Construction method, prediction method and system of grassland carbon sink function prediction model

Abstract

The invention provides a construction method and system of a grassland carbon sink function prediction model, and aims to predict a grassland carbon sink function. The method comprises the following steps: firstly, acquiring environmental information of a research grassland region through a meteorological station and a remote sensing technology, and dividing the region into a plurality of sub-regions by adopting stratified sampling; then, according to evaluation requirements of different scales, a plurality of quadrats of different sizes are set in each sub-region, and environment information, human activity data and carbon sink function index data are collected; a carbon sink function prediction model is constructed by using the data and combining a random forest algorithm. According to the method, spatial variation, a scale effect, environmental factors and human activities are considered, and the prediction precision is improved. In addition, the invention also provides a grassland carbon sink function prediction method, through obtaining prediction data of future climate change and human activities, the model is used to predict a future carbon sink function change trend, and a scientific basis is provided for grassland management and ecological environment improvement.

Description

Technical Field

[0001] The present invention relates to the field of grassland ecosystem information technology, and in particular to a method for constructing a grassland carbon sequestration function prediction model, a prediction method and a system. Background Art

[0002] Ecological and environmental monitoring involves real-time or regular monitoring and assessment of natural ecosystems and their environmental parameters to understand environmental trends and potential ecological and environmental issues. Understanding ecological and environmental information enables timely and effective ecological and environmental protection. Grasslands, as important terrestrial ecosystems, serve as significant carbon sinks, effectively mitigating the increase in atmospheric CO2 concentrations, reducing the greenhouse effect, and improving the ecological environment.

[0003] As a key terrestrial ecosystem, grasslands, through their soils and vegetation, can store large amounts of carbon and, under certain conditions, absorb even more, thus playing a crucial role in the global carbon cycle. Assessing this carbon sink typically involves measuring both carbon sequestration and carbon emission. Sequestration primarily relies on vegetation photosynthesis, while emission is influenced by soil microbial activity and human disturbance. While ground-based observation stations have been deployed across grasslands, and remote sensing technology can be used to obtain large-scale information on vegetation cover, surface temperature, and photosynthetic efficiency, and to estimate carbon storage and sequestration using these data, monitoring grassland carbon sequestration has achieved some success. However, monitoring the carbon sink function of grassland ecosystems is a complex process that requires comprehensive consideration of multiple factors. First, due to the inherent complexity and variability of grassland ecosystems, the acquisition of various data is subject to numerous uncertainties. Basic conditions such as soil type, vegetation cover, and precipitation directly influence carbon sequestration. Human activities also influence grassland carbon sequestration. Data on these factors typically come from different monitoring sites and sensors. These heterogeneous data sources vary significantly in temporal and spatial scales, formats, and accuracy, making them difficult to directly integrate and apply. Secondly, within grassland ecosystems, disturbance levels are also a significant factor influencing carbon sequestration. Currently, there is a lack of effective methods to quantify the impact of disturbance on carbon sequestration. While disturbance levels can be assessed using data on environmental conditions and human activities, converting these data into quantitative indicators remains a challenge. Thirdly, monitoring grassland carbon sequestration faces challenges due to spatial heterogeneity and regional variations. Differences in climate, topography, soils, vegetation, and other factors across grassland ecosystems result in significant spatial variability in carbon sequestration. Traditional grassland carbon sequestration monitoring typically relies on field surveys and plot observations. While these methods can provide information on localized carbon sequestration, they fail to comprehensively reflect the carbon sequestration characteristics of the entire grassland ecosystem. Summary of the Invention

[0004] In order to solve the problems existing in the prior art, the present invention provides a method for constructing a grassland carbon sequestration function prediction model, comprising:

[0005] Using meteorological stations and remote sensing technology to obtain first environmental information of the area to be studied, and using a stratified sampling method to divide the study area into several sub-areas;

[0006] In response to the carbon sequestration function assessment needs at different spatial scales, a multi-scale nested sample layout scheme was adopted, with multiple sample plots of different sizes set up in each sub-area;

[0007] For each sample plot, the secondary environment information, human activity data, and carbon sequestration function indicator data are obtained; and the sub-region type, sample plot, secondary environment information, and human activity data are used as input data, and the carbon sequestration function indicator data is used as output data, combined with the random forest algorithm, to construct a grassland carbon sequestration function prediction model.

[0008] Preferably, the stratified sampling method is used to divide the study area into several sub-areas, including:

[0009] Performing standardization processing on each data of the first environment information;

[0010] Using a clustering algorithm to cluster the data of the first environmental information to obtain a number of sub-areas;

[0011] The first environmental information includes climate factor data, terrain factor data, soil factor data and vegetation feature data.

[0012] Preferably, in response to the carbon sequestration function assessment requirements at different spatial scales, a multi-scale nested sample layout scheme is adopted, and multiple sample plots of different sizes are set in each sub-area, including:

[0013] Nested quadrats of several scales are set in each sub-area, and the number of quadrats of each scale is initialized;

[0014] The coefficient of variation method was used to revise the number of samples at each scale.

