Wetland ecosystem dynamic evaluation method and system based on digital twinning

Abstract

The invention provides a dynamic evaluation method and system for a wetland ecosystem based on digital twinning, and relates to the technical field of ecological environment monitoring and evaluation. According to the method, a unified space framework is constructed, remote sensing time sequence data, in-situ monitoring data and patrol survey data are fused, the wetland digital twinborn body is established, and continuous updating of the ecological state is achieved. A comprehensive evaluation value and a grade conclusion are generated based on the twinborn state sequence, a risk early warning and driving attribution result is output in combination with a threshold library and a management target library, a measure scene parameter set is further introduced to carry out scene evaluation, and a quantitative basis is provided for wetland refined management and protection and decision making.

Description

Technical Field

[0001] This invention relates to the field of ecological environment monitoring and evaluation technology, and in particular to a method and system for dynamic assessment of wetland ecosystems based on digital twins. Background Technology

[0002] In engineering management and scientific research practice, wetland ecosystem assessment typically focuses on hydrological processes, vegetation cover and community structure, water quality changes, and habitat patterns. Data is acquired through methods such as remote sensing time-series interpretation, ground monitoring station observations, quadrat and cross-sectional surveys, and patrol records. This data is then combined with an indicator system to comprehensively determine and grade the wetland's ecological status. In some applications, specialized models for hydrology, hydrodynamics, water quality, or vegetation succession are also introduced to explain the causes of changes or assess the impact of restoration projects. While these technical approaches can support phased evaluations, project acceptance, and annual assessments, the overall work still primarily relies on an "offline aggregation, manual configuration, and periodic report generation" model.

[0003] With the increasing frequency of satellite remote sensing, the routine use of drone patrols, the intensive monitoring by the Internet of Things, and the enhanced capabilities of data processing platforms, wetland assessment is shifting from single data sources to multi-source fusion, from static conclusions to dynamic process characterization, and from post-event statistics to risk early warning and scheduling support. Simultaneously, management departments are paying more attention to the continuous impact of disturbance events (such as changes in water inflow, pollution impacts, and engineering controls) on wetland status, as well as the improvement paths and sustainability of different governance measures on key ecological elements. Driven by this demand, the concept of digital twins, which can simultaneously represent physical objects, process evolution, and be consistently updated with observational data, is gradually entering the field of ecological and environmental governance to support continuously updated state representation and scheme simulation assessment.

[0004] Existing wetland assessments still face common challenges in long-term engineering operation: inconsistencies in the spatiotemporal resolution and quality identification of multi-source data lead to unstable calibers after fusion; the configuration of indicator systems relies on experience, and adjustments to thresholds and weights make it difficult to uniformly interpret results across periods; process models often use one-time calibration, and parameter drift caused by seasonality and human intervention gradually deviates from the actual state, thus weakening the credibility of dynamic assessments; assessment outputs mostly remain at the level of comprehensive grading, making it difficult to form a driving force for management actions and a basis for selecting measures; assessment and the demonstration of governance schemes are often separated, lacking a closed-loop link for simulating and evaluating measures based on the same state expression.

[0005] Therefore, an integrated approach centered on digital twins and oriented towards dynamic updates and scheme simulation evaluation is needed to improve the continuity and consistency of dynamic assessment of wetland ecosystems. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for dynamic assessment of wetland ecosystems based on digital twins. By constructing a digital twin of the wetland and integrating multi-source observation data, the method enables continuous updating and dynamic assessment of the wetland's ecological status. At the same time, it outputs risk warnings, driver attribution, and measure scenario assessment results within a unified spatial framework, providing a reliable basis for the refined management and protection of wetlands.

[0007] To achieve the above objectives, the present invention provides the following solution:

[0008] A method for dynamic assessment of wetland ecosystems based on digital twins, comprising:

[0009] Acquire spatial boundary information and functional zoning information of the wetland under study, and construct an evaluation spatial unit set based on the spatial boundary information and the functional zoning information to obtain a basic spatial framework;

[0010] Collect remote sensing time-series data, in-situ monitoring data, and survey data covering the evaluation spatial unit set; spatiotemporally align the remote sensing time-series data, the in-situ monitoring data, and the survey data and aggregate them into the evaluation spatial unit set to obtain a unified observation dataset;

[0011] A wetland digital twin is constructed based on the aforementioned basic spatial framework and the aforementioned unified observation dataset, and a mapping relationship is established between the observed variables in the unified observation dataset and the state variables in the wetland digital twin; the wetland digital twin includes an object entity model, a process evolution model, and an evaluation rule base.

[0012] The unified observation dataset is input into the wetland digital twin according to a preset update cycle. Based on the mapping relationship, the parameters of the process evolution model are calibrated, the state variables are updated, and the twin state sequence is output.

[0013] The twin state sequence is input into the evaluation rule base to generate a comprehensive evaluation value and a grade conclusion, so as to obtain a dynamic evaluation result;

[0014] Obtain a threshold library and a management target library, compare the dynamic evaluation results with the threshold library and the management target library, and output risk warning results and driving attribution results;

[0015] A set of scenario parameters for measures is constructed, and the set of scenario parameters for measures is input into the wetland digital twin for scenario assessment, and the scenario assessment results are output; the set of scenario parameters for measures includes measure type identifiers and measure implementation parameters.

[0016] Preferably, an evaluation spatial unit set is constructed based on the spatial boundary information and the functional zoning information to obtain a basic spatial framework, including:

[0017] The spatial boundary information is processed to unify the coordinate reference.

[0018] Within the defined range of the spatial boundary information, spatial grid cells are generated according to a preset grid scale;

[0019] The spatial grid units and the functional partition information are spatially overlaid, and each spatial grid unit is assigned a partition identifier to generate the evaluation spatial unit set containing the spatial unit identifier and the partition identifier.

[0020] The evaluation spatial unit set and the spatial reference information corresponding to the evaluation spatial unit set constitute the basic spatial framework.

[0021] Preferably, remote sensing time-series data, in-situ monitoring data, and survey data covering the evaluation spatial unit set are collected. The remote sensing time-series data, in-situ monitoring data, and survey data are then spatiotemporally aligned and aggregated into the evaluation spatial unit set to obtain a unified observation dataset, including:

[0022] The image acquisition time is extracted from the remote sensing time series data and used as the remote sensing acquisition time information; the monitoring acquisition time is extracted from the in-situ monitoring data and used as the in-situ acquisition time information; the survey recording time is extracted from the patrol survey data and used as the patrol acquisition time information; the remote sensing acquisition time information, the in-situ acquisition time information, and the patrol acquisition time information together constitute the acquisition time information.

[0023] Image georegistration information is extracted from the remote sensing time-series data, monitoring point location information is extracted from the in-situ monitoring data, and survey location record information is extracted from the patrol survey data; the image georegistration information, the monitoring point location information, and the survey location record information together constitute spatial positioning information.

