Carbon neutralization evaluation method for ecological park

Abstract

The invention provides a carbon neutralization evaluation method for an ecological park, and the method comprises the steps: obtaining the basic data of various carbon assets in the ecological park, and the basic data comprises the natural ecological system carbon sink capability and human activity carbon emission intensity data; according to the carbon asset basic data, a time sequence analysis method is adopted to carry out modeling on the change trend of the carbon sink capacity and the carbon emission intensity, and a change trend model is obtained; modeling the spatial distribution of the carbon reserves and the carbon flux by utilizing a geographic information system technology and combining a change trend model to obtain a spatial distribution model; key variables are extracted from the spatial distribution model, and the effects of different factors on the carbon assets are quantified by adopting a multiple regression analysis method; acquiring carbon asset monitoring data, evaluating data quality, and if an abnormal value exists, performing correction by adopting a data cleaning algorithm; and according to the corrected carbon asset data, adopting a random forest model to predict the carbon sink capacity and the carbon emission intensity, and generating a dynamic evaluation result.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a carbon neutrality assessment method for an ecological park. Background Art

[0002] When constructing a carbon asset assessment model for an ecological park, many technical challenges are faced. First, there are a wide variety of carbon assets in the park, including natural ecosystems such as forests, grasslands, and wetlands, as well as human activities such as energy consumption and industrial production. The carbon sequestration capacities and carbon emission intensities of different types of carbon assets vary greatly. How to reasonably quantify and balance various carbon assets in the model requires comprehensive consideration of the dynamic change characteristics and interaction mechanisms of carbon assets. Second, the spatio-temporal distribution of carbon assets is uneven, and the variation laws of carbon storage and carbon flux in different regions and different periods are different. The model needs to be able to depict the spatio-temporal heterogeneity of carbon assets and reveal its internal driving mechanism. Third, the changes in carbon assets are affected by both natural and human factors. The model needs to be able to identify and distinguish the action directions and contribution sizes of different influencing factors, and quantitatively evaluate their influence degrees on carbon assets. Finally, the quality of carbon asset monitoring data is uneven, with problems such as inconsistent time frequencies, data missing, and outliers. The model needs to be able to screen and repair low-quality data, and reasonably fill data gaps to ensure the accuracy and continuity of the model input data. At the same time, the model also needs to address issues such as untimely data updates and poor connection between different data sources to ensure the timeliness and traceability of the assessment results. Summary of the Invention

[0003] The present invention provides a carbon neutrality assessment method for an ecological park, mainly including:

[0004] Obtain basic data of various carbon assets in the ecological park, including carbon sequestration capacity of natural ecosystems and carbon emission intensity data of human activities;

[0005] According to the basic data of carbon assets, adopt time series analysis method to model the change trends of carbon sequestration capacity and carbon emission intensity, and obtain a change trend model;

[0006] Utilize geographic information system technology, combined with the change trend model, to model the spatial distribution of carbon storage and carbon flux, and obtain a spatial distribution model;

[0007] Extract key variables from the spatial distribution model, and adopt multiple regression analysis method to quantify the effects of different factors on carbon assets;

[0008] Obtain carbon asset monitoring data, evaluate the data quality, and if there are outliers, use data cleaning algorithms for correction;

[0009] According to the revised carbon asset data, the random forest model is used to predict the carbon sequestration capacity and carbon emission intensity, and generate dynamic assessment results.

