Carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method

Abstract

The invention discloses a carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method, and relates to the technical field of carbon emission early warning, and the method comprises the steps: calculating a carbon emission quota of coupled carbon sink bearing capacity, a carbon sink change rate, a three-layer score of a PSR model, and a coupling coordination degree; the three-layer score of the PSR model comprises a pressure layer score, a state layer score and a response layer score; inputting the carbon emission quota, the carbon sink change rate, the three-layer score of the PSR model, the coupling coordination degree and the carbon emission into a pre-constructed decision tree; and carbon emission early warning grades are output in a grading manner by using the decision tree. According to the method, a carbon sink-quota-PSR deep coupling + adaptive threshold adjustment early warning scheme is constructed, the blank in the prior art is filled, a multi-factor coupling-dynamic judgment-graded early warning-decision support integrated technical scheme is formed, seamless connection from risk identification to management and control execution can be realized, and scientific support is provided for regional carbon emission reduction decision.

Description

Technical Field

[0001] This application relates to the field of carbon emission early warning technology, and in particular to a carbon emission classification and early warning method based on a carbon sink-quota-PSR coupling decision tree. Background Technology

[0002] In the current context, precise monitoring and dynamic early warning of carbon emissions at the regional scale have become a key support for achieving the goal of carbon peaking and carbon neutrality. Regional carbon emissions are affected by a combination of factors, including industrial production, transportation, residential consumption, and the ecological environment, and are characterized by strong spatial and temporal heterogeneity and significant dynamic fluctuations. In particular, at the scale of small and medium-sized cities, due to rapid industrial restructuring and high ecosystem vulnerability, the need for more refined and dynamic carbon emission early warning is even more urgent.

[0003] Currently, existing technologies for regional carbon emission early warning have developed various technical solutions, but they also have significant limitations: traditional early warning methods (such as partial least squares and backpropagation neural networks) rely heavily on single-dimensional indicators (such as carbon emission intensity and CO2 concentration), failing to achieve deep coupling between carbon sinks (ecological regulation), quotas (emission baselines), and PSR (human activity-environmental state-emission reduction response), resulting in incomplete coverage of risk factors and prominent one-sidedness in early warning; existing decision tree applications use fixed threshold judgments, which cannot adapt to dynamic characteristics such as differences in regional ecological carrying capacity and temporal emission fluctuations, and have poor adaptability to the heterogeneity of small and medium-sized cities; although innovative technologies (such as the Internet of Things and digital twins) have made breakthroughs at the monitoring level, they have not formed an integrated framework of "multi-dimensional factor coupling - dynamic threshold judgment - hierarchical early warning," resulting in insufficient accuracy and practicality in early warning. PSR models, as mature ecological and environmental assessment frameworks, have been widely used in ecological security assessments, but they have not been deeply integrated with dynamic carbon sink monitoring and carbon emission quota management; although decision tree models have the advantage of hierarchical decision-making, the traditional static threshold design limits their application in dynamic carbon emission scenarios. Summary of the Invention

[0004] In view of this, embodiments of this application provide a carbon emission classification and early warning method that couples carbon sink, quota, and PSR to accurately predict carbon emission levels.

[0005] One aspect of this application provides a carbon emission classification and early warning method coupled with carbon sink-quota-PSR, the method comprising the following steps:

[0006] Calculate the carbon emission allowance, carbon sink change rate, three-layer score of the PSR model, and coupling coordination degree of the coupled carbon sink carrying capacity; wherein the three-layer score of the PSR model includes the pressure layer score, the state layer score, and the response layer score.

[0007] The carbon emission quota, the carbon sink change rate, the three-layer score of the PSR model, the coupling coordination degree, and the carbon emission amount are input into a pre-constructed decision tree; the decision tree includes a root node, multiple internal nodes, and multiple leaf nodes;

[0008] The root node is used to determine whether the carbon emissions exceed the carbon emission quota.

[0009] If the carbon emissions do not exceed the carbon emission quota, a risk-free warning level is output using the first leaf node; if the carbon emissions exceed the carbon emission quota, the coupling coordination degree is determined using the first internal node to see if it exceeds the first dynamic threshold.

[0010] If the coupling coordination degree does not exceed the first dynamic threshold, the first leaf node is used to output the risk-free warning level; if the coupling coordination degree is greater than the first dynamic threshold, the second internal node is used to determine that the carbon sink change rate exceeds the second dynamic threshold.

[0011] If the carbon sink change rate does not exceed the second dynamic threshold, a low-risk warning level is output using the second leaf node; if the carbon sink change rate exceeds the second dynamic threshold, the pressure layer score is determined using the third internal node to see if it exceeds the third dynamic threshold.

[0012] If the pressure layer score does not exceed the third dynamic threshold, the third leaf node is used to output a medium risk warning level; if the pressure layer score exceeds the third dynamic threshold, the fourth internal node is used to determine whether the state layer score exceeds the fourth dynamic threshold.

[0013] If the state layer score does not exceed the fourth dynamic threshold, the medium-risk warning level is output using the third leaf node; if the state layer score exceeds the fourth dynamic threshold, the response layer score is determined using the fifth internal node.

[0014] If the response layer score does not exceed the fifth dynamic threshold, the medium-risk warning level is output using the third leaf node; if the response layer score exceeds the fifth dynamic threshold, the high-risk warning level is output using the fourth leaf node.

[0015] In some embodiments, the carbon emission allowance is calculated using the following formula:

[0016] Q = Q0 × (1 + λ × R) s );

[0017] Where Q is the carbon emission allowance, Q0 is the basic allowance based on regional economic level and historical emission data, λ is the carbon sink carrying capacity correction coefficient, and R sThis represents the carbon sink change rate.

[0018] In some embodiments, the carbon sink change rate is calculated using the following formula:

[0019] Carbon sequestration change rate = [(current period carbon sequestration – base period carbon sequestration) / base period carbon sequestration] × 100%.

