Land ecological value spatio-temporal data simulation system based on digital twinning

Abstract

The application relates to the technical field of digital twinning and ecological environment information simulation, in particular to a land ecological value space-time data simulation system based on digital twinning. The system comprises the following steps: acquiring environment state data with a first collection frequency and geological bottom layer data with a second collection frequency, and performing space-time alignment processing to construct a digital twinning base model containing space coordinate information; calculating the current resilience residual value of a land ecological system based on the base model; receiving a preset disturbance event sequence, and calculating the disturbance pressure cumulative value of a continuous time sequence; comparing the disturbance pressure cumulative value with the current resilience residual value, outputting a system instability early warning when the value exceeds the limit, converting the value into a value loss evaluation value according to a recovery cost conversion coefficient corresponding to a land use type, and updating the state parameters of the base model and continuing the simulation when the value does not exceed the limit; and the application realizes continuous deduction from present situation display to ecological instability early warning and value quantification.

Description

Technical Field

[0001] This invention relates to the field of digital twin and ecological environment information simulation technology, specifically a land ecological value spatiotemporal data simulation system based on digital twins. Background Technology

[0002] In the demonstration of territorial spatial planning, ecological protection and restoration, and assessment of land development intensity, how to dynamically depict the stress process of land ecosystems under future disturbances and quantitatively analyze changes in ecological status has gradually become an important technical requirement for digital ecological governance.

[0003] Currently, the analysis of land ecological status generally supports single or static processing of remote sensing images, meteorological information, land use information, or some geological monitoring information. If it is necessary to conduct a comprehensive assessment of the surface environment, the underground geological layer, and future planning disturbances at the same time, it is usually necessary to construct a spatiotemporal assessment process through data aggregation, model coupling, and manual rule analysis to obtain the ecological risk assessment results of the corresponding area.

[0004] However, when conducting land ecological simulation and value assessment using the above methods, it is often difficult to achieve unified spatiotemporal alignment of environmental state data and geological bottom layer data collected at different frequencies and spatial scales. This leads to a misalignment between visible surface changes and hidden underground conditions, making it impossible to form a stable and consistent digital foundation. Furthermore, it is difficult to accurately calculate the current buffer capacity based on the coupled processes of material cycling, biological activity, and underground recharge within the ecosystem, resulting in insufficient characterization of the true resilience of the ecosystem. At the same time, this approach typically uses static overlay or linear estimation methods to handle future disturbances such as construction, water intake, and drought, failing to reflect the cumulative amplification of disturbances in the temporal dimension and their penetrating diffusion in the spatial dimension. This results in delayed instability warnings and unclear risk propagation paths.

[0005] Furthermore, existing technologies, when outputting results, mostly focus on displaying changes in ecological status or providing qualitative risk warnings. They are unable to further convert disturbance pressures into comparable objective resource depletion results based on land use types. Consequently, they cannot provide a unified quantitative basis for comparing planning schemes, allocating restoration resources, and making project access decisions, resulting in a low level of application support capabilities. Summary of the Invention

[0006] To address the aforementioned technical problems, and considering the significant shortcomings of existing land ecological status assessment technologies in spatiotemporal alignment of multi-source heterogeneous data, accurate calculation of ecosystem buffer capacity, nonlinear characterization of future spatiotemporal evolution of disturbances, and conversion of ecological pressure into objective quantitative data on resource depletion, this invention discloses a land ecological value spatiotemporal data simulation system based on digital twins. There is an urgent need to propose a land ecological value spatiotemporal data simulation system based on digital twins that can accurately construct a unified digital twin foundation, dynamically calculate resilience residuals, nonlinearly simulate the disturbance penetration process, and achieve spatial instability early warning and objective quantitative assessment of restoration resources. Specifically, the technical solution of this invention is:

[0007] A land ecological value spatiotemporal data simulation system based on digital twins, comprising:

[0008] The data heterogeneous fusion module is used to acquire environmental status data of the first collection frequency and geological bottom layer data of the second collection frequency of the land ecosystem. The environmental status data includes remote sensing image data, meteorological data and land use type data, and the geological bottom layer data includes soil physicochemical property data, groundwater level data and microbial diversity index data.

[0009] Spatiotemporal alignment processing is performed on environmental status data and geological subsurface data to construct a digital twin base model containing spatial coordinate information;

[0010] The buffer capacity calculation module is used to calculate the current resilience residual of the land ecosystem based on the digital twin base model; where the current resilience residual represents the maximum amount of external disturbance that the land ecosystem can absorb without the occurrence of system instability.

[0011] The nonlinear penetration simulation module is used to receive a preset disturbance event sequence, calculate the disturbance pressure accumulation process of the preset disturbance event sequence within the land ecosystem based on a time series simulation algorithm, and generate a continuous time series of disturbance pressure accumulation values.

[0012] The critical warning and value inversion module is used to determine whether the cumulative value of disturbance pressure is greater than the current resilience residual value at each time point in a continuous time series.

[0013] If the cumulative disturbance pressure value is greater than the current resilience residual value, an early warning of system instability will be output, and the recovery cost conversion coefficient corresponding to the land use type will be obtained. Based on the recovery cost conversion coefficient, the cumulative disturbance pressure value will be converted into a value depreciation assessment value.

[0014] If the cumulative value of disturbance pressure is less than or equal to the current toughness residual value, the state parameters of the digital twin base model are updated, and the simulation of subsequent time nodes is continued based on the updated digital twin base model.

[0015] Furthermore, the data heterogeneous fusion module performs spatiotemporal alignment processing on the environmental state data and the geological subsurface data to construct a digital twin base model containing spatial coordinate information, including:

[0016] Extract the first spatial coordinate information and the first timestamp information from the environmental state data;

[0017] Extract the second spatial coordinate information and the second timestamp information from the geological stratum data;

[0018] Based on the first spatial coordinate information, the first timestamp information, the second spatial coordinate information, and the second timestamp information, the environmental state data and the geological bottom layer data are mapped to a preset three-dimensional grid coordinate system to generate the digital twin base model.

[0019] Furthermore, the buffer capacity calculation module, based on the digital twin foundation model, calculates the current resilience residual of the land ecosystem, including:

[0020] Extract physical and biological attribute parameters corresponding to the environmental state data and the geological stratum data from the digital twin base model;

[0021] The physical and biological attribute parameters are input into a preset system dynamics model to construct a set of material and energy cycle equations.

[0022] Solve the aforementioned material energy cycle equations to calculate the current toughness residual value.

[0023] Furthermore, the nonlinear penetration simulation module, based on a time-series simulation algorithm, calculates the accumulation process of disturbance pressure within the land ecosystem of the preset disturbance event sequence, generating continuous time-series accumulated disturbance pressure values, including:

[0024] The preset perturbation event sequence is analyzed in chronological order, and the perturbation intensity value corresponding to each perturbation event is extracted;

[0025] The disturbance intensity value is input into a preset Markov chain model to determine the state transition probability in the time dimension, and combined with a preset cellular automata model to perform neighborhood evolution calculation in the spatial dimension, thereby simulating the spatiotemporal evolution state of the land ecosystem.