[0010] The technical solution provided by the embodiment of the present invention may include the following beneficial effects:

[0011] The present invention discloses a method for evaluating carbon neutrality in an ecological park. The method first obtains basic data such as the carbon sequestration capacity of the natural ecosystem and the carbon emission intensity of human activities in the park, and establishes a change trend model through time series analysis. Combining with geographic information system technology, a spatial distribution model of carbon storage and carbon flux is constructed, and multivariate regression analysis is used to quantify the influencing factors. The present invention also conducts quality assessment and cleaning of the monitoring data, uses the random forest model to predict the carbon sequestration capacity and carbon emission intensity, and generates dynamic assessment results. This method integrates spatio-temporal analysis, data mining and machine learning technologies, realizes the comprehensive evaluation and prediction of carbon assets in the ecological park, provides a scientific basis for carbon asset management and decision-making, and helps to improve the carbon neutrality level of the ecological park. Brief Description of the Drawings

[0012] Figure 1 It is a flowchart of a method for evaluating carbon neutrality in an ecological park of the present invention.

[0013] Figure 2 It is a schematic diagram of a method for evaluating carbon neutrality in an ecological park of the present invention. Detailed Embodiments

[0014] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention.

[0015] Such as Figure 1-2 , a method for evaluating carbon neutrality in an ecological park in this embodiment may specifically include:

[0016] S101. Obtain various types of basic carbon asset data in the ecological park, including the carbon sequestration capacity of the natural ecosystem and the carbon emission intensity data of human activities.

[0017] Obtain the type and area data of the natural ecosystem in the ecological park. Use the object-oriented classification method to interpret the remote sensing images of the ecological park to obtain the spatial distribution vector data of forest, grassland, and wetland natural ecosystems. Conduct accuracy verification on the interpretation results to determine whether the data quality meets the requirements of subsequent analysis. If it meets the requirements, for the forest ecosystem, use the allometric growth model to calculate the biomass of the forest ecosystem based on the input tree diameter at breast height and tree height data. For the grassland and wetland ecosystems, establish a regression model based on the statistical relationship between the normalized difference vegetation index and biomass to estimate the biomass of the grassland and wetland ecosystems. According to the biomass estimation results, combined with the preset carbon content conversion coefficient, calculate the carbon storage and annual carbon sink of various natural ecosystems. Obtain the type and scale data of human activities such as energy consumption, transportation, and waste treatment in the ecological park. Through field investigations and data collection, establish a human activity carbon emission inventory. Use the emission factor method to calculate the carbon emissions of various human activities based on the energy consumption and preset emission factors. Integrate the natural ecosystem carbon sink data with the human activity carbon emission data to analyze the carbon budget balance in the ecological park and obtain the analysis results of the carbon budget balance in the park.

