Grassland carbon sink soil multi-element in-situ monitoring system and method

Abstract

The present application discloses a kind of prairie carbon sink soil multi-element in-situ monitoring system and method, system includes monitor, central control console, at least two foundation pits are provided in grassland, one monitoring system is respectively arranged in each foundation pit, and multiple monitoring systems form control group;Monitoring system includes a monitoring cylinder, monitoring cylinder is placed in foundation pit, several monitors are stretched out in monitoring cylinder, monitor monitors grassland soil condition, and central control console collects the data transmission monitored by monitor.Method includes selecting sample site-excavating foundation pit burying monitoring cylinder-data acquisition-data joint correction and cleaning-soil element data exclusive conversion-data analysis and index derivation calculation.The present application designs monitoring-control double sample site difference analysis system, through homologous baseline correction, environmental interference elimination, statistical difference test, stripping natural environment fluctuation interference, the influence of human disturbance, ecological degradation, repair and increase sink measure on prairie soil can be quantified.

Description

Technical Field

[0001] This invention relates to the field of grassland monitoring, specifically to an in-situ monitoring system and method for multiple elements in grassland carbon sink soil. Background Technology

[0002] Grasslands are core carbon sink carriers in terrestrial ecosystems, and grassland soil carbon sinks are a core component of grassland carbon sinks. Soil nutrient imbalances, micronutrient deficiencies, soil salinization, and heavy metal accumulation directly affect soil organic carbon sequestration capacity and constrain grassland carbon sink stability and storage. Current grassland soil monitoring technologies have significant shortcomings: 1. It relies heavily on manual fixed-point sampling and laboratory digestion and testing, which is time-consuming, costly, and cannot achieve long-term continuous dynamic monitoring. It is difficult to capture the dynamic response of soil elements under seasonal changes in grasslands and rainfall and grazing disturbance, and cannot support the time-series dynamic accounting of carbon sequestration. 2. Existing field sensor monitoring schemes are limited and can only detect conventional NPK or salinity indicators. They cannot cover grassland-specific trace elements such as selenium, molybdenum, and boron, as well as the eight major salt and alkali ions. This makes it difficult to comprehensively analyze the impact mechanism of soil physicochemical properties on carbon sequestration, and there is a lack of data to support the nutrient risk assessment in pastoral areas. 3. In extreme field environments, sensor temperature drift and interference from saline-alkali matrices are severe. Raw data is used directly without calibration, resulting in poor accuracy and low reliability, which cannot meet the accuracy requirements for precise carbon sequestration measurement. 4. The existing data analysis system does not distinguish between available soil elements and total elements, cannot match the patterns of pasture absorption and soil carbon sequestration, and lacks grassland-specific models for assessing salinity, nutrient imbalance, and livestock nutrient deficiency risks, making it difficult to support quantitative analysis of carbon sink influencing factors. 5. Existing technologies cannot achieve in-situ synchronous acquisition of multiple elements and carbon flux in soil. Carbon sequestration accounting relies on discrete sampling and empirical estimation, which cannot continuously and dynamically invert the soil carbon sequestration rate, let alone quantitatively analyze the driving mechanism of element changes on carbon sequestration gains and losses. 6. The existing monitoring system consists of individual plots and independent monitoring, lacking a standardized control plot differential analysis mechanism. It cannot eliminate natural disturbances such as climate, precipitation, soil background, and seasonal fluctuations, making it difficult to distinguish between changes in natural background and actual soil changes caused by human disturbance / ecological restoration / degradation stress. It also cannot quantify the impact of human disturbance on grassland soil carbon sequestration, resulting in large biases in monitoring and evaluation results. This seriously restricts the accuracy of grassland carbon sequestration accounting, carbon sequestration asset management, and assessment of the effectiveness of restoration and carbon sequestration enhancement.

[0003] In summary, existing technologies lack a complete monitoring method that is suitable for long-term unattended operation in grasslands, covers multiple elements, has a dedicated correction algorithm, and includes differential comparative analysis of monitoring and control plots. This makes it impossible to support the scientific and refined management of grassland carbon sequestration. Summary of the Invention

[0004] The purpose of this invention is to provide an in-situ monitoring system and method for multiple elements in grassland carbon sink soil, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a multi-element in-situ monitoring system for grassland carbon sink soil, comprising grassland, monitors, and a central control station, wherein at least two foundation pits are set on the grassland, and a monitoring system is set in each foundation pit, and multiple monitoring systems form a control group; The monitoring system includes a monitoring cylinder placed in the foundation pit, with several monitors extending from the cylinder to monitor the soil conditions of the grassland. The central control unit collects and transmits the data monitored by the monitors. An underground camera is also installed inside the monitoring cylinder to photograph the underground root system. A ground camera is installed on the monitoring cylinder to photograph the growth of the grassland. The central control unit is equipped with a power supply for power supply. The monitoring cylinder has three soil profiles on its sidewalls: 0-10cm, 10-30cm, and 30-60cm. Each profile is equipped with an in-situ soil breathing chamber. One side of the in-situ soil breathing chamber is attached to the original soil profile, and the other side is connected to the non-dispersive infrared gas detection module built into the central control unit through an air guide tube. This module is used to simultaneously collect soil carbon flux data for each soil layer. The central control unit has a built-in data correction module and a carbon sink inversion module, which are used to perform multi-level correction processing on the monitoring data and invert the soil carbon sink rate based on multi-element and carbon flux data.

[0006] Furthermore, at least one of the monitoring systems is equipped with a fence to prevent grassland animals from grazing on the pasture, and at least one monitoring system is not equipped with a fence to allow it to grow naturally; a monitoring port is provided on one side of the monitoring tank, and the monitor extends from the monitoring port and is inserted into the soil; a tank cover is provided on the monitoring tank, an insulation layer is covered on the tank cover, and soil is covered on the insulation layer to restore the grassland to its original state, so that the entire monitoring tank is buried underground.

[0007] Furthermore, the underground camera is installed on an underground column, and an underground light source for illumination is also installed on the underground column. The underground column is fixed in the monitoring cylinder. The underground camera is a macro fixed-focus lens, the underground light source is a parallel surface light source, the transparent inner wall of the monitoring cylinder has a preset scale grid, and the central control panel has a built-in root image segmentation unit for quantitatively calculating root length density, root surface area density, and fine root ratio parameters.

