A soil resource sustainable utilization evaluation and prediction system

Abstract

The present invention provides a soil resource sustainable utilization assessment and prediction system, which relates to the technical field of soil resource sustainable utilization assessment and prediction. It aims to accurately capture the dynamic characteristics of the soil carbon cycle by real-time monitoring of soil respiration rate, carbon content and other environmental data, and combining with a multi-level carbon distribution-respiration rate mapping model. The system integrates MLP and LSTM technologies to dynamically update the multi-level carbon distribution-respiration rate mapping weights to reflect the non-linear mapping relationship between soil carbon storage and respiration rate. By constructing an optimized map of soil resources and a multi-factor comprehensive assessment index, the system can intuitively display the potential impacts of different soil management measures and provide scientific decision-making support for soil resource management.

Description

Technical Field

[0001] The present invention relates to the technical field of sustainable utilization assessment and prediction of soil resources, and particularly to a system for assessing and predicting the sustainable utilization of soil resources. Background Art

[0002] As one of the most important natural resources on the earth, soil resources play a crucial role in agricultural production, ecosystem balance, and global carbon cycle; it is not only the basis for plant growth but also directly affects food security and biodiversity; however, with the acceleration of industrialization and urbanization processes and unreasonable land use, the problem of soil degradation is becoming increasingly serious, posing a severe challenge to the sustainable utilization of soil resources.

[0003] There are significant deficiencies in the existing technologies for assessing and predicting the sustainable utilization of soil resources. Traditional assessment methods rely on periodic soil sample collection and analysis, which cannot monitor the dynamic changes of soil status in real time, resulting in the lack of timeliness of assessment results; existing assessment models often ignore the complex non-linear relationship between soil respiration rate and carbon distribution and fail to fully capture the dynamic characteristics of soil carbon cycle; in addition, existing prediction models usually fail to effectively integrate the impacts of multiple environmental factors and soil management measures, restricting the accuracy and practicality of prediction results. Summary of the Invention

[0004] (I) Technical Problems to be Solved

[0005] Aiming at the deficiencies of the existing technologies, the present invention provides a system for assessing and predicting the sustainable utilization of soil resources, which solves the problem of how to achieve the assessment of the sustainable utilization of soil resources through data processing.

[0006] (II) Technical Solutions

[0007] To achieve the above objectives, the present invention is realized through the following technical solutions: A system for assessing and predicting the sustainable utilization of soil resources, comprising:

[0008] A data collection and preprocessing module, which collects data on soil respiration rate, carbon distribution, and other environmental variables and preprocesses the collected data;

[0009] A multi-level carbon distribution-respiration rate mapping model, which constructs a non-linear mapping relationship between carbon content changes and soil respiration based on the preprocessed data and updates the carbon distribution in real time according to the changes in the respiration rate;

[0010] Soil Resource Optimization Atlas, based on the carbon distribution in the multi-level carbon distribution-respiration rate mapping model, generates the Soil Resource Optimization Atlas through a dynamic optimization algorithm, analyzes future soil changes under different soil management measures, and represents soil resource utilization parameters through nodes in the Soil Resource Optimization Atlas;

[0011] Multi-factor Comprehensive Evaluation Index, used to quantify the sustainability of soil resource utilization, integrates data on carbon distribution, respiration rate, and other environmental variables in the soil, and quantifies the Multi-factor Comprehensive Evaluation Index as a score based on the multi-level carbon distribution-respiration rate mapping model, providing a comprehensive sustainability assessment result.

