Soil and water conservation carbon sink calculation method and system based on artificial intelligence

Abstract

The present application relates to the field of environmental science and technology, and discloses a soil and water conservation carbon sink calculation method and system based on artificial intelligence, which comprises the following steps: obtaining soil parameters, climate parameters, topographic parameters and environmental parameters, performing adaptive uncertainty quantification processing, and constructing a standardized data set; based on the standardized data set, combining a biomass-carbon sink dynamic constraint model, generating forest and grass carbon sink data; through a microtopography-carbon sink dynamic correction model, generating engineering carbon sink data; further through a carbon turnover-cultivation interactive stable feedback model, generating cultivation carbon sink data; performing nonlinear collaborative causal reasoning on the forest and grass carbon sink data, the engineering carbon sink data and the cultivation carbon sink data, and fusing to generate total carbon sink data; and finally performing error correction prediction on the total carbon sink data, and outputting carbon sink prediction results. The present application realizes precise quantification of soil and water conservation carbon sink through multi-measure collaboration and multi-path interweaving, and effectively improves the calculation accuracy and reliability.

Description

Technical Field

[0001] This invention relates to the field of environmental science and technology, and in particular to a method and system for calculating carbon sequestration in soil and water conservation based on artificial intelligence. Background Technology

[0002] In the field of carbon sequestration for soil and water conservation, traditional methods often rely on single-source parameters or static data processing models, lacking the ability to systematically integrate multi-source parameters such as soil, climate, topography, and environment. Existing technologies have not established an effective adaptive uncertainty quantification mechanism, and cannot dynamically identify the fluctuation characteristics and reliability differences of multi-source data. They only eliminate the influence of dimensions through simple normalization, resulting in high uncertainty data interfering with the calculation results. It is difficult to form an accurate standardized dataset, which directly restricts the quality of basic data for carbon sequestration. At the same time, traditional data processing does not consider the dynamic changes of parameter time series, and cannot capture local fluctuations and long-term trends of data through adaptive window adjustment, further aggravating the accumulation of errors in the data preprocessing stage.

[0003] Existing carbon sequestration methods suffer from insufficient separation and measurement of multi-source carbon sequestrations and a lack of synergistic integration mechanisms. Traditional technologies often only calculate carbon sequestrations of a single type, such as forestry, grassland, engineering, or agriculture, ignoring the dynamic interaction effects between different carbon sequestration types. Furthermore, they lack scientific synergistic integration models. Most methods use linear weighting to integrate multi-source carbon sequestration data, failing to quantify the nonlinear causal relationships and dynamic contribution weights between carbon sequestration types. This results in total carbon sequestration data that does not accurately reflect the true carbon sequestration pattern. In addition, existing technologies lack dynamic correction and error feedback mechanisms for the entire carbon sequestration process, making it difficult to effectively correct systematic biases in biomass estimation, carbon sequestration conversion, and integration. Ultimately, this leads to insufficient accuracy in carbon sequestration results, failing to provide reliable data support for soil and water conservation decisions and carbon sequestration management. Therefore, improving the accuracy of carbon sequestration results has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method and system for calculating carbon sequestration in soil and water conservation based on artificial intelligence, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides an artificial intelligence-based method for calculating soil and water conservation carbon sequestration, comprising:

[0006] S1, adaptive uncertainty quantification is performed on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to construct a standardized soil dataset;

[0007] S2, Based on the standardized dataset, the biomass-carbon sequestration dynamic constraint of the soil is calculated using the biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data.

[0008] S3. Based on the standardized dataset, the soil is subjected to micro-topography-carbon sequestration dynamic correction calculation through the micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data.

[0009] S4. Based on the standardized dataset, the carbon turnover-tillage interaction stability feedback of the soil is calculated using the carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data.

[0010] S5, perform nonlinear collaborative causal reasoning on the forest and grassland carbon sink data, engineering carbon sink data and cultivated carbon sink data to generate the total carbon sink data of the soil.

[0011] S6, perform error correction prediction on the total carbon sink data to generate predicted carbon sink data for the soil.

[0012] In a preferred embodiment, the adaptive uncertainty quantification processing of the pre-acquired soil parameters, climate parameters, topographic parameters, and environmental parameters to generate a standardized soil dataset includes:

[0013] S201, perform data cleaning on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to obtain multi-source clean parameters of the soil;

[0014] S202, perform adaptive uncertainty quantification analysis on the multi-source cleaning parameters to obtain the uncertainty quantification value of the soil;

[0015] S203, perform weighted standardization on the uncertainty quantification value and the multi-source cleanliness parameters to obtain the standardized parameter set of the soil;

[0016] S204, Combine the standardized parameter set to construct the standardized dataset of the soil.

[0017] In a preferred embodiment, the adaptive uncertainty quantification analysis of the multi-source cleanliness parameters to obtain the uncertainty quantification value of the soil includes:

[0018] S301, Analyze the parameter sequence change rate of the multi-source cleaning parameters to obtain the soil change rate sequence;

[0019] S302, Adaptively adjust the window size of the rate of change sequence to obtain an adaptive window size sequence;

[0020] S303, based on the adaptive window size sequence, calculate the in-window variance of the multi-source cleaning parameters to obtain the uncertainty quantification value of the soil.

[0021] In a preferred embodiment, based on the standardized dataset, the biomass-carbon sequestration dynamic constraint of the soil is calculated using a biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data, including:

[0022] S401, Based on the vegetation index parameters, soil nutrient parameters and climate moisture parameters in the standardized dataset, construct a biomass-related parameter set;

[0023] S402, Estimate the initial biomass of the biomass-related parameter set to obtain a biomass sequence;

[0024] S403, perform carbon sequestration on the biomass sequence to obtain forest and grassland carbon sequestration;

[0025] S404, based on the biomass-carbon sink dynamic constraint model, the constraint adjustment of the forest and grassland carbon sink amount is carried out to generate forest and grassland carbon sink data.

[0026] In a preferred embodiment, the step of performing micro-topography-carbon sequestration dynamic correction calculations on the soil based on the standardized dataset using a micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data includes:

[0027] S501, Based on the micro-topographical parameters and climate impact parameters in the standardized dataset, construct the correction-related parameters;

[0028] S502, Based on the micro-topography-carbon sink dynamic correction model, the correction-related parameters are dynamically corrected by micro-topography to obtain the correction factor sequence;

[0029] S503, Based on the standardized dataset, the initial engineering carbon sink is obtained through the initial engineering carbon sink calculation model to obtain the engineering carbon sink of the soil;

[0030] S504, Based on the correction factor sequence, the engineering carbon sequestration amount is multiplied by the correction multiplier to perform dynamic correction processing on the engineering carbon sequestration amount, so as to generate engineering carbon sequestration data.

[0031] In a preferred embodiment, the step of calculating the carbon turnover-tillage interaction stability feedback of the soil based on the standardized dataset using a carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data includes:

[0032] S601, extract carbon turnover parameters and tillage management parameters from the standardized dataset to obtain interactively related parameters;

[0033] S602, Based on the aforementioned interactive correlation parameters, analyze the impact of carbon turnover parameters and tillage management parameters on soil carbon sequestration stability to obtain a feedback factor sequence;

[0034] S603, based on the aforementioned feedback factor sequence, the amount of cultivated carbon sink is adjusted through a stable feedback model of carbon turnover-cultivation interaction to generate cultivated carbon sink data.

[0035] In a preferred embodiment, the step of performing nonlinear co-causal inference on the forestry and grassland carbon sequestration data, engineering carbon sequestration data, and cultivated carbon sequestration data to generate the total soil carbon sequestration data includes:

[0036] S701, perform time series alignment processing on the forest and grassland carbon sequestration data, engineering carbon sequestration data and cultivated carbon sequestration data to obtain the aligned carbon sequestration sequence of the soil.

[0037] S702, perform nonlinear collaborative causal reasoning collaborative fusion weight calculation on the aligned carbon sink sequence to obtain a weight sequence;

[0038] S703, perform carbon sink synergistic contribution fusion on the weighted sequence and the aligned carbon sink sequence to generate the total carbon sink data of the soil.