[0020] In some embodiments, the three-level score of the PSR model is calculated through the following steps:

[0021] The score weights of each layer are calculated using the entropy weight method combined with the coupling coordination degree model.

[0022] The scores of each layer are weighted and summed with their corresponding indicators to calculate the scores of the pressure layer, the state layer, and the response layer.

[0023] In some embodiments, the coupling coordination degree is calculated using the following formula:

[0024] C=√[(U1×U2×U3) / (( U1+ U2+ U3) / 3) 3 ];

[0025] Wherein, C is the coupling coordination degree, U1 is the quota adaptation degree, U2 is the carbon sink adjustment degree, and U3 is the PSR synergy degree.

[0026] In some embodiments, the method further includes the following steps:

[0027] A 12-month sliding window is used to update each of the dynamic thresholds monthly.

[0028] Differentiated dynamic thresholds for each benchmark are set based on regional ecological type and economic development level;

[0029] The coupling coordination degree is used to dynamically correct each of the reference dynamic thresholds; wherein each of the reference dynamic thresholds is a reference value for each of the dynamic thresholds.

[0030] In some embodiments, the method further includes the following steps:

[0031] When the accuracy of comparing each warning level with the actual emission exceedance is lower than the accuracy threshold, a threshold optimization step is triggered to adjust the threshold weight of each node.

[0032] Another aspect of this application embodiment provides a carbon sink-quota-PSR coupled decision tree carbon emission classification early warning device, the device comprising:

[0033] The data calculation unit is used to calculate the carbon emission quota, carbon sink change rate, three-layer score of the PSR model, and coupling coordination degree of the coupled carbon sink carrying capacity; wherein, the three-layer score of the PSR model includes the pressure layer score, the state layer score, and the response layer score.

[0034] The tree input unit is used to input the carbon emission quota, the carbon sink change rate, the three-layer score of the PSR model, the coupling coordination degree, and the carbon emission amount into a pre-constructed decision tree; the decision tree includes a root node, multiple internal nodes, and multiple leaf nodes;

[0035] The first judgment unit is used to determine whether the carbon emissions exceed the carbon emission quota using the root node;

[0036] The second judgment unit is used to output a risk-free warning level using the first leaf node if the carbon emissions do not exceed the carbon emission quota; and to determine whether the coupling coordination degree exceeds the first dynamic threshold using the first internal node if the carbon emissions exceed the carbon emission quota.

[0037] The third judgment unit is used to output the risk-free warning level using the first leaf node if the coupling coordination degree does not exceed the first dynamic threshold; and to determine the carbon sink change rate exceeds the second dynamic threshold using the second internal node if the coupling coordination degree is greater than the first dynamic threshold.

[0038] The fourth judgment unit is used to output a low-risk warning level using the second leaf node if the carbon sink change rate does not exceed the second dynamic threshold; and to determine whether the pressure layer score exceeds the third dynamic threshold using the third internal node if the carbon sink change rate exceeds the second dynamic threshold.

[0039] The fifth judgment unit is used to output a medium risk warning level using the third leaf node if the pressure layer score does not exceed the third dynamic threshold; and to determine whether the state layer score exceeds the fourth dynamic threshold using the fourth internal node if the pressure layer score exceeds the third dynamic threshold.

[0040] The sixth judgment unit is used to output the medium-risk warning level using the third leaf node if the state layer score does not exceed the fourth dynamic threshold; and to determine whether the response layer score exceeds the fifth dynamic threshold using the fifth internal node if the state layer score exceeds the fourth dynamic threshold.

[0041] The seventh judgment unit is used to output the medium-risk warning level using the third leaf node if the response layer score does not exceed the fifth dynamic threshold; and to output the high-risk warning level using the fourth leaf node if the response layer score exceeds the fifth dynamic threshold.

[0042] Another aspect of this application embodiment provides an electronic device, including a processor and a memory;

[0043] The memory is used to store programs;

[0044] The processor executes the program to implement any of the methods described above.

[0045] Another aspect of this application provides a computer-readable storage medium storing a program that is executed by a processor to implement the method described in any of the above embodiments.

[0046] This application includes at least the following beneficial effects:

[0047] This application calculates the carbon emission allowance, carbon sink change rate, three-layer score of the PSR model, and coupling coordination degree of the coupled carbon sink carrying capacity. The three-layer score of the PSR model includes the pressure layer score, state layer score, and response layer score. The carbon emission allowance, carbon sink change rate, three-layer score of the PSR model, coupling coordination degree, and carbon emission amount are input into a pre-constructed decision tree. The decision tree is used to output the carbon emission early warning level in a hierarchical manner. This application constructs an early warning scheme with deep coupling of carbon sink-allowance-PSR and adaptive threshold adjustment, filling the gap in existing technology and forming an integrated technical solution of multi-factor coupling-dynamic judgment-hierarchical early warning-decision support. It can achieve seamless connection from risk identification to control execution and provide scientific support for regional carbon emission reduction decision-making. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 An adaptive decision tree early warning rule diagram provided in the embodiments of this application. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0051] Before providing a detailed description of the embodiments of this application, some related technologies involved in the embodiments of this application will be described first, as follows:

[0052] Terminology Explanation:

[0053] 1. Adaptive Decision Tree: A classification model with dynamic threshold adjustment capability. Its structure includes a root node, internal nodes, and leaf nodes. It optimizes the judgment threshold in real time by adapting to time-series data feedback and regional features, thereby realizing dynamic hierarchical early warning of regional carbon emission risks.

[0054] 2. Carbon sink-quota-PSR coupling: A collaborative analysis framework that deeply integrates dynamic changes in carbon sinks, carbon emission quota constraints, and PSR (Pressure-State-Response) multi-dimensional indicators. By quantifying the interaction among the three, it comprehensively covers the entire chain of risk factors from "emission baseline to ecological regulation to human activities to environmental status to emission reduction response".

