Big data-based ecological environment monitoring and early warning method and system

Abstract

The invention discloses an ecological environment monitoring and early warning method and system based on big data, and relates to the field of environment monitoring and early warning, and the method comprises the steps: collecting sea-land ecological monitoring data, and obtaining historical sea-land ecological data; completing index pair matching of all monitoring indexes, and dividing the target sea-land area into an intertidal zone area and a non-intertidal zone area based on an index pair matching result and sea-land ecological monitoring data; constructing an intertidal topological graph, and completing sub-region division of the intertidal zone region based on the intertidal topological graph to obtain a plurality of intertidal sub-regions; calculating ecological area parameters; and combining the sea-land ecological monitoring data and the ecological area parameters to generate an ecological environment early warning report of the target sea-land area. According to the invention, the precision of environment monitoring and early warning in the sea-land boundary area can be effectively improved.

Description

Technical Field

[0001] This application relates to the field of environmental monitoring and early warning, and in particular to an ecological environment monitoring and early warning method and system based on big data. Background Technology

[0002] With the increasing severity of global climate change and human activities, such as industrial pollution, land reclamation projects, and overfishing, the global ecological environment is facing unprecedented pressure. Especially in the land-sea interface region, which serves as a transitional zone between terrestrial and marine ecosystems, exhibiting characteristics of both aquatic and terrestrial ecosystems, it is a biodiversity-rich area and an ecologically sensitive zone. Its ecological environment directly affects the overall ecological balance of land and sea. However, due to the dual impacts of human activities and climate change, problems such as ecological degradation, habitat destruction, and water quality deterioration in the land-sea interface region are becoming increasingly prominent. Therefore, there is an urgent need for a method that can accurately perceive and provide risk warnings for changes in the ecological environment of the land-sea interface region. Existing ecological monitoring and early warning methods are mostly based on single indicators for assessment and warning. While this method can provide some data support, it often suffers from low warning accuracy and high false alarm and missed alarm rates. Summary of the Invention

[0003] This application provides a big data-based method and system for ecological environment monitoring and early warning, which can improve the accuracy of environmental monitoring and early warning in areas where land and sea meet.

[0004] To achieve the above objectives, the embodiments of this application adopt the following technical solutions: Firstly, a big data-based method for ecological environment monitoring and early warning is provided, which includes: Collect marine and terrestrial ecological monitoring data from all marine and terrestrial monitoring points in the target marine and terrestrial area, and obtain historical marine and terrestrial ecological data of the target marine and terrestrial area through a pre-constructed water environment database; Based on historical marine and terrestrial ecological data, complete the indicator pair matching of all monitoring indicators in the historical marine and terrestrial ecological data, and divide the target marine and terrestrial area into intertidal zone and non-intertidal zone based on the indicator pair matching results and marine and terrestrial ecological monitoring data. Based on marine and terrestrial ecological monitoring data, a multidimensional feature vector of the intertidal zone was constructed. The intertidal topology map was constructed by combining the multidimensional feature vector and all marine and terrestrial monitoring points in the intertidal zone. Based on the intertidal topology map, the intertidal zone was divided into sub-regions, resulting in multiple intertidal sub-regions. Calculate the ecological area parameters of the intertidal zone and all intertidal sub-regions within the target land-sea area; An ecological and environmental early warning report for the target marine and land area is generated by combining marine and terrestrial ecological monitoring data and ecological area parameters.

[0005] Optionally, the monitoring indicators include marine and terrestrial indicators. Based on historical marine and terrestrial ecological data, indicator pair matching is performed for all monitoring indicators in the historical marine and terrestrial ecological data. Based on the indicator pair matching results and marine and terrestrial ecological monitoring data, the target marine and terrestrial area is divided into intertidal and non-intertidal zones, including the following steps: Based on historical marine and terrestrial ecological data, the indicator pairs of all marine and terrestrial indicators in the historical marine and terrestrial ecological data were matched to obtain multiple marine indicator pairs and terrestrial indicator pairs. For any pair of marine indicators and pair of terrestrial indicators, extract the historical marine indicator pair data and historical terrestrial indicator pair data corresponding to the marine indicator pair and terrestrial indicator pair data from the historical marine and terrestrial ecological data, respectively. Based on the Gaussian mixture model, marine index distribution models and land index distribution models were constructed respectively. Historical ocean indicators are used to train the ocean indicator distribution model on the data, and multiple ocean component parameter sets are output. For any set of ocean component parameters, a two-dimensional confidence ellipse is defined based on the set of ocean component parameters and using the pre-obtained chi-square distribution critical value; Determine the marine indicator confidence intervals for marine indicator pairs based on the union of all two-dimensional confidence ellipses; A land index distribution model is trained using historical land index data, and multiple land component parameter sets are output through the land index distribution model after the model training is completed. The confidence intervals of land index pairs are calculated based on the land component parameter set and the Mahalanobis distance detection algorithm. By combining the confidence intervals of marine indicators, the confidence intervals of terrestrial indicators, and marine and terrestrial ecological monitoring data, the target marine and terrestrial area is divided into intertidal zone and non-intertidal zone.

[0006] Optionally, based on historical marine and terrestrial ecological data, the indicator pairs of all marine and terrestrial indicators in the historical marine and terrestrial ecological data are matched to obtain multiple marine indicator pairs and terrestrial indicator pairs, including the following steps: Extract historical terrestrial ecological data and historical marine ecological data from historical terrestrial and marine ecological data; Time-series analysis was performed on historical terrestrial ecological data and historical marine ecological data to obtain historical terrestrial sequences and historical marine sequences. The distribution statistics of all ocean indicators in the historical ocean sequence were performed, and the ocean fluctuation sequence of all ocean indicators was calculated based on the distribution statistics results. The land fluctuation sequence of all land indicators in the historical land sequence was calculated based on the clustering algorithm; The volatility synchronicity among all land indicators is calculated based on the land volatility sequence, and the indicator pair matching among all land indicators is completed based on the volatility synchronicity to obtain several land indicator pairs. Based on the ocean wave sequence, index pair matching was completed among all ocean indices, resulting in several ocean index pairs.

[0007] Optionally, calculating the land fluctuation sequence of all land indicators in the historical land sequence based on the clustering algorithm includes the following steps: For any land index, the historical land sequence is divided into all land index subsequences according to time order; For any land indicator subsequence, the land indicator subsequence is divided according to a preset time window to obtain multiple indicator datasets; For any index dataset, a clustering algorithm is used to perform data clustering of the index dataset, and high-density clusters are selected from the data clustering results as core distribution clusters; The land index interval is defined based on the percentile range of the core distribution cluster; Kernel density estimation is performed on the core distribution cluster of the indicator dataset to obtain the indicator probability density function of the indicator dataset; Based on the time window division order, multiple neighborhood datasets of the indicator dataset are determined, the average probability density function of all neighborhood datasets is calculated, and the basic volatility of the indicator dataset is calculated by combining the indicator probability density function and the average probability density function. Calculate the mean base volatility and the mean width of the interval for all indicator datasets to obtain the mean volatility and the mean interval width. Extract the center point of the land index interval of all index datasets, and calculate the center drift between all adjacent index datasets based on the center point extraction results; The land fluctuation sequence of all land indicators is calculated by combining the mean of all fluctuations, the mean of interval width, and the center drift.

[0008] Optionally, dividing the target marine and terrestrial area into intertidal and non-intertidal zones by combining marine indicator confidence intervals, terrestrial indicator confidence intervals, and marine and terrestrial ecological monitoring data includes the following steps: For any land and sea monitoring point, extract all land and sea monitoring data corresponding to all land indicator pairs and all sea indicator pairs from the land and sea ecological monitoring data of the land and sea monitoring point; The distribution characteristics of all ocean monitoring data within the corresponding ocean indicator confidence intervals were statistically determined, and the ocean deviation of the land and sea monitoring points was calculated based on the distribution characteristics. The distribution characteristics of all land-to-land monitoring data within the corresponding land index confidence intervals are statistically determined, and the land deviation of the land-sea monitoring points is calculated based on the distribution characteristics. By combining the ocean deviation and land deviation of all ocean and land monitoring points, the target ocean and land area is divided into intertidal zone and non-intertidal zone, wherein the non-intertidal zone includes ocean area and land area.

[0009] Optionally, a multi-dimensional feature vector of the intertidal zone is constructed based on marine and terrestrial ecological monitoring data. An intertidal topology map is then constructed by combining the multi-dimensional feature vector with all marine and terrestrial monitoring points within the intertidal zone. Based on this topology map, the intertidal zone is further divided into sub-regions, resulting in multiple intertidal sub-regions. This process includes the following steps: Extract intertidal monitoring data from all land and sea monitoring points within the intertidal zone; Time-series analysis was performed on all intertidal monitoring data to obtain multiple intertidal sequences; Tidal cycle data for the target land and sea areas were obtained from a water environment database. Align the timestamps of all intertidal sequences, divide all the intertidal sequences with aligned timestamps into time series based on tidal cycle data, and construct multiple tidal zone datasets located in different tidal time periods based on the time series division results. For any tidal zone dataset, complete the multidimensional feature extraction of the tidal zone dataset to obtain the tidal multidimensional features of all land and sea monitoring points in the intertidal zone; All land and sea monitoring points in the intertidal zone are used as topological nodes, and all adjacent topological nodes are connected to construct an intertidal topological map. Based on the multidimensional tidal features, the node feature similarity between all adjacent topological nodes is calculated using the similarity formula, and the edge weights of the intertidal topological graph are determined based on the node feature similarity. Arrange the intertidal topology maps with all edge weights assigned in chronological order to obtain the intertidal topology map sequence; The dynamic community detection algorithm is used to search for the nodes of the intertidal topology sequence and obtain the community affiliation sequence of all topological nodes. Based on the community affiliation sequence, the community stability and community stability period of all topological nodes are analyzed, and the intertidal zone affiliation of all topological nodes is determined by combining the community stability and community stability period. The intertidal zone is divided into multiple intertidal sub-regions based on the intertidal zone affiliation of all topological nodes. The intertidal sub-regions include the landside transition zone, the core transition zone, and the seaside transition zone.

[0010] Optionally, generating an ecological and environmental early warning report for the target marine and terrestrial area by combining marine and terrestrial ecological monitoring data and ecological area parameters includes the following steps: A marine-terrestrial ecological supermap of the target marine-terrestrial region was constructed based on marine-terrestrial ecological monitoring data. Based on historical marine and terrestrial ecological data and using the Granger causality test algorithm, the ecological causal relationships in the marine and terrestrial ecological hypergraph were identified. Based on the ecological causal relationships and the pre-constructed marine and terrestrial ecological evolution model, the ecological pressure parameter sequence of all ecological supernodes in the marine and terrestrial ecological hypergraph was calculated. The delineation results of various marine and land sub-regions within the target marine and land area divide the ecological pressure parameter sequence into multiple regional pressure parameter sequences. For any regional pressure parameter sequence, the multidimensional pressure features of the regional pressure parameter sequence are extracted. The multidimensional pressure features include pressure trend features, pressure steady-state features, and pressure propagation features. Ecological risk assessment of marine and terrestrial sub-regions was completed based on multidimensional pressure characteristics, and ecological early warning levels and pressure propagation paths of marine and terrestrial sub-regions were generated based on the ecological risk assessment results. The ecological early warning levels and pressure propagation paths of all marine and terrestrial sub-regions are integrated to generate ecological and environmental early warning reports for the target marine and terrestrial regions.

