Rhizosphere sediment carbon flux real-time monitoring method using sensor array

Abstract

The application discloses a rhizosphere sedimentation carbon flux real-time monitoring method using a sensor array, relates to the technical field of soil carbon cycle monitoring, and is based on the three-dimensional distribution characteristics of target plant roots to determine a layered sampling depth and a horizontal concentric circle partition range and map to a preset three-dimensional grid coordinate system; gas sensors and environmental parameter sensors are arranged according to root partition to construct a multidimensional sensor array. The application realizes accurate spatial mapping of rhizosphere carbon flux through a three-dimensional grid coordinate system and a root density gradient model, utilizes vertical layering and horizontal concentric circle partitioning to ensure that the sensor arrangement is highly matched with the root biomass distribution, and the collaborative arrangement of the gas sensors and the environmental parameter sensors can capture root secretion pulse, microbial respiration and other transient changes, generate continuous and consistent carbon flux data sets, and provide a high-resolution dynamic observation basis for the study of plant-soil-microorganism interaction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of soil carbon cycle monitoring, in particular to a rhizosphere sedimentary carbon flux real-time monitoring method using a sensor array. BACKGROUND

[0002] Rhizosphere sedimentary carbon (also known as soil rhizosphere carbon) refers to the dynamic changes of organic carbon in the soil around plant roots, mainly including organic matter secreted by plant roots and organic matter produced by microbial activity, which is transformed and releases carbon in the soil. Rhizosphere sedimentary carbon plays an important role in global carbon cycle. Plants convert atmospheric carbon dioxide into organic carbon through photosynthesis, and transport carbon to the soil through root exudation or plant residues. The accumulation of rhizosphere sedimentary carbon is closely related to soil health, fertility and plant growth. By monitoring carbon flux in real time, agricultural management measures can be optimized to improve crop yield and soil carbon storage capacity, and reduce the negative impact of agricultural activities on the environment.

[0003] In the prior art, rhizosphere sedimentary carbon flux monitoring is of great significance to the study of carbon cycle in ecological systems. Traditional methods such as radioactive tracing and isotope labeling have safety risks, high costs, insufficient accuracy and cannot monitor in real time. Therefore, how to lay out according to the distribution characteristics of root systems, construct a multi-dimensional sensor array, and build an estimation model in the cloud for real-time monitoring and threshold alarm to improve estimation accuracy and help in-depth study of the ecological processes of plant-soil-microorganism systems is the problem to be solved by the present application. Therefore, a rhizosphere sedimentary carbon flux real-time monitoring method using a sensor array is proposed. SUMMARY

[0004] The present application aims to provide a rhizosphere sedimentary carbon flux real-time monitoring method using a sensor array to solve the problems raised in the background art.

[0005] To solve the above technical problems, the technical solution adopted by the present application is:

[0006] A rhizosphere sedimentary carbon flux real-time monitoring method using a sensor array, comprising the following steps:

[0007] S1, based on the three-dimensional distribution characteristics of the target plant root system, determine the layered sampling depth and horizontal concentric circle partition range, and map them to the preset three-dimensional grid coordinate system to provide spatial basis for sensor layout;

[0008] S2, deploy gas sensors and environmental parameter sensors according to the root zone, construct a multi-dimensional sensor array, and form a multi-dimensional monitoring network;

[0009] S3, real-time synchronous collection of gaseous carbon flux and environmental parameter data by various types of sensors of the multi-dimensional monitoring network, to ensure the consistency of spatial and temporal data;

[0010] S4, transmission of gaseous carbon flux and environmental parameter data collected by various sensors to the cloud for preprocessing, the preprocessing process including data filtering and noise reduction and correction of sensor drift, to improve data reliability;

[0011] S5, based on the historical gaseous carbon flux and environmental parameter data after preprocessing, a rhizosphere sediment carbon flux estimation model is built in the cloud;

[0012] S6, combining the rhizosphere sediment carbon flux estimation model with the current gaseous carbon flux and environmental parameter data, the carbon flux trend is displayed, and when the abnormal threshold is triggered, an abnormal alarm is pushed to the user terminal;

[0013] S7, according to the long-term monitoring data and alarm information, iteratively update the model parameters and sensor layout scheme, continuously improve the estimation accuracy of rhizosphere sediment carbon flux and ecological research value.

[0014] Further improvement of the technical scheme of the application is that the S1 specifically comprises:

[0015] Based on the root system configuration (taproot system or fibrous root system) of the target plant, the vertical and horizontal distribution density of the root system is quantified by non-destructive imaging technology, the key parameters including the main root depth, lateral root extension radius and root tip active area range are extracted, and a root system density gradient model is established;

[0016] According to the root system density gradient model, the rhizosphere region is divided into vertical layers and horizontal concentric circle partitions, and the vertical layer depth and horizontal concentric circle partition radius are determined, wherein the vertical layer depth and horizontal concentric circle partition radius are dynamically adjusted according to the root biomass decay threshold, to ensure that the sensors cover the key variation region of carbon flux;

[0017] A three-dimensional grid coordinate system is preset, and the layering and partitioning results are mapped into the three-dimensional grid coordinate system, the deployment positions of various sensor nodes including gas sensors and environmental parameter sensors are determined, and are marked in the three-dimensional coordinate grid.

[0018] Further improvement of the technical scheme of the application is that the process of dividing the rhizosphere region into vertical layers and horizontal concentric circle partitions, and determining the vertical layer depth and horizontal concentric circle partition radius is:

[0019] The vertical layering includes a surface layer, a middle layer and a deep layer, the vertical layer depth is determined by the vertical decay threshold of the root biomass, the surface layer depth is 0-15 cm, corresponding to the high metabolic activity area, the middle layer depth is 15-30 cm, corresponding to the monitoring of carbon migration and transformation, and the deep layer depth is > 30 cm, focusing on the monitoring of steady-state carbon flux;

[0020] The horizontal concentric circle partition includes a near root area and a far root area, and the horizontal concentric circle partition radius is dynamically adjusted according to the radial decrease of the lateral root biomass, the near root area radius is less than or equal to 50% of the lateral root extension radius, and the far root area radius is greater than 50% of the lateral root radius to 1.5 times the radius, so that the near root area covers the main release range of root exudates.