[0055] 3. Regional carbon emissions: refers to the total emissions of greenhouse gases such as carbon dioxide in a specific geographical area (such as a city or county) within a given time period, usually measured in tons.

[0056] 4. Carbon emission quota: The allowable carbon emission amount is estimated based on the regional economic level, ecological carrying capacity (carbon sink capacity) and historical emission data, and serves as the core benchmark value for early warning.

[0057] 5. Carbon sequestration change rate: This indicator reflects the change in carbon sequestration capacity over time. A positive value indicates an increase in carbon sequestration capacity, while a negative value indicates a decrease. It is a core input feature of the ecological regulation dimension.

[0058] 6. Pressure Layer Score: The score is calculated using standardized regional carbon emission intensity, industrial structure, energy consumption, and other indicators. The lower the score, the greater the carbon emission pressure.

[0059] 7. Status layer score: The score is calculated using standardized environmental and socioeconomic indicators such as vegetation cover, air quality, and ecosystem health. The lower the score, the worse the environmental status.

[0060] 8. Response Layer Score: The score is calculated using standardized indicators of government emission reduction investment, enterprise energy efficiency improvement, and renewable energy substitution ratio. The higher the score, the more effective the response measures are.

[0061] 9. Warning levels: divided into four levels: green, yellow, orange, and red, which represent different risk states of the carbon balance system.

[0062] 10. Root node: The starting point of the decision tree, corresponding to the first judgment condition, namely whether the carbon emissions exceed the carbon emission quota of the coupled carbon sink carrying capacity.

[0063] 11. Internal nodes: intermediate decision points in the decision tree, corresponding to dynamic threshold comparisons of carbon sink change rate, pressure layer score, state layer score, and response layer score.

[0064] 12. Leaf node: The terminal node of the decision tree, corresponding to the final warning level output.

[0065] 13. Dynamic threshold: A numerical limit that is adaptively adjusted based on time-series sliding window data, regional feature differences, and early warning feedback results. It is used for conditional judgment in decision trees, which is different from traditional fixed thresholds.

[0066] 14. Coupling Coordination Degree: An indicator that quantifies the degree of interaction and coordination among carbon sinks, quotas, and PSR. The value ranges from 0 to 1, with the closer to 1 indicating stronger coupling. It is used to optimize the weight allocation of input features.

[0067] While existing technologies provide some support for regional carbon emission early warning, the following significant shortcomings still exist for dynamic tiered early warning scenarios at the scale of small and medium-sized cities:

[0068] Insufficient dynamic adaptability: Existing early warning methods are mostly based on fixed thresholds or static data modeling, such as K-means clustering analysis and grey weighted correlation degree method, which do not fully consider the temporal change trend of regional carbon emissions, cannot achieve dynamic tracking of risks and real-time adjustment of early warning levels, and are difficult to adapt to the characteristics of large fluctuations in carbon emissions in small and medium-sized cities.

[0069] The grading logic is unclear: existing decision tree applications have not formed a systematic grading rule for regional carbon emissions, have not combined core benchmarks such as carbon emission quotas and carbon sink change rates, and have not fully integrated the multi-dimensional indicators of stress, state, and response of the PSR model, resulting in a lack of hierarchy and scientificity in the determination of early warning levels, making it difficult to accurately distinguish the control needs of different risk levels.

[0070] Insufficient multi-factor coupling: Existing methods do not deeply integrate carbon sinks (ecological regulation), quotas (emission baselines), and PSR (human activity-environmental status-emission reduction response), and only cover risk factors individually or partially, resulting in a weak ability to systematically determine carbon emission risks;

[0071] Limitations of static thresholds: The judgment thresholds of traditional early warning methods are mostly fixed based on historical data, which cannot adapt to the differences in regional ecological carrying capacity, the fluctuation of time-series emissions, and the dynamic changes in the effects of emission reduction measures, and are prone to misjudgment and omission.

[0072] Poor adaptability to small and medium-sized cities: Existing research mostly focuses on macro-regions such as provinces and urban agglomerations. The indicator system and early warning rules do not fully consider the characteristics of small and medium-sized cities, such as simple industrial structure, high ecological vulnerability, and difficulty in data acquisition. For example, the indicator settings are too complex and the data dependence is strong, which makes it difficult to implement technical solutions in small and medium-sized cities.

[0073] Insufficient accuracy of early warning: Due to the lack of multi-factor collaborative judgment and the limitation of static threshold, existing methods are unable to capture the complex formation mechanism of carbon emission risks, and the early warning results deviate significantly from the actual risk status.

[0074] In view of the above-mentioned deficiencies in the prior art, this application aims to solve the following core technical problems:

[0075] Constructing a deep-coupled input feature system of carbon sink, quota, and PSR: breaking through the limitations of single-dimensional indicators, integrating the entire chain of risk factors of "emission baseline (quota) - ecological regulation (carbon sink) - human activities (pressure) - environmental status (status) - emission reduction response (response)" to achieve the coordinated capture of multi-dimensional risks;

[0076] Develop adaptive decision tree technology: Design a dynamic threshold adjustment mechanism that optimizes the judgment threshold in real time based on time-series sliding window data, regional coupling coordination degree and early warning feedback results, so as to solve the problem of poor adaptability of static thresholds;

[0077] Construct a dynamic hierarchical early warning framework adapted to small and medium-sized cities: break through the limitations of static early warning, combine time-series data changes and real-time indicator monitoring, design dynamically adjusted early warning rules, realize real-time tracking and hierarchical alerts of regional carbon emission risks, and adapt to the characteristics of large carbon emission fluctuations in small and medium-sized cities; at the same time, improve the accuracy of carbon emission early warning for small and medium-sized cities through multi-factor coupling and dynamic threshold adjustment.