[0011] Optionally, constructing a marine-terrestrial ecological supermap of the target marine-terrestrial region based on marine-terrestrial ecological monitoring data includes the following steps: Each marine and terrestrial sub-region within the target marine and terrestrial area is defined as an ecological supernode. The marine and terrestrial sub-regions include marine areas, terrestrial areas, landside transition areas, core transition areas, and seaside transition areas. Extract non-intertidal and intertidal monitoring data from marine and terrestrial ecological monitoring data; The non-intertidal ecological pressure index and non-intertidal resource index of the corresponding ecological supernodes in the non-intertidal zone are calculated based on non-intertidal monitoring data. The intertidal ecological pressure index and intertidal resource index of the intertidal region were calculated by combining intertidal monitoring data and ecological area parameters. The non-intertidal ecological pressure index, non-intertidal resource index, intertidal ecological pressure index, and intertidal resource index are integrated into supernode attribute parameters, and the supernode attribute parameters are mapped to each ecological supernode. Based on marine and terrestrial ecological monitoring data and historical marine and terrestrial ecological data, regional ecological interaction chains between all marine and terrestrial sub-regions were extracted, and the regional ecological interaction chains were defined as ecological superborders. Among them, regional ecological interaction chains include pollutant runoff chains, climate impact chains, and biological migration chains. By combining the non-intertidal ecological pressure index, the non-intertidal resource index, the intertidal ecological pressure index, and the intertidal resource index, the hyperedge attribute parameters of all ecological hyperedges are calculated, and the hyperedge attribute parameters are mapped to the corresponding ecological hyperedges. A marine-terrestrial ecological supergraph is constructed based on all ecological supernodes that have completed supernode attribute parameter mapping and all ecological superedges that have completed superedge attribute parameter mapping.

[0012] Optionally, based on historical marine and terrestrial ecological data and using the Granger causality test algorithm to identify ecological causal relationships in the marine and terrestrial ecological hypergraph, and based on these ecological causal relationships and a pre-constructed marine and terrestrial ecological evolution model, the ecological pressure parameter sequence of all ecological supernodes in the marine and terrestrial ecological hypergraph is calculated, including the following steps: Based on historical marine and terrestrial ecological data, the historical ecological pressure index and historical ecological activity index of all ecological supernodes in the marine and terrestrial ecological supermap were calculated respectively. By combining historical stress index and historical activity index and using Granger causality test algorithm, ecological causal relationships between all ecological supernodes and all ecological superedges are identified. Based on ecological causality, the ecological propagation parameters between all ecological hyperedges and all ecological supernodes are calculated using a vector autoregression model. Historical pollution time data of the target marine and land areas were obtained from the water environment database. A pollution attenuation model of the target marine and land areas was fitted by combining the historical pollution time data and the historical ecological pressure index. The non-intertidal ecological pressure index or intertidal ecological pressure index of all ecological supernodes is input into the pollution attenuation model, and the ecological restoration parameters of all ecological supernodes are output sequentially through the pollution attenuation model. Ecological propagation parameters and ecological restoration parameters are mapped onto a marine-terrestrial ecological supergraph, and the mapped marine-terrestrial ecological supergraph is input into a pre-constructed marine-terrestrial ecological evolution model. Based on the marine-terrestrial ecological evolution model, the ecological pressure parameter sequence of all ecological supernodes is iteratively calculated. The marine-terrestrial ecological evolution model is constructed based on a supergraph propagation dynamics model.

[0013] Secondly, this application provides an ecological environment monitoring system based on big data, comprising: The memory is configured to store instructions; and The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the big data-based ecological environment monitoring and early warning method according to the first aspect.

[0014] The above technical solution first completes the matching of all monitoring indicators using historical marine and terrestrial ecological data. Based on the matching results, it captures the differences in ecological environment characteristics between non-intertidal and intertidal areas, laying a data foundation for delineating the non-intertidal and intertidal zones of the target marine and terrestrial region. By dividing the target marine and terrestrial regions, regional early warning can be implemented. This is because, within the target marine and terrestrial region, the intertidal zone has lower ecological stability than the non-intertidal zone, making it more susceptible to human activities and natural factors. Furthermore, there are significant differences in the ecological environment between the intertidal and non-intertidal zones. Therefore, different early warning standards and parameter types are needed for regional early warning to prevent underestimation of intertidal risks or over-warning of non-intertidal risks. Simultaneously, when abnormal ecological conditions occur in the target marine and terrestrial region, the affected areas can be quickly identified, and targeted governance measures can be implemented. In summary, this application can effectively improve the accuracy of environmental monitoring and early warning in the land-sea boundary area, thereby enabling timely containment of further deterioration of ecological problems in the target land-sea area at the initial stage of risk, and reducing the cost of ecological governance in the target land-sea area.

[0015] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a big data-based ecological environment monitoring and early warning method provided in this application embodiment; Figure 2 A flowchart illustrating a method for dividing intertidal and non-intertidal zones according to an embodiment of this application; Figure 3 This is a flowchart illustrating an ecological environment early warning method for a target land and sea area, provided as an embodiment of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0018] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0019] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0020] Figure 1 The illustration shows a flowchart of a big data-based ecological environment monitoring and early warning method according to an embodiment of this application. Figure 1 As shown in the figure, this application provides an ecological environment monitoring and early warning method based on big data, which may include the following steps: S101. Collect marine and terrestrial ecological monitoring data from all marine and terrestrial monitoring points in the target marine and terrestrial area, and obtain historical marine and terrestrial ecological data of the target marine and terrestrial area through a pre-constructed water environment database. S102. Based on historical marine and terrestrial ecological data, complete the indicator pair matching of all monitoring indicators in the historical marine and terrestrial ecological data, and divide the target marine and terrestrial area into intertidal zone and non-intertidal zone based on the indicator pair matching results and marine and terrestrial ecological monitoring data. S103. Based on marine and terrestrial ecological monitoring data, a multidimensional feature vector of the intertidal zone is constructed. The intertidal topology map is constructed by combining the multidimensional feature vector and all marine and terrestrial monitoring points in the intertidal zone. Based on the intertidal topology map, the sub-regions of the intertidal zone are divided to obtain multiple intertidal sub-regions. S104. Calculate the ecological area parameters of the intertidal zone and all intertidal sub-regions within the target land and sea area; S105. Generate an ecological and environmental early warning report for the target marine and land areas by combining marine and terrestrial ecological monitoring data and ecological area parameters.

[0021] In this embodiment, the target marine-terrestrial area refers to the transition zone between land and sea. Marine-terrestrial ecological monitoring data includes soil moisture content, vegetation cover, soil pH, water temperature, benthic species abundance, and sediment particle size. A marine-terrestrial monitoring point is the smallest sampling unit within the target marine-terrestrial area. At each monitoring point, various sensors and sampling devices are deployed to collect marine-terrestrial ecological monitoring data for a specific sampling area within the target marine-terrestrial area. The closer the monitoring point is to the coastline, the smaller its area. The water environment database stores historical marine-terrestrial ecological data, historical ecological early warning information, and historical environmental governance information for the target marine-terrestrial area.

[0022] Historical terrestrial and marine ecological data from the historical land-sea ecological data are arranged chronologically to obtain historical terrestrial and marine sequences. Then, using the same steps as calculating the terrestrial fluctuation sequence, the mean fluctuation, mean interval width, and center drift of the marine indicators are calculated. After normalization, a weighted sum is calculated to obtain the marine fluctuation sequence of the marine indicators. Principal component analysis or entropy weighting can be used to assign weights, or equal weights can be directly assigned. Next, the fluctuation synchronicity among all terrestrial indicators is calculated based on the terrestrial fluctuation sequence. If the fluctuation synchronicity between two terrestrial indicators is greater than a preset synchronicity threshold, the two terrestrial indicators are matched as a terrestrial indicator pair. The same method is used to calculate the fluctuation synchronicity among all marine indicators. If the fluctuation synchronicity between two marine indicators is greater than a preset synchronicity threshold, the two marine indicators are matched as a marine indicator pair.

[0023] Historical marine and terrestrial indicator pairs were extracted from historical marine and terrestrial ecological data. Distribution models for these pairs were constructed based on Gaussian mixture models. Outlier removal was then performed on both sets of historical marine and terrestrial indicator pairs to obtain historical sample data. The marine indicator distribution model was trained using these historical sample data to obtain the final model parameters, i.e., the set of marine component parameters for each marine indicator pair. This set includes the number of Gaussian components, mean, covariance matrix, and component weights of the marine indicator distribution model. A two-dimensional confidence ellipse was constructed for each marine indicator pair using these parameter sets, and this ellipse was used as the confidence interval for each pair. Similarly, a terrestrial indicator distribution model was trained using historical sample data. The resulting terrestrial indicator distribution model outputs a set of terrestrial component parameters, which was then used to construct a two-dimensional confidence ellipse for each terrestrial indicator pair. This ellipse was then used as the confidence interval for each terrestrial indicator pair. Based on the confidence intervals of marine and terrestrial indicators, it is determined whether the sampling areas corresponding to all marine and terrestrial monitoring points are intertidal zones. For each sampling area that has completed the intertidal zone determination, an intertidal zone probability distribution map is generated using Kriging interpolation. Then, spline curves are used to fit and connect the boundaries to form continuous intertidal zones, terrestrial zones, and marine zones.

[0024] Intertidal monitoring data collected from all land and sea monitoring points within the intertidal zone was extracted from the land and sea monitoring data. The intertidal monitoring data was arranged chronologically to obtain multiple intertidal sequences. An intertidal sequence refers to the time-series data of intertidal zones collected from each land and sea monitoring point within the intertidal zone. Next, the timestamps of all intertidal sequences were aligned, and the data was divided chronologically using a preset first time window to obtain multiple intertidal sub-sequences. Then, intertidal sub-sequences located within the same first time window but collected from different land and sea monitoring points were integrated into a tidal zone dataset. For each tidal zone dataset, feature extraction was performed on each intertidal sub-sequence, extracting features such as variance, extreme values, linear trend slope, and number of abrupt change points for each monitoring indicator. These features were then normalized and concatenated to form the tidal multidimensional features of each land and sea monitoring point. Finally, all land and sea monitoring points within the intertidal zone were used as topological nodes, and all adjacent topological nodes were connected to construct an intertidal topology map.

[0025] After constructing the intertidal topology map by connecting all adjacent topological nodes, the node feature similarity between all adjacent topological nodes is calculated based on tidal multidimensional features and a similarity formula. Edge weights are then assigned to the topological edges connecting two adjacent topological nodes based on node feature similarity; the higher the node feature similarity, the higher the assigned edge weight. The intertidal topology maps with all edge weights assigned are arranged chronologically to construct an intertidal topology map sequence. Next, a dynamic community detection algorithm is used to identify and track the evolution of community structures in the intertidal topology map sequence, recording the community to which each land-sea monitoring point belongs during each tidal period. Finally, the community affiliation sequence of all topological nodes is output. The community affiliation sequence contains which intertidal subregion each topological node belongs to during each tidal period. Commonly used dynamic community detection algorithms include the Leiden algorithm, multi-target immune algorithm, and label propagation method (LFM). The intertidal zone affiliation of each topological node is determined based on the community affiliation sequence. The intertidal zone affiliation corresponding to the topological nodes is mapped to the corresponding land and sea monitoring points to form an intertidal zone distribution map. The community boundaries in the intertidal zone distribution map are fitted using spline curves to form continuous landside transition areas, core transition areas, and seaside transition areas.

[0026] Based on the intertidal zone delineation results, ecological area parameters for the intertidal region are calculated. These parameters include the intertidal zone area and sub-region area parameters, with each sub-region's area and percentage of its total area included. Finally, an ecological and environmental early warning report for the target marine and terrestrial region is generated by combining marine and terrestrial ecological monitoring data and these ecological area parameters. The report includes the ecological warning level and pressure propagation path for all disturbed or degraded stable-state marine and terrestrial sub-regions, as well as marine and terrestrial ecological monitoring data for all sub-regions (including those in a healthy stable-state).

[0027] The above methods enable regional environmental early warning for the target land and sea areas, allowing for targeted assessment of the health status of each area within the target land and sea region, thus avoiding bias in the overall assessment. Furthermore, the target land and sea area is a transitional zone between land and sea, and its intertidal zone has a more fragile and sensitive ecosystem compared to terrestrial and marine areas. If the intertidal zone and non-intertidal zone are mixed for overall ecological assessment, it is possible that the target land and sea area's ecological environment is deemed acceptable, but in reality, only the non-intertidal zone's ecological environment may be acceptable, while the intertidal zone's ecological environment may already be abnormal. This could lead to missed monitoring of early ecological anomalies, resulting in missed opportunities for ecological environment remediation and indirectly increasing the cost of subsequent ecological environment remediation.