[0026] Based on the spatiotemporal evolution state, calculate the nonlinear transmission path of the disturbance intensity value within the land ecosystem;

[0027] The disturbance intensity value is accumulated along the nonlinear transmission path to obtain the cumulative disturbance pressure value.

[0028] Furthermore, the critical warning and value inversion module outputs a system instability warning, including:

[0029] Based on the difference between the current resilience residual and the cumulative disturbance pressure, and the time rate of change of the cumulative disturbance pressure, the estimated remaining time of instability is calculated;

[0030] Based on the estimated remaining time of instability and the spatial coordinate information of the digital twin base model, a vulnerability penetration heat map is generated;

[0031] The output includes the estimated remaining time of instability and the vulnerability penetration heatmap, which is the system instability warning.

[0032] Further, the step of obtaining the restoration cost conversion coefficient corresponding to the land use type, and converting the cumulative disturbance pressure value into a value depreciation assessment value based on the restoration cost conversion coefficient, includes:

[0033] Based on the land use type, the corresponding restoration cost conversion coefficient is obtained from the preset economic evaluation database;

[0034] The estimated recovery cost is calculated by multiplying the cumulative value of the disturbance pressure by the recovery cost conversion factor.

[0035] The estimated recovery cost is output as the value depreciation assessment value.

[0036] Furthermore, solving the matter-energy cycle equations to calculate the current resilience residual includes:

[0037] The material-energy cycle equations are processed using a pre-defined catastrophe theory algorithm to extract the critical point threshold of the material-energy cycle equations.

[0038] Extract the current state value of the land ecosystem from the digital twin base model;

[0039] The current resilience residual is determined based on the difference between the critical point threshold and the current state value.

[0040] Furthermore, the preset disturbance event sequence includes human development planning data and climate anomaly event data.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. This system acquires environmental state data and geological bottom layer data at different collection frequencies through a data heterogeneous fusion module. Based on the extracted spatial coordinate information and timestamp information, the above data is mapped to a preset three-dimensional grid coordinate system for spatiotemporal alignment. This design effectively overcomes the problem of inconsistent spatiotemporal dimensions of multi-source heterogeneous ecological data, realizes high-precision three-dimensional mapping of land ecosystem, and constructs an accurate digital twin base model, providing solid data support for subsequent buffer capacity calculation and penetration simulation.

[0043] 2. This system constructs a set of material and energy cycle equations and extracts critical point thresholds using catastrophe theory algorithms to accurately calculate the current resilience residual value of the system under the premise of no instability. At the same time, it uses Markov chain models and cellular automata models to calculate the evolution of preset disturbance events such as human development or climate anomalies in the spatiotemporal dimension, thereby obtaining the cumulative value of disturbance pressure along the nonlinear transmission path. Combined with the calculation of remaining time and the generation of vulnerability penetration heatmaps, the system can intuitively display the remaining time and spatial vulnerability distribution of instability, providing an accurate early warning means for the prevention and dynamic monitoring of ecological and environmental risks.

[0044] 3. When the critical early warning and value inversion module of this system determines that the cumulative value of disturbance pressure is greater than the current resilience residual value, it not only outputs an instability warning, but also matches the corresponding restoration cost conversion coefficient based on the land use type. The abstract cumulative value of disturbance pressure is multiplied by this coefficient and converted into a specific restoration cost estimate, which is then output as a value depreciation assessment value. This design realizes a direct mapping between the ecosystem's resistance to disturbance and the actual economic restoration cost, and intuitively quantifies the degree of ecological damage into economic indicators. Attached Figure Description

[0045] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0046] Figure 1 This is a system block diagram of a land ecological value spatiotemporal data simulation system based on digital twins, according to an embodiment of this application. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0048] Example 1:

[0049] Please see Figure 1A land ecological value spatiotemporal data simulation system based on digital twins. The system includes: a data heterogeneous fusion module, used to acquire environmental status data of the first acquisition frequency and geological bottom layer data of the land ecosystem of the second acquisition frequency. The environmental status data includes remote sensing image data, meteorological data and land use type data. The geological bottom layer data includes soil physicochemical property data, groundwater level data and microbial diversity index data.

[0050] Spatiotemporal alignment processing is performed on environmental status data and geological subsurface data to construct a digital twin base model containing spatial coordinate information; a buffer capacity calculation module is used to calculate the current resilience residual value of the land ecosystem based on the digital twin base model;

[0051] Among them, the current resilience residual value represents the maximum amount of external disturbance that a land ecosystem can absorb without causing system instability;

[0052] The nonlinear penetration simulation module is used to receive a preset disturbance event sequence, calculate the disturbance pressure accumulation process of the preset disturbance event sequence within the land ecosystem based on a time series simulation algorithm, and generate a continuous time series of disturbance pressure accumulation values.

[0053] The critical warning and value inversion module is used to determine whether the cumulative value of disturbance pressure is greater than the current resilience residual value at each time point in a continuous time series.

[0054] If the cumulative disturbance pressure value is greater than the current resilience residual value, an early warning of system instability will be output, and the recovery cost conversion coefficient corresponding to the land use type will be obtained. Based on the recovery cost conversion coefficient, the cumulative disturbance pressure value will be converted into a value depreciation assessment value.

[0055] If the cumulative value of disturbance pressure is less than or equal to the current toughness residual value, the state parameters of the digital twin base model are updated, and the simulation of subsequent time nodes is continued based on the updated digital twin base model.

[0056] This embodiment provides a land ecological value spatiotemporal data simulation mechanism based on digital twins; specifically, taking the land space planning demonstration of the proposed construction of an advanced manufacturing industrial park in the riverside new area as the main scenario, the total area of ​​the planning area can be divided into several basic grid units, such as a 100m×100m surface grid, and extended underground to form a three-dimensional columnar unit.

[0057] The system continuously receives two types of data at different frequencies around the planned area: one is high-frequency environmental status data, and the other is low-frequency geological bottom layer data. The former can be updated daily, weekly, or every ten days, while the latter can be updated monthly, quarterly, or semi-annually. The system first writes the above two types of data into the digital twin base, then calculates the current disturbance margin that the ecosystem can still withstand, and then advances future planning and construction, extreme weather, and other disturbances in a time sequence to determine whether they will exceed the margin. When the margin is exceeded, an instability warning and an assessment of ecological value loss are given.

[0058] During the heterogeneous data fusion stage, the system receives remote sensing image data, meteorological data, land use type data, as well as soil physicochemical property data, groundwater level data, and microbial diversity index data. For example, the planning area is assumed to contain only 4 surface grids G1, G2, G3, and G4, where G1 and G2 are located in the proposed factory area, G3 is located in the farmland buffer zone, and G4 is located in the riverbank wetland area.

[0059] At a given moment, high-frequency environmental data can be represented as follows: G1 surface temperature 32℃, vegetation cover index 0.28, land use type is reserved land for construction; G3 rainfall 12mm, land use type is irrigated farmland; low-frequency geological data can be represented as follows: G1 surface soil organic matter 18g / kg, groundwater depth 7.2m, microbial diversity index 1.8; G4 surface soil organic matter 32g / kg, groundwater depth 2.1m, microbial diversity index 3.6;

[0060] The system connects these raw data according to a unified spatial grid, forming a three-dimensional record structure of grid number - timestamp - above-ground status - underground status; the first and second acquisition frequencies here can be different, for example, remote sensing and meteorological data are updated every 7 days, while groundwater and microbial data are updated every 90 days.