[0008] Furthermore, the ground camera is fixed to a ground column, which is fixed to the grass; the outside of the in-situ soil breathing chamber is provided with a PVC blade structure undisturbed soil column isolation ring, which is embedded in the soil to isolate gas diffusion from different soil layers and ensure the independence of stratified carbon flux detection.

[0009] A monitoring method for a multi-element in-situ monitoring system of grassland carbon sink soil, the steps of which are as follows: S1. Selecting sample plots: Select at least two sample plots with consistent site conditions in the grassland, one as the monitoring sample plot and the other as the control sample plot; S2. Excavate the foundation pit and bury the monitoring cylinder, deploy the monitor, central control station, underground camera, above-ground camera and in-situ soil breathing chamber, cover the soil to restore the original state of the grassland, and deploy the fence as needed; the monitor includes an ion-selective electrode array, soil temperature and humidity and EC-pH integrated probe, and field spectral acquisition module, and collect soil data in three layers. S3. Data Acquisition: Simultaneously acquire soil multi-element time-series data, soil physicochemical parameters, soil total element and heavy metal data, stratified soil carbon flux data, and aboveground vegetation and underground root system image data; specifically: continuously acquire soil available nitrogen, phosphorus, potassium, calcium, magnesium, sulfur, selenium, boron, molybdenum, and eight water-soluble ions (salt and alkali) data through an ion-selective electrode array; simultaneously acquire soil moisture content, temperature, pH, and electrical conductivity time-series data through an integrated probe; and precisely acquire soil total element and heavy metal data through a portable spectral module. S4. Data joint correction and cleaning: The raw data is sequentially screened for extreme values, corrected for temperature compensation, corrected for multi-ion cross-interference, corrected for sensor slow drift, removed outliers and filled with missing values, and then uniformly converted to standard concentration units. S5. Soil element data conversion: The ion concentration of soil solution detected by the electrode is converted into the content of available elements in the soil solid phase by combining soil bulk density and volumetric water content parameters; matrix correction and detection limit correction are performed on the spectral data to obtain the total elements and heavy metal content of the soil. S6. Data Analysis and Indicator Derivation Calculation: Based on the corrected dataset, baseline difference analysis of two sample plots is performed to calculate grassland-specific element evaluation indicators; simultaneously, the carbon sequestration rate of stratified soils is inverted to quantify the carbon sequestration gains and losses caused by anthropogenic disturbances and complete the attribution evaluation. Based on the corrected dataset, the differences between fenced and unfenced sample plots are compared and analyzed to calculate the influence of grassland-specific element characteristics, including nutrient imbalance coefficient, sodium adsorption ratio (SAR), exchangeable sodium percentage (ESP), heavy metal single-factor pollution index and Nemerow comprehensive pollution index, and pastoral area trace element deficiency risk index.

[0010] Furthermore, the ion-selective electrode array described in S2 includes nitrate nitrogen electrode, ammonium nitrogen electrode, phosphate electrode, potassium ion electrode, calcium ion electrode, magnesium ion electrode, sodium ion electrode, chloride ion electrode, and sulfate electrode, which are suitable for complex outdoor substrates with high salinity, high organic matter, and high calcium carbonate content in grasslands.

[0011] Furthermore, the data joint correction and cleaning described in S4 includes: performing full-volume ion temperature compensation correction based on 25℃, establishing a salt-alkali matrix interference coefficient model, eliminating the detection interference of bicarbonate and sodium ions on phosphorus, calcium, and sulfur elements; and using the 3σ criterion combined with soil gradient consistency verification to remove abnormal data. The multi-ion cross-interference correction is a water content coupled full-ion interference matrix correction: the interference coefficient of each target ion under different water contents is determined in advance by spiked recovery experiment, and an 8×8 ion interference coefficient matrix is ​​constructed that is dynamically updated with volume water content. Multi-ion cross-antagonistic interference is eliminated by matrix operation. The sensor slow drift correction is an adaptive correction based on the conservation of deep soil: taking the average element concentration of the soil in the initial 30-60cm depth as the benchmark, the current average concentration in the deep layer is calculated periodically and the drift correction coefficient is obtained, and the detection data of the upper soil layer is corrected simultaneously.

[0012] Furthermore, the formula for the soil element data-specific conversion described in S5 is as follows: ; in The concentration of available elements in the soil solid phase is mg / kg. To detect the soil solution concentration in mg / L using electrodes, Soil volumetric water content, Soil bulk density; The spectral data were corrected using an organic matter-calcium carbonate combined matrix absorption correction model, and data below the detection limit were assigned unbiased values ​​using half the detection limit.

[0013] Furthermore, the steps for inverting the stratified soil carbon sequestration rate in S6 are as follows: 1) Decompose the heterotrophic respiration components of stratified soil respiration to obtain the soil organic carbon decomposition rate. ; 2) Based on the effective nitrogen, phosphorus, and potassium concentrations, root length density, and temperature and humidity limiting factors, a carbon input rate estimation model was constructed to calculate the carbon input rates of root exudates and litter. ; 3) The stratified net carbon sink rate is The total soil carbon sink is obtained by integrating over time with the soil layer.

[0014] In S6, the baseline difference between the two sample plots uses a dynamic time warping algorithm to match and align time-series data, and calculates the incremental carbon sink gains and losses caused by human disturbances after eliminating environmental response lag errors. A carbon sink-element coupled attribution matrix is ​​constructed to quantitatively analyze the contribution ratio of nutrient changes, salinity stress, trace element deficiency, and heavy metal accumulation to carbon sink gains and losses, and a graded evaluation of the effectiveness of grassland ecological restoration and carbon sink enhancement is conducted based on the results.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention enables long-term, unattended, in-situ monitoring of core elements in grassland soil, overcoming the bottlenecks of traditional manual sampling, such as short cycles, discrete data, and inability to dynamically monitor. Through stratified in-situ soil breathing chambers, it achieves point-level synchronous acquisition of multiple elements and carbon flux in the same profile, providing an in-situ data foundation for quantitative carbon sequestration and solving the technical pain points of traditional carbon sequestration accounting, which relies on discrete sampling and has poor timeliness.