[0039] In a preferred embodiment, the aligned carbon sink sequence undergoes nonlinear co-causal inference and co-fusion weight calculation to obtain a weight sequence, including:

[0040] S801, The initial weight values ​​are calculated using a nonlinear cocausal inference algorithm on the aligned carbon sink sequence to obtain an initial weight sequence. The mathematical expression of the nonlinear cocausal inference algorithm is as follows:

[0041] ;

[0042] In the formula: It is a type of carbon sink. At the point of time The initial weight values, It is an exponential function. It is a type of carbon sink. The weighting coefficients, It is the natural logarithm function. It is a type of carbon sink. At the point of time carbon sequestration value, It is a very small positive number. It is the synergistic effect coefficient. It is a type of carbon sink. For carbon sink types The causal influence coefficient, Indicates the index of the current carbon sink type. Index representing a point in time. Indexes representing other carbon sink types, It is a type of carbon sink. and At the point of time The product term;

[0043] S802, the initial weight sequence is normalized to obtain a weight sequence.

[0044] In a preferred embodiment, the step of performing error correction prediction on the total carbon sink data to generate predicted carbon sink data for the soil includes:

[0045] S901, Perform historical error analysis on the total carbon sink data to obtain the error sequence of the soil;

[0046] S902, The error sequence is corrected using an error correction algorithm to obtain the corrected value sequence of the soil.

[0047] S903, Based on the correction value sequence, the total carbon sink data is corrected to generate the predicted carbon sink data for the soil.

[0048] To address the above problems, this invention also provides an artificial intelligence-based carbon sequestration calculation system for soil and water conservation, the system comprising:

[0049] The adaptive uncertainty quantization module performs adaptive uncertainty quantization on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to construct a standardized soil dataset.

[0050] The forest and grassland carbon sequestration module, based on the standardized dataset, calculates the dynamic constraints of biomass and carbon sequestration in the soil using a biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data.

[0051] The engineering carbon sequestration module, based on the standardized dataset, performs micro-topography-carbon sequestration dynamic correction calculations on the soil using a micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data.

[0052] The tillage carbon sequestration module, based on the standardized dataset, uses a carbon turnover-tillage interaction stability feedback model to calculate the soil carbon turnover-tillage interaction stability feedback, thereby generating tillage carbon sequestration data.

[0053] The total carbon sink module performs nonlinear co-causal inference on the forestry and grassland carbon sink data, engineering carbon sink data and cultivated carbon sink data to generate the total carbon sink data of the soil.

[0054] The error correction prediction module performs error correction prediction on the total carbon sink data to generate the predicted carbon sink data for the soil.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. At the data processing level, this invention significantly improves the quality and reliability of basic data for carbon sink measurement by constructing an adaptive uncertainty quantification and dynamic weighted standardization data processing mechanism. This mechanism first performs targeted data cleaning on multi-source parameters such as soil, climate, topography, and environment to eliminate missing values, outliers, and format differences. Then, it adaptively adjusts the sliding window size by analyzing the rate of change of parameter sequences and accurately calculates the variance within the window to quantify data uncertainty. Finally, it dynamically assigns weights based on the uncertainty values ​​and completes the standardization process. This process can not only effectively identify and reduce the impact of highly volatile and unreliable data, but also capture the local dynamics and long-term trends of data through adaptive windows, avoiding the problem of error accumulation in traditional static data processing. It provides a standardized dataset with strong consistency and high reliability for subsequent carbon sink measurement, ensuring the accuracy of the measurement from the source.

[0057] 2. In terms of carbon sink measurement and integration, this invention relies on a multi-source carbon sink dynamic measurement and nonlinear pattern collaborative integration mechanism to achieve accurate measurement and scientific integration of carbon sink data. On the one hand, dynamic measurement logic is designed for three types of carbon sinks: forestry and grassland, engineering, and cultivation. Forestry and grassland carbon sink data are generated based on biomass-carbon sink dynamic constraints, engineering carbon sink data is obtained by combining micro-topography-carbon sink dynamic correction, and cultivation carbon sink data is obtained through carbon turnover-cultivation interaction and stable feedback. This ensures that the measurement of each type of carbon sink can fit its ecological processes and influencing factors. On the other hand, the traditional linear weighting method is abandoned. A unified time benchmark for multi-source carbon sinks is achieved by aligning time series, and dynamic weights are calculated using a nonlinear collaborative causal reasoning algorithm. This fully considers the nonlinear interaction effects and causal relationships between carbon sink types, and finally achieves the collaborative contribution integration of multi-source carbon sinks. This mechanism avoids the one-sidedness of single carbon sink measurement and solves the defect that traditional integration methods cannot reflect the true pattern of carbon sinks, greatly improving the accuracy of total carbon sink data. At the same time, the results are further optimized by combining error correction prediction, providing a more reliable decision-making basis for soil and water conservation carbon sink assessment and management. Attached Figure Description

[0058] Figure 1 A flowchart illustrating an artificial intelligence-based carbon sequestration calculation method for soil and water conservation provided in an embodiment of the present invention;

[0059] Figure 2 A functional block diagram of an artificial intelligence-based carbon sequestration calculation system for soil and water conservation provided in an embodiment of the present invention;

[0060] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0061] 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.

[0062] This application provides an artificial intelligence-based method for calculating carbon sequestration in soil and water conservation. The executing entity of this AI-based method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the AI-based method for calculating carbon sequestration in soil and water conservation can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0063] Reference Figure 1 The diagram shown is a flowchart illustrating an artificial intelligence-based carbon sequestration calculation method for soil and water conservation according to an embodiment of the present invention. In this embodiment, the artificial intelligence-based carbon sequestration calculation method for soil and water conservation includes:

[0064] S1, adaptive uncertainty quantification is performed on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to construct a standardized soil dataset;

[0065] In this embodiment of the invention, the adaptive uncertainty quantification processing of the pre-acquired soil parameters, climate parameters, topographic parameters, and environmental parameters to generate a standardized soil dataset includes:

[0066] S201, perform data cleaning on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to obtain multi-source clean parameters of the soil;

[0067] S202, perform adaptive uncertainty quantification analysis on the multi-source cleaning parameters to obtain the uncertainty quantification value of the soil;

[0068] S203, perform weighted standardization on the uncertainty quantification value and the multi-source cleanliness parameters to obtain the standardized parameter set of the soil;

[0069] S204, Combine the standardized parameter set to construct the standardized dataset of the soil.

[0070] It should be noted that data cleaning is based on pre-acquired soil, climate, topography, and environmental parameters. It involves missing value imputation, outlier detection, and format unification to eliminate data noise and generate clean parameters from multiple sources. Specifically, missing value imputation uses linear interpolation to fill in null values, outlier detection uses box plots to identify and remove data points exceeding 1.5 times the interquartile range, and format unification converts parameters from different sources into floating-point format and stores them in an in-memory array to ensure the integrity and consistency of the input data. This provides a clean foundation for subsequent uncertainty quantification analysis and prevents the propagation of errors.

[0071] It should be noted that the multi-source clean parameters include soil moisture and organic carbon content extracted from soil parameters, temperature and precipitation rate extracted from climate parameters, elevation and slope extracted from topographic parameters, and vegetation index and cultivation intensity extracted from environmental parameters. These parameters are cleaned and stored in time series form as direct inputs for adaptive uncertainty quantification analysis, comprehensively reflecting the potential impact of soil, climate, topography and environmental factors on carbon sinks, and providing multi-dimensional data support for subsequent calculations.

[0072] It should be noted that adaptive uncertainty quantification analysis is based on time series data of multi-source cleaning parameters. It calculates parameter variance by dynamically adjusting the sliding window size to quantify measurement uncertainty. Its core lies in the adaptive change of window size according to the parameter change rate. When the change rate is large, a small window is used to capture local fluctuations, and when the change rate is small, a large window is used to smooth noise. This is used to accurately assess the reliability of each parameter, identify data volatility, provide weighting basis for weighted standardization processing, and improve the robustness of carbon sink measurement.

[0073] It should be noted that the uncertainty quantification value is the output of adaptive uncertainty quantification analysis, representing the variance of each parameter within the sliding window. The unit is the square of the original unit of the parameter. For example, the variance of soil moisture is expressed as percentage square. The role of the uncertainty quantification value is to directly reflect the dispersion of parameter measurement. The larger the value, the higher the uncertainty. It is then used as the basis for weighting the standardization process to adjust the contribution of the parameter in the dataset.