[0024] Set data source identifiers and quality identifiers for the remote sensing time series data, the in-situ monitoring data, and the survey data, respectively;

[0025] Based on the spatial positioning information, the remote sensing time series data, the in-situ monitoring data, and the survey data are collected into the spatial unit identifier corresponding to the evaluation spatial unit set;

[0026] Based on the acquisition time information, data collected to the same spatial unit identifier are time-aligned to generate the unified observation dataset; the unified observation dataset includes spatial unit identifier, acquisition time information, observation variable name, observation variable data, data source identifier, and quality identifier.

[0027] Preferably, a wetland digital twin is constructed based on the basic spatial framework and the unified observation dataset, and a mapping relationship is established between the observed variables in the unified observation dataset and the state variables in the wetland digital twin, including:

[0028] A wetland digital twin is constructed based on the aforementioned basic spatial framework and the aforementioned unified observation dataset; the wetland digital twin includes an object entity model, a process evolution model, and an evaluation rule base;

[0029] In the wetland digital twin, state variable names and spatial unit identifiers are set for state variables;

[0030] In the unified observation dataset, set the observation variable name, spatial unit identifier, and acquisition time information for the observation variables;

[0031] Establish variable correspondences; the variable correspondences include corresponding entries between the names of observed variables and the names of state variables;

[0032] Establish spatial unit matching rules; the spatial unit matching rules include matching entries between spatial unit identifiers in the unified observation dataset and spatial unit identifiers in the evaluation spatial unit set of the basic spatial framework;

[0033] Establish time matching rules; the time matching rules include matching entries between the acquisition time information in the unified observation dataset and the preset update cycle;

[0034] The variable correspondence, the spatial unit matching rule, and the time matching rule are used together as the mapping relationship.

[0035] Preferably, the application steps of the object entity model include:

[0036] A set of wetland object entities is established based on the aforementioned basic spatial framework; the set of wetland object entities consists of multiple wetland object entities.

[0037] Each wetland object entity is assigned an object entity identifier and an object type identifier; the object type identifier includes a water body type identifier, a vegetation type identifier, and a shoreline type identifier.

[0038] Spatial unit association information is set for each of the wetland object entities; the spatial unit association information includes the correspondence between the object entity identifier and at least one spatial unit identifier;

[0039] Configure a set of state variable names corresponding to each spatial unit identifier, and associate the set of state variable names with the object entity identifier that has the spatial unit association information, so that the twin state sequence can be output according to the spatial unit identifier and associated with the object entity identifier and the object type identifier.

[0040] Preferably, the unified observation dataset is input into the wetland digital twin according to a preset update cycle, the process evolution model is calibrated based on the mapping relationship, the state variables are updated, and the twin state sequence is output, including:

[0041] Within each preset update cycle, observation variable data corresponding to the state variable name is extracted from the unified observation dataset according to the mapping relationship, and the corresponding state variable data output by the process evolution model within the same preset update cycle is obtained.

[0042] Calculate the difference value based on the observed variable data and the corresponding state variable data;

[0043] A calibration threshold set is established in the wetland digital twin; the calibration threshold set contains corresponding entries of state variable names and difference thresholds;

[0044] When the difference value is greater than the difference threshold corresponding to the state variable name, the model parameters corresponding to the state variable name in the process evolution model are adjusted, and the state variable is updated;

[0045] The updated state variables are aggregated according to a preset update cycle to obtain the twin state sequence.

[0046] Preferably, the evaluation rule base includes an indicator rule set and a level discrimination rule set; the indicator rule set contains multiple indicator entries, each indicator entry containing an indicator item name, a corresponding set of state variable names, a standardized rule entry, and an indicator weight entry; the level discrimination rule set contains multiple level entries, each level entry containing a level name and a comprehensive evaluation value range; when the twin state sequence is input into the evaluation rule base, the comprehensive evaluation value is generated according to the indicator rule set, and the level conclusion is generated according to the level discrimination rule set.

[0047] Preferably, a threshold library and a management target library are obtained, and the dynamic evaluation results are compared with the threshold library and the management target library to output risk warning results and driving attribution results, including:

[0048] Establish a threshold library; the threshold library contains threshold entries corresponding to the indicator item names, and the threshold entries contain threshold ranges;

[0049] Establish a management target library; the management target library contains corresponding entries of partition identifiers and target ranges, wherein the partition identifier is the partition identifier used to characterize the functional partition in the functional partition information;

[0050] The comprehensive evaluation value and the level conclusion in the dynamic evaluation results are matched with the threshold library and the management target library respectively to generate the risk warning result;

[0051] Set a preset time window, where the preset time window is a continuous preset update cycle range corresponding to the risk warning result;

[0052] Within the preset time window, the change in state variable data in the twin state sequence between adjacent preset update cycles is calculated;

[0053] An attribution threshold set is established in the wetland digital twin; the attribution threshold set contains corresponding entries of state variable names and change thresholds;

[0054] State variable names whose changes are greater than the change threshold corresponding to the state variable name are selected to generate the driving attribution results.

[0055] Preferably, a set of scenario parameters for the measures is constructed, the set of scenario parameters is input into the wetland digital twin for scenario assessment, and the scenario assessment results are output, including:

[0056] Construct a set of scenario parameters for measures; the set of scenario parameters for measures includes a measure type identifier and measure implementation parameters; the measure implementation parameters include an action spatial unit identifier, action start time information, action end time information and action intensity parameters, wherein the action spatial unit identifier is the spatial unit identifier in the evaluation spatial unit set;

[0057] The scenario parameter set of the measures is input into the wetland digital twin, so that the process evolution model identifies the spatial unit of the action within the time information from the start time of the action to the end time of the action and updates the state variables according to the action intensity parameter to obtain the scenario twin state sequence.

[0058] The scenario twin state sequence is input into the evaluation rule base to generate the scenario evaluation result.

[0059] A dynamic assessment system for wetland ecosystems based on digital twins, comprising:

[0060] A spatial framework construction unit is used to acquire spatial boundary information and functional zoning information of the wetland under study, and to construct an evaluation spatial unit set based on the spatial boundary information and the functional zoning information to obtain a basic spatial framework.

[0061] The multi-source observation data fusion unit is used to collect remote sensing time-series data, in-situ monitoring data and survey data covering the evaluation spatial unit set, and to align the remote sensing time-series data, the in-situ monitoring data and the survey data in time and space and collect them into the evaluation spatial unit set to obtain a unified observation dataset.

[0062] The wetland digital twin construction and mapping unit is used to construct a wetland digital twin based on the basic spatial framework and the unified observation dataset, and to establish a mapping relationship between the observed variables in the unified observation dataset and the state variables in the wetland digital twin; the wetland digital twin includes an object entity model, a process evolution model, and an evaluation rule base.

[0063] The twin state update unit is used to input the unified observation dataset into the wetland digital twin according to a preset update cycle, calibrate the parameters of the process evolution model based on the mapping relationship, update the state variables, and output the twin state sequence.