[0078] An integrated technical solution of "multi-factor coupling, dynamic judgment, hierarchical early warning and decision support" has been formed: achieving seamless connection from risk identification to control and management execution, and providing scientific support for regional carbon emission reduction decision-making.

[0079] Enhance the practicality and operability of technical solutions: Optimize the indicator system and calculation process based on the data characteristics of small and medium-sized cities to ensure data accessibility; at the same time, establish a corresponding mechanism of "early warning level - control recommendations" to accurately locate high-risk links and provide direct support for regional emission reduction decisions.

[0080] This application provides a carbon emission classification and early warning method coupled with carbon sink-quota-PSR, specifically including the following steps S100~S180:

[0081] S100: Calculate the carbon emission allowance, carbon sink change rate, three-layer score of the PSR model, and coupling coordination degree of the coupled carbon sink carrying capacity; wherein, the three-layer score of the PSR model includes the pressure layer score, the state layer score, and the response layer score.

[0082] S110: Input the carbon emission quota, the carbon sink change rate, the three-layer score of the PSR model, the coupling coordination degree, and the carbon emission amount into a pre-constructed decision tree; the decision tree includes a root node, multiple internal nodes, and multiple leaf nodes;

[0083] S120: Use the root node to determine whether the carbon emissions exceed the carbon emission quota;

[0084] S130: If the carbon emissions do not exceed the carbon emission quota, then the first leaf node outputs a risk-free warning level; if the carbon emissions exceed the carbon emission quota, then the first internal node determines whether the coupling coordination degree exceeds the first dynamic threshold.

[0085] S140: If the coupling coordination degree does not exceed the first dynamic threshold, the risk-free warning level is output using the first leaf node; if the coupling coordination degree is greater than the first dynamic threshold, the carbon sink change rate is determined to exceed the second dynamic threshold using the second internal node.

[0086] S150: If the carbon sink change rate does not exceed the second dynamic threshold, then the second leaf node is used to output a low-risk warning level; if the carbon sink change rate exceeds the second dynamic threshold, then the third internal node is used to determine whether the pressure layer score exceeds the third dynamic threshold.

[0087] S160: If the pressure layer score does not exceed the third dynamic threshold, the risk warning level is output using the third leaf node; if the pressure layer score exceeds the third dynamic threshold, the state layer score is determined using the fourth internal node.

[0088] S170: If the state layer score does not exceed the fourth dynamic threshold, the medium-risk warning level is output using the third leaf node; if the state layer score exceeds the fourth dynamic threshold, the response layer score is determined using the fifth internal node.

[0089] S180: If the response layer score does not exceed the fifth dynamic threshold, the medium-risk warning level is output using the third leaf node; if the response layer score exceeds the fifth dynamic threshold, the high-risk warning level is output using the fourth leaf node.

[0090] The following section will provide a detailed introduction and explanation of the solutions in the embodiments of this application, using specific application examples.

[0091] Figure 1 This is an adaptive decision tree early warning rule graph. The root node uses "whether carbon emissions exceed the quota of coupled carbon sink carrying capacity" as the initial judgment condition, while also filtering low-risk scenarios based on coupling coordination degree. Internal nodes sequentially compare dynamic thresholds of carbon sink change rate, pressure layer score, state layer score, and response layer score to achieve progressive judgment of multi-dimensional risk factors. Leaf nodes output four-level early warning levels, clarifying the risk combination characteristics corresponding to each level, providing accurate basis for the formulation of control measures. All dynamic thresholds can be adaptively adjusted based on regional characteristics and time series data.

[0092] Specifically, this embodiment provides an adaptive decision tree-based dynamic grading and early warning method for regional carbon emissions, deeply coupled with carbon sinks, quotas, and PSR. The core innovation lies in constructing a full-dimensional input feature system through deep coupling of carbon sinks, quotas, and PSR, combined with the dynamic threshold adjustment technology of adaptive decision trees, to achieve refined and dynamic grading and early warning of regional carbon emission risks. This method constructs a multi-level adaptive decision tree model, using the coupled features of carbon sinks, quotas, and PSR as input, and outputs four warning levels—green, yellow, orange, and red—through dynamic conditional judgment, providing targeted control suggestions for each level.

[0093] The warning levels are indicated as follows:

[0094] Green Alert: Carbon emissions are ≤ the quota for coupled carbon sink carrying capacity, the carbon sink change rate is positive, and the scores for the pressure layer, state layer, and response layer are all within a reasonable range, indicating that the carbon balance system is operating well. It is recommended to maintain the existing production plan and emission reduction measures, and to continue monitoring key indicators.

[0095] Yellow alert: Carbon emissions are approaching the quota coupled with carbon sink carrying capacity, or the carbon sink change rate has slightly decreased, or a single PSR indicator is approaching the dynamic threshold, posing potential risks. It is recommended to strengthen real-time monitoring of core indicators and optimize the targeting of emission reduction measures.

[0096] Orange alert: Carbon emissions exceed the quota for coupled carbon sink carrying capacity, the carbon sink change rate has decreased significantly, the pressure layer score is below the dynamic threshold and the state layer score is poor, and response layer measures are partially effective. It is recommended to immediately activate the emergency plan, implement controls on high-emission processes, and increase investment in emission reduction.

[0097] Red Alert: Carbon emissions significantly exceed the quota for coupled carbon sink carrying capacity; the carbon sink change rate has dropped drastically (negative and with a large absolute value); both the pressure layer and state layer scores are significantly below the dynamic threshold; and response layer measures have failed. It is recommended to initiate comprehensive emergency control measures, suspend high-emission projects, and implement the strictest emission reduction measures.

[0098] Specifically, the following technical solutions are included:

[0099] 1. Construction of an input feature system with deep coupling of carbon sink, quota, and PSR.

[0100] The input feature system aims to "maximize coupling coordination," integrating three core dimensions: carbon sequestration, quotas, and PSR. It quantifies interaction relationships through a coupling matrix, specifically including:

[0101] Core benchmark feature: carbon emission allowance (Q) coupled with carbon sink carrying capacity.

