A method for predicting the weekly methane emission flux of a rice field based on sensor monitoring data

Abstract

The application discloses a paddy field methane weekly emission flux prediction method based on sensor monitoring data, and relates to the field of agricultural greenhouse gas emission prediction, and comprises the following steps: obtaining paddy field basic records, field management records, sensor monitoring records and methane flux measured records to form a weekly event chain; locking a methane release window according to the weekly event chain and forming a methane release window sequence; calling window-level emission reference records to determine a methane weekly emission flux prediction result; playing back a candidate irrigation and drainage scheme to the weekly event chain to re-form a corresponding prediction result and determine a water-saving and emission-reducing irrigation and drainage scheme; and correcting release window identification basis, reference record calling mode and irrigation and drainage scheme playback results according to actual monitoring and measured results. The application realizes the associated processing of paddy field methane release window identification, weekly emission prediction and irrigation and drainage emission reduction control.

Description

Technical Field

[0001] This invention relates to the field of agricultural greenhouse gas emission prediction, and in particular to a method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data. Background Technology

[0002] Methane emissions from paddy fields are an important research subject in agricultural greenhouse gas accounting and emission reduction management. Current research typically combines static box sampling, automatic flux monitoring, water layer observation, soil moisture content, electrical conductivity, temperature, and field irrigation and drainage records to analyze methane flux changes under different growth stages, straw return methods, and water management conditions. This data is then used for emission inventory compilation, emission reduction effect evaluation, and farmland water regulation reference.

[0003] However, existing methods still have two limitations: First, methane flux prediction is mostly based on weekly average data, fixed sampling results, or single state parameters, which makes it difficult to correspond to the release period formed by the combined effects of irrigation and drainage, aeration, salinity changes, and straw carbon supply within the week. Second, there is insufficient connection between emission prediction results and water-saving irrigation and drainage schemes, making it difficult to reflect the impact of different irrigation and drainage arrangements on methane release windows, soil moisture boundaries, and salinity control conditions during the prediction stage. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data, which solves the problems of insufficient windowed characterization of the release process and difficulty in supporting irrigation and drainage emission reduction regulation in existing technologies.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data. The method includes: acquiring basic paddy field records, field management records, sensor monitoring records, and measured methane flux records; using a seven-day period as the methane emission prediction cycle; time-marking records related to moisture, salinity, temperature, straw return to the field, and irrigation / drainage to form a weekly event chain; locking methane release windows based on the weekly event chain; forming a methane release window sequence according to the window start and end times, trigger sources, termination sources, and window connection relationships; and calling the corresponding window-level emission reference record based on the window type, duration, status fields, and window connection relationships in the methane release window sequence, and determining the methane emission flux based on window truncation, window correction, and the relationship between the output to be verified. The methane weekly emission flux prediction results are used to generate candidate irrigation and drainage schemes based on the original planned irrigation and drainage schedule. These schemes are then replayed into the weekly event chain to reconstruct the methane release window sequence and methane weekly emission flux prediction results corresponding to the candidate schemes. Water-saving and emission-reducing irrigation and drainage schemes are determined by combining the release window changes, agronomic safety boundaries, salinity control conditions, and irrigation water consumption. After the methane emission prediction cycle ends, the actual weekly event chain and actual methane release window sequence are generated based on the actual methane flux measurement records, actual irrigation and drainage execution records, and complete sensor monitoring records. Based on the correspondence between the actual methane release window sequence and the predicted release window sequence, the release window identification criteria, the window-level emission reference record recall method, and the irrigation and drainage scheme replay results are corrected.

[0008] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data described in this invention, the formation of the weekly event chain includes: taking a paddy field plot or a paddy field zone under the same irrigation and drainage facility as the prediction object, binding soil volumetric water content sensor, soil conductivity sensor, water depth sensor, soil temperature sensor, irrigation inlet, drainage outlet, straw return record, and methane flux measurement record to the prediction object, and writing the sensor acquisition time, irrigation and drainage execution time, straw return time, and methane flux measurement time into the same continuous seven-day prediction cycle according to a unified time granularity.

[0009] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data described in this invention, the formation of the weekly event chain includes: generating candidate carbon supply events based on the amount of straw returned to the field, the date of straw return to the field, the depth of straw burial, and the range of straw decomposition time; generating flooding events, waterlogging without aeration events, and aeration events based on water layer depth, soil volumetric water content, saturated water content, field water holding capacity, and water layer identification boundaries; generating candidate salinity events based on conductivity correction values, salinization background values, irrigation and drainage events, and response time ranges; and generating sampling verification events based on sensor range, sampling time missing status, continuous missing measurement time, and the time correspondence between water layer depth and irrigation and drainage records.

[0010] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data according to the present invention, the step of locking the methane release window according to the weekly event chain and forming a methane release window sequence according to the window start and end time, trigger source, termination source, and window connection relationship includes: when locking the methane release window, locking the time period when the carbon supply candidate event coincides with the flooding event or the water-induced non-aeration event and the soil temperature reaches the effective temperature condition for methane production as the water-carbon superposition release window; locking the time period covered by the flooding event or the water-induced non-aeration event but without forming a carbon supply candidate event as the water maintenance release window; locking the time period formed by the salinity candidate event within the response time range after irrigation, drainage, rainfall replenishment, or field drying as the water-salt disturbance release window; locking the time period when the aeration event continues to reach the aeration judgment duration as the release reduction window; and locking the time period when the sampling and verification event covers the key monitoring fields as the verification release window.

[0011] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data described in this invention, the step of forming a methane release window sequence according to the window start and end times, trigger source, termination source, and window connection relationship includes: arranging each release window according to its start and end times; when a release reduction window is adjacent to a water-carbon superimposed release window or a water-maintaining release window, the start time of the release reduction window is used as the termination correction time of the adjacent release window; when a water-salt disturbance release window overlaps with a water-carbon superimposed release window or a water-maintaining release window, the water-salt disturbance release window is written into the salinity correction field of the corresponding release window; when the release window to be nulled covers water layer depth, soil volumetric water content, electrical conductivity correction value, or soil temperature, the corresponding release window is marked as a window to be nulled.

[0012] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data described in this invention, the step of calling the corresponding window-level emission reference record and determining the predicted methane weekly emission flux based on window truncation, window correction, and the relationship between the output to be verified includes: for water and carbon superposition release windows, calling the corresponding reference record according to the amount of straw returned to the field, the number of days after straw return, the duration of water status, and the soil temperature status field; for water maintenance release windows, calling the corresponding reference record according to the rice growth period, the duration of water status, and the type of irrigation and drainage events; for release windows with salinity correction fields, calling the corresponding reference record according to the direction of salinity change, the type of irrigation and drainage events, and the response time range; and for the window to be verified, calling the previous valid window of the same prediction object, the window of the same type of adjacent prediction object, and the reference window of the same growth period to form the methane flux prediction range.

[0013] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data according to the present invention, the step of generating candidate irrigation and drainage schemes according to the original planned irrigation and drainage schedule and replaying the candidate irrigation and drainage schemes into the weekly event chain includes: when replaying the candidate irrigation and drainage schemes into the weekly event chain, replacing or supplementing the original planned irrigation and drainage event fields according to the irrigation time, drainage time, water layer target value, water drop duration, and irrigation recovery time in the candidate irrigation and drainage schemes, regenerating the water layer event, aeration event, salinity candidate event, and methane release window sequence under the candidate scheme conditions, and redetermining the methane weekly emission flux prediction result corresponding to the candidate scheme according to the window-level emission reference record.

[0014] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data described in this invention, the step of determining a water-saving and emission-reducing irrigation and drainage scheme by combining changes in the release window, agronomic safety boundaries, salinity control conditions, and irrigation water volume includes: when determining a water-saving and emission-reducing irrigation and drainage scheme, verifying the methane release window sequence formed by candidate irrigation and drainage schemes; if the duration of the water-carbon superposition release window or the water-maintained release window of a candidate irrigation and drainage scheme is less than the duration of the originally planned corresponding window, and it does not exceed the minimum production safety moisture content control boundary, does not form a salinity upward shift judgment record, and meets the allowable operation record during the rice growth period, the candidate irrigation and drainage scheme is retained; if the predicted weekly methane emission flux corresponding to the candidate irrigation and drainage scheme is less than the predicted result of the originally planned scheme, and the irrigation water volume is less than the originally planned irrigation water volume, it is determined as a water-saving and emission-reducing irrigation and drainage scheme.

[0015] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data according to the present invention, the step of generating the actual weekly event chain and the actual methane release window sequence based on the actual methane flux measurement record, the actual irrigation and drainage execution record, and the complete sensor monitoring record includes: when generating the actual weekly event chain after the prediction period ends, the actual methane flux measurement record is assigned to the corresponding time segment according to the sampling time or flux monitoring time, the actual irrigation and drainage execution record is assigned to the corresponding time segment according to the event start time and the event end time, the complete sensor monitoring record is assigned to the corresponding time segment according to a uniform time granularity, and the actual methane release window sequence is re-formed based on the actual carbon supply candidate event, the actual flooding event, the actual water-induced non-aeration event, the actual aeration event, the actual salinity candidate event, and the actual sampling pending event.

[0016] As a preferred embodiment of the method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data according to the present invention, the correction of the release window identification criteria, the window-level emission reference record retrieval method, and the irrigation and drainage scheme playback results includes: when correcting the release window identification criteria, the window-level emission reference record retrieval method, and the irrigation and drainage scheme playback results, the actual methane release window sequence and the predicted release window sequence are matched according to window type, window start time, window end time, trigger source, and termination source; when the actual methane release window sequence and the predicted release window sequence do not match according to window type, window start time, window end time, trigger source, and termination source, a window identification difference field is generated; the release window start and end identification criteria are corrected according to the window identification difference field; the window-level emission reference record is supplemented according to the actual methane flux measurement record; and the formation results of water layer events, aeration events, and salinity candidate events during the playback of candidate irrigation and drainage schemes are corrected according to the actual irrigation and drainage execution record.

[0017] The beneficial effects of this invention are as follows: By locking the methane release window based on the weekly event chain and forming a window sequence, the carbon supply, water aeration, salinity changes, and sampling events to be verified are windowed together in the time dimension, which is used to clarify the start and end, connection, and verification relationship of methane release, thereby improving the process characterization accuracy of the weekly emission flux prediction object; by replaying the candidate irrigation and drainage schemes to the weekly event chain, the water-saving and emission-reduction irrigation and drainage schemes are determined, and the irrigation and drainage arrangements are re-inferred from the release window sequence and prediction results, which is used to support the coordinated verification of water content boundaries, salinity control, and water consumption, thereby improving the supporting effect of the prediction results on the water-saving and emission-reduction regulation of paddy fields. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of a method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data.

[0020] Figure 2 This is a flowchart for predicting methane emissions.

