Landslide risk evaluation method based on landslide dynamic evolution law

Abstract

The invention discloses a landslide risk evaluation method based on a landslide dynamic evolution law, and the method comprises the following steps: carrying out the evolution law-driven risk evaluation data modeling, collecting landslide geological evolution characteristic data and multi-field associated data, and constructing a geological-social economic two-dimensional data pool; analyzing risk factor contribution degrees of different evolution stages by adopting an interpretability algorithm, and taking the contribution degrees as factor weights of subsequent risk quantification; performing risk quantification of evolution stage adaptation, taking a Gumbel-Hougaard Copula model as a basis, taking the landslide dynamic evolution stage as a model core parameter, embedding a multi-subject risk preference weight, and performing calculation to obtain a fusion risk level; on the basis of risk level fusion and risk loss quantification in the three fields of urban and rural development, agricultural economy and supply chain, single data fluctuation misjudgment is avoided through an evolution stage fault-tolerant mechanism, Copula model staged parameter adjustment is combined, and data modeling, a risk quantification whole process and an evolution state are deeply coupled.

Description

Technical Field

[0001] This invention belongs to the field of geological disaster prevention and control technology, and in particular relates to a landslide risk assessment method based on the dynamic evolution law of landslides. Background Technology

[0002] Currently, landslide risk assessment technology has been widely applied in the field of geological disaster prevention and control, forming a complete technical system covering data collection, risk quantification, decision output, early warning notification, and fund supervision. In the data collection stage, the industry commonly uses equipment such as Beidou positioning receivers, tipping bucket rain gauges, and water level sensors to achieve real-time monitoring of key parameters such as landslide displacement rate, rainfall intensity, and water level fluctuations, with data accuracy reaching centimeter-level displacement and 0.1mm-level rainfall. For risk quantification, algorithms such as the Copula model and EWMA control charts are often introduced, combined with historical landslide data to construct risk assessment models and classify low, medium, and high risk levels. In the decision-making and early warning stage, by integrating data from urban planning, agricultural planting, transportation and logistics, decision-making suggestions such as development restrictions, crop adjustments, and route optimization are output, and early warning information is pushed through government platforms, enterprise systems, and SMS notifications. In the area of ​​fund supervision and ecological compensation, a model of manual supervision to verify project progress and offline negotiation to allocate compensation funds is often adopted to ensure the use of prevention and control funds and cross-regional ecological collaborative governance. These existing technologies have played a crucial foundational role in reducing landslide disaster losses and ensuring regional safety.

[0003] However, existing landslide risk assessment technologies still have significant limitations, making it difficult to meet the needs of refined and dynamic prevention and control, which contrasts sharply with the beneficial effects of this invention: First, existing technologies mostly use static thresholds to determine the landslide evolution stage, failing to deeply couple with the evolution law, and are prone to misjudgment of stages due to single data anomalies, unlike this invention which can achieve dynamic adaptation through fault-tolerant mechanisms and stage-based parameter adjustments; Second, risk quantification models often use fixed parameters, failing to correlate the "triggering factor-displacement response" correlation strength, resulting in large quantification errors, making it difficult to achieve the precise calculation of losses in different areas as described in this invention; Third... The risk preference weights of multiple stakeholders are fixed in the long term and cannot be dynamically adjusted according to the evolutionary stage, making it difficult for decision-making schemes to balance public safety, corporate interests, and farmers' rights, and their adaptability is far lower than that of this invention; fourth, the early warning push has no priority design, which is prone to information congestion and delays in emergency response for key groups (such as residents), and its efficiency is not as good as the "personnel safety first" push logic of this invention; fifth, fund supervision relies on manual verification and compensation relies on offline negotiation, which has problems such as long disbursement cycle, high risk of misappropriation, and many distribution disputes, and cannot achieve the safe and fair effect of automatic supervision by smart contracts and transparent compensation by blockchain of this invention. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a landslide risk assessment method based on the dynamic evolution law of landslides, which solves the problems of disconnection between landslide risk assessment and evolution, low accuracy and insufficient coordination in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The landslide risk assessment method based on the dynamic evolution law of landslides includes the following steps:

[0007] S1: Evolutionary law-driven risk assessment data modeling, collecting landslide geological evolution characteristic data and multi-domain related data, constructing a geological-socio-economic dual-dimensional data pool; using interpretable algorithms to analyze the contribution of risk factors at different evolution stages, and using this contribution as the factor weight for subsequent risk quantification;

[0008] S2: Risk quantification adapted to the evolution stage is based on the Gumbel-Hougaard Copula model, which takes the dynamic evolution stage of landslides as the core parameter of the model and embeds the risk preference weights of multiple subjects to calculate the integrated risk level; based on the integrated risk level, the risk loss in the three major areas of urban and rural development, agricultural economy and supply chain is quantified.

[0009] S3: Risk assessment-oriented scenario-based decision output. Based on the integrated risk level and risk loss in different fields obtained from S2, the output includes urban and rural development decisions, agricultural adaptation decisions, and cross-regional ecological compensation decisions that correspond one-to-one with the evolution stage.

[0010] S4: Intelligent early warning push based on evolutionary features. An evolutionary stage-risk early warning linkage system is built based on the EWMA control chart. Dynamic early warning thresholds are set according to the evolutionary stage. Customized early warning information containing the current evolutionary stage, risk factor contribution, integrated risk level, and subject implementation suggestions is generated and pushed to the corresponding subject.

[0011] S5: Smart contract supervision of risk performance is based on the risk level of S2 and the decision-making scheme of S3 to construct a fund disbursement contract and an insurance claim contract; the contract dynamically triggers fund disbursement or claim according to the evolution stage, and the validity of performance is confirmed through preset verification standards.

[0012] The dynamic evolution stages of landslides are divided into four categories based on displacement rate: stable stage with a displacement rate of less than 3 mm / month, slow deformation stage with a displacement rate of 3-5 mm / month, accelerated deformation stage with a displacement rate of more than 5 mm / month, and warning stage with a single displacement of more than 10 mm. The integrated risk level is divided into the following categories based on the corresponding evolution stages: low risk corresponds to the stable stage, medium risk corresponds to the slow deformation stage, high risk corresponds to the accelerated deformation stage, and extremely high risk corresponds to the warning stage.

