at 2.5Hz, the main frequency of the blasting disturbance). The energy ridge inflection point is detected by the frequency band segmentation threshold (to avoid the loss of weak signals caused by artificially preset frequency bands). Then, a compactly supported orthogonal wavelet filter group is generated, where the filters are as follows: , in, is the filter function of the nth frequency band, z is the roll-off shape function, γ n is the half-width of the transition band, ω is the angular frequency variable, is the left boundary frequency of the nth frequency band, is the right boundary frequency of the nth frequency band.

[0030] This step extracts modal components by convolving the time-frequency representation with the filter bank, and then screens effective modes based on the energy entropy criterion to eliminate background noise components. The adaptive detection mechanism of the frequency band boundary in this step (whether it is greater than the preset threshold) solves the defect of traditional wavelet decomposition relying on the preset scale. Combined with the energy entropy criterion, intelligent discrimination between noise and weak ecological responses is achieved, and the effective modal capture rate is improved compared with the fixed threshold method.

[0031] Step S33: Perform Hilbert spectrum analysis on the multi-scale modal subset, characterize the dynamic response characteristics of the community at different time scales after the disturbance by calculating the instantaneous frequency and instantaneous energy curves, and obtain a multi-scale modal set of post-disaster ecological response.

[0032] It can be understood that this step is based on the denoised multi-scale modal subset, and quantifies the dynamic response characteristics of the post-disaster community through the Hilbert spectrum analysis technology. In this step, a feature extraction transformation with spatiotemporal intensity field constraints is first implemented on each mode - the instantaneous frequency and energy envelope of the oscillation mode are extracted using the Hilbert transform, wherein the instantaneous frequency is time-varyingly weighted by integrating the disturbance gradient (such as giving a high weight within the 5-minute window after the blasting event), and a spatial gradient attenuation factor is applied to the instantaneous energy to suppress random fluctuations unrelated to the disturbance (such as instrument noise). This step injects the disturbance spatiotemporal pattern into the instantaneous attribute analysis (such as giving a higher weight to the area around the blasting point), which solves the defect of insufficient representation of coupling effects by traditional methods.

[0033] Step S4: performing low-rank dynamics identification processing on the multi-scale modal set to obtain an approximate operator and control gain pair of the ecological response; It can be understood that this step extends the modal time series to a high-dimensional phase space based on the Hankel matrix, uses singular value decomposition (SVD) to extract the dominant eigenvectors, removes noise interference (retains 10-dimensional principal components with variance contribution rates > 95%), forms low-rank state variables, and then introduces the spatiotemporal intensity field as an external control input. Finally, the state transfer matrix and the control gain matrix are solved by dynamic mode decomposition (DMD), and then the ridge regression algorithm is used to suppress overfitting and ensure the sparsity of the transfer matrix and the gain matrix. In this step, step S4 includes step S41, step S42, and step S43.

[0034] Step S41: performing time extension embedding processing on the multi-scale mode set, constructing a Hankel matrix and extracting the dominant mode of the multi-scale mode set using singular value decomposition to obtain a low-rank state space representation; It can be understood that this step reconstructs the ecological dynamics phase space through time extension embedding: multi-scale modal sets (such as high-frequency modes of moss photosynthetic inhibition and low-frequency modes of cyanobacterial succession) are lagged by a preset time (determined by the first zero crossing of the autocorrelation function) to construct a Hankel state matrix, whose row vectors are formed by linear extension of each modal sampling point within a continuous time window. Then, based on singular value decomposition (SVD), noise interference is stripped away—retaining principal components with cumulative contributions greater than 95% to generate low-rank state variables that represent the essential dynamics of the system. This step can compress the original 11-dimensional noise modes into 2-dimensional state variables. This step fully preserves the disturbance-response causal mechanism, laying the foundation for control gain modeling.