[0016] 2. This invention constructs a multi-ion cross-interference matrix correction model coupled with water content. For the scenario of coexistence of eight major ions in the complex matrix of high salinity and alkalinity grassland, it realizes dynamic correction of all-ion cross-antagonistic interference. Compared with conventional single-factor linear correction, the multi-element detection accuracy is improved.

[0017] 3. This invention distinguishes between the available state and the total element data system, which is consistent with the absorption mechanism of forage grass and adapts to the dual evaluation needs of grassland ecology and pastoral breeding. It constructs a stratified soil carbon sequestration rate in-situ inversion model specific to grasslands, realizing the quantitative conversion from soil multi-element data to carbon sequestration rate and carbon sequestration storage. It can continuously and dynamically calculate the temporal changes of soil carbon sequestration and analyze the benefit and loss mechanism of the evolution of soil physicochemical indicators on grassland soil carbon sequestration.

[0018] 4. This invention designs a monitoring-control dual-plot differential analysis system, introducing a dynamic time warping algorithm to complete time-series matching and alignment, eliminating differential errors caused by environmental response lags. Through source baseline correction, environmental interference removal, and statistical difference testing, it accurately isolates natural environmental fluctuations and can quantify the true impact of human disturbances, ecological degradation, and restoration and carbon sequestration measures on grassland soil. The accompanying carbon sequestration-element coupling attribution matrix can locate the core driving factors of carbon sequestration changes, solving the technical pain points of traditional single-plot monitoring results being distorted and unable to evaluate grassland carbon sequestration changes and the effectiveness of ecological restoration and carbon sequestration enhancement, thus improving the accuracy of grassland carbon sequestration accounting and management. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the layout structure of the present invention; Figure 2 This is a schematic diagram of the monitoring method of the present invention.

[0020] In the diagram: 1. Grassland, 101. Excavation pit, 2. Monitoring cylinder, 201. Monitoring port, 202. Cylinder cover, 203. Insulation layer, 3. Monitor, 4. Central control console, 401. Solar panel, 5. Underground camera, 501. Underground light source, 502. Underground column, 6. Above-ground camera, 601. Above-ground column, 7. Fence, 8. In-situ soil breathing chamber, 9. Air duct, 10. In-situ soil column isolation ring. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 This invention provides a technical solution: an in-situ monitoring system for multiple elements in grassland carbon sink soil, comprising a grassland 1, a monitor 3, and a central control station 4. At least two pits 101 are set on the grassland 1, each pit 101 containing one monitoring system, with multiple monitoring systems forming a control group. At least one monitoring system is equipped with a fence 7 to prevent grassland animals from grazing on the pasture, while at least one monitoring system is not equipped with a fence 7, allowing it to grow naturally. The solar panels 401, power supply, and monitor 3 all use existing models.

[0023] The monitoring system includes a monitoring cylinder 2, which is placed in the foundation pit 101. A monitoring port 201 is set on one side of the monitoring cylinder 2. The probe of the monitor 3 extends from the monitoring port 201 and inserts into the soil. Several monitors 3 extend out of the monitoring cylinder 2 and monitor the soil condition of the grassland 1. The central control station 4 collects the data monitored by the monitors 3, processes and transmits it. An underground camera 5 is also set inside the monitoring cylinder 2 to photograph the underground root system. The underground camera 5 is mounted on an underground column 502. An underground light source 501 for illumination is also installed on the underground column 502. The underground column 502 is fixed in the monitoring cylinder 2. The underground camera 5 uses a macro fixed-focus lens. The underground light source 501 is a parallel surface light source. The transparent inner wall of the monitoring cylinder 2 has a preset scale grid. The central control station 4 has a built-in root image segmentation unit, which can perform pixel-level semantic segmentation on the captured root images, automatically calculate the root length density, root surface area density, and fine root ratio parameters of each soil layer, and output root growth time series data on a daily basis.

[0024] A ground camera 6 is installed on the monitoring tank 2 to photograph the growth of the grassland. The ground camera 6 is fixed to a ground post 601, which is in turn fixed to the grassland 1. A power supply is provided in the central control unit 4, powered by a solar panel 401. A tank cover 202 is installed on the monitoring tank 2, covered with an insulation layer 203, and then covered with soil to restore the grassland 1 to its original state, thus burying the entire monitoring tank 2 underground. The monitoring tank 2 is made of a transparent material, such as acrylic glass, to prevent water loss from the pit 101 and its impact on the ecosystem, while also facilitating photography.

[0025] The monitoring cylinder 2 has three soil profiles on its sidewalls: a 0-10cm surface layer, a 10-30cm root-dense layer, and a 30-60cm deep layer. Each layer is equipped with an in-situ soil respiration chamber 8. One side of the in-situ soil respiration chamber 8 is attached to the original soil profile, and the other side is connected to the non-dispersive infrared (NDIR) CO2 concentration detection module built into the central control unit 4 via an air guide tube 9. The chamber is equipped with a miniature air pump and a solenoid valve to achieve intermittent closed-loop measurement in each layer, simultaneously collecting the heterotrophic respiration rate of each soil layer. An original soil column isolation ring 10 is set on the outside of the in-situ soil respiration chamber 8. It adopts a PVC blade structure and is embedded in each soil layer to isolate gas diffusion between different soil layers and ensure the independence of layered carbon flux detection. Electrode guide holes are reserved on the sidewall of the chamber, which are coaxially arranged with the original ion electrodes and temperature and humidity probes to achieve point-level synchronous acquisition of multi-element data and carbon flux data of the same soil profile.

[0026] Please see Figure 2 A method for in-situ monitoring of multiple elements in grassland carbon sink soil, the steps of which are as follows: S1. Selecting Sample Plots: Select at least two sample plots on grassland 1, with similar growth conditions and sizes. Select grassland areas subject to grazing disturbance, soil degradation, salinization stress, ecological restoration intervention, or industrial / mining disturbance as monitoring sample plots, i.e., sample plots with fencing (7). Simultaneously, within 50-200m of the monitoring sample plots, select native stable grassland with completely identical site conditions and no human disturbance or restoration intervention as blank control sample plots, i.e., sample plots without fencing (7), using their long-term monitoring data as the baseline for regional natural changes. The paired monitoring and blank control sample plots form a homogeneous monitoring model. By strictly unifying site conditions, systematic errors caused by differences in soil background, topography, climate, and vegetation substrate are eliminated, achieving the separation of human disturbance from natural changes. All monitoring and control sample plots follow the principles of similar landforms, soil parent material, altitude slope, climate background, and vegetation type.