[0015] Preferably, for each sample plot, obtaining the second environmental information, human activity data, and carbon sequestration function indicator data, and using the sub-region type, sample plot, second environmental information, and human activity data as input data, and the carbon sequestration function indicator data as output data, in combination with the random forest algorithm, to construct a grassland carbon sequestration function prediction model includes:

[0016] Based on each quadrat in each sub-region, second environmental information and carbon sequestration function index data are obtained, and an environmental factor evaluation index is constructed based on the second environmental information and carbon sequestration function index data to determine a number of lower-level environmental factors;

[0017] obtaining human activity data for each sub-region, and determining a quantitative relationship between human activity and carbon sink function based on the human activity data and carbon sink function indicator data;

[0018] Based on the quantitative relationship between sub-region types, sample plot scales in each sub-region, environmental factor evaluation indicators and human activities on carbon sequestration function, the hierarchical analysis method is used to determine the weights of various environmental factor evaluation indicators and human activities, and combined with the random forest algorithm, a grassland carbon sequestration function prediction model is constructed.

[0019] Preferably, the obtaining of second environmental information and carbon sequestration function index data based on each quadrat of each sub-region, and constructing environmental factor evaluation indicators based on the second environmental information and carbon sequestration function index data to determine several lower-level environmental factors include:

[0020] Obtain data on various environmental factors and carbon sequestration indicators in each sample plot in each sub-region;

[0021] Determine the regression coefficient corresponding to each environmental factor using a multiple linear regression model based on the environmental factor data and the carbon sink indicator data; extract a number of environmental factors from the environmental factors based on the regression coefficient and set weight coefficients;

[0022] Based on the extracted environmental factors and weights, an environmental factor evaluation index system is established.

[0023] Preferably, the acquiring of human activity data of each sub-region and determining the quantitative relationship between human activity and carbon sequestration function based on the human activity data and carbon sequestration function indicator data includes:

[0024] Obtain monitoring data on various indicators of human activities in each sub-region;

[0025] After annotating the monitoring data of various human activity indicators, variance analysis was used to determine the significance of carbon sink function differences under different human activity levels.

[0026] Based on the determined significance of carbon sink function differences, human activity indicators with significant differences were selected as input features, carbon sink function data were used as output variables, and a nonlinear prediction model of human activity and carbon sink function was constructed using the support vector machine algorithm.

[0027] And the carbon sink index value corresponding to the value range of each index value of human activity is determined according to the nonlinear prediction model.

[0028] Preferably, the grassland carbon sink function prediction model is constructed based on the quantitative relationship between the sub-region type, the scale of each sample plot in each sub-region, the environmental factor evaluation index and human activities on the carbon sink function, and the weight of each environmental factor evaluation index and human activities is determined by using the hierarchical analysis method, and combined with the random forest algorithm, including:

[0029] The judgment matrix was constructed using the analytic hierarchy process to calculate the weights of each indicator in the environmental factor evaluation index system and each human activity indicator in each sub-region type and at each sample scale, and then the carbon sink function evaluation index system was constructed.

[0030] Based on the sub-region types, the training and test sets were constructed using the scales of various plots, the values ​​and weights of each indicator in the carbon sequestration function evaluation index system as input data, and the carbon sequestration function indicator values ​​as output data. The random forest algorithm was then used to construct a carbon sequestration function prediction model.

[0031] Setting the number of decision trees and the feature sampling ratio, training the carbon sequestration function prediction model based on the training set, and optimizing the parameters in the carbon sequestration function prediction model;

[0032] The trained carbon sink function prediction model was subjected to k-fold cross validation based on the test set, and the final carbon sink function prediction model was obtained after passing the validation.

[0033] Based on the same inventive concept, the present application provides a system for constructing a grassland carbon sequestration function prediction model, comprising:

[0034] A sub-region division module is used to obtain first environmental information of the area to be studied by using a meteorological station and remote sensing technology, and to divide the study area into several sub-regions by using a stratified sampling method;

[0035] The sample setting module is used to assess the carbon sequestration function at different spatial scales. It adopts a multi-scale nested sample layout scheme and sets multiple sample plots of different sizes in each sub-area.

[0036] The model building module is used to obtain secondary environmental information, human activity data, and carbon sequestration function indicator data for each sample plot; and uses sub-region type, sample plot, environmental factor data, and human activity data as input data, and carbon sequestration function indicator data as output data, combined with the random forest algorithm, to construct a grassland carbon sequestration function prediction model.

[0037] Based on the same inventive concept, the present application provides a method for predicting grassland carbon sequestration function, comprising:

[0038] Obtain forecast data on climate change and human activities in the study area under multiple future scenarios;

[0039] Based on each set of prediction data, the grassland carbon sink function prediction model is used to predict the changing trend of carbon sink function under each set of scenarios in the future;

[0040] The grassland carbon sink function prediction model is obtained by a construction method of a grassland carbon sink function prediction model provided by the present invention.

[0041] Based on the same inventive concept, the present application provides a grassland carbon sink function prediction system, comprising:

[0042] The data acquisition module is used to obtain forecast data on climate change and human activities in the study area under multiple scenarios in the future;

[0043] The prediction module is used to predict the trend of carbon sequestration function changes under various scenarios in the future based on each set of prediction data and using the grassland carbon sequestration function prediction model;

[0044] The grassland carbon sink function prediction model is obtained according to a method for constructing a grassland carbon sink function prediction model provided by the present invention.

