Space-air-ground collaborative multi-greenhouse gas carbon budget dynamic accounting method and system

Abstract

The application relates to a space-air-ground coordinated multi-greenhouse gas carbon budget dynamic accounting method and system, which comprises the following steps: obtaining farmland operation related data, identifying a farmland operation event, and constructing an event accounting unit based on the farmland operation event; determining an event response time window; obtaining multi-source data in the corresponding event response time window, and forming a state vector corresponding to the event accounting unit after quality control of the multi-source data; performing state consistency screening on the obtained multi-source data according to crop growth stages, soil states, nitrogen states and meteorological conditions, and constructing a same-state flux baseline; calculating the observability score of the event accounting unit, obtaining the estimation results of carbon dioxide, methane and nitrous oxide based on the multi-source data, calculating the total event greenhouse gas emission and the event carbon sequestration, and obtaining the event net greenhouse gas budget by deducting the carbon sequestration from the total event greenhouse gas emission.

Description

Technical Field

[0001] This invention relates to the field of agricultural carbon accounting technology, and in particular to a dynamic accounting method and system for multi-greenhouse gas carbon balance involving space, air, and ground. Background Technology

[0002] Agricultural ecosystems have the ability to fix carbon through crop photosynthesis, crop biomass accumulation, and soil organic carbon accumulation. At the same time, they also generate carbon dioxide, methane, and nitrous oxide emissions from agricultural operations such as tillage, sowing, fertilization, irrigation, drainage, straw return to the field, pest and disease control, harvesting, and agricultural machinery fuel consumption.

[0003] Current methods for calculating greenhouse gas emissions and carbon budgets in farmland typically employ ground flux observations, remote sensing inversion, process model simulations, or statistical emission factor methods. While ground flux observations and process model methods offer some process interpretation capabilities, they struggle to consistently cover all farmland plots and peak emission periods following agricultural operations. Remote sensing methods can achieve regional coverage, but they usually focus on spatial pixels, plots, or regional carbon sink status, making it difficult to accurately identify short-term greenhouse gas pulse changes caused by specific farmland operations such as fertilization, irrigation, tillage, drainage, and straw return. Statistical emission factor methods are mostly based on annual or seasonal inputs, making it difficult to reflect emission differences in the same plot under different crop growth stages, soil hydrothermal conditions, and management practices.

[0004] The existing technology has at least the following problems: Most farmland carbon sink or carbon budget accounting methods use spatial pixels, plots or regions as basic accounting boundaries, and do not take specific farmland operation events such as fertilization, irrigation, tillage, drainage and straw return to the field as direct driving factors of greenhouse gas budget changes, making it difficult to reliably attribute carbon budget changes to specific management measures. Existing multi-source data fusion typically processes data based on the spatial resolution of satellites, drones, and ground sensors, fixed calendar cycles, or pixel matching relationships, without considering the timing of farmland operations, operation intensity, operation coverage, crop growth stage, and post-operation greenhouse gas response to determine the accounting window. Existing process models, remote sensing estimates, and ground observation data are usually fused using simple weighting, statistical regression, data assimilation, or residual correction methods. They lack a joint correction mechanism that is constrained by carbon conservation, nitrogen balance, water status, and operational event response patterns. This can easily lead to situations where the calculation results appear to match the observation data but do not meet the boundaries of agricultural processes. Existing supplementary monitoring mechanisms are usually determined based on routine monitoring plans, missing remote sensing data, spatial resolution or pixel registration issues, without prioritizing supplementary monitoring based on key observation gaps during the operational event response period and their contribution to the event's net greenhouse gas budget uncertainty. The evaluation of existing management practices usually uses fertilizer application, irrigation amount or farming methods as ordinary input variables and evaluates the contribution through feature importance, correlation analysis or limiting factor diagnosis. It is difficult to obtain auditable event-level carbon budget differences formed by replacing only the target operation measures in the same plot, the same crop growth stage and similar environmental conditions. Existing outputs are typically regional carbon sink distribution maps, annual summary values ​​for plots, or crop season summary values. They lack event-level carbon budget ledgers that can be traced back to specific operational events, original observation data, baseline sources, accounting boundaries, constraint verification results, model versions, and uncertainty levels.

[0005] Therefore, it is necessary to provide a space-air-ground collaborative dynamic accounting system and method for multiple greenhouse gas carbon budgets, which uses farmland operation events as the accounting index, greenhouse gas dynamic response processes as the observation boundary, accounting uncertainty contributions as the basis for supplementary measurements, carbon-nitrogen-water process constraints as correction conditions, and event substitution counterfactual scenarios as the attribution basis. Summary of the Invention

[0006] In view of this, the purpose of this invention is to provide a dynamic accounting method and system for multi-greenhouse gas carbon budget in a space-air-ground coordinated manner, so as to solve the problems of existing farmland carbon budget accounting which is difficult to correspond to specific farmland operation events, different observation scales are difficult to serve the monitoring of key response periods after operations, multi-source data fusion lacks agricultural process constraints, observation resources are difficult to schedule for accounting uncertainty, the contribution of management measures is difficult to reliably attribute, and the accounting results are difficult to trace.

[0007] The technical solution adopted by this invention to solve the above-mentioned technical problems is: a dynamic accounting method and system for multi-greenhouse gas carbon budget with space-air-ground coordination, comprising the following steps: S1. Obtain relevant data on farmland operations, identify farmland operation events, and construct event accounting units based on farmland operation events; S2. Determine the event response time window. The event response time window includes the pre-event baseline period, the event occurrence period, and the post-event response period. The endpoint of the event response time window is determined based on the degree of regression of the target greenhouse gas flux relative to the baseline flux in the same state. S3. For the event accounting unit, acquire multi-source data within the corresponding event response time window. The multi-source data includes space-based observation data, air-based observation data, ground-based observation data, and operational activity data. After quality control, the multi-source data forms a state vector corresponding to the event accounting unit. S4. The acquired multi-source data are screened for consistency according to crop growth stage, soil condition, nitrogen status and meteorological conditions, and a flux baseline with the same status is constructed. S5. Calculate the observability score of the event accounting unit. When the observability score is less than the preset observability threshold, identify the missing key observations in the event accounting unit and trigger the targeted supplementary measurement task based on the contribution of the missing key observations to the uncertainty of the event's net greenhouse gas balance. S6. Based on multi-source data, obtain the estimated results of carbon dioxide, methane and nitrous oxide, calculate the total greenhouse gas emissions and carbon sequestration of the event, and subtract the carbon sequestration from the total greenhouse gas emissions to obtain the net greenhouse gas balance of the event.

[0008] Furthermore, in step S1, acquiring farmland operation-related data and identifying farmland operation events specifically includes: acquiring the plot boundaries, crop types, crop growth stages, agricultural machinery operation trajectories, agricultural machinery working status, agricultural logs, input records, irrigation and drainage records, and agricultural machinery fuel or energy consumption records of the target farmland area. Based on at least two types of information from agricultural machinery operation trajectories, agricultural machinery working status, input records, irrigation and drainage records, and agricultural logs, identify one or more agricultural operation events among tillage, sowing, fertilization, irrigation, drainage, straw return to the field, pest and disease control, harvesting, and agricultural machinery operations; The event accounting unit based on farmland operation events specifically includes: For each farmland operation event, an event accounting unit is constructed. The event accounting unit is determined by at least the associated plot or operation coverage sub-region, crop growth stage, operation event type and operation intensity, event response time window, spatial range of operation event impact, and initial soil-crop-meteorological state at the time of the event.

[0009] Furthermore, in step S2, the baseline period before the event is the effective observation period of 1 to 7 consecutive days before the event occurs; when there is insufficient effective observation data before the event or there is significant disturbance before the event, effective observation data of the same historical period, the same crop growth stage and matching status of the same plot are used as supplementary baseline data. When the same event involves carbon dioxide, methane and nitrous oxide, the overall response period endpoint of the event accounting unit is the latest time among the response period endpoints of the three gases; or, for different event categories, the response period endpoint of a preset priority gas is selected as the overall response period endpoint.

[0010] Furthermore, in step S3, after quality control, the multi-source data forms a state vector corresponding to the event accounting unit, specifically including: Perform radiometric correction, atmospheric correction, geometric correction, cloud and shadow removal, and anomalous pixel removal on space-based observation data; Attitude correction, ground control point correction, image stitching, spatial registration, image sharpness screening, and abnormal flight data removal are performed on the airborne observation data. Perform instrument drift checks, outlier removal, data integrity checks, and timestamp corrections on ground-based observation data; The data from operational activities are processed to remove sensor fault values, remove duplicate records, mark missing records, and synchronize time. The multi-source data that has undergone quality control is subjected to coordinate benchmark unification, time synchronization, spatial clipping, event association, and state variable extraction to form a state vector corresponding to the event accounting unit.

[0011] Furthermore, in step S4, the state consistency screening includes at least the following rules: When the crop growth stage of the candidate data is inconsistent with that of the target event data, it is determined that the states are inconsistent. When the candidate data and the target event data are at different soil moisture levels, they are judged to be inconsistent in state; For nitrogen application events, when the soil inorganic nitrogen status of candidate data and target event data crosses the preset inorganic nitrogen level, it is determined that the status is inconsistent; For irrigation, precipitation, or drainage events, when the candidate data and the target event data are in different irrigation / drainage duration intervals, water accumulation state intervals, or redox state intervals, they are judged to be inconsistent in state. When the difference between the soil temperature, canopy coverage, or meteorological conditions of the candidate data and the target event data exceeds the corresponding preset threshold, it is determined to be an inconsistency.