[0061] During the buffer capacity calculation phase, the system extracts state variables reflecting the ecological self-stabilization capacity from the digital twin base to form the current resilience residual value. For ease of explanation, the resilience residual value output by the system is represented by a uniform perturbation unit RU. For G1, considering factors such as low soil fertility, large groundwater depth, and weak microbial diversity, the current resilience residual value can be calculated to be 85RU. For G4, due to the strong self-purification capacity of the wetland, the current resilience residual value can be 160RU. This value is not a visual representation of the surface, but rather a characterization of how much additional pressure the current ecosystem can absorb without system instability.

[0062] In the nonlinear penetration simulation stage, the system receives a sequence of future disturbance events. For example, the first phase of construction in the industrial park will lead to surface hardening, increased water intake, and earthwork disturbance. The second year may be superimposed with high temperatures and low rainfall. The third year may see continued expansion. For ease of demonstration, it is assumed that the system receives disturbance events at three consecutive time points: construction disturbance of 30RU at time T1, industrial water intake disturbance of 40RU at time T2, and extreme drought disturbance of 35RU at time T3. The system does not simply add the three together, but obtains the cumulative value of disturbance pressure in the continuous time series based on the transmission and amplification relationship within the land ecosystem.

[0063] For example, in G1, the cumulative value for T1 is 28 RU; by T2, considering insufficient groundwater recharge and soil compaction, the cumulative value increases to 76 RU; by T3, due to drought further weakening the buffering capacity, the cumulative value rises to 98 RU; the system is then compared hourly with the resilience residual value of 85 RU.

[0064] During the critical warning and value inversion stage, if the cumulative value of disturbance pressure at a certain time point is greater than the current resilience residual value, a system instability warning will be output; for example, if 98RU is greater than 85RU at T3, it means that the grid has exceeded the tolerable range.

[0065] The system reads the restoration cost conversion coefficient according to the land use type corresponding to the grid. If the coefficient corresponding to the reserved construction land is 12 standard restoration resource units / RU, the value depreciation assessment value can be 98×12=1176 standard restoration resource units, or calculated as 156 standard restoration resource units based on the excess part of 13RU. The two can be determined by the deployment unit using the preset method.

[0066] If the limits are not exceeded at a certain time point, for example, if the cumulative disturbance pressure value of G4 in the same stage is only 88RU, which is less than its 160RU toughness residual value, the system updates its state parameters such as soil moisture, groundwater recharge status, and vegetation recovery degree, and continues to advance subsequent simulations. In terms of abnormal data processing, if a certain high-frequency data is missing at the same time point, such as cloud cover causing missing remote sensing images, the system can call interpolation results from adjacent time points or ground station observations to fill in the missing data.

[0067] If low-frequency groundwater data exceeds the preset validity period, for example, if groundwater monitoring in a certain grid has not been updated for more than 180 days, the grid status is marked as low-confidence data, and a confidence level label is attached to the subsequent output; if multiple abnormal events occur simultaneously in a certain grid at the same time point, the system can first standardize them according to the event type, and then merge them into the same time slice to avoid duplicate measurement; if the recovery cost database does not find a completely matching land use type, it will fall back to the previous level type according to the preset mapping rules, for example, riverside ecological farmland will fall back to the farmland category, and the fallback log will be recorded.

[0068] During the feasibility study of the Yanjiang New Area project, the relevant authorities planned to compare the ecological consequences of the original plan and the plan to reduce water intake by 20%. After simulating the original plan, the system found that G1 and G2, which are close to the plant area, entered the over-limit state in the third year, predicting that the groundwater would continue to decline and cause the risk of soil salinization. After simulating the reduction plan, the maximum cumulative disturbance pressure of G1 dropped to 81RU, which did not exceed the 85RU threshold, and G2 also changed from over-limit to critical edge. Therefore, the system output a more suitable plan on the digital twin interface and gave the value loss difference of each grid.

[0069] The purpose of this step is to integrate visible changes on the surface with hidden resilience underground into the same digital twin link, thereby enabling continuous deduction from the current situation to early warning of ecological instability and value quantification.

[0070] The land ecological value spatiotemporal data simulation system based on digital twins uses a data heterogeneous fusion module to perform spatiotemporal alignment processing on environmental state data and geological bottom layer data, and constructs a digital twin base model containing spatial coordinate information, including: extracting the first spatial coordinate information and the first timestamp information from the environmental state data; and extracting the second spatial coordinate information and the second timestamp information from the geological bottom layer data.

[0071] Based on the first spatial coordinate information, the first timestamp information, the second spatial coordinate information, and the second timestamp information, the environmental state data and the geological bottom layer data are mapped to a preset three-dimensional grid coordinate system to generate a digital twin base model.

[0072] This embodiment provides a mechanism for spatiotemporal alignment of multi-source data. Specifically, in the aforementioned riverside new area planning scenario, simply overlaying high-frequency surface data with low-frequency underground data will result in a discrepancy between the observed time and the measured time. For example, if the remote sensing image was acquired on May 1st, while the groundwater data was detected on April 15th, without processing, the system will mistakenly assume that the two describe the same ecological state, leading to distortion of the base model. Therefore, this embodiment introduces a dual alignment process of spatial coordinates and timestamps to uniformly map heterogeneous data to a preset three-dimensional grid coordinate system.

[0073] The system first extracts the first spatial coordinate information and the first timestamp information from the environmental status data; remote sensing images typically correspond to an area and a sampling time, for example, an image covers a coordinate range. , The shooting time was 10:00 on May 1st; meteorological data can come from stations, for example, station P1 is located at coordinates (80m, 120m) and the recording time is 12:00 on May 1st; land use type patches can be accompanied by their polygon boundaries and the date of the most recent interpretation; the system converts this information into a unified plane coordinate reference and projects it onto the surface grid layer through rasterization;

[0074] The system extracts the second spatial coordinate information and the second timestamp information from the geological stratum data; the groundwater level is usually obtained from the monitoring well, for example, well W1 is located at coordinates (90m, 110m), the detection depth is 8m, and the detection time is April 15;

[0075] Soil physicochemical property data can be obtained from profile sampling points. For example, S1 is located at coordinates (100m, 100m), with sampling depths of 0.5m and 1.5m, and the detection time is April 20th. The microbial diversity index can correspond to specific soil layer sampling points. The system converts these point or profile data into three-dimensional coordinate points, that is, adds the depth direction coordinate Z to the original planar position.

[0076] To illustrate the mapping process, assume that the preset three-dimensional grid coordinate system consists of a 2×2 planar grid and 3 layers of depth units, forming 12 basic solid units; the upper layer unit G1 can be represented as coordinates (1,1,1), the middle layer as coordinates (1,1,2), and the deep layer as coordinates (1,1,3); if the center of a pixel in the remote sensing image falls into the planar position 1,1, then its surface state is written into the surface associated slot at coordinates (1,1,1);

[0077] If groundwater monitoring well W1 is located in the same plane unit and reflects the water level in the middle and deep layers, its data is mapped to coordinates (1,1,2) or (1,1,3). For time alignment, the system can set a unified simulation time of May 1st. If the groundwater data comes from April 15th and the allowed time window is 30 days, the data can be aligned to May 1st through time interpolation or validity period continuation. If it exceeds 30 days, it is marked as outdated data and will not directly participate in high-confidence simulation.