[0045] Based on the same inventive concept, the present application provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0046] The memory is used to store one or more programs;

[0047] When the one or more programs are executed by the at least one processor, a method for constructing a grassland carbon sink function prediction model and / or a method for predicting grassland carbon sink function provided by the present invention are implemented.

[0048] Based on the same inventive concept, the present application provides a readable storage medium having an execution program stored thereon. When the execution program is executed, it implements a method for constructing a grassland carbon sink function prediction model and / or a grassland carbon sink function prediction method provided by the present invention.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] The present application provides a method and system for constructing a grassland carbon sink function prediction model, including: using meteorological stations and remote sensing technology to obtain first environmental information of the area to be studied, and using a stratified sampling method to divide the study area into several sub-areas; according to the carbon sink function assessment needs of different spatial scales, a multi-scale nested sample layout scheme is adopted, and multiple sample plots of different sizes are set in each sub-area; for each sample plot, second environmental information, human activity data, and carbon sink function indicator data are obtained; and sub-area type, sample plot, environmental factor data, and human activity data are used as input data, and carbon sink function indicator data is used as output data, combined with a random forest algorithm to construct a grassland carbon sink function prediction model. The present invention divides the study area into several relatively uniform sub-areas, and sets sample plots in each sub-area, fully considering the characteristics of spatial variation; by estimating the carbon storage and sequestration of sample plots of different scales, the scale effect of the carbon sink function of the grassland ecosystem can be revealed; the present invention can improve the accuracy of the model in predicting the carbon sink function by comprehensively considering factors such as spatial heterogeneity, scale effect, environmental factors and human activities.

[0051] The present invention also provides a method for predicting the carbon sink function of grassland, including: obtaining prediction data on climate change and human activities in the study area under multiple sets of scenarios in the future; based on each set of prediction data, using a grassland carbon sink function prediction model to predict the changing trend of the carbon sink function under each set of scenarios in the future, which can comprehensively reflect the carbon sink function characteristics of the entire grassland ecosystem and provide a scientific basis for grassland management, reducing the greenhouse effect, and improving the ecological environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a flow chart of a method for constructing a grassland carbon sequestration function prediction model according to the present invention;

[0053] Figure 2 This is a schematic diagram of the system structure for constructing a grassland carbon sequestration function prediction model of the present invention;

[0054] Figure 3 This is a schematic flow chart of a method for predicting grassland carbon sequestration function according to the present invention;

[0055] Figure 4 This is a schematic structural diagram of a grassland carbon sequestration function prediction system according to the present invention;

[0056] Figure 5 This is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0057] In order to better understand the present invention, the present invention is further described below with reference to the accompanying drawings and examples.

[0058] Example 1

[0059] The present invention provides a method for constructing a grassland carbon sink function prediction model, such as Figure 1 Shown, including:

[0060] S1. using a meteorological station and remote sensing technology to obtain first environmental information of the area to be studied, and using a stratified sampling method to divide the study area into several sub-areas;

[0061] S2. To meet the carbon sequestration function assessment needs at different spatial scales, a multi-scale nested sample layout scheme is adopted, with multiple sample plots of different sizes set up in each sub-area;

[0062] S3. For each sample plot, obtain the secondary environmental information, human activity data, and carbon sequestration function indicator data; and use the sub-region type, sample plot, secondary environmental information, and human activity data as input data, and the carbon sequestration function indicator data as output data, combined with the random forest algorithm, to construct a grassland carbon sequestration function prediction model.

[0063] Specifically:

[0064] In step S1, because grassland ecosystems have strong spatial heterogeneity, their carbon sequestration function is affected by multiple factors such as climate, topography, soil, and vegetation. To improve the representativeness and accuracy of carbon sequestration assessment, this paper uses a stratified sampling method to divide the study area into several sub-regions with relatively uniform climate, topography, soil, and vegetation conditions. The specific division process is as follows:

[0065] Step S1-1: Obtain climate factors such as multi-year average temperature, precipitation, and wind speed through meteorological station data; extract terrain factors such as slope, aspect, and altitude through remote sensing images (Landsat8, Sentinel-2) and digital elevation models (DEM); use soil profile surveys and remote sensing inversion technology to obtain soil factors such as soil texture, organic carbon content, pH value, and other information; calculate vegetation characteristics such as vegetation cover, vegetation type, and biomass through vegetation surveys and remote sensing data.

[0066] Step S1-2: Standardize the above data to eliminate dimensional differences; use the K-means clustering algorithm or hierarchical clustering method to divide the study area into several sub-areas based on climate, topography, soil and vegetation factors; adjust the sub-area division based on the field survey results to ensure good uniformity within each sub-area.

[0067] Step S1-3: Select a typical area for on-site measurement verification to ensure the rationality of the sub-area division; adjust the sub-area boundaries based on the measured data and optimize the area division scheme.

[0068] Specifically, the Xilin Gol grassland covers an area of ​​about 5000km 2For example, the average annual temperature in the area is 1.5℃, the annual precipitation is about 350mm, and the average altitude of the area is 1200-1500m, the slope is <5°, the soil organic carbon content is 1.2%-2.5%, the pH is 6.5-7.5, the vegetation cover is 30%-60%, and the aboveground biomass is 120-250g / m 2 The study area was divided into four sub-areas using the K-means clustering algorithm: lowland grassland, typical grassland, desertified grassland and degraded grassland.