[0012] Furthermore, in step S5, observability evaluation is based on the effective coverage, temporal proximity, data quality, spatial representativeness, and state consistency of the data source during the response period. When the score is lower than a preset threshold, the missing key observations are identified, and the contribution of each missing key observation to the net greenhouse gas budget uncertainty of the event is determined through univariate perturbation, local sensitivity analysis, conditional Monte Carlo sampling, or variance decomposition. Based on the contribution value, the timeliness of supplementary measurement, and the cost of supplementary measurement, priority is given to performing supplementary measurement tasks that can be completed within the response period. Supplementary measurement tasks include UAV supplementary flights, ground flux densification sampling, soil parameter supplementary measurement, work record supplementary recording, and supplementary identification of work boundaries.

[0013] Furthermore, the event emissions are the cumulative difference between the target gas flux and the baseline flux under the same conditions during the event response period, and selectively include emissions from input production, agricultural machinery fuel, and energy consumption; carbon sequestration is determined using either the flux-driven accounting model or the inventory change accounting model. Under the flux-driven model, the integral result of the net ecosystem carbon exchange within the event response time window is used as the event carbon sequestration. Under the inventory change accounting model, when continuous net ecosystem carbon exchange observation data are lacking, the carbon sequestration of events is based on crop biomass carbon increment, soil organic carbon inventory changes, and straw or other organic matter inputs.

[0014] Furthermore, it also includes: S7. Construct a process constraint set including carbon conservation constraints, nitrogen balance constraints, water state constraints, and operational event response constraints. Determine the feasible region of the event response state based on the process constraint set. Determine the adaptive fusion weights of the process model estimation results, ground observation results, and remote sensing-aided estimation results based on observability scores, data quality, and uncertainties of each data source to obtain fusion candidate states. Then, restrict or project the fusion candidate states to the feasible region that satisfies the process constraint set to obtain the final event response state. S8. Construct actual operation event scenarios and counterfactual management baseline scenarios for the same event accounting unit. The counterfactual management baseline scenarios keep the plot, crop type, crop growth stage, initial soil condition, meteorological conditions, non-target state variables and non-target operation activities corresponding to the event accounting unit unchanged, and only replace the target farmland operation events. Calculate the net greenhouse gas balance under the actual operation event scenarios and the counterfactual management baseline scenarios respectively to obtain the event-level carbon balance contribution of the target management measures. S9. Generate event-level carbon income and expenditure ledger entries for each event accounting unit, and summarize and output the event-level carbon income and expenditure ledger entries according to four levels: event, plot, crop season, and year.

[0015] Furthermore, the carbon conservation constraint is used to define the boundary relationships between crop carbon accumulation, soil organic carbon change, carbon dioxide exchange, organic carbon input, and organic carbon output. The nitrogen balance constraint is used to define the boundary relationships between nitrogen fertilizer input, changes in soil inorganic nitrogen stock, crop nitrogen uptake, nitrogen leaching, ammonia volatilization, and nitrous oxide emissions. The water state constraint is used to define the boundary relationships between irrigation amount, drainage amount, precipitation, evapotranspiration, field water level, soil moisture content, water accumulation state, redox potential and methane and nitrous oxide emission responses. The job event response constraints include at least one of the following: For nitrogen application events, the jointly corrected nitrous oxide flux status should be consistent with the soil inorganic nitrogen content, soil moisture content, and soil temperature status. For irrigation, precipitation, or drainage events, the combined corrected methane and nitrous oxide flux states should be consistent with soil moisture content, waterlogging status, drainage status, or redox potential status. For tillage or straw return events, the jointly corrected carbon dioxide flux status should be consistent with the bare soil ratio, soil temperature, tillage depth, straw input, or organic matter input status. The difference in gas flux between the treated area and the untreated control area shall not exceed the preset boundary allowed by differences in crop growth stage, soil condition and meteorological condition. For complex events, the gas response time series corresponding to each sub-event is consistent with the order of operation, the time interval between operations, and the changes in state variables.

[0016] This application also provides a space-air-ground integrated dynamic accounting system for multiple greenhouse gas carbon budgets, based on the above-mentioned space-air-ground integrated dynamic accounting method for multiple greenhouse gas carbon budgets, including: The operation event recognition module is used to receive data such as plot boundaries, crop types, crop growth stages, agricultural machinery operation trajectories, agricultural machinery working status, agricultural logs, input records, irrigation and drainage records, and fuel or energy consumption records, to identify farmland operation events and generate event description vectors. The event accounting unit construction module is used to construct event accounting units based on operation events, related plots or operation coverage sub-areas, crop growth stages, operation intensity, operation impact range, initial soil-crop-meteorological state, and dynamic response time windows, and to perform independent accounting, truncation processing, or composite event merging processing on overlapping events. The air-space-ground integrated observation module is used to acquire satellite remote sensing data, UAV remote sensing data, ground flux observation data, soil observation data, meteorological observation data, and operational activity data; The unified state representation and quality control module is used to unify coordinate benchmarks, synchronize time, remove outliers, control quality, prune space, associate events, and construct state vectors for multi-source data, and output the event state vectors corresponding to the event accounting units. The module for constructing state consistency and same-state flux baselines is used to screen candidate observation data for state consistency based on crop growth stage, soil moisture level, soil temperature level, inorganic nitrogen status, duration after irrigation and drainage, meteorological conditions and duration after operation, and construct same-state flux baselines according to preset priorities. The observability assessment and targeted supplementary measurement module is used to calculate the observability score of the event accounting unit; when the observability score is insufficient, it determines the contribution value of the missing observations to the uncertainty of the event's net greenhouse gas balance sheet, and generates targeted supplementary measurement tasks based on the contribution value, the timeliness of supplementary measurement, and the cost of supplementary measurement. The multi-greenhouse gas process accounting module is used to obtain process model estimation results, ground observation results and remote sensing-aided estimation results for carbon dioxide, methane and nitrous oxide based on event state vectors, and to calculate the target gas emission increment, input-related equivalent emissions, event carbon sequestration and event net greenhouse gas balance. The process constraint joint correction module is used to receive the event accounting unit, event state vector, same-state flux baseline, observability score and candidate flux state output by the multi-greenhouse gas process accounting module. It determines the fusion weight based on the data source uncertainty, observability score and data quality, and restricts the fusion candidate state to the feasible region that satisfies the constraints of carbon conservation, nitrogen balance, moisture state and operational event response. It outputs the jointly corrected event response state. The counterfactual attribution module is used to construct actual operational event scenarios and counterfactual management baseline scenarios for the same event accounting unit, keeping the plot, crop growth stage, initial soil condition, meteorological conditions and non-target operational activities consistent, only replacing the target operational events, and calculating the event-level carbon budget contribution of management measures; The event-level ledger and reporting module is used to generate auditable event-level carbon budget ledgers and output multi-level greenhouse gas carbon budget results, quality grades, management action contribution results, maps, and standardized accounting reports at the event level, plot level, crop season level, and annual level.

[0017] The beneficial effects of this application are as follows: 1. This application constructs an event accounting unit by combining farmland operation events, crop growth stages, operation coverage, operation intensity, initial state and dynamic response time window, which can track changes in greenhouse gas balance for key disturbances such as fertilization, irrigation, drainage, tillage and straw return to the field.

[0018] 2. This application determines the endpoint of the operational event response period by the degree of regression of carbon dioxide, methane and nitrous oxide fluxes relative to the baseline fluxes under the same conditions, avoiding the omission of emission peaks, unreasonable window truncation or invalid observations caused by fixed calendar windows.

[0019] 3. This application performs state matching in terms of crop growth stage, soil moisture, soil temperature, inorganic nitrogen status, post-irrigation and drainage duration and meteorological conditions to avoid directly mixing data from different farmland management states for flux estimation and baseline construction.

[0020] 4. This application prioritizes supplementary measurements based on their contribution to the uncertainty of the net greenhouse gas balance of the event during the event response period, so that supplementary UAV flights, ground-based intensive sampling, soil supplementary measurements, and supplementary recording of work records can serve the weak links in the accounting.

[0021] 5. This application does not simply weight or superimpose residuals from process models, remote sensing-aided estimations, and ground observations. Instead, it restricts the fusion of candidate states to the feasible domain of constraints on carbon conservation, nitrogen balance, water status, and operational event response, thereby reducing the risk of violating agricultural process boundaries.

[0022] 6. This application can simultaneously perform event-level accounting for carbon dioxide, methane, nitrous oxide, and carbon sequestration, and improve the reliability of the event-based net greenhouse gas balance through accounting boundary consistency verification and duplicate measurement elimination processing.

[0023] 7. This application, while maintaining consistency in non-target operational activities and environmental conditions, only replaces target operational events to reduce the confounding effects caused by different sites, growth stages, hydrothermal conditions, and nitrogen status, thereby obtaining recalcible event-level emission reduction or carbon sequestration contributions.

[0024] 8. This application links the accounting results to specific operational events, operational coverage, raw data, same-state flux baseline, counterfactual management baseline, accounting boundaries, process constraint verification, model version, and uncertainty level, which can support agricultural carbon sink project verification, low-carbon agronomic optimization, and carbon asset management.

[0025] 9. This invention transforms air-space-ground observation from general spatial pixel fusion into adaptive collaborative observation and auditable accounting for the impact period of farmland operations through closed-loop processing of operation event identification, event accounting unit construction, dynamic response window determination, state consistency screening, same-state flux baseline construction, observability evaluation and uncertainty-driven supplementary measurement, preliminary calculation of multiple greenhouse gas budgets, joint correction of process constraints, event substitution counterfactual attribution, and event-level ledger output. Attached Figure Description

[0026] Figure 1 This is the overall flowchart of the method of the present invention.