[0064] The dynamic evaluation generation unit is used to input the twin state sequence into the evaluation rule base, generate a comprehensive evaluation value and a grade conclusion, so as to obtain the dynamic evaluation result;

[0065] The risk warning and attribution analysis unit is used to acquire a threshold library and a management target library, compare the dynamic evaluation results with the threshold library and the management target library, and output risk warning results and driving attribution results.

[0066] The scenario assessment unit is used to construct a set of scenario parameters for measures, input the set of scenario parameters for measures into the wetland digital twin for scenario assessment, and output the scenario assessment results; the set of scenario parameters for measures includes measure type identifiers and measure implementation parameters.

[0067] The present invention discloses the following technical effects:

[0068] This invention constructs an assessment spatial unit set under a unified basic spatial framework, and structurally integrates the spatial boundary information and functional zoning information of wetlands. This ensures that all subsequent observation data, state variables and assessment results are organized and expressed using the same spatial unit as the carrier, avoiding the problems of inconsistent spatial benchmarks of multi-source data and frequent adjustments of assessment units as projects change in existing technologies. This ensures the consistency and comparability of wetland ecological assessments at the spatial scale from the source.

[0069] This invention unifies and aggregates remote sensing time-series data, in-situ monitoring data, and survey data into a unified observation dataset. It retains multi-source observation information within the same data structure, overcoming the shortcomings of existing technologies where multi-source data are processed independently and are difficult to utilize collaboratively. This allows different observation methods to participate in the expression of wetland ecological status within the same time axis and spatial unit, improving the comprehensive ability of the assessment results to reflect actual ecological changes.

[0070] This invention constructs a wetland digital twin that includes an object entity model, a process evolution model, and an evaluation rule base, and establishes a mapping relationship between observed variables and state variables. This enables the observed data to continuously drive the twin's state update, effectively solving the problems of model-real-world disconnect and reliance on one-time calibration in existing technologies. It achieves continuous updating and dynamic expression of wetland ecological status, improving the timeliness and stability of evaluation results.

[0071] This invention generates comprehensive assessment values ​​and grade conclusions based on twin state sequences, and further combines threshold libraries and management target libraries to output risk warning results and driving attribution results. Compared with existing methods that only provide static grades or single indices, this invention can not only reflect the current state of wetland ecosystems, but also reveal the main driving factors of ecological risks, providing a more targeted basis for wetland management and restoration.

[0072] This invention introduces a set of scenario parameters for measures based on dynamic assessment and inputs them into a wetland digital twin for scenario assessment. This allows for the comparison of the effects of different management measures in the same twin environment before implementation, overcoming the problem of the separation between assessment and management decision-making in existing technologies. It realizes an integrated process from status assessment and risk warning to measure effect analysis, significantly enhancing the scientific nature and foresight of wetland ecological management. Attached Figure Description

[0073] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 A flowchart of the method provided in an embodiment of the present invention;

[0075] Figure 2 This is a schematic diagram of the system structure provided in an embodiment of the present invention. Detailed Implementation

[0076] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0077] The purpose of this invention is to provide a method and system for dynamic assessment of wetland ecosystems based on digital twins. On the basis of a unified assessment spatial unit, multi-source observation data is mapped to a wetland digital twin, realizing the integration of dynamic updates of ecological status, comprehensive assessment and scenario analysis, effectively supporting the scientific assessment and management decision-making of wetland ecosystems.

[0078] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0079] Figure 1 The method flowchart provided in the embodiments of the present invention is as follows: Figure 1 As shown, this invention provides a method for dynamic assessment of wetland ecosystems based on digital twins, comprising:

[0080] Step 100: Obtain spatial boundary information and functional zoning information of the wetland under study, and construct an evaluation spatial unit set based on the spatial boundary information and functional zoning information to obtain the basic spatial framework;

[0081] Step 200: Collect remote sensing time-series data, in-situ monitoring data and survey data covering the assessment spatial unit set, align the remote sensing time-series data, in-situ monitoring data and survey data in time and space and collect them into the assessment spatial unit set to obtain a unified observation dataset;

[0082] Step 300: Construct a wetland digital twin based on the basic spatial framework and unified observation dataset, and establish the mapping relationship between the observed variables in the unified observation dataset and the state variables in the wetland digital twin; the wetland digital twin includes an object entity model, a process evolution model, and an evaluation rule base;

[0083] Step 400: Input the unified observation dataset into the wetland digital twin according to the preset update cycle, calibrate the parameters of the process evolution model based on the mapping relationship, update the state variables, and output the twin state sequence;

[0084] Step 500: Input the twin state sequence into the evaluation rule base to generate a comprehensive evaluation value and a grade conclusion to obtain a dynamic evaluation result;

[0085] Step 600: Obtain the threshold library and management target library, compare the dynamic evaluation results with the threshold library and management target library, and output the risk warning results and driving attribution results;

[0086] Step 700: Construct a set of scenario parameters for the measures, input the set of scenario parameters into the wetland digital twin for scenario assessment, and output the scenario assessment results; the set of scenario parameters for the measures includes the measure type identifier and the measure implementation parameters.

[0087] Specifically, in this embodiment, the spatial boundary information and functional zoning information of the wetland under study are first obtained. The spatial boundary information is used to define the outer edge of the wetland in geographic space, and the functional zoning information is used to characterize the zoning range and zoning category of different management and ecological function areas within the wetland. In this embodiment, the spatial boundary information is processed to unify the coordinate reference, so that the spatial boundary information and the functional zoning information are under the same coordinate reference, thereby avoiding spatial overlay deviations caused by inconsistencies in the coordinate reference. In an optional embodiment, the coordinate reference unification process uses the same plane coordinate reference to complete the transformation and consistency verification, and the plane position error of the consistency verification is controlled within 1 meter to ensure the accuracy requirements of subsequent spatial grid generation and spatial overlay.

[0088] This embodiment generates spatial grid cells within the spatial boundary information defined by a unified coordinate reference, according to a preset grid scale. The spatial grid cell is the smallest evaluation unit after discretizing the spatial range of the study wetland, used to carry the subsequent collection of observation data, update of state variables, and output of evaluation results. The preset grid scale is the grid side length or grid resolution parameter used when generating the spatial grid cells. In one optional embodiment, the preset grid scale is set to 30 meters. In this embodiment, the grid is divided sequentially along the east-west and north-south directions of the spatial boundary information defined range, generating a set of spatial grid cells that covers the defined range and does not exceed it, and assigning a unique spatial cell identifier to each spatial grid cell. In another optional embodiment, the spatial cell identifier is numbered consecutively, ranging from 1 to 50000, to facilitate subsequent indexing and collection of the unified observation dataset and the unit-based organization and output of twin state sequences.