[0012] The data acquisition and preprocessing module comprehensively monitors the soil through multiple data acquisition devices to obtain accurate data on soil respiration, carbon storage, and the environment; soil respiration rate sensors are installed in soil layers at different depths to capture the carbon dioxide released during soil microbial activities; infrared sensing is used to detect and record the concentration changes of carbon dioxide in real time to reflect the soil respiration rate; carbon content sensors monitor the carbon storage in each layer of the soil, such as the accumulation and distribution of organic carbon; the change in carbon content is calculated through the analysis of soil samples; the temperature, humidity, and moisture content of the soil are monitored in real time to provide data for soil respiration and carbon storage; after the data acquisition is completed, data preprocessing is carried out. Since soil data is easily disturbed by external environmental changes, the data collected by sensors contains noise; filtering algorithms are used to remove irrelevant high-frequency interference signals, such as electrical noise, wind speed changes, etc.; outliers and errors in the data are reduced through outlier detection, removing mutation points and unreasonable extreme values; missing values are filled through interpolation methods, including linear interpolation or Lagrange interpolation, to keep the data continuous and complete on the time axis; the preprocessed data after standardization will be input into the multi-level carbon distribution-respiration rate mapping model.

[0013] The multi-level carbon distribution-respiration rate mapping model constructs a dynamically updated dynamic model by deeply analyzing the non-linear relationship between soil carbon distribution and respiration rate; a non-linear relationship is established through the carbon distribution data at different depths in the soil and the real-time measured soil respiration rate; the non-linear relationship is expressed as , where R(t) represents the soil respiration rate, which is affected by the soil carbon distribution C(z) at different depths, β is the weight parameter of the model, T and W are environmental variables, which are soil temperature and moisture respectively, is the noise term; where , where sin(C(z)) represents that the carbon changes fluctuatingly with different depths of the soil, It indicates that the metabolic rate of microorganisms has an exponential growth response to temperature. The parameter γ regulates the sensitivity of temperature to the respiration rate and is set between 0.01 - 0.1. ln(W + 1) indicates that the influence of moisture on soil respiration shows a decreasing trend. is the noise term, representing the random error that the model cannot capture. Through the model, the non - linear mapping relationship between soil carbon storage, respiration rate, and the environment is revealed. To establish the non - linear mapping relationship, the model first standardizes the data measured each time. With the processed data of C(z), R(t), T, and W, the model is trained through a neural network to capture the non - linear mapping between these variables. The soil respiration rate R(t) is not only affected by carbon distribution but also jointly influenced by environmental factors such as temperature and moisture. This relationship is comprehensively considered in the model through the weight parameter β. After measuring the soil respiration rate each time, the carbon distribution C(z) is adjusted to update the model in real - time. Using C(z) new =C(z) old +ΔC(z,R(t)) to update the carbon distribution, where ΔC(z, R(t)) represents the carbon change caused by the respiration rate R(t); ΔC(z, R(t)) is obtained through where R avg is the historical average respiration rate, Δt is the time interval, and α is the response coefficient used to regulate the influence of the change in respiration rate on carbon distribution. Through the historical data of soil respiration and carbon dynamics, regression analysis is carried out to determine that the value of the response coefficient is between 0.01 - 0.1. The model dynamically adjusts the carbon distribution according to each new measurement of the respiration rate to reflect the changes in the soil in real - time.

[0014] The model combines MLP and LSTM for the dynamic update of the weight parameter β. MLP is used to capture the basic non - linear relationship between carbon distribution and respiration rate, while LSTM is used to handle the long - term dependencies in time - series data. After each model prediction, the weight parameter β is adjusted according to the error feedback and optimized through where η is the learning rate and L LSTM is the loss function of LSTM, representing the error between the model prediction and the actual observation. By continuously optimizing these weights, the model can gradually capture the influence of the change in soil respiration rate on carbon distribution more accurately, making it more accurate for soil assessment and management.

[0015] The update of the weight parameter β combines MLP and LSTM. Starting from the input layer, MLP receives static feature data, including soil carbon distribution, temperature, and moisture. After undergoing non-linear transformations through multiple hidden layers, the static feature data generates a preliminary predicted value of the respiration rate. LSTM receives the predicted value of the respiration rate from MLP and historical respiration rate data to form an input sequence. Since LSTM can capture the time dependence of the input sequence and process dynamic features, at each time step, LSTM manages data through a gating mechanism to generate the predicted value of the respiration rate R at the current moment. LSTM , after each model run, LSTM calculates the loss function between the predicted value and the actual measured value R observed . .

