A method for evaluating trace element improvement in tobacco-growing soil based on digital twins

By using digital twin models and data-driven methods, the interference of external environment and natural soil evolution in the assessment of micronutrient improvement in tobacco-growing soil was resolved, enabling accurate assessment of the improvement effect and ensuring the physical process constraints and accuracy of the assessment results.

CN123117470APending Publication Date: 2026-11-17LINCANG OF YUNNAN TOBACCO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611437792.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-16
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing methods for assessing micronutrient improvement in tobacco-growing soils are insufficient to separate the effects of improvement interventions from the influence of the external environment and natural soil evolution, and it is difficult to determine whether changes in available soil elements can be converted into actual absorption by tobacco plants.

Method used

A dual-path model based on digital twins is established. By acquiring data before and after the improvement, spatial units are divided, shared parameter sets are calibrated, continuous state sequences are generated, the potential available cumulative exposure in the root region is calculated, and hysteresis matching and material budget residual verification are performed to generate improvement evaluation results.

Benefits of technology

To reduce the interference of external environmental changes and natural soil evolution on improvement assessment, distinguish the difference between the increase in the content of available trace elements in the soil and the increase in the actual available exposure of tobacco roots, determine whether the change in root exposure is converted into actual absorption by tobacco plants, and form improvement assessment results with physical process constraints.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117470A_ABST
    Figure CN123117470A_ABST
Patent Text Reader

Abstract

This invention discloses a digital twin-based method for assessing micronutrient improvement in tobacco-growing soils, relating to the field of tobacco soil quality assessment. The method involves acquiring baseline data before improvement, unimproved calibration data, common exogenous driving data, improvement intervention data, and post-improvement observation data; establishing a real-world improvement evolution path and a virtual unimproved evolution path; inputting the common exogenous driving data into both evolution paths, and applying the improvement intervention data and post-improvement observation correction only to the real-world improvement evolution path to obtain a soil element state sequence; calculating the potential available cumulative exposure in the root zone and the difference in exposure between the two paths based on root distribution; performing lag matching between the phased increments of the exposure difference and the increments of the cumulative elemental exposure in tobacco plants to obtain absorption response consistency; calculating the material budget residuals; and generating improvement assessment results. This invention can separate environmental changes from the effects of improvement interventions and assess the transformation process of soil micronutrient changes into tobacco plant absorption.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of tobacco-growing soil quality assessment technology, specifically to a method for assessing the improvement of trace elements in tobacco-growing soil based on digital twins. Background Technology

[0002] Micronutrients such as zinc, boron, molybdenum, iron, manganese, and copper in tobacco-growing soil participate in photosynthesis, enzymatic reactions, metabolism, and root development in tobacco plants. Micronutrient deficiency, fixation, or irrational spatial distribution can all limit tobacco plant growth. To improve the supply of micronutrients in tobacco-growing soil, soil improvement measures can be implemented, such as applying micronutrient-containing amendments, adjusting soil pH, increasing organic matter, or improving root zone moisture conditions.

[0003] Current methods for assessing micronutrient remediation in tobacco-growing soils typically involve collecting soil samples before and after remediation, comparing the total amount or available content of micronutrients to determine the remediation effect. This approach primarily reflects soil test results at a limited sampling time and struggles to continuously characterize the release of remediation materials, micronutrient adsorption and desorption, fixation and transformation, spatial migration, and root absorption processes. When rainfall, irrigation, soil temperature, and the release of existing soil elements all change simultaneously, the changes in micronutrient content after remediation are not necessarily entirely caused by the remediation intervention. Relying solely on the difference in content before and after remediation is insufficient to separate the influence of the external environment and natural soil evolution.

[0004] Furthermore, an increase in available trace elements in the soil does not equate to actual utilization by tobacco plants. Trace elements released by amendments may concentrate in soil layers not yet reached by roots, or may fail to be effectively absorbed due to unsuitable soil moisture content, temperature, or pH. Existing assessment methods typically lack continuous analysis of the relationship between available element content, root spatial distribution, root activity, and element accumulation in tobacco plants, and also lack a process for verifying the elemental balance of evaluation results. Consequently, it is difficult to identify different scenarios such as insufficient release of amendments, insufficient root zone accessibility, mismatch in tobacco plant absorption response, and boundary losses or observational biases. Summary of the Invention

[0005] The present invention aims to provide a method for evaluating the improvement of trace elements in tobacco-growing soil based on digital twins, in order to solve the problems of existing methods that make it difficult to separate the improvement intervention from the external environment and natural soil evolution, and that it is difficult to determine whether changes in available elements in the soil can be converted into actual absorption by tobacco plants.

[0006] To address the aforementioned technical problems, this invention provides a method for evaluating the improvement of trace elements in tobacco-growing soil based on digital twins, comprising the following steps: S1. Obtain baseline data before improvement, calibration data before improvement, common external driving data, improvement intervention data and observation data after improvement for the tobacco-planted plots. Divide the tobacco-planted plots into spatial units according to the sampling location and soil layer boundary, and determine the common initial state of each spatial unit based on the baseline data before improvement. S2. Based on the unmodified calibration data, calibrate the shared parameter set, establish a dual-path digital twin model including the real modified evolution path and the virtual unmodified evolution path, so that the two evolution paths adopt the same model structure, the common initial state and the shared parameter set, and lock the shared parameter set before the modification intervention begins; S3. Input the common exogenous driving data into the two evolution paths respectively, input the improvement intervention data only into the actual improvement evolution path, and use the improved observation data to correct the actual improvement evolution path to obtain state sequences that characterize the soil element form transformation, spatial migration and root absorption processes in the two evolution paths respectively. S4. Based on the state sequences and corresponding root distributions of the two evolutionary paths, calculate the potential cumulative exposure of each trace element in the root region, and calculate the exposure difference between the two evolutionary paths. S5. Based on the improved observation data, determine the increment of the cumulative element amount of tobacco plants, and perform lag matching between the increment of the cumulative element amount of tobacco plants and the increment of the difference between the exposure amounts to obtain the consistency of the absorption response. S6. Calculate the material budget residual based on the improved intervention data, the state sequence, the cumulative increase of tobacco element and the boundary loss, and generate the improved evaluation result based on the exposure difference, the absorption response consistency and the material budget residual.

[0007] Furthermore, step S1 specifically includes: The baseline data before improvement includes soil layer boundaries, soil dry bulk density, soil moisture content, soil pH, total trace elements and available trace element content at each sampling location, which are used to determine the common initial state of each spatial unit. The unmodified calibration data includes soil environmental data, trace element speciation data, root length density data, and tobacco plant element absorption data for the corresponding spatial unit during the period when no modification materials were applied, which are used to calibrate the shared parameter set. The common exogenous driving data includes rainfall, the concentration of target trace elements carried by rainfall, irrigation amount, the concentration of target trace elements in irrigation water, and soil temperature. The improvement intervention data includes the elemental composition, application amount, application location, and application time of the improvement material. The post-improvement observation data includes post-improvement soil environmental data, trace element speciation data, root length density data, tobacco plant elemental data, evaluation boundary water discharge, and the concentration of target trace elements in the discharge water.