[0027] Figure 2 This is a block diagram of the overall structure of the system of the present invention.

[0028] Figure 3 This is a schematic diagram of the event response joint correction structure based on the process constraint feasible region of the present invention.

[0029] Figure 4 This is a schematic diagram of the event-level carbon income and expenditure ledger and multi-level output of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that, in the description of this invention, unless otherwise stated, "a plurality of" means two or more; the terms "upper," "lower," "left," "right," "inner," "outer," "front end," "rear end," "head," "tail," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0031] Please see Figure 1-4 This application provides a method for dynamic accounting of multi-greenhouse gas carbon budget through a space-air-ground integrated approach, comprising the following steps.

[0032] Step 1: Construction of farmland operation event identification and event accounting unit.

[0033] Acquire information on plot boundaries, crop types, crop growth stages, agricultural machinery operation trajectories, agricultural machinery operating status, agricultural logs, input records, irrigation and drainage records, and agricultural machinery fuel or energy consumption records for the target farmland area.

[0034] Based on at least two types of information from agricultural machinery operation trajectories, agricultural machinery working status, input records, irrigation and drainage records, and manual agricultural logs, identify one or more agricultural operation events among tillage, sowing, fertilization, irrigation, drainage, straw return to the field, pest and disease control, harvesting, and agricultural machinery operations.

[0035] Among them: agricultural machinery operation trajectory is used to determine the operation location, operation time, trajectory continuity and operation coverage; agricultural machinery working status is used to determine the machinery type, working status and operation intensity; input record is used to determine the type, amount, batch and application method of fertilizer, pesticide, straw or other inputs; irrigation and drainage record is used to determine irrigation volume, drainage volume, irrigation and drainage method and irrigation and drainage time; manual agricultural log is used to supplement and confirm the operation category, operation time period and abnormal situation.

[0036] When at least two of the following data meet the preset criteria for the corresponding operation category: agricultural machinery operation trajectory, agricultural machinery working status, input records, irrigation and drainage records, and manual farming logs, a corresponding farmland operation event is generated.

[0037] For each farmland operation event, an event description vector is generated, which includes: event identifier, associated plot identifier, crop type, crop growth stage, event category, event start time and end time, operation intensity, operation coverage, input attributes, irrigation and drainage attributes, agricultural machinery attributes, fuel or energy consumption attributes, and the initial soil-crop-meteorological state at the time of the event.

[0038] The operational coverage area is determined based on the agricultural machinery's operational trajectory line, the width of the operational implements, and the positioning error buffer distance. When the time interval between two adjacent operational events of the same type is less than a preset time interval threshold, the spatial distance between the operational coverage areas is less than a preset spatial distance threshold, and the change in operational intensity does not exceed a preset change threshold, the two adjacent operational events are merged into a single operational event.

[0039] When different types of operational events occur within a preset time range, and their operational coverage overlaps, their expected response time windows overlap, or their corresponding soil-crop-meteorological conditions influence each other, the different types of operational events are merged into a composite event; otherwise, independent event accounting units are established separately.

[0040] For each farmland operation event, an event accounting unit is constructed. The event accounting unit is determined by at least the associated plot or operation coverage sub-region, crop growth stage, operation event type and intensity, event response time window, spatial range of operation event impact, and initial soil-crop-meteorological state at the time of the event.

[0041] Step 2: Determine the dynamic response time window and handle overlapping events.

[0042] The event response time window includes the pre-event baseline period, the event occurrence period, and the post-event response period. The pre-event baseline period can be a valid observation period of 1 to 7 consecutive days before the event. When there is insufficient valid observation data before the event or there is significant disturbance before the event, valid observation data from the same historical period of the same crop growth stage and with matching conditions in the same plot are used as supplementary baseline data.

[0043] The post-event response period does not use a uniform fixed calendar window, but is determined by the degree of regression of the target greenhouse gas flux relative to the baseline flux in the same state, the number of samplings that continuously meet the regression conditions, and the preset maximum observation period.

[0044] Among them, the same-state flux baseline refers to the greenhouse gas flux state that is not affected by the target operation event under the same or similar plots, the same crop growth stage, similar soil moisture state, similar soil temperature state, similar inorganic nitrogen state, and similar meteorological conditions.

[0045] For any target gas among carbon dioxide, methane, and nitrous oxide, the post-event response period for that target gas is considered to have ended when the deviation of its flux from the baseline flux under the same conditions falls below a preset regression threshold multiple times consecutively. If the regression condition is not met within a preset maximum observation period, the end of the preset maximum observation period is taken as the end of the response period for that target gas.

[0046] When the same event involves carbon dioxide, methane, and nitrous oxide, the overall response period endpoint of the event accounting unit can be the latest time among the response period endpoints of the three gases; or, for different event categories, the response period endpoint of a preset priority gas can be selected as the overall response period endpoint.

[0047] When a new farmland operation event occurs within the response period of the current event, and the new farmland operation event causes changes in target gas flux, soil moisture, soil inorganic nitrogen, redox potential, or other key state variables to exceed a preset disturbance threshold, perform any of the following actions: truncate the response period of the current event to before the occurrence of the new farmland operation event, and establish a new event accounting unit for the new farmland operation event as a new independent event; when the operation scope and response time window of the current event and the new farmland operation event overlap and their respective greenhouse gas responses cannot be distinguished, merge the two into a composite event, and reconstruct the event accounting unit and event response time window.

[0048] The preset regression threshold is determined based on the daily variation of the target gas flux and the observation noise level. Specifically, the regression threshold is set to 1.5σ to 3σ, using the standard deviation σ of the target gas flux during the baseline period before the event as a benchmark. The term "consecutive multiple times" refers to at least three consecutive valid observations. If the regression conditions are not met within the preset maximum observation period (e.g., 14 to 30 days after the event), the end of the preset maximum observation period is taken as the end of the response period, and "response period cutoff" and the corresponding uncertainty level are marked in the accounting ledger.

[0049] Step 3: Data acquisition, quality control, and unified status representation through air-space-ground collaboration.

[0050] For each event accounting unit, space-based observation data, airborne observation data, ground-based observation data, and operational activity data are acquired within the corresponding dynamic response time window, ultimately forming multi-source data.

[0051] Space-based observation data includes one or more of the following: multi-temporal satellite multispectral data, hyperspectral data, thermal infrared data, and synthetic aperture radar data. These data are used to obtain information on crop canopy status, vegetation index, surface temperature, soil moisture, water accumulation, bare soil ratio, plot changes, or changes in surface condition before and after operations.

[0052] Airborne observation data includes high-resolution observation data acquired by UAVs equipped with visible light, multispectral, hyperspectral, thermal infrared, or lidar sensors. This data is used to identify operational coverage boundaries, unoperated control areas, canopy changes, straw coverage, bare soil ratio, water accumulation status, surface temperature, and surface heterogeneity.

[0053] Ground-based observation data include one or more of the following: carbon dioxide, methane, and nitrous oxide flux data; soil temperature, soil moisture content, redox potential, soil inorganic nitrogen content, soil organic carbon content, soil texture, soil bulk density, meteorological data, and crop biomass data.

[0054] Operational activity data includes one or more of the following: agricultural machinery operation trajectory, machinery status, type and amount of inputs, fertilization depth, tillage depth, irrigation volume, drainage volume, straw treatment method, fuel consumption, energy consumption, and manual farming logs.

[0055] The data undergoes quality control processing, including: radiometric correction, atmospheric correction, geometric correction, cloud and shadow removal, and outlier pixel removal for satellite remote sensing data; attitude correction, ground control point correction, image stitching, spatial registration, image sharpness screening, and outlier flight data removal for UAV remote sensing data; instrument drift checking, outlier removal, data integrity checking, and timestamp correction for ground flux data; sensor fault value removal, duplicate record removal, missing record marking, and time synchronization processing for soil, meteorological, and operational record data; when there is a conflict between manual agricultural logs and agricultural machinery records, the agricultural machinery records with timestamps and location information are given priority; when equipment records are missing, manual agricultural logs are used and assigned a lower data quality level.

[0056] The multi-source data that has undergone quality control is subjected to coordinate benchmark unification, time synchronization, spatial clipping, event association, and state variable extraction to form a state vector corresponding to the event accounting unit.

[0057] The state vector includes at least crop state variables, soil state variables, meteorological state variables, water and irrigation / drainage state variables, nitrogen state variables, operational activity state variables, and greenhouse gas flux state variables.

[0058] Among them, crop state variables include one or more of the following: crop growth stage, vegetation index, leaf area index, canopy temperature, canopy coverage, and crop biomass; soil state variables include one or more of the following: soil temperature, soil moisture content, soil organic carbon, soil texture, soil bulk density, and redox potential; nitrogen state variables include one or more of the following: nitrogen fertilizer type, nitrogen application rate, soil inorganic nitrogen content, crop nitrogen uptake, ammonia volatilization loss, and nitrogen leaching loss.

[0059] Step 4: State consistency screening and construction of flux baselines in the same state.

[0060] Satellite data, UAV data, ground observation data, and operational activity data within the same event accounting unit are screened for consistency based on crop growth stage, soil moisture level, soil temperature level, inorganic nitrogen status, duration after precipitation or irrigation / drainage, canopy coverage, duration after operation, and meteorological conditions.