[0021] Figure 3 This is a flowchart of candidate irrigation and drainage schemes and water-saving and emission-reduction procedures.

[0022] Figure 4 The timing diagram is corrected after the cycle ends. Detailed Implementation

[0023] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0024] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0026] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data, comprising the following steps:

[0027] S1. Obtain basic records of paddy fields, field management records, sensor monitoring records, and actual methane flux measurement records. Use seven consecutive days as the methane emission prediction cycle, and time-mark records related to moisture, salinity, temperature, straw return to the field, and irrigation and drainage to form a weekly event chain.

[0028] Paddy fields or paddy field zones under the same irrigation and drainage facilities are used as the targets for methane emission prediction. If a paddy field is controlled only by the same irrigation and drainage outlets, and the sensor deployment areas are within the same water layer control range, then that paddy field is considered as a prediction target. If a paddy field contains multiple irrigation outlets, drainage outlets, or sensor deployment areas, the paddy field is divided into zones based on the irrigation and drainage control ranges that can independently form water layer control boundaries. One predicted partition, It is an integer greater than or equal to 1, and This is equal to the number of irrigation and drainage control units within the paddy field that can independently control irrigation, drainage, or runoff. Each prediction zone is assigned a zone number, which is then linked to soil volumetric moisture sensors, soil conductivity sensors, water depth sensors, soil temperature sensors, irrigation inlets, drainage inlets, straw return records, and methane flux measurement records, ensuring that sensor monitoring records, irrigation / drainage execution records, and methane flux measurement records belong to the same prediction object.

[0029] A seven-day period is set as a methane emission prediction cycle, and weekly prediction records for paddy fields are generated. The reason for using a seven-day prediction cycle is that the effects of field management activities such as irrigation, drainage, drying, rainfall replenishment, and flood control on soil moisture, aeration, and salinity migration typically last for several days. The carbon supply process from straw decomposition after straw return to the field also needs to be assessed in conjunction with moisture and soil temperature over several days. If the prediction cycle is too short, it may only record the time of irrigation and drainage actions, failing to cover the changes in moisture and aeration and the electrical conductivity response process that occur after irrigation and drainage. If the prediction cycle is too long, different growth stages, different irrigation and drainage events, and different straw decomposition stages may be mixed into the same prediction cycle, affecting the identification of methane release windows. Therefore, a seven-day period can take into account the field management cycle, the methane flux statistical cycle, and the granularity of release window identification. Within the same project, one of the following methods is used: natural week division or a seven-day period after setting a start date, and these are permanently saved in the project configuration record.

[0030] In this embodiment, the paddy field baseline record is used to characterize the fixed and agronomic attributes of the predicted object, including plot area, soil type, salinity background value, transplanting date, rice variety, growth stage division, sensor deployment information, and irrigation and drainage facility location. Field management records are used to characterize manual or automated field operation processes, including straw return date, straw return amount, straw burial depth, irrigation time, drainage time, field drying time, rainfall replenishment time, and flood drainage time. Sensor monitoring records are used to characterize changes in field conditions within the prediction period, including soil volumetric water content, water layer depth, soil electrical conductivity, and soil temperature. Methane flux measurement records are used to characterize the measured methane emissions of the predicted object within the corresponding time period, and can be derived from automatic flux monitoring equipment, static box sampling, or manual sampling records.

[0031] Obtain basic records for the target prediction object. These records include plot area, soil type, background salinity, transplanting date, rice variety, growth stage division, straw return date, straw return amount, straw burial depth, irrigation and drainage facility location, sensor burial depth, and measured methane flux. Background salinity is determined from continuous conductivity monitoring records before transplanting or straw return; plot saturation water content, field capacity, and minimum safe production water content are determined from soil calibration records, agronomic management records, or project experiment records; rice growth stage division is determined from transplanting date, variety growth stage records, and field management records; straw return amount is recorded per unit area; and straw burial depth is determined from on-site straw return operation records or agricultural machinery operation records.

[0032] Acquire real-time records of the target object during the prediction period. These records include soil volumetric moisture content, soil electrical conductivity, water depth, soil temperature, irrigation time, drainage time, field drying time, rainfall replenishment time, and flood control time. Soil volumetric moisture content, water depth, soil electrical conductivity, and soil temperature are collected by sensors deployed within the target area; irrigation time, drainage time, and flood control time are determined by the opening and closing records of irrigation and drainage control devices, flow records, or records entered by field management personnel; rainfall replenishment time is determined by rainfall monitoring devices or project meteorological records.

[0033] A unified time stamp is applied to real-time records. Sensor-collected data is written to the weekly paddy field prediction record according to the sampling time; irrigation, drainage, field drying, rainfall-induced water replenishment, and flood drainage are written to the weekly paddy field prediction record according to the event start and end times; methane flux measurement records are written to the weekly paddy field prediction record according to the sampling time or flux monitoring time. When the sampling intervals of different sensors are inconsistent, the time granularity predetermined in the project configuration record is used as the unified time granularity, mapping all types of real-time records to the same time segment. For continuously occurring irrigation and drainage events, the same event marker is written within each time segment covered by the event.

[0034] Soil electrical conductivity is corrected for temperature based on sensor calibration records and soil temperature records to generate corrected conductivity values. These corrected values ​​are used to subsequently identify salinity change events, avoiding direct comparison with raw conductivity readings under different temperature conditions. In cases where multiple conductivity monitoring points exist for the same predicted object, each monitoring point is recorded with its own monitoring point number, burial depth, sampling time, and corresponding corrected conductivity value to prevent the direct merging of conductivity changes from different locations or depths.

[0035] A water content status field is generated based on soil volumetric water content. When the soil volumetric water content reaches the saturated water content minus the sensor's allowable error, it is written into the saturated water content status field; when the soil volumetric water content is within the allowable field capacity, it is written into the field capacity status field; when the soil volumetric water content is below the minimum safe production water content control boundary, it is written into the water content control boundary breach field. Saturated water content, field capacity, minimum safe production water content, and sensor allowable error are all derived from soil calibration records, agronomic management records, or sensor calibration records. The water content status field is used for subsequent identification of waterlogging without aeration events, aeration events, and to determine whether candidate irrigation and drainage schemes meet agronomic safety requirements.

[0036] A water layer event field is generated based on the water layer depth. When the water layer depth reaches the water layer identification limit and persists for two or more consecutive sampling periods, it is written into the flooding event field; when the water layer depth does not reach the water layer identification limit, but the soil volumetric water content is already written into the saturated water content field, it is written into the submerged water and not aerated event field; when the water layer depth does not reach the water layer identification limit, but the soil volumetric water content is already written into the field water holding capacity field, it is written into the aerated event field. The water layer identification limit is determined by the water layer depth sensor resolution, on-site water gauge calibration records, or the project monitoring scheme. Two or more consecutive sampling periods are used to eliminate the influence of single sampling jitter on water layer event identification.

[0037] A carbon supply candidate event field is generated based on the straw return-to-field records. No carbon supply candidate event field is generated when the straw return-to-field amount is zero; when the straw return-to-field amount is not zero, the field is generated based on the straw return date, the number of days since return, the straw burial depth, and the straw decomposition time range recorded in the project. The carbon supply candidate event field indicates whether straw decomposition and carbon supply conditions exist for the corresponding time period. Whether a water-carbon superposition release window is formed needs to be determined in subsequent steps in conjunction with flooding events, waterlogging without aeration events, and soil temperature.

[0038] A salinity candidate event field is generated based on the conductivity correction value. The salinity candidate event field is written when the conductivity correction value reaches the conductivity change judgment threshold relative to the salinization background value, or when the conductivity correction value forms a continuous change record in the same direction within the response time range after irrigation, drainage, rainfall replenishment, or field drying events.

[0039] The threshold for determining changes in electrical conductivity is determined based on sensor accuracy records, the background salinization value of the land parcel, and the project monitoring plan. The threshold can be set to more than twice the upper limit of the sensor's electrical conductivity measurement accuracy error, or it can be set to a preset proportional change range of the salinization background value. When the project monitoring plan has a specified value for identifying salinity changes, the specified value in the project monitoring plan shall be followed. This setting allows the threshold for determining changes in electrical conductivity to exclude single-read fluctuations caused by sensor measurement errors and to correspond to the background salinization level of the land parcel.

[0040] The response time range is determined based on the type of irrigation / drainage event, soil type, sensor burial depth, and project monitoring plan. The response time range is calculated from the first valid sampling period after the end of the irrigation, drainage, rainfall replenishment, or field drying event, and continues until the salinity response cutoff time set in the project monitoring plan. If the project monitoring plan does not specify a separate response time range, the response time range is set to a period of at least two consecutive sampling periods and no more than twenty-four hours after the end of the irrigation / drainage event. For conductivity monitoring points at different depths, the response time range is recorded according to the corresponding burial depth.

[0041] A continuous change record in the same direction refers to a conductivity correction value that continuously increases or decreases within the response time range, with at least two consecutive sampling periods for the continuous change. When a sampling event occurs during the continuous change, the continuous change record in the same direction is interrupted, and the corresponding time segment is written to the sampling event field. The salinity candidate event field is used to indicate whether a salinity change process exists within the prediction period, and whether a water-salt disturbance release window is formed. This needs to be determined in subsequent steps by combining irrigation and drainage events with the window connection relationship.

[0042] Real-time records are marked as pending sampling. When sensor sampling values ​​exceed the equipment range, sampling time is missing, continuous missing measurement time reaches the missing measurement judgment duration, or water depth records cannot form a time correspondence with irrigation, drainage, rainfall replenishment, and flood discharge records, the sampling pending event field is written into the corresponding time segment.

[0043] The missing data determination duration is used to determine whether a continuous data interruption has occurred in the sensor monitoring records within the prediction period. The missing data determination duration is determined based on the sensor sampling interval, the equipment's permissible interruption duration, and the project monitoring plan. Specifically, when the sensor sampling interval is a fixed value, the missing data determination duration is set to the duration corresponding to no less than three consecutive sampling intervals; when the project monitoring plan has separate requirements for continuous data integrity, the missing data determination duration is determined according to the permissible data interruption duration specified in the project monitoring plan; when the sensor equipment manual provides a permissible interruption duration, the missing data determination duration is the smaller value between the sensor equipment's permissible interruption duration and the duration specified in the project monitoring plan. This ensures that the missing data determination duration corresponds to the actual sampling frequency, equipment performance, and project monitoring accuracy requirements.

[0044] The sampled event field is not used to directly delete the corresponding data, but to mark the release window to be verified when identifying the release window later. When the release window to be verified covers the key monitoring fields of the water and carbon superposition release window, the moisture maintenance release window, or the water and salt disturbance release window, the field to be verified is written into the covered release window, and the single predicted value is adjusted to the prediction range when the prediction result is output.

