Grassland carbon sink management decision support system based on multi-factor weighted regression model
By integrating multi-source data through a multi-factor weighted regression model (MWRM), automated decision support for grassland carbon sequestration management systems is achieved, overcoming the shortcomings of existing technologies in high-precision assessment and scenario simulation, and improving management efficiency and decision accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- AGRI GENOMICS INST CHINESE ACADEMY OF AGRI SCI
- Filing Date
- 2026-01-20
- Publication Date
- 2026-06-26
Abstract
Description
Technical Field
[0001] This invention provides a grassland carbon sequestration management decision support system based on a multi-factor weighted regression model, belonging to the field of digital agriculture and ecological information decision technology. Background Technology
[0002] Currently, there are some general-purpose Geographic Information Systems (GIS) and remote sensing monitoring platforms that can display basic information such as vegetation indices and land use. There are also some independent soil carbon models. However, they are typically functionally fragmented and their models are isolated: Option 1: Basic GIS platforms, such as ArcGIS or QGIS, can manage and display spatial data, but lack embedded, high-precision grassland carbon sequestration assessment models, and cannot automatically complete the full closed loop from data to carbon sequestration accounting to decision-making recommendations. Option 2: Traditional carbon sequestration monitoring systems. These systems may rely solely on a single NDVI index to estimate biomass and then roughly extrapolate carbon sequestration. Their models are overly simplistic, have limited accuracy, and are not deeply coupled with specific spatial management measures (such as fencing planning).
[0003] The shortcomings of existing technology:
[0004] Disadvantage 1: The model is disconnected from the system. Even with advanced algorithms, they remain only in research papers and have not been engineered into a stable, easy-to-use system module that can interact with business data in real time.
[0005] Disadvantage 2: Weak decision support function. Most existing systems remain at the level of "data visualization" and "current status display", lacking the "what-if" scenario simulation function, and cannot provide managers with the ability to quantify and compare the advantages and disadvantages of different management solutions.
[0006] Disadvantage 3: Non-automated process. From data preprocessing and model calculation to result output, a large amount of manual intervention is required, which is inefficient and makes it difficult to meet the needs of frequent dynamic monitoring and rapid decision-making of carbon sequestration. Summary of the Invention
[0007] To address the aforementioned technical problems, this invention provides a grassland carbon sequestration management decision support system based on a multi-factor weighted regression model, establishing an automated and intelligent decision support system integrating high-precision accounting, scenario simulation, and spatial planning. Specifically, the problems to be solved are:
[0008] 1. How to engineer the advanced carbon storage estimation model (Multi-factor Weighted Regression Model, MWRM) into a stable and efficient system core engine.
[0009] 2. How to achieve automatic aggregation, fusion and processing of multi-source data to form a decision data base.
[0010] 3. How to develop quantifiable decision support functions that can directly guide production practices based on the spatial distribution of carbon sinks (such as the delineation of grazing ban areas, assessment of grazing pressure, and prediction of engineering benefits).
[0011] The specific technical solution provided by this invention is as follows:
[0012] The grassland carbon sink management decision support system based on the multi-factor weighted regression model includes a multi-source data integration and processing module, a MWRM model engine module, a carbon sink accounting and dynamic assessment module, a scenario simulation and decision support module, and a visualization and interactive output module.
[0013] The multi-source data integration and processing module, as the system's data entry point, is responsible for collecting and standardizing heterogeneous data from different channels.
[0014] Input data include: remote sensing data, meteorological data, topographic data, soil type data, field measured SOC sample data, basic geographic information, and human activity data.
[0015] The output generates a standardized raster dataset with the same spatial resolution and coordinate system, which can be directly called by the MWRM model engine module.
[0016] The MWRM model engine module, as the system's processing center, employs a multi-factor weighted regression model algorithm and features callable and configurable microservices.
[0017] It receives standardized data from the multi-source data integration and processing module, automatically performs model training or prediction tasks, and outputs a high-precision, spatially continuous distribution map of soil organic carbon storage.
[0018] It outputs a current status map of carbon storage, as well as a weight distribution map of various environmental factors.
[0019] The carbon sink accounting and dynamic assessment module converts the output of the MWRM model engine module into business language.
[0020] Based on SOC distribution maps and combined with grassland type area data, the system automatically calculates the total carbon sequestration and carbon density of different administrative or natural units. It can also compare carbon sequestration data from different periods, enabling dynamic monitoring and visualization of carbon sequestration changes.
[0021] The scenario simulation and decision support module provides three decision support functions:
[0022] (1) Identification of carbon sink hotspots and priority delineation of protection: Based on the current SOC status map and its changing trends, the system automatically identifies "carbon sink hotspot areas" and "carbon sink vulnerable areas". The output is a suggested spatial layout map of protection and restoration projects.