[0061] The state consistency screening includes at least the following rules: when the crop growth stage of the candidate data and the target event data are inconsistent, it is determined to be inconsistent; when the candidate data and the target event data are at different soil moisture levels, it is determined to be inconsistent; for nitrogen application events, when the soil inorganic nitrogen state of the candidate data and the target event data crosses a preset inorganic nitrogen level, it is determined to be inconsistent; for irrigation, precipitation, or drainage events, when the candidate data and the target event data are in different irrigation / drainage duration intervals, waterlogging state intervals, or redox state intervals, it is determined to be inconsistent; when the difference between the soil temperature, canopy coverage, or meteorological state of the candidate data and the target event data exceeds the corresponding preset threshold, it is determined to be inconsistent.

[0062] State distance can be determined by one or more of the following factors according to preset weights: differences in crop growth stage, soil moisture, soil temperature, inorganic nitrogen status, duration of precipitation or irrigation / drainage, canopy cover, and meteorological status. Only candidate data with state distances less than the preset state distance threshold are used as joint accounting samples and baseline candidate samples to avoid directly mixing data from the same plot or the same spatial location under different farmland management states.

[0063] The same-state flux baseline is used to determine whether the greenhouse gas response period has ended after an operational event; to calculate the increase in greenhouse gas fluxes during the event response period; and to determine the degree of disturbance of the operational event to natural background fluxes. The same-state flux baseline is determined according to the following priorities: valid observational data from the target plot during the baseline period before the event, with the same crop growth stage and consistent state at the target time; valid observational data from the target plot during the same historical period, with the same crop growth stage and consistent state; valid observational data from adjacent non-operational control areas, corrected for crop growth stage, soil condition, and meteorological condition; and simulation results from process models that meet the constraints of carbon conservation, nitrogen balance, water condition, and operational event response. When the number of valid samples under the current priority is insufficient, the next lower priority is used to determine the same-state flux baseline.

[0064] Pre-event baseline data can only be used when the response period is shorter than a certain threshold (e.g., 3 days); a "state distance weighting" method is added to allow the construction of a baseline through weighted regression when the states are not completely consistent; when using process model simulation results as a baseline, the model must be validated in advance with data independent of the current event, and the model version and validation results must be recorded in the ledger.

[0065] Step 5: Observability evaluation and directional supplementary measurements based on uncertainty contribution.

[0066] The observability score of the event accounting unit is calculated based on the effective coverage, temporal proximity, data quality, spatial representativeness, and state consistency of various data sources during the event response period.

[0067] The observability score is calculated as follows: First, the scores of the event accounting unit in the five dimensions of effective data coverage, data temporal proximity, data quality level, spatial representativeness and state consistency during the response period are evaluated respectively, and each score is normalized to the range of zero to one; then, each score is multiplied by its corresponding event category adaptive weight and the results are added together. The weighted sum is the observability score; when the score is lower than 0.6, targeted supplementary testing is triggered.

[0068] Data source weights are determined based on event category, target greenhouse gas type, and the data source's sensitivity to target greenhouse gas estimation. For nitrogen application events, the weights of soil inorganic nitrogen, nitrous oxide flux, soil moisture content, and soil temperature data are increased. For irrigation, precipitation, or drainage events, the weights of soil moisture content, waterlogging status, redox potential, methane flux, and nitrous oxide flux data are increased. For straw return or tillage events, the weights of straw input, tillage depth, bare soil percentage, carbon dioxide flux, and soil temperature data are increased.

[0069] When the observability score is less than the preset observability threshold, the missing key observations in the event accounting unit are identified. Missing key observations include one or more of the following: operation coverage boundary, observations in the unoperated control area, carbon dioxide flux, methane flux, nitrous oxide flux, soil inorganic nitrogen content, soil moisture content, redox potential, crop biomass, irrigation and drainage volume, input batches, and agricultural machinery operation trajectories.

[0070] For each missing key observation, its contribution to the uncertainty of the event's net greenhouse gas balance is determined by at least one of the following methods: Univariate upper and lower limit perturbation method: Under the condition of keeping other state variables unchanged, the upper and lower limits of the uncertainty range of the missing observations are used to participate in the calculation of the net greenhouse gas balance of the event, and the contribution value is determined by the magnitude of the change in balance. Local sensitivity analysis method: Change the value of the missing observation and calculate the sensitivity of the event's net greenhouse gas budget to the change of the missing observation; Conditional Monte Carlo sampling method: Perform multiple samplings within the conditional probability distribution range of the missing observation and calculate the variance contribution of the missing observation to the event's net greenhouse gas budget; Ensemble simulation or variance decomposition method: Determine the contribution value of each missing observation based on the changes in the event's net greenhouse gas budget caused by different combinations of missing observations.

[0071] Based on contribution value, timeliness of retesting, and cost of retesting, candidate retesting tasks are prioritized, and retesting tasks with contribution values ​​higher than the preset contribution threshold and capable of being completed within the event response period are executed first.

[0072] The targeted supplementary monitoring tasks include at least one of the following: conducting drone supplementary flights at the boundary of the operation coverage, the center of the plot, and adjacent unoperated control areas; increasing the sampling frequency of ground fluxes for carbon dioxide, methane, or nitrous oxide during the estimated peak greenhouse gas emission period; increasing sampling of soil inorganic nitrogen, soil moisture content, soil temperature, or redox potential; supplementing the recording of agricultural machinery operation trajectories, machinery working status, input batches, fertilization depth, tillage depth, irrigation and drainage records, fuel consumption, or energy consumption; and supplementing the identification of operation boundaries, straw coverage, water accumulation range, bare soil ratio, or unoperated control areas.

[0073] The triggering basis for targeted supplementary measurements is the observation gap within the event response time window and its contribution to the uncertainty of the event's net greenhouse gas balance, rather than being determined based on the remote sensing pixel alignment window, remote sensing-carbon cycle pixel scale bias, or spatial resolution adjustment.

[0074] The observability assessment is divided into two stages: real-time assessment and post-event assessment. An initial observability assessment is conducted immediately after an event occurs (e.g., within 24 hours). If key data is found to be missing at this time and the response period has not yet ended, supplementary testing is triggered immediately. For data gaps discovered after the response period has ended, the supplementary testing data is not used for the current event accounting, but for model calibration and baseline construction for subsequent events, and is marked in the ledger as "supplementary testing data is used for model updates". A "predictive supplementary testing" mechanism is added at the system level - based on the data gap patterns of similar historical events, supplementary testing resources are pre-deployed before or during the event.

[0075] Step Six: Preliminary calculation of the balance of payments for multiple greenhouse gas events and determination of accounting boundaries.

[0076] Based on the event state vector, process model estimation results, ground observation results, and flux estimation results obtained based on remote sensing-assisted state variables were obtained for carbon dioxide, methane, and nitrous oxide.

[0077] Among them, satellite remote sensing and UAV remote sensing data are used to obtain auxiliary state variables such as crop canopy status, surface temperature, soil moisture, water accumulation status, bare soil ratio, operation coverage, straw coverage, and surface heterogeneity; the auxiliary state variables are input into the process model, flux estimation model, or subsequent joint correction model to generate or correct greenhouse gas flux estimation results.

[0078] For the target gas in the event accounting unit, calculate the gas emission increment relative to the baseline flux under the same conditions during the event response period. The gas emission increment is determined by the cumulative difference between the target gas flux and the baseline flux under the same conditions during the event response period, and selectively includes emissions converted from input production, agricultural machinery fuel consumption, or energy consumption according to the accounting boundary.

[0079] Based on the preset accounting boundaries, one or more of the following are included in the event emissions: emissions from input production, agricultural machinery fuel, and energy consumption. When the accounting boundaries do not include the corresponding emission sources, the corresponding equivalent is zero.

[0080] Carbon dioxide, methane, and nitrous oxide are converted into carbon dioxide equivalents according to the global warming potential parameters applicable at the time of calculation, thus obtaining the total greenhouse gas emissions of the event.

[0081] Event-based carbon sequestration accounting model: Event-based carbon sequestration is determined using one of the following two models.

[0082] Mode 1: Flux-driven accounting mode.

[0083] The integral result of net ecosystem carbon exchange within the event response time window is used as the main basis for the event's carbon sequestration.

[0084] In this model, crop biomass carbon increment, soil organic carbon stock change, and straw or other organic matter input and output are only used to verify the net ecosystem carbon exchange integral results and are not directly superimposed with the net ecosystem carbon exchange.

[0085] Mode 2: Inventory Change Accounting Mode.

[0086] When continuous net ecosystem carbon exchange observation data are lacking, the amount of carbon sequestration by events is determined based on crop biomass carbon increment, changes in soil organic carbon stock, and inputs and outputs of straw or other organic matter.

[0087] In this model, duplicate carbon flows between crop biomass carbon increment, soil organic carbon stock change, organic carbon input, and organic carbon output are deducted to avoid the same carbon flow being repeatedly measured in crop carbon pool, soil carbon pool, and organic input and output.

[0088] The event carbon sequestration amount, after accounting boundary consistency verification and duplicate measurement elimination, is obtained through any of the above modes. The event net greenhouse gas balance is obtained by subtracting the event carbon sequestration amount from the total event greenhouse gas emissions. When the calculation result is positive, it indicates that the event accounting unit is a net greenhouse gas emitter; when the calculation result is negative, it indicates that the event accounting unit is a net greenhouse gas absorber or net carbon sequestrator.