[0008] Furthermore, step S2 specifically includes: Evolution calculations are performed based on the initial calibration state at the first moment of the unmodified calibration data and the common external driving data during the corresponding unmodified period to obtain the unmodified simulation sequence; Calculate the calibration residuals between the effective content of trace elements, the total amount of trace elements, and the cumulative amount of elements in tobacco plants in the unmodified simulated sequence and the unmodified calibration data, respectively; The element release parameters, adsorption-desorption parameters, fixed transformation parameters, spatial migration parameters, and root absorption parameters are iteratively adjusted according to the calibration residual until the calibration residual meets the preset calibration conditions. The adjusted parameters are then determined as the shared parameter set and replicated to form the real improved evolution path and the virtual unimproved evolution path.

[0009] Furthermore, step S3 specifically includes: The water-soluble element reserves, exchangeable element reserves, and fixed element reserves in each spatial unit are set as soil element state quantities, and the element reserves absorbed by the tobacco plants are set as plant element state quantities. The water-soluble element reserves are increased according to the improved material release process, the water-soluble element reserves are transferred between the water-soluble element reserves and the exchangeable element reserves according to the adsorption and desorption process, the water-soluble element reserves are transferred between the exchangeable element reserves and the fixed element reserves according to the fixation and transformation process, and the water-soluble element reserves and plant element state quantities of adjacent spatial units are updated according to the water flux and root absorption process. The available content of trace elements is determined by dividing the sum of the water-soluble element reserves and the exchangeable element reserves converted by the extraction coefficient by the dry mass of the soil in the spatial unit.

[0010] Further, step S3 involves using the improved observation data to correct the actual improved evolutionary path, including: The soil moisture content, soil temperature, and soil pH value in the improved soil environmental data are written into the environmental state quantities of the corresponding time and spatial unit in the actual improvement evolution path. The total amount of observed trace elements is multiplied by the dry mass of the soil in the corresponding spatial unit and mapped to the sum of the three soil element state quantities. The observed available content of trace elements is mapped to the available content of trace elements determined in step S3, and the element data of tobacco plants is mapped to the plant element state quantities. The element observation residuals are obtained respectively. The corresponding state quantities of the actual improved evolutionary path are corrected based on the element observation residuals, and the correction amount is reduced proportionally when the correction amount causes any state quantity to be less than zero; the correction process does not change the virtual unimproved evolutionary path and the shared parameter set.

[0011] Furthermore, in step S4, trace elements In evolutionary path The potential available cumulative exposure in the root zone is calculated using the following formula: in, This refers to the content of trace elements in their available form. For soil dry weight, The root arrival weights are obtained by normalizing the root length density. As the root activity weight, The length of the time interval. The number of spatial units. The number of time intervals. For real-world improved evolutionary paths or virtual unimproved evolutionary paths; the corresponding evolutionary paths are... The difference is determined as the exposure difference value; The root arrival weight is obtained by normalizing the root length density in the corresponding spatial unit and time interval of the corresponding evolution path; the root activity weight is obtained by normalizing the soil moisture content, soil temperature and soil pH value of the corresponding evolution path; the exposure difference is the potential available cumulative exposure of the root zone corresponding to the actual improved evolution path minus the potential available cumulative exposure of the root zone corresponding to the virtual unimproved evolution path.

[0012] Furthermore, step S5 specifically includes: The cumulative element content of tobacco plants at each growth stage is determined by summing the products of the dry matter content of different organs of tobacco plants at each growth stage and the corresponding trace element content, and the increment of the cumulative element content of tobacco plants is obtained by the difference between the cumulative element content of tobacco plants at adjacent growth stages. The exposure difference is divided according to the time intervals corresponding to each reproductive stage to obtain the stage-specific increments of the exposure difference. The incremental increase of the cumulative elemental amount of the tobacco plant and the incremental increase of the difference in exposure amount are normalized to the range of 0 to 1. A preset number of growth stages are moved sequentially and the average absolute deviation between the two is calculated. The amount of movement with the smallest average absolute deviation is determined as the absorption hysteresis. The difference between 1 and the corresponding average absolute deviation is determined as the absorption response consistency.

[0013] Further, in step S6, the residual material balance for each trace element is calculated according to the following formula: in, For external input of trace elements, To assess the change in the sum of the water-soluble element reserves, the exchangeable element reserves, and the stationary element reserves between the termination time and the start time of the assessment, This represents the incremental increase in the cumulative elemental content of the tobacco plant. For boundary loss, A pre-defined positive number having the same dimensions as the aforementioned element quantities. The dimensionless residual of the material balance; The external input amount of trace elements is determined based on the elemental composition and application amount in the improved intervention data, the rainfall and the target trace element concentration carried by the rainfall, the irrigation amount and the target trace element concentration in the irrigation water; the boundary loss amount is determined based on the evaluation boundary outflow and the target trace element concentration in the outflow.

[0014] Further, step S6 generates the improved evaluation result, including: When the difference in the effective state content of trace elements between the actual improved evolution path and the virtual unimproved evolution path is lower than the first threshold, it is determined to be a type with insufficient element release or activation. When the difference in the effective content of the trace elements is not lower than the first threshold and the difference in the exposure is lower than the second threshold, it is determined to be a root region inaccessible type. When the difference in exposure is not lower than the second threshold and the consistency of absorption response is lower than the third threshold, it is determined to be an absorption response mismatch type; when the difference in the effective content of trace elements is not lower than the first threshold, the difference in exposure is not lower than the second threshold and the consistency of absorption response is not lower than the third threshold, it is determined to be an improved effective type. When the material balance residual is higher than the fourth threshold, the aforementioned type will be updated to boundary loss or observation mismatch type. The confidence level of the improved assessment results is determined based on the consistency between the material budget residual and the absorption response.

[0015] Furthermore, after generating the improved evaluation result in step S6, the method further includes: Change the micronutrient availability content, root arrival weight, root activity weight, and boundary loss of the spatial unit to be evaluated, respectively, and repeat steps S4 to S6. Determine the sensitivity based on the change in the improved evaluation results caused by each change. Based on the sensitivity from high to low, select the corresponding spatial units and corresponding data types to generate supplementary sampling positions, supplementary sampling depths and supplementary sampling times; The supplementary observation data obtained from the tobacco-planted plots are input into the actual improvement evolution path, and the supplementary observation data obtained from the unimproved control area are input into the virtual unimproved evolution path. After updating the corresponding state sequence, steps S4 to S6 are executed again to generate the updated improvement evaluation results.