[0102] Calculation formula: Q = Q0 × (1 + λ × R) s );

[0103] Where Q0 is the basic allowance based on regional economic level and historical emission data; λ is the carbon sink carrying capacity correction coefficient (determined based on regional ecosystem type, with a value of 0.1-0.3); R s This represents the carbon sink change rate (average over the past 3 years). This allowance fully considers the ecological regulatory role of carbon sinks, unlike traditional allowance calculation methods that rely solely on economic and historical data.

[0104] Ecological regulation characteristic: Carbon sink change rate. The carbon sink change rate reflects the dynamic change in carbon sink capacity, and its calculation formula is as follows:

[0105] Carbon sequestration change rate (%) = [(current period carbon sequestration – base period carbon sequestration) / base period carbon sequestration] × 100%;

[0106] In the formula, the current period carbon sink is the total carbon sink of the year to be calculated, and the base period carbon sink is the total carbon sink of the previous year, reflecting the dynamic changes in the ecosystem's ability to absorb carbon emissions.

[0107] PSR Synergistic Features: Pressure Layer Score (S p ), State layer score ( ), Response layer score (S) r The entropy weight method combined with a coupling coordination degree model is used to calculate the weight of each score to ensure the synergy between the three factors and carbon sequestration and quotas.

[0108] Pressure level indicators include: carbon emission intensity, proportion of energy-intensive industries, and energy consumption per unit of GDP.

[0109] State-level indicators: vegetation coverage, air quality excellence rate, ecosystem health index, etc.

[0110] Response layer indicators include: the proportion of emission reduction investment, the proportion of renewable energy substitution, and the number of energy efficiency improvement projects.

[0111] Coupling Coordination Index (C): Quantifies the degree of synergy among carbon sequestration, quotas, and PSR, and is used to optimize the branch weights of the decision tree. Calculation formula: C = √[(U1×U2×U3) / ((U1+U2+U3) / 3)] 3 Wherein, U1 is the quota fit degree, U2 is the carbon sink adjustment degree, and U3 is the PSR synergy degree, with a value range of 0-1. C≥0.6 indicates good coupling.

[0112] 2. Construction of adaptive decision tree model.

[0113] The decision tree model, with "multi-dimensional collaborative judgment" as its core, achieves adaptive early warning through dynamic threshold adjustment. Figure 1 As shown, the structure includes a root node, internal nodes, and leaf nodes:

[0114] 2.1 Root node (initial judgment condition).

[0115] Does the carbon emission (E) exceed the carbon emission allowance (Q) of the coupled carbon sink carrying capacity? If E≤Q and the coupling coordination degree C≥0.6, a green warning is directly output; if E>Q or C<0.6, the internal node is judged.

[0116] 2.2 Internal nodes (intermediate judgment conditions).

[0117] The judgment is based on a progressive dynamic threshold, following the order of "carbon sink change rate → pressure layer score → state layer score → response layer score":

[0118] (1) Judgment of carbon sink change rate.

[0119] Set dynamic threshold R sth (Calculated using a sliding window based on nearly 5 years of time-series data, with a window length of 12 months). If R s <-R sth (If carbon sequestration capacity declines significantly), it enters the high-risk category; otherwise, it enters the pressure category for assessment.

[0120] (2) Pressure layer score judgment.

[0121] Dynamic threshold S pth (Adjusted based on regional industrial structure differences). If S p pth If carbon emission pressure is high, proceed to the state layer for judgment; otherwise, proceed to the response layer for judgment.

[0122] (3) State layer score judgment.

[0123] Dynamic threshold S stth (Based on adjustments to regional ecological carrying capacity). If S st stth If the environmental conditions are poor, proceed to the response layer for judgment; otherwise, output a yellow warning.

[0124] (4) Response layer score judgment.

[0125] Dynamic threshold S rth (Based on regional emission reduction baseline adjustments). If S r >S rth If the response measures are effective, an orange alert will be issued; otherwise, a red alert will be issued.

[0126] 2.3 Leaf nodes.

[0127] The system ultimately outputs four warning levels: green, yellow, orange, and red, along with corresponding control recommendations.

[0128] 3. Adaptive threshold adjustment mechanism. ​​

[0129] (1) Timing adaptive.

[0130] A 12-month sliding window is used, and the dynamic threshold is updated monthly to incorporate the latest emissions data and carbon sink trends.

[0131] (2) Region adaptation.

[0132] Differentiated benchmark thresholds are set based on regional ecological types (such as forest-covered areas and industrial areas) and economic development levels, and then dynamically corrected through coupling coordination degree.

[0133] (3) Feedback adaptation.

[0134] When the accuracy rate of the early warning (compared with the actual emission exceedance) is less than 90%, threshold optimization is automatically triggered, and the threshold weights of each node are adjusted.

[0135] 4. Technical Implementation Solution:

[0136] 4.1 Data Acquisition and Processing.

[0137] Data sources: regional statistical bulletins, remote sensing data (such as MOD13A3 vegetation index data), carbon emission accounting results, ecological and environmental monitoring data, government emission reduction annual reports, etc.

[0138] Data processing: Z-score standardization is used to process the data of each indicator to eliminate the influence of units; invalid data is removed by outlier detection (3σ principle); and a time series database is built to support sliding window calculation.

[0139] 4.2 Calculation of coupling characteristics.

[0140] Based on the above formulas, calculate the carbon emission allowance (Q) and carbon sink change rate (R) coupled with carbon sink carrying capacity. s ), PSR three-layer score (S p S st S r ) and coupling coordination degree (C);

[0141] The entropy weight method is used to determine the weights of each feature, ensuring the synergistic effect of multi-dimensional factors.