[0013] Preferably, in step S1: the landslide geological evolution characteristic data includes landslide displacement rate, daily rainfall intensity, and water level fluctuation; the landslide displacement rate is collected by a Beidou positioning device with a sampling frequency of 1 time / hour; the daily rainfall intensity is collected by a tipping bucket rain gauge with an accuracy of 0.1 mm; and the water level fluctuation is collected by a water level sensor with an accuracy of 0.01 m.

[0014] Preferably, in step S1: the multi-domain related data includes data from the fields of urban development, agriculture, ecology, and finance;

[0015] Data in the field of urban development includes the maximum plot ratio, benchmark land price, and land transfer plan. The maximum plot ratio is determined based on the urban planning document, the benchmark land price is determined based on the annual benchmark land price table issued by the local natural resources department, and the land transfer plan is determined based on the local annual land supply plan.

[0016] Agricultural data includes crop yield per mu, planting area, and crop disaster resistance coefficient. Crop yield per mu is determined based on the average of the regional agricultural statistical annual reports over the past three years, planting area is determined based on land ownership data, and crop disaster resistance coefficient is determined based on the crop stress resistance rating standards issued by the agricultural department.

[0017] The ecological data includes prevention and control input standards and regional economic disparity coefficients. The prevention and control input standards are determined based on the quota standards for geological disaster prevention and control projects, and the regional economic disparity coefficients are determined based on the ratio of regional per capita GDP, with a value range of 0.8-1.2.

[0018] Financial data includes insurance coverage and average daily output of enterprises. Insurance coverage is based on the terms of the insurance contract, while average daily output of enterprises is determined by calculating the average value of the enterprise's financial statements for the past 12 months.

[0019] Preferably, in step S1: the interpretability algorithm is the SHAP value analysis algorithm, and the output risk factor contribution is specifically as follows:

[0020] During the stable phase to the slow deformation phase, rainfall accounts for 60%, water level accounts for 30%, and development disturbance accounts for 10%.

[0021] From the accelerated deformation stage to the early warning stage, rainfall accounts for 70%, water level accounts for 25%, and cumulative deformation from the previous period accounts for 5%.

[0022] An error tolerance mechanism is added to the determination of the evolution stage. When the displacement data of a single hour exceeds the threshold of the current stage, but the average daily displacement of the previous 24 hours still meets the requirements of the current stage, continuous monitoring for 24 hours is required.

[0023] If the average displacement rate within 24 hours still meets the requirements of the current stage, then the original evolution stage determination is maintained.

[0024] If the average displacement rate exceeds the threshold of the current stage within 24 hours, then adjust to the corresponding evolution stage.

[0025] Preferably, in step S2: the landslide dynamic evolution stage is used as the core parameter of the model, specifically by adjusting the tail correlation coefficient of the Copula model according to the evolution stage level;

[0026] The evolutionary stages are divided into levels 1-4, with level 1 corresponding to the stable stage, level 2 to the slow deformation stage, level 3 to the accelerated deformation stage, and level 4 to the early warning stage. The tail correlation coefficient increases linearly with the level: level 1 corresponds to 0.5, level 2 to 0.65, level 3 to 0.75, and level 4 to 0.85. The increasing coefficients between levels are: level 1 to level 2 to 0.15, level 2 to level 3 to 0.1, and level 3 to level 4 to 0.1. The correlation strength is the Pearson correlation coefficient of the triggering factors in historical landslide data: rainfall, monthly water level change rate, and monthly landslide displacement change rate. The increase in the tail correlation coefficient is determined based on the statistical results of this correlation strength. A 10% increase in the correlation strength during the accelerated deformation stage corresponds to a 0.1 increase in the coefficient.

[0027] Preferably, in step S2: the multi-subject risk preference weights are determined using the entropy weight method, specifically: government public safety weight 0.35, government fiscal controllability weight 0.15, enterprise cost control weight 0.2, farmer income protection weight 0.18, and resident evacuation convenience weight 0.12. When the evolution stage enters the early warning stage, the government public safety weight is increased to 0.4, and the weights of other subjects are compressed according to the ratio of "original weight × 0.6 ÷ 0.95". After compression, the government fiscal controllability weight is 0.095, the enterprise cost control weight is 0.126, the farmer income protection weight is 0.114, and the resident evacuation convenience weight is 0.076. The multi-subject risk preference weights are updated annually, based on regional policy adjustment documents, enterprise annual risk demand reports, and farmer quarterly feedback questionnaires. The update process involves recalculating the weights based on the new criteria using the entropy weight method. The calculation results are publicized for 7 working days without objection and take effect through the local government website.

[0028] Preferably, in step S2, the specific formula for quantifying risk loss by sector is as follows:

[0029] (1) Risk loss of urban and rural development = benchmark land price × land area × evolution stage delay coefficient, the evolution stage delay coefficient is 0.8 for stable stage, 0.9 for slow deformation stage, 1.2 for accelerated deformation stage, and 1.5 for early warning stage;

[0030] (2) Agricultural economic risk loss = (average yield per mu of original crop - average yield per mu of suitable crop) × planting area + average yield per mu of original crop × yield loss rate in the evolution stage. The yield loss rate in the evolution stage increases linearly with the displacement rate. The displacement rate of 5 mm / month corresponds to 40%, the displacement rate of 10 mm / month corresponds to 60%, the single displacement of 10 mm corresponds to 70%, and the single displacement of 20 mm corresponds to 90%.

[0031] (3) Supply chain risk loss = average daily freight volume ÷ 24 × duration of interruption in evolution stage × unit freight cost × default coefficient. The unit freight cost is determined based on the annual guidance price of the transportation industry. The duration of interruption in the evolution stage increases linearly with the displacement rate. When the displacement rate is 3 mm / month, it corresponds to 12 hours. When the displacement rate is 5 mm / month, it corresponds to 24 hours. When the displacement rate is 10 mm / month, it corresponds to 72 hours. When the single displacement is ≥10 mm, it increases by 24 hours / 10 mm and the upper limit of the interruption duration is 168 hours. The default coefficient is 1.2 for high risk and 1.5 for extremely high risk. High risk corresponds to the accelerated deformation stage and extremely high risk corresponds to the early warning stage.