[0089] The process model selected uses published farmland carbon and nitrogen cycle process models (such as the DNDC model or the DayCent model), or a mechanistic model developed independently after localization parameter calibration. The model includes at least the following core computational frameworks: a crop growth sub-model for calculating crop photosynthetic carbon fixation, biomass accumulation, and phenological development; a soil organic carbon decomposition sub-model for calculating the mineralization and decomposition rate of organic carbon under aerobic and anaerobic conditions; a nitrogen transformation sub-model for calculating the rates of organic nitrogen mineralization, nitrification, denitrification, and ammonia volatilization; and a methane production and oxidation sub-model for calculating the activity intensity of methanogenic bacteria and the consumption rate of methanogenic bacteria under flooded conditions. The model uses soil temperature, soil moisture content, initial soil organic carbon, initial soil inorganic nitrogen, crop type and growth stage, meteorological data, and farmland management practices as input variables, and outputs daily or hourly estimates of carbon dioxide, methane, and nitrous oxide fluxes, as well as the dynamic changes in crop biomass, soil organic carbon pool, and inorganic nitrogen pool.

[0090] The fusion weights are determined based on the uncertainties of each data source: First, the uncertainty of each data source is comprehensively evaluated based on the data quality markers output in step three, the observability scores calculated in step five, and the deviations between the observed values ​​and historical verification data. The smaller the uncertainty of a data source, the greater its weight is allocated in the fusion. Specifically, the reciprocal of the square of the uncertainty of each data source is used as the base value for weight calculation. After summing the base values ​​of all data sources, the ratio of the base value of each data source to the total is used as the final fusion weight of that data source. For nitrogen application events, the weights of ground-based nitrous oxide flux and soil inorganic nitrogen data are appropriately increased; for irrigation or drainage events, the weights of ground-based methane flux and soil moisture data are appropriately increased; for remote sensing inversion data, when ground-based observation data is missing, space-based or air-based observation data bear the main weight, but the weights need to be reduced according to their data quality level.

[0091] The constrained projection process follows these steps: First, the candidate event response states, obtained by weighted fusion of process model estimation results, ground observation results, and remote sensing-aided estimation results, are input one by one into carbon conservation constraints, nitrogen balance constraints, water state constraints, and operational event response constraints for verification to determine whether the candidate state is within the feasible region defined by each constraint. If the candidate state satisfies all constraints, it is directly output as the final event response state. If the candidate state does not satisfy any constraint, a numerical optimization algorithm is used to gradually adjust the candidate state to the feasible region boundary or within the feasible region, with the goal of minimizing the weighted distance between the candidate state and the feasible region boundary. During the adjustment process, the state variables corresponding to high-weight data sources are kept unchanged, and only the state variables corresponding to low-weight data sources are adjusted until all constraints are satisfied. If convergence to the feasible region is still not possible after multiple iterations, the adjusted state closest to the feasible region boundary is output as the final event response state, and "constraints not fully satisfied" and the corresponding uncertainty amplification coefficient are marked in the event-level carbon budget ledger.

[0092] Step 7: Multi-source joint correction based on the feasible region of process constraints.

[0093] Construct a process constraint set that includes at least carbon conservation constraints, nitrogen balance constraints, moisture state constraints, and operational event response constraints, and determine the feasible domain of the event response state based on the process constraint set.

[0094] (1) Carbon conservation constraint.

[0095] The carbon conservation constraint is used to define the boundary relationships between crop carbon accumulation, soil organic carbon change, carbon dioxide exchange, organic carbon input, and organic carbon output.

[0096] For straw return to the field, tillage, or straw return to the field-tillage composite events, within the preset accounting boundary, the difference between the sum of straw carbon input, crop residue carbon input, and other organic carbon input, and the sum of cumulative carbon dioxide release, crop carbon storage change, soil organic carbon inventory change, undecomposed organic carbon residue, and organic carbon output should be within the preset carbon balance tolerance range.

[0097] (2) Nitrogen balance constraint.

[0098] The nitrogen balance constraint is used to define the boundary relationships between nitrogen fertilizer input, changes in soil inorganic nitrogen stock, crop nitrogen uptake, nitrogen leaching, ammonia volatilization, and nitrous oxide emissions.

[0099] For nitrogen application events, within the preset accounting boundaries, the difference between the sum of the amount of nitrogen applied, the soil inorganic nitrogen inventory before the event, the amount of nitrogen input from straw or organic fertilizer, and the amount of nitrogen input from irrigation water, and the sum of the changes in crop nitrogen uptake, the soil inorganic nitrogen inventory after the event, the amount of nitrogen emitted by nitrous oxide, the amount of nitrogen volatilized by ammonia, the amount of nitrogen lost through leaching, and other nitrogen outputs should be within the preset nitrogen balance tolerance range.

[0100] (3) Moisture state constraints.

[0101] The moisture state constraints are used to define the boundary relationships between irrigation, drainage, precipitation, evapotranspiration, field water level, soil moisture content, waterlogging state, redox potential, and methane and nitrous oxide emission responses.

[0102] For irrigation, precipitation, or drainage events, when the soil moisture content does not reach the preset saturation threshold for the corresponding soil texture and there is no water accumulation or low redox potential, the combined corrected methane flux is limited to not exceeding a preset multiple of the baseline flux for the corresponding state.

[0103] When the soil moisture content is in the preset high moisture content range, forms a waterlogged state, or the redox potential is lower than the preset threshold, the changes in methane and nitrous oxide fluxes after joint correction should be consistent with the trends of soil moisture content, waterlogged state, or redox potential.

[0104] (4) Operation event response constraints.

[0105] The operational event response constraints include at least one of the following: For nitrogen application events, the jointly corrected nitrous oxide flux state should be consistent with the soil inorganic nitrogen content, soil moisture content, and soil temperature state; for irrigation, precipitation, or drainage events, the jointly corrected methane and nitrous oxide flux states should be consistent with the soil moisture content, waterlogging state, drainage state, or redox potential state; for tillage or straw return events, the jointly corrected carbon dioxide flux state should be consistent with the bare soil ratio, soil temperature, tillage depth, straw input amount, or organic matter input state; the gas flux difference between the operational coverage area and the unoperated control area should not exceed the preset boundary allowed by differences in crop growth stage, soil state, and meteorological state; for composite events, the gas response time series corresponding to each sub-event should be consistent with the operation occurrence sequence, operation time interval, and state variable changes.

[0106] Based on observability scores, data quality, and uncertainties of each data source, adaptive fusion weights are determined for the process model estimation results, ground observation results, and remote sensing-aided estimation results.

[0107] For a multi-gas state to be corrected, first obtain a fusion candidate state, then restrict or project the fusion candidate state to a feasible region that satisfies the process constraint set to obtain the final event response state.

[0108] Optionally, a regression model, a residual prediction model, or a machine learning model can be used to generate candidate error correction values; the candidate error correction values ​​are only used as initial or candidate values ​​for the fusion candidate states, and the final event response state must satisfy the process constraint set.

[0109] The final event response status obtained after joint correction in step seven is used to correct the total event greenhouse gas emissions and event carbon sequestration in step six, resulting in the final event net greenhouse gas balance. If the correction exceeds a preset threshold (e.g., 20%), it will be marked as "significant correction" and the corresponding reason for the correction in the event-level carbon balance ledger. The supplementary measurement task triggered based on the preliminary calculation results in step five is not affected by subsequent corrections, and the supplementary measurement data will be used simultaneously in both the preliminary calculation and joint correction stages.

[0110] Step 8: Attribution of contributions based on the event substitution counterfactual management baseline.

[0111] For the same event accounting unit, an actual operational event scenario and a counterfactual management baseline scenario are constructed. The counterfactual management baseline scenario is used to evaluate the changes in greenhouse gas balance caused by the target operational event or target management measures, and is different from the same-state flux baseline used to determine the increase in gas flux.

[0112] The counterfactual management baseline scenario keeps the plot, crop type, crop growth stage, initial soil condition, meteorological conditions, non-target state variables, and non-target operation activities corresponding to the event accounting unit unchanged, and only replaces the target farmland operation event.

[0113] The counterfactual management baseline scenario can replace the target farmland operation event with a preset control management measure, a non-target operation measure, a historical low emission management measure, or other management measures to be compared.

[0114] For fertilization events, the counterfactual management baseline scenario can be reduced nitrogen application, multiple fertilizations, slow-release fertilizer substitution, different fertilization depths, no nitrogen application, or other fertilization methods.

[0115] For irrigation events, the counterfactual management baseline scenario can be a scenario of reducing irrigation volume, replacing flood irrigation with drip irrigation, adjusting irrigation timing, not implementing targeted irrigation, or other irrigation methods.

[0116] For straw return to the field and tillage events, the counterfactual management baseline scenario can be a scenario of straw removal from the field, no-tillage, reduced tillage, different amounts of straw returned to the field, or different methods of straw return to the field.

[0117] The counterfactual management baseline is determined in the following order: state matching samples of the same plot during the same historical period and at the same crop growth stage, where no target operation event has occurred; state matching samples of adjacent non-operation control areas; target operation event scenarios formed by extending the pre-event state of the same plot; and counterfactual simulation scenarios of process models that meet the constraints of carbon conservation, nitrogen balance, water status, and operation event response.

[0118] Net greenhouse gas budgets were calculated under both the actual operational event scenario and the counterfactual management baseline scenario to obtain the event-level carbon budget contribution of the target management measures.

[0119] When the net greenhouse gas balance under the actual operating event scenario is higher than that under the counterfactual management baseline scenario, it indicates that the actual target management measures have increased net greenhouse gas emissions relative to the counterfactual management baseline. When the net greenhouse gas balance under the actual operating event scenario is lower than that under the counterfactual management baseline scenario, it indicates that the actual target management measures have contributed to emission reduction or carbon sequestration relative to the counterfactual management baseline.

[0120] The contribution results can be output separately for carbon dioxide, methane, and nitrous oxide, and the total carbon dioxide equivalent contribution and uncertainty range can be output simultaneously.