[0016] The construction of the soil resource optimization atlas uses a multi-level carbon distribution-respiration rate mapping model to extract real-time carbon distribution data and respiration rate prediction values, providing support for different soil management practices. First, the system combines various soil management practices such as tillage depth, irrigation method, and fertilization strategy, and defines each soil management practice as a node, representing the potential impact of different soil management practices on soil resource utilization efficiency. During the atlas construction process, a multi-level carbon distribution-respiration rate mapping model is used to simulate the effects of different soil management practices under specific soil conditions. The model receives data on various soil management practices including tillage depth, irrigation method, and fertilization strategy as input parameters. For example, considering tillage depth, the model receives data on two different management practices: shallow tillage and deep tillage. These data reflect different degrees of soil disturbance, which in turn affect the carbon distribution and respiration rate of the soil. Based on these input data and combined with soil environmental factors such as temperature and humidity, the model calculates the carbon distribution under each management practice. For example, deep tillage increases soil disturbance, thus affecting the carbon distribution in the soil and the respiration rate of microorganisms, while shallow tillage has less impact on soil structure and carbon cycling. The results output by the model, such as changes in carbon storage and respiration rate, provide basic data for evaluating the effects of different management practices. A series of evaluation indicators are defined, which are based on the simulation results of the model to quantify the impact of each management practice on soil resource utilization efficiency. For example, a carbon storage efficiency evaluation indicator is defined, which measures the amount of carbon distributed in the soil per unit area under a management practice. Another evaluation indicator is the respiration rate change rate, which is used to measure the change in soil respiration rate and reflect the change in soil microbial activity. By quantifying these evaluation indicators, the resource utilization efficiency of each management practice is evaluated. For example, if the carbon storage efficiency is high under the deep tillage practice but the respiration rate change rate is also high, it means that deep tillage increases soil carbon storage but also leads to more carbon emissions. By comparing the evaluation indicators under different management practices, it is possible to identify which practices have a positive impact on carbon storage and respiration rate, and which practices need to be adjusted to reduce negative impacts. In the atlas, each soil management practice is represented as a node, and the connections between the nodes represent the interactions and influences between soil management practices.

[0017] During the atlas analysis process, the system uses a multi-objective optimization algorithm to analyze the effects of each management measure; by optimizing the combination of different management parameters, it quantifies the impacts of each soil management on soil carbon storage, respiration rate, and soil fertility in the short and long term, helping users understand the interactions and trade-offs between different measures. For example, tillage depth may affect soil carbon storage but may lead to an increase in respiration rate; while irrigation helps improve soil water retention but has a smaller impact on carbon storage. The system synthesizes these impacts through an optimization algorithm and visualizes them through the paths in the atlas, enabling users to intuitively understand the short-term and long-term benefits of various soil managements. The atlas simulates the management effects under different scenarios, predicts the future trends of soil resources for users, and provides support for agricultural decision-making.

[0018] The multi-factor comprehensive evaluation index obtains the latest data of the soil through real-time monitoring sensors, including carbon distribution, soil respiration rate, and environmental variables. To comprehensively evaluate the health and sustainability of the soil, the system constructs a multi-factor comprehensive evaluation index that integrates various soil data. Each data is weighted and calculated, and the determination of the weight depends not only on the analysis of historical data but also on the knowledge and experience of domain experts. For example, the balance weight of soil carbon input and output is set according to the dynamic characteristics of the carbon cycle, while the weights of soil respiration rate and carbon fixation ability vary according to their impacts on the long-term health of the soil. Each data is weighted and calculated, and based on the non-linear mapping relationship, the weighted data is quantified into a comprehensive sustainability score, which reflects the overall health status of the soil resources and their potential for sustainable utilization in the current and future. The calculation of the comprehensive sustainability score considers the current soil conditions and also simulates the long-term impacts of different management measures, such as tillage methods, irrigation strategies, and fertilization methods on the soil resources. By ranking the comprehensive sustainability scores obtained under different soil management measures, identifying the high-scoring measures through the ranking, and comparing the scores of different soil management measures, the system can identify which measures have high scores, and the soil management measures with high scores can enhance the sustainability of the soil resources. For example, for three soil management measures with comprehensive sustainability scores of 60, 85, and 95 respectively, through the comprehensive sustainability score, the system can identify the optimal one among these measures, thus providing a decision-making basis supported by data for users. Beneficial effects