[0069] Step S2: To meet the needs of carbon sequestration function assessment at different spatial scales, a multi-scale nested sample plot layout scheme is adopted. Multiple sample plots of different sizes are set in each sub-area, such as 1m×1m, 10m×10m, and 100m×100m. By observing and estimating carbon storage and carbon flux of sample plots of different scales, the scale effect of the carbon sequestration function of grassland ecosystems is revealed and the appropriate assessment scale is determined. The specific implementation process is as follows:

[0070] Step S2-1: Initialize the number and scale of quadrats in each sub-area, and use the coefficient of variation (CV) method to revise the optimal number of quadrats at each scale. Specifically, arrange quadrats of different scales (such as 1m 2 、10m 2 , 100m 2 ) and quantity, and after the initial sample measurement, adjust the number of samples at each scale according to the coefficient of variation (CV); when the sample variance is too large, increase the number of samples to improve representativeness.

[0071] The coefficient of variation method was used to calculate the degree of variation of environmental variables (such as soil carbon storage and vegetation carbon flux) in each sub-region;

[0072] According to the calculation formula: Determine the optimal number of quadrats n.

[0073] Where: Z is the confidence level (usually 1.96, corresponding to 95% confidence), E is the allowable error range; CV is the coefficient of variation, which is the ratio of the standard deviation to the mean, and can more intuitively represent the relative degree of variation in the data: σ is the standard deviation, x i For environment variables is the mean of the environmental variable, and n' is the initial number of quadrats. When the CV is small (e.g., less than 10%), the number of quadrats can be reduced. When the CV is large (e.g., greater than 30%), the number of quadrats can be increased to improve statistical accuracy.

[0074] Step S2-2: Different sample layout methods can be used for different types of sub-areas, such as random layout method, systematic layout method or layered layout method. Each layout method has its own advantages. Of course, the same sample layout method can also be used for different sub-areas. Among them, the random layout method is suitable for sub-areas with high uniformity, and the samples are randomly distributed to ensure the representativeness of the data; the systematic layout method is suitable for areas with large terrain changes, and an equidistant layout method is adopted to ensure coverage of the entire sub-area; the layered layout method is suitable for areas with obvious gradient changes in terrain, soil or vegetation, and the samples are distributed according to the gradient, such as laying out samples at the top, middle and foot of the slope to ensure data integrity.

[0075] Step S2-3: Design multi-scale nested quadrats based on different carbon sink function assessment requirements. Small-scale quadrats (1m×1m) are used to measure soil respiration, microbial activity, and fine root biomass; medium-scale quadrats (10m×10m) are used to measure vegetation biomass and analyze soil carbon storage; and large-scale quadrats (100m×100m) are used to validate remote sensing data and estimate large-scale carbon sink function. By nesting quadrats of different scales within each subregion, carbon storage and flux at different scales can be observed, revealing the scale dependence of carbon sink function.

[0076] Step S3: Acquire the second environmental information, human activity data, and carbon sequestration function indicator data. A grassland carbon sequestration function prediction model is constructed using the subregion type, sample plot, second environmental information, and human activity data as input data, and the carbon sequestration function indicator data as output data, combined with a random forest algorithm. The second environmental information here includes climate, topography, soil, and vegetation structure. Specifically, the model includes:

[0077] Step S3-1: For each quadrat in the subregion, obtain data on environmental factors such as climate, topography, soil, and vegetation structure, as well as carbon sequestration function indicators such as carbon storage and carbon flux, construct a multivariate linear regression model, quantitatively analyze the impact of various environmental factors on carbon sequestration function, and use principal component analysis to determine key influencing factors and their weights;

[0078] Step 3-2: Obtain human activity data such as grazing intensity, fence density, and fire frequency. Use variance analysis to determine the significance of differences in carbon sequestration function under different human activity levels. Use a support vector machine algorithm to establish a nonlinear prediction model of human activity and carbon sequestration function. If the model accuracy meets a preset threshold, the quantitative impact of human activity on carbon sequestration function is determined.

[0079] Step 3-3: Taking into account spatial heterogeneity, scale effects, environmental factors, and human activities, the analytic hierarchy process (AHP) was used to determine the weights of each factor. Combined with the random forest algorithm, a grassland carbon sink function prediction model was constructed. The stability and reliability of the model were determined through cross-validation.