[0121] By maintaining consistency between non-targeted operational activities and environmental conditions, and only replacing target operational events, we can reduce confounding errors caused by different plots, different crop growth stages, different hydrothermal conditions, and different nitrogen states, thereby improving the accuracy of estimates of changes in carbon dioxide, methane, and nitrous oxide fluxes and net greenhouse gas budgets caused by target operational events.

[0122] Step Nine: Event-level carbon balance ledger, quality labeling, and visualization output.

[0123] An event-level carbon budget ledger entry is generated for each event accounting unit. This event-level carbon budget ledger entry includes: event identifier, plot identifier, crop type and crop growth stage, event category, whether it is a composite event, start and end times of the operation, operation intensity, operation coverage, event response time window, source of flux baseline in the same state, source of counterfactual management baseline, accounting boundary type, source of original data, data quality marker, observability score, targeted supplementary monitoring task, supplementary monitoring priority and results, separate accounting results for carbon dioxide, methane, and nitrous oxide, total carbon dioxide equivalent, event carbon sequestration, event net greenhouse gas balance, carbon conservation, nitrogen balance, and water state constraint verification results, actual event scenario, counterfactual management baseline scenario, contribution of management measures, data version number, model version number, source of uncertainty, and uncertainty range or uncertainty level.

[0124] The event-level carbon balance sheet is summarized according to four levels: event, plot, crop season, and year, generating a greenhouse gas balance sheet time series diagram, operational event response diagram, plot-level spatial distribution map, management measure contribution comparison diagram, and standardized accounting report.

[0125] This application addresses event overlap and uncertainty handling: When the response time windows of two or more farmland operation events overlap, the system establishes event correlation relationships based on the operation occurrence time, operation space range, operation event category, state variable change magnitude, and degree of mutual influence between events.

[0126] For events that can be distinguished independently, carbon balance is calculated according to their respective dynamic response time windows.

[0127] For events that cannot be distinguished independently, they are merged into composite event accounting units, and the composition of the composite event, the time of occurrence of each sub-event, the scope of operation, and the intensity of operation are recorded in the event-level carbon income and expenditure ledger.

[0128] The system uses a combination of data source uncertainty, parameter uncertainty, model uncertainty, and counterfactual scenario uncertainty to assess the uncertainty range of the event’s net greenhouse gas balance.

[0129] Specifically: data source uncertainty stems from quality labels and observability scores of satellite remote sensing, UAV remote sensing, ground observation, and operational activity data; parameter uncertainty arises from soil parameters, crop parameters, gas transformation parameters, baseline thresholds, and process constraint tolerance ranges; model uncertainty arises from differences between process model estimates, ground observation estimates, and remote sensing-aided estimates; and scenario uncertainty arises from different reasonable settings of counterfactual management baselines. The uncertainty range and uncertainty level of the event's net greenhouse gas balance can be output through Monte Carlo sampling, ensemble simulation, interval propagation, local sensitivity analysis, or confidence interval propagation.

[0130] This application also provides a space-air-ground integrated dynamic accounting system for multiple greenhouse gas carbon budgets, based on the above-mentioned space-air-ground integrated dynamic accounting method for multiple greenhouse gas carbon budgets, including: The operation event recognition module is used to receive data such as plot boundaries, crop types, crop growth stages, agricultural machinery operation trajectories, agricultural machinery working status, agricultural logs, input records, irrigation and drainage records, and fuel or energy consumption records, to identify farmland operation events and generate event description vectors. The event accounting unit construction module is used to construct event accounting units based on operation events, related plots or operation coverage sub-areas, crop growth stages, operation intensity, operation impact range, initial soil-crop-meteorological state, and dynamic response time windows, and to perform independent accounting, truncation processing, or composite event merging processing on overlapping events. The air-space-ground integrated observation module is used to acquire satellite remote sensing data, UAV remote sensing data, ground flux observation data, soil observation data, meteorological observation data, and operational activity data; The unified state representation and quality control module is used to unify coordinate benchmarks, synchronize time, remove outliers, control quality, prune space, associate events, and construct state vectors for multi-source data, and output the event state vectors corresponding to the event accounting units. The module for constructing state consistency and same-state flux baselines is used to screen candidate observation data for state consistency based on crop growth stage, soil moisture level, soil temperature level, inorganic nitrogen status, duration after irrigation and drainage, meteorological conditions and duration after operation, and construct same-state flux baselines according to preset priorities. The observability assessment and targeted supplementary measurement module is used to calculate the observability score of the event accounting unit; when the observability score is insufficient, it determines the contribution value of the missing observations to the uncertainty of the event's net greenhouse gas balance sheet, and generates targeted supplementary measurement tasks based on the contribution value, the timeliness of supplementary measurement, and the cost of supplementary measurement. The multi-greenhouse gas process accounting module is used to obtain process model estimation results, ground observation results and remote sensing-aided estimation results for carbon dioxide, methane and nitrous oxide based on event state vectors, and to calculate the target gas emission increment, input-related equivalent emissions, event carbon sequestration and event net greenhouse gas balance. The process constraint joint correction module is used to receive the event accounting unit, event state vector, same-state flux baseline, observability score and candidate flux state output by the multi-greenhouse gas process accounting module. It determines the fusion weight based on the data source uncertainty, observability score and data quality, and restricts the fusion candidate state to the feasible region that satisfies the constraints of carbon conservation, nitrogen balance, moisture state and operational event response. It outputs the jointly corrected event response state. The counterfactual attribution module is used to construct actual operational event scenarios and counterfactual management baseline scenarios for the same event accounting unit, keeping the plot, crop growth stage, initial soil condition, meteorological conditions and non-target operational activities consistent, only replacing the target operational events, and calculating the event-level carbon budget contribution of management measures; The event-level ledger and reporting module is used to generate auditable event-level carbon budget ledgers and output multi-level greenhouse gas carbon budget results, quality grades, management action contribution results, maps, and standardized accounting reports at the event level, plot level, crop season level, and annual level.

[0131] This invention uses specific farmland operation events as accounting units, determines the dynamic response window based on the regression process of the target gas flux relative to the baseline flux in the same state, and triggers targeted supplementary measurements based on the contribution of missing observations during the event response period to the uncertainty of the event's net greenhouse gas balance. Simultaneously, it performs joint correction by restricting the process model, ground observations, and remote sensing-aided estimation results to a feasible domain that satisfies carbon conservation, nitrogen balance, water status, and the response law of the operation event. Furthermore, while maintaining consistency in plot, growth stage, initial state, and non-target operation activities, it constructs a counterfactual management baseline by replacing only the target operation event, thereby achieving event-level dynamic accounting of carbon dioxide, methane, and nitrous oxide, attribution of management measures' contributions, and traceable ledger output.

[0132] To illustrate this application more clearly, the following embodiments will be used for explanation.

[0133] Example 1: Accounting for nitrogen application-irrigation complex events in winter wheat-summer maize rotation farmland.

[0134] Taking winter wheat-summer maize rotation farmland in the North China Plain as an example, select farmland plots with an area of ​​5 to 30 hectares and clear boundaries as target plots. Soil temperature, soil moisture content, and meteorological observation equipment are set up within the target plots; automatic flux observation devices, static box observation devices, or mobile flux observation devices are set up when conditions permit.

[0135] During the jointing stage of winter wheat, managers apply nitrogen to the target plots and then irrigate within a predetermined timeframe after the nitrogen application.

[0136] The system identifies nitrogen application-irrigation composite events by tracking the fertilizer applicator's positioning trajectory, operating status, controller records, irrigation equipment operation records, and agricultural logs, and obtains the following event information: Event Category: Nitrogen Application-Irrigation Composite Event; Plot identification; crop type; crop growth stage; fertilizer type and amount of pure nitrogen applied; fertilizer location and depth; Irrigation method and irrigation volume; start and end times of the event; operation trajectory and coverage area; agricultural machinery fuel consumption or irrigation energy consumption.

[0137] The system identifies nitrogen application-irrigation composite events based on at least two of the following: fertilization trajectory, fertilizer applicator working status, fertilizer application record, and irrigation record. It also determines the operational coverage area based on the operating range of the fertilizer applicator and irrigation facilities, the width of the equipment, and the positioning error buffer distance.

[0138] The baseline period before the event is selected as the effective observation period of 3 consecutive days before fertilization; when there are inconsistencies in the status due to rainfall, irrigation or other disturbances within 3 consecutive days, the observation data of the same plot at the same historical time, the same winter wheat growth stage and with matching status are used as the supplementary baseline.

[0139] In the nitrogen application-irrigation complex event, soil inorganic nitrogen content, soil moisture content, soil temperature, precipitation, irrigation amount, redox potential, and nitrous oxide flux were used as the main state variables.

[0140] The system acquires nitrous oxide flux more frequently after fertilization and irrigation. The nitrous oxide response period ends when multiple consecutive samples of the flux consistently meet the regression threshold relative to the baseline flux under the same conditions. The response periods for carbon dioxide and methane are determined according to their respective flux regression conditions, and the overall event response period endpoint is the latest of the three gas response period endpoints.

[0141] When a second nitrogen application occurs after the first nitrogen application, and the change in soil inorganic nitrogen concentration relative to the state after the first nitrogen application exceeds a preset disturbance threshold, the system will truncate the independent response period of the first nitrogen application event to before the time of the second nitrogen application. If the two nitrogen application operations overlap and their respective nitrous oxide emission responses cannot be distinguished, the two nitrogen application operations will be merged into a composite nitrogen application event, and the response period will be recalculated.