[0019] The present invention provides a soil resource sustainable utilization evaluation and prediction system, which has the following beneficial effects:

[0020] By real-time monitoring the soil respiration rate and carbon content and combining with the multi-level carbon distribution-respiration rate mapping model, the present invention can more accurately capture the dynamic characteristics of the soil carbon cycle, thereby improving the accuracy of soil resource sustainable utilization evaluation and prediction.

[0021] Through the dynamic soil resource optimization atlas and multi-factor comprehensive evaluation indicators, the present invention can intuitively display the potential impacts of different soil management measures, provide strong decision-making support for soil resource management, and help formulate more effective soil protection and utilization measures.

[0022] By integrating multiple data such as soil carbon distribution, respiration rate, and environmental variables, the present invention provides comprehensive sustainability assessment results, providing a more scientific decision-making basis for soil resource management.

[0023] By accurately assessing soil carbon storage and respiration rate, the present invention helps to monitor and mitigate climate change, while promoting the function of soil as a carbon sink, providing a scientific basis and technical support for global environmental protection and climate change response. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0026] The data acquisition and preprocessing module comprehensively monitors the soil through multiple data acquisition devices to obtain accurate data on soil respiration, carbon storage, and the environment; soil respiration rate sensors are installed in soil layers at different depths to capture the carbon dioxide released during the activities of soil microorganisms; infrared sensing is used to detect and record the concentration changes of carbon dioxide in real time to reflect the soil respiration rate; carbon content sensors monitor the carbon storage in each layer of the soil, such as the accumulation and distribution of organic carbon; the change in its carbon content is calculated by analyzing soil samples; the temperature, humidity, and moisture content of the soil are monitored in real time to provide data for soil respiration and carbon storage.

[0027] After the data collection is completed, these data are preprocessed. Since soil data are easily disturbed by external environmental changes, the data collected by sensors often contain noise. A filtering algorithm is used to remove irrelevant high-frequency interference signals, such as electrical noise and wind speed changes, etc. The filtering algorithm makes the data more stable and reliable by smoothing the data, reducing high-frequency fluctuations and retaining low-frequency trends. Next, outlier detection is carried out to reduce outliers and errors in the data. For example, the standard deviation or the interquartile range IQR is used to identify those data points that significantly deviate from the normal range. Specifically, the mean and standard deviation of the data set are calculated, and then those data points that exceed the mean ± 3 times the standard deviation are regarded as outliers. Or the IQR method is used, that is, the first quartile Q1 and the third quartile Q3 of the data are calculated, and then IQR is defined as Q3 - Q1. Any data point below Q1 - 1.5IQR or above Q3 + 1.5IQR is regarded as an outlier and removed from the data set. After removing the outliers, missing values in the data may be encountered. These missing values are filled by interpolation methods, including linear interpolation and Lagrange interpolation. Linear interpolation estimates the intermediate value by connecting the straight line of two known data points and is applicable to the case where the data changes linearly. Lagrange interpolation is a polynomial interpolation method. It constructs a polynomial function to fit the known data points and uses this polynomial function to estimate the missing values. It is used for the case where the data changes more complexly and can provide more accurate interpolation results. By the above interpolation methods, the missing values in the data are filled to make the data continuous and complete on the time axis. The preprocessed data after standardization will be input into the multi-level carbon distribution-respiration rate mapping model. Standardization is to scale the data to a unified scale, for example, scale all data to between 0 and 1.