[0045] In this embodiment, the moisture-related records include soil volumetric water content, water layer depth, saturated water content, field capacity, and minimum safe production water content; the salinity-related records include soil electrical conductivity, electrical conductivity correction value, salinization background value, electrical conductivity monitoring point records at different depths, and salinity response time range; the temperature-related records include soil temperature, sensor sampling time, and effective temperature conditions for methane production; the straw return-to-field-related records include the straw return-to-field date, number of days after return, amount of straw returned, straw burial depth, and straw decomposition time range; and the irrigation and drainage-related records include irrigation time, drainage time, field drying time, rainfall replenishment time, flood drainage time, irrigation and drainage outlet flow rate, and target water layer value. These records, after being uniformly time-stamped, are used to generate different event fields in the weekly event chain.

[0046] After completing the generation of time stamps, temperature corrections, water content status fields, aquifer event fields, carbon supply candidate event fields, salinity candidate event fields, and sampling verification event fields, the carbon supply candidate event fields, flooding event fields, waterlogging without aeration event fields, aeration event fields, salinity candidate event fields, irrigation event fields, drainage event fields, field drying event fields, rainfall replenishment event fields, flood drainage event fields, and sampling verification event fields are arranged into a weekly event chain according to the chronological order of time segments. Each event unit in the weekly event chain includes the event type, event start time, event end time, source data field, corresponding sensor number or irrigation / drainage facility number, and sampling verification event field.

[0047] S2. Based on the event chain within the week, lock the methane release window and form a methane release window sequence according to the start and end time of the window, the trigger source, the termination source, and the window connection relationship.

[0048] In this embodiment, the methane release window refers to a continuous time period within a methane emission prediction cycle, defined by carbon supply conditions, moisture aeration status, salinity changes, aeration reduction processes, or sampling events awaiting verification. The methane release window is not equivalent to the entire prediction cycle, but rather a specific time range within which methane release conditions are formed, maintained, disturbed, reduced, or await verification during the prediction cycle.

[0049] Based on the established weekly event chain, the following fields are read in chronological order: carbon supply candidate event field, flooding event field, waterlogging and non-aeration event field, aeration event field, salinity candidate event field, irrigation event field, drainage event field, field drying event field, rainfall and water replenishment event field, flood drainage event field, and sampling and verification event field. The continuous time periods related to the formation, maintenance, disturbance, or reduction of methane release within the target prediction period are identified, and the identification results are recorded as methane release window sequence.

[0050] The methane release window sequence is formed by arranging multiple release windows according to their start and end times. Each release window includes a window number, window type, window start time, window end time, trigger event, termination event, moisture status field, salinity status field, carbon supply status field, soil temperature status field, sampling verification field, and adjacent window relationship field. The window number is generated by combining the prediction object number, prediction period number, and window generation order, and is used for subsequent window-level emission reference record retrieval and candidate irrigation and drainage scheme playback verification.

[0051] The start time of the release window is determined by a triggering event. Triggering events include flooding events, water ingress without aeration events, carbon supply candidate events, salinity candidate events, aeration events, and sampling pending verification events. The end time of the release window is determined by a termination event. Termination events include the occurrence of an aeration event, the end of a carbon supply candidate event, the end of a salinity candidate event, the end of a water layer event, the end of a sampling pending verification event, or the end of the prediction period. If no termination event occurs within the same window before the end of the prediction period, the end time of the prediction period is used as the window end time.

[0052] When there is a flooding event or a submerged and unventilated event in the event chain within the week, and there are carbon supply candidate events within the same time range, and the soil temperature reaches the effective temperature condition for methane production, the part of the flooding event or submerged and unventilated event that overlaps with the carbon supply candidate event in time is locked as the water and carbon superposition release window.

[0053] The effective temperature conditions for methane production are determined based on local agronomic experiment records, historical methane flux measurement records, or the project monitoring plan. Specifically, historical methane flux measurement records are read under the same rice growth stage, the same straw return method, and the same moisture conditions. The methane flux during the corresponding growth stage that did not reach a carbon supply candidate event and was within the aeration event coverage period is used as the low-temperature baseline flux. When the soil temperature in the historical records falls within a certain continuous temperature range, and the corresponding methane flux is higher than the low-temperature baseline flux during the continuous observation period, and the corresponding time period simultaneously meets the conditions of either a flooding event or a flooded but non-aerated event, then this continuous temperature range is recorded as the effective temperature conditions for methane production. If there are insufficient local historical methane flux measurement records, the methane production temperature range pre-set in the project monitoring plan is used as the effective temperature conditions for methane production, and this is corrected after subsequent measurement records are supplemented.

[0054] The superimposed water and carbon release window is used to characterize the situation where straw decomposition carbon supply conditions and anaerobic moisture conditions coexist in the same time period.

[0055] The start time of the water-carbon superimposed release window is the start time of the overlapping portion between the flooding event or the non-aeration event and the carbon supply candidate event; the end time of the water-carbon superimposed release window is the end time of the overlapping portion. If an aeration event occurs during the overlapping period, and the duration of the aeration event reaches the aeration judgment time, then the start time of the aeration event is used as the termination correction time of the water-carbon superimposed release window.

[0056] The ventilation determination time is determined based on the sensor sampling interval, soil moisture state change records, and field management requirements. Specifically, the ventilation determination time is set to be no less than the time corresponding to two consecutive sampling cycles, and this consecutive time needs to cover the process of soil volumetric water content changing from saturated water content to field-holding water content. When records have been formed within two consecutive sampling cycles showing that the water layer depth has not reached the water layer identification boundary and the soil volumetric water content has entered the field-holding water content field, the time corresponding to the two consecutive sampling cycles is used as the ventilation determination time. When the time required for soil volumetric water content to change from saturated water content to field-holding water content exceeds two consecutive sampling cycles, the continuous time actually required to complete the state transition is used as the ventilation determination time.

[0057] With this setting, the ventilation determination time corresponds to the sensor sampling frequency and uses the change in soil volumetric water content as the confirmation basis, which can avoid the water and carbon release window being mistakenly cut off due to fluctuations at a single sampling point.

[0058] When a flooding event or a submerged and unaerated event occurs within the weekly event chain, but no carbon-supplying candidate event exists within the same time frame, or the number of days since straw return to the field has exceeded the carbon-supply time frame, the corresponding flooding event or submerged and unaerated event's coverage period is locked as the moisture maintenance release window. The moisture maintenance release window characterizes situations where anaerobic moisture conditions exist, but the straw decomposition carbon-supply stage is not yet in progress. The start time of the moisture maintenance release window is the start time of the flooding event or the submerged and unaerated event, and the end time is the end time of the flooding event or the submerged and unaerated event. If an aeration event occurs during this period and reaches the aeration determination time, the start time of the aeration event is used as the end time of the window.

[0059] When a salinity candidate event field forms within the response time range following an irrigation event, drainage event, rainfall replenishment event, or field drying event, and the conductivity correction value reaches the conductivity change judgment threshold, the time period covered by the salinity candidate event is locked as the water-salt disturbance release window. The response time range, conductivity change judgment threshold, and sampling frequency requirements for continuous changes in the same direction are derived from sensor accuracy records, plot salinization background values, and project monitoring plans. The water-salt disturbance release window is used to record the salinity change process caused by irrigation and drainage events. It does not directly replace the water-carbon superposition release window or the water maintenance release window, but rather serves as a salinity correction field for subsequent emission reference based on the time relationship.

[0060] The start time of the water-salt disturbance release window is taken from the first time segment formed by the salinity candidate event field, and the end time is taken from the end time segment of the salinity candidate event field. If the water-salt disturbance release window overlaps with the water-carbon superimposed release window, a salinity correction field is written into the water-carbon superimposed release window; if the water-salt disturbance release window overlaps with the water maintenance release window, a salinity correction field is written into the water maintenance release window; if the water-salt disturbance release window does not overlap with the water-carbon superimposed release window or the water maintenance release window, it is retained separately as a salinity disturbance record for use in subsequent irrigation and drainage scheme playback to determine the risk of salinity changes.

[0061] When a ventilation event occurs in the weekly event chain, and the duration of the ventilation event reaches the ventilation determination time, the time period covered by the ventilation event is locked as the release reduction window. The release reduction window is used to mark the time period during which a flooding event or a submerged non-ventilation event is interrupted. The release reduction window is not used as an independent emission increase window, but rather to correct the end time of adjacent water and carbon superimposed release windows or moisture maintenance release windows. If the release reduction window is adjacent to the previous water and carbon superimposed release window, the start time of the release reduction window is written into the termination correction field of the previous water and carbon superimposed release window; if the release reduction window is adjacent to the previous moisture maintenance release window, the start time of the release reduction window is written into the termination correction field of the previous moisture maintenance release window.

[0062] When a sampling event exists in the weekly event chain, and this sampling event covers any of the key fields of water depth, soil volumetric water content, electrical conductivity correction value, or soil temperature, the corresponding time period is locked as the release window for verification. The release window is used to mark situations where key fields required for prediction are missing, samples are taken outside the range, or the time correspondence is invalid. The release window does not directly participate in the output of a single flux value, but serves as the basis for subsequent prediction range output and verification field prompts. If the release window covers key fields of the water-carbon superimposed release window, the water maintenance release window, or the water-salt disturbance release window, the verification field is written into the corresponding main window.

[0063] In this embodiment, window connection relationships are used to characterize the temporal and synergistic relationships between multiple methane release windows within the same prediction period, including time overlap, consecutive windows, window truncation, salinity correction, and sampling pending verification. Time overlap indicates that two windows cover the same time segment; consecutive windows indicate that the end time of one window is continuous with the start time of another window; window truncation indicates that a release reduction window causes the premature end of a water-carbon superimposed release window or a moisture-maintaining release window; salinity correction indicates that a water-salt disturbance release window affects the reference record retrieval of the main release window; and sampling pending verification indicates that key monitoring fields are missing or do not meet the temporal correspondence.

[0064] After locking various release windows, the release windows within the same forecast period are sequentially organized. First, they are arranged according to their start time. For windows with the same start time, they are written into the window sequence in the following order: water-carbon superimposed release window, moisture maintenance release window, water-salt disturbance release window, release reduction window, and release window awaiting verification. Windows with overlapping times are not directly deleted; instead, they are written into the main window field, correction field, or awaiting verification field according to their window type. The water-carbon superimposed release window and the moisture maintenance release window are used as the main release windows; the water-salt disturbance release window is used as the salinity correction field; the release reduction window is used as the termination correction field; and the release window awaiting verification is used as the awaiting verification field.

[0065] In the window connection processing, if the start time of the release reduction window is located within the water-carbon superposition release window or the water maintenance release window, the corresponding main release window is truncated with the start time of the release reduction window, and the time period before truncation is retained as the valid release window; if the water-salt disturbance release window covers part of the time period of the main release window, the start time and end time of salinity correction are recorded in the main release window; if the release window to be cored covers the key fields of the main release window, the main release window does not output a single predicted flux value.