[0074] It should be noted that the weighted standardization process is based on the uncertainty quantification value and multi-source cleaning parameters. Weights are calculated and scaled to generate a standardized parameter set, as follows: For each parameter sequence, firstly, the overall mean and standard deviation of the parameter sequence are calculated as a baseline scaling reference. Then, weights are calculated based on the uncertainty quantification value, where the weight is 1 divided by 1 plus the uncertainty quantification value. A larger uncertainty quantification value results in a smaller weight, indicating a reduction in the contribution of highly uncertain parameters. Finally, for each time point parameter value, the overall mean is subtracted, and then the result is divided by the product of the overall standard deviation and the weight to obtain the standardized value. This standardized value is used to dynamically adjust parameter scaling through uncertainty weights, reducing the impact of highly uncertain parameters in the dataset, enhancing data consistency and model input stability, thereby ensuring the accuracy and robustness of subsequent carbon sink calculations.

[0075] It should be noted that the standardized parameter set is the output of the weighted standardization process, containing the standardized values ​​of all parameters, such as standardized soil moisture and standardized temperature. It is stored in the form of a structured array to eliminate differences in parameter dimensions, unify data distribution, and serve as a basic component for generating the standardized dataset, making it easy for subsequent modules such as biomass-carbon sink dynamic constraint calculation to directly call upon it.

[0076] It should be noted that the combined standardized parameter set is based on all standardized parameters. It involves splicing and indexing operations, organizing the standardized values ​​of soil, climate, topography and environmental parameters into a multidimensional array according to time series and spatial identifiers, and adding metadata descriptors to identify the parameter sources. This is used to integrate multi-source parameters into a unified data structure, which facilitates storage, retrieval and transmission, provides standardized input for the entire carbon sink measurement process and improves data processing efficiency.

[0077] It should be noted that the standardized dataset is the output of the combined operation, containing standardized soil parameters, climate parameters, topographic parameters, and environmental parameters, as well as associated timestamps and spatial indices. It serves as the benchmark data source for all subsequent carbon sink measurement steps, ensuring that the measurement process is based on a consistent and reliable data foundation, thereby improving the repeatability and accuracy of the overall method.

[0078] In this embodiment of the invention, the adaptive uncertainty quantification analysis of the multi-source cleanliness parameters to obtain the uncertainty quantification value of the soil includes:

[0079] S301, Analyze the parameter sequence change rate of the multi-source cleaning parameters to obtain the soil change rate sequence;

[0080] S302, Adaptively adjust the window size of the rate of change sequence to obtain an adaptive window size sequence;

[0081] S303, based on the adaptive window size sequence, calculate the in-window variance of the multi-source cleaning parameters to obtain the uncertainty quantification value of the soil.

[0082] It should be noted that the parameter sequence change rate of the multi-source cleaning parameters is analyzed based on the time series data of the multi-source cleaning parameters. Point-by-point differencing and normalization operations are performed to calculate the relative change of the parameter value at each time point relative to the previous time point. Specifically, for the parameter sequence, starting from the second time point, the absolute difference between the parameter value at the current time point and the parameter value at the previous time point is calculated, and this difference is divided by the parameter value at the previous time point to obtain the relative change rate. For the first time point, the change rate is initialized to zero to ensure the integrity of the sequence. The absolute value operation ensures that the change rate is non-negative, and the normalization operation converts the change into a relative proportion by dividing by the parameter value at the previous time point, avoiding the influence of different parameter units.

[0083] Furthermore, the analytical parameter sequence change rate is essentially a high-speed subtraction and division operation. It utilizes pipeline technology to process continuous data points and uses anti-overflow circuitry to prevent division by zero errors, thereby ensuring the accurate generation and computational efficiency of the change rate sequence. The analytical parameter sequence change rate is used to quantify the dynamic changes of parameters over time, identify parameter volatility, and provide input for subsequent adaptive window size adjustment, thereby optimizing uncertainty quantification analysis.

[0084] It should be noted that the rate of change sequence contains the parameter change rate value at each time point. For example, for multi-source cleaning parameters such as soil moisture, temperature, and elevation, the rate of change sequence contains the relative change amplitude value from time point 2 to T, as well as the initial value of 0 at time point 1, stored in the form of a floating-point number sequence. Each value represents the relative change ratio of the parameter at adjacent time points, reflecting the degree of fluctuation of the parameter sequence. The larger the value, the more drastic the parameter change. It is used to dynamically adjust the window size to ensure that a smaller window is used to capture local uncertainties when the parameter changes greatly, and a larger window is used to smooth noise when the change is small, providing key input for adaptive uncertainty quantification analysis.

[0085] It should be noted that the adaptive adjustment of the window size of the rate of change sequence is based on the adaptive adjustment of the window size of the rate of change sequence to generate an adaptive window size sequence, as follows: For the rate of change value at each time point, the window reference parameter is calculated and divided by the maximum value between the rate of change value and a very small positive number, and the result is rounded to obtain the adaptive window size at that time point. The window reference parameter is used to control the reference range of the window size, the very small positive number prevents division by zero error, and the rounding operation ensures that the window size is an integer.

[0086] Furthermore, adaptive window size adjustment is essentially dynamic window adjustment, where the window size is inversely proportional to the rate of change. When the parameter change rate is large, a smaller window is used to capture local uncertainties; when the parameter change rate is small, a larger window is used to smooth noise. Simultaneously, a circular buffer manages window data to ensure real-time processing efficiency and optimizes the window range for variance calculation. This allows uncertainty quantization to respond to rapid changes while smoothing steady-state noise, thereby improving quantization accuracy and providing a reliable foundation for subsequent weighted standardization processing.

[0087] It should be noted that the adaptive window size sequence contains the adaptive window size value for each time point. For example, for each multi-source cleaning parameter, each element in the sequence is an integer representing the sliding window length used for variance calculation at that time point. It is stored in the form of an integer sequence, and the range of values ​​is constrained by preset parameters and rate of change.

[0088] It should be noted that the in-window variance of the multi-source cleaning parameters is calculated based on the adaptive window size sequence and the multi-source cleaning parameter sequence, using a sliding window to calculate the mean and variance. The mathematical expression is as follows:

[0089] ;

[0090] In the formula, It is the first The mean of the parameters within a time window. Represents a point in time Adaptive window size It is the first The variance of parameters within a time window at each time point. These are multi-source cleaning parameter values. It is a summation index. As a quantification of uncertainty.

[0091] Furthermore, the first The mean of the parameters within the window at each time point represents the trend at the center of the window. The variance of parameters within a time window measures the dispersion of the parameters within the window. The multi-source cleaning parameter value is each data point in the parameter sequence. The summation index is used for the summation operation in the variance calculation within the window and indicates the position of the data point within the sliding window.

[0092] It should be noted that the uncertainty quantification value is used to directly reflect the dispersion and reliability of parameter measurement. The larger the value, the higher the uncertainty. It is then used as the basis for weighting the standardization process to reduce the impact of high uncertainty parameters in the standardized dataset, thereby improving the overall accuracy and data consistency of carbon sink measurement.

[0093] S2, Based on the standardized dataset, the biomass-carbon sequestration dynamic constraint of the soil is calculated using the biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data.

[0094] In this embodiment of the invention, based on the standardized dataset, the biomass-carbon sequestration dynamic constraint of the soil is calculated using a biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data, including:

[0095] S401, Based on the vegetation index parameters, soil nutrient parameters and climate moisture parameters in the standardized dataset, construct a biomass-related parameter set;

[0096] S402, Estimate the initial biomass of the biomass-related parameter set to obtain a biomass sequence;

[0097] S403, perform carbon sequestration on the biomass sequence to obtain forest and grassland carbon sequestration;

[0098] S404, based on the biomass-carbon sink dynamic constraint model, the constraint adjustment of the forest and grassland carbon sink amount is carried out to generate forest and grassland carbon sink data.