[0080] Step 3-1: For each plot in the subregion, obtain environmental factor data and carbon sequestration function indicators such as carbon storage and carbon flux, construct a multivariate linear regression model, quantitatively analyze the impact of various environmental factors on carbon sequestration function, and use principal component analysis to determine the key influencing factors and their weights; including:

[0081] Step (1) Obtain environmental factor data for each sample plot, including data such as climate, topography, soil, and vegetation values, and simultaneously determine the carbon storage and carbon flux index data corresponding to the sample plot; here, climate values ​​can be obtained through meteorological station monitoring data and remote sensing meteorological data (MODIS, ERA5), including environmental factors such as average temperature, precipitation, relative humidity, solar radiation, and wind speed. Topography values ​​can be obtained through digital elevation models (DEM, 30m resolution), including environmental factors such as slope, aspect, elevation, and topographic moisture index. Soil values ​​can be obtained through field sampling + laboratory analysis and the ISRIC soil database, including environmental factors such as soil organic carbon content, soil texture (ratio of clay, silt, and sand), soil moisture content, and soil pH. Vegetation values ​​can be obtained through remote sensing images (Sentinel-2, MODIS), field surveys, etc., including environmental factors such as vegetation cover, normalized difference vegetation index, enhanced vegetation index, and vegetation biomass. The determination of carbon storage and carbon flux indicator data includes soil carbon storage determination, vegetation carbon storage determination, soil respiration and carbon flux determination, etc.

[0082] Step (2), based on the acquired climate values, terrain values, soil values, vegetation values ​​and other data, combined with the carbon storage and carbon flux index data, using a multivariate linear regression model to determine the regression coefficient corresponding to each environmental factor;

[0083] Y=β0+β1X1+…β m X m +∈;

[0084] Where Y is carbon storage or carbon flux; β0 is the intercept; β1…β m is the regression coefficient; X1…X m are the values ​​of the respective variables obtained, such as annual average temperature, annual precipitation, relative humidity, altitude, slope, aspect, soil organic matter content, soil pH value, vegetation cover, forest distribution density, etc.; ∈ is the error term.

[0085] Step (3), perform principal component analysis on climate values, terrain values, soil values ​​and vegetation values, extract key influencing factors with greater influence, and calculate the weight value of each factor;

[0086] Step (4): Based on the key influencing factors and their weight values, an environmental factor evaluation index system is established to achieve a quantitative assessment of carbon storage and carbon flux.

[0087] Specifically, when constructing a multivariate linear regression model, it is first necessary to obtain environmental factor data and carbon sequestration function indicators for the sample plots. For example, 100 sample plots were selected. The climate data for each sample plot included annual mean temperature (ranging from 5°C to 25°C) and annual precipitation (ranging from 300 mm to 1500 mm). Topographic data included altitude (ranging from 0 m to 2000 m) and slope (ranging from 0° to 45°). Soil data included soil organic matter content (ranging from 1% to 10%) and soil pH (ranging from 5 to 5). Vegetation structure data included vegetation cover (ranging from 30% to 100%) and stand density (ranging from 100 plants / ha to 1000 plants / ha). Carbon sequestration function indicators included carbon storage (ranging from 10 tC / ha to 200 tC / ha) and carbon flux (ranging from 5 tC / ha / yr to 5 tC / ha / yr). After inputting these data into the model, a multivariate linear regression analysis was performed using the least squares method to obtain the regression coefficients for each environmental factor on carbon sequestration. For example, the regression analysis results showed that the regression coefficients for carbon storage were 35 for annual average temperature, 28 for annual precipitation, -15 for altitude, 12 for slope, 45 for soil organic matter content, 18 for soil pH, 38 for vegetation cover, and 25 for stand density. Next, principal component analysis was used to standardize each environmental factor, calculate its covariance matrix, and then identify the principal components through eigenvalue decomposition. For example, the first principal component explained 45% of the total variance, with weights of 35 and 30 for annual average temperature and annual precipitation, respectively. The second principal component explained 30% of the variance, with weights of 25 and 20 for altitude and slope, respectively. The principal component analysis identified annual average temperature, annual precipitation, and soil organic matter content as key influencing factors, with weights of 35, 30, and 25, respectively. These key influencing factors and their weights provide a scientific basis for further optimizing carbon sequestration.

[0088] Step 3-2: Obtain human activity data from each sample plot, determine the significance of differences in carbon sink function under different human activity levels through variance analysis, and use the support vector machine algorithm to establish a nonlinear prediction model of human activity and carbon sink function. If the model accuracy meets the preset threshold, the quantitative impact relationship between human activity and carbon sink function is determined;

[0089] Step (1) obtains the original monitoring data of various human activity indicators in each sub-area; among them, human activity indicators such as grazing intensity, fence density and fire frequency are pre-processed and standardized to form a normalized data set; here, grazing intensity can be obtained by field survey of the number of livestock per unit area; or by interviewing local herders to obtain the average grazing time, grazing rotation frequency, etc., and the number of livestock per unit area can be calculated based on the sample area. Fence density can be determined by extracting fence distribution from remote sensing images or by measuring the total length of fences per unit area on site. Fire frequency can be determined based on historical fire remote sensing data or by field investigation of fire events in recent years.

[0090] Step (2): Based on the normalized data set, variance analysis was used to calculate the significance of carbon sink function differences under different human activity levels;

[0091] Step (3): Based on the results of the significance analysis, human activity indicators with significant differences are selected as input features, and carbon sink function is selected as the output variable. A nonlinear prediction model of human activity and carbon sink function is constructed using the support vector machine algorithm; the model parameters are optimized by the cross-validation method to obtain the optimal prediction model of the impact of human activity on carbon sink function;

[0092] For nonlinear regression problems, the goal of the support vector machine (SVM) is to find a function f(X) such that:

[0093]

[0094] Where, X i is the input feature, which represents the feature vector of the i-th sample in the training data; X is the feature vector of the new sample to be predicted; K(X i , X) is the kernel function (here we use radial basis function RBF); α i is the weight of the support vector; b is the bias term, and n is the total number of training samples.