[0028] In one embodiment, the monitoring indicators include marine indicators and terrestrial indicators. The process involves matching all monitoring indicators from historical marine and terrestrial ecological data with their corresponding indicator pairs, and then dividing the target marine and terrestrial area into intertidal and non-intertidal zones based on the matching results and the marine and terrestrial ecological monitoring data. This includes the following steps: Based on historical marine and terrestrial ecological data, the indicator pairs of all marine and terrestrial indicators in the historical marine and terrestrial ecological data were matched to obtain multiple marine indicator pairs and terrestrial indicator pairs. For any pair of marine indicators and pair of terrestrial indicators, extract the historical marine indicator pair data and historical terrestrial indicator pair data corresponding to the marine indicator pair and terrestrial indicator pair data from the historical marine and terrestrial ecological data, respectively. Based on the Gaussian mixture model, marine index distribution models and land index distribution models were constructed respectively. Historical ocean indicators are used to train the ocean indicator distribution model on the data, and multiple ocean component parameter sets are output. For any set of ocean component parameters, a two-dimensional confidence ellipse is defined based on the set of ocean component parameters and using the pre-obtained chi-square distribution critical value; Determine the marine indicator confidence intervals for marine indicator pairs based on the union of all two-dimensional confidence ellipses; A land index distribution model is trained using historical land index data, and multiple land component parameter sets are output through the land index distribution model after the model training is completed. The confidence intervals of land index pairs are calculated based on the land component parameter set and the Mahalanobis distance detection algorithm. By combining the confidence intervals of marine indicators, the confidence intervals of terrestrial indicators, and marine and terrestrial ecological monitoring data, the target marine and terrestrial area is divided into intertidal zone and non-intertidal zone.

[0029] In this embodiment, the ocean fluctuation sequence of the ocean indicator and the land fluctuation sequence of the land indicator are first calculated based on historical marine and land ecological data. Then, the indicator pairs of all ocean indicators and land indicators are matched based on the calculated ocean fluctuation sequence and land fluctuation sequence of the land indicator, resulting in multiple ocean indicator pairs and land indicator pairs.

[0030] Extract historical marine indicator pairs and historical terrestrial indicator pairs from historical marine and terrestrial ecological data. For example, historical marine indicator pairs and historical terrestrial indicator pairs from the past many years. Historical marine indicator pairs refer to historical monitoring information of two marine indicators, such as historical monitoring information of soil moisture content and indicator coverage. Historical terrestrial indicator pairs refer to historical monitoring information of two terrestrial indicators, such as historical monitoring information of dissolved oxygen in water and benthic biodiversity index.

[0031] Based on Gaussian mixture models, distribution models for marine and land indicators were constructed. Gaussian mixture models are statistically based probability models used to represent datasets composed of multiple Gaussian distributions. Next, outlier removal was performed on historical marine and land indicator pairs (using the 3σ rule or IQR quartile method) to obtain historical marine and land indicator pair samples. The marine indicator distribution model was trained using these historical marine indicator pair samples. The training of the Gaussian mixture model primarily utilizes the Expectation-Maximization (EM) algorithm for parameter estimation. Its core steps include: first, determining the number of Gaussian components K, which can be selected using the Bayesian Information Criterion (BIC) or the Akaike Information Criterion (AIC). A smaller BIC / AIC value results in a higher model fit and lower complexity; for example, K=2. Then, an initialization step is performed, using the K-means algorithm to cluster the historical marine indicator pair samples, obtaining the initial K cluster centers, which serve as the marine indicator distribution model. The initial mean of each Gaussian component is calculated. Then, the E-step (expectation step) and M-step (maximization step) are performed to complete the iterative training of the ocean index distribution model. The expectation step refers to calculating the posterior probability of each data point in the historical ocean index pair belonging to each Gaussian component. The maximization step refers to updating the mean, covariance, and weight of each Gaussian component. When the change in the log-likelihood value is less than 1e-6, the iteration stops, and the final model parameters are obtained, that is, the ocean component parameter set of each ocean index pair. The ocean component parameter set includes the number of Gaussian components, mean, covariance matrix, and component weights of the ocean index distribution model.

[0032] First, determine the proportion of data contained in the two-dimensional confidence ellipse. A 95% confidence level can be chosen, resulting in a chi-square distribution critical value (or the threshold for the squared Mahalanobis distance) of 5.991. This critical value can be obtained by consulting a chi-square distribution table. The two-dimensional confidence ellipse represents the points (mean 1, mean 2) corresponding to the mean of the Gaussian components. The semi-axis length of the two-dimensional confidence ellipse is proportional to the square root of the eigenvalues, which are obtained by eigenvalue decomposition of the covariance matrix. , The calculation formula includes: ; .in, This is the critical value of the chi-square distribution. and All values ​​are eigenvalues. If there are multiple Gaussian components, the union of the two-dimensional confidence ellipses is used as the confidence interval for the marine indicator pair. Similarly, the Optimum Target Maximization (EM) algorithm is used to train the land indicator distribution model, obtaining the land component parameter set for each land indicator pair. Then, the same steps are used to construct the two-dimensional confidence ellipse for each land component parameter set, thereby determining the land indicator confidence interval for each land indicator pair. Next, based on the calculated confidence intervals for the marine and land indicators and the marine and terrestrial ecological monitoring data, the target marine and terrestrial area is divided into intertidal and non-intertidal areas. The intertidal area includes the intertidal zone, land area, and marine area.

[0033] In one embodiment, matching all marine and terrestrial indicators in historical marine and terrestrial ecological data to obtain multiple marine and terrestrial indicator pairs involves the following steps: Extract historical terrestrial ecological data and historical marine ecological data from historical terrestrial and marine ecological data; Time-series analysis was performed on historical terrestrial ecological data and historical marine ecological data to obtain historical terrestrial sequences and historical marine sequences. The distribution statistics of all ocean indicators in the historical ocean sequence were performed, and the ocean fluctuation sequence of all ocean indicators was calculated based on the distribution statistics results. The land fluctuation sequence of all land indicators in the historical land sequence was calculated based on the clustering algorithm; The volatility synchronicity among all land indicators is calculated based on the land volatility sequence, and the indicator pair matching among all land indicators is completed based on the volatility synchronicity to obtain several land indicator pairs. Based on the ocean wave sequence, index pair matching was completed among all ocean indices, resulting in several ocean index pairs.

[0034] In this embodiment, historical terrestrial ecological data and historical marine ecological data refer to the historical monitoring data corresponding to terrestrial indicators and marine indicators, respectively. Terrestrial indicators refer to terrestrial monitoring indicators for stable terrestrial areas within the target land-sea region, including soil moisture content, vegetation cover, herbaceous plant height, soil element type, and soil element content. Stable terrestrial areas refer to typical terrestrial ecological areas without marine hydrological influence (e.g., areas ≥10km from the coastline). Marine indicators refer to marine monitoring indicators for stable marine areas within the target land-sea region, including water salinity, water temperature, benthic species abundance, and sediment particle size. Stable marine areas refer to typical marine ecological areas that have been continuously covered by seawater for a long period (e.g., areas ≥10km from the coastline).

[0035] Historical terrestrial and marine ecological data were arranged chronologically to obtain historical terrestrial and marine sequences, respectively. Marine indicators refer to marine monitoring indicators of stable marine areas within the target land-sea region, including water salinity, water temperature, benthic species abundance, and sediment particle size. Stable marine areas refer to typical marine ecological areas that have been continuously covered by seawater for a long period (e.g., areas ≥10km from the coastline). Then, using the same steps as calculating the terrestrial fluctuation sequence, the mean fluctuation, mean interval width, and center drift of the marine indicators were calculated, and after normalization, a weighted sum was calculated to obtain the marine fluctuation sequence of the marine indicators. Finally, a clustering algorithm was used to calculate the terrestrial fluctuation sequence of all terrestrial indicators in the historical terrestrial sequence.

[0036] By aligning the land fluctuation sequences of all land indicators with their timestamps, the Pearson correlation coefficient formula can be used to calculate the sequence similarity between the time-stamp-aligned land fluctuation sequences. This sequence similarity can then be used as the fluctuation synchronization degree among all land indicators. Alternatively, change point detection can be used to identify inflection points or trend change points in the land fluctuation sequences, and the synchronization ratio of inflection points or trend change points between different land fluctuation sequences can be determined. This identified synchronization ratio can then be used as the fluctuation synchronization degree among land indicators. Specifically, firstly, time series smoothing techniques, such as moving average, exponential smoothing, or locally weighted scatter smoothing, are used to smooth each land fluctuation sequence. Then, change point detection algorithms are used to identify moments in the smoothed land fluctuation sequence where statistical characteristics (such as mean, variance, and trend) change significantly. Commonly used change point detection algorithms include the cumulative sum algorithm and Bayesian online change point detection. Taking Bayesian online detection as an example, the initial run length is set to 0, and the prior probability distribution parameters are initialized. Next, new data is observed. Whenever new data arrives, the run length and probability distribution parameters are updated, and the predicted probability is calculated. Based on the current probability distribution parameters, the predicted probability density function value of the new data is calculated, followed by the growth probability calculation. If the new data is similar to the previous data, the run length is increased by 1; otherwise, it is reset to 0. Finally, the change point probability is calculated. If the new data differs significantly from the previous data, a change point is considered to have occurred, and the run length is reset to 0. Through these steps, a series of timestamps containing change points are ultimately output. After compiling a set of timestamps for all land indicators' changes, a time window is set, such as 1 week, 2 weeks, or 1 month. The system checks whether changes between any two land indicators occur within the same time window. If so, they are considered synchronization points. The proportion of synchronization points between any two land indicators to all changes is calculated and used as the volatility synchronization degree between the two land indicators. Alternatively, Hausdorff distance or Fréchet distance can be directly used to measure the similarity between the timestamp sets of two changes as the volatility synchronization degree. If the volatility synchronization degree between two land indicators is greater than a preset synchronization threshold, the two land indicators are matched as a single land indicator pair.

[0037] Similarly, after aligning all ocean wave sequences with timestamps, the sequence similarity between the land wave sequences of the ocean indicators that have completed timetamp alignment is calculated using the Pearson correlation coefficient formula, thus obtaining the wave synchronization degree among all ocean indicators. Alternatively, the wave synchronization degree among all ocean indicators can be calculated using change point detection algorithms such as the cumulative sum algorithm and Bayesian online change point detection. If the wave synchronization degree between two ocean indicators is greater than the preset synchronization degree threshold, the two ocean indicators are matched as one ocean indicator pair.

[0038] Both stable terrestrial and stable marine regions are stable ecosystems. Within stable ecosystems, certain ecological indicators (terrestrial and marine indicators) often exhibit long-term, predictable coupling relationships. For example, among terrestrial indicators, soil moisture content is typically positively correlated with terrestrial vegetation cover, while among marine indicators, water salinity may have a specific relationship with the abundance of certain marine organisms. Through the aforementioned indicator pair matching step, the coupling relationships between different ecological indicators can be effectively captured. This lays the data foundation for subsequently defining the normal ranges that the numerical sets of these coupled indicator pairs (terrestrial and marine indicator pairs) should fall within in stable terrestrial and stable marine regions, i.e., the confidence intervals for marine and terrestrial indicators. Influenced by tidal cycles, the intertidal zone experiences periodic inundation or exposure, subjecting it to the dual impacts of both land (freshwater runoff, sediment, nutrients, pollutants) and ocean (saltwater, waves, tides, marine life). This drastic and variable environment prevents the intertidal ecosystem from maintaining a relatively stable, singular, or few predictable ecological patterns, unlike stable terrestrial or marine areas. The relationships between ecological indicators become more complex, dynamic, and may exhibit unique nonlinear patterns. Therefore, the intertidal zone can be identified by whether the values ​​of these coupled indicator pairs fall within normal ranges. As one of the most vulnerable and sensitive areas to environmental changes, the intertidal zone must be assigned a higher risk weight in the process of ecological and environmental early warning for target land and sea areas. Furthermore, compared to non-intertidal areas, the intertidal zone, as a transitional zone between land and sea, is affected by multiple factors such as periodic seawater inundation, tides, runoff, and temperature, resulting in a much faster response to environmental disturbances than terrestrial or offshore areas. By identifying ecological changes in the intertidal zone, early ecological anomalies in the target land and sea areas can be directly captured, allowing for proactive ecological and environmental remediation and preventing further deterioration of the target land and sea areas' ecosystems.