[0078] The system can use a nearest neighbor plus time-weighted alignment method. For example, G1 needs groundwater status on May 1st, but the most recent groundwater measurement was on April 15th, and the previous adjacent measurement point was on January 15th. If the time weight is assigned higher values ​​for more recent data, with April 15th having a weight of 0.8 and January 15th having a weight of 0.2, and the two data points being 7.2m and 6.8m respectively, then the aligned groundwater depth can be taken as 7.2 × 0.8 + 6.8 × 0.2 = 7.12m. This avoids the jump caused by directly using outdated data.

[0079] If spatial coordinates are missing for a certain environmental status data, such as only having the name of an administrative village without a precise location, the system will first call place name resolution or land parcel code to map to an approximate area; if it still cannot be located, it will be temporarily stored in the verification area and will not enter the formal base.

[0080] If multiple conflicting sources exist at the same location at the same time, such as two land use type interpretation results being farmland and construction land respectively, the system can prioritize the selection based on the confidence level of the data source, or mark it as a conflict unit, and the subsequent warning results will include the conflict label; if the underground profile sampling depth is inconsistent with the preset grid layer, such as the measured depth being 0.7m while the grid layer boundary is 0.5m and 1.0m, the system can write to the two adjacent layers according to the depth weight.

[0081] In the riverside wetland on the north side of Yanjiang New Area, remote sensing images obtained on May 1 show that the surface vegetation of the wetland area is still relatively intact, but the most recent underground microbial sample was taken at the end of March, and the groundwater monitoring data was taken in mid-April. After the above spatiotemporal alignment, the system simultaneously records the combination of seemingly stable surface and weakening underground recharge in the same three-dimensional grid, thus providing a consistent data base for subsequent resilience assessment.

[0082] The purpose of this step is to eliminate the misalignment of multi-source heterogeneous data in terms of sampling time, sampling location, and spatial dimension, thereby achieving a unified expression and stable driving of the digital twin base model.

[0083] The land ecological value spatiotemporal data simulation system based on digital twins includes a buffer capacity calculation module that calculates the current resilience residual value of the land ecosystem based on the digital twin base model. This includes: extracting physical and biological attribute parameters corresponding to environmental state data and geological bottom layer data from the digital twin base model; inputting the physical and biological attribute parameters into a preset system dynamics model to construct a set of material and energy cycle equations; and solving the set of material and energy cycle equations to calculate the current resilience residual value.

[0084] This embodiment provides a resilience residual calculation mechanism. Specifically, in the aforementioned scenario, a digital twin base with complete spatiotemporal alignment is insufficient to answer the question of how much development intensity the land can withstand. This is because high surface greenness does not equate to strong system buffering capacity. Slow groundwater recharge, declining soil organic matter, and degradation of microbial communities can all leave the ecosystem in a state of superficial stability but internal fragility. Therefore, this embodiment extracts physical and biological attribute parameters from the base model and inputs them into the system dynamics model to calculate the current resilience residual.

[0085] The system first extracts physical attribute parameters, such as soil moisture content, groundwater depth, rainfall, evapotranspiration intensity, land hardening ratio, and surface runoff coefficient; simultaneously, it extracts biological attribute parameters, such as vegetation cover index, microbial diversity index, soil organic matter content, and aboveground biomass index. For ease of explanation, it is assumed that the current state of grid G1 is: groundwater depth 7.12m, soil organic matter 18g / kg, vegetation cover index 0.28, microbial diversity index 1.8, and land hardening ratio 0.15; while for grid G4, these values ​​are 2.1m, 32g / kg, 0.76, 3.6, and 0.02, respectively.

[0086] The system inputs the above parameters into a preset system dynamics model; this model can establish the coupling relationship between variables around water cycle—soil nutrient cycle—biological activity response—human disturbance feedback;

[0087] To avoid abstraction, this embodiment adopts a simplified sandbox simulation to construct a first-order difference form of the material and energy cycle equations: the model retains three core state variables, namely a water reserve W, a soil mass parameter S, and a biological activity parameter B; the water reserve is constrained by rainfall input, evapotranspiration output, and surface infiltration rate; soil mass is closely related to water and previous organic matter; biological activity is driven by the coupling of water and soil mass; the system iteratively solves the material and energy cycle equations based on the time step, and obtains the steady-state value of the system when the rate of change of each state variable tends to be stable;

[0088] The comprehensive resilience index R is defined as a weighted average of three factors under steady-state conditions, for example... If G1, after solving the equations over multiple time periods and normalizing, has a steady-state value of W=40, S=35, B=30, then R=35.5; if G4, after normalization, has a steady-state value of W=80, S=70, B=85, then R=78.5; the system can further map R to a toughness residual under a uniform disturbance unit, for example, 85RU for G1 and 160RU for G4.

[0089] In practical implementation, the system dynamics model does not perform a single static weighting, but rather continuously evolves and solves based on the aforementioned material and energy cycle equations. For example, the water update term in the equations shows that increased rainfall will increase W, but if the proportion of surface hardening in the physical property parameters is high, the coefficient of rainfall being converted into effective replenishment will decrease significantly. The decrease in soil organic matter will weaken S, and further reduce B by affecting microbial activity through coupling terms. The decrease in B will increase the soil dissipation term, which in turn will reduce S. By cyclically updating these state variables, the system obtains a buffer capacity result that is closer to the real ecological dynamic process.

[0090] If a biological attribute parameter is missing, such as the microbial diversity index not yet being experimentally tested, the system can supplement it based on the historical median value of similar soil types, but simultaneously reduce the confidence level of the results; if there are outliers in a grid, such as negative or out-of-range values ​​in groundwater depth, the system will first perform data verification, remove obvious errors, and then participate in modeling; if the denominator is zero during normalization, such as the range being zero due to the same mean of a batch of samples, the system will directly use a preset constant mapping or switch to quantile normalization to avoid calculation interruption;

[0091] During the environmental assessment prior to the construction of the first phase of the industrial park, the surface of G1 had not yet been hardened over a large area, and the remote sensing images did not show any obvious abnormalities. However, the system extracted data showing that the depth of the underlying groundwater was already relatively deep, and the microbial diversity index had declined for two consecutive quarters. After inputting these parameters into the system dynamics model, the resilience residual value of G1 was found to be only 85RU, which is significantly lower than the 110RU of the surrounding farmland buffer zone and also lower than the 160RU of the riparian wetland. This shows that the same construction disturbance can produce significantly different ecological risks in different grids.

[0092] The purpose of this step is to couple the visible surface state with deep ecological processes, thereby enabling a computable expression of the current ecosystem's buffer capacity.