[0095] The kernel function is defined as: K(X i , X) = exp(-γ||X i -X|| 2 ), where γ is the kernel function parameter.

[0096] Step (4): Based on the optimal prediction model, determine the quantitative impact of human activities on carbon sink function.

[0097] Specifically, after obtaining human activity data such as grazing intensity, fencing density, and fire frequency, the authors first used variance analysis to determine the significance of differences in carbon sink function under different levels of human activity. For example, assuming grazing intensity was categorized as low, medium, and high; fencing density was categorized as sparse, medium, and dense; and fire frequency was categorized as no fire, low-frequency fire, and high-frequency fire. The variance analysis revealed a significant effect of grazing intensity on carbon sink function (F = 134, p < 0.05), an insignificant effect of fencing density on carbon sink function (F = 23, p > 0.05), and a significant effect of fire frequency on carbon sink function (F = 87, p < 0.05). Next, a nonlinear prediction model linking human activity and carbon sink function was constructed using a support vector machine algorithm. Assuming a radial basis function (RBF) kernel function, cross-validation determined the optimal parameters C = 10 and γ = 1. After model training, the model was validated using a test set, achieving a model accuracy of 92%, meeting the pre-set threshold of 90%. Finally, the model results confirmed the quantitative impact of human activity on carbon sink function. For example, model predictions show that for every unit increase in grazing intensity, carbon sequestration decreases by 15 tons per hectare, and for every increase in fire frequency, carbon sequestration decreases by 0.8 tons per hectare. These quantitative relationships provide a scientific basis for developing carbon sequestration management strategies.

[0098] Step 3-3: Comprehensively consider spatial heterogeneity, scale effects, environmental factors, and human activities, use the analytic hierarchy process to determine the weight of each factor, and combine it with the random forest algorithm to construct a grassland carbon sink function prediction model. Through cross-validation, determine the stability and reliability of the model, and obtain a regional scale carbon sink function distribution map, including:

[0099] Step (1) uses the analytic hierarchy process to construct a judgment matrix and calculate the weight parameters of spatial heterogeneity, scale effect, environmental factors and human activities;

[0100] Step (2) Using spatial heterogeneity, scale effect, environmental factors and human activities as input data, a carbon sink function prediction model is constructed using the random forest algorithm, the number of decision trees and feature sampling ratio are set, the model is trained, and the parameter configuration is optimized;

[0101] Step (3), the trained carbon sink function prediction model can be subjected to k-fold cross validation, and the final carbon sink function prediction model is obtained after the validation passes;

[0102] Specifically, in assessing grassland carbon sequestration, the analytic hierarchy process (AHP) was first used to determine the weights of various influencing factors. An evaluation index system encompassing spatial heterogeneity, scale effects, environmental factors, and human activities was constructed. A judgment matrix was constructed using expert scoring, and the calculated consistency ratio (CR) value was 0.12, meeting the requirement of less than 1. The final weights for each index were 35, 25, 20, and 20, respectively. Based on these weights, a carbon sequestration prediction model was constructed using the random forest algorithm, with a set number of 500 decision trees, a maximum feature number of sqrt(n_features), a maximum depth of 10, and a minimum number of leaf samples of 5. The model was trained using a 10-fold cross-validation method. The confusion matrix calculation yielded an average accuracy of 93% and a Kappa coefficient of 89, demonstrating high predictive accuracy and stability. The trained model was then applied to a regional-scale carbon sequestration assessment. Remote sensing imagery with a spatial resolution of 30 meters and environmental factor data, including 12 characteristic variables such as NDVI, surface temperature, and precipitation, were input, and a carbon sequestration level distribution map was output. Through spatial analysis, it was found that high carbon sink areas account for 32% of the study area, mainly distributed in areas with an altitude of 1200-1500 meters and a slope of less than 15 degrees. This is highly consistent with the results of field surveys, verifying the validity and reliability of the model.

[0103] Example 2

[0104] In order to implement the method for constructing a grassland carbon sink function prediction model provided in the above embodiment, the present application also provides a system for constructing a grassland carbon sink function prediction model, such as Figure 2 Shown, including:

[0105] A sub-region division module is used to obtain first environmental information of the area to be studied by using a meteorological station and remote sensing technology, and to divide the study area into several sub-regions by using a stratified sampling method;

[0106] The sample setting module is used to assess the carbon sequestration function at different spatial scales. It adopts a multi-scale nested sample layout scheme and sets multiple sample plots of different sizes in each sub-area.

[0107] The model building module is used to obtain secondary environmental information, human activity data, and carbon sequestration function indicator data for each sample plot; and uses sub-region type, sample plot, environmental factor data, and human activity data as input data, and carbon sequestration function indicator data as output data, combined with the random forest algorithm, to construct a grassland carbon sequestration function prediction model.

[0108] The grassland carbon sink function prediction model construction system provided in this application is to implement the grassland carbon sink function prediction model construction method of the above embodiment. The specific functions implemented by each module can be found in the above embodiment and will not be repeated here.