[0027] S2. Deploy the monitoring system; excavate a foundation pit 101 in the sample plot. The depth of the foundation pit 101 is greater than the height of the monitoring cylinder 2. Place the monitoring cylinder 2 in the foundation pit 101. The monitoring cylinder 2 is made of transparent material, such as acrylic or glass, to facilitate observation and photography while preventing water loss from the inner wall of the foundation pit 101, which would affect the growth of pasture. Install the monitor 3, the central control unit 4, and the underground camera 5 in the monitoring cylinder 2. The power supply of the central control unit 4 is electrically connected to the solar panel 401, the monitor 3, the underground camera 5, and the above-ground camera 6. The control system signal in the central control unit 4 is connected to the monitoring... The monitoring system consists of three components: a ground camera (3), an underground camera (5), and a surface camera (6). The data collected by these components is transmitted back to the central control station (4). After the system is set up, the cylinder cover (202) is sealed, and the insulation layer (203) is placed on top. Then, the grass (1) is covered with soil to restore it to its original state. The surface camera (6) is then set up again. Finally, a fence (7) is set up as needed to surround the area covered by the solar panel (401), the surface camera (6), and the monitoring cylinder (2). The monitoring port (201) is matched with three soil monitoring profiles: a surface layer of 0-10cm, a root dense layer of 10-30cm, and a deep layer of 30-60cm.

[0028] The monitor 3 includes an ion-selective electrode array, an integrated probe for soil temperature, humidity and EC-pH, and a field spectral acquisition module with acquisition depths of 0-10cm, 10-30cm, and 30-60cm. The ion-selective electrode array includes nitrate nitrogen electrode, ammonium nitrogen electrode, phosphate electrode, potassium ion electrode, calcium ion electrode, magnesium ion electrode, sodium ion electrode, chloride ion electrode, and sulfate electrode, which are suitable for complex field substrates with high salinity, high organic matter and high calcium carbonate in grasslands.

[0029] Both sample plots employed a completely identical stratified monitoring structure, uniformly setting up three soil monitoring profiles: a 0-10cm surface layer, a 10-30cm root-dense layer, and a 30-60cm deep layer, matching the grassland root distribution and the main soil layers enriched for soil carbon sinks. Each soil layer was equipped with an integrated sensing unit, including a multi-channel ion-selective electrode array, soil temperature and humidity probes, pH / EC salinity sensors, and a field spectral acquisition interface. All monitoring equipment models, parameter accuracy, and acquisition configurations were completely standardized, equipped with a waterproof and dustproof field protection structure and a solar-powered autonomous power supply system. A unified continuous acquisition frequency was set, enabling year-round unattended synchronous monitoring, ensuring complete temporal and spatial matching of data from the two sample plots, providing a common data foundation for subsequent differential comparative analysis. After equipment deployment, initial background parameters such as soil bulk density, initial organic matter content, background salinity, and basic element content were uniformly measured in both sample plots, constructing a common background database for the two sample plots, providing benchmark data support for subsequent baseline correction and differential analysis.

[0030] S3. Data Acquisition: Using ion electrode arrays in two sample plots, continuous time-series data were collected on available nutrients and water-soluble salt ions in the soil, including ammonium nitrogen, nitrate nitrogen, available phosphorus, available potassium, calcium ions, magnesium ions, sodium ions, chloride ions, and sulfate ions. Data on total soil elements and heavy metals were collected at fixed points using a portable spectrometer module. Simultaneously, real-time data on environmental physicochemical parameters such as soil temperature, volumetric water content, pH value, and electrical conductivity were collected. At fixed time intervals, field spectroscopic sampling was conducted to simultaneously acquire total data on trace elements such as iron, manganese, copper, zinc, and selenium, and heavy metals such as chromium, lead, cadmium, and nickel in the soil of both sample plots. Simultaneously, time-series data on soil respiration CO2 concentration were collected intermittently through eight in-situ soil respiration chambers in each layer, and the stratified soil respiration rate was calculated. This resulted in a multi-dimensional monitoring dataset covering grassland soil fertility, salinity barriers, micronutrients, heavy metal risk, and carbon flux, providing comprehensive data support for subsequent analysis of the impact mechanisms of soil physicochemical properties on grassland carbon sequestration. Data acquisition operations were performed synchronously, at the same frequency, and with the same parameters for both monitoring and control plots to eliminate time differences and operational errors.

[0031] S4. Joint data correction and cleaning: Temperature compensation correction, multi-ion cross-interference correction, sensor slow drift correction, outlier removal and missing value interpolation filling are performed on the original time series data, and the data is uniformly converted into standard soil element concentration units of mg / kg; For complex monitoring scenarios in grassland fields with high salinity, high organic matter, large diurnal temperature range and strong environmental interference, a six-level standardized data purification system is constructed, which includes extreme value threshold screening, time series anomaly discrimination, temperature full-domain compensation, multi-ion cross-interference correction, slow drift adaptive correction, and graded missing value filling. The monitoring plots and control plots are uniformly equipped with the same set of correction parameters to ensure the fairness and data accuracy of the differential comparison between the two plots.

[0032] First, an initial screening of extreme value thresholds is performed on the raw data. Based on the inherent physical boundaries of grassland soil physicochemical parameters, a hard threshold filtering rule is established to remove invalid and abnormal data caused by equipment malfunction, exposed probes, soil cavities, or electromagnetic interference in the field. If the monitored sample data meets any of the following threshold conditions, it is directly marked as invalid and removed: concentration and moisture content indicators less than 0; pH value exceeding the reasonable grassland range of 3.5-11.0; electrical conductivity exceeding the grassland extreme salinity limit of 50 dS / m; and volumetric water content exceeding the soil physical range of 0-0.6.

[0033] Secondly, a time-series anomaly removal algorithm based on the 3σ criterion combined with soil gradient constraints is proposed. To address sensor time-series drift, instantaneous noise, and field disturbance noise, a dual-constraint anomaly identification method combining statistical discrimination and soil stratification physical laws is employed to avoid the erroneous deletion of valid ecological fluctuation data using a single statistical method. This method is applied to single-soil-layer daily time-series sequences. Calculate the time series mean and standard deviation: ; ; Routine anomaly detection: .