[0028] The multi-level carbon distribution-respiration rate mapping model constructs a dynamically updated model by deeply analyzing the non-linear relationship between soil carbon distribution and respiration rate. A non-linear relationship is established through the carbon distribution data C(z) at different depths in the soil and the real-time measured soil respiration rate R(t). The non-linear relationship is expressed as , where R(t) represents the soil respiration rate, which is affected by the soil carbon distribution C(z) at different depths, β is the weight parameter of the model, T and W are environmental variables, which are soil temperature and moisture respectively, is the noise term; where , where sin(C(z)) represents that the carbon changes fluctuatingly with different depths of the soil, represents that the metabolic rate of microorganisms has an exponential growth response to temperature, and the parameter γ regulates the sensitivity of temperature to the respiration rate, set between 0.01 - 0.1, ln(W + 1) represents that the influence of moisture on soil respiration shows a decreasing trend, is the noise term, representing the random error that the model cannot capture; through the model, the non-linear mapping relationship between soil carbon storage and respiration rate and the environment is revealed; to establish the non-linear mapping relationship, the model first standardizes the data measured each time, and through the C(z), R(t), T, and W data obtained after processing, the model is trained by a neural network to capture the non-linear mapping between these variables; for example, under drought conditions, even if the soil carbon content is high, the respiration rate may decrease due to insufficient moisture; the soil respiration rate R(t) is affected not only by the carbon distribution, but also by the combined action of environmental factors such as temperature and moisture, and this relationship is comprehensively considered in the model through the weight parameter β. When β is optimized, the multi-level carbon distribution-respiration rate mapping model not only improves the accuracy of respiration rate prediction, but also more accurately estimates the change of soil carbon distribution; for example, the change of respiration rate may cause the dynamic change of soil carbon content, and the optimized β makes the multi-level carbon distribution-respiration rate mapping model more sensitive to this change; after each new data input, it can dynamically adjust the value of β to accurately update the carbon distribution with the real-time change of soil respiration rate, so as to more truly reflect the carbon cycle process in the soil ecosystem.

[0029] After measuring the soil respiration rate each time, adjust the carbon distribution C(z) to update the model in real time; use C(z) new =C(z)old + ΔC(z,R(t)) to update the carbon distribution, where ΔC(z,R(t)) represents the carbon change caused by the respiration rate R(t); ΔC(z, R(t)) is obtained through obtained, where R avg is the historical average respiration rate, Δt is the time interval, and α is the response coefficient used to adjust the impact of the respiration rate change on the carbon distribution; for example, if the current respiration rate R(t) is significantly higher than the historical average R avg , then ΔC(z,R(t)) will be a relatively large positive value, indicating that the soil carbon content is decreasing due to the higher respiration rate; make the model dynamically adjust the carbon distribution according to each new respiration rate measurement to reflect the change of the soil in real time.

[0030] The multi-level carbon distribution-respiration rate mapping model combines MLP and LSTM for the dynamic update of the weight parameter β. MLP is used to capture the basic non-linear relationship between carbon distribution and respiration rate, while LSTM is used to process the long-term dependence relationship in time series data; after each prediction of the multi-level carbon distribution-respiration rate mapping model, the weight parameter β will be adjusted according to the error feedback and optimized through for optimization, where η is the learning rate, L LSTMis the loss function of LSTM, representing the error between the model prediction and the actual observation; by continuously optimizing these weights, the multi-level carbon distribution-respiration rate mapping model can gradually capture the impact of the change in soil respiration rate on carbon distribution more accurately, making it more accurate in soil assessment and management; for example, assume that within a certain period, due to climate change, the soil temperature and moisture conditions change, which may affect the soil respiration rate; LSTM detects this long-term trend and adjusts the weight parameters together with MLP to reflect the impact of these changes on carbon distribution; if the respiration rate predicted by the multi-level carbon distribution-respiration rate mapping model is lower than the actual observation value, the loss function L of LSTM LSTM will increase, and the gradient ∇L LSTM will point in the direction of reducing the error, thereby adjusting the weight parameters so that the multi-level carbon distribution-respiration rate mapping model can more accurately reflect this change in future predictions.