[0021] Further improvement of the technical scheme of the application is that in S2, the process of constructing a multi-dimensional sensor array to form a multi-dimensional monitoring network is:

[0022] Based on the vertical stratification and horizontal concentric circle partition of the root system, the monitoring requirements of each area are determined to select the sensor type, wherein the gas sensor includes a CO2 sensor and a CH4 sensor, and the environmental parameter sensor includes a temperature and humidity sensor, a pH sensor and a soil conductivity sensor;

[0023] The selected gas sensor and environmental parameter sensor are deployed to the corresponding position according to the three-dimensional coordinate by using the grid layout, so as to ensure good contact with the soil, and the sensors are connected by a wireless networking mode, a multi-dimensional sensor array is constructed, a multi-dimensional monitoring network covering the rhizosphere is formed, and synchronous data acquisition is realized.

[0024] Further improvement of the technical scheme of the application is that S3 specifically includes:

[0025] All sensor nodes are time-synchronized at the millisecond level through the GPS protocol, so as to ensure the uniformity of the data timestamp, and based on the preset three-dimensional grid coordinate system, each sensor is assigned a unique spatial identifier (X / Y / Z+partition ID), and when physically laid out, a laser range finder is used to calibrate the position deviation, so as to realize the alignment of the time and space reference of data acquisition;

[0026] The multi-dimensional monitoring network is started to trigger the gas sensor and the environmental parameter sensor to sample synchronously at a fixed interval, a hardware interrupt mechanism is used to avoid transmission delay, a uniform time-space label is attached to the data packet, data synchronous transmission is performed through a wireless protocol networking (LoRaWAN), and the time and space consistency of the data collected by different sensors is ensured;

[0027] After the cloud receives the data, the data is sorted according to the timestamp and the gaps caused by slight transmission delay are filled, and then the spatial coordinates are mapped into the three-dimensional grid coordinate system to generate a multi-dimensional data set aligned in time and space.

[0028] Further improvement of the technical scheme of the application is that S4 specifically includes:

[0029] The server of the cloud receives the data packet transmitted by each sensor, including the gaseous carbon flux and the environmental parameter data, and parses the time-space label and the original sensor data, and classifies and integrates the sensor data according to the sensor type;

[0030] The original sensor data is filtered by using digital signal processing technology, wherein high-frequency noise is eliminated by a low-pass filter, short-term fluctuations are suppressed by a sliding window smoothing algorithm, and dynamic threshold denoising is performed on CO2 and CH4 data in combination with environmental parameters to eliminate abnormal pulse signals;

[0031] Based on the laboratory calibration data and the preset field reference sensor, a drift compensation model is established to calibrate the baseline of the CO2 sensor, linear regression is used to correct long-term deviation, and the pH and soil conductivity sensors are processed by regular automatic cleaning and standard solution verification to reduce the influence of electrode aging.

[0032] The further improvement of the technical scheme of the application is that the S5 specifically comprises:

[0033] The pre-processed historical gaseous carbon flux and environmental parameter data are collected, multi-dimensional features related to rhizosphere sedimentary carbon flux are extracted, including the time series change rate of gaseous carbon flux (CO2, CH4), the gradient difference of environmental parameters (temperature, humidity, pH, soil conductivity), and the spatial correlation of root zone partition (vertical stratification, horizontal concentric circle), and then a comprehensive data set containing time and space labels is constructed, and the comprehensive data set is divided into a training set and a verification set;

[0034] A hybrid model architecture of LSTM (processing time series dynamics) combined with random forest (analyzing non-linear relationships of environmental factors) is selected to build a rhizosphere sedimentary carbon flux estimation model, the hybrid model architecture is trained using the training set, and the hyperparameters are optimized using the verification set by cross-validation, and then the trained rhizosphere sedimentary carbon flux estimation model is obtained;

[0035] The trained rhizosphere sedimentary carbon flux estimation model is packaged as a cloud API to receive real-time sensor data streams and output the calculated rhizosphere sedimentary carbon flux estimation value.

[0036] The further improvement of the technical scheme of the application is that the calculation process of the rhizosphere sedimentary carbon flux estimation value is:

[0037] The CO2 concentration instantaneous value at the current time and the concentration difference value of the previous period (5 minutes ago) are obtained, the CO2 concentration change rate is obtained by dividing the time interval, the CO2 change rate is temperature corrected by using an exponential function, the index of the exponential function is the difference between the current soil temperature and the reference temperature divided by 10, which is used to normalize the biological activity at the actual environmental temperature to the reference temperature level, and then the respiration and temperature coupling term is obtained by integrating the CO2 concentration change rate and the temperature correction result.

[0038] The ratio of the CH4 concentration to the CO2 concentration is calculated, a natural logarithm is taken after adding 1, and a relative ratio of the current soil conductivity is calculated based on a reference soil conductivity; the ratio is taken as a base, a difference between the pH value and a reference pH value is taken as an index for power operation, and then the natural logarithm and the power operation results are integrated to obtain an anaerobic metabolism and soil environment interaction term;

[0039] The result of the respiration and temperature coupling term is added to the result of the anaerobic metabolism and soil environment interaction term to obtain an estimated value of the rhizosphere sedimentary carbon flux.

[0040] The further improvement of the technical scheme of the present application is that the S6 specifically comprises:

[0041] The cloud receives a sensor data stream, parses a time stamp, a spatial coordinate and an environmental parameter, maps the data according to a preset three-dimensional grid, performs standardization preprocessing, and calls a rhizosphere sedimentary carbon flux estimation model to calculate a current estimated value of the rhizosphere sedimentary carbon flux;

[0042] The estimated values of the rhizosphere sedimentary carbon flux in three consecutive fixed time periods are analyzed to generate a change trend curve, and the change trend of the estimated value of the rhizosphere sedimentary carbon flux is analyzed to determine its rising, falling or stable state;

[0043] The current estimated value of the rhizosphere sedimentary carbon flux is compared with a preset abnormal threshold value, if the current estimated value exceeds the abnormal threshold value, it is determined as an abnormal situation, and an abnormal alarm information is automatically generated, and an alarm notification containing abnormal information, time, location and change trend of the estimated value of the rhizosphere sedimentary carbon flux is pushed to a user terminal.