[0032] Preferably, in step S4: the dynamic early warning threshold of the evolution stage-risk early warning linkage system is set as follows: a single displacement greater than 8mm triggers an early warning in the stable stage; a single displacement greater than 6mm triggers an early warning in the slow deformation stage; a single displacement greater than 5mm triggers an early warning in the accelerated deformation stage; and a single displacement greater than 3mm triggers an early warning in the early warning stage. The early warning information push is prioritized, specifically in the following order: push notification of resident evacuation trigger conditions, push notification of emergency personnel dispatch suggestions, push notification of farmer subsidy application portal, and push notification of alternative enterprise logistics routes. The early warning information push time is no more than 15 minutes, and the push channels include government service platforms, enterprise management systems, farmer-specific apps, resident SMS, and community notices. Community notices are displayed on electronic screens and broadcast.

[0033] Preferably, in step S5: the triggering conditions and verification standards for the prevention and control of fund disbursement contracts are as follows:

[0034] (1) Medium risk corresponds to the slow deformation stage: triggering the allocation of 30% of the total prevention and control funds; the verification standard is that the installation rate of monitoring equipment is not less than 90%, verified by drone aerial photography, the drone flight altitude is not more than 100m, the heading overlap is not less than 80%, the lateral overlap is not less than 60%, the shooting resolution is not less than 0.1m, the shooting time is within 24 hours after the completion of the project, and at least 1 monitoring equipment point is set up for every 100m²;

[0035] (2) High risk corresponds to accelerated deformation stage: triggering the allocation of 60% of the total prevention and control funds; the verification standard is that the progress of temporary anti-slide engineering is not less than 80%, which is confirmed by comparing the engineering supervision report with drone aerial photography, and the completion rate of anti-slide pile pouring and the anchor bolt implantation rate are not less than 80%;

[0036] (3) The warning stage corresponding to extremely high risk: the remaining 40% of the prevention and control funds will be disbursed; the verification standard is that the landslide displacement rate drops to below 5 mm / month, and the average displacement rate is less than 5 mm / month after continuous monitoring for 72 hours; 5% of the total prevention and control funds will be frozen as a quality guarantee deposit. If the evolution stage is stable to the slow deformation stage and there is no re-slide after 1 year after the funds are disbursed, the quality guarantee deposit will be unlocked; the re-slide judgment standard is that the single displacement is greater than 10 mm or the monthly displacement rate is greater than 5 mm / month within 1 year.

[0037] Preferably, in step S3: the cross-regional ecological compensation decision-making uses blockchain technology to record the correlation between the evolutionary stages of upstream and downstream regions, with the protection correlation between the upstream early warning stage and the downstream stable stage calculated at 80%; the blockchain records include the start time of upstream prevention and control projects, the completion time of upstream prevention and control projects, details of upstream prevention and control investment, water quality monitoring data of downstream protected areas, and soil stability monitoring data of downstream protected areas; specific indicators of water quality monitoring data of downstream protected areas include pH value 6.0-9.0 and chemical oxygen demand ≤50mg / L, and specific indicators of soil stability monitoring data of downstream protected areas include soil moisture content 15%-25% and internal friction angle ≥25°. Both types of monitoring data are collected once a month; the blockchain adopts a consortium blockchain architecture, with nodes including upstream and downstream local governments, ecological and environmental departments, and financial departments; data modifications must be confirmed by no less than 2 / 3 of the nodes before taking effect.

[0038] The technical effects and advantages of the landslide risk assessment method based on the dynamic evolution law of landslides in this invention are as follows:

[0039] 1. This invention takes the dynamic evolution stage of landslides as the core anchor point, avoids misjudgment of single data fluctuations through the "evolution stage fault tolerance mechanism" (continuous monitoring mean judgment stage), and combines "Copula model stage parameter adjustment" (dynamic parameter setting based on historical triggering factors - displacement response intensity) to deeply couple the entire process of data modeling and risk quantification with the evolution state, replacing the traditional "static threshold, fixed parameter" evaluation mode, significantly reducing stage judgment error and risk quantification deviation, so that the evaluation results always fit the actual evolution trend of landslides, and provide accurate stage basis for risk management.

[0040] 2. On the one hand, this invention establishes quantitative formulas for risk losses in urban and rural development, agricultural economy, and supply chain sectors, accurately calculating the degree of risk impact on different sectors, replacing the traditional general calculation, and providing reliable data support for decisions such as development restrictions, agricultural subsidies, and logistics adjustments; on the other hand, the risk preference weights of multiple stakeholders are dynamically adjusted according to the evolution stage, breaking the limitation of the traditional fixed weights that "take one thing into account but lose another", so that the decision-making scheme not only fits the core contradictions at different stages, but also takes into account the reasonable demands of the government, enterprises, farmers, and residents, thereby improving the acceptance and implementation of decisions.

[0041] 3. This invention designs an early warning push mechanism based on the priority of personnel safety, followed by emergency dispatch, and then rights protection, replacing the traditional "indiscriminate push" mode. This ensures that key early warning information such as evacuation instructions for residents is delivered first, shortening the emergency response time for high-risk groups. At the same time, it avoids response delays caused by congestion of non-priority information. While improving the efficiency of early warning transmission, it strengthens the protection of personnel life safety and effectively reduces the risk of casualties in landslide emergency scenarios.

[0042] 4. This invention enables the automatic allocation of prevention and control funds according to the evolution stage and verification standards through smart contracts, reducing manual intervention, shortening the allocation cycle, and reducing the risk of fund misappropriation. At the level of interest coordination, it uses blockchain to record the correlation of cross-regional landslide evolution, details of prevention and control investment, and monitoring data of protected areas. Through joint confirmation by multiple nodes, it achieves transparent allocation of compensation funds, breaking the limitations of traditional cross-regional compensation that is "mainly based on negotiation and lacks data transparency," reducing interest disputes, and taking into account both resource utilization efficiency and regional fairness.

[0043] 5. This invention deeply integrates the various stages of precise evaluation, scientific decision-making, efficient early warning, and standardized supervision with the evolutionary phases, forming a closed-loop management system from data collection to interest coordination, replacing the traditional risk governance model of "disconnected links and fragmented advancement"; through the coordinated optimization of each link, it not only solves the pain points of single links (such as stage misjudgment and inefficient funding), but also realizes the transformation of landslide risk governance from "passive response" to "proactive prevention and control", significantly improving the systematicness and effectiveness of overall governance. Attached Figure Description

[0044] Figure 1 This is a flowchart of the landslide risk assessment method based on the dynamic evolution law of landslides proposed in this invention. Detailed Implementation

[0045] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0046] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, 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 a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0047] Example 1

[0048] refer to Figure 1 This embodiment provides a landslide risk assessment method based on the dynamic evolution law of landslides, which is used for determining and warning of landslide evolution stages in mountainous areas. Specific implementation details include:

[0049] Implementation scenario:

[0050] A landslide in a mountainous area in southern China (28°15′N, 118°30′E) has historically been stable (monthly displacement of 2.2 mm). Displacement fluctuations occurred during the rainy season in June 2024. A risk assessment needs to be completed using a five-step method to avoid misjudgment at each stage.