[0142] 4.3 Adaptive Decision Tree Training and Validation:

[0143] Training process: A training set is constructed using nearly 10 years of historical data. The optimization objectives are early warning accuracy and coupling coordination. The decision tree branch structure is optimized through cross-validation (5 folds).

[0144] Threshold calibration: Based on pilot data from different regions (such as industrial areas and ecological protection areas), calibrate the regionally differentiated parameters of the dynamic threshold;

[0145] Validation process: The confusion matrix was used to evaluate the model performance, with key metrics including accuracy (≥90%), recall (≥88%), and F1 score (≥89%).

[0146] 4.4 Early Warning Output:

[0147] Real-time output of early warning level, core risk factors (such as "carbon sink decline + pressure layer exceeding the standard") and coupling coordination degree;

[0148] By binding targeted control recommendations, a closed-loop output of "early warning level - risk factor - control measures" is formed.

[0149] In summary, this embodiment includes the following key technical solutions:

[0150] 1. Input feature system with deep coupling of carbon sink, quota, and PSR.

[0151] This includes methods for calculating carbon emission quotas that couple carbon sink carrying capacity, a synergistic integration mechanism for carbon sink change rate and PSR indicators, and a coupled and coordinated quantitative model, achieving full-chain coverage of risk factors from "emission baseline to ecological regulation to human activities to environmental status to emission reduction response".

[0152] 2. Dynamic threshold adjustment mechanism of adaptive decision tree.

[0153] It encompasses a time-series sliding window threshold update algorithm, a regionally differentiated threshold calibration method, and a threshold optimization logic driven by early warning feedback. Unlike traditional fixed threshold designs, it enables real-time adaptive adjustment of the threshold.

[0154] 3. Multi-dimensional collaborative decision tree branching rules.

[0155] With "whether carbon emissions exceed the coupled quota" as the root node, internal nodes are designed according to the progressive logic of "carbon sink change rate → pressure layer score → state layer score → response layer score", forming a multi-condition judgment rule based on dynamic thresholds to ensure the scientific nature and distinguishability of the early warning level.

[0156] 4. Early warning-control linkage mechanism.

[0157] Each warning level corresponds to the identification of core risk factors and targeted control recommendations, achieving a seamless connection from risk warning to decision-making and execution, and improving the practicality of technical solutions.

[0158] The beneficial effects of this embodiment include:

[0159] 1. Deep coupling of carbon sequestration, quotas, and PSR enables the coordinated capture of risk factors across the entire chain.

[0160] Breaking through the limitations of existing single-dimensional indicators, this method is the first to deeply integrate carbon sinks (ecological regulation), quotas (emission baselines), and PSR (human activity-environmental status-emission reduction response) to construct a full-chain input feature system of "emission-ecology-response". By coupling and coordinating the measurement of the interaction among the three, it avoids the one-sidedness of a single indicator, comprehensively covers the core driving factors of carbon emission risk, and solves the problem of incomplete risk factor coverage in existing methods.

[0161] 2. Adaptive decision tree technology breaks through the bottleneck of static threshold adaptability.

[0162] An innovative dynamic threshold adjustment mechanism with time-adaptive, region-adaptive, and feedback-adaptive design is proposed. Based on sliding window data, regional feature differences, and early warning feedback results, the judgment threshold is optimized in real time. This effectively adapts to the characteristics of large carbon emission fluctuations and strong ecological heterogeneity in small and medium-sized cities, solves the defects of misjudgment and missed judgment of traditional fixed thresholds, and improves the robustness of the early warning model.

[0163] 3. The accuracy of early warning has been significantly improved, supporting refined emission reduction decision-making.

[0164] Through multi-dimensional factor collaborative judgment and dynamic threshold adjustment, the model's early warning accuracy is ≥90%, which is 15%-20% higher than traditional methods. It can accurately locate core risk factors (such as "carbon sink decline + pressure exceeding the standard") and bind targeted control suggestions to different early warning levels, forming a closed loop of "risk identification - factor positioning - measure formulation". This solves the pain point of the existing disconnect between early warning and control, and provides scientific support for carbon emission reduction decisions in small and medium-sized cities.

[0165] 4. It is highly adaptable to small and medium-sized cities and has low implementation costs.

[0166] The indicator system and data sources are optimized by using publicly available statistical bulletins and remote sensing data to reduce the difficulty of data acquisition. The decision tree model has an intuitive structure, and dynamic threshold adjustment does not require complex computing power, making it easy for local governments to implement. At the same time, the regionally differentiated threshold design fully considers the industrial and ecological characteristics of small and medium-sized cities, solving the problem of poor adaptability of existing technologies to small and medium-sized cities.

[0167] 5. Systematic innovation in the technical framework, filling gaps in the field.

[0168] For the first time, an integrated early warning framework of "deep coupling of carbon sink-quota-PSR + adaptive decision tree" has been constructed. It integrates multiple technologies such as coupling coordination degree model, entropy weight method, and sliding window algorithm to achieve full-process optimization of "data collection-coupled feature calculation-dynamic judgment-tiered early warning-decision support". It fills the technical gap of refined carbon emission early warning at the scale of small and medium-sized cities and provides a quantitative tool for the scientific implementation of "dual carbon" goals at the grassroots level.

[0169] One specific implementation example is as follows:

[0170] Taking the carbon emissions of a certain city as the research object, a regional carbon emission early warning study was conducted from 2015 to 2021.

[0171] Step 1: Data preparation and coupling feature calculation.

[0172] 1. Basic data collection:

[0173] Carbon emissions (E) in 2021: 12.74 million tons;

[0174] Basic quota (Q0): Based on the average annual carbon emission quota of a certain district in the city and the city's GDP share in 2021, it is estimated to be 8,709,940 tons;

[0175] Carbon sink data: 1.56 million tons of carbon sinks in 2020 and 1.42 million tons of carbon sinks in 2021;

[0176] PSR indicator data includes 12 indicators such as industrial carbon emission intensity, vegetation coverage, and the proportion of government investment in emission reduction.