[0035] Step S42: performing dynamic mode decomposition processing according to the low-rank state space representation, decomposing the state transfer matrix by the least squares method and introducing external disturbance as a control variable to obtain the Koopman operator and the initial value of the control gain of the low-rank state space representation; This step, based on a low-rank state-space representation, constructs an initial model of the ecosystem's dynamics using control-enhanced dynamic modal decomposition (CADD). First, the state variables and their time derivatives are used to construct an augmented matrix (introducing the spatiotemporal intensity field as an external control input to construct a least-squares problem). QR decomposition and Moore-Penrose pseudo-inverse are then used to solve the state transfer matrix and control gain matrix. In this step, the CAD-enhanced least-squares architecture incorporates the intensity field as an independent control term, addressing the problem of ecosystem modeling distortion.

[0036] Step S43: performing regularization optimization processing based on the approximate operator and the initial value of the control gain, suppressing noise interference and maintaining system sparsity through the ridge regression method, and obtaining the approximate operator and control gain pair of the ecological response.

[0037] It can be understood that this step targets the initial values ​​of the dynamic modal decomposition (state transition matrix and control gains) and uses a ridge regression regularization algorithm to achieve robustness upgrades in the ecological dynamics model. This process constrains the coefficients of non-critical terms to approach zero (for example, the weak coupling term between travertine porosity and insect communities is compressed to the order of 10⁻¹) while preserving core ecological mechanisms (for example, the strong control gain of blast vibration on photosynthesis). This addresses the noise sensitivity and ecological interpretability issues of the initial model.

[0038] Step S5: performing sparse regression modeling based on the multi-scale modal set, the approximate operator of the ecological response, and the control gain pair to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios; It can be understood that this step constructs a nonlinear ecological dynamics model through sparse regression modeling: first, a feature dictionary matrix is ​​constructed based on a multi-scale mode set (high-frequency modes of photosynthetic inhibition and low-frequency modes of successional trends) and control gain pairs. A minimum angle regression algorithm combined with L1 regularization (with a preset weight of 0.01) is used to screen for significant ecological process terms (for example, only retaining the interaction term between photosynthetic inhibition and blast intensity), and a Bayesian information criterion is introduced to penalize complexity. Finally, a set of nonlinear state equations is generated through ridge regression stable parameter estimation to predict community evolution trajectories under different disturbance scenarios (for example, changes in moss recovery period when blast intensity is ±20%). In this step, step S5 includes steps S51, S52, and S53.

[0039] Step S51: constructing a feature dictionary based on the multi-scale modal set and the control gain pair to obtain a feature dictionary matrix of the community state variable; It can be understood that this step first takes each component represented by the low-rank state space (such as photosynthetic inhibition intensity, cyanobacteria ratio) as the core independent variable; secondly, based on the prior knowledge of ecological mechanisms, a nonlinear function set of state variables is constructed. Then, the control gain matrix is ​​combined with the spatiotemporal intensity field to generate disturbance-state cross terms. Then, the instantaneous energy of each mode is introduced as a weight factor to construct weighted features, and finally the feature dictionary matrix is ​​integrated, whose column vectors correspond to potential ecological process terms.

[0040] Step S52: performing sparse regression modeling based on the feature dictionary matrix, the approximate operator of the ecological response, and the control gain pair, selecting significant nonlinear terms through a minimum angle regression algorithm combined with an L1 regularization constraint, and introducing the Bayesian information criterion to suppress overfitting, thereby obtaining a sparse nonlinear dynamic expression; It can be understood that this step first takes the ecological response approximation operator as the baseline target, iteratively selects the characteristic basis function along the minimum angle path through the LARS algorithm, and imposes the L1 regularization constraint to force the coefficients of non-critical items (such as the quadratic term of rainfall) to zero; then, for the candidate model set on the LARS path, the optimal complexity model is selected, and finally, ridge regression optimization (penalty coefficient of 10000) is implemented on the selected feature subset to suppress the coefficient sensitivity and generate the final nonlinear equation: , in, represents the time derivative of the state vector, Represents the filtered feature item set, is the optimized coupling coefficient, is a nonlinear characteristic function, λ is the perturbation coefficient, Λ scenario is the set of perturbation coefficients.

[0041] Step S53: Parameter optimization is performed according to the nonlinear dynamic expression, the sparse intensity is determined by cross-validation, and the stability coefficient is estimated using ridge regression to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios.