[0016] By adopting the above technical solution, the present invention has the following beneficial effects: 1. This invention establishes a real-world improved evolution path and a virtual unimproved evolution path based on the same model structure, common initial state, and shared parameter set. Common external driving data are simultaneously input into the two evolution paths, while improvement intervention data is only input into the real-world improved evolution path. This can generate control evolution results without improvement intervention under the same rainfall, irrigation, and soil temperature conditions, reducing the interference of external environmental changes and natural soil evolution on improvement assessment. 2. This invention generates a continuous state sequence based on soil element form transformation, spatial migration and root absorption process, and correlates the available content of trace elements with root distribution, root activity and action time to obtain the potential available cumulative exposure in the root zone. It can distinguish the difference between the increase in available content of trace elements in the soil and the increase in actual available exposure in the root zone of tobacco plants. 3. This invention performs lag matching on the incremental changes in the cumulative amount of elements in tobacco plants and the incremental changes in the difference between the amount of exposure. It uses the consistency of absorption response to determine whether the changes in root zone exposure have been converted into actual absorption by tobacco plants. It also verifies the degree of closure between element input, soil storage changes, tobacco plant absorption and boundary loss through the material budget residual, thereby forming an improved evaluation result with physical process constraints. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method for evaluating the improvement of trace elements in tobacco-growing soil based on digital twins, according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the data input and evolution relationship of the homologous dual-path digital twin model in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the relationship between trace element morphology transformation, spatial migration, and root absorption within a spatial unit in an embodiment of the present invention. Figure 4 This is a schematic diagram illustrating the calculation relationship between the potential available cumulative exposure in the root region, the consistency of absorption response, and the residual of material budget in an embodiment of the present invention. Figure 5 This is a flowchart illustrating how the present invention determines supplementary sampling locations and updates improved evaluation results based on sensitivity. Detailed Implementation

[0018] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0019] like Figure 1 As shown, this invention provides a method for evaluating the improvement of trace elements in tobacco-growing soil based on digital twins. The method includes the following steps: S1. Obtain baseline data before improvement, calibration data before improvement, common external driving data, improvement intervention data, and post-improvement observation data for tobacco-planted plots. Divide the tobacco-planted plots into spatial units according to sampling locations and soil layer boundaries, and determine the common initial state of each spatial unit based on the baseline data before improvement.

[0020] Step S1 forms the data foundation for the dual-path digital twin model. First, the planar boundaries, evaluation depth, and unified spatial coordinates of the tobacco-planted plots are determined. Then, spatial units are divided based on sampling locations and soil layer boundaries. Pre-modification baseline data represents the soil state before the modification intervention, used to determine the initial state shared by both evolutionary paths. Unmodified calibration data represents the natural evolution of soil micronutrients and tobacco plant uptake processes before the application of the modification materials to be evaluated, used to calibrate the shared parameter set. Common exogenous driving data represents rainfall, irrigation, the concentration of target micronutrients in rainfall and irrigation water, and soil temperature experienced by both evolutionary paths. Modification intervention data represents material inputs occurring only in the actual modification evolution path. Post-modification observation data represents soil, root length density, tobacco plant, and evaluation boundary outflow data obtained from the actual modified plots.

[0021] S2. Based on the unmodified calibration data, calibrate the shared parameter set, establish a dual-path digital twin model including the real modified evolution path and the virtual unmodified evolution path, so that the two evolution paths adopt the same model structure, common initial state and shared parameter set, and lock the shared parameter set before the modification intervention begins.

[0022] like Figure 2 As shown, step S2 establishes a real-world improved evolutionary path and a virtual unimproved evolutionary path. The real-world improved evolutionary path describes the actual state changes of a tobacco-growing plot after receiving improvement intervention; the virtual unimproved evolutionary path describes the counterfactual state changes of the same tobacco-growing plot under the same external environment without the improvement intervention. Both evolutionary paths use the same model structure, common initial state, and shared parameter set. The shared parameter set is calibrated and locked before the improvement intervention begins, and the shared parameter set is not modified in reverse during the evaluation period using the differences between the two paths.

[0023] S3. Input the common exogenous driving data into the two evolutionary paths respectively, input the improvement intervention data only into the actual improvement evolutionary path, and use the improved observation data to correct the actual improvement evolutionary path to obtain the state sequence that represents the soil element form transformation, spatial migration and root absorption process in the two evolutionary paths respectively.

[0024] Step S3 drives two evolutionary paths with a unified time step. Common exogenous driving data is input into both evolutionary paths simultaneously, while improved intervention data is input only into the current improved evolutionary path. At each time step, the soil and plant element states for the next time step are calculated based on the state of the previous time step, the current driving data, and the shared parameter set. Improved observation data only corrects the current improved evolutionary path; the virtual unimproved evolutionary path continues to evolve based on the common initial state, common exogenous driving data, and the shared parameter set.

[0025] S4. Based on the state sequences of the two evolutionary paths and the corresponding root distributions, calculate the potential cumulative exposure of each trace element in the root region, and calculate the exposure difference between the two evolutionary paths. Step S4 corrects the potential effective element mass of each spatial unit for root arrival and root activity, and then accumulates it along the time axis to obtain the potential cumulative exposure of the root region.

[0026] S5. Based on the improved observation data, determine the increment of cumulative elemental content in tobacco plants, and perform lag matching between the increment of cumulative elemental content in tobacco plants and the periodic increment of the difference between exposure levels to obtain the consistency of absorption response. Step S5 compares the increment of cumulative elemental content in tobacco plants with the periodic increment of the difference between exposure levels, and determines the absorption lag and the consistency of absorption response through lag shifting.

[0027] S6. Calculate the material budget residual based on the improvement intervention data, state sequence, increase in cumulative elemental content of tobacco plants, and boundary loss. Generate the improvement assessment results based on the exposure difference, absorption response consistency, and material budget residual. Step S6 establishes the budget relationship between elemental input, soil storage changes, tobacco plant absorption, and boundary loss, and generates the improvement assessment results by combining the exposure difference, absorption response consistency, and material budget residual.

[0028] Steps S1 to S6 constitute a continuous data chain: Step S1 outputs the common initial state and the sorted input data; Step S2 outputs the shared parameter set and two evolution paths; Step S3 outputs two state sequences; Step S4 outputs the exposure difference; Step S5 outputs the absorption hysteresis and absorption response consistency; Step S6 outputs the material budget residual, mismatch type and confidence level.