[0034] Simultaneously, a priori constraint on grassland soil gradient is added: grassland soil nutrients and trace elements generally follow a pattern of enrichment at the surface and decrease layer by layer. If the element concentration in the lower layer (30-60cm) is significantly higher than that in the surface layer (0-10cm) and exceeds the statistical range, it is judged as a probe anomaly and is removed. This constraint can improve the accuracy of anomaly identification in grassland field data.

[0035] Third, an ion electrode temperature compensation correction model based on the Nernst equation. Diurnal temperature variations in the field can cause ion electrode potential drift, easily leading to detection errors. A global temperature correction formula is derived based on the Nernst equation, uniformly calibrating all ion detection data to a standard temperature of 25℃, eliminating systematic errors caused by temperature fluctuations. Standard Nernst equation: ; In the formula: The measured electrode potential (mV) This is the intrinsic constant potential of the electrode; This is the universal gas constant; To monitor the real-time absolute temperature (K) at the monitoring point; The valence state of the target ion; It is Faraday's constant; This represents the molar concentration of ions.

[0036] At a constant temperature of 298.15 K, the potential drift at a fixed ion concentration is only related to Linear correlation leads to the temperature-corrected concentration formula: ; In the formula: This represents the actual ion concentration at a standard temperature of 25℃. For real-time temperature The original concentration detected by the lower sensor; For each ion's specific temperature correction factor, subscript Different ions such as nitrate nitrogen, ammonium nitrogen, potassium, calcium, sodium, and sulfate are distinguished by calibration using gradient standard solutions.

[0037] By combining the standard Nernst equation with temperature variables, a temperature correction formula specifically for grassland soil is derived. Dedicated temperature correction coefficients are configured for different ions such as ammonium nitrogen, nitrate nitrogen, potassium, calcium, sodium, and sulfate to achieve multi-ion temperature compensation.

[0038] Fourth, we performed multi-ion cross-interference matrix correction coupled with water content. Eight major ions coexist in grassland soils, exhibiting complex antagonistic and synergistic interference among them. Furthermore, the interference intensity dynamically changes with soil water content, leading to large errors in traditional single-factor linear correction. We pre-determined the interference coefficients of each target ion with the other seven coexisting ions under different water content gradients through spiked recovery experiments in grassland soils, constructing an 8×8 ion interference coefficient matrix. Where i is the target ion and j is the interfering ion, and the coefficient varies with the volumetric water content. Dynamic interpolation update. The correction formula is: ; In the formula, This is the corrected true concentration. This represents the original concentration of the i-th ion after temperature correction. The interference coefficient is dynamically changed with volumetric water content, where i is the target ion and j is the interfering ion; The original concentration of the j-th coexisting ion after temperature compensation correction is given. Additionally, for grassland high-calcium carbonate matrix, a calcium carbonate dissociation correction term is introduced: the degree of carbonate dissociation is calculated in real-time based on soil pH and temperature to correct for the precipitation interference of carbonate / bicarbonate on phosphate and calcium ions.

[0039] Fifth, adaptive correction of sensor drift based on deep soil conservation. The 30-60cm deep soil layer in grasslands is minimally affected by external disturbances and seasonal changes, and its element concentrations remain stable and conserved over a long period. Using this as a natural reference layer, adaptive correction of surface sensor drift without calibration liquid can be achieved. The average elemental concentration at a depth of 30-60 cm during the initial month of deployment was used as the long-term baseline value. Calculate the average value of deep element detection within the current window every 15 days. Calculate the drift correction factor: ; This is the drift correction factor; This coefficient was used to simultaneously correct all ion detection data from the 0-10cm and 10-30cm layers of the same cross-section: ; The values ​​at time t are the actual ion detection values ​​at 0-10cm and 10-30cm after slow drift correction (unit: mg / L). The original detection concentration value (unit: mg / L) output by the shallow ion-selective electrode at time t; outlier constraints are applied using the 3σ criterion, if... If the value exceeds the range of 0.85-1.15, it is considered a sensor malfunction and marked as abnormal.

[0040] Sixth, a hierarchical adaptive missing value imputation algorithm is implemented. For data missing due to scenarios such as power outages in the field, electrode maintenance, and flooding caused by rain and snow, a differentiated imputation strategy is adopted to ensure the continuity and authenticity of time-series data. Missing data within a short period of 6 hours is imputed using time-series linear interpolation; for medium- to long-term missing data exceeding 6 hours, corrected means of the same soil layer, vegetation, and time period are used to imput the missing data, based on baseline data from control plots. Compared to traditional single-mean imputation, this effectively eliminates interference from seasonal and climatic fluctuations.

[0041] Short-term missing (duration of missing data) Two-point linear interpolation ; In the formula: The effective observation times before and after the missing interval; These are the corrected index values ​​at the corresponding time points; The time to be filled.

[0042] Medium- to long-term deficiency ( ): Baseline correction mean of control plots with the same soil layer, same vegetation type, and same time period is filled to avoid smoothing out seasonal natural carbon sink fluctuations by traditional global mean.

[0043] Finally, the data synchronization and normalization rules for the two sample plots were established. The same set of extreme value thresholds, 3σ judgment parameters, and ion temperature correction coefficients were reused throughout the monitoring and control sample plots. Matrix interference coefficient Interpolation rules ensure that the two datasets complete the entire correction process synchronously, with no systematic bias in the correction parameters. This provides a standardized paired dataset for subsequent differential stripping of natural fluctuations and quantification of human-induced carbon sequestration gains and losses.

[0044] S5. Specific conversion of soil element data: The ion concentration of soil solution detected by the electrode is converted into the content of available elements in the soil solid phase by combining soil bulk density and volumetric water content parameters; matrix correction and detection limit correction are performed on the spectral data to obtain the total element and heavy metal content of the soil; the ion concentration of pore water solution is converted into available elements in the soil solid phase. Known parameters: The ion concentration in the pore water solution, in units of ; Soil volumetric water content, dimensionless ; For soil bulk density, unit ; The content of effective elements in the solid phase of dry soil, in units : Mass of water per unit volume of soil: ; Mass of dry soil per unit volume: ; By simultaneously eliminating the dimensions of volume, we obtain the standardized conversion formula: ; This formula has no empirical fitting term and is uniquely determined by physical logic. It is the core conversion formula for aligning in-situ field data with laboratory soil testing standards. It is used to quantify the available soil nutrient reserves and analyze the supporting strength of nutrient supply for vegetation carbon sequestration and soil organic carbon storage.