[0051] Implementation steps:

[0052] S1, Evolutionary Law-Driven Risk Assessment Data Modeling:

[0053] Data Acquisition: Equipment configuration: BDS-3 Beidou positioning receiver (sampling frequency 1 time / hour, accuracy ±2cm) to collect displacement, SL3-1 tipping bucket rain gauge (accuracy 0.1mm) to collect daily rainfall, and submersible water level sensor (accuracy 0.01m) to collect water level; Monitoring data on June 10: single-hour displacement 4.2mm (abnormal fluctuation), average daily displacement of 0.035mm in the 24 hours prior to monitoring (monthly rate 1.14mm), daily rainfall 35mm, water level fluctuation 0.8m.

[0054] Fault tolerance mechanism execution: Due to the single-hour displacement exceeding the stable stage threshold (3mm / month → hourly average 0.04mm), continuous monitoring for 24 hours (14:00 on June 10th to 14:00 on June 11th) was conducted. During this period, the 24-hour average displacement was 0.038mm (monthly rate 1.14mm), and the "stable stage" judgment was maintained.

[0055] Risk factor contribution: Based on SHAP value analysis, the contribution of the stable phase is output as follows: rainfall 60%, water level 30%, and development disturbance 10%, which are used as factor weights for subsequent risk quantification.

[0056] S2, Risk Quantification of Evolutionary Stage Adaptation:

[0057] Since it was determined to be in a stable phase (low risk), the low risk level quantification rule was applied according to the correlation of the integration risk level. Based on the Gumbel-Hougaard Copula model (tail correlation coefficient 0.5), the integration risk level was calculated to be "low risk". The loss quantification result was the basic monitoring value (the loss of urban and rural development / agriculture / supply chain was 0 yuan). The risk factor weights were recorded for subsequent monitoring.

[0058] S3, Risk Assessment-Oriented Scenario-Based Decision Output:

[0059] Stable phase (low risk): Based on the decision-making logic, the urban and rural development decision is to "maintain the current development intensity (plot ratio 2.0) and retain 5% of the prevention and control reserve fund", without adjusting the planting or compensation plan.

[0060] S4, intelligent early warning push linked to evolutionary features:

[0061] Warning threshold: The warning threshold for the stable phase is a single displacement > 8mm. The current single displacement is 4.2mm, which does not trigger a warning. Only a "Stable Phase Risk Monitoring Report" is generated and pushed to the government (natural resources department) and surrounding enterprises to remind them to pay close attention to changes in rainfall.

[0062] S5, Smart Contract Regulation for Risk-Based Performance:

[0063] During the low-risk phase, there is no need to trigger fund disbursement or claims contracts. Instead, the monitoring data and decision-making results of S1-S4 are recorded through contracts to form a traceability file.

[0064] Implementation results:

[0065] The accuracy rate of evolution stage determination is 100%, avoiding the misjudgment of "stable → slow deformation" caused by traditional fault-tolerant mechanisms (which can reduce the loss of 2 million yuan in business downtime). The contribution of SHAP value is 95% consistent with the actual monitoring factors.

[0066] Example 2

[0067] This embodiment provides a landslide risk assessment method based on the dynamic evolution law of landslides, which is used for landslide risk quantification and decision-making around reservoirs. The specific implementation content includes:

[0068] Implementation scenario:

[0069] A landslide occurred around a reservoir (water level fluctuation ±3m / month). In August 2024, it entered a slow deformation stage (monthly displacement 4mm). The landslide affected farmland planting (50 mu of camellia trees) and reservoir ecological protection. The risk needs to be quantified and a decision made through a five-step method.

[0070] Implementation steps:

[0071] S1, Evolutionary Law-Driven Risk Assessment Data Modeling:

[0072] Data collection across multiple fields: urban development data (no direct development, using surrounding plots with a floor area ratio of 1.6), agricultural data (average yield of 3,000 yuan per mu for camellia trees, planting area of ​​50 mu, disaster resistance coefficient of 0.8), ecological data (prevention and control investment standard of 800,000 yuan / km², regional economic difference coefficient of 1.0), and geological data (monthly displacement of 4 mm, average daily rainfall of 45 mm, water level fluctuation of 2.5 m).

[0073] Risk factor contribution: SHAP value analysis outputs the contribution of the slow deformation stage - rainfall 65%, water level 25%, reservoir water level fluctuation 10%; the correlation strength (Pearson correlation coefficient between monthly rainfall variation rate and monthly displacement variation rate) is 0.45, which is used as the basis for adjusting the Copula parameter.

[0074] S2, Risk Quantification of Evolutionary Stage Adaptation:

[0075] Copula parameter settings: The slow deformation stage is at evolution level 2, with a tail correlation coefficient of 0.65 (0.5 for level 1, increasing to 0.15). The parameters are determined based on historical data from the past 10 years (the correlation strength between rainfall and displacement increases by 5% during the slow deformation stage).

[0076] Multi-entity weighting: weighted according to the slow transformation stage - government public safety 0.35, government financial controllability 0.15, enterprise (reservoir operation) cost control 0.2, farmers' income protection 0.18, and residents' evacuation convenience 0.12.

[0077] Loss quantification: Agricultural economic risk loss = (3000-3000)×50+3000×50×0 (yield loss rate during slow deformation stage is 0) = 0 yuan; potential ecological loss = 80×0.5 (landslide affected area 0.5km²)×0.5 (medium risk coefficient) = 200,000 yuan.

[0078] Fusion risk level: calculated as "medium risk", which corresponds perfectly to the slow deformation stage.

[0079] S3, Risk Assessment-Oriented Scenario-Based Decision Output:

[0080] Agricultural adaptation decision: For medium-risk planting of camellia trees with a disaster resistance coefficient of 0.8, the subsidy is 0 × 20% = 0 yuan, and there is no need to change the planting.

[0081] Ecological compensation decision: The downstream reservoir benefits 60%, and the compensation amount is 20 × 0.5 (medium risk benefit coefficient) × 1.0 = 100,000 yuan. The downstream bears 60% (60,000 yuan).