[0142] The system acquires satellite multispectral, thermal infrared, and synthetic aperture radar data to characterize crop canopy status, surface temperature, and soil moisture; it conducts UAV multispectral and thermal infrared observations at fertilizer coverage boundaries, plot centers, and adjacent untreated control areas to identify operational boundaries, differences in canopy water stress, waterlogging status, and surface heterogeneity; and it acquires carbon dioxide, methane, nitrous oxide fluxes, soil inorganic nitrogen, soil moisture content, soil temperature, and meteorological data through ground equipment.

[0143] When drone imagery is missing due to weather conditions, and the missing operational boundary or canopy moisture observations cause the uncertainty contribution of the event's net greenhouse gas balance to exceed a preset contribution threshold, the system triggers a drone re-flight mission.

[0144] When the ground flux sampling frequency is insufficient during periods of high nitrous oxide levels, and the uncertainty contribution of nitrous oxide flux is higher than that of other missing observations such as soil temperature or canopy status, the system will prioritize triggering a nitrous oxide flux intensified sampling task.

[0145] During the process constraint joint correction phase, the system establishes nitrogen balance constraints to ensure that the differences between applied nitrogen amount, pre-event soil inorganic nitrogen inventory, irrigation water nitrogen input and crop nitrogen uptake, post-event soil inorganic nitrogen inventory, nitrous oxide emissions, ammonia volatilization, and nitrogen leaching are less than the preset nitrogen balance tolerance; establishes water state constraints to ensure that changes in irrigation, precipitation, evapotranspiration, and soil moisture content are consistent; and establishes nitrogen application-irrigation event response constraints to ensure that changes in nitrous oxide flux are consistent with changes in soil inorganic nitrogen, soil moisture content, and soil temperature.

[0146] The system restricts the candidate states generated by process models, ground observations, and remote sensing-assisted estimations to the feasible domain that satisfies the above constraints, and obtains the carbon dioxide, methane, and nitrous oxide emissions, total carbon dioxide equivalent emissions, event carbon sequestration, and net greenhouse gas balance of the nitrogen application-irrigation composite event.

[0147] To analyze the contribution of this nitrogen application-irrigation measure, a counterfactual management baseline scenario was constructed: keeping the target plot, winter wheat jointing stage, initial soil conditions, meteorological conditions, and other non-target agricultural activities consistent, only replacing the actual nitrogen application-irrigation scheme with a reduced nitrogen application scheme, a multi-stage fertilization scheme, a drip irrigation instead of flood irrigation scheme, or a scheme that does not implement the target nitrogen application-irrigation.

[0148] By comparing the net greenhouse gas budget difference between the actual event scenario and the counterfactual management baseline scenario, we can obtain the carbon dioxide equivalent contribution of the nitrogen application-irrigation combined management measures and its uncertainty range.

[0149] Example 2: Accounting for the combined event of straw return to the field and tillage after summer corn harvest.

[0150] After the summer corn harvest, the target plots will be subjected to straw crushing and returning to the field, as well as rotary tillage.

[0151] The system identifies straw return-tillage composite events by tracking harvesting machinery trajectories, the status of straw crushing and returning equipment, the status of rotary tillage equipment, tillage depth records, and agricultural logs.

[0152] In response to this event, the system uses straw input amount, straw carbon-nitrogen ratio, straw coverage, uniformity of straw return to the field, tillage depth, soil temperature, soil moisture content, bare soil ratio, and post-operation duration as the main state variables.

[0153] UAV observations are used to identify straw coverage, operation boundaries, and the proportion of bare soil; satellite remote sensing data are used to obtain changes in surface conditions before and after the operation; ground observations are used to obtain data on carbon dioxide, methane, nitrous oxide flux, soil temperature, soil moisture, soil inorganic nitrogen, and soil organic carbon.

[0154] Under the carbon conservation constraint, the system considers straw carbon input, changes in soil organic carbon, carbon dioxide release, undecomposed straw carbon residue, and organic matter output; under the nitrogen balance constraint, the system considers straw nitrogen input, changes in soil inorganic nitrogen, and nitrous oxide emissions; under the water state constraint, the system considers the effects of precipitation, soil moisture content, and post-tillage soil aeration on carbon dioxide, methane, and nitrous oxide fluxes.

[0155] When the cumulative carbon dioxide emissions estimated by the process model exceed the allowable range of straw carbon input, soil organic carbon change, and organic matter output, the estimation results of the process model are restricted; when the straw coverage identified by the UAV is significantly inconsistent with the straw return to the field input by the process model, the corresponding data source weights are re-determined based on data quality and uncertainty.

[0156] The counterfactual management baseline scenario can be set as: a scenario where straw is removed from the field but not returned to the field under similar hydrothermal conditions on the same plot of land; No-till scenario with the same amount of straw returned to the field but no tillage; scenario with the same amount of straw returned to the field but with reduced tillage; scenario with different amounts of straw returned to the field or different straw treatment methods.

[0157] The system outputs the combined CO2 equivalent contribution of straw return-to-field tillage practices to short-term greenhouse gas emissions and medium-term soil carbon sequestration changes.

[0158] Example 3: Dynamic Accounting of Methane and Nitrous Oxide in Paddy Field Irrigation and Drainage Events Taking rice paddies with significant methane emissions as an example, operational events such as irrigation, drainage, and intermittent irrigation are identified during the rice tillering or heading stages.

[0159] The system acquires irrigation and drainage records, water pump operation records, field water level, soil moisture content, water accumulation status, redox potential, soil temperature, methane flux, nitrous oxide flux, and rice canopy status.

[0160] For irrigation events, when water accumulates in the field, the soil redox potential drops to a preset range, and the methane flux increases significantly relative to the baseline flux under the same conditions, it is determined to be the methane response phase; when the water accumulates and disappears after drainage, the redox potential rises again, and the methane flux continuously meets the regression threshold, the methane response period of the irrigation and drainage event is determined to have ended.

[0161] In this embodiment, soil moisture content reaching 80% or more of field capacity or the formation of identifiable waterlogging in the field can be used as an example of a high-moisture response state; redox potential below a preset redox threshold can be used as an example of a high-risk methane emission state. Specific thresholds can be set according to paddy field soil type, crop variety, and regional climate conditions.

[0162] When the contribution of missing methane flux observations to the uncertainty of the event’s net greenhouse gas balance is higher than that of other missing observations, the system prioritizes intensive sampling of ground methane flux; when the water accumulation area is not fully identified and its uncertainty contribution is high, the system triggers a UAV thermal infrared or visible light supplementary flight mission.

[0163] In the joint correction, irrigation and drainage, precipitation, evapotranspiration, field water level, soil moisture content, water accumulation status, and redox potential together constitute the water state constraints; the joint correction results of methane flux and nitrous oxide flux must be consistent with the changes in the water state.

[0164] Counterfactual management baseline scenarios can be set as continuous flooding, different drainage timings, different intermittent irrigation cycles, or no targeted drainage measures, in order to obtain the event-level contribution of irrigation and drainage management measures to the methane, nitrous oxide, and total carbon dioxide equivalent budget.

[0165] It should be noted that the above embodiments are only used to illustrate the present invention, but the present invention is not limited to the above embodiments. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A dynamic accounting method for multi-greenhouse gas carbon budget using a space-air-ground integrated approach, characterized in that, Includes the following steps: S1. Obtain relevant data on farmland operations, identify farmland operation events, and construct event accounting units based on farmland operation events; S2. Determine the event response time window. The event response time window includes the pre-event baseline period, the event occurrence period, and the post-event response period. The endpoint of the event response time window is determined based on the degree of regression of the target greenhouse gas flux relative to the baseline flux in the same state. S3. For the event accounting unit, acquire multi-source data within the corresponding event response time window. The multi-source data includes space-based observation data, air-based observation data, ground-based observation data, and operational activity data. After quality control, the multi-source data forms a state vector corresponding to the event accounting unit. S4. The acquired multi-source data are screened for consistency according to crop growth stage, soil condition, nitrogen status and meteorological conditions, and a flux baseline with the same status is constructed. S5. Calculate the observability score of the event accounting unit. When the observability score is less than the preset observability threshold, identify the missing key observations in the event accounting unit and trigger the targeted supplementary measurement task based on the contribution of the missing key observations to the uncertainty of the event's net greenhouse gas balance. S6. Based on multi-source data, obtain the estimated results of carbon dioxide, methane and nitrous oxide, calculate the total greenhouse gas emissions and carbon sequestration of the event, and subtract the carbon sequestration from the total greenhouse gas emissions to obtain the net greenhouse gas balance of the event.

2. The method for dynamic accounting of multi-greenhouse gas carbon budget through space-air-ground collaboration according to claim 1, characterized in that, In step S1, acquiring farmland operation-related data and identifying farmland operation events specifically includes: acquiring the plot boundaries, crop types, crop growth stages, agricultural machinery operation trajectories, agricultural machinery working status, agricultural logs, input records, irrigation and drainage records, and agricultural machinery fuel or energy consumption records of the target farmland area. Based on at least two types of information from agricultural machinery operation trajectories, agricultural machinery working status, input records, irrigation and drainage records, and agricultural logs, identify one or more agricultural operation events among tillage, sowing, fertilization, irrigation, drainage, straw return to the field, pest and disease control, harvesting, and agricultural machinery operations; The event accounting unit based on farmland operation events specifically includes: For each farmland operation event, an event accounting unit is constructed. The event accounting unit is determined by at least the associated plot or operation coverage sub-region, crop growth stage, operation event type and operation intensity, event response time window, spatial range of operation event impact, and initial soil-crop-meteorological state at the time of the event.