[0045] XRF combined matrix absorption correction model for total elemental organic matter and calcium carbonate. High organic matter (organic carbon) and calcium carbonate components in grasslands absorb X-rays, resulting in inflated values ​​for trace elements and heavy metals, interfering with carbon sequestration risk assessment. A dual-matrix coupling correction formula is constructed: ; In the formula: To correct the true content of total elements in the soil; This refers to the raw spectral readings. The proportion of soil organic matter is directly related to soil organic carbon storage; These are the absorption correction coefficients for organic matter and calcium carbonate matrix, respectively. This represents the percentage of calcium carbonate in the soil by mass.

[0046] Unbiased standardized assignment algorithm for data below the instrument detection limit. XRF trace and heavy metal indicators often appear below the detection limit (ND). Directly assigning a value of 0 would underestimate the degree of elemental deficiency and pollution stress. Therefore, a statistically unbiased estimation is used: ; In the formula: The corresponding element is the instrument detection limit; The value is assigned to samples below the detection limit for inclusion in the statistics.

[0047] The monitoring plots and control plots used exactly the same methods. , , , The parameters are synchronously converted, and two sets of standardized paired datasets with completely consistent dimensions and statistical rules are output as the input basis for S6 dual-plot difference analysis.

[0048] S6. Data Analysis and Indicator Derivation Calculation: Based on the corrected dataset, compare and analyze the difference between the plots with fence 7 and the plots without fence 7, and calculate the influence of grassland-specific element characteristics through the difference, including nutrient imbalance coefficient, sodium adsorption ratio SAR, exchangeable sodium percentage ESP, heavy metal single-factor pollution index and Nemerow comprehensive pollution index, and pastoral area trace element deficiency risk index.

[0049] Baseline differential perturbation stripping model for dual-plot sites: Define timing variables: for Monitor the corrected and converted index values ​​of the sample plots at all times; for The index values ​​were corrected and converted by comparing the sample plots at all times. The initial background values ​​(including initial organic carbon background) were set for the monitoring and control plots, respectively.

[0050] Regional fluctuations in natural environment during the same period: ; Initial background deviation coefficient between the two sample plots: ; The true change in anthropogenic disturbances after removing all natural fluctuations: ; All temporal fluctuations in the control plots were attributed to the baseline of natural evolution of climate, season, and soil. Eliminating initial background differences, ultimately It only reflects changes in soil indicators caused by human activities and can be directly used to quantify the gains and losses of soil carbon sequestration caused by human activities.

[0051] To address the discrepancy in carbon sequestration rates, a Dynamic Time Warping (DTW) algorithm is introduced for time series matching and alignment, resolving the asynchrony in ecological responses between the two sample plots caused by grassland rainfall pulses and temperature fluctuations. Using the carbon sequestration rate time series of the control plot as a reference sequence and the time series of the monitoring plot as the sequence to be matched, the DTW algorithm calculates the optimal time series matching path, eliminating time series misalignments caused by rainfall lag and temperature conduction differences. After alignment, the incremental carbon sequestration gains and losses due to anthropogenic disturbances are calculated. ; in The net carbon sink gain / loss increment at time t after DTW time series matching and alignment, stripped of natural fluctuations (unit: gC·m). -2 ·d -1 A positive value represents carbon dioxide emissions due to human activities, while a negative value represents carbon losses. The measured net carbon sink rate (unit: gC·m³) of the sample plot at the corresponding time t′ after matching by the DTW algorithm was monitored. -2 ·d -1 ); The measured net carbon sequestration rate of the control plot at the original time t (unit: gC·m). -2 ·d -1 ); The corresponding time after DTW alignment is used to eliminate the lag error in the ecological response of the two sample plots caused by environmental pulses such as rainfall and temperature. The initial background correction coefficient is used to eliminate the inherent minor differences in the initial soil organic carbon background and site conditions between the two sample plots.

[0052] Secondly, a quantitative evaluation algorithm for grassland nutrient imbalance is developed: the ratio of nitrogen and phosphorus, and nitrogen and potassium supply in the soil directly determines the rate of photosynthetic carbon fixation in vegetation, and the deviation of the ratio after differential analysis can quantify the nutrient limitation caused by human degradation.

[0053] Nutrient Stoichiometry: ; ; In the formula: This refers to the nitrogen-phosphorus nutrient ratio in the soil. This refers to the nitrogen-potassium nutrient ratio in the soil. The content of available nitrogen in the soil (mg / kg); The content of available phosphorus in the soil (mg / kg); This represents the available potassium content in the soil (mg / kg).

[0054] Monitoring-control nutrient imbalance deviation: ; In the formula: The degree of imbalance in the nitrogen-phosphorus ratio; The degree of imbalance in the nitrogen-potassium ratio; These represent the nitrogen-phosphorus ratio and nitrogen-potassium ratio of the monitored sample plots, respectively. These represent the nitrogen-phosphorus ratio and nitrogen-potassium ratio of the control plots, respectively.

[0055] When the vegetation deviates from the optimal range for grassland, it is considered to be under nutrient stress, which corresponds to a decrease in vegetation carbon sequestration and a reduction in soil carbon sink input.

[0056] Third, perform quantitative calculations of salt-alkali stress: Sodium adsorption ratio SAR (Grassland alkalization grading standard index): ; In the formula: The sodium adsorption ratio is used to characterize the degree of soil alkalization stress. This refers to the concentration of water-soluble sodium ions in the soil. This refers to the concentration of water-soluble calcium ions in the soil. This refers to the concentration of water-soluble magnesium ions in the soil.

[0057] Formulas for calculating total salt content based on the conservation of eight major ions: ; In the formula: This refers to the total water-soluble salt content of the soil. , , , These represent the concentrations of water-soluble potassium, sodium, calcium, and magnesium cations in the soil. , , , These represent the concentrations of water-soluble bicarbonate, carbonate, sulfate, and chloride anions in the soil.

[0058] By combining the differential model, the increase in secondary salinity caused by human activity is obtained, the natural salinity background is distinguished from the increased salinity caused by human disturbance, and the soil organic carbon loss caused by salinity stress is quantitatively calculated.