[0089] This embodiment spatially overlays the spatial grid units and the functional partition information, assigning a partition identifier to each spatial grid unit to generate an evaluation spatial unit set containing both spatial unit identifiers and partition identifiers. The partition identifier is identification information used to characterize the functional partition to which the spatial grid unit belongs, determined by the partition range and partition category recorded in the functional partition information. When a spatial grid unit overlaps with multiple functional partition ranges, this embodiment determines the partition identifier based on the coverage ratio of the spatial grid unit within each functional partition range, assigning the partition identifier corresponding to the functional partition with the largest coverage ratio to that spatial grid unit. In an optional implementation, when the coverage ratio of the functional partition with the largest coverage ratio is not less than 50%, the partition identifier corresponding to the functional partition with the largest coverage ratio is used as the partition identifier of that spatial grid unit. Further, this embodiment uses the evaluation spatial unit set and the spatial reference information corresponding to the evaluation spatial unit set to form a basic spatial framework. The spatial reference information is used to record the coordinate reference, preset grid scale, and arrangement relationship of the spatial grid units in the evaluation spatial unit set, thereby enabling the basic spatial framework to serve as the spatial foundation for subsequent unified observation dataset collection and wetland digital twin construction.

[0090] Specifically, in step 200 of this embodiment, remote sensing time-series data, in-situ monitoring data, and patrol survey data covering the assessment spatial unit set are collected to form an observational basis for subsequent updates of the wetland digital twin. Remote sensing time-series data consists of image sequences obtained by imaging the study wetland at different acquisition times. In-situ monitoring data consists of record sequences obtained by continuous or timed collection of hydrological, water quality, or vegetation-related elements from monitoring points deployed within the study wetland. Patrol survey data consists of record information generated from on-site patrols or sample point surveys conducted within the study wetland. In this embodiment, the image acquisition time is extracted from the remote sensing time-series data as remote sensing acquisition time information, the monitoring acquisition time is extracted from the in-situ monitoring data as in-situ acquisition time information, and the survey record time is extracted from the patrol survey data as patrol acquisition time information. The remote sensing acquisition time information, in-situ acquisition time information, and patrol acquisition time information are combined as acquisition time information. In an optional implementation, the acquisition time information is organized and recorded with the day as the smallest statistical unit, covering a continuous 365 days to ensure that the seasonal processes of wetland ecological changes are reflected.

[0091] This embodiment further extracts image georegistration information from remote sensing time-series data, monitoring point location information from in-situ monitoring data, and survey location record information from patrol survey data, and uses the image georegistration information, monitoring point location information, and survey location record information together as spatial positioning information. Image georegistration information is used to characterize the correspondence between remote sensing images and geographic space; monitoring point location information is used to characterize the location of monitoring points in geographic space; and survey location record information is used to characterize the geographical location where the patrol or survey occurred. In one optional embodiment, the monitoring point location information includes location records of no less than 10 monitoring points, and the patrol survey data includes no less than 20 survey location records to ensure coverage of spatial differences in the studied wetlands. This embodiment sets data source identifiers and quality identifiers for remote sensing time-series data, in-situ monitoring data, and patrol survey data, respectively. The data source identifier is used to distinguish whether the data comes from remote sensing time-series data, in-situ monitoring data, or patrol survey data, and the quality identifier is used to characterize the validity status of the data. In one optional embodiment, the quality identifier includes at least three states, used to characterize valid data, missing data, and abnormal data, respectively.

[0092] This embodiment aggregates remote sensing time-series data, in-situ monitoring data, and survey data into spatial unit identifiers corresponding to the evaluation spatial unit set based on spatial positioning information, so that data from different sources can form a unified representation within the same spatial unit. During aggregation, this embodiment establishes a correspondence between the image coverage area corresponding to the image georegistration information and the spatial unit identifier, the monitoring point location information and the spatial unit identifier, and the survey location record information and the spatial unit identifier. Subsequently, this embodiment performs time alignment on the data aggregated to the same spatial unit identifier based on the acquisition time information to generate a unified observation dataset. Time alignment is used to form aligned records of data from different sources under the same spatial unit identifier at the same time scale, thereby facilitating subsequent input into the wetland digital twin according to a preset update cycle. In an optional embodiment, the time scale of time alignment is consistent with the preset update cycle and is set to 7 days, so that the unified observation dataset forms corresponding aligned records within each preset update cycle. The unified observation dataset includes at least a spatial unit identifier, acquisition time information, observation variable name, observation variable data, data source identifier, and quality identifier. The observation variable name is used to characterize the wetland element name corresponding to the observation variable data.

[0093] Optionally, in step 300 of this embodiment, a wetland digital twin is constructed based on the basic spatial framework formed in step 100 and the unified observation dataset obtained in step 200. The wetland digital twin is used to uniformly represent the structure, state change process, and evaluation rules of wetland objects at the scale of evaluation spatial units. The wetland digital twin includes three parts: an object entity model, a process evolution model, and an evaluation rule base. In this embodiment, the evaluation spatial unit set already determined in the basic spatial framework serves as the spatial carrier of the wetland digital twin, enabling the wetland digital twin to maintain a corresponding set of state variables under each spatial unit identifier. In one specific implementation, the evaluation spatial unit set contains approximately 10,000 spatial unit identifiers, thereby ensuring the ability to characterize the spatial heterogeneity of the studied wetland.

[0094] In constructing a digital twin of a wetland, this embodiment first sets state variable names and spatial unit identifiers for state variables within the digital twin. State variables characterize the state features of the wetland ecosystem in different spatial units; state variable names distinguish different state items; and spatial unit identifiers indicate the assessment spatial unit to which the state variable belongs. This embodiment organizes state variables according to spatial unit identifiers, ensuring that each spatial unit identifier corresponds to at least six state variable names, thereby providing a basic state representation for subsequent generation of twin state sequences and dynamic assessment.

[0095] This embodiment further sets the observation variable name, spatial unit identifier, and collection time information for the observation variables in the unified observation dataset. The observation variable is a variable item in the unified observation dataset used to describe the category of observation data. The observation variable name is used to characterize the wetland element type corresponding to the observation data, the spatial unit identifier is used to indicate the evaluation spatial unit to which the observation variable data is collected, and the collection time information is used to indicate the collection time or time period of the observation variable data. In this embodiment, the observation variable name, spatial unit identifier, and collection time information are set for each observation record in the unified observation dataset. In one specific implementation, the time scale of the collection time information is set to 7 days to ensure that subsequent observation variables and state variables are matched under a unified time scale.

[0096] Based on this, this embodiment establishes a mapping relationship between observed variables and state variables. Specifically, this embodiment establishes a variable correspondence relationship to limit the one-to-one correspondence between the names of observed variables and the names of state variables, enabling observed variable data to be used to update or calibrate the corresponding state variables; in one specific implementation, the variable correspondence relationship contains no fewer than 6 corresponding entries. Simultaneously, this embodiment establishes spatial unit matching rules to limit the matching relationship between spatial unit identifiers in the unified observation dataset and spatial unit identifiers in the evaluation spatial unit set of the basic spatial framework, ensuring that observed variable data and state variables are associated under the same spatial unit identifier. Furthermore, this embodiment also establishes time matching rules to limit the matching relationship between acquisition time information and a preset update cycle, enabling observed variable data to be merged into the corresponding twin update cycle; in one specific implementation, the preset update cycle is also set to 7 days. This embodiment uses the variable correspondence relationship, spatial unit matching rules, and time matching rules to constitute a mapping relationship, and stores this mapping relationship in the wetland digital twin.