[0044] The further improvement of the technical scheme of the present application is that the S7 specifically comprises:

[0045] Based on historical monitoring data and alarm records, the spatio-temporal distribution characteristics of model estimation error are analyzed, the hidden layer structure of the LSTM and the feature weight of the random forest are adjusted by using Bayesian optimization for the high-frequency error area, the reference benchmark of the drift compensation model is updated synchronously, the influence of sensor aging is corrected by using long-term calibration data, and the model is dynamically adapted to local environmental changes;

[0046] In combination with the spatial clustering result of the abnormal event, a monitoring blind area or a redundant node is identified, the sensor density is increased in the area where the rhizosphere sedimentary carbon flux rises, the node density is reduced in the steady state area, the horizontal and vertical arrangement ratio is optimized, and the node spacing and relay position of the wireless protocol networking are reconfigured to balance the data quality and network life;

[0047] The optimized rhizosphere deposition carbon flux estimation model and the layout scheme are applied to the new cycle monitoring, the deviation of the rhizosphere deposition carbon flux estimation value and the laboratory measured value is cross-validated, the coupling law of the root exudates-microbial metabolism-environmental factors is revealed through long-term trend analysis, and then the verification conclusion is fed back to the model optimization and layout adjustment, a closed-loop improvement system is formed, and the ecological mechanism analysis ability is continuously improved.

[0048] Due to the adoption of the above technical solutions, the technical progress achieved by the present application relative to the prior art is:

[0049] 1. The present application provides a real-time monitoring method for rhizosphere deposition carbon flux using a sensor array, which realizes accurate spatial mapping of rhizosphere carbon flux through a three-dimensional grid coordinate system and a root density gradient model, ensures high matching between sensor layout and root biomass distribution by vertical stratification and horizontal concentric circle partitioning, and can capture transient changes such as root exudate pulses and microbial respiration through the coordinated deployment of gas sensors and environmental parameter sensors, thereby generating continuous and consistent carbon flux data sets and providing a high-resolution dynamic observation basis for studying plant-soil-microbial interactions.

[0050] 2. The present application provides a real-time monitoring method for rhizosphere deposition carbon flux using a sensor array, which builds a rhizosphere deposition carbon flux estimation model in the cloud by integrating multi-dimensional data of gaseous carbon flux and environmental parameters, combining a hybrid model architecture of LSTM and random forest, and capturing time series dynamics and nonlinear environmental responses respectively, thereby significantly reducing estimation errors through multi-source data collaboration compared to single sensors or models, and revealing the regulation mechanism of environmental thresholds on carbon flux. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0052] Figure 1 A workflow schematic diagram of the real-time monitoring method for rhizosphere deposition carbon flux using a sensor array of the present application;

[0053] Figure 2 A method flow schematic diagram of the real-time monitoring method for rhizosphere deposition carbon flux using a sensor array of the present application. DETAILED DESCRIPTION

[0054] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0055] As shown in Embodiment 1, Figure 1 , Figure 2 The present application provides a rhizosphere sedimentary carbon flux real-time monitoring method using a sensor array, comprising the following steps:

[0056] S1, based on the three-dimensional distribution characteristics of the target plant root system, the stratified sampling depth and the horizontal concentric circle partition range are determined and mapped into a preset three-dimensional grid coordinate system, providing a spatial basis for sensor layout, based on the root system configuration (taproot system or fibrous root system) of the target plant, the distribution density of the root system in the vertical and horizontal directions is quantified by non-destructive imaging technology, the key parameters including the main root depth, lateral root extension radius and root tip active area range are extracted, the root system density gradient model is established, the rhizosphere area is divided into vertical stratification and horizontal concentric circle partition according to the root system density gradient model, and the vertical stratification depth and the horizontal concentric circle partition radius are determined, wherein the vertical stratification depth and the horizontal concentric circle partition radius are dynamically adjusted according to the root biomass decay threshold value to ensure that the sensors cover the key variation area of carbon flux, a three-dimensional grid coordinate system is preset, and the stratified partition results are mapped into the three-dimensional grid coordinate system, the deployment positions of each sensor node including gas sensors and environmental parameter sensors are determined, and are marked in the three-dimensional coordinate grid, wherein the gas sensors are preferentially arranged in the high biomass area of the root-soil interface, and the environmental parameter sensors are uniformly distributed to capture the gradient change, and the layout scheme needs to be compatible with the plant growth dynamics and reserve adjustable space;

[0057] In addition, the process of dividing the rhizosphere area into vertical stratification and horizontal concentric circle partition, and determining the vertical stratification depth and the horizontal concentric circle partition radius is as follows:

[0058] The vertical stratification includes the surface layer, the middle layer and the deep layer, the vertical stratification depth is determined by the vertical decay threshold value of the root biomass, the surface layer depth is 0-15 cm, corresponding to the high metabolic activity area, the middle layer depth is 15-30 cm, corresponding to the monitoring of carbon migration and transformation, and the deep layer depth is > 30 cm, focusing on the monitoring of steady-state carbon flux; the horizontal concentric circle partition includes the near-root area and the far-root area, the horizontal concentric circle partition radius is dynamically adjusted according to the radial decrease of the lateral root biomass, the near-root area radius ≤ 50% of the lateral root extension radius, and the far-root area radius > 50% of the lateral root radius to 1.5 times the radius, to ensure that the near-root area covers the main release range of root exudates;

[0059] Specific work content is: based on the root system configuration of target plant taproot system or fibrous root system, using non-destructive imaging technology (X-ray tomography or laser confocal microscopy) to quantify the distribution density of root system in vertical and horizontal directions, extracting key parameters such as main root depth, lateral root extension radius and root tip active area range through image analysis algorithm, establishing root density gradient model, taking root length density as index to describe the distribution law of root system in three-dimensional space, among them, taproot system (main root dominated) adopts X-ray tomography (X-ray CT), resolution ≤50μm, penetration depth >1m, identifying the vertical trend of main root and the branching angle of lateral root; fibrous root system (fibrous root) adopts laser confocal microscopy (CLSM) + fluorescence staining, resolution ≤1μm, quantifying the spatial topology of fine root (diameter <0.2mm); according to the root density gradient model, the rhizosphere region is divided into vertical stratification (surface layer, middle layer, deep layer) and horizontal concentric circle partition (near root area, far root area), among them, the surface layer corresponds to the high metabolic activity area, which is the most vigorous area of root growth and metabolic activity, and the carbon release is relatively active, the middle layer and the deep layer focus on carbon migration and transformation and steady carbon flux monitoring, reflecting the long-term carbon fixation and release of root system in deeper soil layer, in horizontal direction, the near root area needs to ensure covering the main release range of root exudates, in order to accurately capture the carbon flux change produced by the interaction between root system and soil microorganisms, the far root area is used as a control area for comparative analysis of carbon flux difference between rhizosphere and non-rhizosphere environment; the stratification and partition results are mapped to the preset three-dimensional grid coordinate system, and the theoretical deployment positions of each sensor node including gas sensor and environmental parameter sensor are marked, among them, the gas sensor is preferentially arranged in the high biomass area of root-soil interface to capture carbon release hotspots, and the environmental parameter sensor is uniformly arranged according to the gradient distribution principle, covering the rhizosphere and non-rhizosphere transition zone, the deployment scheme needs to reserve extensible interface to support subsequent dynamic adjustment of node position according to plant growth or model feedback, such as adding deep monitoring points, grid coordinates are recorded in cloud at the same time, realizing the spatial correlation of sensor position and monitoring data;