[0066] S3. Based on the window type, duration, status field, and window connection relationship in the methane release window sequence, call the corresponding window-level emission reference record, and determine the methane weekly emission flux prediction result based on window truncation, window correction, and the relationship between the output to be verified.

[0067] Based on the formed methane release window sequence, the window type, start time, end time, window duration, trigger event, termination event, moisture status field, salinity correction field, carbon supply status field, soil temperature status field, and sampling verification field are read one by one according to the window number. Based on the reading results, the window-level emission reference record is called to determine the methane flux prediction results corresponding to each release window within the target prediction period.

[0068] In this embodiment, the window-level emission reference record is emission reference data formed by recording confirmed methane release windows within the historical prediction period, and does not use the whole-week average as the sole recording object. The window-level emission reference record at least saves the reference window type, window start and end time, window duration, rice growth stage, carbon supply status, moisture status, salinity status, irrigation and drainage event type, sampling fields to be verified, and corresponding measured methane flux, which is used to provide flux prediction basis for similar release windows within the target prediction period.

[0069] Window-level emission reference records are formed from continuous historical time periods for which methane flux measurements have been completed. Each window-level emission reference record includes at least the reference window number, reference plot number, reference window type, rice growth stage, straw return amount segment, number of days after return segment, straw burial depth segment, duration of moisture status segment, duration of aeration segment, salinity change direction, irrigation / drainage event type, soil temperature status field, sampling verification field, and measured methane flux. The reference window is not recorded in whole weeks, but rather in continuous time periods during which methane release conditions are actually formed.

[0070] When establishing a window-level emission reference record, first read the confirmed release windows within the historical prediction period, and then assign the measured methane flux records within the corresponding time period of the release window. If the measured methane flux records are continuous monitoring data, then the corresponding measured flux is extracted according to the start and end times of the release window; if the measured methane flux records are from static chamber sampling or manual sampling, then the sampling results are assigned to release windows within the same time range according to the sampling time, sampling duration, and field conditions covered by the sampling.

[0071] For cases where the sampling time spans two release windows, the assigned window is determined based on the coverage ratio of the sampling duration within each release window. If the coverage ratio of the sampling duration within a certain release window exceeds half of the total sampling duration, the sampling result is assigned to that release window. If the coverage ratio of the sampling duration within neither release window exceeds half, the release window corresponding to the longest continuous coverage duration is used for assignment. If the longest continuous coverage durations are the same, the sampling result is not assigned to a single release window, but is written into the cross-window sampling field in the reference record, and the corresponding measured flux is used as the reference data when forming the prediction range. The cross-window sampling field includes the sampling start time, sampling end time, the covered release window number, the coverage duration of each release window, and the corresponding window type, for identification when subsequent window-level emission reference records are called.

[0072] When calling window-level emission reference records, a first-level screening is performed based on the target window type. When the target window is a water-carbon superimposed release window, the reference records must simultaneously meet the following requirements: consistent window type, consistent straw return amount segments, consistent number of days after return, consistent duration of water status segments, and consistent soil temperature status field. When the target window is a water maintenance release window, the reference records must meet the following requirements: consistent window type, consistent or adjacent rice growth stages, consistent duration of water status segments, and consistent irrigation / drainage event types. When the target window is a water-salt disturbance release window, the reference records must meet the following requirements: consistent window type, consistent direction of salinity change, consistent irrigation / drainage event types, and consistent response time range. When the target window is a release reduction window, the emission increase reference record is not directly called; instead, the main release window truncation correction is performed. When the target window is a release window awaiting verification, a single reference flux value is not called; instead, the prediction range output is used.

[0073] After the first layer of screening, a second layer of screening is conducted based on window duration, sensor burial depth, soil salinization background value, and rice growth stage. If multiple reference records exist after the second layer of screening, the reference record is selected according to the sampling coverage integrity of the measured methane flux record, the number of fields to be verified, and the time correspondence. Sampling coverage integrity refers to the proportion of the measured methane flux record covering the start and end times of the reference window; the number of fields to be verified refers to the number of fields in the reference window that affect the judgment of methane release; and the time correspondence refers to whether the irrigation / drainage events, salinity change events, and ventilation events in the reference window fall within the corresponding response time range.

[0074] For water and carbon superimposed release windows, if a salinity correction field exists within the window, the window-level emission reference record with the same salinity change direction and the same irrigation / drainage event type is invoked; if no corresponding salinity correction field exists, the water and carbon superimposed release reference record without a salinity correction field is invoked. In this way, salinity change does not replace the water and carbon superimposed release window alone, but rather participates in the reference record invocation as a correction condition for the water and carbon superimposed release window.

[0075] For the water maintenance release window, if the window is truncated by the release reduction window, the end time of the water maintenance release window is first replaced with the start time of the release reduction window, and then reference records are called segment by segment according to the duration of the truncated window. If a salinity correction field exists within the water maintenance release window, the reference record corresponding to the salinity change direction, irrigation / drainage event type, and water state duration segment is called. If the water maintenance release window is not truncated and no salinity correction field exists, the reference record corresponding to the rice growth stage and water state duration segment is called.

[0076] For water-salt disturbance release windows, if the water-salt disturbance release window overlaps with the water-carbon superimposed release window or the water maintenance release window in time, it is not used as an independent main window for flux calculation. Instead, it is used as the salinity correction field of the main window for reference record retrieval. If the water-salt disturbance release window does not overlap with the water-carbon superimposed release window or the water maintenance release window in time, a salinity disturbance record is generated. The salinity disturbance record includes the salinity disturbance start time, salinity disturbance end time, salinity change direction, triggering irrigation / drainage event, conductivity correction value change range, corresponding conductivity monitoring point number, and response time range.

[0077] The salinity disturbance record does not form a separate emission flux prediction value, but is written into the methane weekly emission flux prediction result of the target prediction period as a salinity status description field output. When replaying candidate irrigation and drainage schemes, the salinity disturbance record is used to determine whether the candidate irrigation and drainage scheme will form a new water and salt disturbance release window or reach the salinity upward movement judgment condition. During the actual measurement correction, the salinity disturbance record is used to correspond with the actual conductivity correction value change record, and when the two do not correspond, the basis for the formation of salinity candidate events, the salinity response time range, or the salinization background value is corrected.

[0078] For the release reduction window, read the start time, duration, and corresponding ventilation event fields. If the release reduction window is adjacent to the previous water-carbon superposition release window or water maintenance release window, the period from the start time of the previous main release window to the start time of the release reduction window is used as the corrected effective release time period, and then reference records are called in segments according to the corrected duration. If the release reduction window is not adjacent to the main release window, it is retained as a ventilation event record.

[0079] For the release window to be verified, the key field types covered by the sampled fields to be verified are read. If the sampled fields to be verified cover any of the following fields: water layer depth, soil volumetric water content, electrical conductivity correction value, or soil temperature, the corresponding main release window will not output a single flux prediction value. Instead, it will call the previous valid window of the same prediction object, adjacent prediction object windows of the same type, and reference windows of the same growth period to form a prediction range, and simultaneously output the list of fields to be verified. The list of fields to be verified includes the missing field name, missing time segment, source sensor number, affected release window number, and affected candidate irrigation and drainage scheme number.

[0080] The list of fields to be verified is written into the methane weekly emission flux prediction results and used as a limiting field for scheme determination in the verification of candidate irrigation and drainage schemes. When the list of fields to be verified covers water layer depth or soil volumetric water content, candidate irrigation and drainage schemes involving short-term water droplet aeration, delayed irrigation, or intermittent irrigation are not directly determined as water-saving and emission-reducing irrigation and drainage schemes, but are output as candidate schemes to be verified; when the list of fields to be verified covers conductivity correction values, candidate irrigation and drainage schemes involving salinity upward movement judgment conditions are not directly verified by salinity, but are output as salinity-verified schemes; when the list of fields to be verified covers soil temperature, flux prediction results involving water and carbon superposition release windows are retained in the form of prediction ranges.

[0081] After the target prediction period ends, the list of fields to be verified is used in the actual correction process. Based on the source sensor number, missing time segment, and affected release window number in the list of fields to be verified, the supplemented complete sensor monitoring records, equipment maintenance records, or manual review records are read. When the supplemented records can restore the key fields of the corresponding time segment, the corresponding actual release window is regenerated, and the window-level emission reference record is updated. When the supplemented records still cannot restore the corresponding key fields, the corresponding release window is retained as a release window to be verified, and the priority of the corresponding reference record is reduced in subsequent prediction periods.

[0082] Through this setting, the list of fields to be verified not only serves as information for displaying prediction results, but also participates in the verification of candidate irrigation and drainage schemes, correction of measured data, and control of subsequent reference record calls, so that the impact of missing sampling or inconsistent fields on emission prediction and irrigation and drainage scheme determination can be traced and controlled.

[0083] In this embodiment, the window truncation relationship is used to handle the case where the release reduction window is adjacent to the main release window, and the start time of the release reduction window serves as the end correction time of the main release window. The window correction relationship is used to handle the case where the water-salt disturbance release window overlaps with the main release window, and the water-salt disturbance release window participates in the window-level emission reference record call as the salinity correction field of the main release window. The output to be verified relationship is used to handle the case where the output to be verified window covers key monitoring fields. The covered main release window does not output a single flux prediction value, but outputs the methane flux prediction range and the list of fields to be verified.

[0084] After retrieving reference records for all release windows, the methane emission prediction results for the target prediction period are summarized in chronological order of the release window sequence. The output window flux prediction values ​​for main release windows not covered by the release windows to be nucleated are: the output window flux prediction range for main release windows covered by the release windows to be nucleated; the release reduction window is only used to correct the duration of adjacent main release windows; and the water-salt disturbance release window is retained as a salinity correction field or a salinity disturbance record based on the overlap relationship.

[0085] The predicted weekly methane emission flux for the target prediction period is formed by summing the cumulative emission contributions of each main release window. The window flux prediction value obtained by calling the window-level emission reference record is the representative value of the flux per unit time. The representative value of the flux per unit time can be derived from the average flux per unit time of the same type of reference window, or from the flux conversion caliber specified in the project monitoring plan. For main release windows not covered by the release windows to be verified, the cumulative emission contribution of that main release window is formed by multiplying the predicted window flux value by the window duration.

[0086] For a main release window covered by a pending release window, instead of outputting a single window flux prediction value, the window flux prediction range is output. The window flux prediction range includes a lower limit and an upper limit, formed by the previous effective window for the same prediction object, windows of the same type for adjacent prediction objects, and reference windows of the same growth period, respectively. The lower limit of the window flux prediction is multiplied by the window duration to form the lower limit of the cumulative emission contribution of the window; the upper limit of the window flux prediction is multiplied by the window duration to form the upper limit of the cumulative emission contribution of the window.