[0031] The update of the weight parameter β combines MLP and LSTM. Starting from the input layer, MLP receives static feature data, including soil carbon distribution, temperature, and moisture; the static feature data is imported through the input layer and processed by the non-linear activation functions of multiple hidden layers, enabling it to gradually extract the deep features related to the soil respiration rate layer by layer. The output of each layer gradually approaches the complex mapping relationship of the actual respiration rate through non-linear transformation, and finally generates a preliminary respiration rate prediction value. LSTM receives the prediction value from MLP and historical respiration rate data to form an input sequence; because LSTM can capture the time dependence of the input sequence and process dynamic features, at each time step, LSTM manages the data through the gating mechanism to generate the respiration rate prediction value R at the current moment LSTM , after each run of the multi-level carbon distribution-respiration rate mapping model, LSTM calculates the prediction value R LSTM and the actual measured value R observed between the loss function .

[0032] The soil resource optimization map extracts real-time carbon distribution data and respiration rate prediction values through the multi-level carbon distribution-respiration rate mapping model, and then provides decision-making support for different soil management measures; the soil resource optimization map shows the relationship between different soil management measures, reveals the effect differences of these measures under different soil conditions, and provides a scientific basis for future soil resource management.

[0033] Define nodes when constructing the soil resource optimization map. The nodes include soil management measures such as tillage depth, irrigation method, and fertilization strategy. Each node represents the potential impact of different soil management measures on the utilization efficiency of soil resources. For example, the node of tillage depth is related to the impact of soil turning on carbon storage and respiration rate, while the node of fertilization strategy is related to the changes in soil fertility and crop yield. During the construction of the soil resource optimization map, the system will simulate the effects of these soil management measures under specific soil conditions. For example, assume there is a piece of soil with low carbon content and slow respiration rate. The system will simulate the impact of increasing tillage depth, changing irrigation frequency, or applying different types of fertilizers to this soil. These measures optimize carbon storage, increase soil respiration rate, or improve soil health. The system considers the effects of each soil management measure in these aspects and compares them with historical data and real-time monitoring data to evaluate the resource utilization efficiency of each node. For example, the system finds that increasing tillage depth can increase soil carbon storage because it can mix more organic matter into the soil, but it may also lead to an increase in respiration rate because tillage will damage the soil structure and increase microbial activity. On the other hand, the system finds that increasing irrigation can improve soil water retention, which helps crop growth, but has less impact on carbon storage. To more accurately evaluate the impact of these soil management measures on soil resources, the system uses a multi-objective optimization algorithm to analyze the effects of each soil management measure, helping to quantify the impact of each soil management measure on soil carbon storage, respiration rate, and soil fertility in the short and long term. By optimizing the combination of different management parameters, the system helps users understand the interactions and trade-offs between different soil management measures. For example, an increase in tillage depth will be combined with an adjustment of the fertilization strategy to balance the relationship between carbon storage and respiration rate.

[0034] The system synthesizes these effects through a multi-objective optimization algorithm and visualizes them through the paths in the soil resource optimization map, enabling users to intuitively understand the long-term and short-term benefits of various soil management. For example, the soil resource optimization map will show that increasing tillage depth and fertilization in the short term can increase crop yield, but in the long term, an irrigation strategy may need to be combined to maintain the balance of soil health and carbon storage. By simulating the management effects under different conditions, the soil resource optimization map predicts the future change trends of soil resources for users and provides support for agricultural decision-making. For example, if the prediction shows that future climate change may lead to drought, the soil resource optimization map will recommend that users adopt water-saving irrigation measures and adjust tillage and fertilization strategies to maintain soil carbon storage and fertility.