[0060] S2, the gas sensor and the environmental parameter sensor are deployed according to root system zoning, a multi-dimensional sensor array is constructed, and a multi-dimensional monitoring network is formed, wherein the gas sensor includes a CO2 sensor and a CH4 sensor, the environmental parameter sensor includes a temperature and humidity sensor, a pH sensor and a soil conductivity sensor, based on the zoning characteristics of vertical stratification and horizontal concentric circle zoning of root system, the monitoring requirements of each region are determined to select the type of sensor, wherein the gas sensor includes a CO2 sensor and a CH4 sensor, the environmental parameter sensor includes a temperature and humidity sensor, a pH sensor and a soil conductivity sensor, the CO2 sensor (NDIR technology, ±50ppm) is preferentially deployed in the high-biomass root-soil interface, the CH4 sensor (TDLAS technology) focuses on anaerobic areas, and the environmental sensor (temperature and humidity, pH, soil conductivity) covers the rhizosphere and non-rhizosphere transition zone according to the gradient, so as to ensure that the spatial resolution matches the root density decay law, a grid layout is adopted, the selected gas sensor and environmental parameter sensor are deployed to the corresponding position according to the three-dimensional coordinate, good contact with the soil is ensured, and each sensor is connected through wireless networking to construct a multi-dimensional sensor array and form a multi-dimensional monitoring network covering the rhizosphere, and synchronous data acquisition is realized;

[0061] Specific work content: based on the zoning characteristics of vertical stratification (surface layer, middle layer, deep layer) and horizontal concentric circle zoning (near root area, far root area) of root system, the monitoring requirements of each region and the sensor deployment strategy need to match the spatial differentiation of root ecological function, wherein in the vertical direction, the surface layer (0-15cm) is as a high metabolic activity area, a high-precision CO2 sensor (NDIR technology) needs to be preferentially deployed to capture the transient changes of root exudates and microbial respiration; the middle layer (15-30cm) focuses on carbon migration process monitoring, the CO2 sensor density can be reduced, and the CH4 sensor (TDLAS technology) is supplemented to detect potential anaerobic microenvironment; the deep layer (>30cm) mainly monitors the steady flux, and the environmental parameter sensor (temperature and humidity, pH, soil conductivity) is arranged according to the gradient to cover the rhizosphere-non-rhizosphere transition zone; in the horizontal direction, the near root area needs to densely deploy the gas sensor, and the far root area establishes a background reference through the environmental parameter sensor to ensure that the spatial resolution is consistent with the root biomass decay gradient; a three-dimensional grid coordinate system is adopted, the selected sensors are mapped to the theoretical position according to the vertical stratification and horizontal concentric circle zoning, and the physical layout is matched with the spatial model of root density, wherein the gas sensor (CO2, CH4) is preferentially embedded in the high-biomass area of the root-soil interface, and when installed, the probe needs to be ensured to be in close contact with the soil to avoid air gap interference; the environmental parameter sensor is uniformly distributed according to the gradient, the temperature and humidity sensor and the pH / soil conductivity sensor are cooperatively arranged to reveal the coupling effect of environmental factors on carbon flux, all sensors are networked through a low-power wireless protocol (LoRaWAN), the node spacing is optimized according to the signal strength and power consumption, and time stamp calibration is synchronously adopted to ensure the spatio-temporal consistency of multi-source data;

[0062] S3, real-time synchronous collection of gaseous carbon flux and environmental parameter data by various types of sensors of the multi-dimensional monitoring network, ensuring spatial and temporal data consistency, millisecond-level time synchronization of all sensor nodes through GPS protocol, ensuring uniformity of data timestamps, based on a preset three-dimensional grid coordinate system, assigning a unique spatial identifier (X / Y / Z+partition ID) to each sensor, using a laser range finder to calibrate positional deviation during physical layout, achieving spatial and temporal reference alignment of data collection, starting the multi-dimensional monitoring network, triggering gaseous sensors and environmental parameter sensors to sample synchronously at fixed intervals, using a hardware interrupt mechanism to avoid transmission delays, adding a uniform time-space label to data packets, synchronously transmitting data through wireless protocol networking (LoRaWAN), ensuring spatial and temporal consistency of data collected by different sensors, after the cloud receives the data, sorting by timestamp and filling in the gaps caused by minor transmission delays, and then mapping the data to a three-dimensional grid coordinate system according to the spatial coordinates, generating a multi-dimensional data set with spatial and temporal alignment;

[0063] The specific work content is: high-precision time synchronization of all sensor nodes, millisecond-level time calibration of all nodes using GPS protocol (such as PPS pulse signal), eliminating clock drift between devices, ensuring uniformity of data timestamps, and using a preset three-dimensional grid coordinate system to assign a unique spatial identifier (X / Y / Z coordinates and partition ID) to each sensor, clearly defining its vertical stratification and horizontal concentric circle partition, using a laser range finder to calibrate the deviation (error controlled within ±1 cm) between the actual position of the sensor and the theoretical coordinates during physical layout, ensuring that the spatial distribution of the sensor network matches the root density gradient model; after starting the multi-dimensional monitoring network, all sensors are triggered to collect data synchronously at fixed sampling intervals, at the same time, a hardware interrupt mechanism is used to avoid communication delays, ensuring that the sampling times of gaseous sensors and environmental parameter sensors are aligned, each data packet is attached with a uniform time-space label (timestamp + three-dimensional coordinates), and is transmitted to the cloud through wireless protocol networking (LoRaWAN), wherein the embedding of the time-space label enables the data of different sensors to be accurately aligned in time and space dimensions in the cloud, avoiding data misplacement caused by transmission delays; after the cloud receives the data, it is sorted by timestamp and the gaps caused by minor transmission delays are filled in by linear interpolation, the data is mapped to a three-dimensional grid coordinate system according to the spatial coordinates of the sensors, generating a multi-dimensional data set with spatial and temporal alignment, including gaseous carbon flux, environmental parameters and corresponding position information;