[0177] 2. Coupling characteristic calculation:

[0178] Carbon sink change rate (R s ): [(142-156) / 156]×100%=-8.97%;

[0179] Carbon sink carrying capacity correction factor (λ): The ecological type of a certain city is mainly grassland, and λ is taken as 0.2;

[0180] Carbon emission allowance (Q) coupled with carbon sink carrying capacity: 870.994 × (1 + 0.2 × (-8.97%)) ≈ 8,542,300 tons;

[0181] PSR three-level score (calculated using entropy weight method): S p =0.0184, S st =0.1030, S r =0.2246;

[0182] Coupling coordination degree (C): 0.42 (<0.6, weak coupling).

[0183] Step 2: Application of the adaptive decision tree model.

[0184] 1. Dynamic threshold determination (based on sliding window data from 2015-2020):

[0185] 2. Decision tree decision process:

[0186] Root node judgment: E=12.74 million tons > Q=8.5423 million tons, and C=0.42<0.6, enter the internal node;

[0187] Carbon sink change rate assessment: R s =-8.97% <-5%, entering a high-risk branch;

[0188] Pressure layer score assessment: S p =0.0184<0.2, proceed to state layer for judgment;

[0189] State layer score determination: =0.1030<0.2, proceed to the response layer for judgment;

[0190] Response layer score judgment: S r =0.2246>0.2, output an orange warning.

[0191] Step 3: Analysis of early warning results.

[0192] In 2021, the city's carbon emission warning level was orange, with the core risk factors being "excessive carbon emissions (49.1% over the limit) + significant decline in carbon sink capacity + exceeding pressure level limits + poor status level," and some response measures were effective. This result highly aligns with the city's actual situation: rapid industrial emissions growth led to quota overruns, and grassland ecological degradation resulted in a decline in carbon sink capacity. Although the city had implemented emission reduction measures, they had not yet fully offset the risks. Based on the warning results, it is recommended to activate the emergency response plan, focus on controlling high-energy-consuming industrial enterprises, and simultaneously increase investment in grassland ecological restoration to enhance carbon sink capacity.

[0193] Through this embodiment, the method successfully achieved dynamic classification and early warning of carbon emission risks in the city, verifying the feasibility and effectiveness of the method.

[0194] This application provides a carbon emission classification and early warning device with carbon sink-quota-PSR coupling decision tree, including:

[0195] The data calculation unit is used to calculate the carbon emission quota, carbon sink change rate, three-layer score of the PSR model, and coupling coordination degree of the coupled carbon sink carrying capacity; wherein, the three-layer score of the PSR model includes the pressure layer score, the state layer score, and the response layer score.

[0196] The tree input unit is used to input the carbon emission quota, the carbon sink change rate, the three-layer score of the PSR model, the coupling coordination degree, and the carbon emission amount into a pre-constructed decision tree; the decision tree includes a root node, multiple internal nodes, and multiple leaf nodes;

[0197] The first judgment unit is used to determine whether the carbon emissions exceed the carbon emission quota using the root node;

[0198] The second judgment unit is used to output a risk-free warning level using the first leaf node if the carbon emissions do not exceed the carbon emission quota; and to determine whether the coupling coordination degree exceeds the first dynamic threshold using the first internal node if the carbon emissions exceed the carbon emission quota.

[0199] The third judgment unit is used to output the risk-free warning level using the first leaf node if the coupling coordination degree does not exceed the first dynamic threshold; and to determine the carbon sink change rate exceeds the second dynamic threshold using the second internal node if the coupling coordination degree is greater than the first dynamic threshold.

[0200] The fourth judgment unit is used to output a low-risk warning level using the second leaf node if the carbon sink change rate does not exceed the second dynamic threshold; and to determine whether the pressure layer score exceeds the third dynamic threshold using the third internal node if the carbon sink change rate exceeds the second dynamic threshold.

[0201] The fifth judgment unit is used to output a medium risk warning level using the third leaf node if the pressure layer score does not exceed the third dynamic threshold; and to determine whether the state layer score exceeds the fourth dynamic threshold using the fourth internal node if the pressure layer score exceeds the third dynamic threshold.

[0202] The sixth judgment unit is used to output the medium-risk warning level using the third leaf node if the state layer score does not exceed the fourth dynamic threshold; and to determine whether the response layer score exceeds the fifth dynamic threshold using the fifth internal node if the state layer score exceeds the fourth dynamic threshold.

[0203] The seventh judgment unit is used to output the medium-risk warning level using the third leaf node if the response layer score does not exceed the fifth dynamic threshold; and to output the high-risk warning level using the fourth leaf node if the response layer score exceeds the fifth dynamic threshold.

[0204] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0205] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this application are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0206] Furthermore, although this application is described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding this application. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional technology for an engineer. Therefore, those skilled in the art can implement the application set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of this application, which is determined by the full scope of the appended claims and their equivalents.

[0207] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0208] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0209] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0210] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0211] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0212] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