[0035] The system obtains the latest data of the soil through real-time monitoring sensors, including carbon distribution, soil respiration rate, and environmental variables; these are jointly constituted into a multi-factor comprehensive evaluation index to comprehensively reflect the health status and sustainable utilization potential of soil resources; each data is quantified for its impact on soil health and sustainability through weighted calculation, and the determination of the weight depends on the analysis of historical data and the combination of the knowledge and experience of domain experts; for example, the balance weight of soil carbon input and output is set according to the dynamic characteristics of the carbon cycle. If historical data shows that the balance between, for example, the decomposition of plant residues and the release of carbon dioxide through soil respiration has a significant impact on soil carbon storage, then this weight is set high; similarly, the weights of soil respiration rate and carbon fixation ability vary according to their impact on the long-term health of the soil; if expert experience indicates that an increase in soil respiration rate leads to a decrease in carbon storage, then the weight of this data will be adjusted accordingly; by multiplying each data by the corresponding weight, and then based on the non-linear mapping relationship, the weighted data is quantified into a comprehensive sustainability score, which reflects the overall health status of the soil resources and also predicts its potential for current and future sustainable utilization; assume that the original data collected by the system at a certain moment is carbon distribution C(z)=0.8, temperature T = 20, and moisture W = 0.5; the weighted data is, weighted carbon distribution C(z) 加权 =0.8×0.4 = 0.32, weighted temperature T 加权 =20×0.4 = 8, weighted moisture W 加权 =0.5×0.4 = 0.2;

[0036] Calculate the non-linear respiration rate, substitute the weighted data into the non-linear mapping relationship, and assume β = 0.5 and γ = 0.1 to calculate Rweighted and then generate a score for the respiration rate.

[0037] The calculation of the comprehensive sustainability score takes into account the current soil conditions and also simulates the long-term impacts of different soil management practices on soil resources; for example, the system simulates changes in tillage methods, such as the transition from conventional tillage to conservation tillage, which may reduce soil erosion, increase soil organic matter content, and thus enhance the soil's carbon sequestration capacity; the adjustment of irrigation strategies, such as the transition from flood irrigation to drip irrigation, reduces water waste and improves water use efficiency, thereby affecting the soil's respiration rate and nutrient loss; the optimization of fertilization methods, such as precision fertilization, improves fertilizer use efficiency, reduces nutrient loss, and simultaneously increases crop yields; by comparing the comprehensive sustainability scores obtained under different soil management practices, the system identifies which practices can enhance the sustainability of soil resources; for example, on a farm, the manager compared three soil management practices: conventional tillage, organic tillage, and conservation tillage; through real-time monitoring and evaluation, the conservation tillage obtained the highest comprehensive sustainability score and was thus identified as the most effective soil management practice, which helps to enhance the sustainability of soil resources.

[0038] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A soil resource sustainable utilization assessment and prediction system, characterized in that: include: The data collection and preprocessing module collects data on soil respiration rate, carbon distribution and other environmental variables, and preprocesses the collected data; A multi-level carbon distribution-respiration rate mapping model constructs a nonlinear mapping relationship between carbon content changes and soil respiration based on preprocessed data, and updates carbon distribution in real time according to changes in respiration rate; Soil resource optimization map, based on the carbon distribution in the multi-level carbon distribution-respiration rate mapping model, generates a soil resource optimization map through a dynamic optimization algorithm, analyzes future soil changes under different soil management measures, and represents different soil management measures through nodes in the soil resource optimization map; A multi-factor comprehensive evaluation index is used to quantify the sustainability of soil resource utilization. It integrates data on carbon distribution, respiration rate and other environmental variables in the soil, quantifies the multi-factor comprehensive evaluation index into a score based on a multi-level carbon distribution-respiration rate mapping model, and provides a comprehensive sustainability evaluation result. The multi-level carbon distribution-respiration rate mapping model constructs a nonlinear mapping relationship through preprocessed data to establish a nonlinear mapping relationship , where R(t) represents soil respiration rate, which is affected by soil carbon distribution C(z) at different depths, β is the weight parameter of the model, T and W are environmental variables, namely soil temperature and moisture, and ϵ is the noise term; , where sin(C(z)) indicates that carbon distribution fluctuates due to the influence of different soil depths. It indicates that the metabolic rate of microorganisms responds exponentially to temperature. The parameter γ adjusts the sensitivity of temperature to respiration rate and is set at 0.01-0.