[0023] (2) Grazing pressure optimization analysis: The system overlays and analyzes the current livestock distribution data with the spatial distribution of carbon sinks. When it detects that "high grazing pressure" and "high carbon sink loss risk" overlap in a certain area, the system will automatically issue an early warning and provide quantitative suggestions for livestock relocation or grassland rotation.
[0024] (3) Preliminary assessment of the benefits of management measures: The user draws a virtual fenced area. The system will simulate and predict the potential growth of carbon storage in the area in the next 1-5 years based on the MWRM model engine module and historical data trends, and display the input-output benefits in the form of charts to provide data support for project establishment.
[0025] The visualization and interactive output module provides a graphical interface on the web or desktop, presenting the results of all the above modules in the form of thematic maps, statistical charts, and evaluation reports.
[0026] The technical effects of the present invention are as follows:
[0027] 1. Scientific and precise decision-making: Upgrading management decisions from "experience-driven" to "data-driven and model-driven". Through the heterogeneity mapping of carbon sink space, precise management of "one policy for one location" has been achieved, avoiding resource waste and greatly improving the pertinence and effectiveness of management measures.
[0028] 2. Significantly improved efficiency: It has achieved full automation from data to decision-making, reducing the manual analysis work that originally required several weeks to the hour level, making high-frequency dynamic monitoring of carbon sequestration and rapid decision response possible.
[0029] 3. Foresight and practicality of functions: The unique "scenario simulation" function gives managers the ability to "foresee the future" and can conduct quantitative pre-assessment of the ecological benefits (carbon sink increment) before project implementation, which significantly reduces investment risk and optimizes capital allocation.
[0030] 4. System Integration and Engineering Value: For the first time, the model was successfully transformed from a "scientific research tool" into the "core of the business system," solving the "last mile" problem of technology and application, and providing a practical solution for the digital governance of grassland carbon sequestration. Detailed Implementation
[0031] The unique architecture of the grassland carbon sink management decision support system based on a multi-factor weighted regression model provided by this invention lies in the closed-loop collaboration of five major modules and the design of unidirectional data flow.
[0032] The system architecture and specific module functions are as follows:
[0033] 1. Multi-source data integration and processing module:
[0034] Function: As the system's data entry point, it is responsible for automatically collecting, cleaning, and standardizing heterogeneous data from different channels.
[0035] Inputs: Remote sensing data (MODIS / Landsat NDVI), meteorological data (precipitation, temperature), topographic data (DEM), soil type data, field measured SOC sample data, basic geographic information (administrative divisions, roads, settlements), and human activity data (grazing intensity, fence location, distribution of ecological engineering).
[0036] Output: Generates a standardized raster dataset with unified spatial resolution and coordinate system, which can be directly called by the MWRM model engine module.
[0037] 2. MWRM Model Engine Module:
[0038] Function: As the processing center of the system, it adopts a multi-factor weighted regression model algorithm, but its key feature is that it is designed as a callable and configurable microservice.
[0039] Workflow: Receive standardized data from the multi-source data integration and processing module → Automatically execute model training or prediction tasks → Output a high-precision, spatially continuous soil organic carbon storage distribution map.
[0040] Unique design: This engine not only outputs a current carbon storage map, but also a weight distribution map of various environmental factors, providing decision-makers with a scientific basis for understanding the dominant factors in the spatial differentiation of carbon sinks.
[0041] 3. Carbon Sequestration Accounting and Dynamic Assessment Module:
[0042] Function: Converts the output of the MWRM model engine module into business language.
[0043] Workflow: Based on SOC distribution maps and combined with data such as grassland type area, the system automatically calculates the total carbon sequestration and carbon density of different administrative units (counties, townships, villages) or natural units (watersheds, grassland type zones). It can also compare carbon sequestration data from different periods, enabling dynamic monitoring and visualization of carbon sequestration changes.
[0044] 4. Scenario Simulation and Decision Support Module (The core innovation of this invention):
[0045] Functionality: This is the most crucial module that distinguishes the system from pure research models; it provides three unique decision support functions:
[0046] Function A (Carbon Sink Hotspot Identification and Protection Priority Determination): Based on the current SOC status map and its changing trends, the system automatically identifies "carbon sink hotspot areas" (high reserves, high stability) and "carbon sink vulnerable areas" (low reserves, high degradation risk). The output is a suggested spatial layout map for protection and restoration projects (e.g., strict grazing bans should be implemented in area A, and light utilization should be implemented in area B).