[0029] In one embodiment, in step S1, the baseline data before improvement includes soil layer boundaries, soil dry bulk density, soil moisture content, soil pH, total trace element content, and available trace element content at each sampling location. Sampling locations are set along the ridge direction, perpendicular to the ridge direction, and soil depth of the tobacco-planted plot, and the sampling locations are assigned to corresponding spatial units. The soil dry bulk density is multiplied by the volume of the spatial unit to obtain the soil dry mass, which is used to convert the trace element content in mass concentration form into element reserves.

[0030] Soil dry mass is calculated based on soil dry bulk density and spatial unit volume. The initial total elemental reserves are calculated based on the total trace element content. Then, based on speciation determination results, adsorption-desorption tests, or unmodified calibration data, the initial total elemental reserves are divided into water-soluble elemental reserves, exchangeable elemental reserves, and fixed elemental reserves. The sum of these three types of soil elemental reserves should correspond to the initial total elemental reserves obtained from the total trace element content.

[0031] Unmodified calibration data can be derived from continuous monitoring data of the same plot before modification or from a concurrent unmodified control area. When using a control area, the soil texture, tobacco variety, planting density, basic fertilization regime, and irrigation method of the control area and the modified treatment area should meet the preset matching requirements. Soil environmental data in the unmodified calibration data are used to determine the environmental correction relationship; trace element speciation data are used to calibrate element transformation and migration parameters; root length density data are used to determine the root arrival weight under unmodified conditions; and tobacco plant element absorption data are used to calibrate root absorption parameters.

[0032] Common externally driven data are organized according to the model's time step. Rainfall and irrigation amounts are summed over their corresponding time intervals, and soil temperature is calculated as a time-weighted average over its corresponding time interval. When concentrated rainfall or irrigation events occur, the start and end times of the event and the amount of water are retained, and data spanning multiple events is not over-smoothed. The concentrations of target trace elements carried by rainfall and those in irrigation water are measured from rainfall and irrigation water samples obtained over their respective time intervals, and multiplied by rainfall and irrigation amounts before being included in the external input of trace elements.

[0033] In addition to the total amount of amendment materials applied, the data on soil improvement interventions also include the target trace element mass fraction, material moisture content, and actual coverage area. Element input quantities are determined based on the application amount and target trace element mass fraction, allocated to corresponding spatial units according to application location, and written into the corresponding calculation time step according to application time. Improved soil environmental data are used to correct environmental state quantities, improved trace element speciation data are used to correct soil element state quantities, and tobacco plant element data are used to determine the incremental increase in cumulative element content in tobacco plants. Improved root length density data are used to determine the root arrival weight of the actual improvement evolution path; the evaluation boundary water discharge and the concentration of target trace elements in the discharge are used to determine the boundary loss.

[0034] In one embodiment, in step S2, the shared parameter set includes at least element release parameters, adsorption-desorption parameters, fixation-conversion parameters, spatial migration parameters, and root absorption parameters. Element release parameters characterize the rate at which fixed elements or elements not yet released from the modified material transform into migratory forms; adsorption-desorption parameters characterize the bidirectional transfer between water-soluble and exchangeable elements; fixation-conversion parameters characterize the transfer of exchangeable elements to fixed elements; spatial migration parameters characterize the movement of water-soluble elements with water between adjacent spatial units; and root absorption parameters characterize the rate at which tobacco roots absorb trace elements.

[0035] The initial range of parameters can be determined through indoor incubation experiments, soil column leaching experiments, adsorption-desorption experiments, or historical data from similar soils. All transfer parameters and root absorption parameters are set to non-negative numbers. The upper limit of the spatial migration parameters is determined based on soil porosity, saturated hydraulic conductivity, and the calculation time step, ensuring that the mass of elements migrating out within a time step does not exceed the original mass of migratable elements in the spatial unit.

[0036] Using the initial calibration state of the unmodified calibration data at the first moment as the starting point of the unmodified simulation, the common external driving data corresponding to the unmodified period are sequentially input into the model to obtain the unmodified simulation sequence. The unmodified simulation sequence is compared with the unmodified calibration data according to the same sampling time and spatial unit. The calibration residuals for the available content of trace elements constrain water-soluble and extractable exchangeable elements; the calibration residuals for the total amount of trace elements constrain the total soil element reserves; and the calibration residuals for the cumulative element content in tobacco plants constrain the root absorption process.

[0037] Parameter calibration can employ a weighted least squares algorithm with upper and lower bound constraints. The calibration residuals are scaled according to the detection standard deviations of the three types of observation data, and the shared parameter set is iteratively adjusted within the allowable parameter range, causing the sum of squares of the scaled calibration residuals to decrease successively. In one implementation, the relative change in the sum of squares of the calibration residuals between two adjacent iterations is no greater than [missing value]. Alternatively, the iteration can be stopped when the number of iterations reaches 200.

[0038] After calibration, nonnegativity and mass conservation checks are performed. If a negative elemental reserve occurs at any time step, the computation time step is shortened or the corresponding transfer flux is limited. If the difference between the elemental input during the simulation and changes in soil reserves, tobacco plant uptake, and boundary losses exceeds the calibration tolerance, the current parameters are not included in the shared parameter set.

[0039] The calibrated shared parameter set was copied to both the real improved evolutionary path and the virtual unimproved evolutionary path. Both paths used the same state variables, spatial unit connectivity, time steps, and parameter values. The parameter version was recorded and locked at the start of the improvement intervention; subsequent observational corrections only changed the state variables of the real improved evolutionary path.

[0040] In one embodiment, such as Figure 3 As shown, in step S3, each spatial unit is configured with water-soluble element reserves, exchangeable element reserves, and stationary element reserves, and the amount of elements absorbed by the tobacco plant is set as the plant element state quantity. All four state quantities use element mass units. Water-soluble element reserves represent the mass of elements present in the soil solution and capable of migrating with water; exchangeable element reserves represent the mass of elements adsorbed on the surface of soil colloids and capable of desorption; stationary element reserves represent the mass of elements that are not easily desorbed directly within the evaluation timescale; and plant element state quantities represent the mass of elements already absorbed by the tobacco plant.

[0041] The modification material release process only affects the actual modification evolution path. For soluble modification materials, the release amount of the target trace element increases the water-soluble element reserves in the applied spatial unit; for slow-release or mineral-based modification materials, the water-soluble element reserves are gradually increased based on the element release parameters and the amount of remaining unreleased elements. The release amount at each time step must not exceed the mass of the target element that has not yet been released in the modification material.

[0042] Water-soluble elements are transferred to exchangeable elements according to adsorption parameters, and exchangeable elements are returned to water-soluble elements according to desorption parameters. The fixation transformation process transfers some exchangeable elements to fixed elements. When the corresponding trace elements and soil conditions permit, some fixed elements can also be transferred back to exchangeable elements based on element release parameters. The transfer rates are adjusted according to soil pH, soil moisture content, soil temperature, and current element reserves.