[0039] In one embodiment, calculating the land fluctuation sequence of all land indicators in the historical land sequence based on a clustering algorithm includes the following steps: For any land index, the historical land sequence is divided into all land index subsequences according to time order; For any land indicator subsequence, the land indicator subsequence is divided according to a preset time window to obtain multiple indicator datasets; For any index dataset, a clustering algorithm is used to perform data clustering of the index dataset, and high-density clusters are selected from the data clustering results as core distribution clusters; The land index interval is defined based on the percentile range of the core distribution cluster; Kernel density estimation is performed on the core distribution cluster of the indicator dataset to obtain the indicator probability density function of the indicator dataset; Based on the time window division order, multiple neighborhood datasets of the indicator dataset are determined, the average probability density function of all neighborhood datasets is calculated, and the basic volatility of the indicator dataset is calculated by combining the indicator probability density function and the average probability density function. Calculate the mean base volatility and the mean width of the interval for all indicator datasets to obtain the mean volatility and the mean interval width. Extract the center point of the land index interval of all index datasets, and calculate the center drift between all adjacent index datasets based on the center point extraction results; The land fluctuation sequence of all land indicators is calculated by combining the mean of all fluctuations, the mean of interval width, and the center drift.

[0040] In this embodiment, the land index sequence corresponding to each land index in the historical land sequence is extracted. Then, the land index sequence is divided into multiple land index subsequences. A land index subsequence is a time-series sequence of data where the same land index data is arranged in chronological order. For example, the land index subsequence corresponding to soil moisture content is... , This represents the soil moisture content dataset collected from all land and sea monitoring points within the stable land area at time point t. Next, the land indicator subsequence is divided into multiple indicator datasets using a preset time window. The time window can be determined based on the time range covered by the land and sea ecological monitoring data, such as one week. An indicator dataset refers to the same land indicator data within different time periods. For example, the land indicator dataset includes... , , ..., .

[0041] Next, clustering algorithms are used to cluster the data for each indicator dataset. Commonly used clustering algorithms include DBSCAN, OPTICS, and hierarchical clustering algorithms. Taking DBSCAN as an example, DBSCAN is a density-based spatial clustering algorithm. Its core steps include: determining core points and calculating the number of samples in the ε-neighborhood (a circular region with radius ε centered on the point) of each point. If the number of samples is greater than or equal to the preset minimum number of contained points, then the point is regarded as a core point; expanding the cluster by recursively adding core points in the neighborhood of the core point to the same cluster; marking boundary points by marking non-core points that are located in the neighborhood of the core point as boundary points and merging them into the corresponding cluster; and identifying noise points by considering isolated points that are neither core points nor boundary points as noise. The above steps of clustering can yield multiple data clusters. Then, the intra-cluster density of all data clusters is calculated: intra-cluster density = number of core points / total number of samples within the cluster. High-density clusters with intra-cluster density greater than a preset density threshold are selected as core distribution clusters. Next, the sample data within each core distribution cluster are sorted, and then its percentile range is calculated. For example, the 5th percentile can be calculated as the lower limit, meaning that 5% of the sample data within the core distribution cluster is less than this lower limit. The 95th percentile can be calculated as the upper limit, meaning that 95% of the sample data within the core distribution cluster is less than this upper limit. The interval formed by the calculated upper and lower limits is used as the initial index interval. The initial index intervals of all core distribution clusters are calculated, and the initial index interval with the largest coverage is selected as the land index interval.

[0042] Kernel density estimation (KDE) is a nonparametric method used to estimate the probability density function of a random variable. It works by applying a kernel function (e.g., a Gaussian kernel) to each sample data point within all core distribution clusters, and then summing the kernel functions of all samples to obtain a smooth density estimation curve, i.e., the indicator probability density function. This indicator probability density function characterizes the probability density of the land indicator taking different values ​​over a certain time period. A neighborhood dataset refers to a dataset that is temporally close to the indicator dataset. For example, the two indicator datasets preceding and following a given indicator dataset can be considered its neighborhood dataset. The neighborhood dataset is By integrating the sample data within the core distribution cluster of the neighborhood dataset into a neighborhood sample set, and then performing kernel density estimation on this neighborhood sample set, the average probability density function of the entire neighborhood dataset can be obtained.

[0043] Next, the Kullback-Leibler (KL) divergence measure can be used to quantify the distributional difference between the index probability density function and the average probability density function, and this difference can be used as the basic volatility of the index dataset. Specifically, the Kullback-Leibler (KL) divergence is an asymmetric measure of the difference between two probability distributions, also known as relative entropy. It is used to quantify the information loss or information gain that occurs when an approximate distribution is used to replace the true distribution. By substituting the index probability density function and the average probability density function into the KL divergence formula, the distributional difference between the two can be calculated, and this calculated difference can be used as the basic volatility of the index dataset.

[0044] After calculating the base volatility and land indicator intervals for all indicator datasets using the same method, the mean of all base volatility is calculated to obtain the mean volatility. The mean volatility can be used to represent the average change in the local data distribution of a land indicator relative to its neighborhood over the entire historical period. Simultaneously, the interval width of all land indicator intervals is extracted (interval width = upper limit - lower limit), and the mean of all interval widths is calculated to obtain the mean interval width. The mean interval width reflects the overall volatility range of the land indicator. Next, the center points of the land indicator intervals for all indicator datasets are extracted. The center point of a land indicator interval refers to the data located at the center of the interval, i.e., the central data. The average absolute change of all central data is calculated; that is, first, the absolute value of the difference between central data points is calculated, and then the mean of all absolute values ​​of the difference is calculated to obtain the average absolute change. This average absolute change is used as the center drift between all adjacent indicator datasets. The center drift reflects the average movement of the land indicator's baseline value over the historical period. After normalizing the mean volatility, mean interval width, and center drift, a weighted average is applied to calculate the land volatility of a specific subsequence of the land index. Arranging the land volatility in chronological order yields the land volatility sequence of that land index. The same method can be used to calculate the ocean volatility sequences of all ocean indices.

[0045] In one embodiment, reference is made to Figure 2 The process of dividing the target marine and terrestrial areas into intertidal and non-intertidal zones by combining confidence intervals for marine indicators, confidence intervals for terrestrial indicators, and marine and terrestrial ecological monitoring data includes the following steps: S201. For any land-sea monitoring point, extract all land-sea indicator pairs and all marine-sea indicator pairs corresponding to the land-sea ecological monitoring data of the land-sea monitoring point. S202. Statistically determine the distribution characteristics of all ocean monitoring data within the corresponding ocean indicator confidence intervals, and calculate the ocean deviation of the land and sea monitoring points based on the distribution characteristics. S203. Statistically determine the distribution characteristics of all land-to-land monitoring data within the corresponding land index confidence interval, and calculate the land deviation of the land-sea monitoring points based on the distribution characteristics. S204. Combining the ocean deviation and land deviation of all ocean and land monitoring points, the target ocean and land area is divided into intertidal zone and non-intertidal zone, wherein the non-intertidal zone includes ocean area and land area.

[0046] In this embodiment, the land-sea monitoring point is the smallest sampling unit of the target land-sea area. At each monitoring point, various sensors and sampling devices are deployed to collect land-sea ecological monitoring data for a specific sampling area within the target land-sea area. The land-sea monitoring point corresponds to a sampling area within the target land-sea area, and the closer the sampling area is to the coastline, the smaller its area. Some monitoring indicators within the sampling area are collected on an ad-hoc basis, such as soil moisture content, soil pH value, and water temperature, while other monitoring indicators are collected periodically, such as daily, weekly, or monthly, such as vegetation coverage and herbaceous plant height.

[0047] Extract all land indicator pairs and all corresponding marine indicator pairs from the marine and terrestrial ecological monitoring data of all land and marine indicator pairs. Analyze whether each land indicator pair's monitoring data falls within the confidence interval of its corresponding land indicator. Specifically, for any land indicator pair's monitoring data, map it to a two-dimensional space to form an indicator pair data point. For example, if the indicator pair data point is (R1, T1), R1 and T1 represent the monitoring data of land indicators R and T at time point 1, respectively. Next, the squared Mahalanobis distance between the data points of the indicator pair and the mean of the nearest Gaussian component is calculated based on the mean and covariance matrix of the two-dimensional confidence ellipse. If the squared Mahalanobis distance is less than the chi-square critical value, the data point of the indicator pair is determined to be within the confidence interval of the land indicator. The number of all indicator pairs whose squared Mahalanobis distance is greater than or equal to the chi-square critical value is counted, and the proportion of the number of indicator pairs is calculated to obtain the land indicator deviation of the land indicator pair. In the same way, the ocean indicator deviation of all ocean indicator pairs is calculated, and weights are assigned to them according to the fluctuation synchronicity of the ocean indicator pairs. The greater the fluctuation synchronicity, the higher the weight, but the total weight is 1. The weighted sum of all ocean indicator deviations is used to obtain the ocean deviation of the land-sea monitoring point. The land deviation of the land-sea monitoring point is calculated in the same way. If both the ocean deviation and the land deviation of the land-sea monitoring point are greater than the preset corresponding deviation threshold, the sampling area corresponding to the land-sea monitoring point is determined to be the intertidal zone. The above method is used to determine whether the sampling areas corresponding to all land-sea monitoring points are intertidal zones. For each sampling area where intertidal zone determination has been completed, Kriging interpolation is used to generate an intertidal zone probability distribution map, and then spline curves are used to fit and connect the boundaries to form a continuous intertidal zone, land area, and marine area.

[0048] In one embodiment, a multidimensional feature vector of the intertidal zone is constructed based on marine and terrestrial ecological monitoring data. An intertidal topology map is then constructed by combining the multidimensional feature vector with all marine and terrestrial monitoring points within the intertidal zone. Based on this topology map, the intertidal zone is further divided into sub-regions to obtain multiple intertidal sub-regions, including the following steps: Extract intertidal monitoring data from all land and sea monitoring points within the intertidal zone; Time-series analysis was performed on all intertidal monitoring data to obtain multiple intertidal sequences; Tidal cycle data for the target land and sea areas were obtained from a water environment database. Align the timestamps of all intertidal sequences, divide all the intertidal sequences with aligned timestamps into time series based on tidal cycle data, and construct multiple tidal zone datasets located in different tidal time periods based on the time series division results. For any tidal zone dataset, complete the multidimensional feature extraction of the tidal zone dataset to obtain the tidal multidimensional features of all land and sea monitoring points in the intertidal zone; All land and sea monitoring points in the intertidal zone are used as topological nodes, and all adjacent topological nodes are connected to construct an intertidal topological map. Based on the multidimensional tidal features, the node feature similarity between all adjacent topological nodes is calculated using the similarity formula, and the edge weights of the intertidal topological graph are determined based on the node feature similarity. Arrange the intertidal topology maps with all edge weights assigned in chronological order to obtain the intertidal topology map sequence; The dynamic community detection algorithm is used to search for the nodes of the intertidal topology sequence and obtain the community affiliation sequence of all topological nodes. Based on the community affiliation sequence, the community stability and community stability period of all topological nodes are analyzed, and the intertidal zone affiliation of all topological nodes is determined by combining the community stability and community stability period. The intertidal zone is divided into multiple intertidal sub-regions based on the intertidal zone affiliation of all topological nodes. The intertidal sub-regions include the landside transition zone, the core transition zone, and the seaside transition zone.