[0059] Fourth, conduct the Micronutrient Deficiency Coupling Risk Index (MSI) for pastoral areas: Se, Cu, Zn, and Mo regulate soil microbial activity, and microorganisms participate in soil organic carbon fixation and decomposition cycles, constructing a weighted comprehensive nutrient deficiency index: ; In the formula: A risk index for the coupling of trace element deficiencies in pastoral areas; Standardized weighting coefficients for grassland pastoral areas were determined using the analytic hierarchy process combined with grassland forage demand characteristics. The critical standard value for effective absorption by forage grass is determined based on the diagnostic standards for soil nutrition in grassland pastoral areas. , , , The measured contents of available selenium, copper, zinc and molybdenum in the soil solid phase after correction and conversion using step S5 are shown in mg / kg. The lower the MSI value, the stronger the inhibition of microbial activity and the higher the risk of soil carbon cycle imbalance. After baseline differentiation of the control plots, the background of natural nutrient deficiency in the region can be eliminated, and the risk of trace element loss and carbon sink depletion caused by human degradation can be identified.

[0060] Fifth, conduct quantitative assessments of heavy metal pollution, as heavy metals toxicize soil carbon-sequestering microorganisms and continuously reduce soil carbon sinks. Single-factor pollution index: ; In the formula: For the first Single-factor pollution index of heavy metals; For the first in the soil The total measured content of all heavy metals; These are the screening standard limits for the risk of heavy metal pollution in agricultural land soil.

[0061] Nemerow Comprehensive Pollution Index: ; In the formula: The Nemerow Comprehensive Pollution Index; This is the arithmetic mean of all single-factor pollution indices for heavy metals; It represents the maximum value among all single-factor pollution indices for heavy metals.

[0062] By combining the difference model to deduct the natural background heavy metals in the soil, the carbon sink loss caused by chemical mining and transportation pollution is quantified.

[0063] Sixth, conduct in-situ inversion of stratified soil carbon sequestration rates: 1) Respiration component decomposition: Heterotrophic respiration and autotrophic respiration were decomposed using stratified soil respiration data. Combined with root biomass data and fine root proportion parameters, the proportion of root autotrophic respiration was calculated to obtain the soil organic carbon decomposition rate. (z represents the depth of the soil layer).

[0064] 2) Carbon input rate estimation: A carbon input rate estimation sub-model was constructed, which estimated the carbon input rate of root exudates and litter based on available nitrogen, phosphorus and potassium concentrations, root length density, and aboveground vegetation growth (coverage and plant height extracted by aboveground cameras). : ; In the formula, The carbon input rate of soil layer z at time tt (unit: g C·m) -2 ·d -1 This includes carbon input from root exudates and litter; The root length density of soil layer z; This corresponds to the effective nitrogen, phosphorus, and potassium concentrations in the soil layer. Temperature and humidity are the limiting factors, and the Arrhenius formula is used for fitting; a, b, c, and d are grassland type-specific fitting coefficients, which are calibrated through field measurements in sample plots.

[0065] 3) Net carbon sequestration rate calculation: The net carbon sequestration rate of stratified soil is the difference between carbon input and heterotrophic respiration. ; 4) Cumulative Carbon Sequestration: The cumulative total soil carbon sequestration is the integral of the three-layer carbon sequestration rate over time and soil depth. ; In the formula To correspond to the soil layer thickness, It is the unit weight of the soil layer.

[0066] Seventh, conduct carbon sink-element coupling attribution and effectiveness evaluation: Based on the differential carbon sink gain / loss data and element change data, a carbon sink-element coupled attribution matrix is ​​constructed. Through multivariate regression analysis, the contribution ratio of nitrogen and phosphorus nutrient changes, increased salinization, trace element deficiency, and heavy metal accumulation to carbon sink gain / loss is quantitatively analyzed, thus achieving accurate attribution of carbon sink changes.

[0067] Eighth, use the paired-samples t-test to determine the significance of the difference: Paired t-tests were used to distinguish between natural fluctuations and anthropogenic disturbances, providing statistical support for carbon sequestration assessment results. (Difference sequences) , : ; In the formula: This is the paired test statistic; The average value of the paired differences of indicators between the monitoring plot and the control plot; The sample standard deviation of the paired difference sequence; This represents the total number of paired time-series samples.

[0068] according to Value corresponding The results are divided into four levels: no significant natural fluctuations, slight human disturbances, significant ecological degradation (carbon sink depletion), and significant ecological restoration and carbon sink increase.

[0069] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-element in-situ monitoring system for grassland carbon sink soil, comprising grassland (1), a monitor (3), and a central control station (4), characterized in that: At least two foundation pits (101) are set up on the grassland (1), and a monitoring system is set up in each foundation pit (101). Multiple monitoring systems form a control group. The monitoring system includes a monitoring cylinder (2), which is placed in the foundation pit (101). Several monitors (3) extend from the monitoring cylinder (2). The monitors (3) monitor the soil conditions of the grassland (1). The central control unit (4) collects and transmits the data monitored by the monitors (3). An underground camera (5) is also installed inside the monitoring cylinder (2) to photograph the underground root system. An above-ground camera (6) is installed on the monitoring cylinder (2) to photograph the growth of the grassland. A power supply for power supply is installed in the central control unit (4). The monitoring cylinder (2) has three soil profiles corresponding to 0-10cm, 10-30cm and 30-60cm on its sidewall. Each layer is equipped with an in-situ soil breathing chamber (8). One side of the in-situ soil breathing chamber (8) is attached to the original soil profile, and the other side is connected to the non-dispersive infrared gas detection module built into the central control unit (4) through the air guide tube (9) for synchronously collecting soil carbon flux data of each soil layer. The central control unit (4) has a built-in data correction module and a carbon sink inversion module for performing multi-level correction processing on the monitoring data and inverting the soil carbon sink rate based on multi-element and carbon flux data.

2. The grassland carbon sink soil multi-element in-situ monitoring system according to claim 1, characterized in that: At least one of the monitoring systems is equipped with a fence (7) to prevent grassland animals from grazing on the grass, and at least one monitoring system is not equipped with a fence (7) to allow it to grow naturally; a monitoring port (201) is provided on one side of the monitoring tank (2), and the monitor (3) extends out from the monitoring port (201) and is inserted into the soil; a tank cover (202) is provided on the monitoring tank (2), and an insulation layer (203) is covered on the tank cover (202), and soil is covered on the insulation layer (203) to restore the grassland (1) to its original state, so that the entire monitoring tank (2) is buried underground.