[0042] This step, understandably, divides the parameters into training and test sets. Minimizing the root mean square error (RMSE) on the test set is the goal. The L1 regularization coefficient is optimized, two core ecological mechanisms (photosynthetic inhibition and accelerated succession) are selected, and noise terms are removed. A perturbation parameter (e.g., blast intensity ±20%) is then injected to generate a set of nonlinear equations of state. This ridge regression algorithm ensures model stability even when highly coupled environmental factors are present.

[0043] Step S6: performing distribution displacement measurement processing based on the set of nonlinear state equations and the preset sample baseline distribution to obtain a comprehensive assessment result of the biodiversity impact.

[0044] It can be understood that this step uses the set of nonlinear state equations established in the previous step, and combines it with the preset sample baseline distribution to carry out distribution displacement measurement, and characterizes the difference between the ecological response distribution and the baseline distribution under the disturbance scenario through the optimal transmission theory. This process can not only capture the overall shift trend of species composition and functional traits under the influence of disturbance, but also comprehensively measure the changes in ecological stability and functional diversity at multiple scales, thereby obtaining more accurate and systematic biodiversity impact assessment results. This step breaks through the limitations of traditional static index methods and realizes the quantitative measurement of multi-scenario, dynamic and nonlinear community evolution processes, providing a scientific basis for ecological risk management and restoration strategy formulation.

[0045] Among them, the Monge-Kantorovich transport problem is constructed with the preset sample baseline distribution (undisturbed state) as the target: , in, is the distance measure between probability distributions, In order to find the minimum transmission cost among all possible coupling modes, is a random variable under disturbance state, is a random variable under the baseline state, is the integral of the coupled measure.

[0046] Then, the Sinkhorn algorithm was used to solve the Wasserstein distance, and weights (determined by the entropy weight method) were assigned to different community functional traits (photosynthetic efficiency, species richness) to obtain a comprehensive assessment index of biodiversity.

[0047] Example 2

[0048] As the instruction manual Figure 2 As shown, this embodiment provides a biodiversity impact comprehensive assessment system based on multi-source ecological factors, see Figure 2 The system includes an acquisition unit 701, a processing unit 702, a decomposition unit 703, an identification unit 704, a modeling unit 705 and an evaluation unit 706.

[0049] An acquisition unit 701 is used to acquire multi-source ecological factor data in the area affected by geological disaster control, wherein the multi-source ecological factor data includes a spatiotemporal event sequence of construction and natural disturbance, a time series of community functional traits, and a hydrological observation sequence; The processing unit 702 is configured to perform maximum likelihood estimation and kernel function learning processing on the multi-source ecological factor data to obtain a spatiotemporal intensity field that describes the disturbance triggering effect; The processing unit 702 includes a first processing sub-unit 7021, a second processing sub-unit 7022 and a third processing sub-unit 7023; The first processing sub-unit 7021 is used to construct spatiotemporal events from the multi-source ecological factor data by identifying the mutation time point for each data stream, using the preset altitude stratification and travertine beach bar location mutation as a single triggering event, to obtain a spatiotemporal event sequence, where the time and position accuracy of each event is corrected using a preset drone image grid; The second processing subunit 7022 is configured to perform a preliminary fitting process based on maximum likelihood estimation according to the spatiotemporal event sequence to obtain preliminary parameterized kernel coefficients and a spatial background field for non-parametric kernel learning; The third processing sub-unit 7023 is used to perform non-parametric learning and uncertainty quantification processing of the kernel function based on the preliminary parameterized kernel coefficients and the spatial background field, and obtain the spatiotemporal intensity field that characterizes the disturbance triggering effect by constructing a space-time basis function and introducing time monotonicity and non-negativity constraints.