[0049] In this embodiment, intertidal monitoring data collected from all land and sea monitoring points within the intertidal zone is extracted from the land and sea monitoring data. The intertidal monitoring data is arranged chronologically to obtain multiple intertidal sequences. An intertidal sequence refers to the time-series data of intertidal zones collected from each land and sea monitoring point within the intertidal zone. Next, the timestamps of all intertidal sequences are aligned, and the data is divided chronologically using a preset first time window to obtain multiple intertidal sub-sequences. Then, intertidal sub-sequences located within the same first time window (tidal period) but collected from different land and sea monitoring points are integrated into a tidal zone dataset. For each tidal zone dataset, feature extraction is performed on each intertidal sub-sequence, extracting features such as variance, extreme values, linear trend slope, and number of abrupt change points for each monitoring indicator. These features are then normalized and concatenated to form the tidal multidimensional features of each land and sea monitoring point. Next, all land and sea monitoring points in the intertidal zone are used as topological nodes. All adjacent topological nodes are connected to construct an intertidal topology map. Adjacent topological nodes can be defined by methods such as Delaunay triangulation, K-nearest neighbor graph, distance threshold, and Voronoi graph adjacency. Taking K-nearest neighbor graph as an example, the K nearest nodes to each topological node are selected as the adjacent topological nodes of that topological node. Alternatively, taking distance threshold as an example, topological nodes whose Euclidean distance from the topological node is less than the preset distance threshold are selected as adjacent topological nodes.

[0050] After constructing the intertidal topology map by connecting all adjacent topological nodes, the node feature similarity between all adjacent topological nodes is calculated based on tidal multidimensional features and using similarity formulas, such as the cosine similarity formula or Pearson correlation coefficient. Edge weights are assigned to the topological edges connecting two adjacent topological nodes based on node feature similarity; the higher the node feature similarity, the higher the assigned edge weight. The intertidal topology maps with all edge weights assigned are arranged in chronological order to construct an intertidal topology map sequence. Then, a dynamic community detection algorithm is used to identify and track the evolution of community structures in the intertidal topology map sequence, recording the community to which each land-sea monitoring point belongs during each tidal period. Finally, the community affiliation sequence of all topological nodes is output. The community affiliation sequence contains which intertidal subregion each topological node belongs to during each tidal period. Commonly used dynamic community detection algorithms include the Leiden algorithm, multi-target immune algorithm, and label propagation method (LFM).

[0051] Taking the Leiden algorithm as an example, it divides nodes into communities with dense internal connections and sparse external connections by optimizing modularity. Specifically, it initializes the community structure in the first time step, treating each topological node in the first intertidal topology as an independent community. Then, it moves each topological node to the community of its neighboring node. If the move improves the modularity of the entire intertidal topology, it proceeds. This process is repeated until moving any topological node no longer improves the modularity, at which point the process stops. Next, each detected community is treated as a supernode, and the process is repeated until the modularity can no longer be improved. Then, the community structure from the first time step is used as prior information to continue iterating the node movement steps for subsequent intertidal topologies. Furthermore, during iteration, if two communities become highly similar or tightly connected, they may be merged; conversely, if a community's internal connections weaken or its external connections strengthen, it may split. After completing the community search for the entire intertidal topology, the community affiliation sequence of all topological nodes is output.

[0052] The proportion of each topological node belonging to different communities is calculated based on the community affiliation sequence. The highest proportion is used as the community consistency index for each topological node. For example, if the proportions of a topological node belonging to communities A, B, and C are 0.6, 0.3, and 0.1 respectively, then the community consistency index for that topological node is 0.6. Next, the community switching frequency of each topological node is calculated. For example, if a topological node belongs to community A during tidal period 'a' and belongs to community B during the next tidal period 'b', then the community switching frequency of that topological node is incremented by 1. The community consistency index and community switching frequency are combined to determine the community stability of that topological node. Finally, for each topological node, the longest period during which it continuously remains in the same community in its community affiliation sequence is identified, resulting in the community stability period. Finally, if the community consistency index is greater than a preset threshold and the community switching frequency is less than a preset frequency threshold, then the community with the highest proportion is directly selected as the intertidal zone affiliation of that topological node. For example, if a topological node's community consistency index is greater than the preset threshold, its community switching frequency is less than the preset frequency threshold, and its proportions belonging to community A, community B, and community C are 0.8, 0.1, and 0.1 respectively, then the topological node's intertidal zone affiliation is determined to be community A. If a topological node's community consistency index is greater than the preset threshold, but its community switching frequency is greater than or equal to the preset frequency threshold, or its community switching frequency is less than the preset frequency threshold, but its community consistency index is less than the preset threshold, then the community corresponding to its stable period is selected as the topological node's intertidal zone affiliation. If a topological node's community consistency index is less than or equal to the preset threshold, but its community switching frequency is greater than or equal to the preset frequency threshold, then the topological node's intertidal zone affiliation is marked as a pending community.

[0053] The intertidal zone affiliation of topological nodes is mapped to the corresponding land-sea monitoring points, initially dividing the primary landside transition zone, primary core transition zone, and primary seaside transition zone into an intertidal zone distribution map. For example, if community B is located in the middle area of ​​the intertidal zone in the intertidal zone distribution map, it means that the land-sea monitoring point corresponding to community B is located in the core transition zone. The sampling area where the land-sea monitoring point corresponding to community B is located is initially classified as the primary core transition zone. For land-sea monitoring points corresponding to undetermined communities in the intertidal zone distribution map, the nearest neighbor method is used, and the undetermined community is assigned to the community closest to it. In addition, there is a case where the land-sea monitoring point corresponding to an undetermined community is located between two sub-regions, such as the boundary between the primary landside transition zone and the primary core transition zone. In this case, its intertidal zone affiliation is determined based on the highest proportion. When the number of undetermined communities is less than a preset threshold, spline curves are used to fit the community boundaries in the intertidal zone distribution map to form continuous landside transition zones, core transition zones, and seaside transition zones.

[0054] In one embodiment, reference is made to Figure 3 The process of generating an ecological and environmental early warning report for a target marine and terrestrial area by combining marine and terrestrial ecological monitoring data and ecological area parameters includes the following steps: S301. Construct a marine-terrestrial ecological supermap of the target marine-terrestrial region based on marine-terrestrial ecological monitoring data; S302. Based on historical marine and terrestrial ecological data and using the Granger causality test algorithm, identify the ecological causal relationships in the marine and terrestrial ecological hypergraph, and calculate the ecological pressure parameter sequence of all ecological supernodes in the marine and terrestrial ecological hypergraph based on the ecological causal relationships and the pre-constructed marine and terrestrial ecological evolution model. S303. The division results of each marine and land sub-region within the target marine and land area divide the ecological pressure parameter sequence into multiple regional pressure parameter sequences. S304. For any regional pressure parameter sequence, extract the multidimensional pressure features of the regional pressure parameter sequence, where the multidimensional pressure features include pressure trend features, pressure steady-state features, and pressure propagation features. S305. Based on the multidimensional pressure characteristics, complete the ecological risk assessment of the marine and terrestrial sub-regions, and generate the ecological early warning level and pressure propagation path of the marine and terrestrial sub-regions based on the ecological risk assessment results; S306. Integrate the ecological early warning levels and pressure propagation paths of all marine and land sub-regions to generate an ecological and environmental early warning report for the target marine and land region.

[0055] In this embodiment, a marine-terrestrial ecological hypergraph of the target marine-terrestrial region is first constructed based on marine-terrestrial ecological monitoring data. Then, using the Granger causality test algorithm and a vector autoregression model, the ecological restoration parameters of all ecological supernodes and the ecological propagation parameters between all ecological superedges and all ecological supernodes are calculated. These ecological restoration and propagation parameters are then mapped onto the marine-terrestrial ecological hypergraph and updated. Based on the updated parameters in the hypergraph and using a marine-terrestrial ecological evolution model, the cumulative ecological pressure propagation data of all ecological supernodes in the hypergraph over a future period is calculated, i.e., the ecological pressure parameter sequence. Finally, an ecological environment early warning report for the target marine-terrestrial region is generated based on the ecological pressure parameter sequence. Specifically, for any regional pressure parameter sequence, its multidimensional pressure characteristics are first extracted. These multidimensional pressure characteristics include pressure trend characteristics, pressure steady-state characteristics, and pressure propagation characteristics. Pressure trend characteristics refer to the pressure growth rate and pressure accumulation in the marine-terrestrial sub-region. Pressure steady-state characteristics include steady-state deviation and disturbance state labels (T=0 / 1 / 2, corresponding to healthy steady-state / disturbed state / degraded steady-state). If the regional pressure parameter sequence exhibits a single-peak decay and then stabilizes at a low ESI, it corresponds to a healthy steady-state. If the regional pressure parameter sequence fluctuates repeatedly between the low and high pressure index ranges, it corresponds to a disturbed state. If the sequence remains stable in the high pressure index range for a long period (fluctuation ≤0.01), it corresponds to a degraded steady-state. The steady-state deviation is the average difference between all pressure indices in the regional pressure parameter sequence and the average pressure index in the healthy state. The average pressure index in the healthy state can be calculated by selecting healthy periods from historical marine-terrestrial ecological data, characterized by no pollution events, no extreme weather, and low human activity disturbance. The low and high pressure index ranges correspond to the pressure index ranges in the healthy state and the pressure index ranges in the polluted state, respectively. Pressure propagation characteristics include the hyperedge coordination coefficient and pressure propagation delay. The hyperedge coordination coefficient refers to the degree of synchronization in pressure propagation among multiple ecological supernodes within an ecological hyperedge. It can be calculated by aligning the timestamps of multiple regional pressure parameter sequences of multiple ecological supernodes within the ecological hyperedge, and then calculating the average Pearson correlation coefficient among the multiple ecological supernodes. This average value is used as the hyperedge coordination coefficient of the ecological hyperedge. The calculation steps for the pressure propagation delay include: first, extracting the peak pressure index of the regional pressure parameter sequences of multiple ecological supernodes within the ecological hyperedge; then, calculating the average time difference between the peak pressure index values; and finally, using this average time difference as the pressure propagation delay of the ecological hyperedge.

[0056] If the disturbance state label of an ecological supernode is T=0, it indicates a healthy steady state, meaning the corresponding marine-terrestrial sub-region is in a healthy state and no risk warning is needed. If the disturbance state label of an ecological supernode is T=1, it indicates a disturbed state, meaning the ecological condition of the corresponding marine-terrestrial sub-region is disordered, possibly due to climate change or environmental pollution. In this case, the steady-state deviation, pressure growth rate, pressure accumulation, and super-edge synergy coefficient need to be normalized and then weighted (using entropy weighting or expert scoring to determine the weights) to obtain the regional ecological risk level. Then, the ecological warning level is determined based on the regional ecological risk level and a pre-obtained risk threshold. If the regional ecological risk level is greater than the corresponding risk threshold, the risk warning level is determined to be level two; if the regional ecological risk level is less than or equal to the corresponding risk threshold, the risk warning level is determined to be level three. Furthermore, the risk threshold varies for different marine-terrestrial sub-regions. The intertidal zone is more sensitive to environmental degradation, therefore its risk threshold is lower. The risk threshold can be determined based on the regional ecological risk level under healthy and polluted conditions. If the disturbance status label of an ecological supernode is T=2, it indicates that the marine and terrestrial sub-regions corresponding to the ecological supernode have experienced environmental degradation. Therefore, the risk warning level of the ecological supernode is directly determined to be Level 1, and ecological environment governance needs to be started immediately to prevent further deterioration of the ecological environment.

[0057] Next, candidate propagation paths are initially determined based on the order of appearance of pressure index peaks in the regional pressure parameter sequence. Then, the candidate propagation paths are verified based on the ecological causal relationships between all ecological supernodes and all ecological superedges. If there are ecological causal relationships between adjacent ecological supernodes and ecological superedges in the candidate propagation path, it indicates that the pressure transmission between supernodes is causally driven rather than accidental synchronization. Therefore, the candidate propagation path can be output as the pressure propagation path.