[0099] It should be noted that the biomass-related parameter set is constructed based on the vegetation index parameters, soil nutrient parameters, and climate moisture parameters in the standardized dataset. Through data retrieval and combination operations performed by the processor, the vegetation index parameters, soil nutrient parameters, and climate moisture parameters are extracted from the standardized dataset and integrated into a structured parameter set. Each parameter is stored in the form of a standardized value. The vegetation index parameter ranges from -1 to 1, the soil nutrient parameter is in milligrams per kilogram, and the climate moisture parameter is in percentage or millimeters. This set provides multi-dimensional input parameters for the initial biomass estimation, comprehensively reflecting the impact of vegetation growth, soil fertility, and climate conditions on biomass, and ensuring the comprehensiveness and accuracy of subsequent biomass measurements.

[0100] It should be noted that the initial biomass estimation is based on a set of biomass-related parameters and is calculated using a multiple linear regression model. The mathematical expression is as follows:

[0101] ;

[0102] In the formula, It is the first Initial biomass at each time point It is a vegetation index parameter. These are soil nutrient parameters. It is a climate moisture parameter; , , It is the regression coefficient. It is the intercept term.

[0103] Furthermore, the first The initial biomass unit at each time point is kilograms per hectare, representing the dry matter weight of the vegetation. The regression coefficients are obtained through training on historical datasets and represent the contribution weights of each parameter to biomass. The intercept term adjusts the model bias. The initial biomass estimate is used to quantify vegetation biomass and generate a biomass sequence, providing basic data for carbon sink conversion.

[0104] It should be noted that the biomass sequence contains the initial biomass value at each time point. For example, for forest and grassland vegetation, each element in the sequence is a biomass value in kilograms per hectare, stored in the form of a floating-point sequence, representing the weight of vegetation dry matter that changes over time. This is used as a direct input for carbon sink conversion, recording the dynamic changes in biomass, which is convenient for subsequent carbon sink measurement and data analysis. At the same time, the time series format supports trend assessment and model validation.

[0105] It should be noted that carbon sink conversion is based on biomass sequences and is performed linearly using a carbon density factor, as shown in the following expression:

[0106] ;

[0107] In the formula, It is the first Forest and grassland carbon sequestration at a given time point It is a carbon conversion factor. It is the first Biomass values ​​at each time point.

[0108] Furthermore, the first The forest and grassland carbon sink at each time point is expressed in tons of carbon per hectare, representing carbon storage. The carbon conversion factor ranges from 0.4 to 0.5, representing the average proportion of carbon in biomass. It is determined using a table lookup method based on vegetation type and regional characteristics. Biomass values ​​at each time point, carbon sink conversion is used to convert biomass into carbon sink, quantify carbon storage, and provide input for dynamic constraint adjustments.

[0109] It should be noted that forestry and grassland carbon sequestration is the output of carbon sequestration conversion, representing the amount of carbon stored at each point in time, in tons of carbon per hectare. It is stored in the form of a floating-point sequence, such as a sequence, and is used as the baseline input for the biomass-carbon sequestration dynamic constraint model. It reflects the unadjusted carbon sequestration dynamics and is used for subsequent constraint adjustments to generate more stable forestry and grassland carbon sequestration data.

[0110] It should be noted that the biomass-carbon sink dynamic constraint model is based on the forest and grassland carbon sink sequence and is calculated using a nonlinear model to generate forest and grassland carbon sink data. Specifically, for each time point, the forest and grassland carbon sink is multiplied by an exponential decay function, which is calculated based on the decay coefficient and time index to simulate the long-term natural decay of the carbon sink. At the same time, the rate of change of the carbon sink is calculated, that is, the difference between the carbon sink at the current time point and the carbon sink at the previous time point, and this rate of change is multiplied by an adjustment weight. Finally, the results of the above two parts are added together to obtain the constrained adjusted forest and grassland carbon sink data. Here, the decay coefficient represents the rate of natural decay of the carbon sink over time, and the adjustment weight controls the degree of influence of the rate of change on the final result, which is used to simulate the dynamic changes and stability of the carbon sink. The robustness of the carbon sink data is enhanced by adjusting the decay and rate of change, and errors caused by short-term fluctuations are prevented.

[0111] It should be noted that the constraint adjustment is based on scalar calculations of the output of the biomass-carbon sink dynamic constraint model. The forest and grassland carbon sink amount is adjusted to forest and grassland carbon sink data, and exponential decay and rate of change weighting are applied to generate a smoother and more stable carbon sink sequence. This is used to optimize the carbon sink data, so that it more accurately reflects the carbon storage dynamics of the forest and grassland ecosystem, and improves the reliability of the data and the accuracy of subsequent calculations.

[0112] It should be noted that the forestry and grassland carbon sink data includes the constrained carbon sink amount at each time point, in tonnes of carbon per hectare, stored in time series form. Each value in the series represents the carbon storage related to forestry and grassland vegetation. It combines biomass estimation and dynamic constraint effects and is used as one of the inputs for subsequent micro-topography-carbon sink dynamic correction calculation and carbon turnover-tillage interaction stability feedback calculation, providing the contribution of forestry and grassland carbon sink and being used to generate total carbon sink data.

[0113] S3. Based on the standardized dataset, the soil is subjected to micro-topography-carbon sequestration dynamic correction calculation through the micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data.

[0114] In this embodiment of the invention, the step of performing micro-topography-carbon sequestration dynamic correction calculations on the soil based on the standardized dataset using a micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data includes:

[0115] S501, Based on the micro-topographical parameters and climate impact parameters in the standardized dataset, construct the correction-related parameters;

[0116] S502, Based on the micro-topography-carbon sink dynamic correction model, the correction-related parameters are dynamically corrected by micro-topography to obtain the correction factor sequence;

[0117] S503, Based on the standardized dataset, the initial engineering carbon sink is obtained through the initial engineering carbon sink calculation model to obtain the engineering carbon sink of the soil;

[0118] S504, Based on the correction factor sequence, the engineering carbon sequestration amount is multiplied by the correction multiplier to perform dynamic correction processing on the engineering carbon sequestration amount, so as to generate engineering carbon sequestration data.

[0119] It should be noted that the construction of correction-related parameters is based on the micro-topographic parameters and climate impact parameters in the standardized dataset. Data retrieval and combination operations are performed to extract and integrate the micro-topographic parameters and climate impact parameters from the standardized dataset into a structured parameter set. Each parameter is stored in the form of a standardized value, such as slope in degrees, elevation in meters, and climate impact parameters in meters per second. This set provides dedicated input parameters for the micro-topographic-carbon sink dynamic correction model, comprehensively reflecting the potential impact of micro-topographic features and climate conditions on carbon sinks, and ensuring the accuracy and comprehensiveness of subsequent correction calculations.

[0120] It should be noted that the micro-topography-carbon sink dynamic correction model is based on the correction of relevant parameters and is calculated using a nonlinear regression algorithm to generate correction factors. Specifically, for each time point, the slope parameter is incremented by 1, the natural logarithm is taken, and then multiplied by the regression coefficient. This is then added to the elevation parameter multiplied by the regression coefficient, and finally to the climate impact parameter multiplied by the regression coefficient to obtain the correction factor value for that time point. Here, the slope parameter is in degrees, the elevation parameter is in meters, and the climate impact parameters, such as wind speed or precipitation intensity, are in meters per second. The regression coefficients are obtained through training on historical datasets and represent the contribution weight of each parameter to the correction factor. The natural logarithm function is used to smooth the nonlinear response of the slope impact, prevent numerical instability caused by extreme values, and quantify the dynamic impact of micro-topography and climate factors on carbon sinks. This generates a correction factor sequence, providing an adjustment basis for dynamic correction processing, thereby ensuring that engineering carbon sink data accurately reflects the micro-topography effect.

[0121] It should be noted that the correction factor sequence is the output of the micro-topography dynamic correction, containing the correction factor value at each time point. It is dimensionless and stored in the form of a floating-point number sequence. Each value in the sequence represents the degree of correction of carbon sink by micro-topography and climate. Positive values ​​indicate enhanced carbon sink and negative values ​​indicate weakened carbon sink. It is used as input for dynamic correction processing to adjust the engineering sink and quantify the changes in carbon storage under engineering measures.