[0097] This embodiment further executes the application steps of the object entity model. First, a wetland object entity set is established based on the basic spatial framework. The wetland object entity set consists of multiple wetland object entities, which are used to abstractly represent entity objects in the studied wetland. In one specific implementation, the wetland object entity set contains no less than 300 wetland object entities. Subsequently, this embodiment sets an object entity identifier and an object type identifier for each wetland object entity. The object entity identifier is used to uniquely indicate the corresponding wetland object entity, and the object type identifier is used to characterize whether the wetland object entity belongs to the water body type, vegetation type, or shoreline type. Next, this embodiment sets spatial unit association information for each wetland object entity. The spatial unit association information includes the correspondence between the object entity identifier and at least one spatial unit identifier, which is used to characterize the spatial coverage of the wetland object entity. Finally, this embodiment configures a set of state variable names corresponding to each spatial unit identifier, and associates the set of state variable names with object entity identifiers that have spatial unit association information. This allows the twin state sequence to be output according to the spatial unit identifier, while also being associated with the object entity identifier and the object type identifier, providing object-level support for subsequent process evolution analysis and dynamic evaluation.

[0098] As an example, in this embodiment, when establishing the mapping relationship, the observation variable name and the state variable name are first matched one-to-one according to the variable correspondence relationship, and the unified observation dataset is collected into the "spatial unit identifier-update cycle" granularity according to the spatial unit matching rule and the time matching rule, so as to obtain the observation sequence for updating.

[0099]

[0100] in, To identify the spatial unit as Update time is The name of the observed variable is Observation set under the given conditions; To unify the first in the observation dataset Observational variable data; For the first The spatial unit identifier corresponding to each observation record; For the first The data collection time information corresponding to each observation record; The preset update cycle length; For the first The observation variable name identifier of each observation record, and the observation variable name identifier is associated with the state variable name identifier through the variable correspondence. correspond.

[0101] In this embodiment, to avoid the inability to directly compare data from multiple sources with different frequencies within the same update cycle, the observation set is filtered according to the quality identifier to form representative observations within the update cycle, which are then used for subsequent difference calculation and status update.

[0102]

[0103] in, To identify the spatial unit as Update time is The state variable name is identified as Representative periodic observations under certain conditions; The set of observations defined in the above formula; To compare with observed variable data The validity coefficient, derived from the corresponding quality label conversion, ranges from 0 to 1, where the quality label is valid data. When the quality identifier indicates missing data When the quality is identified as abnormal data Use a preset value between 0 and 1.

[0104] In this embodiment, the process evolution model outputs the predicted value of the state variable in each update cycle and calculates the difference with the representative observation of the cycle. When the difference exceeds the calibration threshold, this embodiment performs calibration update on the state variable, thereby outputting a twin state sequence.

[0105]

[0106] in, The difference value; Representative observations of the period; The process evolution model is identified as a spatial unit. Update time is The state variable name is identified as Predicted values ​​of state variables output under given conditions; The updated state variable values ​​are then entered into the twin state sequence; The calibration coefficient has a value ranging from 0 to 1 and is determined jointly by the calibration threshold set and the quality identifier, such that when... Not greater than and When the corresponding difference threshold is taken ,when Greater than and When the corresponding difference threshold is taken ,in To and The corresponding effectiveness coefficient.

[0107] In this embodiment, when the twin state sequence is input into the evaluation rule base to generate a comprehensive evaluation value and a grade conclusion, the indicator value is generated by using the indicator items defined by the indicator rule set as the unit, and the comprehensive evaluation value is obtained by equal weighting without introducing a large amount of manual weighting, and then the grade conclusion is output by the grade discrimination rule set.

[0108]

[0109] in, To identify the spatial unit as Update time is Comprehensive evaluation value under the given conditions; The number of indicator items is limited by the evaluation rule base; For the first Each indicator item is identified as a spatial unit. Update time is The index values ​​under the given conditions are obtained from the state variables in the twin state sequence according to the index definitions in the evaluation rule base; The conclusion is a graded conclusion; This is the interval mapping function corresponding to the set of level discrimination rules, used to map the comprehensive evaluation value to the level name in the preset level set.

[0110] Further, in step 400 of this embodiment, the unified observation dataset is input into the wetland digital twin according to a preset update cycle to drive state updates. In specific implementation, this embodiment uses the preset update cycle as the time organization unit, dividing the unified observation dataset into several consecutive time periods; in a preferred embodiment, the preset update cycle is set to 7 days. Within each preset update cycle, this embodiment extracts the observation variable data corresponding to each state variable name from the unified observation dataset according to the mapping relationship, and synchronously acquires the corresponding state variable data output by the process evolution model within the same preset update cycle; wherein, the mapping relationship is used to limit the correspondence between the observation variable name and the state variable name, as well as the matching relationship between the spatial unit identifier and the acquisition time information, thereby ensuring that the observation variable data and state variable data participating in the update are within the same spatial unit identifier and the same time period. In a specific embodiment, the number of observation variable data corresponding to the same spatial unit identifier and the same state variable name within a preset update cycle is not less than 3, in order to reduce the impact of single observation fluctuations on the update results.

[0111] After pairing the observed variable data with the state variable data, this embodiment merges the observed variable data within the same spatial unit identifier, the same state variable name, and the same preset update period to obtain the representative observed value within the preset update period. The representative observed value is then compared with the corresponding state variable data output by the process evolution model to calculate the difference value. The difference value characterizes the degree of deviation between the representative observed value and the output result of the process evolution model. This embodiment establishes a calibration threshold set in the wetland digital twin. The calibration threshold set contains a one-to-one correspondence between state variable names and difference thresholds. In one specific implementation, the calibration threshold set contains at least six difference threshold entries, each corresponding to a different state variable name, and the difference threshold values ​​are set between 0.05 and 0.30, enabling different state variables to trigger or suppress calibration operations under a unified discrimination logic.

[0112] When the difference value exceeds the difference threshold corresponding to the state variable name, this embodiment adjusts the model parameters corresponding to the state variable name in the process evolution model and updates the state variable based on the adjusted model parameters. The model parameters are a set of parameters in the process evolution model used to describe the evolution relationship of the state variable over time. Their adjustment is used to correct the model output towards the observed representative value. In one specific implementation, the magnitude of a single parameter adjustment is limited to no more than 10% of the original parameter value to avoid excessive correction leading to a decrease in model stability. After completing the parameter adjustment, this embodiment writes the updated state variable into the wetland digital twin and aggregates the continuously updated state variables according to a preset update cycle, thereby forming a twin state sequence. In one specific implementation, the twin state sequence contains at least 52 consecutive preset update cycle state variable records to support subsequent dynamic and scenario assessment processes.