[0093] A land ecological value spatiotemporal data simulation system based on digital twins uses a nonlinear penetration simulation module based on time series simulation algorithms to calculate the accumulation process of disturbance pressure within the land ecosystem for a preset sequence of disturbance events, generating continuous time-series accumulated disturbance pressure values, including:

[0094] The system analyzes a preset sequence of disturbance events in chronological order and extracts the disturbance intensity value corresponding to each disturbance event. The disturbance intensity value is then input into a preset Markov chain model to determine the state transition probability in the time dimension. In combination with a preset cellular automata model, the system performs neighborhood evolution calculations in the spatial dimension, thereby simulating the spatiotemporal evolution of the land ecosystem.

[0095] Based on the spatiotemporal evolution, the nonlinear transmission path of disturbance intensity values ​​within the land ecosystem is calculated; the disturbance intensity values ​​are accumulated along the nonlinear transmission path to obtain the cumulative disturbance pressure value.

[0096] This embodiment provides a nonlinear penetration simulation mechanism. Specifically, in the aforementioned scenario, even if the resilience residual of a certain grid has been obtained, significant deviations will occur if all future disturbances are directly added together for prediction. This is because the land ecosystem is not a linear container; some disturbances decay over time, while others amplify each other and propagate along adjacent grids. For example, the impact of water intake from the factory on the groundwater level may first manifest in the factory area and then gradually spread to farmland and wetlands. Therefore, this embodiment introduces a Markov chain model in the time dimension and a cellular automata model in the spatial dimension to perform a penetration simulation of the spatiotemporal evolution of disturbance pressure.

[0097] The system first analyzes the preset disturbance event sequence in chronological order. Taking the aforementioned three-year plan for the riverside new area as an example, we can set T1 as road and factory foundation construction with a disturbance intensity of 30RU; T2 as industrial production operation and increased groundwater extraction with a disturbance intensity of 40RU; and T3 as an extreme drought year with a disturbance intensity of 35RU. The disturbance intensity value here can be obtained by standardizing factors such as construction area, water extraction scale, and rainfall deviation.

[0098] The system uses a Markov chain model to determine the state transition probabilities in the time dimension. For ease of explanation, the ecological state is only divided into stable state A, stressed state B, and critical state C. If the current state is A, after a moderate disturbance, the probability of transitioning to B at the next time node is 0.6, and the probability of remaining in A is 0.4. If the current state is B and then suffers from drought, the probability of transitioning to C is 0.7. Taking G1 as an example, the initial state is A. After T1, there is a high probability of entering B. After T2, the state continues to remain in B or approaches C. After T3, the state may enter C. Thus, the time dimension is no longer a rigid jump, but rather has path dependence.

[0099] The system uses a cellular automata model to perform spatial neighborhood evolution. G1, G2, G3, and G4 can be considered as four adjacent cells, where G1 is adjacent to G2, G2 is adjacent to G3, and G3 is adjacent to G4. If at a certain moment, G1 experiences an increase in groundwater pressure due to construction and water extraction, not only will G1 itself deteriorate, but it may also spread to adjacent cells in a certain proportion. For example, G1 transmits 0.3 times the pressure to G2, G2 then transmits 0.2 times to G3, and G3 transmits 0.1 times to G4.

[0100] If G2 itself is simultaneously subjected to road hardening disturbance, its final state is not a simple superposition, but is determined by the state of its neighboring area and its own state. In order to achieve transparent solution of this evolution process, the system performs spatiotemporal state update calculation according to the following business logic at each time step: the cumulative value of the disturbance pressure of the target grid at the next time step is equal to the product of its existing pressure value at the current time step and the retention coefficient extracted from the local Markov state transition probability, plus the direct intensity of the newly occurring disturbance event at the next time step, and then plus the sum of the pressure increments that have permeated and diffused from the surrounding adjacent grids through the cellular automaton rules.

[0101] Through the above-mentioned structured causal decomposition, the implicit computational limitations of the spatiotemporal transmission amplification effect are effectively broken; based on the above-mentioned spatiotemporal evolution state, the system further calculates the nonlinear transmission path;

[0102] For ease of explanation, assume that at time T1, 20 RUs of the 30 RUs in G1 remain locally, 8 RUs are transferred to G2, and 2 RUs are dissipated; at time T2, an additional 40 RUs are added, but because G1 is already under pressure, it is more sensitive to the new disturbances, so the local retention ratio increases, forming a local pressure of 50 RUs, while transferring another 12 RUs to G2; by time T3, the extreme drought adds 35 RUs to the already pressured system, triggering an amplification effect, causing the new effective pressure in G1 to no longer be 35 RUs, but 45 RUs.

[0103] The cumulative perturbation pressure values ​​for G1 in the three time periods can be 28RU, 76RU, and 98RU; while for G2 they can be 10RU, 29RU, and 51RU respectively. This reflects the nonlinear characteristic that the closer to the critical point, the greater the effective pressure caused by the same subsequent perturbation.

[0104] If no new disturbance events occur at a certain time point, the system does not force the pressure to remain unchanged, but performs attenuation calculations based on natural recovery parameters; for example, if there are no new disturbances and the rainfall is normal, the pressure of the previous cycle can drop by 10% to 20%; if multiple disturbance events occur at the same time, such as construction and high temperature arriving on the same day, the system first sets priority or parallel synthesis rules according to the event category to prevent repeated amplification.

[0105] If there is no actual hydrological or ecological connectivity between a grid and its neighboring grids, such as the presence of an isolation and seepage prevention structure, the neighborhood transfer coefficient in the cellular automaton can be directly set to zero.

[0106] In grid G1 of the central factory area of ​​Yanjiang New Area, system simulation found that although the first phase of construction only caused limited surface changes, its inhibitory effect on soil compaction and groundwater recharge would be significantly amplified during the industrial water intake in the second year.

[0107] The drought in the third year was not an isolated event, but continued to penetrate along the path of factory area-farmland-wetland, causing the originally safe G3 to approach the critical edge. Without this kind of spatiotemporal penetration simulation, it would be difficult to discover the risk that was slowly accumulating underground based on a single-period static assessment.

[0108] The purpose of this step is to characterize the real process of disturbances accumulating over time, spreading spatially, and amplifying near the critical point within the ecosystem, thereby enabling the solution of continuous temporal disturbance pressure accumulation.

[0109] The land ecological value spatiotemporal data simulation system based on digital twins outputs system instability warnings from the critical early warning and value inversion module, including: calculating the estimated remaining time of instability based on the difference between the current resilience residual value and the cumulative value of disturbance pressure, as well as the time change rate of the cumulative value of disturbance pressure.

[0110] Based on the estimated remaining time of instability and the spatial coordinate information of the digital twin base model, a vulnerability penetration heatmap is generated; the output includes a system instability warning containing the estimated remaining time of instability and the vulnerability penetration heatmap.

[0111] This embodiment provides a critical early warning output mechanism; specifically, in the aforementioned scenario, although judging whether the limit has been exceeded can provide a risk conclusion, it is still insufficient to support planning decisions.

[0112] Because the competent authorities are usually more concerned with two issues: first, how long before the problem occurs, and second, where the problem will first appear and where it will spread; therefore, this embodiment, based on comparing the toughness residual value and the cumulative value of disturbance pressure, further calculates the estimated remaining time of instability and generates a vulnerability penetration heat map; the system first reads the difference between the toughness residual value and the cumulative value of disturbance pressure of each grid at the current time node, and estimates the estimated remaining time of instability by combining the time change rate of the cumulative value of disturbance pressure.