[0050] A decomposition unit 703 is configured to perform adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits to obtain a multi-scale mode set of post-disaster ecological responses; The decomposition unit 703 includes a first decomposition subunit 7031, a second decomposition subunit 7032 and a third decomposition subunit 7033; The first decomposition subunit 7031 is used to perform multi-resolution wavelet pre-decomposition processing based on the spatiotemporal intensity field and the community functional trait time series, extract the disturbance response components of different frequency bands through continuous wavelet transform, and redistribute the time-frequency energy using the synchronous compression method to obtain a preliminary time-frequency localization representation; The second decomposition subunit 7032 is used to perform empirical wavelet decomposition processing based on the time-frequency localization representation, separate the dominant mode of the community functional trait sequence from the background noise by adaptively constructing a filter function, and screen the effective modes in combination with the energy entropy criterion to obtain a noise-reduced multi-scale mode subset; The third decomposition subunit 7033 is used to perform Hilbert spectrum analysis based on the multi-scale modal subset, characterize the dynamic response characteristics of the community at different time scales after the disturbance by calculating the instantaneous frequency and instantaneous energy curves, and obtain a multi-scale modal set of post-disaster ecological responses.

[0051] an identification unit 704 for performing low-rank dynamics identification processing on the multi-scale modal set to obtain an approximate operator and a control gain pair of the ecological response; The identification unit 704 includes a first identification subunit 7041, a second identification subunit 7042 and a third identification subunit 7043; The first identification subunit 7041 is configured to perform time extension embedding processing on the multi-scale mode set, extract the dominant mode of the multi-scale mode set by constructing a Hankel matrix and using singular value decomposition to obtain a low-rank state space representation; The second identification subunit 7042 is configured to perform dynamic mode decomposition processing according to the low-rank state space representation, decompose the state transfer matrix by the least squares method, and introduce external disturbances as control variables to obtain the Koopman operator and initial control gain values ​​represented by the low-rank state space; The third identification subunit 7043 is used to perform regularization optimization processing based on the approximate operator and the initial value of the control gain, suppress noise interference and maintain system sparsity through the ridge regression method, and obtain the approximate operator and control gain pair of the ecological response.

[0052] A modeling unit 705 is configured to perform sparse regression modeling based on the multi-scale modal set, the ecological response approximation operator, and the control gain pair to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios; The modeling unit 705 includes a first modeling subunit 7051, a second modeling subunit 7052 and a third modeling subunit 7053; The first modeling subunit 7051 is configured to construct a feature dictionary based on the multi-scale mode set and the control gain pair to obtain a feature dictionary matrix of the community state variables; The second modeling subunit 7052 is configured to perform sparse regression modeling based on the feature dictionary matrix, the ecological response approximation operator, and the control gain pair, select significant nonlinear terms using a minimum angle regression algorithm combined with an L1 regularization constraint, and introduce a Bayesian information criterion to suppress overfitting, thereby obtaining a sparse nonlinear dynamic expression. The third modeling subunit 7053 is used to perform parameter optimization processing according to the nonlinear dynamic expression, determine the sparse intensity through cross-validation and use ridge regression stability coefficient estimation to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios.

[0053] The evaluation unit 706 is configured to perform distribution displacement measurement processing based on the set of nonlinear state equations and a preset quadrat baseline distribution to obtain a comprehensive evaluation result of the biodiversity impact.

[0054] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0055] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

[0056] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A comprehensive biodiversity impact assessment method based on multi-source ecological factors, characterized by: include: Obtain multi-source ecological factor data for the area affected by geological disaster management, including the spatiotemporal event sequence of construction and natural disturbance, the time series of community functional traits, and the hydrological observation sequence; The multi-source ecological factor data are processed by maximum likelihood estimation and kernel function learning to obtain a spatiotemporal intensity field that describes the disturbance triggering effect; Performing adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits to obtain a multi-scale modal set of post-disaster ecological responses; Performing low-rank dynamic identification processing on the multi-scale mode set to obtain an approximate operator and a control gain pair of the ecological response; Based on the multi-scale modal set, the approximate operator of the ecological response and the control gain pair, a sparse regression modeling process is performed to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios; Based on the set of nonlinear state equations and the preset sample baseline distribution, distribution displacement measurement processing is performed to obtain a comprehensive assessment result of the biodiversity impact.