[0113] Further, in step 500 of this embodiment, the twin state sequence obtained in step 400 is input into the evaluation rule base to generate a dynamic evaluation result. The evaluation rule base is used to unify the composition of constraint indicators, the generation criteria of indicator values, and the classification criteria. The evaluation rule base includes an indicator rule set and a classification rule set. The indicator rule set consists of multiple indicator entries, each of which defines the generation method of an indicator item. The indicator entry includes at least an indicator item name, a corresponding set of state variable names, a standardization rule entry, and an indicator weight entry. The indicator item name uniquely identifies the indicator item, the corresponding set of state variable names limits which state variables are extracted from the twin state sequence to participate in the generation of the indicator item, the standardization rule entry limits the unified processing criteria for state variables with different dimensions or different value ranges before generating the indicator item, and the indicator weight entry limits the contribution ratio of the indicator item in the comprehensive evaluation value. In one specific embodiment, the indicator rule set contains no fewer than six indicator entries, enabling the comprehensive evaluation value to cover the main state dimensions of the wetland ecosystem.

[0114] When generating the comprehensive evaluation value, this embodiment uses the spatial unit identifier and the preset update cycle as the organizational granularity. It extracts the set of state variable names corresponding to each indicator item from the twin state sequence, and standardizes these state variables according to the standardization rule entries in each indicator item to obtain indicator item values ​​at the same scale. The standardization rule entries are rules used to map the values ​​of state variables to a unified scale. This embodiment can use interval normalization to perform the standardization process, that is, mapping the values ​​of state variables to a fixed interval according to preset upper and lower limits. In one specific implementation, the fixed interval is set to 0 to 1, thereby making different indicator items comparable. Subsequently, this embodiment performs a weighted summation of the indicator item values ​​according to the indicator weight entries to generate a comprehensive evaluation value corresponding to the spatial unit identifier and the corresponding preset update cycle. In one specific implementation, the indicator weight entries adopt a constraint of non-negative weights and a weight sum of 1 to ensure the numerical scale of the comprehensive evaluation value is stable and facilitates cross-period comparison.

[0115] When generating the grading conclusion, this embodiment uses a grading discrimination rule set to perform interval discrimination on the comprehensive evaluation value. The grading discrimination rule set consists of multiple grading entries, each of which includes a grading name and a comprehensive evaluation value range. The grading name is used to characterize the grading category of the wetland's ecological state, and the comprehensive evaluation value range is used to define the range of comprehensive evaluation values ​​belonging to that grading name. This embodiment matches the comprehensive evaluation value with the comprehensive evaluation value range corresponding to each grading entry and outputs the grading name corresponding to the matched range as the grading conclusion, thereby forming a dynamic evaluation result. In one specific implementation, the grading discrimination rule set sets four grading entries, and the comprehensive evaluation value range from 0 to 1 is sequentially divided into continuous intervals of 0 to 0.25, 0.25 to 0.50, 0.50 to 0.75, and 0.75 to 1, so that the dynamic evaluation result can be output with a unified standard and used for subsequent risk warning and scenario assessment.

[0116] Furthermore, in step 600 of this embodiment, after obtaining the dynamic evaluation result output in step 500, a threshold library and a management target library are acquired, and the dynamic evaluation result is compared with the threshold library and the management target library to generate a risk warning result. The threshold library is used to define the threshold range of indicator items under different risk levels. When establishing the threshold library in this embodiment, the names of indicator items already determined in the indicator rule set are used as indexes, and a corresponding threshold entry is set for each indicator item name. The threshold entry contains a threshold range, so that different indicator items have their own risk discrimination criteria. In a specific implementation, the threshold library covers no less than 6 indicator item names, and the threshold range corresponding to each indicator item name is set to 3 consecutive intervals to achieve graded discrimination of risk level. The management target library is used to define the management target range of different functional zones. In this embodiment, when establishing the management target library, the partition identifier is extracted based on the functional zone information in step 100. The partition identifier is the partition identifier used to characterize the functional zone in the functional zone information. A target range is set for each partition identifier, so that the same dynamic evaluation result has a differentiated comparison benchmark under different functional zones. In a specific implementation, the management target library contains no less than 3 partition identifiers, and the target range of each partition identifier is in the range of 0 to 1.

[0117] When generating risk warning results, this embodiment uses spatial unit identifiers and preset update cycles as the organizational granularity, and performs matching and discrimination on the comprehensive evaluation value and level conclusion in the dynamic evaluation results. Specifically, this embodiment first matches the comprehensive evaluation value with the threshold range of the corresponding indicator item name in the threshold library to determine the threshold range in which the comprehensive evaluation value falls, and forms warning level information accordingly. At the same time, this embodiment compares the level conclusion with the target range of the corresponding partition identifier in the management target library. When the comprehensive evaluation value corresponding to the level conclusion does not meet the target range, a management deviation identifier is generated. The partition identifier is determined by the correspondence between spatial unit identifiers and functional partition information, so that each spatial unit identifier can obtain a corresponding partition identifier. In one specific implementation, when the comprehensive evaluation value falls into the lowest threshold range or the management deviation identifier appears continuously for two consecutive preset update cycles, this embodiment marks the risk warning result of the corresponding spatial unit identifier as a high-priority warning, so that subsequent attribution analysis can focus on high-risk areas.

[0118] When generating the driving attribution results, this embodiment sets a preset time window, which is a continuous preset update cycle range corresponding to the risk warning results; in one specific embodiment, the preset time window includes four consecutive preset update cycles. Within the preset time window, this embodiment calculates the change in state variable data in the twin state sequence for adjacent preset update cycles to characterize the intensity of change in state variables in the short term. To achieve a unified attribution discrimination standard, this embodiment establishes an attribution threshold set in the wetland digital twin, which includes corresponding entries for state variable names and change thresholds; in one specific embodiment, the attribution threshold set includes no fewer than six corresponding entries, and the change threshold value range is set to 0.10 to 0.30. This embodiment filters state variable names whose change is greater than the change threshold corresponding to the state variable name, and outputs the filtered state variable names as driving attribution results, thereby enabling the risk warning results to be associated with state variables that have changed significantly within the preset time window.

[0119] Furthermore, in step 700 of this embodiment, to achieve scenario assessment of management measures, this embodiment constructs a set of measure scenario parameters and inputs the set of measure scenario parameters into the wetland digital twin. The set of measure scenario parameters is used to parameterize the management measures to be taken. The set of measure scenario parameters includes measure type identifiers and measure implementation parameters; wherein, the measure type identifier is used to characterize the category of management measures, and the measure implementation parameters are used to characterize the scope, implementation period, and implementation intensity of the management measures. The measure implementation parameters include spatial unit identifiers, start time information, end time information, and intensity parameters. The spatial unit identifiers are spatial unit identifiers in the assessment spatial unit set, so that management measures can be limited to specific spatial units. In one specific embodiment, the number of spatial unit identifiers is set to be no less than 50 to cover multiple spatial unit identifiers within the same functional zone, and the interval between the start time information and the end time information is set to be no less than 28 days to meet the time scale requirements of wetland ecological response.