1. ln(W+1) represents the effect of water on soil respiration. This model captures the nonlinear mapping relationship between soil respiration rate and carbon distribution and environmental variables. After each respiration rate measurement, the model uses C(z) new =C(z) old +ΔC(z,R(t)) updates the carbon distribution, where C(z) old is the carbon distribution before the update, ΔC(z,R(t)) is calculated by ΔC(z,R(t))=α(R(t)−R avg )Δt calculation, R avg is the historical average respiration rate, Δt is the time interval, and α is the response coefficient; the update of the weight parameter β combines the multilayer perceptron MLP and LSTM to dynamically learn the nonlinear mapping relationship between soil respiration and carbon distribution, through β new =β old +η∇L LSTM Update the weights, where β old is the old weight parameter, η is the learning rate, L LSTM is the loss function, which represents the difference between the model prediction value and the actual value, ∇L LSTM is the gradient of the LSTM loss function.

2. A soil resource sustainable utilization assessment and prediction system according to claim 1, characterized in that: The data acquisition and preprocessing module performs real-time monitoring and collection through multiple data acquisition devices, including soil respiration rate sensors installed in the soil at different depths to capture carbon dioxide in the soil during respiration; carbon content sensors monitor soil carbon storage at different levels, and environmental parameter sensors collect temperature, humidity and soil moisture; the data acquisition and preprocessing module preprocesses the collected data, removes interference signals and reduces noise, and uses interpolation to fill in any missing values ​​to keep the data complete; The standardized preprocessed data will be input into the multi-level carbon distribution-respiration rate mapping model.

3. A soil resource sustainable utilization assessment and prediction system according to claim 1, characterized in that: The update of the weight parameter β combines MLP and LSTM, starting from the input layer, where MLP receives static feature data including soil carbon distribution, temperature, and moisture; After the static feature data is transformed nonlinearly through multiple hidden layers, a preliminary respiratory rate prediction value is generated. The LSTM receives the output respiratory rate prediction value from the MLP and the historical respiratory rate data to form an input sequence. Because LSTM can capture the time dependency of the input sequence and process dynamic features, at each time step, LSTM manages the data through a gating mechanism to generate a predicted value R of the respiratory rate at the current moment. LSTM After each model run, LSTM calculates the predicted value R LSTM Compared with the actual measured value R observed The loss function between .

4. The soil resource sustainable utilization assessment and prediction system according to claim 1, characterized in that: The soil resource optimization map extracts real-time carbon distribution data and respiratory rate prediction values ​​through a multi-level carbon distribution-respiration rate mapping model. The soil resource sustainable utilization assessment and prediction system combines a variety of soil management measures, including tillage depth, irrigation method and fertilization strategy, and defines soil management measures as nodes. In the process of map construction, the carbon distribution under different soil management measures is obtained by inputting data of different soil management measures into the multi-level carbon distribution-respiration rate mapping model, thereby simulating the effects of different soil management measures under a soil condition. Evaluation indicators are defined, and the evaluation indicators quantify the impact of each management measure on soil resource utilization efficiency based on the simulation results. By evaluating the resource utilization efficiency of each node and comparing the evaluation indicators under different management measures, the impact of different measures on carbon storage and respiratory rate can be identified. By combining previous soil data and real-time monitoring data, the system generates a dynamic soil resource optimization map to show the relationship and effect between various soil management measures. Through a multi-objective optimization algorithm, the potential effect of each management measure is quantitatively analyzed, and the analysis results are visualized through the path in the soil resource optimization map, so that users can intuitively understand the long-term and short-term benefits of various management measures.

5. The soil resource sustainable utilization assessment and prediction system according to claim 1, characterized in that: The multi-factor comprehensive evaluation index obtains the latest soil status data through real-time monitoring sensors, constructs a multi-factor comprehensive evaluation index, and integrates multiple soil data, including soil carbon distribution, soil respiration rate and environmental variables; each data is calculated in a weighted manner, and the weight is determined by historical data analysis and expert knowledge. The weighted data is quantified into a comprehensive sustainability score based on a nonlinear mapping relationship, which reflects the overall health status and sustainable utilization potential of soil resources; by comparing the comprehensive sustainability scores under different soil management measures, the system identifies which measures can improve the sustainability of soil resources; through the visualization of the evaluation results of the soil resource optimization map, the user can intuitively understand the impact of different soil management measures on the sustainability of soil resources.

Images