[0058] Ecological and environmental early warning reports for the target marine and land areas are generated by integrating the ecological early warning levels and pressure propagation paths of all marine and land sub-regions. The ecological and environmental early warning reports include the ecological early warning levels and pressure propagation paths of all marine and land sub-regions that are in a disturbed or degraded stable state, as well as marine and land ecological monitoring data of all marine and land sub-regions (including marine and land sub-regions that are in a healthy stable state).

[0059] In one embodiment, constructing a marine-terrestrial ecological supermap of a target marine-terrestrial region based on marine-terrestrial ecological monitoring data includes the following steps: Each marine and terrestrial sub-region within the target marine and terrestrial area is defined as an ecological supernode. The marine and terrestrial sub-regions include marine areas, terrestrial areas, landside transition areas, core transition areas, and seaside transition areas. Extract non-intertidal and intertidal monitoring data from marine and terrestrial ecological monitoring data; The non-intertidal ecological pressure index and non-intertidal resource index of the corresponding ecological supernodes in the non-intertidal zone are calculated based on non-intertidal monitoring data. The intertidal ecological pressure index and intertidal resource index of the intertidal region were calculated by combining intertidal monitoring data and ecological area parameters. The non-intertidal ecological pressure index, non-intertidal resource index, intertidal ecological pressure index, and intertidal resource index are integrated into supernode attribute parameters, and the supernode attribute parameters are mapped to each ecological supernode. Based on marine and terrestrial ecological monitoring data and historical marine and terrestrial ecological data, regional ecological interaction chains between all marine and terrestrial sub-regions were extracted, and the regional ecological interaction chains were defined as ecological superborders. Among them, regional ecological interaction chains include pollutant runoff chains, climate impact chains, and biological migration chains. By combining the non-intertidal ecological pressure index, the non-intertidal resource index, the intertidal ecological pressure index, and the intertidal resource index, the hyperedge attribute parameters of all ecological hyperedges are calculated, and the hyperedge attribute parameters are mapped to the corresponding ecological hyperedges. A marine-terrestrial ecological supergraph is constructed based on all ecological supernodes that have completed supernode attribute parameter mapping and all ecological superedges that have completed superedge attribute parameter mapping.

[0060] In this embodiment, the marine and terrestrial sub-regions defined in the above steps are used as ecological supernodes. These sub-regions include marine areas, terrestrial areas, landside transition areas, core transition areas, and seaside transition areas. The ecological supernodes include marine supernodes, terrestrial supernodes, landside transition supernodes, core transition supernodes, and seaside transition supernodes. Then, based on the distribution of marine and terrestrial monitoring points, non-intertidal monitoring data and intertidal monitoring data are extracted from the marine and terrestrial ecological monitoring data. Non-intertidal monitoring data refers to monitoring data collected from land and sea monitoring points located in non-intertidal areas. Non-intertidal monitoring data includes land area data and marine area data. Land area data includes soil moisture content, vegetation coverage, soil pH value, etc., while marine area data includes water temperature, salinity, plankton abundance, and marine pollutant concentration (such as petroleum and heavy metals), etc. Intertidal monitoring data refers to monitoring data collected from land and sea monitoring points located in intertidal areas. Intertidal monitoring data includes data from the landside transition zone, core transition zone, and seaside transition zone. It mainly includes salinity fluctuation values, benthic organism (such as shellfish and crabs) abundance, sediment particle size, tidal inundation duration, intertidal vegetation (such as mangroves and salt marshes) coverage, and land-based pollutant transit concentration, etc.

[0061] Next, based on non-intertidal monitoring data, the non-intertidal ecological pressure index and non-intertidal resource index of the corresponding ecological supernodes in the non-intertidal zone are calculated. Specifically, the non-intertidal ecological pressure index of the marine supernode is calculated first. Based on marine monitoring experience, pressure-related parameters that affect marine ecology are extracted from the marine regional data, such as marine pollutant concentrations (e.g., petroleum hydrocarbons, heavy metals), coral bleaching area, marine nitrogen and phosphorus concentrations, phytoplankton decline rate (calculated as the percentage difference between the current phytoplankton abundance and the historical average abundance), and benthic decline rate (calculated as the percentage difference between the current benthic biomass and the historical average biomass). After standardizing the pressure-related parameters, multiple standardized pressure parameters are obtained. Then, the entropy weight method or expert scoring method can be used to assign weights to each standardized pressure parameter, and then a weighted average is performed to obtain the non-intertidal ecological pressure index of the marine supernode. Pressure-related parameters in the terrestrial regional data include soil pH, vegetation cover loss rate, soil organic matter decline rate, and terrestrial animal habitat reduction rate. The non-intertidal ecological pressure index of the terrestrial supernode is calculated using the same method. Next, the non-intertidal resource index was calculated. Resource-related parameters from the non-intertidal monitoring data were selected, including marine and terrestrial resource parameters. These parameters are those that can improve the ecological resilience of terrestrial or marine areas and are conducive to natural ecological restoration. For example, marine resource parameters include ocean current intensity, seawater exchange rate, plankton abundance, and marine vegetation coverage (the proportion of seagrass beds and algae beds). Terrestrial resource parameters include soil organic matter content, vegetation type diversity, and terrestrial vegetation coverage. After standardizing the marine and terrestrial resource parameters, a weighted average was calculated (the weights can be determined using the entropy weight method or expert scoring method) to obtain the non-intertidal resource indices for marine and terrestrial supernodes, respectively.

[0062] Next, combining intertidal monitoring data and ecological area parameters, the intertidal ecological pressure index and intertidal resource index of the intertidal region are calculated. The ecological area parameters include the intertidal zone area and sub-region area parameters. The sub-region area parameters include the sub-region area and its proportion. The sub-region area refers to the area of ​​the landside transition zone, the core transition zone, and the seaside transition zone. The sub-region area proportion refers to the percentage of the intertidal zone area occupied by each of these three areas. After completing the intertidal zone and sub-region division, the intertidal zone area and sub-region area can be calculated using software such as ArcGIS and QGIS. Then, the proportion of each sub-region area is obtained by dividing the area of ​​each sub-region by the area of ​​the intertidal zone.

[0063] First, the intertidal ecological pressure index of each intertidal subregion is calculated based on the pressure-related parameters of the three subregions. The pressure-related parameters for the three regions include the salinity fluctuation exceedance rate, sediment erosion rate, intertidal biodiversity decline rate, pollutant concentration exceedance rate, and area change coefficient, etc. Sediment erosion rate and area change coefficient are also included. ,in This is the area variation coefficient. This represents the proportion of the changed area of ​​a sub-region to the corresponding historical sub-region. The preset adjustment coefficients are used. The pressure-related parameters of each intertidal subregion are standardized and then weighted to obtain the intertidal ecological pressure index for each subregion. Next, the intertidal resource index is calculated based on the resource-related parameters of each subregion, including intertidal vegetation cover, sediment stability, biological community integrity, and subregion area. These parameters are standardized and then weighted to obtain the intertidal resource index for each subregion. Finally, the non-intertidal ecological pressure index, non-intertidal resource index, intertidal ecological pressure index, and intertidal resource index are used as supernode attribute parameters for each supernode and mapped to each ecological supernode.

[0064] Based on marine and terrestrial ecological monitoring data and historical marine and terrestrial ecological data, regional ecological interaction chains among all marine and terrestrial sub-regions are extracted and defined as ecological superedges. These regional ecological interactions include pollutant runoff chains, climate impact chains, and biological migration chains. Regional ecological interaction chains refer to complex ecological processes or environmental impact pathways involving two or more supernodes, formed between marine and terrestrial sub-regions through material flow, energy transfer, information exchange, or pressure transmission. These chains represent high-order, multi-regional collaborative interactions. Pollutant runoff chains can be simulated using Geographic Information Systems (GIS) combined with hydrological models to model runoff paths and confluence areas in terrestrial regions. By overlaying land-based pollution sources with runoff paths, the potential paths and intensities of pollutant transport from terrestrial regions to landside transition zones and further impacting core transition zones can be analyzed. The spatiotemporal correlation of pollutant concentrations in different regions can be analyzed based on the concentration gradient of the same pollutant, confirming the cascading transport effect of pollutants along the runoff direction. These chains can be defined as ecological superedges connecting terrestrial supernodes, landside transition supernodes, and core transition supernodes. Climate impact chains can be analyzed by examining the synergistic impact patterns of climate change factors such as sea-level rise, ocean acidification, and marine heat waves on marine regions, seaside transition zones, and core transition zones in historical marine and terrestrial ecological data. This allows for the assessment of the cascading effects of extreme weather events on these regional ecosystems; for example, how storm surges impact seaside transition zones from marine regions, thereby affecting the physical structure and biological communities of core transition zones. These chains can be defined as ecological superedges connecting marine supernodes, core transition supernodes, and seaside transition supernodes. Biological migration chains, on the other hand, can be identified by analyzing species migration data and habitat use patterns, identifying biological migration routes that depend on multiple marine and terrestrial subregions. These can serve as ecological superedges linking marine supernodes, terrestrial supernodes, terrestrial transition supernodes, core transition supernodes, and seaside transition supernodes. Next, the superedge attribute parameters for each superedge are calculated, including superedge order, higher-order propagation rate, and superedge activity. The superedge order refers to the number of ecological supernodes connected by each ecological superedge.

[0065] The calculation steps for the higher-order propagation rate include: statistically analyzing historical marine and terrestrial ecological data on ecological pressure propagation events, such as pressure diffusion from land-based pollution to land-side transitional supernodes and then to core transitional supernodes, and biological migration causing synchronous changes in pressure across the entire region. Then, for each ecological pressure propagation event, the time difference between the peak pressure at the source supernode and the peak pressure at the terminal supernode is calculated. For example, the time difference between when a certain pollution appears at the land supernode and when it appears at the core transitional supernode. Based on this time difference and the change in pollution concentration, the basic propagation rate is calculated. Then, it is weighted according to the super-edge order (the higher the order, the stronger the co-propagation, and the higher the basic higher-order propagation rate). Finally, the average of the basic propagation rates of all ecological pressure propagation events is taken as the basic higher-order propagation rate. Next, the relative area proportions of each land-sea sub-region will be calculated. First, the distribution area of ​​land-sea monitoring points will be calculated. Then, the relative land area and relative ocean area will be calculated based on the intertidal zone area. Next, the relative area proportions of each land-sea sub-region will be calculated. Then, the relative area index of each ecological superedge will be calculated. The relative area index = Σ(supernode resource index × supernode relative area proportion) / number of associated supernodes. The supernode resource index refers to the intertidal resource index or the non-intertidal resource index. The product between the relative area index and the basic propagation rate will be used as the higher-order propagation rate of the ecological superedge.

[0066] Hyperedge activity refers to the propagation intensity of ecological pressure propagation events occurring within an ecological superedge. The calculation steps for hyperedge activity include: calculating the pressure growth rate of each ecological supernode within different time windows based on the non-intertidal and intertidal ecological pressure indices calculated above. The time window is defined according to the monitoring time step of the land and sea monitoring points within the target land-sea area. Next, the synchronization of pressure growth among the ecological nodes connected by each ecological superedge is calculated, thereby determining its pressure coordination coefficient. The percentage of supernodes with a pressure growth rate greater than zero within each ecological superedge is then multiplied by the ratio of the average pressure growth rate to the peak pressure growth rate of all ecological supernodes connected to that ecological superedge, yielding the pressure coordination coefficient of that ecological superedge. Hyperedge activity is a comprehensive reflection of propagation efficiency and synchronous propagation status; therefore, the calculated higher-order propagation rate and pressure coordination coefficient need to be normalized and their average calculated as the hyperedge activity. Finally, the hyperedge attribute parameters are mapped to the corresponding ecological superedges, and all ecological supernodes that have completed the hyperedge attribute parameter mapping are connected to complete the mapping of supernode attribute parameters, thus constructing a land-sea ecological supergraph.