[0087] Within the same consecutive seven-day forecast period, the cumulative emission contributions of each non-pending main release window are summed, and the lower and upper limits of the cumulative emission contributions of each pending main release window are included in the summary. When there are no pending main release windows within the forecast period, the predicted weekly methane emission flux is output; when there are pending main release windows within the forecast period, the predicted range of weekly methane emission flux is output, and a list of fields to be verified is output simultaneously.

[0088] The output results should include at least the prediction period number, prediction object number, main release window number, main release window type, window start time, window end time, window duration, reference record number, window flux prediction value or window flux prediction range, window cumulative emission contribution or window cumulative emission contribution range, salinity correction field, release reduction correction field, list of fields to be verified, and methane weekly emission flux prediction value or methane weekly emission flux prediction range.

[0089] S4. Generate candidate irrigation and drainage schemes based on the original planned irrigation and drainage schedule. Replay the candidate irrigation and drainage schemes into the weekly event chain to regenerate the methane release window sequence and methane weekly emission flux prediction results corresponding to the candidate schemes. Combine the changes in release window, agronomic safety boundary, salinity control conditions and irrigation water consumption to determine the water-saving and emission-reducing irrigation and drainage scheme.

[0090] After obtaining the predicted weekly methane emission flux under the original planned conditions, the system reads the intra-weekly event chain, methane release window sequence, window-level emission reference record retrieval results, and the original planned irrigation and drainage schedule within the current prediction period. Using the original planned irrigation and drainage schedule as a reference, candidate irrigation and drainage schemes are generated. The original planned irrigation and drainage schedule includes planned irrigation time, planned drainage time, planned water level target value, planned field drying time, planned resumption of irrigation time, irrigation and drainage outlet flow records, and field management notes.

[0091] Candidate irrigation and drainage schemes are used to alter water layer events, aeration events, and salinity candidate events within the current prediction period, thereby changing the methane release window sequence, without exceeding agronomic safety boundaries and salinity control conditions. Agronomic safety boundaries include the minimum safe production water content control boundary and permissible operation records for the rice growth period. The minimum safe production water content control boundary limits the lower limit of soil volumetric water content after the implementation of candidate irrigation and drainage schemes. Permissible operation records for the rice growth period limit whether short-term irrigation and aeration, delayed irrigation, field drying, resumption of irrigation, or maintaining shallow water within the current growth period are permissible. Salinity control conditions include the time range for determining salinity upward movement, salinity upward movement determination conditions, salinization background values, and records of changes in conductivity monitoring points at different depths.

[0092] In this embodiment, the original planned irrigation and drainage arrangement refers to the field water management arrangement formed before the start of the target prediction period, including the planned irrigation time, planned drainage time, planned water level target value, planned field drying time, planned resumption of irrigation time, and irrigation / drainage inlet flow rate. Candidate irrigation and drainage schemes are alternative arrangements generated based on the original planned irrigation and drainage arrangement, real-time monitoring status, and rice growth stage requirements. Replay refers to rewriting the irrigation time, drainage time, water level target value, water drop duration, and resumption of irrigation time from the candidate irrigation and drainage schemes into the weekly event chain, and regenerating the corresponding water level events, aeration events, salinity candidate events, and methane release window sequences.

[0093] Candidate irrigation and drainage schemes are generated based on the permissible operation records during the rice growth period, the original planned irrigation and drainage schedule, weather forecast records, soil volumetric water content status field, electrical conductivity correction value, salinization background value, and the control capacity of irrigation and drainage facilities. When generating candidate irrigation and drainage schemes, the permissible operation records during the rice growth period are first read to determine the types of irrigation and drainage operations allowed to be performed within the current forecast period; then, the original planned irrigation and drainage schedule is read to determine the original planned irrigation time, drainage time, water layer target value, and resumption irrigation time; next, weather forecast records are used to determine if there is a period of rainfall replenishment within the current forecast period; then, the difference between the current water content status and the minimum production safety water content control boundary is determined based on the soil volumetric water content status field; next, the electrical conductivity correction value and salinization background value are used to determine whether a salinity candidate event or a salinity upward shift judgment record has been formed; finally, the control capacity of irrigation and drainage facilities is used to determine the irrigation inlet, drainage outlet, irrigable and drainable volume per unit time, and adjustable water layer range.

[0094] A shallow water conservation plan is generated when the water requirement record in the rice growth period operation record requires maintaining a water layer, or when the difference between the soil volumetric moisture content and the minimum safe production moisture content control boundary reaches the water replenishment judgment threshold. The water replenishment judgment threshold is determined by the permissible error of the soil volumetric moisture content sensor, soil calibration records, and field management plan. The shallow water conservation plan uses the water layer maintenance range in the water requirement record as the target water layer value, the originally planned irrigation time or the time when the water replenishment judgment threshold is reached as the irrigation time, and determines the irrigation duration based on the irrigation and drainage inlet flow records.

[0095] When water control is permitted in the rice growth period operation record, and the soil volumetric moisture content does not fall below the minimum safe production moisture content control boundary during the recovery period, an intermittent irrigation plan is generated. The intermittent irrigation plan takes the original planned irrigation time as the starting point, divides the current prediction period into an irrigation maintenance period and a drainage and aeration period, and determines the irrigation recovery time based on historical irrigation and drainage response records, so that the soil volumetric moisture content after drainage and aeration does not fall below the minimum safe production moisture content control boundary.

[0096] When the rice growth period allows for short-term irrigation in the operational record, and there is a water-carbon superposition release window or a water maintenance release window in the weekly event chain, a short-term irrigation and aeration plan is generated. The short-term irrigation and aeration plan uses the start time of the corresponding release window or the originally planned drainage time as the drainage reference time. The start time of irrigation is determined based on the historical water layer receding time record after drainage. The duration of irrigation is determined based on the aeration judgment duration and the minimum production safety moisture content control boundary. The re-release conditions for the release reduction window are generated based on the irrigation recovery time. When a salt uplift judgment record is formed during the candidate re-release process, a short-term irrigation and aeration plan is not generated, or the irrigation duration is shortened and written into the candidate plan to be verified.

[0097] When weather forecast records indicate that there is a period of rainfall replenishment within the current forecast period, and the predicted rainfall reaches the rainfall replenishment threshold, and the original planned irrigation and drainage arrangements have water level maintenance arrangements in place before the rainfall replenishment, a pre-rainfall drainage plan is generated. The rainfall replenishment threshold is determined by project meteorological records, field water level target values, and drainage management requirements. The pre-rainfall drainage plan uses the executable irrigation and drainage time before the start of rainfall replenishment as the drainage time. The drainage duration is determined based on the water level target value, irrigation and drainage inlet flow records, and water level depth records, ensuring that the water level depth after rainfall replenishment does not exceed the water level target value. During playback, corresponding water level event fields and salinity candidate event fields are generated simultaneously.

[0098] When rainfall replenishment records show that rainfall replenishment has occurred within the current forecast period and the water layer depth exceeds the target water layer value, or when the conductivity correction value forms a salinity candidate event within the response time range after rainfall replenishment, a post-rainfall control and drainage plan is generated. The post-rainfall control and drainage plan uses the executable irrigation and drainage time after the end of rainfall replenishment as the drainage time, and determines the drainage duration based on the target water layer value, drainage outlet flow records, and salinity upward movement determination records. If a salinity upward movement determination record is formed after the post-rainfall control and drainage plan is replayed, the drainage time period in the post-rainfall control and drainage plan is adjusted or added to the candidate plan to be verified.

[0099] When delayed irrigation is permitted in the rice growth period operation record, and the soil volumetric moisture content field indicates that the post-delayed moisture content will not fall below the minimum safe production moisture content control boundary, a delayed irrigation plan is generated. The delayed irrigation plan uses the original planned irrigation time as a reference, postponing the irrigation time to within the time range specified in the permitted water control record. It also determines whether an aeration event occurs during the delay period based on historical moisture content change records. If a salt uplift determination record or a moisture content control boundary breach occurs during the delay period, no delayed irrigation plan is generated, or the delay time is shortened and added to the candidate plan for verification.

[0100] Each candidate irrigation and drainage scheme records the scheme number, scheme type, planned irrigation time, planned drainage time, target water level, duration of water drop, resumption of irrigation time, corresponding irrigation / drainage outlet number, estimated irrigation water consumption, and applicable growth period. The estimated irrigation water consumption is determined by the irrigation / drainage outlet flow records and irrigation duration, while the target water level is determined by field management records, irrigation / drainage facility control accuracy, and water requirement records from the permissible operation records for the rice growth period. Scheme types that do not meet the corresponding generation conditions are not included in the candidate irrigation and drainage schemes; schemes that meet the generation conditions but have missing fields for sampling verification, missing conductivity correction values ​​required for salinity upward movement determination, missing monitoring point numbers, insufficient continuous sampling records, or insufficient irrigation / drainage response records are included in the candidate schemes for verification.

[0101] After generating candidate irrigation and drainage plans, each candidate plan is replayed to the weekly event chain. During replay, the original planned irrigation and drainage event fields are replaced or supplemented according to the irrigation time, drainage time, water level target value, water drop duration, and irrigation recovery time in the candidate irrigation and drainage plans. The irrigation event fields, drainage event fields, field drying event fields, rainfall replenishment event fields, flood drainage event fields, water level event fields, and aeration event fields under the candidate plan conditions are regenerated.

[0102] For sensor-measured fields that remain unchanged from candidate irrigation and drainage schemes, the soil temperature field, carbon supply candidate event field, and sampling verification event field from the original weekly event chain are retained. For the water layer depth, soil volumetric water content, and electrical conductivity correction values ​​affected by irrigation and drainage actions, segmented playback processing is performed according to the irrigation and drainage response records. The irrigation and drainage response records include records of water layer depth changes, soil volumetric water content changes, and electrical conductivity correction value changes for the same predicted object or similar predicted objects under the same growth period, the same irrigation and drainage event type, the same initial water content state, and the same salinity background range.

[0103] The segmented playback process includes water layer playback, water content state playback, and salinity state playback. During water layer playback, the start and end times of water layer events are determined based on irrigation time, target water layer value, irrigation / drainage outlet flow records, and historical water layer arrival time records in the candidate irrigation / drainage schemes. If the candidate irrigation / drainage scheme includes drainage, flooding, or field drying arrangements, the start time of the flooding or aeration event is determined based on drainage time, flooding duration, and historical water layer receding time records. During water content state playback, the soil volumetric water content state field before the execution of the candidate irrigation / drainage scheme is used as the initial state. Based on historical records of the transition from water content state to saturated water content state after irrigation and the transition from saturated water content state to field capacity state after drainage, the saturated water content state field, field capacity state field, and water content control boundary breach field under the candidate scheme conditions are regenerated. When replaying the salinity status, the irrigation, drainage, rainfall replenishment, or field drying events corresponding to the candidate irrigation and drainage schemes are read. Based on the response time range, change direction, and salinity upward movement judgment conditions of the historical conductivity correction value after the same event, the salinity candidate event field under the candidate scheme conditions is regenerated.