[0109] Example 3

[0110] Based on the same inventive concept, this application provides a method for predicting grassland carbon sink function, such as Figure 3 Shown, including:

[0111] Step 1: Obtain forecast data on climate change and human activities in the study area under multiple future scenarios;

[0112] Step 2: Based on each set of prediction data, use the grassland carbon sink function prediction model to predict the trend of carbon sink function changes under each set of scenarios in the future;

[0113] The grassland carbon sink function prediction model is obtained by a construction method of a grassland carbon sink function prediction model provided by the present invention.

[0114] For example, in this application, the prediction data of climate change and human activities in the area to be studied under multiple sets of future scenarios can be obtained based on existing technologies.

[0115] At the same time, in order to accurately predict the carbon sink function and discover the factors that affect the carbon sink function, climate change prediction data and human activity prediction data are provided from multiple angles from the optimal development direction and the worst development direction.

[0116] Specifically, based on the grassland carbon sink function prediction model, the current carbon sink status of the grassland ecosystem is first obtained through remote sensing data and ground monitoring data. For example, using MODIS satellite data, combined with carbon storage data from ground sampling points, it is concluded that the current grassland carbon sink is 6 tons per hectare. On this basis, different climate change scenarios are set, such as RCP6 and RCP5, representing low-emission and high-emission scenarios, respectively. The temperature and precipitation changes over the next 50 years are predicted through climate models and input into the grassland carbon sink function prediction model to predict future trends in carbon sink function changes. For example, under the RCP6 scenario, grassland carbon sinks are expected to increase to 8 tons per hectare in 2050, while under the RCP5 scenario, carbon sinks may drop to 2 tons per hectare. At the same time, considering human activity scenarios, such as grazing intensity and land use changes, different combinations of grazing intensities (low, medium, and high) and land use types (grazing land, farmland, and construction land) were set, and the grassland carbon sink function prediction model was used for analysis. The analysis results showed that grazing intensity had the most significant impact on carbon sink function. When the grazing intensity exceeded 5 sheep units per hectare, the carbon sink amount decreased significantly. In addition, among land use changes, the conversion of grassland to farmland would lead to a 30% reduction in carbon sink amount. Based on these quantitative analysis results, a scientific basis for grassland management is provided, such as the recommendation to control grazing intensity to below 0 sheep units per hectare and give priority to protecting grassland resources to maintain and enhance grassland carbon sink function, thereby reducing the greenhouse effect and improving the ecological environment.

[0117] The description of the above embodiments is only used to help understand the technical solutions and core ideas of this application; ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0118] Example 4

[0119] In order to realize the grassland carbon ash function prediction method provided in this application, this application also provides a grassland carbon sink function prediction system, such as Figure 4 Shown, including:

[0120] The data acquisition module is used to obtain forecast data on climate change and human activities in the study area under multiple scenarios in the future;

[0121] The prediction module is used to predict the trend of carbon sequestration function changes under various scenarios in the future based on each set of prediction data and using the grassland carbon sequestration function prediction model;

[0122] The grassland carbon sink function prediction model is obtained according to a method for constructing a grassland carbon sink function prediction model provided by the present invention. The specific functions implemented by each module are described in the above embodiment and will not be repeated here.

[0123] Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0124] Example 5

[0125] like Figure 5 As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0126] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of a hybrid energy storage capacity optimization configuration method in the above embodiment.

[0127] Example 6

[0128] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in the electronic device and, of course, extended storage media supported by the electronic device. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a hybrid energy storage capacity optimization configuration method in the above embodiment.

[0129] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0130] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0131] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0132] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0133] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for constructing a grassland carbon sink function prediction model, characterized in that: include: Using meteorological stations and remote sensing technology to obtain first environmental information of the area to be studied, and using a stratified sampling method to divide the study area into several sub-areas; In response to the carbon sequestration function assessment needs at different spatial scales, a multi-scale nested sample layout scheme was adopted, with multiple sample plots of different sizes set up in each sub-area; For each sample plot, the secondary environment information, human activity data, and carbon sequestration function indicator data are obtained; and the sub-region type, sample plot, secondary environment information, and human activity data are used as input data, and the carbon sequestration function indicator data is used as output data, combined with the random forest algorithm, to construct a grassland carbon sequestration function prediction model.

2. The method according to claim 1, wherein The stratified sampling method is used to divide the study area into several sub-areas including: Performing standardization processing on each data of the first environment information; Using a clustering algorithm to cluster the data of the first environmental information to obtain a number of sub-areas; The first environmental information includes climate factor data, terrain factor data, soil factor data and vegetation feature data.

3. The method according to claim 1, wherein In response to the carbon sequestration function assessment needs at different spatial scales, a multi-scale nested sample layout scheme is adopted, and multiple sample plots of different sizes are set in each sub-area, including: Nested quadrats of several scales are set in each sub-area, and the number of quadrats of each scale is initialized; The coefficient of variation method was used to revise the number of samples at each scale.