[0067] In one embodiment, the ecological causal relationships in the marine-terrestrial ecological hypergraph are identified based on historical marine-terrestrial ecological data and using the Granger causality test algorithm. The ecological pressure parameter sequence of all ecological supernodes in the marine-terrestrial ecological hypergraph is then calculated based on these ecological causal relationships and a pre-constructed marine-terrestrial ecological evolution model, including the following steps: Based on historical marine and terrestrial ecological data, the historical ecological pressure index and historical ecological activity index of all ecological supernodes in the marine and terrestrial ecological supermap were calculated respectively. By combining historical stress index and historical activity index and using Granger causality test algorithm, ecological causal relationships between all ecological supernodes and all ecological superedges are identified. Based on ecological causality, the ecological propagation parameters between all ecological hyperedges and all ecological supernodes are calculated using a vector autoregression model. Historical pollution time data of the target marine and land areas were obtained from the water environment database. A pollution attenuation model of the target marine and land areas was fitted by combining the historical pollution time data and the historical ecological pressure index. The non-intertidal ecological pressure index or intertidal ecological pressure index of all ecological supernodes is input into the pollution attenuation model, and the ecological restoration parameters of all ecological supernodes are output sequentially through the pollution attenuation model. Ecological propagation parameters and ecological restoration parameters are mapped onto a marine-terrestrial ecological supergraph, and the mapped marine-terrestrial ecological supergraph is input into a pre-constructed marine-terrestrial ecological evolution model. Based on the marine-terrestrial ecological evolution model, the ecological pressure parameter sequence of all ecological supernodes is iteratively calculated. The marine-terrestrial ecological evolution model is constructed based on a supergraph propagation dynamics model.

[0068] In this embodiment, the calculation method for the historical ecological pressure index is similar to that for the intertidal or non-intertidal ecological pressure index, except that recently collected marine and terrestrial ecological monitoring data are replaced with historical marine and terrestrial ecological data from different time periods. The historical ecological activity index is used to quantify the activity level of driving factors or ecological processes that may lead to changes in ecological pressure at different points in the past for ecological supernodes. Ecological activity parameters corresponding to each ecological supernode in the historical marine and terrestrial ecological data are extracted. For example, ecological activity parameters corresponding to terrestrial supernodes include the frequency of historical land-based pollutant emissions, the rate of change in indicator coverage, and the rate of change in soil organic matter; ecological activity parameters corresponding to marine supernodes include the rate of change in marine nitrogen and phosphorus concentrations and the rate of phytoplankton decay; and ecological activity parameters corresponding to the ecological supernodes in the three intertidal sub-regions include the frequency of salinity fluctuations exceeding standards, the rate of sediment erosion, and the frequency of changes in intertidal biological abundance. Similarly, for any ecological supernode, its ecological activity parameters are normalized and then weighted (weights can be assigned using the entropy weight method or expert scoring method) to obtain the ecological activity parameters for each ecological supernode.

[0069] Historical stress and activity indices for different time periods were arranged chronologically to construct historical stress and activity time series, respectively. These time series were then used to verify the ecological causal relationships between all ecological supernodes and all ecological superedges. The core definition of the Granger causality test algorithm is: if introducing historical data of variable X improves the prediction accuracy of future values ​​of variable Y, then X is called a Granger cause of Y. This method establishes a regression model with lag terms and uses the F-test to compare the sum of squared residuals of constrained and unconstrained models to determine the causal direction. For any pair of ecological supernodes and ecological superedges that are correlated in the marine-terrestrial ecological hypergraph, they are considered as potential causal pairs. Then, stationarity tests are performed on the historical stress time series and historical activity time series of each potential causal pair. If the historical stress time series or historical activity time series exhibits non-stationarity, it is adjusted to meet the stationarity requirement through differencing (first-order or second-order). When all series meet the stationarity requirement, the optimal lag order is determined. Considering the lag characteristics of the marine-terrestrial ecological environment, the minimum value of the Akaike Information Criterion is generally used as the lag order. This fully captures the lagged impact of ecological superedges on ecological supernodes while avoiding excessive model complexity. Next, two sets of models are constructed: a baseline model and an extended model. The baseline model uses the lag order and the historical stress time series or historical activity time series corresponding to the ecological supernodes to predict the stress index or activity index of the ecological supernode at the current time point. Then, the sum of squared residuals between the model predictions and the actual observed values ​​is calculated. The extended model, based on the baseline model, incorporates historical stress or activity time series corresponding to the ecological superedge to jointly predict the stress or activity index of the ecological supernode at the current time point. Then, the sum of squared residuals between the model's predicted values ​​and the actual observed values ​​is calculated. Next, an F-test is used to compare the sums of squared residuals of the two models. If the sum of squared residuals significantly decreases after adding the superedge variable, and the F-statistic is greater than the critical value for the corresponding degrees of freedom (obtained from the F-distribution table), and the p-value is less than 0.05, it is preliminarily determined that there is a statistically significant causal relationship between the superedge and the corresponding index of the supernode, i.e., an ecological causal relationship.

[0070] Next, based on ecological causality, the ecological propagation parameters between all ecological superedges and all ecological supernodes are calculated using a vector autoregression (VAR) model. Specifically, the variable set for the VAR model is first determined by incorporating the historical pressure and activity time series of ecological supernodes and superedges with ecological causal relationships into the VAR model as the variable set. Then, the lag order of the VAR model is determined. Similar to the Granger causality test, the VAR model also needs to choose an appropriate lag order to capture the time dependence between variables. This can also be determined using information criteria (such as AIC or BIC). Next, the VAR model is constructed and estimated. The basic idea of ​​the VAR model is to represent the current value of each variable as a linear combination of its own past value and the past values ​​of all other variables included in the model. Statistical methods such as least squares can be used to predict the parameters in the VAR model, including the lag regression coefficients and the disturbance term parameters. The lag regression coefficients refer to the weighting coefficients corresponding to different lag orders of the input variables, and the disturbance term parameters refer to the variance and covariance of the error term (disturbance term) in the regression equation. The vector autoregression model constructed through the above steps can accurately capture the degree of influence between ecological supernodes and ecological superedges, i.e., the ecological propagation parameters.

[0071] The historical pollution time data is obtained by extracting the time of occurrence of pollutants in the target marine and land areas from the water environment database. Then, data overlapping with or adjacent to the historical pollution time data are extracted from the historical ecological pressure index to obtain the pre-pollution pressure index, the pollution-intermediate pressure index, and the post-pollution pressure index. During the screening process, if data is missing, linear interpolation or the mean of adjacent time points can be used to fill in the gaps. Outliers caused by monitoring errors (such as values ​​deviating from adjacent data by more than three standard deviations) are removed. The series is smoothed using a moving average method to ensure the continuity of the ecological pressure index's changing trend, forming a continuous dataset of pressure change trajectory from pre-pollution baseline value to peak pressure and subsequent recovery period decay.

[0072] Next, the pollution attenuation model is fitted. First, an appropriate model type is selected based on the attenuation trajectory characteristics of the pressure change trajectory dataset. For example, if the attenuation trajectory shows a rapid decline in the early stage, a gradual slowdown in the later stage, and finally approaching the baseline value, an exponential attenuation model can be chosen. If the attenuation trajectory shows an S-shaped characteristic of a slow decline in the early stage, a faster decline in the middle stage, and a slowdown in the later stage, a logistic model can be chosen. After determining the model type, the least squares method is used for parameter estimation. With time as the independent variable and historical ecological pressure index observations as the dependent variable, an error function (i.e., the sum of squared errors between the theoretically calculated value and the observed value) is constructed. The parameters in the model (such as the attenuation coefficient in the exponential model and the growth coefficient in the logistic model) are adjusted through iterative calculations until the error function reaches its minimum value. The parameters at this minimum value are the optimal estimates, thus obtaining the pollution attenuation model. The non-intertidal or intertidal ecological pressure indices of all ecological supernodes are input into the pollution attenuation model. The ecological restoration parameters corresponding to the ecological supernodes are fitted using the pollution attenuation model. The ecological restoration parameters represent the rate of ecological pressure decline of the ecological supernode after pollution. The larger the ecological restoration parameters, the faster the ecological pressure declines after pollution, and the stronger the ecological restoration capacity of the supernode. Since there are significant differences in the ecological restoration capacity between the intertidal zone and the non-intertidal zone, different pollution attenuation models can be constructed for the two zones to achieve more accurate estimation of ecological restoration parameters.

[0073] Ecological propagation parameters and ecological restoration parameters are mapped as attributes of each ecological supernode or ecological superedge to complete the update of the marine-terrestrial ecological supergraph. Then, the marine-terrestrial ecological supergraph with completed parameter mapping is input into a marine-terrestrial ecological evolution model constructed based on the supergraph propagation dynamics model. Based on the marine-terrestrial ecological evolution model, the ecological pressure parameter sequence of all ecological supernodes is iteratively calculated. The marine-terrestrial ecological evolution model is as follows:

[0074] The first item on the right It is a valid recovery item, among which, These are the ecological restoration parameters of ecological supernode i. It is the ecological pressure index (non-intertidal ecological pressure index or intertidal ecological pressure index) of ecological supernode i at time point t. The second term on the right is the pressure propagation term, which represents the process by which the ecological supernode continuously increases and accumulates due to the pressure from other ecological supernodes propagating through the ecological superedge. This indicates the potential for accumulated pressure on ecological supernodes. This represents the set of all ecological supernodes connected to the ecological supernode. The effective higher-order propagation rate of ecological superedge h is expressed as: Effective higher-order propagation rate = Higher-order propagation rate × Ecological propagation parameter × (1 - Mean resource index). The mean resource index refers to the mean of the non-intertidal resource index or intertidal resource index of all ecological supernodes associated with ecological superedge h. The order of the ecological superedge h represents the order of the superedge. The higher the order of the superedge, the stronger the ability of multi-node collaborative propagation. The sum of the ecological pressure indices of all ecological supernodes (excluding ecological supernode i) within ecological superedge h represents the synergistic effect of pressure within ecological superedge h. That is, the higher the pressure index of other ecological supernodes connected to the ecological superedge, the higher the pressure propagated to that ecological supernode. By inputting the parameters from the marine-terrestrial ecological hypergraph into the aforementioned marine-terrestrial ecological evolution model, the sequence of ecological pressure parameters over a future period can be solved. The solution process simulates the continuous propagation and accumulation of ecological pressure.

[0075] This application also provides a big data-based ecological environment monitoring system, including: The memory is configured to store instructions; and The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement a big data-based ecological environment monitoring and early warning method according to any of the above.

[0076] The processor can be a central processing unit (CPU). Of course, depending on the actual use, it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not limit it.

[0077] The memory can be an internal storage unit of a computer device, such as a hard disk or RAM, or an external storage device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD), or flash memory card (FC) provided on the computer device. Furthermore, the memory can be a combination of internal storage units and external storage devices of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or will be output. This application does not limit this.

[0078] This application also provides a machine-readable storage medium storing instructions that cause a machine to execute the above-described big data-based ecological environment monitoring and early warning method.

[0079] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0081] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0082] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0083] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0084] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0085] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0086] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0087] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. An ecological environment monitoring and early warning method based on big data, characterized in that, The method comprises the following steps: Collecting marine and terrestrial ecological monitoring data of all marine and terrestrial monitoring points in the target marine and terrestrial region, and obtaining historical marine and terrestrial ecological data of the target marine and terrestrial region through a pre-constructed water environment database; According to the historical marine and terrestrial ecological data, all monitoring indicators in the historical marine and terrestrial ecological data are matched, and the target marine and terrestrial region is divided into an intertidal zone and a non-intertidal zone based on the matching results and the marine and terrestrial ecological monitoring data; Based on the marine and terrestrial ecological monitoring data, a multi-dimensional feature vector of the intertidal zone is constructed, and an intertidal topology graph is constructed in combination with the multi-dimensional feature vector and all marine and terrestrial monitoring points in the intertidal zone. The intertidal zone is divided into sub-regions based on the intertidal topology graph, and a plurality of intertidal sub-regions are obtained; The ecological area parameters of the intertidal zone and all intertidal sub-regions in the target marine and terrestrial region are calculated; The ecological environment early warning report of the target marine and terrestrial region is generated in combination with the marine and terrestrial ecological monitoring data and the ecological area parameters.