[0122] It should be noted that the initial engineering carbon sequestration calculation model is based on the micro-topographic parameters and climate impact parameters in the standardized dataset. It uses a multiple linear regression model to calculate the initial engineering carbon sequestration, specifically as follows: The model uses the micro-topographic parameters and climate impact parameters in the standardized dataset as input variables, and calculates the initial engineering carbon sequestration at each time point through a linear combination. The model includes an intercept term and multiple regression coefficients. The intercept term is used to adjust the overall model bias, and the regression coefficients represent the contribution weights of the micro-topographic parameters and climate impact parameters to the initial engineering carbon sequestration. The model is trained using the least squares method. Based on historical engineering carbon sequestration data, these coefficients are optimized to ensure model fitting accuracy and generalization ability. Matrix operations and coefficient applications are performed to generate an initial engineering carbon sequestration sequence, which is stored in time series form in tons of carbon per hectare. This sequence serves as input for subsequent micro-topography-carbon sequestration dynamic correction processing, providing an independent benchmark carbon sequestration value for micro-topography-carbon sequestration dynamic correction calculation. By acquiring data independently, this method avoids data cross-contamination, provides a reliable foundation for the dynamic correction of subsequent correction factor sequences, improves the accuracy and robustness of engineering carbon sequestration data, and ultimately supports the accurate fusion of total carbon sequestration data and soil and water conservation decision-making.

[0123] It should be noted that the dynamic correction process is based on the correction factor sequence and the engineering carbon sink quantity, and performs a multiplication operation to generate engineering carbon sink data. Specifically, for each time point, the engineering carbon sink quantity is multiplied by the correction multiplier, where the correction multiplier is 1 plus the correction factor for that time point to obtain the engineering carbon sink data for that time point. Here, the engineering carbon sink data represents the carbon storage under micro-topographic engineering measures, with units of ton carbon per hectare. The engineering carbon sink quantity represents the uncorrected carbon storage, and the correction factor represents the correction effect of micro-topography on carbon sink, which is used to apply the micro-topographic correction effect to the engineering carbon sink quantity to generate engineering carbon sink data. This data truly reflects the dynamics of carbon storage under engineering measures and is achieved through high-speed calculation via an arithmetic logic unit, improving processing efficiency.

[0124] It should be noted that the engineering carbon sink data includes the dynamically corrected carbon sink amount at each time point, in tonnes of carbon per hectare, stored in time series form. Each value in the series represents the carbon storage under micro-topographic engineering measures. It combines forestry and grassland carbon sink data with micro-topographic correction effects and is used as one of the inputs for subsequent nonlinear co-causal inference to provide the contribution of engineering carbon sinks and to generate total carbon sink data, thereby enhancing the completeness and practicality of carbon sink measurement.

[0125] S4. Based on the standardized dataset, the carbon turnover-tillage interaction stability feedback of the soil is calculated using the carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data.

[0126] In this embodiment of the invention, the step of calculating the carbon turnover-tillage interaction stability feedback of the soil based on the standardized dataset using a carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data includes:

[0127] S601, extract carbon turnover parameters and tillage management parameters from the standardized dataset to obtain interactively related parameters;

[0128] S602, Based on the aforementioned interactive correlation parameters, analyze the impact of carbon turnover parameters and tillage management parameters on soil carbon sequestration stability to obtain a feedback factor sequence;

[0129] S603, Based on the standardized dataset, the initial tillage carbon sequestration is obtained through the initial tillage carbon sequestration calculation model to obtain the tillage carbon sequestration of the soil.

[0130] S604, Based on the aforementioned feedback factor sequence, the amount of cultivated carbon sink is adjusted through a stable feedback model of carbon turnover-cultivation interaction to generate cultivated carbon sink data.

[0131] It should be noted that the extraction of carbon turnover parameters and tillage management parameters is based on the standardized dataset. Data retrieval operations are performed to extract carbon turnover parameters and tillage management parameters from the soil parameter and environmental parameter subsets of the standardized dataset. The carbon turnover parameters are stored in the form of standardized values ​​with dimensionless units, and the tillage management parameters are stored in the form of standardized values ​​with dimensionless units. These parameters are used to provide dedicated input parameters for subsequent carbon turnover-tillage interaction stability feedback calculations, quantify the potential impact of carbon turnover processes and tillage activities on soil carbon sinks, and ensure the comprehensiveness and data foundation of the calculations.

[0132] It should be noted that the interaction-related parameters are a combination of extracted carbon turnover parameters and tillage management parameters, stored in the form of a structured parameter set. The sequence contains carbon turnover parameter values ​​and tillage management parameter values, which are used as inputs for the carbon turnover-tillage interaction stability feedback measurement. This comprehensively reflects the interaction effect between carbon turnover and tillage management, facilitating subsequent analysis of its impact on soil carbon sequestration stability and providing multi-dimensional data support for generating feedback factor sequences.

[0133] It should be noted that the analysis of the impact of carbon turnover parameters and tillage management parameters on soil carbon sink stability is based on interaction-related parameters. Calculations are performed using a dynamic feedback model embedded in the processor to generate a feedback factor sequence, as follows: For each time point, the carbon turnover rate parameter is first calculated by multiplying it by the natural logarithm function, then by the model coefficient. Simultaneously, the model coefficient is calculated by multiplying it by 1 and subtracting the exponential decay function, where the exponential decay function is calculated based on the model coefficient and the time decay parameter. Finally, the results of the above two parts are added together to obtain the feedback factor value for that time point. The natural logarithm function is used to smooth the nonlinear response of the tillage management parameter's influence, preventing numerical instability caused by extreme values. The exponential decay function is used to simulate the time-dependent nature of the carbon turnover-tillage interaction effect, i.e., the interaction effect gradually weakens over time, quantifying the dynamic impact of the carbon turnover and tillage management interaction effect on carbon sink stability, generating a feedback factor sequence, and providing a basis for stable feedback adjustments.

[0134] It should be noted that the feedback factor sequence is the output of the analysis of the impact of carbon turnover parameters and tillage management parameters on soil carbon sequestration stability. It contains the feedback factor values ​​at each time point, is dimensionless, and is stored in the form of a floating-point number sequence. Each value in the sequence represents the degree of adjustment of carbon turnover-tillage interaction on carbon sequestration stability. Positive values ​​indicate enhanced stability, while negative values ​​indicate weakened stability. It is used as input for stability feedback adjustment to adjust tillage carbon sequestration, quantify the carbon sequestration stability effect under tillage management, and ensure that tillage carbon sequestration data truly reflects the dynamics of the interaction.

[0135] It should be noted that the initial cultivated carbon sequestration calculation model is based on the carbon turnover parameters and cultivated management parameters in the standardized dataset. It uses a nonlinear regression model to calculate the initial cultivated carbon sequestration, specifically as follows: The model uses the carbon turnover parameters and cultivated management parameters in the standardized dataset as input variables. It calculates the initial cultivated carbon sequestration at each time point through nonlinear transformation and interaction terms. The model applies the natural logarithm function to process the carbon turnover parameters to smooth their nonlinear effects and adds a very small positive number to prevent taking the logarithm for zero or negative values. The model includes an intercept term, linear coefficients, and interaction term coefficients. The intercept term is used to adjust model bias, the linear coefficients represent the individual contribution weights of the carbon turnover parameters and cultivated management parameters, and the interaction term coefficients represent the synergistic effect of the two on the initial cultivated carbon sequestration. The training method employs a gradient descent algorithm to optimize these coefficients based on historical cultivated carbon sequestration data, ensuring the capture of dynamic relationships between parameters. Nonlinear calculations and weighted summation are performed to generate an initial cultivated carbon sequestration sequence, stored in time-series format in tons of carbon per hectare. This sequence serves as input for subsequent carbon turnover-cultivation interaction stabilization feedback adjustments, providing an independent benchmark carbon sequestration value for carbon turnover-cultivation interaction stabilization feedback calculations. The initial cultivated carbon sequestration quantifies the cultivated carbon sequestration state before the application of carbon turnover and cultivation interaction effects. This method, obtained independently, provides a foundation for the stabilization feedback adjustment of subsequent feedback factor sequences, allowing for individual optimization of cultivated carbon sequestration dynamics, improving the accuracy of carbon sequestration calculations, and supporting the scientific fusion of total carbon sequestration data. Ultimately, this enhances the overall effectiveness and decision support capabilities of soil and water conservation carbon sequestration assessment.