[0113] For ease of explanation, let's assume that G1's current resilience residual is 85RU, and the cumulative values ​​of disturbance pressure in the most recent three periods are 52RU, 68RU, and 80RU, respectively. Then, the rate of change in the most recent period can be approximated as an increase of 12RU per period. At this point, G1 is still 5RU away from the critical point, and the estimated remaining time for instability is about 5 / 12 periods, or about 0.42 periods. If converted to quarters, this is about 1.3 months. For another example, G3's resilience residual is 110RU, the current cumulative value is 74RU, and the growth rate in the most recent period is 6RU / per period. Then, the estimated remaining time for instability is about 6 periods, significantly later than G1.

[0114] After obtaining the estimated remaining time of instability for each grid, the system generates a vulnerability penetration heatmap by combining the spatial coordinates in the digital twin base. The heatmap can be expressed using hierarchical color bands; for example, less than one cycle remaining is marked as red, one to three cycles as orange, three to six cycles as yellow, and more than six cycles as green. If the same grid has not yet exceeded the limit, but its growth rate continues to rise, it can also be marked as an early warning edge state. For visualization, let G1 be red, G2 be orange, G3 be yellow, and G4 be green. If further superimposed with underground layer data, the system can also show in the three-dimensional scene that shallow soil vulnerability and deep groundwater vulnerability are not in the same location, thus forming a penetration heatmap.

[0115] Subsequently, the system outputs an instability warning result. The warning content includes not only the word "exceeding the limit," but also the warning level, the estimated remaining time of instability, the main risk factor, and the heat map number. For example, for G1, the system may output: Level 1 warning, expected to enter the instability zone within 1.3 months, the main factor being water intake disturbance combined with drought, and it is recommended to immediately reduce water intake intensity and increase ecological water replenishment; for G2, the system may output: Level 2 warning, expected to approach the critical point within 2.8 quarters, and it is recommended to reduce the hardened area and set up an infiltration buffer zone.

[0116] If the time change rate of a certain grid is close to zero or negative, it indicates that the disturbance pressure has entered a plateau or recovery period. At this time, the estimated remaining time of instability can be marked as not triggering or infinite, instead of giving an erroneous maximum value. If the difference is already negative, it means that the system has crossed the critical point. In this case, there is no need to estimate the remaining time, but to directly mark it as unstable. If there is insufficient confidence in the data in the same grid, such as the recent lack of underground data, the warning output can be accompanied by a low confidence label to remind the manager to prioritize resampling.

[0117] At the demonstration meeting for the Yanjiang New Area plan, the system marked two grids on the northwest side of the factory area in red, one of which was expected to have only 1.3 months left before instability, and the other 2.1 months; the farmland buffer zone was marked in yellow, with an estimated remaining time of about 1.5 years; and most of the riverside wetlands were marked in green. Managers could intuitively see that the risks did not arrive simultaneously, but would first penetrate from the edge of the factory area and then gradually approach the farmland and wetlands; this result was more suitable for guiding planning adjustments than a single numerical report.

[0118] The purpose of this step is to upgrade the judgment of whether instability will occur to a spatial early warning of when instability will occur and where instability will occur first, thereby achieving a more operational risk presentation.

[0119] Example 2:

[0120] A land ecological value spatiotemporal data simulation system based on digital twins acquires restoration cost conversion coefficients corresponding to land use types. Based on these conversion coefficients, the accumulated disturbance pressure is converted into a value depreciation assessment value, including:

[0121] Based on land use type, the corresponding restoration cost conversion coefficient is obtained from the pre-set economic evaluation database; the cumulative value of disturbance pressure is multiplied by the restoration cost conversion coefficient to calculate the estimated restoration cost; and the estimated restoration cost is output as the value depreciation assessment value.

[0122] This embodiment provides a value inversion mechanism. Specifically, in the aforementioned scenario, simply outputting an early warning of ecological instability is insufficient to directly support the quantitative assessment of project access, because management departments need to convert ecological risks into comparable and calculable objective resource consumption. Therefore, this embodiment extracts the restoration cost conversion coefficient from a preset economic evaluation database based on land use type and converts the cumulative value of disturbance pressure into a value depreciation assessment value.

[0123] The system presets restoration cost conversion coefficients for various land use types in the database. For illustration, the coefficient for reserved construction land is set at 12 standard restoration resource units / RU, the coefficient for irrigated farmland is set at 18 standard restoration resource units / RU, and the coefficient for riparian wetlands is set at 35 standard restoration resource units / RU. These coefficients can be comprehensively formed by the resources required for historical restoration projects, regional ecological compensation base, soil improvement engineering consumption, wetland restoration engineering consumption, etc., and can be further subdivided according to administrative region, soil type, or governance difficulty.

[0124] When the system detects that a grid has exceeded the limit or is in a high-risk state, it first reads its land use type and then retrieves the corresponding coefficient from the database. Taking G1, G3, and G4 as examples, assuming that the current cumulative disturbance pressure values ​​of the three are 98RU, 74RU, and 88RU, respectively, the corresponding land use types are construction reserve land, irrigated farmland, and riverbank wetland.

[0125] The estimated recovery cost can be calculated as 1176 standard repair resource units, 1332 standard repair resource units, and 3080 standard repair resource units, respectively. If the system is configured to calculate based on the excess portion, the G1 resilience residual value is 85RU, exceeding the limit by only 13RU, and its loss value is 156 standard repair resource units. If G4 does not exceed the limit, the loss value can be omitted, and only the potential risk value can be output. Both methods can be used as optional strategies in engineering deployment.

[0126] To make the assessment more operational, the system can also add correction factors; for example, if a wetland is located in a drinking water source protection area, its restoration will consume more resources, so the protection level correction factor of 1.3 can be multiplied by the basic coefficient of 35 standard restoration resource units / RU to form 45.5 standard restoration resource units / RU; another example is that if a farmland has already built a salt drainage canal and an ecological isolation belt, the actual restoration difficulty is slightly lower, so it can be multiplied by a correction factor of 0.9.

[0127] In this way, the value inversion results are closer to the actual level of regional governance resource consumption. In terms of the anomaly handling mechanism, if the coefficient of a certain sub-category does not exist in the database, such as the salt marsh type riparian wetland not having a separate database, the system can backtrack to the coefficient of the riparian wetland category according to the preset mapping relationship. If there is still no available coefficient after backtracking, it can be temporarily estimated according to the average coefficient of the same region, and the temporary value is marked in the output.

[0128] If a grid contains two land use attributes simultaneously, such as overlapping of map boundaries resulting in half being farmland and half being reserved land for future construction, then a mixing coefficient can be calculated based on the area proportion. For example, calculating... One standard repair resource unit / RU;

[0129] In comparing the original and optimized plans for the Yanjiang New Area, the original plan showed a combined loss of 320 standard restoration resource units for the two reserved construction sites G1 and G2, and a potential loss of 960 standard restoration resource units for the G3 farmland buffer zone. The optimized plan, by reducing water intake and increasing the permeable green belt, ensured that G1 did not exceed the limits, and the extent of exceeding the limits for G2 was also significantly reduced.