[0082] S4-S5: Warnings and Contracts

[0083] S4, the warning threshold for the slow deformation stage is a single displacement > 6mm. Currently, there is no threshold exceeding the current value, so a "medium risk monitoring notification" is pushed.

[0084] S5 triggered the ecological compensation preparatory contract, freezing 60,000 yuan of compensation funds for downstream users, which will be disbursed after quarterly monitoring.

[0085] Implementation results:

[0086] The model achieves 100% matching between risk level and evolution stage, with a quantification error of only 5% for the Copula model (compared to 35% for traditional fixed-parameter models), and a 90% implementation rate for decision-making solutions.

[0087] Example 3

[0088] This embodiment provides a landslide risk assessment method based on the dynamic evolution law of landslides, which is used for landslide early warning and subsidy decision-making in urban-rural fringe areas. The specific implementation content includes:

[0089] Implementation scenario:

[0090] A landslide occurred in a suburban area (100 mu of residential development land and 50 mu of farmland). In September 2024, due to continuous rainfall, the landslide entered the warning stage (single displacement of 12 mm), and the five-step response method needed to be activated urgently.

[0091] Implementation steps:

[0092] S1, Data Modeling:

[0093] Data: Single displacement 12mm, daily rainfall 85mm, water level fluctuation 1.8m;

[0094] Data from multiple sectors: benchmark land price for residential land is 5 million yuan per mu, average yield per mu for camellia oleifera trees is 3,000 yuan, and average yield per mu for pasture is 800 yuan.

[0095] Stage determination: Warning stage (extremely high risk), SHAP contribution: rainfall 70%, water level 25%, development disturbance 5%.

[0096] S2, Risk Quantification:

[0097] Multi-entity weighting: During the early warning stage, government public safety 0.4, government financial controllability 0.095, enterprise cost control 0.126, farmers' income protection 0.114, and residents' evacuation convenience 0.076.

[0098] Loss calculation:

[0099] Urban and rural development losses = 500 × 100 × 1.5 (early warning delay coefficient) = 750 million yuan;

[0100] Agricultural loss = (3000-800)×50 + 3000×50×80% = 230000 yuan.

[0101] S3, Decision Output:

[0102] Urban and rural areas: Development is suspended, and 20 mu of land is requisitioned for material stockpiling; Agriculture: Pasture planting is changed, and the subsidy is 230,000 × 120% = 276,000 yuan.

[0103] S4: Alert Push Notifications

[0104] Threshold: During the warning phase, a single displacement > 3mm (triggered), notifications will be pushed out according to priority within 15 minutes.

[0105] Residents: Evacuation routes (3 routes, the longest being 2km), resettlement sites (capacity of 500 people);

[0106] Government: Additional funding of 3 million yuan (total investment 1000 × (0.9 - 0.6)) and deployment of 100 people (5000 × 0.02);

[0107] For businesses: Alternative routes (20km detour), claims channel (LC20240915).

[0108] Implementation results:

[0109] The early warning was pushed out within 12 minutes, the resident evacuation completion rate was 100%, the subsidy was received within 7 days, and the implementation rate was 95%.

[0110] Example 4

[0111] This embodiment provides a landslide risk assessment method based on the dynamic evolution law of landslides, which is used for supply chain losses and contract supervision in landslides along major transportation routes. Specific implementation details include:

[0112] Implementation scenario:

[0113] A landslide occurred beside a national highway (average daily freight volume of 5,000 tons, unit freight cost of 0.1 yuan / ton·km). In October 2024, it entered an accelerated deformation stage (monthly displacement of 6mm). It is necessary to quantify the supply chain losses and initiate fund disbursement.

[0114] Implementation steps:

[0115] S1, Data Modeling:

[0116] Data: Monthly displacement 6mm, interruption duration 48h, freight volume 5000 tons, freight cost 0.1 yuan / ton·km.

[0117] S2, Risk Quantification:

[0118] Supply chain loss = 5000 ÷ 24 × 48 × 0.1 × 1.2 (default coefficient) = 28800 yuan, integrated risk level "high risk".

[0119] S3, Decision Output:

[0120] Logistics companies: Activate alternative routes; Government: Initiate temporary anti-skid works (budget 8 million yuan).

[0121] S4, Alert Push Notification:

[0122] When the acceleration phase threshold is greater than 5mm for a single displacement (already triggered), a loss report and engineering plan will be pushed out.

[0123] S5, Contract Supervision:

[0124] Fund disbursement: 60% of the total funds (4.8 million yuan) will be disbursed within 24 hours if the project progress is verified by drone aerial photography (80m altitude, 0.08m resolution) to be 85% (≥80%).

[0125] Insurance claim: Triggering corporate claim = 28800 × 90% = 25920 yuan, which will be credited to the account within 12 hours.

[0126] Implementation results:

[0127] Funds were not misappropriated, and logistics losses were reduced by 80%.

[0128] Example 5

[0129] This embodiment provides a landslide risk assessment method based on the dynamic evolution law of landslides, which is used for landslide monitoring funding contracts in industrial parks. Specific implementation details include:

[0130] Implementation scenario:

[0131] A landslide near an industrial park (total prevention and control funds of 20 million yuan) entered a slow deformation stage in November 2024, requiring the installation of monitoring equipment and the disbursement of the first installment of funds.

[0132] Implementation steps:

[0133] S1, Data Modeling:

[0134] Equipment installation: Install 20 BDS-3 receivers (sampling once per hour) to cover the entire landslide area.

[0135] S5, Contract Supervision:

[0136] Triggering conditions: 30% (6 million yuan) disbursement during the slow deformation stage (medium risk), drone verification and installation rate of 95% (≥90%), automatic fund disbursement, and blockchain recording of verification reports.

[0137] Implementation results:

[0138] Equipment installation rate reached 100%, and fund disbursement efficiency improved by 60%.

[0139] Example 6

[0140] This embodiment provides a landslide risk assessment method based on the dynamic evolution law of landslides, which is used for cross-basin landslide ecological compensation blockchain. The specific implementation content includes:

[0141] Implementation scenario:

[0142] A cross-basin landslide (upstream mountainous area, downstream irrigation area) entered the early warning stage in December 2024, requiring compensation to be calculated and distributed through blockchain.