[0120] After constructing the scenario parameter set, this embodiment inputs the scenario parameter set into the wetland digital twin, enabling the process evolution model to update state variables based on the action intensity parameter and the spatial unit identifier within the action start time to action end time, thereby obtaining the scenario twin state sequence. The action intensity parameter characterizes the degree of influence of management measures on the update magnitude of state variables. In this embodiment, the action intensity parameter is used as an external driving condition for the process evolution model, causing the spatial unit identifier under action to undergo a state change matching the measure type identifier during the implementation period. In one specific implementation, the action intensity parameter is limited to the range of 0.10 to 0.50 to control the magnitude of the impact of scenario updates on state variables and avoid unrealistic excessive disturbances to the process evolution model. The scenario twin state sequence is a sequence of state variables output under the constraints of the scenario parameter set. Its organization is consistent with the twin state sequence, both using spatial unit identifiers and preset update cycles as the organizational granularity. In one specific implementation, the scenario twin state sequence covers more than four consecutive preset update cycles to ensure the comparability of scenario assessments.

[0121] This embodiment inputs the scenario twin state sequence into the evaluation rule base, generates a scenario comprehensive evaluation value based on the indicator rule set, and generates a scenario level conclusion based on the level discrimination rule set, thereby forming a scenario evaluation result. The scenario evaluation result is used to characterize the evaluation status of the wetland ecosystem under the constraints of the measure scenario parameter set. This embodiment further compares the scenario evaluation result with the dynamic evaluation result output in step 500 to form scenario change information. The scenario change information includes at least the change in the comprehensive evaluation value and the change in the level conclusion. In a specific implementation, when the scenario comprehensive evaluation value is improved by at least 0.05 or the level conclusion is improved by at least 1 level compared to the dynamic evaluation result, this embodiment marks the scenario evaluation result corresponding to the measure type identifier as the preferred scenario result, which is used to support the selection and implementation ranking of subsequent management measures.

[0122] Corresponding to the above methods, such as Figure 2 As shown, this embodiment also provides a wetland ecosystem dynamic assessment system based on digital twins, including:

[0123] A spatial framework construction unit is used to acquire spatial boundary information and functional zoning information of the wetland under study, and to construct an evaluation spatial unit set based on the spatial boundary information and the functional zoning information to obtain a basic spatial framework.

[0124] The multi-source observation data fusion unit is used to collect remote sensing time-series data, in-situ monitoring data and survey data covering the evaluation spatial unit set, and to align the remote sensing time-series data, the in-situ monitoring data and the survey data in time and space and collect them into the evaluation spatial unit set to obtain a unified observation dataset.

[0125] The wetland digital twin construction and mapping unit is used to construct a wetland digital twin based on the basic spatial framework and the unified observation dataset, and to establish a mapping relationship between the observed variables in the unified observation dataset and the state variables in the wetland digital twin; the wetland digital twin includes an object entity model, a process evolution model, and an evaluation rule base.

[0126] The twin state update unit is used to input the unified observation dataset into the wetland digital twin according to a preset update cycle, calibrate the parameters of the process evolution model based on the mapping relationship, update the state variables, and output the twin state sequence.

[0127] The dynamic evaluation generation unit is used to input the twin state sequence into the evaluation rule base, generate a comprehensive evaluation value and a grade conclusion, so as to obtain the dynamic evaluation result;

[0128] The risk warning and attribution analysis unit is used to acquire a threshold library and a management target library, compare the dynamic evaluation results with the threshold library and the management target library, and output risk warning results and driving attribution results.

[0129] The scenario assessment unit is used to construct a set of scenario parameters for measures, input the set of scenario parameters for measures into the wetland digital twin for scenario assessment, and output the scenario assessment results; the set of scenario parameters for measures includes measure type identifiers and measure implementation parameters.

[0130] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0131] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A wetland ecosystem dynamic evaluation method based on digital twinning, characterized in that, The method comprises the following steps: acquiring spatial boundary information and functional partition information of a wetland, constructing an evaluation spatial unit set based on the spatial boundary information and the functional partition information, and obtaining a basic spatial framework; collecting remote sensing time series data, in-situ monitoring data, and patrol survey data covering the evaluation spatial unit set, spatiotemporally aligning and collecting the remote sensing time series data, the in-situ monitoring data, and the patrol survey data to the evaluation spatial unit set, and obtaining a unified observation data set; constructing a wetland digital twin based on the basic spatial framework and the unified observation data set, and establishing a mapping relationship between observation variables in the unified observation data set and state variables in the wetland digital twin; the wetland digital twin comprises an object entity model, a process evolution model, and an evaluation rule base; inputting the unified observation data set into the wetland digital twin according to a preset update period, calibrating parameters of the process evolution model based on the mapping relationship, updating the state variables, and outputting a twin state sequence; inputting the twin state sequence into the evaluation rule base, generating a comprehensive evaluation value and a grade conclusion, and obtaining a dynamic evaluation result; acquiring a threshold library and a management target library, comparing the dynamic evaluation result with the threshold library and the management target library, and outputting a risk early warning result and a driving attribution result; constructing a measure scenario parameter set, inputting the measure scenario parameter set into the wetland digital twin for scenario evaluation, and outputting a scenario evaluation result; the measure scenario parameter set comprises a measure type identifier and a measure implementation parameter.

2. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, characterized in that, The method comprises the following steps: performing coordinate reference unification processing on the spatial boundary information; generating spatial grid units according to a preset grid scale within the range defined by the spatial boundary information; spatially overlaying the spatial grid units and the functional partition information, assigning a partition identifier to each spatial grid unit, and generating the evaluation spatial unit set comprising spatial unit identifiers and partition identifiers; combining the evaluation spatial unit set and spatial reference information corresponding to the evaluation spatial unit set to form the basic spatial framework.

3. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, characterized in that, The method comprises the following steps: extracting image acquisition times from the remote sensing time series data as remote sensing acquisition time information, extracting monitoring acquisition times from the in-situ monitoring data as in-situ acquisition time information, and extracting investigation record times from the patrol survey data as patrol acquisition time information; the remote sensing acquisition time information, the in-situ acquisition time information, and the patrol acquisition time information together constitute acquisition time information; extracting image geographical registration information from the remote sensing time series data, monitoring point position information from the in-situ monitoring data, and investigation position record information from the patrol investigation data; the image geographical registration information, the monitoring point position information, and the investigation position record information jointly constitute spatial positioning information; setting data source identifiers and quality identifiers for the remote sensing time series data, the in-situ monitoring data, and the patrol investigation data, respectively; according to the spatial positioning information, collecting the remote sensing time series data, the in-situ monitoring data, and the patrol investigation data to spatial unit identifiers corresponding to the evaluation spatial unit set; according to the collection time information, time-aligning the data collected to the same spatial unit identifier, and generating the unified observation data set; the unified observation data set includes spatial unit identifiers, collection time information, observation variable names, observation variable data, data source identifiers, and quality identifiers.

4. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, characterized in that, based on the basic spatial framework and the unified observation data set, constructing a wetland digital twin, and establishing a mapping relationship between observation variables in the unified observation data set and state variables in the wetland digital twin, including: based on the basic spatial framework and the unified observation data set, constructing a wetland digital twin; the wetland digital twin includes an object entity model, a process evolution model, and an evaluation rule base; setting state variable names and spatial unit identifiers for state variables in the wetland digital twin; setting observation variable names, spatial unit identifiers, and collection time information for observation variables in the unified observation data set; establishing a variable correspondence relationship; the variable correspondence relationship includes corresponding entries of observation variable names and state variable names; establishing a spatial unit matching rule; the spatial unit matching rule includes matching entries between spatial unit identifiers in the unified observation data set and spatial unit identifiers of the evaluation spatial unit set in the basic spatial framework; establishing a time matching rule; the time matching rule includes matching entries between collection time information in the unified observation data set and the preset update period; the variable correspondence relationship, the spatial unit matching rule, and the time matching rule jointly serve as the mapping relationship.

5. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, wherein, the application steps of the object entity model include: based on the basic spatial framework, establishing a wetland object entity set; the wetland object entity set is composed of multiple wetland object entities; setting object entity identifiers and object type identifiers for each wetland object entity; the object type identifier includes a water body type identifier, a vegetation type identifier, and a shoreline type identifier; setting spatial unit association information for each wetland object entity; the spatial unit association information includes a correspondence relationship between the object entity identifier and at least one spatial unit identifier; configuring a set of state variable names corresponding to the spatial unit identifier for each spatial unit identifier, and associating the set of state variable names with the object entity identifier having the spatial unit association information, so that the twin state sequence can be output according to the spatial unit identifier and associated to the object entity identifier and the object type identifier.

6. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, wherein, inputting the unified observation dataset into the wetland digital twin according to a preset update period, performing parameter calibration on the process evolution model based on the mapping relationship, updating the state variables, and outputting a twin state sequence, including: in each of the preset update periods, extracting observation variable data corresponding to the state variable name from the unified observation dataset according to the mapping relationship, and obtaining corresponding state variable data output by the process evolution model in the same preset update period; calculating a difference value based on the observation variable data and the corresponding state variable data; establishing a calibration threshold set in the wetland digital twin; the calibration threshold set includes corresponding entries of state variable names and difference threshold values; when the difference value is greater than the difference threshold value corresponding to the state variable name, adjusting the model parameters in the process evolution model corresponding to the state variable name, and updating the state variables; collecting the updated state variables according to the preset update period to obtain the twin state sequence.

7. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, wherein, The evaluation rule base includes an index rule set and a grade judgment rule set; the index rule set includes a plurality of index entries, each index entry includes an index item name, a corresponding set of state variable names, a standardization rule entry, and an index weight entry; the grade judgment rule set includes a plurality of grade entries, each grade entry includes a grade name and a comprehensive evaluation value interval; when the twin state sequence is input into the evaluation rule base, the comprehensive evaluation value is generated according to the index rule set, and the grade conclusion is generated according to the grade judgment rule set.

8. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 7, characterized in that, obtain a threshold library and a management target library, compare the dynamic evaluation result with the threshold library and the management target library, output a risk warning result and a driving attribution result, including: establishing a threshold library; the threshold library includes threshold entries corresponding to the index item names, and the threshold entries include threshold intervals; establishing a management target library; the management target library includes corresponding entries of partition identifiers and target intervals, and the partition identifier is a partition identifier used to represent a functional partition in the functional partition information; matching the comprehensive evaluation value in the dynamic evaluation result and the grade conclusion with the threshold library and the management target library respectively to generate the risk warning result; setting a preset time window, the preset time window is a range of continuous preset update periods corresponding to the risk warning result; in the preset time window, calculate the change amount of the state variable data in the twin state sequence in adjacent preset update periods; establishing an attribution threshold set in the wetland digital twin; the attribution threshold set includes corresponding entries of state variable names and change threshold values; screening state variable names with a change amount greater than the change threshold value corresponding to the state variable name to generate the driving attribution result.

9. The digital-twin-based dynamic assessment method of a wetland ecosystem according to claim 1, wherein, constructing a measure scenario parameter set, inputting the measure scenario parameter set into the wetland digital twin for scenario evaluation, and outputting a scenario evaluation result, including: constructing a measure scenario parameter set, the measure scenario parameter set comprising a measure type identifier and a measure implementation parameter, the measure implementation parameter comprising an action space unit identifier, an action start time information, an action end time information, and an action intensity parameter, the action space unit identifier being a space unit identifier in the evaluation space unit set; inputting the measure scenario parameter set into the wetland digital twin, and causing the process evolution model to update a state variable for the action space unit identifier and according to the action intensity parameter within the action start time information to the action end time information, to obtain a scenario twin state sequence; inputting the scenario twin state sequence into the evaluation rule base to generate the scenario evaluation result.

10. A wetland ecosystem dynamic assessment system based on digital twinning, characterized in that, The method comprises the following steps: a spatial framework construction unit is configured to obtain spatial boundary information and functional partition information of a study wetland, and construct an evaluation space unit set based on the spatial boundary information and the functional partition information to obtain a basic spatial framework; a multi-source observation data fusion unit is configured to collect remote sensing time series data, in-situ monitoring data, and patrol survey data covering the evaluation space unit set, and perform spatio-temporal alignment and collection of the remote sensing time series data, the in-situ monitoring data, and the patrol survey data to the evaluation space unit set to obtain a unified observation data set; a wetland digital twin construction and mapping unit is configured to construct a wetland digital twin based on the basic spatial framework and the unified observation data set, and establish a mapping relationship between an observation variable in the unified observation data set and a state variable in the wetland digital twin; the wetland digital twin comprises an object entity model, a process evolution model, and an evaluation rule base; a twin state updating unit is configured to input the unified observation data set into the wetland digital twin at a preset update period, calibrate parameters of the process evolution model based on the mapping relationship, update the state variable, and output a twin state sequence; a dynamic evaluation generation unit is configured to input the twin state sequence into the evaluation rule base to generate a comprehensive evaluation value and a level conclusion, to obtain a dynamic evaluation result; a risk early warning and attribution analysis unit is configured to obtain a threshold library and a management target library, compare the dynamic evaluation result with the threshold library and the management target library, and output a risk early warning result and a driving attribution result; a scenario evaluation unit is configured to construct a measure scenario parameter set, input the measure scenario parameter set into the wetland digital twin for scenario evaluation, and output a scenario evaluation result; the measure scenario parameter set comprises a measure type identifier and a measure implementation parameter.

Images