2. The method of claim 1, wherein, The monitoring indicators include marine indicators and land indicators. According to the historical marine and terrestrial ecological data, all monitoring indicators in the historical marine and terrestrial ecological data are matched, and the target marine and terrestrial region is divided into an intertidal zone and a non-intertidal zone based on the matching results and the marine and terrestrial ecological monitoring data, which comprises the following steps: According to the historical marine and terrestrial ecological data, all marine indicators and land indicators in the historical marine and terrestrial ecological data are matched respectively, and a plurality of marine indicator pairs and land indicator pairs are obtained; For any marine indicator pair and land indicator pair, the corresponding historical marine indicator pair data and historical land indicator pair data in the historical marine and terrestrial ecological data are extracted respectively; Based on the Gaussian mixture model, a marine indicator distribution model and a land indicator distribution model are constructed respectively; The model training of the marine indicator distribution model is completed by using the historical marine indicator pair data, and a plurality of marine component parameter sets are output; For any marine component parameter set, a two-dimensional confidence ellipse is defined according to the marine component parameter set and by using a pre-obtained chi-square distribution critical value; The marine indicator confidence interval of the marine indicator pair is determined according to the union set of all two-dimensional confidence ellipses; The land indicator distribution model is trained by using the historical land indicator pair data, and a plurality of land component parameter sets are output by the trained land indicator distribution model; Based on the land component parameter set and by using the Mahalanobis distance detection algorithm, the land indicator confidence interval of the land indicator pair is calculated; The target marine and terrestrial region is divided into an intertidal zone and a non-intertidal zone in combination with the marine indicator confidence interval, the land indicator confidence interval, and the marine and terrestrial ecological monitoring data.

3. The method of claim 2, wherein, The matching of all marine indicators and land indicators in the historical marine and terrestrial ecological data is completed respectively according to the historical marine and terrestrial ecological data, and a plurality of marine indicator pairs and land indicator pairs are obtained, which comprises the following steps: The historical land ecological data and historical marine ecological data in the historical marine and terrestrial ecological data are extracted; Time series analysis is performed on the historical land ecological data and the historical marine ecological data respectively, and historical land sequences and historical marine sequences are obtained; Distribution statistics are performed on all marine indicators in the historical marine sequence, and marine fluctuation sequences of all marine indicators are calculated according to the distribution statistics results; Calculate the land fluctuation sequence of all land indexes in the historical land sequence based on the clustering algorithm; Calculate the fluctuation synchronization degree between all land indexes according to the land fluctuation sequence, and complete the index pair matching between all land indexes according to the fluctuation synchronization degree, to obtain a plurality of land index pairs; Complete the index pair matching between all marine indexes based on the marine fluctuation sequence, to obtain a plurality of marine index pairs.

4. The method of claim 3, wherein, The calculation of the land fluctuation sequence of all land indexes in the historical land sequence based on the clustering algorithm comprises the following steps: For any land index, divide the historical land sequence into all land index subsequences of the land index in chronological order; For any land index subsequence, divide the land index subsequence according to a preset time window to obtain a plurality of index data sets; For any index data set, use the clustering algorithm to complete data clustering of the index data set, and filter out high-density clustering clusters as core distribution clusters from the data clustering results; Define the land index interval of the land index according to the percentile range of the core distribution cluster; Perform kernel density estimation on the core distribution cluster of the index data set to obtain the index probability density function of the index data set; Determine a plurality of neighborhood data sets of the index data set according to the time window division order, calculate the average probability density function of all neighborhood data sets, and combine the index probability density function and the average probability density function to calculate the basic fluctuation degree of the index data set; Calculate the mean value of the basic fluctuation degree of all index data sets and the mean value of the width of the land index interval to obtain the fluctuation mean value and the interval width mean value; Extract the center points of the land index intervals of all index data sets, and calculate the center drift degree between all adjacent index data sets according to the center point extraction results; Combine all fluctuation mean values, interval width mean values and center drift degrees to calculate the land fluctuation sequence of all land indexes.

5. The method of claim 2, wherein, The combination of the marine index confidence interval, the land index confidence interval and the marine-terrestrial ecological monitoring data divides the target marine-terrestrial area into an intertidal zone and a non-intertidal zone, which comprises the following steps: For any marine-terrestrial monitoring point, extract all land pairs of monitoring data and marine pairs of monitoring data corresponding to all land index pairs and marine index pairs in the marine-terrestrial ecological monitoring data of the marine-terrestrial monitoring point; Statistically analyze the distribution characteristics of all marine pairs of monitoring data within the corresponding marine index confidence interval, and calculate the marine deviation degree of the marine-terrestrial monitoring point according to the distribution characteristics; Statistically analyze the distribution characteristics of all land pairs of monitoring data within the corresponding land index confidence interval, and calculate the land deviation degree of the marine-terrestrial monitoring point according to the distribution characteristics; Combine the marine deviation degrees and land deviation degrees of all marine-terrestrial monitoring points to divide the target marine-terrestrial area into an intertidal zone and a non-intertidal zone, wherein the non-intertidal zone includes a marine area and a land area.

6. The method of claim 1, wherein, The construction of the multi-dimensional feature vector of the intertidal zone based on the marine-terrestrial ecological monitoring data, the construction of the intertidal topology graph combining the multi-dimensional feature vector and all marine-terrestrial monitoring points in the intertidal zone, and the sub-region division of the intertidal zone based on the intertidal topology graph to obtain a plurality of intertidal sub-regions comprise the following steps: Extracting intertidal zone monitoring data collected by all sea-land monitoring points in the intertidal zone area from the sea-land monitoring data; Respectively, the time series analysis of all intertidal zone monitoring data is performed to obtain multiple intertidal zone sequences; Obtaining the tidal period data of the target sea-land area through the water environment database; Aligning the timestamps of all intertidal zone sequences, and performing time series division on all intertidal zone sequences with completed timestamp alignment according to the tidal period data, and constructing multiple tidal zone data sets located in different tidal time periods according to the time series division results; For any tidal zone data set, complete multi-dimensional feature extraction of the tidal zone data set to obtain the tidal multi-dimensional features of all sea-land monitoring points in the intertidal zone area; Connecting all adjacent topological nodes to construct an intertidal topological graph; According to the tidal multi-dimensional features and using the similarity formula, the node feature similarity between all adjacent topological nodes is calculated, and the edge weight of the intertidal topological graph is determined according to the node feature similarity; Arranging all intertidal topological graphs with completed edge weight distribution in time sequence to obtain an intertidal topological graph sequence; Using a dynamic community detection algorithm to complete the node community search of the intertidal topological graph sequence to obtain the community attribution sequence of all topological nodes; According to the community attribution sequence, the community stability and community stable period of all topological nodes are analyzed, and the intertidal zone attribution of all topological nodes is determined in combination with the community stability and community stable period; According to the intertidal zone attribution of all topological nodes, the intertidal zone area is divided into multiple intertidal sub-areas, including a land-side transition area, a core transition area, and a sea-side transition area.

7. The method of claim 1, wherein, The steps of generating an ecological environment early warning report of a target sea-land area by combining sea-land ecological monitoring data and ecological area parameters include: Constructing a sea-land ecological hypergraph of the target sea-land area based on the sea-land ecological monitoring data; Identifying the ecological causal relationship in the sea-land ecological hypergraph according to the historical sea-land ecological data and using the Granger causality test algorithm, and calculating the ecological pressure parameter sequence of all ecological hypernodes in the sea-land ecological hypergraph based on the ecological causal relationship and the pre-constructed sea-land ecological evolution model; The division results of each sea-land sub-area in the target sea-land area divide the ecological pressure parameter sequence into multiple regional pressure parameter sequences; For any regional pressure parameter sequence, extracting the multi-dimensional pressure features of the regional pressure parameter sequence, wherein the multi-dimensional pressure features include pressure trend features, pressure steady-state features, and pressure propagation features; According to the multi-dimensional pressure features, the ecological risk assessment of the sea-land sub-area is completed, and the ecological warning level and the pressure propagation path of the sea-land sub-area are generated according to the ecological risk assessment results; Integrating the ecological warning levels and pressure propagation paths of all sea-land sub-areas to generate an ecological environment early warning report of the target sea-land area.

8. The method of claim 7, wherein, The steps of constructing a sea-land ecological hypergraph of a target sea-land area based on sea-land ecological monitoring data include: Defining each sea-land sub-area in the target sea-land area as an ecological hypernode, wherein the sea-land sub-area includes a marine area, a land area, a land-side transition area, a core transition area, and a sea-side transition area; extracting non-tidal zone monitoring data and intertidal zone monitoring data from the marine and terrestrial ecological monitoring data; calculating non-tidal ecological stress indexes and non-tidal resource indexes of the ecological super-nodes corresponding to the non-tidal zone regions according to the non-tidal zone monitoring data; calculating intertidal ecological stress indexes and intertidal resource indexes of the intertidal zone regions in combination with the intertidal zone monitoring data and the ecological area parameters; integrating the non-tidal ecological stress indexes, the non-tidal resource indexes, the intertidal ecological stress indexes and the intertidal resource indexes into super-node attribute parameters, and mapping the super-node attribute parameters to the ecological super-nodes; extracting regional ecological interaction relationship chains between all marine and terrestrial sub-regions based on the marine and terrestrial ecological monitoring data and historical marine and terrestrial ecological data, and defining the regional ecological interaction relationship chains as ecological super-edges, wherein the regional ecological interaction relationship includes pollutant runoff chains, climate impact chains and biological migration chains; calculating super-edge attribute parameters of all ecological super-edges in combination with the non-tidal ecological stress indexes, the non-tidal resource indexes, the intertidal ecological stress indexes and the intertidal resource indexes, and mapping the super-edge attribute parameters to the corresponding ecological super-edges; constructing a marine and terrestrial ecological super-graph based on all the ecological super-nodes with completed super-node attribute parameter mapping and all the ecological super-edges with completed super-edge attribute parameter mapping.

9. The method of claim 8, wherein, The ecological causal relationships in the marine and terrestrial ecological super-graph are identified according to the historical marine and terrestrial ecological data and by using a Granger causality test algorithm, and the ecological stress parameter sequences of all the ecological super-nodes in the marine and terrestrial ecological super-graph are calculated based on the ecological causal relationships and a pre-constructed marine and terrestrial ecological evolution model, including the following steps: calculating historical ecological stress indexes and historical ecological activity indexes of all the ecological super-nodes in the marine and terrestrial ecological super-graph according to the historical marine and terrestrial ecological data; identifying ecological causal relationships between all the ecological super-nodes and all the ecological super-edges in combination with the historical stress indexes and the historical activity indexes and by using a Granger causality test algorithm; calculating ecological propagation parameters between all the ecological super-edges and all the ecological super-nodes according to the ecological causal relationships and by using a vector autoregression model; obtaining historical pollution time data of a target marine and terrestrial region from a water environment database, fitting a pollution decay model of the target marine and terrestrial region in combination with the historical pollution time data and the historical ecological stress indexes; inputting the non-tidal ecological stress indexes or the intertidal ecological stress indexes of all the ecological super-nodes into the pollution decay model, and sequentially outputting ecological restoration parameters of all the ecological super-nodes by the pollution decay model; mapping the ecological propagation parameters and the ecological restoration parameters to the marine and terrestrial ecological super-graph, and inputting the marine and terrestrial ecological super-graph with completed parameter mapping into the pre-constructed marine and terrestrial ecological evolution model, iteratively calculating the ecological stress parameter sequences of all the ecological super-nodes based on the marine and terrestrial ecological evolution model, wherein the marine and terrestrial ecological evolution model is constructed based on a super-graph propagation dynamics model.

10. An ecological environment monitoring system based on big data, characterized in that, comprise: a memory configured to store instructions; and a processor configured to call the instructions from the memory and enable implementation of the big data-based ecological environment monitoring and early warning method according to any one of claims 1 to 9 when the instructions are executed.

Images