[0136] It should be noted that the stabilization feedback adjustment is based on the feedback factor sequence and the amount of cultivated carbon sink, and a multiplication operation is performed to generate cultivated carbon sink data, as follows: For each time point, the cultivated carbon sink is multiplied by the adjustment multiplier, where the adjustment multiplier is 1 plus the feedback factor for that time point to obtain the cultivated carbon sink data for that time point. The cultivated carbon sink data represents the carbon storage per ton of carbon per hectare under cultivated management, the cultivated carbon sink amount represents the unadjusted carbon storage, and the feedback factor represents the degree of influence of carbon turnover-cultivation interaction on carbon sink stability. The role of the stabilization feedback adjustment is to apply the stabilization effect of carbon turnover-cultivation interaction to the cultivated carbon sink amount to generate cultivated carbon sink data, which truly reflects the carbon storage dynamics of cultivated practices.

[0137] It should be noted that the tillage carbon sink data includes the carbon sink amount adjusted for stable feedback at each time point, in tonnes of carbon per hectare, stored in time series form. Each value in the series represents the carbon storage under tillage management. It combines the tillage carbon sink amount and the stable feedback effect of carbon turnover-tillage interaction, and is used as one of the inputs for subsequent nonlinear co-causal inference to provide the tillage carbon sink contribution and to generate total carbon sink data, thereby enhancing the completeness and practicality of carbon sink measurement.

[0138] S5, perform nonlinear collaborative causal reasoning on the forest and grassland carbon sink data, engineering carbon sink data and cultivated carbon sink data to generate the total carbon sink data of the soil.

[0139] In this embodiment of the invention, the step of performing nonlinear co-causal inference on the forestry and grassland carbon sequestration data, engineering carbon sequestration data, and cultivated carbon sequestration data to generate the total soil carbon sequestration data includes:

[0140] S701, perform time series alignment processing on the forest and grassland carbon sequestration data, engineering carbon sequestration data and cultivated carbon sequestration data to obtain the aligned carbon sequestration sequence of the soil.

[0141] S702, perform nonlinear collaborative causal reasoning collaborative fusion weight calculation on the aligned carbon sink sequence to obtain a weight sequence;

[0142] S703, perform carbon sink synergistic contribution fusion on the weighted sequence and the aligned carbon sink sequence to generate the total carbon sink data of the soil.

[0143] It should be noted that the time series alignment processing is based on the timestamps of the forestry and grassland carbon sequestration data, engineering carbon sequestration data, and cultivated carbon sequestration data. Interpolation and smoothing operations are performed to unify carbon sequestration data with different time resolutions to a common time point and generate an aligned carbon sequestration sequence. The interpolation operation uses linear interpolation to fill in missing time points, and the smoothing operation uses a moving average filter to reduce noise, ensuring that the data is consistent in the time dimension. This is used to eliminate errors caused by time mismatch, provide standardized input for subsequent nonlinear co-causal inference, and improve the accuracy and efficiency of data processing.

[0144] It should be noted that the aligned carbon sink sequence is the output of time series alignment processing. It contains the aligned values ​​of forestry and grassland carbon sink data, engineering carbon sink data, and cultivated carbon sink data at a common time point. It is stored in time series form, with each time point corresponding to the forestry and grassland carbon sink value, engineering carbon sink value, and cultivated carbon sink value. The unit is ton carbon per hectare. It is used as a direct input for nonlinear co-causal inference, providing a consistent time benchmark, facilitating weight calculation and fusion, and ensuring that subsequent calculations are based on synchronized data.

[0145] It should be noted that the nonlinear co-causal inference co-fusion weight calculation is based on aligned carbon sink sequences, dynamically quantifying the nonlinear co-causal relationship between forestry and grassland carbon sink data, engineering carbon sink data, and cultivated carbon sink data. By calculating the weight sequence, the contribution ratio of each carbon sink type in the total carbon sink is adjusted. Specifically, this calculation captures the dynamic dependence between carbon sink types through nonlinear transformation and causal interaction terms. The weight sequence is used to fairly allocate the contribution of each carbon sink type in the subsequent carbon sink co-contribution fusion, reducing the impact of highly volatile or low-reliability carbon sink data, improving the accuracy and robustness of the total carbon sink data, ensuring that carbon sink measurement can comprehensively reflect the dynamic interaction effect of multi-source carbon sinks, and ultimately improving the reliability and practicality of the overall method.

[0146] It should be noted that the weight sequence is the output of the nonlinear co-causal reasoning co-fusion weight calculation. It contains the weight value at each time point and is a three-dimensional vector sequence. The sequence represents the weights of forestry, grassland, engineering, and cultivated carbon sink data respectively, with values ​​ranging from 0 to 1 and summing to 1. It is used as the input for carbon sink synergistic contribution fusion to guide the weighted fusion process and make the total carbon sink data more accurately reflect the synergistic effect of each carbon sink source.

[0147] It should be noted that the carbon sink synergy contribution fusion is based on a weighted summation operation of the weighted sequence and the aligned carbon sink sequence to generate total carbon sink data, as follows: For each time point, the forestry and grassland carbon sink data, engineering carbon sink data and cultivated carbon sink data are multiplied by their corresponding weight values, and then these weight values ​​are added together to obtain the total carbon sink data for that time point. This data is used to fuse the various carbon sink data according to their weight ratios to generate comprehensive total carbon sink data, which truly reflects the overall carbon storage dynamics of the soil.

[0148] It should be noted that the total carbon sink data is the output of the synergistic contribution fusion of carbon sinks, containing the total carbon sink value at each time point, in tons of carbon per hectare, stored in time series form. Each value in the series represents the overall carbon storage of the soil, combining the synergistic contributions of forestry, grassland, engineering, and tillage carbon sinks. It is used as input for subsequent error correction prediction, providing the final carbon sink measurement results for decision support, trend analysis, and environmental assessment, enhancing the practicality and completeness of the method.

[0149] In this embodiment of the invention, the alignment of the carbon sink sequence undergoes nonlinear co-causal inference and co-fusion weight calculation to obtain a weight sequence, including:

[0150] S801, The initial weight values ​​are calculated using a nonlinear cocausal inference algorithm on the aligned carbon sink sequence to obtain an initial weight sequence. The mathematical expression of the nonlinear cocausal inference algorithm is as follows:

[0151] ;

[0152] In the formula: It is a type of carbon sink. At the point of time The initial weight values, It is an exponential function. It is a type of carbon sink. The weighting coefficients, It is the natural logarithm function. It is a type of carbon sink. At the point of time carbon sequestration value, It is a very small positive number. It is the synergistic effect coefficient. It is a type of carbon sink. For carbon sink types The causal influence coefficient, Indicates the index of the current carbon sink type. Index representing a point in time. Indexes representing other carbon sink types, It is a type of carbon sink. and At the point of time The product term;

[0153] S802, the initial weight sequence is normalized to obtain a weight sequence.

[0154] It should be noted that carbon sink types At the point of time The initial weights, dimensionless, represent the relative contribution of this carbon sink type before normalization. An exponential function is used to ensure the weights are positive, avoiding negative values ​​that could negatively impact subsequent normalization. (Carbon sink type) The weight coefficients are obtained through training on historical datasets, are dimensionless, and represent the linear influence of the carbon sink type itself on the weights. The natural logarithm function is used to smooth the carbon sink data. To mitigate fluctuations and prevent numerical instability caused by extreme values, carbon sink types... At the point of time The carbon sink value is measured in tons of carbon per hectare, derived from aligned carbon sink sequences. Minimal positive values ​​are used to prevent mathematical errors when the value is zero. The synergy coefficient is obtained through training on historical datasets, is dimensionless, and controls the overall impact of interaction effects between carbon sink types on the weights. (Carbon sink type) For carbon sink types The causal influence coefficients are obtained through training on historical datasets, are dimensionless, and represent... right Causal contribution level, carbon sink type and At the point of time The product term is used to quantify the nonlinear synergistic effect between the two.

[0155] It should be noted that the initial weight sequence is the output of the nonlinear cocausal inference algorithm that calculates the initial weight values. It contains the initial weight values ​​of all carbon sink types at all time points and is stored in the form of a structured array. Each element in the sequence represents the unnormalized weight of the carbon sink type at the time point, and the value range is positive real number. The initial weight sequence is used to directly reflect the original contribution of each carbon sink type to the total carbon sink, quantify the nonlinear cocausal relationship between carbon sink data, and serve as the input for normalization processing. By retaining the original weight information, it supports subsequent fair fusion, thereby improving the accuracy and robustness of the total carbon sink data.