2. The biodiversity impact comprehensive assessment method based on multi-source ecological factors according to claim 1 is characterized in that , the multi-source ecological factor data is processed by maximum likelihood estimation and kernel function learning to obtain the spatiotemporal intensity field that describes the disturbance triggering effect, including: The multi-source ecological factor data is processed into spatiotemporal events. By identifying the mutation time point for each data stream, the preset altitude stratification and travertine beach bar location mutation is used as a single triggering event to obtain a spatiotemporal event sequence, where the time and position accuracy of each event is corrected using a preset drone image grid; Performing a preliminary fitting process based on maximum likelihood estimation according to the spatiotemporal event sequence to obtain preliminary parameterized kernel coefficients and a spatial background field for non-parametric kernel learning; Based on the preliminary parameterized kernel coefficients and the spatial background field, non-parametric learning and uncertainty quantification of the kernel function are performed. By constructing a space-time basis function and introducing time monotonicity and non-negativity constraints, a space-time intensity field that characterizes the disturbance triggering effect is obtained.

3. The biodiversity impact comprehensive assessment method based on multi-source ecological factors according to claim 1 is characterized in that , performing adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits, including: Multi-resolution wavelet pre-decomposition is performed on the spatiotemporal intensity field and the time series of community functional traits, and the disturbance response components of different frequency bands are extracted by continuous wavelet transform. The time-frequency energy is redistributed using the synchronous compression method to obtain a preliminary time-frequency localization representation; Performing empirical wavelet decomposition processing based on the time-frequency localization representation, separating the dominant mode of the community functional trait sequence from the background noise by adaptively constructing a filter function, and screening effective modes in combination with the energy entropy criterion to obtain a denoised multi-scale mode subset; Hilbert spectrum analysis is performed on the multi-scale modal subset, and the dynamic response characteristics of the community at different time scales after the disturbance are characterized by calculating the instantaneous frequency and instantaneous energy curves, thereby obtaining a multi-scale modal set of post-disaster ecological responses.

4. The biodiversity impact comprehensive assessment method based on multi-source ecological factors according to claim 1 is characterized in that , the multi-scale modal set is subjected to low-rank dynamic identification processing to obtain an approximate operator and control gain pair of the ecological response, including: Performing time extension embedding processing on the multi-scale mode set, extracting the dominant mode of the multi-scale mode set by constructing a Hankel matrix and using singular value decomposition to obtain a low-rank state space representation; Performing dynamic modal decomposition processing according to the low-rank state space representation, decomposing the state transfer matrix by the least squares method and introducing external disturbance as a control variable to obtain the Koopman operator and initial value of the control gain represented by the low-rank state space; Regularized optimization processing is performed based on the approximate operator and the initial value of the control gain, and the noise interference is suppressed and the sparsity of the system is maintained by the ridge regression method, so as to obtain the approximate operator and the control gain pair of the ecological response.

5. The biodiversity impact comprehensive assessment method based on multi-source ecological factors according to claim 1 is characterized in that , performing sparse regression modeling based on the multi-scale modal set, the approximate operator of the ecological response and the control gain, including: Performing feature dictionary construction processing based on the multi-scale mode set and the control gain pair to obtain a feature dictionary matrix of the community state variable; Based on the feature dictionary matrix, the approximate operator of the ecological response, and the control gain pair, sparse regression modeling is performed, significant nonlinear terms are selected through a minimum angle regression algorithm combined with an L1 regularization constraint, and the Bayesian information criterion is introduced to suppress overfitting, thereby obtaining a sparse nonlinear dynamic expression; Parameter optimization is performed according to the nonlinear dynamic expression, the sparse intensity is determined by cross-validation and the stability coefficient is estimated using ridge regression to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios.

6. A comprehensive biodiversity impact assessment system based on multi-source ecological factors, characterized by: include: An acquisition unit is used to obtain multi-source ecological factor data in the area affected by geological disaster control, wherein the multi-source ecological factor data includes a spatiotemporal event sequence of construction and natural disturbance, a time series of community functional traits, and a hydrological observation sequence; A processing unit, configured to perform maximum likelihood estimation and kernel function learning processing on the multi-source ecological factor data to obtain a spatiotemporal intensity field that describes the disturbance triggering effect; A decomposition unit is used to perform adaptive time-frequency decomposition processing based on the time series of the spatiotemporal intensity field and the community functional traits to obtain a multi-scale mode set of post-disaster ecological response; an identification unit, configured to perform low-rank dynamics identification processing on the multi-scale modal set to obtain an approximate operator and a control gain pair of the ecological response; A modeling unit is used to perform sparse regression modeling based on the multi-scale modal set, the approximate operator of the ecological response, and the control gain pair to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios; An evaluation unit is used to perform distribution displacement measurement processing based on the set of nonlinear state equations and a preset sample baseline distribution to obtain a comprehensive evaluation result of the biodiversity impact.