[0064] S4, transmit the gaseous carbon flux and environmental parameter data collected by each sensor to the cloud for preprocessing, the preprocessing process including data filtering and noise reduction and correction of sensor drift, improving data reliability, the server of the cloud receiving the data packet including the gaseous carbon flux and environmental parameter data transmitted by each sensor, and analyzing the time-space label and the original sensor data, classifying and integrating according to the sensor type, using digital signal processing technology, filtering the original sensor data, wherein high-frequency noise is eliminated by a low-pass filter, short-term fluctuations are suppressed by a sliding window smoothing algorithm, for CO2 and CH4 data, dynamic threshold denoising is performed in combination with environmental parameters, abnormal pulse signals are removed, a drift compensation model is established based on laboratory calibration data and a preset reference sensor on site, the CO2 sensor is baseline calibrated, long-term deviation is corrected by linear regression, the pH and soil conductivity sensors are processed by regular automatic cleaning and standard solution verification to reduce the influence of electrode aging;

[0065] The specific work content is that the server of the cloud receives the data packet from the sensor node through the wireless communication protocol, each data packet contains a time stamp, a three-dimensional space coordinate, a partition identification and an original sensor reading, the server analyzes the data packet, extracts the time-space label, ensures the time synchronization and spatial uniqueness of the data, and stores them according to the sensor type, and establishes a structured database, wherein the time stamp adopts an international standard format, the space coordinate is mapped to a preset three-dimensional grid coordinate system, the partition identification is associated with the vertical stratification (surface layer, middle layer, deep layer) and horizontal partition (near-root zone, far-root zone) of the root, after data classification, a multidimensional data set is generated, and the data source node ID is recorded at the same time; the original sensor data is processed by digital signal processing technology, high-frequency noise is eliminated by a low-pass filter (Butterworth filter), the cutoff frequency is dynamically adjusted according to the sensor characteristics, short-term fluctuations are suppressed by a sliding window smoothing algorithm (window size 5-10 sampling points) to retain the trend changes, for CO2 and CH4 data, dynamic threshold denoising is performed in combination with environmental parameters: if the data point exceeds the range of historical mean value ± 3 times standard deviation and does not match the environmental parameter change, it is determined as an abnormal pulse signal and is removed, the filtered data retains the time-space label to ensure the spatio-temporal consistency of subsequent analysis; based on the laboratory calibration data and the preset reference sensor (deployed in the non-rhizosphere area), a drift compensation model is established, for the CO2 sensor, the baseline deviation is corrected by linear regression, the deviation between the sensor output and the reference value is fitted regularly, the calibration coefficient is dynamically updated, the pH and soil conductivity sensors reduce the attachment of pollutants by an automatic cleaning mechanism, and are regularly injected with standard solution for online verification, the measurement deviation caused by electrode aging is corrected, the calibrated data and the original value are stored together, the calibration time and method are labeled to ensure data traceability, the sensor health status is automatically recorded, and a maintenance alarm is triggered to maintain long-term monitoring accuracy;

[0066] S5, based on the pre-processed historical gaseous carbon flux and environmental parameter data, a rhizosphere sediment carbon flux estimation model is built in the cloud;

[0067] S6, combining the rhizosphere sediment carbon flux estimation model with the current gaseous carbon flux and environmental parameter data, the carbon flux change trend is displayed, and when the abnormal threshold is triggered, an abnormal alarm is pushed to the user terminal;

[0068] S7, according to the long-term monitoring data and alarm information, the model parameters and sensor layout scheme are iteratively updated, and the rhizosphere sediment carbon flux estimation accuracy and ecological research value are continuously improved.

[0069] Embodiment 2, as shown in Figure 1 , Figure 2 The application provides a technical solution based on embodiment 1: preferably, S5 specifically includes:

[0070] The pre-processed historical gaseous carbon flux and environmental parameter data are collected, and the multi-dimensional features related to the rhizosphere sediment carbon flux are extracted, including the time series change rate of gaseous carbon flux (CO2, CH4), the gradient difference of environmental parameters (temperature and humidity, pH, soil conductivity), and the spatial correlation of root zone partition (vertical stratification, horizontal concentric circle). Then, a comprehensive data set containing time-space labels is constructed, and the comprehensive data set is divided into a training set and a validation set. A hybrid model architecture of LSTM (processing time series dynamics) combined with random forest (analyzing non-linear relationships of environmental factors) is selected to build a rhizosphere sediment carbon flux estimation model. The hybrid model architecture is trained using the training set, and the hyperparameters are optimized using the validation set through cross-validation. The evaluation indicators include root mean square error and spatial correlation coefficient to ensure that the model has robust prediction ability in time dimension and spatial partition. Then, the trained rhizosphere sediment carbon flux estimation model is obtained. The trained rhizosphere sediment carbon flux estimation model is encapsulated as a cloud API to receive real-time sensor data stream and output the calculated rhizosphere sediment carbon flux estimation value;

[0071] In addition, the calculation process of the rhizosphere sediment carbon flux estimation value is:

[0072] The CO2 concentration instantaneous value at the current time and the concentration difference value of the previous period (5 minutes ago) are obtained, and the CO2 concentration change rate is obtained by dividing the time interval, and the exponential function is used to correct the CO2 change rate, wherein the CO2 concentration change rate reflects the instantaneous intensity of root exudates and microbial respiration, and a positive value indicates that the carbon release is accelerated, and a negative value reflects the temporary pause of photosynthetic product transport, and since the temperature is increased by 10°C, the microbial metabolic rate is increased by about 2-3 times, so the index of the exponential function is the difference between the current soil temperature and the reference temperature divided by 10, which is used to normalize the biological activity at the actual environmental temperature to the reference temperature level, and then the CO2 concentration change rate and the temperature correction result are integrated to obtain the respiration and temperature coupling term; the ratio of CH4 concentration to CO2 concentration is calculated, 1 is added, and the natural logarithm is taken, and the relative ratio of the current soil conductivity is calculated based on the reference soil conductivity, and the ratio is taken as the base, and the difference between the pH value and the reference pH value is taken as the index to perform power operation, and then the natural logarithm and the power operation result are integrated to obtain the anaerobic metabolism and soil environment interaction term, wherein the ratio of CH4 concentration to CO2 concentration itself reflects the redox state of the soil, and the larger the value, the more significant the anaerobic metabolism, and when the pH value is higher than the reference pH value, the promoting effect of the soil conductivity on the carbon flux is amplified; when the pH value is lower than the reference pH value, the effect of the soil conductivity is inhibited, reflecting the synergistic regulation of the soil ionic strength and the acid-base degree on the microbial electron transfer chain; the results of the respiration and temperature coupling term and the results of the anaerobic metabolism and soil environment interaction term are added to obtain the rhizospheric sedimentation carbon flux estimation value, the respiration and temperature coupling term dominates the carbon release under normal aerobic conditions, and the contribution of the anaerobic metabolism and soil environment interaction term increases in anaerobic environments such as water accumulation or dense soil layers, and then when the soil is completely aerobic (CH4≈0), the anaerobic metabolism and soil environment interaction term approaches 0, and the rhizospheric sedimentation carbon flux estimation value is determined by the respiration and temperature coupling term; when the temperature drops suddenly or the CO2 change rate is negative, the respiration and temperature coupling term can be close to 0, and the rhizospheric sedimentation carbon flux estimation value reflects the background anaerobic process;