3. The grassland carbon sink soil multi-element in-situ monitoring system according to claim 1, characterized in that: The underground camera (5) is installed on the underground column (502), and an underground light source (501) for lighting is also installed on the underground column (502). The underground column (502) is fixed in the monitoring cylinder (2). The underground camera (5) is a macro fixed-focus lens, the underground light source (501) is a parallel surface light source, the transparent inner wall of the monitoring cylinder (2) has a preset scale grid, and the central control console (4) has a built-in root image segmentation unit for quantitatively calculating the root length density, root surface area density and fine root ratio parameters.

4. The grassland carbon sink soil multi-element in-situ monitoring system according to claim 1, characterized in that: The ground camera (6) is fixed on the ground column (601), and the ground column (601) is fixed on the grass (1); the in-situ soil breathing chamber (8) is provided with a PVC blade structure original soil column isolation ring (10) on the outside, which is embedded in the soil to isolate the gas diffusion of different soil layers and ensure the independence of layered carbon flux detection.

5. The monitoring method of the grassland carbon sink soil multi-element in-situ monitoring system according to any one of claims 1-4, characterized in that, The steps are as follows: S1. Selecting sample plots; Select at least two sample plots with consistent site conditions in grassland (1), and use them as monitoring sample plots and control sample plots respectively; S2. Excavate the foundation pit and bury the monitoring cylinder (2), and deploy the monitoring device (3), central control unit (4), underground camera (5), ground camera (6) and in-situ soil breathing chamber (8), cover the soil to restore the original state of the grassland, and deploy the fence (7) as needed; the monitoring device (3) includes an ion-selective electrode array, soil temperature and humidity and EC-pH integrated probe, and field spectral acquisition module, and collects soil data in three layers; S3. Data Acquisition: Simultaneously acquire soil multi-element time-series data, soil physicochemical parameters, total soil and heavy metal data, stratified soil carbon flux data, and aboveground vegetation and underground root system image data. S4. Data joint correction and cleaning: The raw data is sequentially screened for extreme values, corrected for temperature compensation, corrected for multi-ion cross-interference, corrected for sensor slow drift, removed outliers and filled with missing values, and then uniformly converted to standard concentration units. S5. Soil element data conversion: The ion concentration of soil solution detected by the electrode is converted into the content of available elements in the soil solid phase by combining soil bulk density and volumetric water content parameters; matrix correction and detection limit correction are performed on the spectral data to obtain the total elements and heavy metal content of the soil. S6. Data analysis and index derivation calculation: Based on the corrected dataset, perform baseline difference analysis of two sample plots to calculate grassland-specific element evaluation indicators; at the same time, invert the carbon sequestration rate of stratified soils, quantify the carbon sequestration gains and losses caused by human disturbances, and complete the attribution evaluation.

6. The monitoring method of the grassland carbon sink soil multi-element in-situ monitoring system according to claim 5, characterized in that: The ion-selective electrode array described in S2 includes nitrate nitrogen electrode, ammonium nitrogen electrode, phosphate electrode, potassium ion electrode, calcium ion electrode, magnesium ion electrode, sodium ion electrode, chloride ion electrode, and sulfate electrode, which are suitable for complex outdoor substrates with high salinity, high organic matter, and high calcium carbonate in grasslands.

7. The monitoring method of the grassland carbon sink soil multi-element in-situ monitoring system according to claim 5, characterized in that: The data joint correction and cleaning described in S4 includes: full ion temperature compensation correction based on 25℃, establishing a salt-alkali matrix interference coefficient model, eliminating the detection interference of bicarbonate and sodium ions on phosphorus, calcium and sulfur elements; and using the 3σ criterion combined with soil gradient consistency verification to remove abnormal data. The multi-ion cross-interference correction is a water content coupled full-ion interference matrix correction: the interference coefficient of each target ion under different water contents is determined in advance by spiked recovery experiment, and an 8×8 ion interference coefficient matrix is ​​constructed that is dynamically updated with volume water content. Multi-ion cross-antagonistic interference is eliminated by matrix operation. The sensor slow drift correction is an adaptive correction based on the conservation of deep soil: taking the average element concentration of the soil in the initial 30-60cm depth as the benchmark, the current average concentration in the deep layer is calculated periodically and the drift correction coefficient is obtained, and the detection data of the upper soil layer is corrected simultaneously.

8. The monitoring method of the grassland carbon sink soil multi-element in-situ monitoring system according to claim 5, characterized in that: The formula for the soil element data conversion described in S5 is as follows: ； in The concentration of available elements in the soil solid phase is mg / kg. To detect the soil solution concentration in mg / L using electrodes, Soil volumetric water content, Soil bulk density; The spectral data were corrected using an organic matter-calcium carbonate combined matrix absorption correction model, and data below the detection limit were assigned unbiased values ​​using half the detection limit.

9. The monitoring method of the grassland carbon sink soil multi-element in-situ monitoring system according to claim 5, characterized in that: The steps for inverting the stratified soil carbon sequestration rate in S6 are as follows: 1) Decompose the heterotrophic respiration components of stratified soil respiration to obtain the soil organic carbon decomposition rate. ; 2) Based on the effective nitrogen, phosphorus, and potassium concentrations, root length density, and temperature and humidity limiting factors, a carbon input rate estimation model is constructed to calculate the carbon input rates of root exudates and litter. ; 3) The stratified net carbon sink rate is The total soil carbon sink is obtained by integrating over time with the soil layer.

10. The monitoring method of the grassland carbon sink soil multi-element in-situ monitoring system according to claim 5, characterized in that: In S6, the baseline difference between the two sample plots uses a dynamic time warping algorithm to match and align time-series data, and calculates the incremental carbon sink gains and losses caused by human disturbances after eliminating environmental response lag errors. A carbon sink-element coupled attribution matrix is ​​constructed to quantitatively analyze the contribution ratio of nutrient changes, salinity stress, trace element deficiency, and heavy metal accumulation to carbon sink gains and losses, and a graded evaluation of the effectiveness of grassland ecological restoration and carbon sink enhancement is conducted based on the results.

Images