[0043] Spatial migration is deducted from water-soluble element reserves. The migration amount between adjacent spatial units is determined based on water flux, element concentrations in the soil solution of upstream spatial units, and spatial migration parameters. The element mass deducted from the spatial unit migrating out equals the element mass added to the spatial unit migrating in; the element mass flowing out of the evaluation boundary is included in the boundary loss.

[0044] The root absorption process deducts the amount of absorbable elements from the spatial units with root distribution and adds the same amount of elements to the plant element state quantity. The amount of root absorption at each time step is determined based on the water-soluble element reserves, the desorbable portion of exchangeable elements, root length density, root activity, and root absorption parameters, and must not exceed the element mass that the current spatial unit can supply.

[0045] The available content of trace elements is determined by dividing the sum of water-soluble element reserves and exchangeable element reserves converted by the extraction coefficient by the dry mass of the soil in the spatial unit. Extraction coefficients are element-specific parameters between 0 and 1, and can be calibrated using speciation and available content extraction measurements from the same batch of soil samples. Extraction coefficients are determined separately for different trace elements or different extraction methods.

[0046] Furthermore, in step S3, the actual improvement evolution path is corrected using the improved observation data. This includes: soil moisture content, soil temperature, and soil pH values, after outlier checking and instrument calibration, are written into the same sampling time and spatial unit in the actual improvement evolution path. When the observation location is between multiple spatial units, spatial interpolation can be performed using the inverse distance weight; when the observation time is between two calculation times, temporal interpolation can be used.

[0047] The observed total amount of trace elements is first multiplied by the dry mass of the soil in the spatial unit to obtain the observed total elemental storage value, which is then compared with the sum of water-soluble elemental storage, exchangeable elemental storage, and fixed elemental storage. The observed available content of trace elements is compared with the available content calculated based on soil elemental state parameters, extraction coefficients, and soil dry mass. Tobacco plant elemental data are converted into the cumulative amount of tobacco plant elements through organ dry matter and organ elemental content, and then compared with plant elemental state parameters.

[0048] The element observation residual is the difference between the observed quantity and the model mapping quantity. The total residual of trace elements constrains the sum of the three soil element states; the residual of available trace element content mainly corrects for water-soluble element reserves and exchangeable element reserves; the residual of tobacco plant element accumulation corrects for plant element states.

[0049] Furthermore, correction weights are determined based on the contribution ratio of each state variable to the observed quantity and the detection error. The elemental observation residuals are then allocated to the corresponding state variables according to these correction weights. After correction, it is checked whether each state variable is non-negative and whether the sum of the three soil element state variables meets the trace element total constraint. If the initial correction causes any state variable to be less than zero, the correction is reduced according to the maximum common proportion that keeps all state variables non-negative. The remaining observation residuals are retained as evaluation uncertainty.

[0050] Observational correction only changes the corresponding state variables of the real improved evolutionary path, without changing the virtual unimproved evolutionary path and the shared parameter set, thus maintaining the independence of the virtual unimproved evolutionary path as a counterfactual reference.

[0051] In one embodiment, such as Figure 4 As shown, in step S4, trace elements In evolutionary path The potential available cumulative exposure in the root zone is calculated using the following formula:

[0052] in, Index the target trace elements; For evolutionary path indexing, This represents the actual path of improvement and evolution. Indicates a virtual, unmodified evolutionary path; For spatial unit indexing; For time interval index; This refers to the number of spatial units; This represents the number of time intervals.

[0053] For the first Within the time interval, the first The effective content of trace elements in a spatial unit can be expressed as the average value of the effective content at the start and end of a time interval. For the first Soil dry mass of each spatial unit; Weights for root system reach; The root activity weights are obtained by normalizing based on soil moisture content, soil temperature, and soil pH. For the first The duration of each time interval.

[0054] The value is obtained by normalizing the root length density, ranging from 0 to 1. Root length density can be obtained through root drilling, root scanning, or micro-root canal imaging. Spatial cells with no root distribution are assigned a value of 0. The root activity weight is determined based on the location of soil moisture content, soil temperature, and soil pH relative to the suitable range of tobacco plant roots, with values ​​ranging from 0 to 1. Three environmental response coefficients can be determined separately, and then multiplied together to obtain the root activity weight.

[0055] For each evolutionary path and time interval, the root arrival weight of a spatial unit is calculated by dividing the root length density of that spatial unit by the sum of the root length densities of all spatial units within the same evolutionary path and time interval; when the denominator is zero, the root arrival weight of each spatial unit is zero. The root length density of the actual improved evolutionary path is interpolated temporally and spatially based on the improved observation data, while the root length density of the virtual unimproved evolutionary path is determined based on the data of the unimproved control area during the same period; when no unimproved control area is set during the same period, it is determined by interpolation according to the growth stage based on the unimproved calibration data. The environmental response coefficients corresponding to soil moisture content, soil temperature, and soil pH are taken as one within the suitable range, linearly decreasing to zero between the boundary of the suitable range and the tolerance boundary, and taken as zero when exceeding the tolerance boundary.

[0056] This represents the potential effective element mass within the corresponding spatial unit. Both root arrival weight and root activity weight are dimensionless. Multiplying by the time interval length... The dimension of is the product of element mass and time. The two evolutionary paths correspond to... The difference is the exposure difference. In this embodiment, the exposure difference is the potential available cumulative exposure of the root region of the actual improved evolutionary path minus the potential available cumulative exposure of the root region of the virtual unimproved evolutionary path.

[0057] In one embodiment, in step S5, root, stem, and leaf samples of tobacco plants are collected at each preset growth stage, and the dry matter mass and target trace element content of each organ are measured. The dry matter mass of each organ is multiplied by the corresponding trace element content and then summed to obtain the cumulative element amount of the tobacco plant at the corresponding growth stage. When the number of sampled plants is greater than one, it is uniformly converted to the cumulative amount per plant or per unit area. The difference in the cumulative element amount of tobacco plants between adjacent growth stages is the increment of the cumulative element amount of the tobacco plant.

[0058] Exposure differences are determined based on the sampling time of tobacco plants. The cumulative exposure difference at the end of each time interval is subtracted from the cumulative exposure difference at the beginning of that time interval to obtain the period increment of the exposure difference. The cumulative elemental increment of tobacco plants and the period increment of the exposure difference are normalized to 0-1 using the minimum and maximum values, respectively. When the maximum and minimum values ​​of the sequence are the same, a preset positive number is added to the denominator or the sequence is set to 0 to avoid division by zero.