[0073] The calculation expression of the rhizospheric sedimentation carbon flux estimation value is as follows:

[0074] ;

[0075] In the formula, is the rhizospheric sedimentation carbon flux estimation value, the unit is μgC·cm -2 ·h -1 , which represents the total carbon release per unit area per unit time; is the instantaneous value of the carbon dioxide concentration, the unit is ppm, which is the main control factor and reflects the root respiration and microbial decomposition intensity; is the CO2 concentration time change rate, the unit is ppm / min, which is used to capture the pulse type secretion event; is the soil temperature, the unit iso C, adjusting microbial activity by Arrhenius effect; Reference temperature (take 15 o C, unit is o C, indicating the reference of normalized temperature effect; Methane concentration, unit is ppb, indicating anaerobic metabolic process, and CO2 ratio reflects redox state; Soil conductivity, unit is μS / cm, related to ionic strength, affecting microbial membrane transport efficiency; Reference soil conductivity (take 100 μS / cm), unit is μS / cm, reference of normalized soil conductivity effect; Soil pH, affecting carbon conversion pathway by regulating enzyme activity; Reference pH (take 6.5), indicating neutral pH reference;

[0076] The specific work content is: based on the pretreated historical data, focusing on gaseous carbon flux and environmental parameter data, extracting multi-dimensional features related to rhizosphere sediment carbon flux, in time dimension, calculating the time series change rate of gaseous carbon flux, in space dimension, analyzing the gradient difference of environmental parameters, quantifying the transition characteristics of rhizosphere and non-rhizosphere, combining with the root zone information, constructing a spatial correlation matrix to represent the carbon flux interaction of different partitions, and then integrating time-space labels to form a comprehensive data set containing gaseous flux, environmental factors and partition attributes, and dividing them into training set and verification set in time sequence to ensure the time sequence continuity of data distribution; a mixed architecture combining LSTM and random forest is used to build a model to estimate the rhizosphere sediment carbon flux, wherein the LSTM network processes the time series dynamics of gaseous flux, captures the periodicity of root exudates release, and the random forest analyzes the nonlinear relationship between environmental parameters and carbon flux, quantifies the contribution weight of different partition environmental factors, in the training stage, the time series cross-validation is used to optimize the hyperparameters, the evaluation indicators include root mean square error in time dimension and correlation coefficient in spatial partition, through joint training, the model maintains robustness in time series prediction and spatial partition, and the constructed rhizosphere sediment carbon flux estimation model is output; the trained model is packaged as a cloud RESTful API to receive real-time sensor data stream (including timestamp, spatial coordinates and environmental parameters), the API internally performs standardization preprocessing, and calls the rhizosphere sediment carbon flux estimation model to estimate the carbon flux, and the output result is the estimated value of rhizosphere sediment carbon flux;

[0077] S6 specifically includes:

[0078] The cloud receives the sensor data stream, parses the time stamp, spatial coordinates and environmental parameters, maps the data according to a preset three-dimensional grid, performs standardization preprocessing, and calls a rhizosphere sediment carbon flux estimation model to calculate the current rhizosphere sediment carbon flux estimation value, analyzes the rhizosphere sediment carbon flux estimation values of three consecutive fixed time periods, generates a change trend curve, analyzes the change trend of the rhizosphere sediment carbon flux estimation value, judges its rising, falling or stable state, compares the current rhizosphere sediment carbon flux estimation value with a preset abnormal threshold value, if it exceeds the abnormal threshold value, it is determined as an abnormal situation, and an abnormal alarm information is automatically generated, and an alarm notification containing abnormal information, time, location and change trend of rhizosphere sediment carbon flux estimation value is pushed to the user terminal;

[0079] The specific work content is: the server of the cloud receives the data stream from the sensor node through the wireless communication protocol, each data packet contains time stamp, three-dimensional spatial coordinates, environmental parameters (and node identifier), the server parses the data packet, extracts time-space label, ensures the time synchronization and spatial uniqueness of the data, the data is subjected to standardization preprocessing, including filtering and noise reduction, sensor drift compensation and environmental parameter cross verification, to eliminate short-term interference and hardware errors, after preprocessing, the rhizosphere sediment carbon flux estimation model is called, the current gaseous carbon flux and environmental parameter data are combined, and the rhizosphere sediment carbon flux estimation value is calculated in real time; the server performs dynamic trend calculation on the rhizosphere sediment carbon flux estimation values of three consecutive fixed periods based on time series analysis algorithm, eliminates high-frequency noise through sliding window smoothing processing, extracts the trend change of carbon flux, calculates the flux difference value of adjacent time points, and combines linear regression analysis slope k to judge the current trend state: if k>0 and the flux difference value is continuously positive, it is determined as an upward trend; if k<0 and the flux difference value is continuously negative, it is determined as a downward trend; if k tends to 0 and the flux difference value fluctuates within a preset tolerance range, it is marked as a stable state, and the trend analysis result is visualized in the form of a curve, and the trend duration is recorded; compare the current rhizosphere sediment carbon flux estimation value with the preset abnormal threshold value, the threshold value is dynamically set according to historical data statistics (mean ± 3 times standard deviation), if the rhizosphere sediment carbon flux estimation value exceeds the abnormal threshold value, the abnormal detection algorithm is triggered immediately, the cooperative change of environmental parameters is verified, the sensor failure or transient interference is excluded, after confirming the abnormality, a structured alarm information is generated, including abnormal type, occurrence time, three-dimensional coordinates, current flux value and change trend curve, the alarm information is pushed to the user terminal (mobile phone APP / mailbox) through the message queue, and at the same time, the system automatically records the abnormal event to the log library;