[0104] When a corresponding irrigation / drainage response record exists for the same predicted object, the record for the same predicted object is used first. If there are insufficient records for the same predicted object, records for adjacent predicted objects with the same soil type, salinization background value, growth period, and irrigation / drainage event type are used. If there are still insufficient records for adjacent predicted objects, the candidate irrigation / drainage scheme does not directly enter the water-saving and emission-reduction irrigation / drainage scheme determination process, but instead generates a candidate scheme to be verified, retaining the irrigation / drainage response fields that need to be supplemented. Through water layer playback, water content state playback, and salinity state playback, an intra-week event chain for the candidate scheme is formed.

[0105] After the event chain of candidate schemes is formed within the week, the methane release window sequence is re-locked according to the release window locking rules. If a candidate irrigation / drainage scheme causes an aeration event within the original water-carbon superimposed release window, and the aeration duration reaches the aeration judgment time, the start time of the aeration event is used as the termination correction time of the water-carbon superimposed release window, and a release reduction window is generated. If a candidate irrigation / drainage scheme causes the original water maintenance release window to become a time period covered by the aeration event, the water maintenance release window is truncated with the start time of the aeration event. If a candidate irrigation / drainage scheme causes the conductivity correction value to reach the salinity upward judgment condition within the salinity upward judgment time range, the salinity risk field is written into the candidate scheme window sequence. If the candidate irrigation / drainage scheme does not change the start and end times of the main release window, nor does it change the salinity correction field and the field to be verified, the original window sequence correspondence is retained.

[0106] After generating a new methane release window sequence for each candidate scheme, the predicted weekly methane emission flux under the candidate scheme conditions is determined according to the window-level emission reference record retrieval method in S3. The prediction results corresponding to the candidate scheme include at least the candidate scheme number, candidate scheme window sequence, number of main release windows, duration of each main release window, number of release reduction windows, salinity risk field, fields to be verified, window-level emission reference record number, and the predicted weekly methane emission flux of the candidate scheme. When there are release windows to be verified for a candidate scheme, a single predicted value is not output, but a prediction range is output, and the list of fields to be verified is retained in the scheme verification results.

[0107] In this embodiment, changes in the release window include changes in the number of main release windows, changes in the duration of the main release windows, the formation of release reduction windows, and the formation of water-salt disturbance release windows. Agronomic safety boundaries include the minimum safe production moisture content control boundary and records of permissible operations during the rice growth period. Salinity control conditions include the time range for determining salt uplift, the conditions for determining salt uplift, salinization background values, and records of changes in conductivity monitoring points at different depths. Irrigation water volume is determined based on irrigation and drainage inlet flow records and irrigation duration, and is used for comparison with the water volume corresponding to the original planned irrigation and drainage schedule.

[0108] After predicting the candidate irrigation and drainage schemes, the schemes are verified. The verification includes verification of changes in the methane release window, verification of the minimum safe production water content control boundary, verification of salinity upward movement conditions, verification of irrigation water consumption, and verification of suitability for the growing season.

[0109] The methane release window change verification reads the original planned window sequence and the candidate scheme window sequence. If a candidate irrigation / drainage scheme results in a shorter duration for the superimposed water and carbon release window than the corresponding duration in the original plan, or a shorter duration for the water maintenance release window than the corresponding duration in the original plan, and no new superimposed water and carbon release windows are generated, the release window shortening verification is considered passed. If a candidate irrigation / drainage scheme results in a greater number of main release windows than the original planned number, or a longer duration for the superimposed water and carbon release window than the corresponding duration in the original plan, the release window shortening verification is considered failed.

[0110] The minimum safe production moisture content control boundary verification is performed by reading the soil volumetric moisture content projection records and the minimum safe production moisture content control boundary under the candidate scheme conditions. If the candidate scheme causes the soil volumetric moisture content to fall below the minimum safe production moisture content control boundary in any time segment, the moisture content control boundary verification is considered unsuccessful; if the candidate scheme does not cause the soil volumetric moisture content to fall below the minimum safe production moisture content control boundary in any time segment, the moisture content control boundary verification is considered successful. The minimum safe production moisture content control boundary is derived from the plot's agronomic management records, soil calibration records, or project test records.

[0111] The verification of salinity upward movement conditions involves reading the corrected conductivity values ​​at different depths of conductivity monitoring points under candidate scheme conditions, records of conductivity changes before and after irrigation and drainage, and background salinization values. If, within the salinity upward movement judgment time range, the candidate scheme results in an upward movement judgment record for the surface conductivity correction value relative to the lower layer conductivity correction value, or meets the salinity upward movement judgment condition relative to the salinization background value, the salinity upward movement verification is considered failed; otherwise, it is considered passed. The salinity upward movement judgment time range and conditions are determined by the project monitoring scheme, plot salinization background records, and sensor accuracy records.

[0112] The calibration of salt uplift conditions involves reading the corrected conductivity values, conductivity changes before and after irrigation and drainage, and salinization background values ​​at different depths of conductivity monitoring points under candidate scheme conditions. Surface conductivity monitoring points correspond to the burial depth in the upper part of the tillage layer, while lower conductivity monitoring points correspond to the burial depth in the lower part of the tillage layer or the lower part of the root zone. The specific burial depth is determined by the sensor deployment records.

[0113] A salinity upward shift determination record is formed when the difference between the surface conductivity correction value and the lower layer conductivity correction value reaches the conductivity upward shift determination threshold and continuously reaches the salinity upward shift determination sampling number; or, a salinity upward shift determination record is formed when the deviation of the surface conductivity correction value from the salinization background value reaches the background deviation determination threshold and continuously reaches the salinity upward shift determination sampling number. The conductivity upward shift determination threshold, background deviation determination threshold, and salinity upward shift determination sampling number are determined by sensor accuracy records, plot salinization background value, salinity response time range, and project monitoring plan; the salinity upward shift determination sampling number shall not be less than three valid sampling points.

[0114] The irrigation water usage verification process involves reading the projected irrigation water usage of candidate schemes and the original projected irrigation water usage. If the projected irrigation water usage of a candidate scheme is less than the original projected irrigation water usage, the water usage verification is considered passed; if the projected irrigation water usage of a candidate scheme is not less than the original projected irrigation water usage, the water usage verification is considered failed. For candidate schemes aimed at flood drainage, salinity control, or water layer restoration, the water usage verification results can be retained, but they are not used as the sole basis for elimination. Instead, a combined verification of methane release window changes and salinity upward shift conditions is used for the final decision.

[0115] The suitability verification for the rice growth period involves reading the rice growth period field and the candidate scheme type. Candidate schemes involving short-term irrigation and aeration, delayed irrigation, or field drying must correspond to the permitted water control records in the rice growth period permitted operation records; candidate schemes involving restored irrigation or shallow water maintenance must correspond to the water requirement records in the rice growth period permitted operation records. If a candidate scheme type does not fall within the permitted operation records for the growth period, the suitability verification is considered failed; if a candidate scheme type falls within the permitted operation records for the growth period, the verification is considered passed. Permitted operation records for the growth period are derived from variety growth period records, field management plans, or project agronomic records.

[0116] The selection of candidate solutions is determined based on the results of each verification. A candidate solution that simultaneously passes the verifications for shortening the release window, water content control boundary, salinity shift, and suitability for the reproductive period is included in the water conservation and emission reduction candidate set.

[0117] A candidate scheme is designated as a water-saving and emission-reduction scheme if its predicted weekly methane emission flux is lower than the original predicted weekly methane emission flux and its expected irrigation water consumption is lower than the original expected irrigation water consumption. A candidate scheme is designated as a water-saving alternative scheme if its predicted weekly methane emission flux is not higher than the original predicted weekly methane emission flux and its expected irrigation water consumption is lower than the original expected irrigation water consumption. A candidate scheme is designated as an emission-reduction alternative scheme if its predicted weekly methane emission flux is lower than the original predicted weekly methane emission flux but its expected irrigation water consumption is not lower than the original expected irrigation water consumption.

[0118] When a water-saving and emission-reduction scheme exists in the candidate set, the water-saving and emission-reduction irrigation and drainage scheme for the target prediction period is determined based on changes in the release window duration, changes in expected irrigation water consumption, and the operating conditions of irrigation and drainage facilities. When no water-saving and emission-reduction scheme exists, but alternative water-saving or emission-reduction schemes exist, the alternative water-saving or emission-reduction schemes are not directly identified as water-saving and emission-reduction irrigation and drainage schemes, but are instead added to the candidate scheme list.

[0119] The candidate scheme list should include at least the candidate scheme number, scheme type, predicted weekly methane emission flux, estimated irrigation water consumption, methane release window change records, water content control boundary verification results, salinity shift verification results, growing season suitability verification results, and the reasons for not being selected as a water-saving and emission-reducing irrigation and drainage scheme. The candidate scheme list is used for manual review, selection of implementation schemes by field managers, or as a reference record when generating candidate irrigation and drainage schemes in the next forecast cycle.

[0120] When the next forecast period and the current forecast period belong to the same forecast object, and the rice growth period, salinization background value and irrigation and drainage facility conditions have not changed, the water-saving and emission-reduction alternatives in the candidate alternative list can be used as inputs for generating candidate irrigation and drainage schemes for the next forecast period; when the monitoring records of the next forecast period show that the water content, salinity, or release window sequence has changed, the candidate alternative list is only retained as a manual review record and is not directly used as the output of candidate irrigation and drainage schemes.

[0121] When multiple candidate schemes are included in the water-saving and emission-reduction candidate set, schemes that fail the water content control boundary verification, salinity shift verification, or fertile period applicability verification are first eliminated. Then, schemes that reduce the duration of the water-carbon superposition release window, do not increase the water-salt disturbance release window, and whose expected irrigation water consumption is less than the originally planned irrigation water consumption are prioritized. If multiple schemes still meet the same conditions, the scheme whose execution time matches the control capacity of existing irrigation and drainage facilities is selected. The final output is a water-saving and emission-reduction irrigation and drainage scheme for the target prediction cycle.

[0122] The target prediction cycle water-saving and emission-reduction irrigation and drainage plan should include at least the plan number, prediction object number, irrigation time, drainage time, target water level value, duration of water drop, resumption of irrigation time, corresponding irrigation / drainage outlet number, expected irrigation water consumption, candidate plan window sequence, predicted weekly methane emission flux, water content control boundary verification results, salinity upward shift verification results, fertile season suitability verification results, and an execution prompt field. The execution prompt field is used to prompt irrigation and drainage personnel or automatic irrigation and drainage control devices to perform irrigation, drainage, water drop and aeration, or resumption of irrigation operations according to the plan's time.