[0156] It should be noted that the normalization process is based on the initial weight sequence and involves weight scaling to generate a new weight sequence. Specifically, for each time point and carbon sink type, the normalized weight value is equal to the initial weight value of that carbon sink type at that time point divided by the sum of the initial weight values ​​of all carbon sink types at that time point. This is used to convert the initial weight sequence into a probability distribution form, ensuring that the total weight at each time point is 1. This allows the weight sequence to fairly reflect the contribution ratio of each carbon sink type to the total carbon sink, avoiding weight bias from affecting the fusion result. At the same time, high-speed scalar calculation is achieved through division, improving processing efficiency and providing standardized input for the fusion of carbon sink collaborative contributions.

[0157] S6, perform error correction prediction on the total carbon sink data to generate predicted carbon sink data for the soil.

[0158] In this embodiment of the invention, the step of performing error correction prediction on the total carbon sink data to generate predicted carbon sink data for the soil includes:

[0159] S901, Perform historical error analysis on the total carbon sink data to obtain the error sequence of the soil;

[0160] S902, The error sequence is corrected using an error correction algorithm to obtain the corrected value sequence of the soil.

[0161] S903, Based on the correction value sequence, the total carbon sink data is corrected to generate the predicted carbon sink data for the soil.

[0162] It should be noted that historical error analysis is based on the time series of total carbon sink data. Differential and statistical calculations are performed to generate an error series, specifically as follows: First, the historical records of total carbon sink data are obtained, including the carbon sink value at each time point. This carbon sink value represents the total carbon sink data. Then, for each time point, the difference between the carbon sink value at that time point and the reference carbon sink value is calculated to obtain the error value. The reference carbon sink value is calculated using historical measured data or a benchmark model, and is expressed in tons of carbon per hectare. It serves as the benchmark for error calculation. The error value represents the deviation between the actual carbon sink value and the reference value, also expressed in tons of carbon per hectare. Its magnitude and direction reflect the accuracy and systematic error of the measurement results, quantifying the deviation pattern of the total carbon sink data, identifying error trends, and providing input for subsequent error correction, thereby improving the reliability and accuracy of carbon sink prediction.

[0163] It should be noted that the error series is the output of historical error analysis, containing the error value at each time point and stored in time series form. Each element in the series represents the measurement deviation of the total carbon sink data at that time point, in tons of carbon per hectare. The error series is used to directly reflect the error dynamics of historical measurements, recording the magnitude and direction of the deviation, and serving as the direct input to the error correction algorithm. By analyzing error patterns, it supports the calculation of correction values, thereby optimizing the subsequent prediction process.

[0164] It should be noted that the error correction algorithm is based on the error sequence and is calculated using a time series prediction model to generate a sequence of correction values. For each time point, the correction value consists of two parts: the first part is the current error value multiplied by the current error weight coefficient, and the second part is the weighted sum of historical error values. The historical error values ​​are multiplied by corresponding decay coefficients and weight coefficients. Both the current error weight coefficient and the historical error weight coefficient are trained using historical datasets and are used to control the influence of the current error and historical error on the correction value, respectively. The weighted sum of historical errors involves summing the errors from multiple past time points to capture the time dependence of the errors. The decay coefficient is calculated using an exponential decay function, representing the rate at which the influence of historical errors decays over time. The error correction algorithm is implemented through weighted summation and exponential operations, processing data from multiple time points and managing historical error data through a circular buffer to ensure real-time correction efficiency. This algorithm is used to predict future correction values ​​based on the error sequence, dynamically adjust total carbon sink data, reduce systematic biases, and thus improve the accuracy and robustness of carbon sink prediction data.

[0165] It should be noted that the correction value sequence is the output of the error correction algorithm, containing the correction value at each time point and stored in time series form. Each element in the sequence represents the adjustment amount to the total carbon sink data, in tons of carbon per hectare. The correction value sequence is used to quantify the magnitude and direction of error correction and serves as the direct input for the correction process. By applying the correction values ​​to optimize the carbon sink data, the predicted results are made closer to the true values, enhancing the practicality of the method.

[0166] It should be noted that the correction process is based on the correction value sequence and total carbon sink data. It generates carbon sink prediction data through an addition operation, as follows: For each time point, the total carbon sink data value is added to the correction value for that time point to obtain the carbon sink prediction data value. The correction process is essentially a high-speed addition operation, and the processing efficiency is improved through pipeline technology. It is used to apply the error correction effect to the total carbon sink data to generate optimized carbon sink prediction data, reduce the impact of historical bias, and thus improve the accuracy and reliability of the calculation results, providing a reliable basis for environmental decision-making.

[0167] It should be noted that the carbon sink prediction data is the output of the correction process, containing the predicted carbon sink value at each time point, in tons of carbon per hectare, stored in time series form. Each value in the series represents the corrected carbon storage of the soil, and combines timestamps and spatial index metadata. The carbon sink prediction data is used to provide the final optimized carbon sink calculation results, supporting soil and water conservation management and carbon sink trend analysis. By correcting errors, the credibility and practicality of the data are enhanced, ultimately improving the overall effectiveness of the AI-based soil and water conservation carbon sink calculation method.

[0168] like Figure 2 The diagram shown is a functional block diagram of an artificial intelligence-based soil and water conservation carbon sink calculation system provided in an embodiment of the present invention.

[0169] The artificial intelligence-based soil and water conservation carbon sequestration calculation system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the artificial intelligence-based soil and water conservation carbon sequestration calculation system 100 may include an adaptive uncertainty quantification module 101, a forestry and grassland carbon sequestration module 102, an engineering carbon sequestration module 103, a cultivated carbon sequestration module 104, a total carbon sequestration module 105, and an error correction and prediction module 106. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0170] In this embodiment, the functions of each module / unit are as follows:

[0171] The adaptive uncertainty quantization module performs adaptive uncertainty quantization processing on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to construct a standardized soil dataset.

[0172] The forest and grassland carbon sequestration module, based on the standardized dataset, performs biomass-carbon sequestration dynamic constraint calculation on the soil using a biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data.

[0173] The engineering carbon sequestration module, based on the standardized dataset, performs micro-topography-carbon sequestration dynamic correction calculations on the soil using a micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data.

[0174] The tillage carbon sink module, based on the standardized dataset, performs carbon turnover-tillage interaction stability feedback calculation on the soil through a carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data.

[0175] The total carbon sink module performs nonlinear collaborative causal reasoning on the forestry and grassland carbon sink data, engineering carbon sink data, and tillage carbon sink data to generate the total carbon sink data of the soil.

[0176] The error correction prediction module performs error correction prediction on the total carbon sink data to generate the predicted carbon sink data for the soil.

[0177] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0178] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0179] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0180] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0181] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0182] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for calculating carbon sequestration in soil and water conservation based on artificial intelligence, characterized in that, The method includes: S1, adaptive uncertainty quantification is performed on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to construct a standardized soil dataset; S2, Based on the standardized dataset, the biomass-carbon sequestration dynamic constraint of the soil is calculated using the biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data. S3. Based on the standardized dataset, the soil is subjected to micro-topography-carbon sequestration dynamic correction calculation through the micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data. S4. Based on the standardized dataset, the carbon turnover-tillage interaction stability feedback of the soil is calculated using the carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data. S5, perform nonlinear collaborative causal reasoning on the forest and grassland carbon sink data, engineering carbon sink data and cultivated carbon sink data to generate the total carbon sink data of the soil. S6, perform error correction prediction on the total carbon sink data to generate predicted carbon sink data for the soil; Based on the standardized dataset, the carbon turnover-tillage interaction stability feedback of the soil is calculated using a carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data, including: S601, extract carbon turnover parameters and tillage management parameters from the standardized dataset to obtain interactively related parameters; S602, Based on the aforementioned interactive correlation parameters, analyze the impact of carbon turnover parameters and tillage management parameters on soil carbon sequestration stability to obtain a feedback factor sequence; S603, Based on the aforementioned feedback factor sequence, the amount of cultivated carbon sink is adjusted through a stable feedback model of carbon turnover-cultivation interaction to generate cultivated carbon sink data. The step of performing nonlinear co-causal inference on the forestry and grassland carbon sequestration data, engineering carbon sequestration data, and cultivated carbon sequestration data to generate the total soil carbon sequestration data includes: S701, perform time series alignment processing on the forest and grassland carbon sequestration data, engineering carbon sequestration data and cultivated carbon sequestration data to obtain the aligned carbon sequestration sequence of the soil. S702, perform nonlinear collaborative causal reasoning collaborative fusion weight calculation on the aligned carbon sink sequence to obtain a weight sequence; S703, perform carbon sink synergistic contribution fusion on the weighted sequence and the aligned carbon sink sequence to generate the total carbon sink data of the soil; The aligned carbon sink sequence undergoes nonlinear co-causal inference and co-fusion weight calculation to obtain a weight sequence, including: S801, The initial weight values ​​are calculated using a nonlinear cocausal inference algorithm on the aligned carbon sink sequence to obtain an initial weight sequence. The mathematical expression of the nonlinear cocausal inference algorithm is as follows: ； In the formula: It is a type of carbon sink. At the point of time The initial weight values, It is an exponential function. It is a type of carbon sink. The weighting coefficients, It is the natural logarithm function. It is a type of carbon sink. At the point of time carbon sequestration value, It is a very small positive number. It is the synergistic effect coefficient. It is a type of carbon sink. For carbon sink types The causal influence coefficient, Indicates the index of the current carbon sink type. An index representing a point in time. Indexes representing other carbon sink types, It is a type of carbon sink. and At the point of time The product term; S802, the initial weight sequence is normalized to obtain a weight sequence.