[0080] S7 specifically comprises:

[0081] Based on historical monitoring data and alarm records, the spatio-temporal distribution characteristics of model estimation error are analyzed. For high-frequency error areas, Bayesian optimization is used to adjust the hidden layer structure of LSTM and the feature weight of random forest, and the reference benchmark of the drift compensation model is updated synchronously. Long-term calibration data is used to correct the influence of sensor aging, ensuring that the model dynamically adapts to local environmental changes. Combined with the spatial clustering results of abnormal events, monitoring blind spots or redundant nodes are identified. In areas with rising rhizosphere sediment carbon flux, the density of sensors in the vertical and horizontal directions is increased to improve spatial resolution. In stable areas, the density of nodes is reduced to reduce energy consumption. At the same time, the node spacing and relay position of the LoRaWAN network are reconfigured to ensure communication reliability in high-frequency data areas. The optimized layout scheme is deployed accurately through a three-dimensional grid coordinate system, ensuring spatial matching between new nodes and the root density gradient model, and preserving interfaces for subsequent iterative adjustments. The optimized rhizosphere sediment carbon flux estimation model and layout scheme are applied to the new monitoring period. Through cross-validation, the deviation between model estimated values and laboratory measured values is compared to quantify the improvement effect. Long-term trend analysis reveals the coupling rules of root exudates, microbial metabolic activity, and environmental factors, and the verification conclusions are fed back to the model optimization and network adjustment process, forming a closed-loop improvement system of monitoring-analysis-optimization-verification. Multidimensional ecological data is continuously accumulated, and mechanism models of plant-soil-microorganism interactions are gradually constructed to improve the ability to analyze key processes of carbon cycling.

[0082] Specific work content: Based on historical monitoring data and alarm records, the spatio-temporal distribution characteristics of model estimation error are analyzed. High-frequency error areas and their potential causes are identified. For time dimension errors, Bayesian optimization algorithm is used to adjust the hidden layer structure of LSTM network to optimize its time response ability to root exudate pulse events. For spatial dimension errors, the weights of environmental parameters in the random forest model are redistributed, and the reference benchmark value of the drift compensation model is updated synchronously. Long-term calibration data is used to correct the baseline shift caused by sensor aging, ensuring that model parameters dynamically adapt to local environmental changes. Combined with the spatial clustering results of abnormal events, monitoring blind spots or redundant nodes are identified. The sensor layout scheme is dynamically optimized. In hotspots with rising rhizosphere sediment carbon flux, the density of sensors in the vertical and horizontal directions is increased to improve spatial resolution. In stable areas, the density of nodes is reduced to reduce energy consumption. At the same time, the node spacing and relay position of the LoRaWAN network are reconfigured to ensure communication reliability in high-frequency data areas. The optimized layout scheme is deployed accurately through a three-dimensional grid coordinate system, ensuring spatial matching between new nodes and the root density gradient model, and preserving interfaces for subsequent iterative adjustments. The optimized rhizosphere sediment carbon flux estimation model and layout scheme are applied to the new monitoring period. Through cross-validation, the deviation between model estimated values and laboratory measured values is compared to quantify the improvement effect. Long-term trend analysis reveals the coupling rules of root exudates, microbial metabolic activity, and environmental factors, and the verification conclusions are fed back to the model optimization and network adjustment process, forming a closed-loop improvement system of monitoring-analysis-optimization-verification. Multidimensional ecological data is continuously accumulated, and mechanism models of plant-soil-microorganism interactions are gradually constructed to improve the ability to analyze key processes of carbon cycling.

[0083] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array, characterized in that, Includes the following steps: S1. Based on the three-dimensional distribution characteristics of the target plant root system, determine the stratified sampling depth and the range of horizontal concentric circle partitions, and map them to a preset three-dimensional grid coordinate system, specifically including: Based on the root system architecture of the target plant, the root distribution density in the vertical and horizontal directions is quantified by non-destructive imaging technology. Key parameters including taproot depth, lateral root expansion radius and root tip active zone range are extracted to establish a root density gradient model. Based on the root density gradient model, the rhizosphere region is divided into vertical stratification and horizontal concentric circle partitions, and the depth of vertical stratification and the radius of horizontal concentric circle partitions are determined. The depth of vertical stratification and the radius of horizontal concentric circle partitions are dynamically adjusted according to the root biomass decay threshold. A three-dimensional mesh coordinate system is preset, and the hierarchical partitioning results are mapped to the three-dimensional mesh coordinate system to determine the deployment location of each sensor node, including gas sensors and environmental parameter sensors, and mark them in the three-dimensional coordinate mesh; The process of dividing the rhizosphere region into vertical strata and horizontal concentric circle partitions, and determining the depth of the vertical strata and the radius of the horizontal concentric circle partitions, is as follows: The vertical stratification includes a surface layer, a middle layer, and a deep layer. The depth of the vertical stratification is determined by the vertical decay threshold of root biomass. The surface layer has a depth of 0-15cm, corresponding to the high metabolic activity zone; the middle layer has a depth of 15-30cm, corresponding to monitoring carbon migration and transformation; and the deep layer has a depth of >30cm, focusing on monitoring steady-state carbon flux. The horizontal concentric circle partitions include a proximal root zone and a distal root zone. The radius of the horizontal concentric circle partitions is dynamically adjusted according to the radial decrease in lateral root biomass. The radius of the proximal root zone is ≤ 50% of the lateral root expansion radius, and the radius of the distal root zone is > 50% to 1.5 times the lateral root radius. S2. Deploy gas sensors and environmental parameter sensors according to root system zones to construct a multi-dimensional sensor array and form a multi-dimensional monitoring network; S3. Real-time synchronous acquisition of gaseous carbon flux and environmental parameter data using various types of sensors in a multi-dimensional monitoring network; S4. Transmit the gaseous carbon flux and environmental parameters collected by each sensor to the cloud for preprocessing; S5. Based on the preprocessed historical gaseous carbon flux and environmental parameter data, a rhizosphere sedimentary carbon flux estimation model is built in the cloud. S6. Combine the rhizosphere sedimentary carbon flux estimation model with the current gaseous carbon flux and environmental parameter data to display the carbon flux change trend and push an abnormal alarm to the user terminal when the abnormal threshold is triggered. S7. Based on long-term monitoring data and alarm information, iteratively update the model parameters and sensor deployment scheme.