[0130] The total loss value has decreased to below 90 standard restoration resource units; management departments can then directly incorporate ecological risks into the comprehensive consideration of planning resources, avoiding one-sided judgments based solely on the scale of construction. The purpose of this step is to transform the abstract ecological pressure results into quantifiable and comparable objective resource loss indicators, thereby achieving a unified measurement between ecological risks and the allocation of planning resources.

[0131] A land ecological value spatiotemporal data simulation system based on digital twins is used to solve the material and energy cycle equations and calculate the current resilience residual value. This includes: using a pre-set catastrophe theory algorithm to process the material and energy cycle equations and extracting the critical point threshold of the material and energy cycle equations.

[0132] The current state value of the land ecosystem is extracted from the digital twin base model; the current resilience residual value is determined based on the difference between the critical point threshold and the current state value.

[0133] This embodiment provides a resilience residual solution mechanism based on critical point extraction. Specifically, in the aforementioned scenario, if the buffer capacity is evaluated solely based on the current output value of the system dynamics model, although it can reflect the general trend, it may still underestimate the risk when approaching the edge of ecological mutation. This is because some land ecosystems may exhibit slow changes over a long period of time, but once they cross the critical point, they may quickly slide into salinization, degradation, or vegetation disintegration.

[0134] To address this bottleneck, this embodiment further introduces a catastrophe theory algorithm to extract the critical point threshold from the material-energy cycle equations, and then subtracts it from the current state value to obtain a resilience residual value that better reflects the meaning of the pre-catastrophe margin. The system first establishes a comprehensive state surface based on the aforementioned state variables such as water reserve W, soil quality parameter S, and biological activity B. To avoid the opaque calculation logic of the parameter mapping process, the system uses a preset dimensionality reduction transformation mechanism to extract control variables.

[0135] Specifically, the difference between the preset upper limit of the safe threshold for water reserves and the current water reserves is multiplied by the corresponding first characteristic coefficient to obtain the water pressure factor P; the difference between the soil quality parameter and the biological activity quantity and their corresponding upper limit of the safe threshold is multiplied by their respective characteristic coefficients and summed to obtain the biological degradation factor Q; as P and Q increase dynamically, the system state value H, which characterizes the macroscopic energy deformation potential, will gradually approach the critical region from the stable region.

[0136] The system uses a pre-defined catastrophe theory algorithm to map the water pressure factor P to an asymmetric factor and the biological degradation factor Q to a bifurcation factor, constructing a cusp catastrophe potential function with the system state value H as the state variable; the system then extracts the equilibrium surface equation based on this potential function. ;

[0137] The system simultaneously forms the equilibrium surface equations and its first-order partial derivative equations with respect to the system state value H. After eliminating H, the bifurcation set equations that determine the transition behavior of the system are obtained. During simulation, the system scans the state change trajectory under different P and Q combinations. When the input P and Q combination crosses the boundary defined by the divergence equation in the parameter space, the response of the state value to the control variable changes from gradual to steep. The system extracts the corresponding environmental disturbance pressure value as the critical point threshold of the grid.

[0138] Here's an example deduction: Suppose that after solving, the G1 mesh exhibits a significant transition near the control combination P=70 and Q=60, crossing the boundary determined by the divergence equation, and the corresponding comprehensive critical point threshold can be converted to 100RU; while the current state value of this mesh in the digital twin base is converted to 15RU of occupied pressure, then the current resilience residual value can be recorded as 100-15=85RU;

[0139] For example, the critical threshold of the G4 wetland grid is relatively high, at 190RU, and the current state value is 30RU, so the resilience residual value is 160RU. The current state value here can be understood as the internal pressure that the system has already endured but has not yet released, and is not equal to the single disturbance value observed externally.

[0140] Catastrophe theory algorithms can help identify grids where system state changes are gradual but are actually close to critical. For example, G2 may look similar to G3 in a normal linear model, but after critical point extraction, it is found that the control parameters of G2 are close to the branch boundary, and the calculated critical point threshold is only 92RU, while the current state value has reached 70RU, so its toughness residual is only 22RU. Although G3 also has a current state value of 60RU, its critical point threshold is 110RU due to the greater distance of the control parameters, and its toughness residual is still 50RU.

[0141] Therefore, the critical point threshold is not a fixed constant, but is determined by the internal material and energy cycle structure of the grid and its position in the catastrophe potential function. In terms of the abnormal situation handling mechanism, if no obvious catastrophe characteristics appear in the solution of the equation system of a certain grid, such as maintaining a monotonically smooth change within the observation range, the system can revert to the conventional stability threshold method and use the empirical threshold or historical quantile as the critical point threshold.

[0142] If the current state value is greater than the critical threshold, the resilience residual is recorded as zero or negative and directly marked as out of bounds; if there are multiple local critical points in the equation system, the system can prioritize the first critical point reached as the safety boundary to improve the conservatism of the early warning.

[0143] In the southern farmland-factory transition zone of the Yanjiang New Area, some plots still show normal texture on remote sensing, but catastrophe theory algorithms show that their groundwater-microbe coupling system is close to the edge of the potential function's divergence set, and the critical point threshold has been significantly reduced. Based on this, the system calculates that the toughness residual value of these plots is much lower than that of the surrounding plots, thus exposing the high-risk zone for future salinization expansion in advance.

[0144] The purpose of this step is to identify the critical boundary of an ecosystem shifting from continuous change to rapid instability, thereby enabling the calculation of resilience residuals that more closely approximate the actual critical risks.

[0145] The land ecological value spatiotemporal data simulation system based on digital twins pre-sets a series of disturbance events, including human development planning data and climate anomaly event data.

[0146] This embodiment provides a mechanism for constructing a perturbation event sequence. Specifically, in the aforementioned scenario, if future simulations only input development plans without considering climate anomalies, the system may draw overly optimistic conclusions. Conversely, if only climate risks are considered while actual construction activities are ignored, the system cannot reflect the consumption of ecological resilience by the planning decisions themselves.

[0147] Therefore, in this embodiment, the preset disturbance event sequence is explicitly constructed as a combined sequence that includes at least human development planning data and climate anomaly event data;

[0148] Human development planning data may include the area of ​​the proposed factory buildings, the proportion of paved roads, the scale of groundwater extraction, the duration of construction, the depth of earthwork excavation, and the proportion of green space restoration. For illustration, let's assume that the first phase of the industrial park starts at T1, with an additional 20 hectares of paved area, corresponding to a disturbance intensity of 20RU; after T2 is put into operation, the annual water extraction will increase by 500,000 cubic meters, corresponding to a disturbance intensity of 40RU; if T3 continues to expand, it will increase by another 15RU.