[0143] Implementation steps:

[0144] S1, Data Modeling:

[0145] Data: Upstream prevention and control investment of 10 million yuan, benefit coefficient of 0.9 in the early warning stage, coefficient of 1.2 in the western region, and downstream benefit area of ​​100 km².

[0146] S2, Risk Quantification:

[0147] The compensation amount is 1000 × 0.9 × 1.2 = 10.8 million yuan, with downstream beneficiaries accounting for 80% and bearing 6.48 million yuan.

[0148] S3, Decision Output:

[0149] The downstream allocated 6.48 million yuan to the upstream for the anti-slide pile project.

[0150] S5, Contract Supervision:

[0151] The blockchain records the upstream project details (5 million yuan for anti-slide piles) and the downstream water quality (pH 7.2, COD 35mg / L). Payment is made after confirmation by 3 / 4 nodes (upstream and downstream local governments, ecological and environmental departments, and finance departments). The data is tamper-proof.

[0152] Implementation results:

[0153] The compensation dispute rate was 0%, and the downstream water quality improved by 15%.

[0154] Comparative Example 1

[0155] This comparison provides a stage-based determination mechanism for traditional fault-tolerant mechanisms, specifically including:

[0156] Comparison scenarios:

[0157] In the same mountain landslide as in Example 1, the traditional method does not implement the fault tolerance mechanism in step S1.

[0158] Comparison steps:

[0159] Without continuous monitoring for 24 hours, a displacement of 4.2 mm per hour was identified as the "slow deformation stage," and the decision to "reduce the volume ratio to 1.5" in step S3 was executed, resulting in the suspension of the development project.

[0160] Comparison results:

[0161] The company suffered a loss of 2 million yuan, with a 100% misjudgment rate; however, the present invention implemented the S1 fault tolerance mechanism, resulting in no loss.

[0162] Comparative Example 2

[0163] The comparative model provides traditional fixed-parameter Copula quantization, including:

[0164] Comparison scenarios:

[0165] Similar to the reservoir landslide in Example 2, the traditional method does not perform the Copula parameter adjustment in step S2.

[0166] Comparison steps:

[0167] With a fixed Copula coefficient of 0.7 and without considering the Pearson correlation coefficient between the triggering factors and the displacement response, the risk level was calculated to be "medium to high," requiring an additional 3 million yuan in prevention and control funds.

[0168] Comparison results:

[0169] The quantization error is 35%; this invention performs S2 parameter adjustment, with an error of 5%, saving 3 million yuan.

[0170] Compared with Examples 1-6 and Comparative Examples 1-2, Examples 1-6 of this invention focus on the core technical feature of "landslide risk assessment method based on the dynamic evolution law of landslides," selecting typical scenarios such as mountainous areas, reservoir peripheries, and urban-rural fringe areas, and verifying the feasibility of the scheme through the entire process of "data modeling - risk quantification - decision output - early warning push - contract supervision." Comparative Examples 1-2, on the other hand, adopt traditional landslide risk assessment methods, focusing on the two core pain points of "evolution stage determination" and "risk quantification," highlighting the technical advantages of this invention. The specific comparison is as follows:

[0171] From the perspective of evolution stage determination, Example 1 addresses the issue of abnormal hourly displacement fluctuations in landslides in mountainous areas of southern China. Using a fault-tolerant mechanism, it continuously monitors for 24 hours and determines the stage based on the average daily displacement of the preceding 24 hours, ultimately maintaining a "stable stage" determination and preventing unintended project shutdowns. Comparative Example 1, employing the traditional "single-hour threshold triggering" mode, directly misjudged abnormal fluctuations as a "slow deformation stage," resulting in a 2 million yuan loss for the company due to work stoppage. The comparison demonstrates that the "24-hour average verification" mechanism of this invention can reduce the error rate of evolution stage determination to 0, while the traditional method has an error rate exceeding 30%.

[0172] Regarding the accuracy of risk quantification, Example 2, targeting landslides around reservoirs, used the Copula model's phased parameter adjustment rules to set a tail correlation coefficient of 0.65 for the slow deformation stage (evolution level 2). Combined with the Pearson correlation coefficient between triggering factors and displacement response, a "medium risk" level was calculated, perfectly matching the actual evolution state. Comparative Example 2, using a traditional fixed coefficient (0.7), was mistakenly calculated as "medium-high risk," resulting in an overpayment of 3 million yuan in prevention and control funds. Data shows that the risk quantification error of this invention is only 5%, while the traditional method has an error of 35%, confirming the crucial role of "phased parameter adjustment" in improving quantification accuracy.

[0173] Regarding the adaptability of decision output, Example 3, targeting landslides in the early warning stage of urban-rural fringe areas, dynamically adjusted the government's public safety weight to 0.4 according to the multi-stakeholder weighting rule, outputting a decision of "suspending development + full agricultural subsidies," achieving a 95% implementation rate and over 90% satisfaction from both farmers and the government. Traditional methods, using fixed weights (government 0.3, enterprises 0.3, farmers 0.4), biased decisions towards farmer benefits, leading to delayed public safety control and failure to address three potential hazards in a timely manner. Therefore, the "dynamic weighting with stage" mechanism of this invention can achieve a balance of multi-stakeholder needs, while traditional fixed-weight decisions are prone to domain imbalances.

[0174] In terms of early warning delivery efficiency, Example 4, targeting landslides near major transportation routes, uses a priority-based delivery rule (residents → government → businesses), completing the delivery to all affected entities within 12 minutes. This results in a 100% evacuation rate for residents and an 80% reduction in logistics losses for businesses through alternative routes. In contrast, the traditional method, with its lack of priority and indiscriminate delivery, leads to information congestion, causing a 1.5-hour delay in resident evacuation and over 50,000 yuan in logistics losses for businesses. The comparison shows that this invention improves early warning response time by 80% and reduces losses by over 70%, while the traditional method suffers from delayed response and inefficient loss management.

[0175] Regarding the security of fund supervision, Example 5 addresses landslides in industrial parks. Using a smart contract mechanism, after verifying the installation rate of monitoring equipment (95%) via drone aerial photography, 30% of the prevention and control funds are automatically disbursed within 24 hours, without human intervention and with full traceability. Traditional methods employ a "manual supervision + offline disbursement" model, with verification taking 3 days and the disbursement process exceeding 7 days, while also carrying a 20% risk of fund misappropriation. This invention improves fund disbursement timeliness by 90% and reduces the risk of misappropriation to zero. Traditional methods are inefficient and lack security.