3. The method for dynamic accounting of multi-greenhouse gas carbon budget through space-air-ground collaboration according to claim 1, characterized in that, In step S2, the baseline period before the event is the effective observation period of 1 to 7 consecutive days before the event occurs; when there is insufficient effective observation data before the event or there is significant disturbance before the event, effective observation data of the same historical period, the same crop growth stage and matching status of the same plot are used as supplementary baseline data. When the same event involves carbon dioxide, methane and nitrous oxide, the overall response period endpoint of the event accounting unit is the latest time among the response period endpoints of the three gases; or, for different event categories, the response period endpoint of a preset priority gas is selected as the overall response period endpoint.

4. The method for dynamic accounting of multi-greenhouse gas carbon budget through space-air-ground collaboration according to claim 1, characterized in that, In step S3, after quality control, the multi-source data forms a state vector corresponding to the event accounting unit, specifically including: Perform radiometric correction, atmospheric correction, geometric correction, cloud and shadow removal, and anomalous pixel removal on space-based observation data; Attitude correction, ground control point correction, image stitching, spatial registration, image sharpness screening, and abnormal flight data removal are performed on the airborne observation data. Perform instrument drift checks, outlier removal, data integrity checks, and timestamp corrections on ground-based observation data; The data from operational activities are processed to remove sensor fault values, remove duplicate records, mark missing records, and synchronize time. The multi-source data that has undergone quality control is subjected to coordinate benchmark unification, time synchronization, spatial clipping, event association, and state variable extraction to form a state vector corresponding to the event accounting unit.

5. The method for dynamic accounting of multi-greenhouse gas carbon budget through space-air-ground collaboration according to claim 1, characterized in that, In step S4, the state consistency screening includes at least the following rules: When the crop growth stage of the candidate data is inconsistent with that of the target event data, it is determined that the states are inconsistent. When the candidate data and the target event data are at different soil moisture levels, they are judged to be inconsistent in state; For nitrogen application events, when the soil inorganic nitrogen status of candidate data and target event data crosses the preset inorganic nitrogen level, it is determined that the status is inconsistent; For irrigation, precipitation, or drainage events, when the candidate data and the target event data are in different irrigation / drainage duration intervals, water accumulation state intervals, or redox state intervals, they are judged to be inconsistent in state. When the difference between the soil temperature, canopy coverage, or meteorological conditions of the candidate data and the target event data exceeds the corresponding preset threshold, it is determined to be an inconsistency.

6. The method for dynamic accounting of multi-greenhouse gas carbon budget through space-air-ground collaboration according to claim 1, characterized in that, In step S5, observability evaluation is based on the effective coverage, temporal proximity, data quality, spatial representativeness, and state consistency of the data source during the response period. When the score is lower than a preset threshold, the missing key observations are identified, and the contribution of each missing key observation to the net greenhouse gas budget uncertainty of the event is determined by univariate perturbation, local sensitivity analysis, conditional Monte Carlo sampling, or variance decomposition. Based on the contribution value, the timeliness of retesting, and the cost of retesting, priority is given to retesting tasks that can be completed within the response period. Retesting tasks include UAV re-flight, ground flux densification sampling, soil parameter retesting, work record supplementation, and work boundary supplementation identification.

7. The method for dynamic accounting of multi-greenhouse gas carbon budget through space-air-ground collaboration according to claim 1, characterized in that, Event emissions are the cumulative difference between the target gas flux and the baseline flux under the same conditions during the event response period, and selectively include emissions from input production, agricultural machinery fuel, and energy consumption. Carbon sequestration is determined using either the flux-driven accounting model or the inventory change accounting model. Under the flux-driven model, the integral result of the net ecosystem carbon exchange within the event response time window is used as the event carbon sequestration. Under the inventory change accounting model, when continuous net ecosystem carbon exchange observation data are lacking, the carbon sequestration of events is based on crop biomass carbon increment, soil organic carbon inventory changes, and straw or other organic matter inputs.

8. A dynamic accounting method for multi-greenhouse gas carbon budget with space-air-ground coordination according to any one of claims 1-7, characterized in that, Also includes: S7. Construct a process constraint set including carbon conservation constraints, nitrogen balance constraints, water state constraints, and operational event response constraints. Determine the feasible region of the event response state based on the process constraint set. Determine the adaptive fusion weights of the process model estimation results, ground observation results, and remote sensing-aided estimation results based on observability scores, data quality, and uncertainties of each data source to obtain fusion candidate states. Then, restrict or project the fusion candidate states to the feasible region that satisfies the process constraint set to obtain the final event response state. S8. Construct actual operation event scenarios and counterfactual management baseline scenarios for the same event accounting unit. The counterfactual management baseline scenarios keep the plot, crop type, crop growth stage, initial soil condition, meteorological conditions, non-target state variables and non-target operation activities corresponding to the event accounting unit unchanged, and only replace the target farmland operation events. Calculate the net greenhouse gas balance under the actual operation event scenarios and the counterfactual management baseline scenarios respectively to obtain the event-level carbon balance contribution of the target management measures. S9. Generate event-level carbon income and expenditure ledger entries for each event accounting unit, and summarize and output the event-level carbon income and expenditure ledger entries according to four levels: event, plot, crop season, and year.

9. A dynamic accounting method for multi-greenhouse gas carbon budget with space-air-ground coordination according to claim 8, characterized in that, The carbon conservation constraint is used to limit the boundary relationships between crop carbon accumulation, soil organic carbon change, carbon dioxide exchange, organic carbon input, and organic carbon output. The nitrogen balance constraint is used to define the boundary relationships between nitrogen fertilizer input, changes in soil inorganic nitrogen stock, crop nitrogen uptake, nitrogen leaching, ammonia volatilization, and nitrous oxide emissions. The water state constraint is used to define the boundary relationships between irrigation amount, drainage amount, precipitation, evapotranspiration, field water level, soil moisture content, water accumulation state, redox potential and methane and nitrous oxide emission responses. The job event response constraints include at least one of the following: For nitrogen application events, the jointly corrected nitrous oxide flux status should be consistent with the soil inorganic nitrogen content, soil moisture content, and soil temperature status. For irrigation, precipitation, or drainage events, the combined corrected methane and nitrous oxide flux states should be consistent with soil moisture content, waterlogging status, drainage status, or redox potential status. For tillage or straw return events, the jointly corrected carbon dioxide flux status should be consistent with the bare soil ratio, soil temperature, tillage depth, straw input, or organic matter input status. The difference in gas flux between the treated area and the untreated control area shall not exceed the preset boundary allowed by differences in crop growth stage, soil condition and meteorological condition. For complex events, the gas response time series corresponding to each sub-event is consistent with the order of operation, the time interval between operations, and the changes in state variables.

10. A space-air-ground integrated dynamic accounting system for multiple greenhouse gas carbon budgets, based on the space-air-ground integrated dynamic accounting method for multiple greenhouse gas carbon budgets as described in claim 9, characterized in that, include: The operation event recognition module is used to receive data such as plot boundaries, crop types, crop growth stages, agricultural machinery operation trajectories, agricultural machinery working status, agricultural logs, input records, irrigation and drainage records, and fuel or energy consumption records, to identify farmland operation events and generate event description vectors. The event accounting unit construction module is used to construct event accounting units based on operation events, related plots or operation coverage sub-areas, crop growth stages, operation intensity, operation impact range, initial soil-crop-meteorological state, and dynamic response time windows, and to perform independent accounting, truncation processing, or composite event merging processing on overlapping events. The air-space-ground integrated observation module is used to acquire satellite remote sensing data, UAV remote sensing data, ground flux observation data, soil observation data, meteorological observation data, and operational activity data; The unified state representation and quality control module is used to unify coordinate benchmarks, synchronize time, remove outliers, control quality, prune space, associate events, and construct state vectors for multi-source data, and output the event state vectors corresponding to the event accounting units. The module for constructing state consistency and same-state flux baselines is used to screen candidate observation data for state consistency based on crop growth stage, soil moisture level, soil temperature level, inorganic nitrogen status, duration after irrigation and drainage, meteorological conditions and duration after operation, and construct same-state flux baselines according to preset priorities. The observability assessment and targeted supplementary measurement module is used to calculate the observability score of the event accounting unit; when the observability score is insufficient, it determines the contribution value of the missing observations to the uncertainty of the event's net greenhouse gas balance sheet, and generates targeted supplementary measurement tasks based on the contribution value, the timeliness of supplementary measurement, and the cost of supplementary measurement. The multi-greenhouse gas process accounting module is used to obtain process model estimation results, ground observation results and remote sensing-aided estimation results for carbon dioxide, methane and nitrous oxide based on event state vectors, and to calculate the target gas emission increment, input-related equivalent emissions, event carbon sequestration and event net greenhouse gas balance. The process constraint joint correction module is used to receive the event accounting unit, event state vector, same-state flux baseline, observability score and candidate flux state output by the multi-greenhouse gas process accounting module. It determines the fusion weight based on the data source uncertainty, observability score and data quality, and restricts the fusion candidate state to the feasible region that satisfies the constraints of carbon conservation, nitrogen balance, moisture state and operational event response. It outputs the jointly corrected event response state. The counterfactual attribution module is used to construct actual operational event scenarios and counterfactual management baseline scenarios for the same event accounting unit, keeping the plot, crop growth stage, initial soil condition, meteorological conditions and non-target operational activities consistent, only replacing the target operational events, and calculating the event-level carbon budget contribution of management measures; The event-level ledger and reporting module is used to generate auditable event-level carbon budget ledgers and output multi-level greenhouse gas carbon budget results, quality grades, management action contribution results, maps, and standardized accounting reports at the event level, plot level, crop season level, and annual level.

Images