4. The method according to claim 1, wherein The method of obtaining the second environmental information, human activity data, and carbon sequestration function indicator data for each sample plot, and constructing a grassland carbon sequestration function prediction model using the sub-region type, sample plot, second environmental information, and human activity data as input data and the carbon sequestration function indicator data as output data in combination with a random forest algorithm includes: Based on each quadrat in each sub-region, second environmental information and carbon sequestration function index data are obtained, and an environmental factor evaluation index is constructed based on the second environmental information and carbon sequestration function index data to determine a number of lower-level environmental factors; obtaining human activity data for each sub-region, and determining a quantitative relationship between human activity and carbon sink function based on the human activity data and carbon sink function indicator data; Based on the quantitative relationship between sub-region types, sample scales of various types in each sub-region, environmental factor evaluation indicators and human activities on carbon sequestration function, the hierarchical analysis method is used to determine the weights of various environmental factor evaluation indicators and human activities, and combined with the random forest algorithm, a grassland carbon sequestration function prediction model is constructed.

5. The method according to claim 4, wherein The second environmental information and carbon sequestration function index data are obtained based on each quadrat in each sub-region, and the environmental factor evaluation index is constructed based on the second environmental information and carbon sequestration function index data to determine a number of lower-level environmental factors, including: Obtain data on various environmental factors and carbon sequestration indicators in each sample plot in each sub-region; Determine the regression coefficient corresponding to each environmental factor using a multiple linear regression model based on the environmental factor data and the carbon sink indicator data; extract a number of environmental factors from the environmental factors based on the regression coefficient and set weight coefficients; Based on the extracted environmental factors and weights, an environmental factor evaluation index system is established; Preferably, the acquiring of human activity data of each sub-region and determining the quantitative relationship between human activity and carbon sequestration function based on the human activity data and carbon sequestration function indicator data includes: Obtain monitoring data on various indicators of human activities in each sub-region; After annotating the monitoring data of various human activity indicators, variance analysis was used to determine the significance of carbon sink function differences under different human activity levels. Based on the determined significance of carbon sink function differences, human activity indicators with significant differences were selected as input features, carbon sink function data were used as output variables, and a support vector machine algorithm was used to construct a nonlinear prediction model of human activity and carbon sink function. and determining the carbon sink index value corresponding to the value range of each index value of human activities according to the nonlinear prediction model; Preferably, the grassland carbon sink function prediction model is constructed based on the quantitative relationship between the sub-region type, the scale of each sample plot in each sub-region, the environmental factor evaluation index and human activities on the carbon sink function, and the weight of each environmental factor evaluation index and human activities is determined by using the hierarchical analysis method, and combined with the random forest algorithm, including: The judgment matrix was constructed using the analytic hierarchy process to calculate the weights of each indicator in the environmental factor evaluation index system and each human activity indicator in each sub-region type and at each sample scale, and then the carbon sink function evaluation index system was constructed. Based on the sub-region types, the training and test sets were constructed using the scales of various plots, the values ​​and weights of each indicator in the carbon sequestration function evaluation index system as input data, and the carbon sequestration function indicator values ​​as output data. The random forest algorithm was then used to construct a carbon sequestration function prediction model. Setting the number of decision trees and the feature sampling ratio, training the carbon sequestration function prediction model based on the training set, and optimizing the parameters in the carbon sequestration function prediction model; The trained carbon sink function prediction model was subjected to k-fold cross validation based on the test set, and the final carbon sink function prediction model was obtained after passing the validation.

6. A system for constructing a grassland carbon sink function prediction model, characterized in that: include: A sub-region division module is used to obtain first environmental information of the area to be studied by using a meteorological station and remote sensing technology, and to divide the study area into several sub-regions by using a stratified sampling method; The sample setting module is used to assess the carbon sequestration function at different spatial scales. It adopts a multi-scale nested sample layout scheme and sets multiple sample plots of different sizes in each sub-area. The model building module is used to obtain secondary environmental information, human activity data, and carbon sequestration function indicator data for each sample plot; and uses sub-region type, sample plot, environmental factor data, and human activity data as input data, and carbon sequestration function indicator data as output data, combined with the random forest algorithm, to construct a grassland carbon sequestration function prediction model.

7. A method for predicting grassland carbon sequestration function, characterized in that: include: Obtain forecast data on climate change and human activities in the study area under multiple future scenarios; Based on each set of prediction data, the grassland carbon sink function prediction model is used to predict the changing trend of carbon sink function under each set of scenarios in the future; The grassland carbon sink function prediction model Obtained according to the method for constructing a carbon sink function prediction model according to any one of claims 1-7.

8. A grassland carbon sink function prediction system, characterized in that: include: The data acquisition module is used to obtain forecast data on climate change and human activities in the study area under multiple scenarios in the future; The prediction module is used to predict the trend of carbon sequestration function changes under various scenarios in the future based on each set of prediction data and using the grassland carbon sequestration function prediction model; The grassland carbon sink function prediction model Obtained according to the method for constructing a carbon sink function prediction model according to any one of claims 1-5.

9. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the method for constructing a grassland carbon sink function prediction model according to any one of claims 1 to 5 and / or the method for predicting grassland carbon sink function according to claim 7 are implemented.

10. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, it implements the method for constructing a grassland carbon sink function prediction model according to any one of claims 1 to 5 and / or implements a grassland carbon sink function prediction method according to claim 7.

Images