[0176] Regarding the fairness of cross-regional compensation, Example 6 addresses cross-basin landslides by using a blockchain mechanism to record monitoring data such as water quality (pH 7.2, COD 35 mg / L) and soil moisture content (20%). Compensation allocation is completed after confirmation by 3 / 4 of the nodes (upstream and downstream local governments, environmental protection departments, and finance departments), resulting in a 0% dispute rate. Traditional methods rely on upstream and downstream negotiations, leading to disputes over compensation amounts exceeding two months due to data opacity, and downstream ecological improvement rates are only 8%. This invention achieves a 0% compensation dispute rate and a 15% ecological improvement rate, while traditional methods have a dispute rate exceeding 60%, fully supporting transparent and fair design.

[0177] In summary, through full-process implementation verification, Examples 1-6 demonstrate that the present invention is significantly superior to traditional methods (Comparative Examples 1-2) in terms of accuracy in determining the evolution stage, precision in risk quantification, adaptability of decision-making, efficiency in early warning, security of funds, and fairness of compensation.

[0178] The above embodiments can be implemented in whole or in part by software, hardware, firmware or other arbitrary combinations. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0179] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0180] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0181] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.

[0182] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A landslide risk assessment method based on the dynamic evolution law of landslides, characterized in that, Includes the following steps: S1: Evolutionary law-driven risk assessment data modeling, collecting landslide geological evolution characteristic data and multi-domain related data, and constructing a geological-socio-economic dual-dimensional data pool; An interpretable algorithm is used to analyze the contribution of risk factors at different evolution stages, and this contribution is used as the factor weight for subsequent risk quantification. S2: Risk quantification adapted to the evolution stage is based on the Gumbel-Hougaard Copula model, which takes the dynamic evolution stage of landslides as the core parameter of the model, embeds the risk preference weights of multiple subjects, and calculates the fused risk level. Based on the risk level of integration, the risk losses in the three major areas of urban and rural development, agricultural economy and supply chain are quantified. S3: Risk assessment-oriented scenario-based decision output. Based on the integrated risk level and risk loss in different fields obtained from S2, the output includes urban and rural development decisions, agricultural adaptation decisions, and cross-regional ecological compensation decisions that correspond one-to-one with the evolution stage. S4: Intelligent early warning push based on evolutionary features. An evolutionary stage-risk early warning linkage system is built based on the EWMA control chart, and dynamic early warning thresholds are set according to the evolutionary stage. Generate customized early warning information that includes the current evolution stage, risk factor contribution, integrated risk level, and subject implementation recommendations, and push it to the corresponding subject; S5: Smart contract supervision of risk performance is based on the risk level of S2 and the decision-making scheme of S3 to construct a fund disbursement contract and an insurance claim contract; the contract dynamically triggers fund disbursement or claim according to the evolution stage, and the validity of performance is confirmed through preset verification standards. The dynamic evolution stages of landslides are divided into four categories according to displacement rate: the stable stage has a displacement rate of less than 3 mm / month, the slow deformation stage has a displacement rate of 3-5 mm / month, the accelerated deformation stage has a displacement rate of more than 5 mm / month, and the early warning stage has a single displacement of more than 10 mm. The risk levels of fusion are divided according to the evolution stage as follows: low risk corresponds to the stable stage, medium risk corresponds to the slow deformation stage, high risk corresponds to the accelerated deformation stage, and extremely high risk corresponds to the early warning stage.

2. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S1: the geological evolution characteristic data of the landslide includes the landslide displacement rate, daily rainfall intensity, and water level fluctuation. The landslide displacement rate is collected by Beidou positioning equipment at a sampling frequency of 1 time / hour. The daily rainfall intensity is collected by a tipping bucket rain gauge with an accuracy of 0.1 mm. The water level fluctuation is collected by a water level sensor with an accuracy of 0.01 m.

3. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S1: the multi-domain related data includes data from the fields of urban development, agriculture, ecology, and finance; Data in the field of urban development includes the maximum plot ratio, benchmark land price, and land transfer plan. The maximum plot ratio is determined based on the urban planning document, the benchmark land price is determined based on the annual benchmark land price table issued by the local natural resources department, and the land transfer plan is determined based on the local annual land supply plan. Agricultural data includes crop yield per mu, planting area, and crop disaster resistance coefficient. Crop yield per mu is determined based on the average of the regional agricultural statistical annual reports over the past three years, planting area is determined based on land ownership data, and crop disaster resistance coefficient is determined based on the crop stress resistance rating standards issued by the agricultural department. The ecological data includes prevention and control input standards and regional economic disparity coefficients. The prevention and control input standards are determined based on the quota standards for geological disaster prevention and control projects, and the regional economic disparity coefficients are determined based on the ratio of regional per capita GDP, with a value range of 0.8-1.

2. Financial data includes insurance coverage and average daily output of enterprises. Insurance coverage is based on the terms of the insurance contract, while average daily output of enterprises is determined by calculating the average value of the enterprise's financial statements for the past 12 months.

4. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S1: the interpretability algorithm is the SHAP value analysis algorithm, and the output risk factor contribution is specifically as follows: During the stable phase to the slow deformation phase, rainfall accounts for 60%, water level accounts for 30%, and development disturbance accounts for 10%. From the accelerated deformation stage to the early warning stage, rainfall accounts for 70%, water level accounts for 25%, and cumulative deformation from the previous period accounts for 5%. An error tolerance mechanism is added to the determination of the evolution stage. When the displacement data of a single hour exceeds the threshold of the current stage, but the average daily displacement of the previous 24 hours still meets the requirements of the current stage, continuous monitoring for 24 hours is required. If the average displacement rate within 24 hours still meets the requirements of the current stage, then the original evolution stage determination is maintained. If the average displacement rate exceeds the threshold of the current stage within 24 hours, then adjust to the corresponding evolution stage.

5. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S2: The dynamic evolution stage of the landslide is used as the core parameter of the model, specifically by adjusting the tail correlation coefficient of the Copula model according to the evolution stage level; The evolutionary stages are divided into levels 1-4, with level 1 corresponding to the stable stage, level 2 to the slow deformation stage, level 3 to the accelerated deformation stage, and level 4 to the early warning stage. The tail correlation coefficient increases linearly with the level: level 1 corresponds to 0.5, level 2 to 0.65, level 3 to 0.75, and level 4 to 0.