[0059] The candidate range for absorption lags was determined based on the number of tobacco growth stages and the sampling interval. For each candidate lag, the phased increment sequence of the moving exposure difference was used to retain the sequence items that shared a common time range with the cumulative elemental increments in the tobacco plant, and the mean absolute deviation was calculated. The moving quantity with the smallest mean absolute deviation was determined as the absorption lag, and the absorption response consistency was obtained by subtracting the corresponding mean absolute deviation from one. The normalized sequence and its mean absolute deviation were both within the range of 0 to 1, therefore the absorption response consistency was also within the range of 0 to 1.

[0060] In one embodiment, in step S6, the residual material balance of each trace element is calculated according to the following formula:

[0061] in, The external input of trace elements into the evaluation boundary of tobacco-growing plots during the evaluation period includes the mass of target trace elements carried by amendment materials, irrigation water, and rainfall; To evaluate the change in the sum of water-soluble element reserves, exchangeable element reserves, and fixed element reserves between the end time and the start time of the evaluation; This represents the increase in the cumulative elemental content of tobacco plants during the evaluation period; The target trace element mass flowing out of the evaluation boundary; A pre-defined positive number having the same dimensions as the aforementioned element quantities; It is a dimensionless material balance residual.

[0062] The total soil element reserves at the end of the calculation are subtracted from the total soil element reserves at the beginning of the calculation. A positive value indicates an increase in soil reserves, while a negative value indicates a decrease. Plant element state quantities are not included. This avoids double counting of tobacco plant absorption. This includes leaching at the lower boundary of the root zone and surface runoff loss, which can be obtained by multiplying the outflow volume by the target trace element concentration of the outflow water body and summing them over time. It can be determined based on the limit of quantitation of the detection method and the dry weight of the soil in the smallest sampling unit.

[0063] The numerator represents the elemental mass not yet explained by changes in soil storage, tobacco plant uptake, and boundary losses; the denominator scales the elemental mass. When the value is negative, the molecule contains The release of the original elemental reserves in the soil is included in the sources of elements that can be absorbed by tobacco plants or lost at the boundary, and the symbolic relationship conforms to the law of conservation of matter.

[0064] In one embodiment, step S6 generates an improvement evaluation result, including: when the difference in the effective content of trace elements between the actual improved evolution path and the virtual unimproved evolution path is lower than a first threshold, it is determined to be an insufficient element release or activation type. When the difference in effective content is not lower than the first threshold and the difference in exposure is lower than a second threshold, it is determined to be a root region inaccessible type. When the difference in exposure is not lower than the second threshold and the consistency of absorption response is lower than a third threshold, it is determined to be an absorption response mismatch type.

[0065] When the difference in available content is not lower than the first threshold, the difference in exposure is not lower than the second threshold, and the consistency of absorption response is not lower than the third threshold, it is determined to be an improved and effective type. When the residual of material budget is higher than the fourth threshold, the aforementioned type is updated to boundary loss or observation mismatch type. The difference in available content of trace elements is the difference in available content of trace elements by soil dry mass weighted average between the actual improved evolution path and the virtual unimproved evolution path in the root zone spatial unit at the end of the evaluation.

[0066] The first threshold can be determined based on the statistical distribution of the difference between the observed effective content and the simulated effective content in the unmodified calibration data; the second threshold can be determined based on the error distribution of the exposure difference during the unmodified period; the third threshold can be determined based on the distribution of the consistency of the absorption response under normal absorption conditions; and the fourth threshold can be determined based on the distribution of the material balance residuals during the unmodified calibration period or the combined uncertainty of the detection error and the flow measurement error.

[0067] The classification is performed in the order of element activation, root zone arrival, tobacco plant absorption, and material balance verification. When the material balance residual exceeds the fourth threshold, regardless of the type of the previous judgment, it is updated to boundary loss or observation mismatch. If the material balance residual is not higher than the fourth threshold and the absorption response consistency is not lower than the third threshold, it is determined to be of high confidence level; if only one of the two conditions is met, it is determined to be of medium confidence level; if neither condition is met, it is determined to be of low confidence level.

[0068] The improved assessment results should include at least the target trace element, the difference in the effective content of trace elements, the difference in exposure, the absorption hysteresis, the consistency of absorption response, the material budget residual, the assessment type and confidence level, and retain the corresponding spatial unit, time interval and the associated index of the input data.

[0069] In one embodiment, such as Figure 5 As shown, after generating the improved evaluation results, the trace element effective content, root arrival weight, root activity weight, and boundary loss of the spatial unit to be evaluated are changed respectively, and steps S4 to S6 are repeated. The sensitivity is determined based on the change in the improved evaluation results. The perturbation amplitude can be ±5% of the detection standard deviation, model estimation standard deviation, or baseline value, and a uniform rule is used for the same analysis. The change in the evaluation results is the maximum of the absolute changes before and after the perturbation of the exposure difference, absorption response consistency, and material budget residuals, which are normalized according to the minimum and maximum values ​​in the unimproved calibration data, respectively; when the maximum and minimum values ​​are the same, one is used as the normalization denominator; when the evaluation type changes, the change in the evaluation results is taken as one. The sensitivity is the ratio of the change in the evaluation results to the relative perturbation amplitude of the corresponding input quantity; when the baseline value of the corresponding input quantity is zero, the relative perturbation amplitude is determined by dividing the absolute value of the perturbation by the preset allowable range of the input quantity.

[0070] Each time, only one type of data in a spatial unit is changed, while other input data and the shared parameter set remain unchanged. When the sensitivity of available micronutrient content is high, soil samples are collected again in the corresponding spatial unit and soil layer; when the sensitivity of root arrival weight is high, root drilling or micro-root canal observation is supplemented; when the sensitivity of root activity weight is high, soil moisture content, soil temperature, soil pH, or root activity are measured again; when the sensitivity of boundary loss is high, water flux and target micronutrient concentration are measured again at the lower boundary of the root zone or the surface runoff outlet.

[0071] Supplementary observation data obtained from tobacco-planted plots are input into the actual improved evolution path, while supplementary observation data obtained from unimproved control areas are input into the virtual unimproved evolution path. When no concurrent unimproved control area is set, the virtual unimproved evolution path maintains its original state sequence and does not accept supplementary observation data from the improved treatment areas. After updating the corresponding state sequence, steps S4–S6 are re-executed to generate updated improvement evaluation results.

[0072] Specific application examples: Example 1: This example uses zinc in tobacco-grown soil as the evaluation object. A tobacco-grown plot with an area of ​​600 m² was selected and divided into 24 planar areas with a planar scale of 5 m × 5 m. Each planar area was then divided into two soil layers: 0–20 cm and 20–40 cm, forming a total of 48 spatial units.