[0047] Function B (Grazing Pressure Optimization Analysis): The system overlays and analyzes the current livestock distribution data with the spatial distribution of carbon sinks. When it detects that "high grazing pressure" and "high carbon sink loss risk" overlap spatially in a certain area, the system will automatically issue an early warning and provide quantitative suggestions for livestock relocation or pasture rotation.
[0048] Function C (Preliminary Assessment of the Benefits of Management Measures): Users can draw a virtual fenced area (simulating grazing ban). The system will simulate and predict the potential increase in carbon storage in the area over the next 1-5 years based on the MWRM model engine module and historical data trends, and display the input-output benefits in the form of charts to provide data support for project initiation.
[0049] 5. Visualization and Interactive Output Module:
[0050] Functionality: Provides a graphical interface on web or desktop, visually presenting the results of all the above modules in the form of thematic maps, statistical charts, evaluation reports, etc. Administrators can directly invoke the scenario simulation function through simple operations such as "clicking on a drawing frame".
[0051] System data flow: Multi-source data integration and processing module (data) → MWRM model engine module (core calculation) → Carbon sink accounting and dynamic assessment module (business quantification) → Scenario simulation and decision support module (intelligent decision-making) → Visualization and interactive output module (result presentation).
[0052] Multi-source data integration and processing module (data) → MWRM model engine module (core computing) → Carbon sink accounting and dynamic assessment module (business quantification) → Scenario simulation and decision support module (intelligent decision-making) → Visualization and interactive output module (result presentation).
[0053] This invention is a unidirectional, hierarchical data flow that gradually refines raw data into final decision knowledge, forming a complete closed loop.
Claims
1. A grassland carbon sequestration management decision support system based on a multi-factor weighted regression model, characterized in that, It includes a multi-source data integration and processing module, a MWRM model engine module, a carbon sink accounting and dynamic assessment module, a scenario simulation and decision support module, and a visualization and interactive output module; The multi-source data integration and processing module, as the system's data entry point, is responsible for collecting and standardizing heterogeneous data from different channels; The MWRM model engine module, as the system's processing center, adopts a multi-factor weighted regression model algorithm and has callable and configurable microservices. The carbon sink accounting and dynamic assessment module converts the output of the MWRM model engine module into business language. The scenario simulation and decision support module provides three decision support functions: (1) Identification of carbon sink hotspots and priority delineation of protection: Based on the current status map and trend of SOC, the system automatically identifies "carbon sink hotspot areas" and "carbon sink vulnerable areas"; the output is a suggested spatial layout map of protection and restoration projects; (2) Grazing pressure optimization analysis: The system overlays and analyzes the current livestock distribution data with the spatial distribution of carbon sinks; when it detects that "high grazing pressure" and "high carbon sink loss risk" overlap in a certain area, the system will automatically issue an early warning and provide quantitative suggestions for livestock transfer or grassland rotation. (3) Preliminary assessment of the benefits of management measures: The user draws a virtual fenced area. The system will simulate and predict the potential growth of carbon storage in the area in the next 1-5 years based on the MWRM model engine module and historical data trends, and display the input-output benefits in the form of charts to provide data support for project establishment. The visualization and interactive output module provides a graphical interface on the web or desktop to present the results of all the above modules.
2. The grassland carbon sequestration management decision support system based on a multi-factor weighted regression model according to claim 1, characterized in that, The multi-source data integration and processing module receives input data including: remote sensing data, meteorological data, topographic data, soil type data, field measured SOC sample data, basic geographic information, and human activity data. The output generates a standardized raster dataset with the same spatial resolution and coordinate system, which can be directly called by the MWRM model engine module.
3. The grassland carbon sequestration management decision support system based on a multi-factor weighted regression model according to claim 1, characterized in that, The MWRM model engine module receives standardized data from the multi-source data integration and processing module, automatically performs model training or prediction tasks, and outputs a high-precision, spatially continuous soil organic carbon storage distribution map. It outputs a current status map of carbon storage, as well as a weight distribution map of various environmental factors.
4. The grassland carbon sequestration management decision support system based on a multi-factor weighted regression model according to claim 1, characterized in that, The carbon sequestration and dynamic assessment module, based on the SOC distribution map and combined with grassland type area data, automatically calculates the total carbon amount and carbon density of different administrative units or natural units; and can compare carbon sequestration data at different times to achieve dynamic monitoring and visualization of carbon sequestration changes.
5. The grassland carbon sequestration management decision support system based on a multi-factor weighted regression model according to claim 1, characterized in that, The visualization and interactive output module presents information intuitively in the form of thematic maps, statistical charts, and evaluation reports.