2. The method for calculating soil and water conservation carbon sequestration based on artificial intelligence as described in claim 1, characterized in that, The adaptive uncertainty quantification process performed on the pre-acquired soil parameters, climate parameters, topographic parameters, and environmental parameters to generate a standardized soil dataset includes: S201, perform data cleaning on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to obtain multi-source clean parameters of the soil; S202, perform adaptive uncertainty quantification analysis on the multi-source cleaning parameters to obtain the uncertainty quantification value of the soil; S203, perform weighted standardization on the uncertainty quantification value and the multi-source cleanliness parameters to obtain the standardized parameter set of the soil; S204, Combine the standardized parameter set to construct the standardized dataset of the soil.

3. The method for calculating soil and water conservation carbon sequestration based on artificial intelligence as described in claim 2, characterized in that, The adaptive uncertainty quantification analysis of the multi-source cleanliness parameters to obtain the uncertainty quantification value of the soil includes: S301, Analyze the parameter sequence change rate of the multi-source cleaning parameters to obtain the soil change rate sequence; S302, Adaptively adjust the window size of the rate of change sequence to obtain an adaptive window size sequence; S303, based on the adaptive window size sequence, calculate the in-window variance of the multi-source cleaning parameters to obtain the uncertainty quantification value of the soil.

4. The method for calculating soil and water conservation carbon sequestration based on artificial intelligence as described in claim 1, characterized in that, Based on the standardized dataset, the biomass-carbon sequestration dynamic constraint of the soil is calculated using a biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data, including: S401, Based on the vegetation index parameters, soil nutrient parameters and climate moisture parameters in the standardized dataset, construct a biomass-related parameter set; S402, Estimate the initial biomass of the biomass-related parameter set to obtain a biomass sequence; S403, perform carbon sequestration on the biomass sequence to obtain forest and grassland carbon sequestration; S404, based on the biomass-carbon sink dynamic constraint model, the constraint adjustment of the forest and grassland carbon sink amount is carried out to generate forest and grassland carbon sink data.

5. The method for calculating soil and water conservation carbon sequestration based on artificial intelligence as described in claim 1, characterized in that, Based on the standardized dataset, the soil is dynamically corrected using a micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data, including: S501, Based on the micro-topographical parameters and climate impact parameters in the standardized dataset, construct the correction-related parameters; S502, Based on the micro-topography-carbon sink dynamic correction model, the correction-related parameters are dynamically corrected by micro-topography to obtain the correction factor sequence; S503, Based on the standardized dataset, the initial engineering carbon sink is obtained through the initial engineering carbon sink calculation model to obtain the engineering carbon sink of the soil; S504, Based on the correction factor sequence, the engineering carbon sequestration amount is multiplied by the correction multiplier to perform dynamic correction processing on the engineering carbon sequestration amount, so as to generate engineering carbon sequestration data.

6. The method for calculating soil and water conservation carbon sequestration based on artificial intelligence as described in claim 1, characterized in that, The step of performing error correction prediction on the total carbon sink data to generate predicted carbon sink data for the soil includes: S901, Perform historical error analysis on the total carbon sink data to obtain the error sequence of the soil; S902, The error sequence is corrected using an error correction algorithm to obtain the corrected value sequence of the soil. S903, Based on the correction value sequence, the total carbon sink data is corrected to generate the predicted carbon sink data for the soil.

7. A soil and water conservation carbon sequestration calculation system based on artificial intelligence, characterized in that, The system includes: The adaptive uncertainty quantification module performs adaptive uncertainty quantification on the pre-acquired soil parameters, climate parameters, topographic parameters and environmental parameters to construct a standardized soil dataset. The forest and grassland carbon sequestration module, based on the standardized dataset, calculates the dynamic constraints of biomass and carbon sequestration in the soil using a biomass-carbon sequestration dynamic constraint model to generate forest and grassland carbon sequestration data. The engineering carbon sequestration module, based on the standardized dataset, performs micro-topography-carbon sequestration dynamic correction calculations on the soil using a micro-topography-carbon sequestration dynamic correction model to generate engineering carbon sequestration data. The tillage carbon sequestration module, based on the standardized dataset, uses a carbon turnover-tillage interaction stability feedback model to calculate the soil carbon turnover-tillage interaction stability feedback in order to generate tillage carbon sequestration data. The total carbon sink module performs nonlinear co-causal inference on the forestry and grassland carbon sink data, engineering carbon sink data and cultivated carbon sink data to generate the total carbon sink data of the soil. The error correction prediction module performs error correction prediction on the total carbon sink data to generate the predicted carbon sink data for the soil. Based on the standardized dataset, the carbon turnover-tillage interaction stability feedback of the soil is calculated using a carbon turnover-tillage interaction stability feedback model to generate tillage carbon sink data, including: S601, extract carbon turnover parameters and tillage management parameters from the standardized dataset to obtain interactively related parameters; S602, Based on the aforementioned interactive correlation parameters, analyze the impact of carbon turnover parameters and tillage management parameters on soil carbon sequestration stability to obtain a feedback factor sequence; S603, Based on the aforementioned feedback factor sequence, the amount of cultivated carbon sink is adjusted through a stable feedback model of carbon turnover-cultivation interaction to generate cultivated carbon sink data. The step of performing nonlinear co-causal inference on the forestry and grassland carbon sequestration data, engineering carbon sequestration data, and cultivated carbon sequestration data to generate the total soil carbon sequestration data includes: S701, perform time series alignment processing on the forest and grassland carbon sequestration data, engineering carbon sequestration data and cultivated carbon sequestration data to obtain the aligned carbon sequestration sequence of the soil. S702, perform nonlinear collaborative causal reasoning collaborative fusion weight calculation on the aligned carbon sink sequence to obtain a weight sequence; S703, perform carbon sink synergistic contribution fusion on the weighted sequence and the aligned carbon sink sequence to generate the total carbon sink data of the soil; The aligned carbon sink sequence undergoes nonlinear co-causal inference and co-fusion weight calculation to obtain a weight sequence, including: S801, The initial weight values ​​are calculated using a nonlinear cocausal inference algorithm on the aligned carbon sink sequence to obtain an initial weight sequence. The mathematical expression of the nonlinear cocausal inference algorithm is as follows: ； In the formula: It is a type of carbon sink. At the point of time The initial weight values, It is an exponential function. It is a type of carbon sink. The weighting coefficients, It is the natural logarithm function. It is a type of carbon sink. At the point of time carbon sequestration value, It is a very small positive number. It is the synergistic effect coefficient. It is a type of carbon sink. For carbon sink types The causal influence coefficient, Indicates the index of the current carbon sink type. An index representing a point in time. Indexes representing other carbon sink types, It is a type of carbon sink. and At the point of time The product term; S802, the initial weight sequence is normalized to obtain a weight sequence.

Images