[0073] Before applying the amendment materials, soil dry bulk density, soil moisture content, soil pH, total zinc content, and available zinc content were measured. The average soil pH in the 0–20 cm soil layer was 6.2, and the average available zinc content was 0.72 mg / kg. Soil dry mass was calculated based on spatial unit volume and soil dry bulk density. Then, based on the total zinc content, available zinc content, and speciation results, the water-soluble zinc reserves, exchangeable zinc reserves, and fixed zinc reserves were determined to form a common initial state.

[0074] Continuous monitoring data from an unmodified control area with the same soil type were used as calibration data for unmodified soil. A shared parameter set was calibrated using a weighted least squares algorithm with boundary constraints. Subsequently, both the actual modified evolution path and the virtual unmodified evolution path were replicated and the shared parameter set was locked. The two evolution paths used a one-day computation time step, with an evaluation period of 120 days.

[0075] The application rate of zinc-containing amendments was 75 kg per hectare, with a zinc mass fraction of 2%, resulting in a zinc input of 90 g per 600 m² plot. The amendments were applied in strips into the 0–20 cm soil layer. Only the actual amendment evolution path was entered into the intervention data. Actual rainfall, irrigation volume, and soil temperature were simultaneously entered into two evolution paths. During the evaluation period, zinc concentrations in rainfall and irrigation water samples were measured concurrently. Results below the limit of quantitation were counted as zero as the external zinc input.

[0076] Soil samples were collected on days 30, 60, 90, and 120 after application. Root length density was measured on days 30, 60, and 90. Tobacco plant organ samples were collected during the rosette stage, vigorous growth stage, and maturity stage. Soil environmental data and elemental observation data were corrected for the actual improvement evolution path according to the aforementioned mapping relationship. Simultaneously, water outflow was recorded at the lower boundary of the root zone and at the surface runoff outlet, and the zinc concentration in the outflow was measured to calculate boundary losses.

[0077] The calculated potential cumulative exposure of the root region of the actual improved evolutionary path is as follows: mg·d, the potential cumulative exposure available in the root region of the virtual unmodified evolutionary pathway is mg·d, the difference in exposure is mg·d. The phased increments of the difference between the cumulative zinc content in tobacco plants and the exposure were normalized, and the candidate lags for 0, 1, and 2 growth stages were compared. The mean absolute deviation was smallest when the lag was for 1 growth stage, and the corresponding absorption response consistency was 0.88.

[0078] During the evaluation period, the external zinc input was 90,000 mg, the increase in soil zinc reserves was 81,000 mg, the increase in cumulative zinc in tobacco plants was 3,200 mg, and the boundary loss was 5,100 mg. Based on the material balance residual formula, the numerator is 700 mg, and the denominator, ignoring smaller amounts... The value was 179300 mg, and the residual of the substance balance was approximately 0.0039. The calculated result was consistent with the dimensions of all data.

[0079] In this embodiment, if the difference in the effective zinc content is higher than the first threshold, the difference in exposure is higher than the second threshold, the consistency of absorption response is higher than the third threshold, and the residual of material balance is lower than the fourth threshold, it is determined to be an improved effective type, and a high-confidence improvement evaluation result is generated.

[0080] Example 2: This example illustrates a situation where the availability of trace elements increases but root zone accessibility is insufficient. The improved material is applied via hole application, primarily at a depth of 25–35 cm. The tobacco-planted plots are divided into soil spatial units of 0–10 cm, 10–20 cm, 20–30 cm, and 30–40 cm, and a common initial state, shared parameter set, and two evolutionary paths are determined according to the aforementioned method.

[0081] The evaluation results showed that the available content of trace elements in the 20-40cm soil layer increased significantly, and the difference in available content between the actual improved evolution path and the virtual unimproved evolution path reached the first threshold. However, during the evaluation period, the tobacco plant roots were mainly distributed in the 0-20cm soil layer, and the root reach weight in the areas where the improved materials were concentrated was low. Some deep spatial units had low root activity weights due to high water content.

[0082] After combining the available micronutrient content with root access weight, root activity weight, and time interval length, the exposure difference was found to be below the second threshold, indicating a root-inaccessible type in the improvement assessment. Sensitivity analysis showed that the root access weight in the 20–30 cm soil layer had a significant impact on the assessment results; therefore, root drilling was increased in the corresponding soil layer, and root length density was measured again during the vigorous growth period. After supplementing the observation data with the actual improvement evolution path, the corresponding state sequence was updated, and the improvement assessment results were regenerated.

[0083] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for evaluating the improvement of trace elements in tobacco-growing soil based on digital twinning, characterized by, include: S1. Obtain baseline data before improvement, calibration data before improvement, common external driving data, improvement intervention data and observation data after improvement for the tobacco-planted plots. Divide the tobacco-planted plots into spatial units according to the sampling location and soil layer boundary, and determine the common initial state of each spatial unit based on the baseline data before improvement. S2. Based on the unmodified calibration data, calibrate the shared parameter set, establish a dual-path digital twin model including the real modified evolution path and the virtual unmodified evolution path, so that the two evolution paths adopt the same model structure, the common initial state and the shared parameter set, and lock the shared parameter set before the modification intervention begins; S3. Input the common exogenous driving data into the two evolution paths respectively, input the improvement intervention data only into the actual improvement evolution path, and use the improved observation data to correct the actual improvement evolution path to obtain state sequences that characterize the soil element form transformation, spatial migration and root absorption processes in the two evolution paths respectively. S4. Based on the state sequences and corresponding root distributions of the two evolutionary paths, calculate the potential cumulative exposure of each trace element in the root region, and calculate the exposure difference between the two evolutionary paths. S5. Based on the improved observation data, determine the increment of the cumulative element amount of tobacco plants, and perform lag matching between the increment of the cumulative element amount of tobacco plants and the increment of the difference between the exposure amounts to obtain the consistency of the absorption response. S6. Calculate the material budget residual based on the improved intervention data, the state sequence, the cumulative increase of tobacco element and the boundary loss, and generate the improved evaluation result based on the exposure difference, the absorption response consistency and the material budget residual.

2. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, Step S1 specifically includes: The baseline data before improvement includes soil layer boundaries, soil dry bulk density, soil moisture content, soil pH, total trace elements and available trace element content at each sampling location, which are used to determine the common initial state of each spatial unit. The unmodified calibration data includes soil environmental data, trace element speciation data, root length density data, and tobacco plant element absorption data for the corresponding spatial unit during the period when no modification materials were applied, which are used to calibrate the shared parameter set. The common exogenous driving data includes rainfall, the concentration of target trace elements carried by rainfall, irrigation amount, the concentration of target trace elements in irrigation water, and soil temperature. The improvement intervention data includes the elemental composition, application amount, application location, and application time of the improvement material. The post-improvement observation data includes post-improvement soil environmental data, trace element speciation data, root length density data, tobacco plant elemental data, evaluation boundary water discharge, and the concentration of target trace elements in the discharge water.

3. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, Step S2 specifically includes: Evolution calculations are performed based on the initial calibration state at the first moment of the unmodified calibration data and the common external driving data during the corresponding unmodified period to obtain the unmodified simulation sequence; Calculate the calibration residuals between the effective content of trace elements, the total amount of trace elements, and the cumulative amount of elements in tobacco plants in the unmodified simulated sequence and the unmodified calibration data, respectively; The element release parameters, adsorption-desorption parameters, fixed transformation parameters, spatial migration parameters, and root absorption parameters are iteratively adjusted according to the calibration residual until the calibration residual meets the preset calibration conditions. The adjusted parameters are then determined as the shared parameter set and replicated to form the real improved evolution path and the virtual unimproved evolution path.

4. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, Step S3 specifically includes: The water-soluble element reserves, exchangeable element reserves, and fixed element reserves in each spatial unit are set as soil element state quantities, and the element reserves absorbed by the tobacco plants are set as plant element state quantities. The water-soluble element reserves are increased according to the improved material release process, the water-soluble element reserves are transferred between the water-soluble element reserves and the exchangeable element reserves according to the adsorption and desorption process, the water-soluble element reserves are transferred between the exchangeable element reserves and the fixed element reserves according to the fixation and transformation process, and the water-soluble element reserves and plant element state quantities of adjacent spatial units are updated according to the water flux and root absorption process. The available content of trace elements is determined by dividing the sum of the water-soluble element reserves and the exchangeable element reserves converted by the extraction coefficient by the dry mass of the soil in the spatial unit.

5. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 4, characterized in that, Step S3 involves using the improved observation data to correct the actual improved evolutionary path, including: The soil moisture content, soil temperature, and soil pH value in the improved soil environmental data are written into the environmental state quantities of the corresponding time and spatial unit in the actual improvement evolution path. The total amount of observed trace elements is multiplied by the dry mass of the soil in the corresponding spatial unit and mapped to the sum of the three soil element state quantities. The observed available content of trace elements is mapped to the available content of trace elements determined in step S3, and the element data of tobacco plants is mapped to the plant element state quantities. The element observation residuals are obtained respectively. The corresponding state quantities of the actual improved evolutionary path are corrected based on the element observation residuals, and the correction amount is reduced proportionally when the correction amount causes any state quantity to be less than zero; the correction process does not change the virtual unimproved evolutionary path and the shared parameter set.

6. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, In step S4, trace elements In evolutionary path The potential available cumulative exposure in the root zone is calculated using the following formula: in, This refers to the content of trace elements in their available form. For soil dry weight, The root arrival weights are obtained by normalizing the root length density. As the root activity weight, The length of the time interval. The number of spatial units. The number of time intervals. For real-world improved evolutionary paths or virtual unimproved evolutionary paths; the corresponding evolutionary paths are... The difference is determined as the exposure difference value; The root arrival weight is obtained by normalizing the root length density in the corresponding spatial unit and time interval of the corresponding evolution path; the root activity weight is obtained by normalizing the soil moisture content, soil temperature and soil pH value of the corresponding evolution path; the exposure difference is the potential available cumulative exposure of the root zone corresponding to the actual improved evolution path minus the potential available cumulative exposure of the root zone corresponding to the virtual unimproved evolution path.

7. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, Step S5 specifically includes: The cumulative element content of tobacco plants at each growth stage is determined by summing the products of the dry matter content of different organs of tobacco plants at each growth stage and the corresponding trace element content, and the increment of the cumulative element content of tobacco plants is obtained by the difference between the cumulative element content of tobacco plants at adjacent growth stages. The exposure difference is divided according to the time intervals corresponding to each reproductive stage to obtain the stage-specific increments of the exposure difference. The incremental increase of the cumulative elemental amount of the tobacco plant and the incremental increase of the difference in exposure amount are normalized to the range of 0 to 1. A preset number of growth stages are moved sequentially and the average absolute deviation between the two is calculated. The amount of movement with the smallest average absolute deviation is determined as the absorption hysteresis. The difference between 1 and the corresponding average absolute deviation is determined as the absorption response consistency.

8. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, In step S6, the residual material balance for each trace element is calculated according to the following formula: in, For external input of trace elements, To assess the change in the sum of the water-soluble element reserves, the exchangeable element reserves, and the stationary element reserves between the termination time and the start time of the assessment, This represents the incremental increase in the cumulative elemental content of the tobacco plant. For boundary loss, A pre-defined positive number having the same dimensions as the aforementioned element quantities. The dimensionless residual of the material balance; The external input amount of trace elements is determined based on the elemental composition and application amount in the improved intervention data, the rainfall and the target trace element concentration carried by the rainfall, the irrigation amount and the target trace element concentration in the irrigation water; the boundary loss amount is determined based on the evaluation boundary outflow and the target trace element concentration in the outflow.

9. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, Step S6 generates the improved evaluation result, including: When the difference in the effective state content of trace elements between the actual improved evolution path and the virtual unimproved evolution path is lower than the first threshold, it is determined to be a type with insufficient element release or activation. When the difference in the effective content of the trace elements is not lower than the first threshold and the difference in the exposure is lower than the second threshold, it is determined to be a root region inaccessible type. When the difference in exposure is not lower than the second threshold and the consistency of absorption response is lower than the third threshold, it is determined to be an absorption response mismatch type; when the difference in the effective content of trace elements is not lower than the first threshold, the difference in exposure is not lower than the second threshold and the consistency of absorption response is not lower than the third threshold, it is determined to be an improved effective type. When the material balance residual is higher than the fourth threshold, the aforementioned type will be updated to boundary loss or observation mismatch type. The confidence level of the improved assessment results is determined based on the consistency between the material budget residual and the absorption response.

10. The method for evaluating trace element improvement in tobacco-growing soil based on digital twins according to claim 1, characterized in that, After generating the improved evaluation result in step S6, the method further includes: Change the micronutrient availability content, root arrival weight, root activity weight, and boundary loss of the spatial unit to be evaluated, respectively, and repeat steps S4 to S6. Determine the sensitivity based on the change in the improved evaluation results caused by each change. Based on the sensitivity from high to low, select the corresponding spatial units and corresponding data types to generate supplementary sampling positions, supplementary sampling depths and supplementary sampling times; The supplementary observation data obtained from the tobacco-planted plots are input into the actual improvement evolution path, and the supplementary observation data obtained from the unimproved control area are input into the virtual unimproved evolution path. After updating the corresponding state sequence, steps S4 to S6 are executed again to generate the updated improvement evaluation results.