[0123] S5. After the methane emission prediction cycle ends, generate the actual intra-week event chain and the actual methane release window sequence based on the actual methane flux measurement record, the actual irrigation and discharge execution record and the complete sensor monitoring record. Based on the correspondence between the actual methane release window sequence and the predicted release window sequence, correct the release window identification criteria, the window-level emission reference record calling method and the irrigation and discharge scheme playback results.

[0124] After the target prediction period ends, the measured methane flux, actual irrigation and drainage execution records, and complete sensor monitoring records for the target prediction object within the same prediction period are acquired to form the target prediction period measurement record. The target prediction period measurement record includes the actual weekly methane emission flux, actual water layer depth change record, actual soil volumetric water content change record, actual conductivity correction value change record, actual soil temperature record, actual irrigation time, actual drainage time, actual field drying time, actual rainfall replenishment time, actual flood drainage time, actual straw decomposition status field, and sampling verification field.

[0125] In this embodiment, the actual intra-week event chain is a reconstructed event chain based on the complete sensor monitoring records, actual irrigation and drainage execution records, and measured methane flux records obtained after the end of the target prediction cycle. The actual methane release window sequence is a window sequence re-locked based on the actual intra-week event chain. The predicted release window sequence is a window sequence formed during the prediction phase of the target prediction cycle based on the playback of the original planned irrigation and drainage arrangements or candidate irrigation and drainage schemes. The correspondence between the actual methane release window sequence and the predicted release window sequence is used to determine whether the release window identification criteria and the playback results of the irrigation and drainage schemes need to be corrected.

[0126] Time merging is performed on the measured records for the target prediction period. Methane flux measured records are assigned to corresponding time segments based on sampling time or flux monitoring time; actual irrigation and drainage execution records are assigned to corresponding time segments based on event start and end times; sensor monitoring records are assigned to corresponding time segments according to a defined uniform time granularity. If the methane flux measured record comes from an automatic flux monitoring device, the corresponding measured flux is extracted according to the start and end times of the release window; if the methane flux measured record comes from static chamber sampling or manual sampling, it is assigned according to sampling time, sampling duration, and moisture, salinity, and carbon supply events within the sampling period.

[0127] Based on the actual measurement records of the target prediction period, the actual weekly event chain is reconstructed. The actual weekly event chain includes actual carbon supply candidate events, actual flooding events, actual water-induced non-aeration events, actual aeration events, actual salinity candidate events, actual irrigation events, actual drainage events, actual field drying events, actual rainfall-induced water replenishment events, actual flood drainage events, and actual sampling pending verification events.

[0128] Based on the actual weekly event chain, the actual methane release window sequence is regenerated according to the release window locking rules. The actual methane release window sequence includes the actual water-carbon superposition release window, the actual water maintenance release window, the actual water-salt disturbance release window, the actual release reduction window, and the actual release window to be verified. Each actual release window records the window number, window type, window start time, window end time, trigger event, termination event, water status field, salinity correction field, carbon supply status field, soil temperature status field, and sampling verification field.

[0129] The locked predicted release window sequence, the candidate release window sequence obtained from playback, and the actual methane release window sequence generated at the end of the current prediction period are mapped to each other. The mapping relationship is determined according to the prediction object number, prediction period number, window type, window start time range, window end time range, trigger event, and termination event. If the windows in the actual methane release window sequence and the windows in the predicted release window sequence cannot be mapped to each other in terms of window type, start time, end time, or trigger event, a window identification difference field is generated. The window identification difference field includes at least the difference window number, difference type, involved time segment, involved data field, and source record number.

[0130] When the actual start time of the water and carbon superimposed release window is earlier than the predicted start time, and the difference time reaches the window start correction judgment duration, the start identification criteria for the water and carbon superimposed release window are corrected; when the actual end time of the water and carbon superimposed release window is earlier or later than the predicted end time, and the difference time reaches the window end correction judgment duration, the end identification criteria for the water and carbon superimposed release window are corrected. The window start correction judgment duration and the window end correction judgment duration are determined by the sampling interval, the measured frequency of methane flux, and the project monitoring plan.

[0131] When the actual water maintenance release window and the predicted water maintenance release window are inconsistent in duration segmentation, and the difference originates from water depth events or soil volumetric water content status fields, the duration determination criteria for flooding events, water-induced non-aeration events, or aeration events are revised. If the difference originates from inconsistencies between irrigation and drainage execution records and planned irrigation and drainage records, the water event determination criteria are not revised, and the actual irrigation and drainage execution time is written into the irrigation and drainage playback verification record.

[0132] When the actual water and salt disturbance release window differs from the predicted water and salt disturbance release window in terms of salinity change direction, response time range, or duration, the actual conductivity correction value, conductivity monitoring point change records at different depths, irrigation and drainage event times, and salinization background values ​​are read. If the difference is related to the change in conductivity correction value relative to salinization background value, the basis for forming salinity candidate events is updated; if the difference is related to the response time of irrigation and drainage events, the salinity response time range of the corresponding predicted object is updated; if the difference originates from the sampling field to be verified covering the conductivity correction value, the original judgment basis is retained, and the corresponding time segment is written into the verification record.

[0133] When the actual release reduction window does not appear in the candidate scheme playback results, or when the candidate scheme playback results show a release reduction window but the actual record does not show an aeration event, the actual drainage time, water drop duration, irrigation recovery time, and soil volumetric moisture content change records are read. If the difference stems from the irrigation / drainage execution time deviating from the candidate scheme time, the actual execution time is written into the irrigation / drainage playback verification record; if the difference stems from the soil volumetric moisture content response being inconsistent with similar records, the irrigation / drainage response record of the corresponding prediction object is updated.

[0134] The window-level emission reference record is updated based on actual methane flux measurements. For measured fluxes that correspond to actual release windows, the actual release window number, window type, start time, end time, rice growth stage, straw decomposition state, moisture state, salinity state, irrigation / drainage event type, window duration, fields to be sampled for verification, and measured methane flux are written into the window-level emission reference record. For actual release windows with fields to be sampled for verification, these fields are retained when writing to the reference record and are not prioritized for recall, but can be used for prediction range formation.

[0135] When the predicted results for the same window type do not correspond to the measured results in two or more consecutive prediction periods, and the differences do not originate from the fields to be verified in the sampling, the window-level emission reference record retrieval method for the corresponding window type is updated. The updates include the reference record filtering field order, window duration segmentation, salinity correction field retrieval method, release reduction window truncation method, and field to be verified output method. The basis for using two or more consecutive prediction periods as the update trigger condition is that differences in a single prediction period may originate from manual sampling time, temporary irrigation / drainage operations, or equipment maintenance, while consecutive differences reflect a need to adjust the window identification criteria for the corresponding plot or management method.

[0136] When the conductivity correction value reaches the conductivity background deviation judgment limit relative to the original salinization background value, a background deviation marker is written in the corresponding time segment. The conductivity background deviation judgment limit is determined by the sensor accuracy record, the historical salinization background fluctuation range, and the project monitoring plan. Specifically, the conductivity background deviation judgment limit can be an absolute difference limit or a relative rate of change limit; the absolute difference limit is not less than twice the upper limit of the conductivity sensor measurement accuracy error, and the relative rate of change limit is determined based on the fluctuation range of the historical salinization background value under no irrigation and drainage disturbance, and is fixedly recorded in the project monitoring plan. If the project monitoring plan has specified the conductivity background deviation judgment limit, the value specified in the project monitoring plan shall be followed.

[0137] Continuous deviation refers to the deviation of the same conductivity monitoring point from the previous one. A sampling point or a series of sampling points Within each time segment, the conductivity correction value reached the conductivity background deviation judgment limit, and the deviation direction was consistent. Update the sample number to determine the background. The number should be no less than three, with the specific value determined by the sensor sampling interval, the salinity response time range, and the project monitoring plan. When the sensor sampling interval is less than a uniform time granularity... The number of continuous time segments is determined based on the conversion to a unified time granularity; when the project monitoring plan already specifies the frequency of salinity background updates. The monitoring should be determined according to the project monitoring plan; if the project monitoring plan does not specify otherwise. Take a number of sampling points that can cover at least three valid sampling points within a salt response time range.

[0138] Only in continuous A sampling point or a series of sampling points Salinization background value updates are triggered only when all time segments reach the conductivity background deviation judgment threshold, and the deviation process falls within the same response time range as irrigation, drainage, rainfall replenishment, or field drying events. After the salinization background value is updated, the salt candidate event field, water-salt disturbance release window, and salt upshift verification in subsequent prediction periods all use the updated salinization background value.

[0139] When the conductivity correction value shows a continuous deviation, and this continuous deviation occurs within the same response time range as irrigation, drainage, rainfall replenishment, or field drying events, the salinization background value of the predicted object is updated. When updating the salinization background value, the effective conductivity correction value that caused the continuous deviation is used as the update basis, and the original salinization background value, the updated salinization background value, the corresponding conductivity monitoring point number, the start and end times of the continuous deviation, and the triggering irrigation / drainage events are retained. After the salinization background value is updated, the salt candidate event field, water-salt disturbance release window, and salt upshift verification in subsequent prediction cycles all use the updated background value.

[0140] When, under the same straw return method, the actual methane flux changes occur concentrated outside the originally set carbon supply time range, the carbon supply time range corresponding to the straw decomposition state is adjusted. After the carbon supply time range is adjusted, the carbon supply candidate event field and the water-carbon superposition release window in subsequent prediction cycles will both adopt the adjusted time range. The straw return method includes straw type, return amount per unit area, plowing depth, and return date.

[0141] When a candidate irrigation and drainage scheme has been executed, the actual execution results are recorded in the irrigation and drainage scheme execution record. The execution record includes at least the candidate scheme number, actual irrigation time, actual drainage time, actual water drop duration, actual irrigation recovery time, actual irrigation water consumption, actual release window sequence, actual weekly methane emission flux, water cut control boundary verification results, salinity shift verification results, and fertile season suitability verification results. The irrigation and drainage scheme execution record is used to correct the formation results of aquifer events, aeration events, and salinity change events when replaying similar candidate irrigation and drainage schemes.

[0142] In this embodiment, the correction of the release window identification criteria includes adjusting the starting and ending identification conditions or duration segments of the water-carbon superposition release window, moisture maintenance release window, water-salt disturbance release window, and release reduction window. The correction of the window-level emission reference record retrieval method includes adjusting the order of reference record filtering fields, window duration segments, salinity correction field retrieval method, release reduction window truncation method, and output method of fields to be verified. The correction of the irrigation and drainage scheme playback results includes correcting the formation results of water layer events, aeration events, and salinity candidate events during the candidate irrigation and drainage scheme playback process based on the actual irrigation and drainage execution time, actual water layer changes, actual soil volumetric water content changes, and actual conductivity correction value changes.