[0149] Climate anomaly data can include drought, rainstorms, high temperatures, cold waves, etc. Among them, those closely related to the resilience of land ecosystems can be given priority in simulation; for example, T2 is superimposed with a summer high temperature and drought, corresponding to 20RU; T3 is an extreme low rainfall, corresponding to 35RU;

[0150] When constructing sequences, the system can uniformly encode two types of events into a structure of time node—event type—disturbance intensity—affected area; for example: the first record is T1, construction hardening, 20RU, and a combination of G1-G2; the second record is T2, increased water intake, 40RU, and a combination of G1-G3; the third record is T2, high temperature and drought, 20RU, and a combination of G1-G4; the fourth record is T3, expansion, 15RU, and a combination of G2-G3; the fifth record is T3, extreme low rainfall, 35RU, and a combination of all grids; the subsequent nonlinear penetration simulation module can directly read this sequence for time-series progression;

[0151] Because the two types of events have different mechanisms of action, the system can also set different transmission parameters for them; development planning events are more likely to create persistent pressure in local areas, such as hardening and water extraction affecting groundwater recharge; climate anomaly events are more likely to have synchronous effects across regions, such as drought reducing water reserves over a large area; thus, with the same 20RU, construction hardening may mainly be concentrated in G1 and G2, while high temperature and drought may act on G1 to G4 simultaneously, and exhibit a longer recovery time lag in the wetland grid;

[0152] If there is no formal development plan data for a certain year, the system can generate several candidate event sequences based on feasibility study schemes, historical development intensity, or multiple scenario assumptions, such as three schemes: conservative development, moderate development, and high-intensity development. If there is uncertainty about future climate anomalies, the system can select multiple scenarios such as normal, drought, and extreme drought from the scenario library provided by the meteorological department for parallel simulation.

[0153] If two types of events completely conflict at the same time point, such as a plan to reduce water withdrawal but actual climate causes an abnormal rise in groundwater, the system will store them in the database according to the event type, without prior cancellation, and leave the cancellation relationship to be calculated by the subsequent spatiotemporal simulation module.

[0154] Before the approval of the project in the Yanjiang New Area, the competent authorities input three sets of event sequences: the original development plan, the original development plan under normal climate, the optimized development plan under two consecutive years of drought, and the optimized development plan under two consecutive years of drought. The system simulation results showed that under normal climate, the original plan was still on the edge of control, but once the continuous drought was superimposed, G1 and G2 would enter instability in advance. Even if the optimized plan was superimposed with drought, it could still keep most grids below the critical value.

[0155] Therefore, incorporating human development planning data and climate anomaly event data into the event sequence can more accurately reflect the future stress path of land ecosystems. The purpose of this step is to cover the combined sources of human disturbance and natural anomalies, thereby achieving a complete contextualized input of future ecological risks.

[0156] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A land ecological value spatiotemporal data simulation system based on digital twins, characterized in that, The system includes: The data heterogeneous fusion module is used to acquire environmental status data of the first collection frequency and geological bottom layer data of the second collection frequency of the land ecosystem. The environmental status data includes remote sensing image data, meteorological data and land use type data, and the geological bottom layer data includes soil physicochemical property data, groundwater level data and microbial diversity index data. Spatiotemporal alignment processing was performed on environmental status data and geological subsurface data to construct a digital twin base model containing spatial coordinate information; The buffer capacity calculation module is used to calculate the current resilience residual of the land ecosystem based on the digital twin base model; where the current resilience residual represents the maximum amount of external disturbance that the land ecosystem can absorb without the occurrence of system instability. The nonlinear penetration simulation module is used to receive a preset disturbance event sequence, calculate the disturbance pressure accumulation process of the preset disturbance event sequence within the land ecosystem based on a time series simulation algorithm, and generate a continuous time series of disturbance pressure accumulation values. The critical warning and value inversion module is used to determine whether the cumulative value of disturbance pressure is greater than the current resilience residual value at each time point in a continuous time series. If the cumulative disturbance pressure value is greater than the current resilience residual value, an early warning of system instability will be output, and the recovery cost conversion coefficient corresponding to the land use type will be obtained. Based on the recovery cost conversion coefficient, the cumulative disturbance pressure value will be converted into a value depreciation assessment value. If the cumulative value of disturbance pressure is less than or equal to the current toughness residual value, the state parameters of the digital twin base model are updated, and the simulation of subsequent time nodes is continued based on the updated digital twin base model.

2. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 1, characterized in that, The data heterogeneous fusion module performs spatiotemporal alignment processing on the environmental state data and the geological subsurface data to construct a digital twin base model containing spatial coordinate information, including: Extract the first spatial coordinate information and the first timestamp information from the environmental state data; Extract the second spatial coordinate information and second timestamp information from the geological stratum data; Based on the first spatial coordinate information, the first timestamp information, the second spatial coordinate information, and the second timestamp information, the environmental state data and the geological bottom layer data are mapped to a preset three-dimensional grid coordinate system to generate the digital twin base model.

3. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 1, characterized in that, The buffer capacity calculation module, based on the digital twin foundation model, calculates the current resilience residual of the land ecosystem, including: Extract physical and biological attribute parameters corresponding to the environmental state data and the geological stratum data from the digital twin base model; The physical and biological attribute parameters are input into a preset system dynamics model to construct a set of material and energy cycle equations. Solve the aforementioned material energy cycle equations to calculate the current toughness residual value.

4. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 1, characterized in that, The nonlinear penetration simulation module, based on a time-series simulation algorithm, calculates the accumulation process of disturbance pressure within the land ecosystem of the preset disturbance event sequence, generating continuous time-series accumulated disturbance pressure values, including: The preset perturbation event sequence is analyzed in chronological order, and the perturbation intensity value corresponding to each perturbation event is extracted; The disturbance intensity value is input into a preset Markov chain model to determine the state transition probability in the time dimension, and combined with a preset cellular automata model to perform neighborhood evolution calculation in the spatial dimension, thereby simulating the spatiotemporal evolution state of the land ecosystem. Based on the spatiotemporal evolution state, calculate the nonlinear transmission path of the disturbance intensity value within the land ecosystem; The disturbance intensity value is accumulated along the nonlinear transmission path to obtain the cumulative disturbance pressure value.

5. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 1, characterized in that, The critical warning and value inversion module outputs system instability warnings, including: Based on the difference between the current resilience residual and the cumulative disturbance pressure, and the time rate of change of the cumulative disturbance pressure, the estimated remaining time of instability is calculated; Based on the estimated remaining time of instability and the spatial coordinate information of the digital twin base model, a vulnerability penetration heat map is generated; The output includes the estimated remaining time of instability and the vulnerability penetration heatmap, which is the system instability warning.

6. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 1, characterized in that, Obtain the restoration cost conversion factor corresponding to the land use type, and convert the cumulative disturbance pressure value into a value depreciation assessment value based on the restoration cost conversion factor, including: Based on the land use type, the corresponding restoration cost conversion coefficient is obtained from the preset economic evaluation database; The estimated recovery cost is calculated by multiplying the cumulative value of the disturbance pressure by the recovery cost conversion factor. The estimated recovery cost is output as the value depreciation assessment value.

7. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 3, characterized in that, Solving the aforementioned material-energy cycle equations to calculate the current resilience residual value includes: The material-energy cycle equations are processed using a pre-defined catastrophe theory algorithm to extract the critical point threshold of the material-energy cycle equations. Extract the current state value of the land ecosystem from the digital twin base model; The current resilience residual is determined based on the difference between the critical point threshold and the current state value.

8. The land ecological value spatiotemporal data simulation system based on digital twins according to claim 1, characterized in that, The preset perturbation event sequence includes human development planning data and climate anomaly event data.

Images