85. The increasing coefficients between levels are: level 1 to level 2 to 0.15, level 2 to level 3 to 0.1, and level 3 to level 4 to 0.

1. The correlation strength is the Pearson correlation coefficient of the triggering factors in historical landslide data: rainfall, monthly water level change rate, and monthly landslide displacement change rate. The increase in the tail correlation coefficient is determined based on the statistical results of this correlation strength. A 10% increase in the correlation strength during the accelerated deformation stage corresponds to a 0.1 increase in the coefficient.

6. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S2: the risk preference weights of multiple stakeholders are determined using the entropy weight method, specifically: government public safety weight 0.35, government fiscal controllability weight 0.15, enterprise cost control weight 0.2, farmer income protection weight 0.18, and resident evacuation convenience weight 0.

12. When the evolutionary stage enters the early warning stage, the government public safety weight is increased to 0.4, and the weights of other stakeholders are compressed according to the ratio of "original weight × 0.6 ÷ 0.95". After compression, the government fiscal controllability weight is 0.095, the enterprise cost control weight is 0.126, the farmer income protection weight is 0.114, and the resident evacuation convenience weight is 0.

076. The risk preference weights of multiple stakeholders are updated annually, based on regional policy adjustment documents, enterprise annual risk demand reports, and farmer quarterly feedback questionnaires. The update process involves recalculating the weights using the entropy weight method based on the new criteria. The results will be publicized for 7 working days without objection and will take effect on the local government's official website.

7. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S2: the specific formula for quantifying risk loss by sector is as follows: (1) Risk loss of urban and rural development = benchmark land price × land area × evolution stage delay coefficient, the evolution stage delay coefficient is 0.8 for stable stage, 0.9 for slow deformation stage, 1.2 for accelerated deformation stage, and 1.5 for early warning stage; (2) Agricultural economic risk loss = (average yield per mu of original crop - average yield per mu of suitable crop) × planting area + average yield per mu of original crop × yield loss rate in the evolution stage. The yield loss rate in the evolution stage increases linearly with the displacement rate. The displacement rate of 5 mm / month corresponds to 40%, the displacement rate of 10 mm / month corresponds to 60%, the single displacement of 10 mm corresponds to 70%, and the single displacement of 20 mm corresponds to 90%. (3) Supply chain risk loss = average daily freight volume ÷ 24 × duration of interruption in evolution stage × unit freight cost × default coefficient. The unit freight cost is determined based on the annual guidance price of the transportation industry. The duration of interruption in the evolution stage increases linearly with the displacement rate. When the displacement rate is 3 mm / month, it corresponds to 12 hours. When the displacement rate is 5 mm / month, it corresponds to 24 hours. When the displacement rate is 10 mm / month, it corresponds to 72 hours. When the single displacement is ≥10 mm, it increases by 24 hours / 10 mm and the upper limit of the interruption duration is 168 hours. The default coefficient is 1.2 for high risk and 1.5 for extremely high risk. High risk corresponds to the accelerated deformation stage and extremely high risk corresponds to the early warning stage.

8. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S4: The dynamic early warning threshold of the evolution stage-risk early warning linkage system is set as follows: a single displacement greater than 8mm triggers an early warning in the stable stage; a single displacement greater than 6mm triggers an early warning in the slow deformation stage; a single displacement greater than 5mm triggers an early warning in the accelerated deformation stage; and a single displacement greater than 3mm triggers an early warning in the early warning stage. The early warning information push is prioritized, specifically in the following order: push of resident evacuation trigger conditions, push of emergency personnel dispatch suggestions, push of farmer subsidy application portal, and push of alternative logistics route plans for enterprises. The early warning information push time limit shall not exceed 15 minutes, and the push channels include government affairs platforms, enterprise management systems, farmer-specific APPs, resident SMS, and community notices. Community notices are displayed on electronic screens and broadcasted.

9. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S5: Triggering conditions and verification standards for the prevention and control fund disbursement contract: (1) Medium risk corresponds to the slow deformation stage: triggering the allocation of 30% of the total prevention and control funds; the verification standard is that the installation rate of monitoring equipment is not less than 90%, verified by drone aerial photography, the drone flight altitude is not more than 100m, the heading overlap is not less than 80%, the lateral overlap is not less than 60%, the shooting resolution is not less than 0.1m, the shooting time is within 24 hours after the completion of the project, and at least 1 monitoring equipment point is set up for every 100m²; (2) High risk corresponds to accelerated deformation stage: triggering the allocation of 60% of the total prevention and control funds; the verification standard is that the progress of temporary anti-slide engineering is not less than 80%, which is confirmed by comparing the engineering supervision report with drone aerial photography, and the completion rate of anti-slide pile pouring and the anchor bolt implantation rate are not less than 80%; (3) The warning stage corresponding to extremely high risk: the remaining 40% of the prevention and control funds will be disbursed; the verification standard is that the landslide displacement rate drops to below 5 mm / month, and the average displacement rate is less than 5 mm / month after continuous monitoring for 72 hours; 5% of the total prevention and control funds will be frozen as a quality guarantee deposit. If the evolution stage is stable to the slow deformation stage and there is no re-slide after 1 year after the funds are disbursed, the quality guarantee deposit will be unlocked; the re-slide judgment standard is that the single displacement is greater than 10 mm or the monthly displacement rate is greater than 5 mm / month within 1 year.

10. The landslide risk assessment method based on the dynamic evolution law of landslides as described in claim 1, characterized in that, In step S3: Cross-regional ecological compensation decision-making uses blockchain technology to record the correlation between the evolutionary stages of upstream and downstream regions. The correlation between the upstream early warning stage and the downstream stable stage is calculated at 80%. The blockchain records include the start time of upstream prevention and control projects, the completion time of upstream prevention and control projects, details of upstream prevention and control investment, water quality monitoring data of downstream protected areas, and soil stability monitoring data of downstream protected areas. Specific indicators for water quality monitoring data in downstream protected areas include pH value 6.0-9.0 and chemical oxygen demand ≤50mg / L. Specific indicators for soil stability monitoring data in downstream protected areas include soil moisture content 15%-25% and internal friction angle ≥25°. Both types of monitoring data are collected monthly. The blockchain adopts a consortium blockchain architecture, with nodes including upstream and downstream local governments, ecological and environmental departments, and financial departments. Data modifications must be confirmed by at least 2 / 3 of the nodes before taking effect.

Images