[0143] After the correction is completed, the current forecast period correction record is output. The current forecast period correction record includes at least the forecast object number, forecast period number, actual weekly methane emission flux, actual methane release window sequence, window identification difference field, window-level emission reference record update content, salinization background value update content, carbon supply time range update content, irrigation and discharge playback verification record, and subsequent forecast period call flag.

[0144] Subsequent forecast cycle recall markers are used to determine the recall position of corrected records in the next consecutive seven-day forecast cycle. Subsequent forecast cycle recall markers include release window identification recall markers, reference record recall markers, irrigation / discharge replay recall markers, and retrospective verification recall markers. The release window identification recall marker is used to load the corrected release window start and end identification criteria, ventilation determination duration, salinity response time range, and carbon supply time range into the methane release window locking process of the next forecast cycle; the reference record recall marker is used to load the supplemented window-level emission reference record into the window-level emission reference record recall process of the next forecast cycle; the irrigation / discharge replay recall marker is used to load the actual irrigation / discharge execution record and irrigation / discharge response record into the candidate irrigation / discharge scheme replay process of the next forecast cycle; and the retrospective verification recall marker is used to retain the window identification difference field, sampling verification field, and source record number as the basis for manual verification or equipment maintenance verification.

[0145] When the next forecast period belongs to the same forecast object, the correction record of the current forecast period is used first. When the next forecast period belongs to an adjacent forecast object or a forecast object of the same type, only the window-level emission reference record and irrigation and drainage playback verification record that match the soil type, salinization background value range, growth period, and irrigation and drainage event type are used. Through the subsequent forecast period call marker, the correction results generated in the current forecast period can be used in the weekly event chain reconstruction, methane release window locking, window-level emission reference record call, and candidate irrigation and drainage scheme playback process of the next forecast period, instead of just being saved as historical display information.

[0146] In summary, this invention locks the methane release window based on the weekly event chain and forms a window sequence, realizing the windowed aggregation of carbon supply, water aeration, salinity changes, and sampling events in the time dimension. This is used to clarify the start and end, connection, and nullification relationships of methane release, thereby improving the process characterization accuracy of the weekly emission flux prediction object. By replaying candidate irrigation and drainage schemes back to the weekly event chain, water-saving and emission-reducing irrigation and drainage schemes are determined, enabling the re-examination of the release window sequence and prediction results by irrigation and drainage arrangements. This supports the coordinated verification of water content boundaries, salinity control, and water consumption, thereby enhancing the supporting effect of prediction results on water-saving and emission-reducing regulation of paddy fields.

[0147] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data, characterized in that, include: Acquire basic records of paddy fields, field management records, sensor monitoring records, and actual methane flux measurement records. Use seven consecutive days as the methane emission prediction cycle, and time-mark records related to moisture, salinity, temperature, straw return to the field, and irrigation and drainage to form a weekly event chain. Based on the event chain within the week, methane release windows are identified, and a methane release window sequence is formed according to the start and end times of the windows, the triggering source, the terminating source, and the window connection relationships. Based on the window type, duration, status field and window connection relationship in the methane release window sequence, the corresponding window-level emission reference record is called, and the methane weekly emission flux prediction result is determined according to the window truncation, window correction and the relationship of the output to be verified. Based on the original irrigation and drainage schedule, candidate irrigation and drainage schemes are generated. The candidate irrigation and drainage schemes are then replayed into the weekly event chain to re-form the methane release window sequence and methane weekly emission flux prediction results corresponding to the candidate schemes. Combined with changes in the release window, agronomic safety boundary, salinity control conditions and irrigation water consumption, water-saving and emission-reducing irrigation and drainage schemes are determined. After the methane emission prediction cycle ends, the actual intra-week event chain and actual methane release window sequence are generated based on the actual methane flux measurement records, actual irrigation and discharge execution records, and complete sensor monitoring records. Based on the correspondence between the actual methane release window sequence and the predicted release window sequence, the release window identification criteria, the window-level emission reference record recall method, and the irrigation and discharge scheme playback results are corrected.

2. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1, characterized in that, The formation of the intra-week event chain includes: Using paddy fields or paddy field zones under the same irrigation and drainage facilities as the prediction objects, soil volumetric moisture content sensors, soil electrical conductivity sensors, water depth sensors, soil temperature sensors, irrigation inlets, drainage outlets, straw return records, and methane flux measurement records are bound to the prediction objects. According to the uniform time granularity, the sensor acquisition time, irrigation and drainage execution time, straw return time, and methane flux measurement time are written into the same continuous seven-day prediction cycle.

3. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1 or 2, characterized in that, The formation of the intra-week event chain includes: Candidate carbon supply events are generated based on the amount of straw returned to the field, the date of straw return to the field, the depth of straw plowing and burying, and the range of straw decomposition time. Flooding events, water-induced non-aeration events, and aeration events are generated based on water depth, soil volumetric water content, saturated water content, field capacity, and water layer identification boundaries. Salinity candidate events are generated based on conductivity correction values, salinization background values, irrigation and drainage events, and response time ranges. Based on the sensor range, sampling time missing status, continuous missing time, and the time correspondence between water layer depth and irrigation / drainage records, sampling events awaiting verification are generated.

4. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1, characterized in that, The step of locking the methane release window based on the intra-week event chain, and forming the methane release window sequence according to the window start and end times, trigger source, termination source, and window connection relationship includes: When locking the methane release window, the time period in which the carbon supply candidate event coincides with the flooding event or the water-without-aeration event and the soil temperature reaches the effective temperature condition for methane production is locked as the water-carbon superposition release window. The time period covered by flooding events or submerged and unventilated events, where no carbon supply candidate events have been formed, is defined as the moisture sustaining release window; The time period formed within the response time range of salinity candidate events after irrigation, drainage, rainfall replenishment, or field drying is defined as the water and salt disturbance release window. The time period during which a ventilation event continues to reach the ventilation determination duration is locked as the release reduction window; The time period during which the sampled events cover key monitoring fields is locked as the release window for verification.

5. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 4, characterized in that, The formation of the methane release window sequence according to the window start and end times, trigger source, termination source, and window connection relationship includes: Arrange the release windows according to their start and end times; When the release reduction window is adjacent to the water-carbon superimposed release window or the water-sustaining release window, the start time of the release reduction window is used as the termination correction time of the adjacent release window. When the water-salt disturbance release window overlaps with the water-carbon superposition release window or the water maintenance release window, the water-salt disturbance release window is written into the salinity correction field of the corresponding release window. When the release window to be nucleated covers the water layer depth, soil volumetric water content, electrical conductivity correction value, or soil temperature, the corresponding release window will be marked as the release window to be nucleated.

6. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1, characterized in that, The process of calling the corresponding window-level emission reference record and determining the methane weekly emission flux prediction result based on window truncation, window correction, and the relationship between the output to be verified includes: For the water and carbon release window, the corresponding reference records are called according to the amount of straw returned to the field, the number of days after returning to the field, the duration of water status, and the soil temperature status field. For the water maintenance release window, corresponding reference records are called according to the rice growth period, the duration of water status, and the type of irrigation and drainage event; For release windows with salinity correction fields, the corresponding reference records are called according to the direction of salinity change, irrigation / drainage event type, and response time range. For the window to be tested, the previous valid window of the same prediction object, the window of the same type of adjacent prediction objects, and the reference window of the same growth period are called to form the methane flux prediction range.

7. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1, characterized in that, The process of generating candidate irrigation and drainage schemes based on the original planned irrigation and drainage schedule, and then replaying these candidate schemes into the weekly event chain, includes: When replaying candidate irrigation and drainage schemes to the weekly event chain, the original planned irrigation and drainage event fields are replaced or supplemented based on the irrigation time, drainage time, water layer target value, water drop duration, and irrigation recovery time in the candidate irrigation and drainage schemes. The water layer events, ventilation events, salinity candidate events, and methane release window sequences under the candidate scheme conditions are regenerated. The methane weekly emission flux prediction results corresponding to the candidate schemes are re-determined according to the window-level emission reference records.

8. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1, characterized in that, The method for determining water-saving and emission-reduction irrigation and drainage schemes by combining changes in the release window, agronomical safety boundaries, salinity control conditions, and irrigation water consumption includes: When determining water-saving and emission-reducing irrigation and drainage schemes, the methane release window sequence formed by the candidate irrigation and drainage schemes is verified. If the duration of the candidate irrigation and drainage scheme is shorter than the original planned duration of the water and carbon release window or the water maintenance release window, and does not exceed the minimum production safety moisture content control boundary, does not form a salt uplift judgment record, and meets the allowable operation record for the rice growth period, the candidate irrigation and drainage scheme shall be retained. If the predicted weekly methane emission flux corresponding to the candidate irrigation and drainage scheme is less than the predicted result of the original plan, and the irrigation water consumption is less than the original planned irrigation water consumption, then the scheme is determined to be a water-saving and emission-reducing irrigation and drainage scheme.

9. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 8, characterized in that, The process of generating the actual weekly event chain and actual methane release window sequence based on actual methane flux measurement records, actual irrigation and drainage execution records, and complete sensor monitoring records includes: When generating the actual intra-week event chain after the prediction period ends, the actual methane flux measurement records are assigned to the corresponding time segments according to the sampling time or flux monitoring time, the actual irrigation and drainage execution records are assigned to the corresponding time segments according to the event start time and event end time, and the complete sensor monitoring records are assigned to the corresponding time segments according to a uniform time granularity. Based on the actual carbon supply candidate events, actual flooding events, actual water-induced non-aeration events, actual aeration events, actual salinity candidate events, and actual sampling pending verification events, the actual methane release window sequence is re-formed.

10. The method for predicting weekly methane emission flux from paddy fields based on sensor monitoring data as described in claim 1 or 9, characterized in that, The criteria for identifying the corrected release window, the method for recalling window-level emission reference records, and the results of the irrigation and drainage scheme playback include: When revising the criteria for identifying release windows, the method for recalling window-level emission reference records, and the playback results of irrigation and drainage schemes, the actual methane release window sequence and the predicted release window sequence will be matched according to window type, window start time, window end time, trigger source, and termination source; When the actual methane release window sequence does not correspond to the predicted release window sequence in terms of window type, window start time, window end time, trigger source, and termination source, a window identification difference field is generated. The criteria for identifying the start and end of the release window are corrected based on the window identification difference field. Window-level emission reference records are supplemented based on actual methane flux measurement records. The formation results of water layer events, aeration events, and salinity candidate events during the playback of candidate irrigation and drainage schemes are corrected based on actual irrigation and drainage execution records.

Images