[0213] The above is a detailed description of the preferred embodiments of this application, but this application is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A decision tree carbon emission grading early warning method coupled with carbon sink-quota-PSR, characterized in that, The method comprises the following steps: The method comprises the following steps: The carbon emission quota, the carbon sink change rate, the three-layer scores of the PSR model, and the coupling coordination degree are calculated; wherein the three-layer scores of the PSR model comprise a pressure layer score, a state layer score, and a response layer score; The carbon emission quota, the carbon sink change rate, the three-layer scores of the PSR model, the coupling coordination degree, and the carbon emission amount are input into a pre-constructed decision tree; the decision tree comprises a root node, multiple internal nodes, and multiple leaf nodes; It is judged by the root node whether the carbon emission amount exceeds the carbon emission quota; If the carbon emission amount does not exceed the carbon emission quota, a first leaf node is used to output a no-risk early warning level; if the carbon emission amount exceeds the carbon emission quota, a first internal node is used to judge whether the coupling coordination degree exceeds a first dynamic threshold value; If the coupling coordination degree does not exceed the first dynamic threshold value, the first leaf node is used to output the no-risk early warning level; if the coupling coordination degree is greater than the first dynamic threshold value, a second internal node is used to judge whether the carbon sink change rate exceeds a second dynamic threshold value; If the carbon sink change rate does not exceed the second dynamic threshold value, a second leaf node is used to output a low-risk early warning level; if the carbon sink change rate exceeds the second dynamic threshold value, a third internal node is used to judge whether the pressure layer score exceeds a third dynamic threshold value; If the pressure layer score does not exceed the third dynamic threshold value, a third leaf node is used to output a medium-risk early warning level; if the pressure layer score exceeds the third dynamic threshold value, a fourth internal node is used to judge whether the state layer score exceeds a fourth dynamic threshold value; If the state layer score does not exceed the fourth dynamic threshold value, the third leaf node is used to output the medium-risk early warning level; if the state layer score exceeds the fourth dynamic threshold value, a fifth internal node is used to judge whether the response layer score exceeds a fifth dynamic threshold value; 2. The carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method according to claim 1, characterized in that, If the response layer score does not exceed the fifth dynamic threshold value, the third leaf node is used to output the medium-risk early warning level; if the response layer score exceeds the fifth dynamic threshold value, a fourth leaf node is used to output a high-risk early warning level. Q = Q0x (1 + λ x R s ); Wherein Q is the carbon emission quota, Q0 is the basic quota based on regional economic level and historical emission data, λ is the carbon sink carrying capacity correction coefficient, R s is the carbon sink change rate.

3. The carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method according to claim 1, characterized in that, The carbon emission quota is calculated by the following formula: The carbon sink change rate is calculated by the following formula:

4. The carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method according to claim 1, characterized in that, Carbon sink change rate = [(current period carbon sink amount - base period carbon sink amount) / base period carbon sink amount] x 100%. The three-layer scores of the PSR model are calculated by the following steps: The entropy weight method is used to calculate the score weights of each layer in combination with the coupling coordination degree model; 5. The carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method according to claim 1, characterized in that, The score weights of each layer are weighted and summed with the corresponding indicators to calculate the pressure layer score, the state layer score, and the response layer score. C = V [(U1 x U2 x U3) / ((U1 + U2 + U3) / 3) 3 ]; The coupling coordination degree is calculated by the following formula:

6. The carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method according to claim 1, characterized in that, Wherein, C is the coupling coordination degree, U1 is the quota adaptation degree, U2 is the carbon sink adjustment degree, and U3 is the PSR coordination degree. The method further comprises the following steps: Each dynamic threshold value is updated once a month using a 12-month sliding window; Setting different reference dynamic thresholds according to regional ecological types and economic development levels; Each reference dynamic threshold is a reference value of each dynamic threshold.

7. The carbon sink-quota-PSR coupled decision tree carbon emission grading early warning method according to any one of claims 1 to 6, characterized in that, The method further comprises the following steps: When the accuracy of comparing each early warning level with the actual emission exceeding standard is lower than an accuracy threshold, triggering a threshold optimization step to adjust the node threshold weight.

8. The decision tree carbon emission grading early warning device coupled with carbon sink-quota-PSR, characterized in that, The device comprises: A data calculation unit configured to calculate a carbon emission quota coupled with a carbon sink bearing capacity, a carbon sink change rate, three-layer scores of a PSR model, and a coupling coordination degree; wherein the three-layer scores of the PSR model include a pressure layer score, a state layer score, and a response layer score; A tree input unit configured to input the carbon emission quota, the carbon sink change rate, the three-layer scores of the PSR model, the coupling coordination degree, and the carbon emission amount into a pre-constructed decision tree; the decision tree includes a root node, multiple internal nodes, and multiple leaf nodes; A first judgment unit configured to determine whether the carbon emission amount exceeds the carbon emission quota by using the root node; A second judgment unit configured to output a no-risk early warning level by using a first leaf node if the carbon emission amount does not exceed the carbon emission quota, and determine whether the coupling coordination degree exceeds a first dynamic threshold by using a first internal node if the carbon emission amount exceeds the carbon emission quota; A third judgment unit configured to output the no-risk early warning level by using the first leaf node if the coupling coordination degree does not exceed the first dynamic threshold, and determine whether the carbon sink change rate exceeds a second dynamic threshold by using a second internal node if the coupling coordination degree is greater than the first dynamic threshold; A fourth judgment unit configured to output a low-risk early warning level by using a second leaf node if the carbon sink change rate does not exceed the second dynamic threshold, and determine whether the pressure layer score exceeds a third dynamic threshold by using a third internal node if the carbon sink change rate exceeds the second dynamic threshold; A fifth judgment unit configured to output a medium-risk early warning level by using a third leaf node if the pressure layer score does not exceed the third dynamic threshold, and determine whether the state layer score exceeds a fourth dynamic threshold by using a fourth internal node if the pressure layer score exceeds the third dynamic threshold; A sixth judgment unit configured to output the medium-risk early warning level by using the third leaf node if the state layer score does not exceed the fourth dynamic threshold, and determine whether the response layer score exceeds a fifth dynamic threshold by using a fifth internal node if the state layer score exceeds the fourth dynamic threshold; A seventh judgment unit configured to output the medium-risk early warning level by using the third leaf node if the response layer score does not exceed the fifth dynamic threshold, and output a high-risk early warning level by using a fourth leaf node if the response layer score exceeds the fifth dynamic threshold.

9. An electronic device, comprising: The electronic device comprises a processor and a memory; The memory is configured to store a program; The processor executes the program to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a program, and the program is executed by the processor to implement the method in any one of claims 1 to 7.

Images