2. The method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 1, characterized in that: In step S2, the process of constructing a multi-dimensional sensor array to form a multi-dimensional monitoring network is as follows: Based on the zoning characteristics of vertical root stratification and horizontal concentric circle partitioning, the sensor types were selected to meet the monitoring needs of each region. Among them, gas sensors included CO2 sensors and CH4 sensors, and environmental parameter sensors included temperature and humidity sensors, pH sensors and soil conductivity sensors. A grid-based layout is adopted, in which selected gas sensors and environmental parameter sensors are deployed to corresponding locations according to three-dimensional coordinates, and the sensors are connected by wireless networking to build a multi-dimensional sensor array, forming a multi-dimensional monitoring network covering the rhizosphere.

3. The method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 1, characterized in that: S3 specifically includes: All sensor nodes are synchronized in time using the GPS protocol. Based on a preset three-dimensional grid coordinate system, each sensor is assigned a unique spatial identifier. During physical deployment, a laser rangefinder is used to calibrate the position deviation. A multi-dimensional monitoring network is activated to trigger synchronous sampling of gas sensors and environmental parameter sensors at fixed intervals. A hardware interrupt mechanism is used to avoid transmission delays. Data packets are attached with a unified time-space tag and data is transmitted synchronously through a wireless protocol network. After receiving the data in the cloud, it sorts it by timestamp and fills in the gaps caused by minor transmission delays. Then, it maps the data to a three-dimensional grid coordinate system based on spatial coordinates to generate a spatiotemporally aligned multidimensional dataset.

4. The method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 1, characterized in that: S4 specifically includes: The cloud-based server receives data packets from various sensors, including gaseous carbon flux and environmental parameter data, and parses out time-space labels and raw sensor data, classifying and integrating them according to sensor type. Digital signal processing technology is used to filter the raw sensor data; Based on laboratory calibration data and preset field reference sensors, a drift compensation model was established to perform baseline calibration on the CO2 sensor. Long-term offset was corrected using linear regression. The pH and soil conductivity sensors were processed through regular automatic cleaning and standard solution calibration.

5. A method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 1, characterized in that: S5 specifically includes: We collected preprocessed historical gaseous carbon flux and environmental parameter data, extracted multi-dimensional features related to rhizosphere deposited carbon flux, including the temporal variation rate of gaseous carbon flux, gradient differences of environmental parameters, and spatial correlation of root system partitions, and then constructed a comprehensive dataset containing time-space labels. At the same time, the comprehensive dataset was divided into training set and validation set. A hybrid model architecture combining LSTM and random forest was chosen to build a rhizosphere sedimentary carbon flux estimation model. The hybrid model architecture was trained using the training set, and the hyperparameters were optimized by cross-validation using the validation set, thus obtaining the trained rhizosphere sedimentary carbon flux estimation model. The trained rhizosphere sedimentary carbon flux estimation model is encapsulated as a cloud API, which receives real-time sensor data streams and outputs the calculated rhizosphere sedimentary carbon flux estimates.

6. A method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 5, characterized in that: The calculation process for the estimated rhizosphere sedimentary carbon flux is as follows: The difference between the instantaneous CO2 concentration at the current moment and the concentration at the previous time period is obtained. Dividing the difference by the time interval yields the CO2 concentration change rate. An exponential function is used to correct the CO2 change rate for temperature. The exponent of the exponential function is the difference between the current soil temperature and the reference temperature divided by 10. This is used to standardize biological activity at the actual ambient temperature to the reference temperature level. Finally, by combining the CO2 concentration change rate and the temperature correction results, the respiration-temperature coupling term is obtained. Calculate the ratio of CH4 concentration to CO2 concentration, add 1 and take the natural logarithm. Using the reference soil conductivity as a benchmark, calculate the relative ratio of the current soil conductivity. Use this ratio as the base and the difference between the pH value and the reference pH value as the exponent for power operation. Then, combine the natural logarithm and power operation results to obtain the interaction term between anaerobic metabolism and soil environment. The results of the respiration-temperature coupling term are added to the results of the anaerobic metabolism-soil environment interaction term to obtain the estimated value of rhizosphere deposited carbon flux.

7. A method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 6, characterized in that: S6 specifically includes: The cloud receives sensor data streams, parses timestamps, spatial coordinates and environmental parameters, maps the data according to a preset three-dimensional grid, performs standardized preprocessing, and calls the rhizosphere sedimentary carbon flux estimation model to calculate the current estimated value of rhizosphere sedimentary carbon flux. Analyze the estimated values ​​of rhizosphere deposited carbon flux over three consecutive fixed time periods, generate trend curves, analyze the trend of the estimated values ​​of rhizosphere deposited carbon flux, and determine whether it is rising, falling, or stable. The current estimated value of rhizosphere deposited carbon flux is compared with a preset abnormal threshold. If the value exceeds the abnormal threshold, it is determined to be an abnormal situation, and an abnormal alarm message is automatically generated. An alarm notification containing abnormal information, time, location, and the changing trend of the estimated value of rhizosphere deposited carbon flux is pushed to the user terminal.

8. A method for real-time monitoring of rhizosphere deposited carbon flux using a sensor array according to claim 7, characterized in that: Specifically, S7 includes: Based on historical monitoring data and alarm records, the spatiotemporal distribution characteristics of the model estimation error are analyzed. For high-frequency error regions, Bayesian optimization is used to adjust the hidden layer structure of LSTM and the feature weights of random forest, and the reference benchmark of the drift compensation model is updated synchronously. Long-term calibration data is used to correct the effects of sensor aging. By combining the spatial clustering results of abnormal events, we can identify monitoring blind spots or redundant nodes, increase sensor density in areas where rhizosphere carbon flux increases, reduce node density in steady-state areas, and reconfigure the node spacing and relay positions of the wireless protocol network. The optimized rhizosphere carbon flux estimation model and deployment scheme were applied to the new monitoring cycle to cross-validate the deviation between the estimated rhizosphere carbon flux and the laboratory measured values. Through long-term trend analysis, the coupling law of root exudates-microbial metabolism-environmental factors was revealed. The validation conclusions were then fed back to model optimization and deployment adjustment to form a closed-loop improvement system.

Images