7. The biodiversity impact comprehensive assessment system based on multi-source ecological factors according to claim 6 is characterized in that: The processing unit includes: The first processing sub-unit is used to construct spatiotemporal events from the multi-source ecological factor data by identifying the mutation time point for each data stream, using the preset altitude stratification and travertine beach bar location mutation as a single triggering event, and obtaining a spatiotemporal event sequence, wherein the time and position accuracy of each event is corrected using a preset drone image grid; A second processing subunit is configured to perform a preliminary fitting process based on maximum likelihood estimation according to the spatiotemporal event sequence to obtain preliminary parameterized kernel coefficients and a spatial background field for non-parametric kernel learning; The third processing sub-unit is used to perform non-parametric learning and uncertainty quantification processing of the kernel function based on the preliminary parameterized kernel coefficients and the spatial background field. By constructing a space-time basis function and introducing time monotonicity and non-negativity constraints, a space-time intensity field that characterizes the disturbance triggering effect is obtained.

8. The biodiversity impact comprehensive assessment system based on multi-source ecological factors according to claim 6 is characterized in that: The decomposition unit comprises: The first decomposition subunit is used to perform multi-resolution wavelet pre-decomposition processing based on the spatiotemporal intensity field and the community functional trait time series, extract the disturbance response components of different frequency bands through continuous wavelet transform, and redistribute the time-frequency energy using the synchronous compression method to obtain a preliminary time-frequency localization representation; The second decomposition subunit is used to perform empirical wavelet decomposition processing based on the time-frequency localization representation, separate the dominant mode of the community functional trait sequence from the background noise by adaptively constructing a filter function, and screen the effective modes in combination with the energy entropy criterion to obtain a noise-reduced multi-scale mode subset; The third decomposition subunit is used to perform Hilbert spectrum analysis based on the multi-scale modal subset, characterize the dynamic response characteristics of the community at different time scales after the disturbance by calculating the instantaneous frequency and instantaneous energy curves, and obtain a multi-scale modal set of post-disaster ecological responses.

9. The biodiversity impact comprehensive assessment system based on multi-source ecological factors according to claim 6 is characterized in that: The identification unit includes: A first identification subunit is configured to perform time extension embedding processing on the multi-scale modal set, extract the dominant mode of the multi-scale modal set by constructing a Hankel matrix and using singular value decomposition to obtain a low-rank state space representation; A second identification subunit is configured to perform dynamic mode decomposition processing according to the low-rank state space representation, decompose the state transfer matrix by a least squares method, and introduce an external disturbance as a control variable to obtain a Koopman operator and an initial value of a control gain represented by the low-rank state space; The third identification subunit is used to perform regularization optimization processing based on the approximate operator and the initial value of the control gain, suppress noise interference and maintain system sparsity through the ridge regression method, and obtain the approximate operator and control gain pair of the ecological response.

10. The biodiversity impact comprehensive assessment system based on multi-source ecological factors according to claim 6, characterized in that: The modeling unit comprises: A first modeling subunit is configured to perform a feature dictionary construction process based on the multi-scale mode set and the control gain pair to obtain a feature dictionary matrix of the community state variable; A second modeling subunit is configured to perform sparse regression modeling based on the feature dictionary matrix, the approximate operator of the ecological response, and the control gain pair, select significant nonlinear terms using a minimum angle regression algorithm combined with an L1 regularization constraint, and introduce a Bayesian information criterion to suppress overfitting, thereby obtaining a sparse nonlinear dynamic expression; The third modeling subunit is used to perform parameter optimization processing according to the nonlinear dynamic expression, determine the sparse intensity through cross-validation and use ridge regression stability coefficient estimation to obtain a set of nonlinear state equations for generating community evolution trajectories under different disturbance scenarios.

Images