[0018] Specifically, the object-oriented classification method is an advanced remote sensing image interpretation technology. It not only considers the spectral information of individual pixels but also comprehensively takes into account features such as the shape, texture, and spatial relationships of objects in the image, thus more accurately identifying land cover types. For example, when identifying forests, traditional pixel-based classification methods may misclassify shrubs with similar spectral characteristics as forests, while the object-oriented classification method can distinguish them from shrubs based on features such as the generally larger area and regular shape of forest patches. In the interpretation of remote sensing images of ecological parks, the image can first be segmented into homogeneous objects, and then features such as the spectrum, shape, and texture of each object are extracted. Using classifiers such as decision trees and support vector machines, these objects are classified to obtain the spatial distribution of different ecosystems such as forests, grasslands, and wetlands. Allometric models describe the proportional relationships between the growth of different parts of trees and are often used to estimate forest biomass. For example, the allometric model of a certain tree species is B = a * D^b * H^c, where B represents biomass, D represents diameter at breast height, H represents tree height, and a, b, and c are model parameters. Through field surveys, the average diameter at breast height of this tree species is measured to be 20 cm, and the average tree height is 15 m. Combining the existing model parameters, the average biomass per individual tree of this species can be calculated. Suppose the average biomass per individual tree of a dominant tree species in the park calculated by the allometric model is 0.5 tons, and the number of trees of this species is 10,000, then the total biomass of this tree species is 5000 tons. The Normalized Difference Vegetation Index (NDVI) is an important indicator reflecting vegetation growth status and has a significant correlation with biomass. For grassland and wetland ecosystems, a regression model between the two can be established using the measured biomass data and the corresponding NDVI values of remote sensing images. For example, through field surveys, the biomass of multiple grassland quadrats is measured at different locations in the park, and the corresponding NDVI values of these quadrats are obtained. Using these data, a linear regression model of NDVI and biomass is established, such as B = m * NDVI + n, where B represents biomass, NDVI is the Normalized Difference Vegetation Index, and m and n are model parameters. Suppose the average biomass of grasslands in the park calculated by the regression model is 3 tons per hectare, and the average biomass of wetlands is 4 tons per hectare. The carbon content conversion coefficient refers to the mass of carbon elements contained in the dry weight of unit biomass, and the carbon content conversion coefficients of different ecosystems are different. For example, the carbon content conversion coefficient of forest ecosystems is usually around 0.5, and those of grassland and wetland ecosystems are slightly lower. Combining the biomass estimation results and the carbon content conversion coefficient, the carbon storage of various ecosystems can be calculated. For example, if the forest biomass in the park is 5000 tons and the carbon content conversion coefficient is 0.5, then the carbon storage of the forest ecosystem is 2500 tons. The annual carbon sink is the amount of carbon dioxide absorbed by the ecosystem each year and can be estimated through the annual growth of biomass and the carbon content conversion coefficient.The establishment of a carbon emission inventory of human activities requires a detailed investigation of the types and scales of various human activities within the park. Through on-site investigations and data collection, detailed data on these activities can be obtained. For example, the annual electricity consumption of the office building is 1 million kWh, the annual driving mileage of employee commuting and official vehicles is 50,000 km, and the annual output of domestic waste is 100 tons. The emission factor method is a commonly used method for calculating carbon emissions from human activities. Its basic principle is to calculate the carbon emissions based on the consumption of various activities and the corresponding emission factors. For example, the emission factor for electricity consumption is 0.6 kg CO₂eq per kWh, the emission factor for gasoline is 2.3 kg CO₂eq per liter, and the emission factor for landfill is 0.5 tons CO₂eq per ton. Based on the consumption of various activities within the park and the corresponding emission factors, the carbon emissions of various human activities can be calculated. For example, if the annual electricity consumption of the office building in the park is 1 million kWh and the emission factor for electricity consumption is 0.6 kg CO₂eq per kWh, then the annual carbon emissions of the office building are 600 tons CO₂eq. By integrating the carbon sink data of natural ecosystems with the carbon emission data of human activities, the carbon budget balance within the park can be analyzed. For example, if the total carbon sink of forests, grasslands, and wetland ecosystems within the park is 100 tons of carbon per year, and the carbon emissions from human activities are 80 tons CO₂eq per year, then the net carbon sink of the park is 20 tons of carbon per year, indicating that the park as a whole is a carbon sink and has a positive effect on mitigating climate change. If the carbon emissions from human activities are greater than the carbon sink of the natural ecosystem, then the park is a carbon source and measures need to be taken to reduce emissions or increase the carbon sink. Through carbon budget analysis, data support can be provided for the management of ecological parks, scientific and reasonable carbon emission reduction and carbon sink enhancement strategies can be formulated, and the green and low-carbon development of the park can be promoted.

[0019] S102. Based on the basic carbon asset data, use the time series analysis method to model the changing trends of carbon sink capacity and carbon emission intensity, and obtain a changing trend model.

[0020] Specifically, based on the carbon asset basic data, use the Pandas library in Python to perform preprocessing operations on the data, such as cleaning, removing duplicates, and filling missing values, to ensure data quality and consistency. Split the preprocessed carbon asset data according to the time dimension, extract key indicators such as carbon sink capacity and carbon emission intensity, and construct a time series dataset. Use the ARIMA model in the statsmodels library of Python to perform modeling analysis on the constructed time series dataset. Optimize the model hyperparameters through grid search and cross-validation to obtain the final change trend prediction model. Use the change trend prediction model to predict the future change trends of carbon sink capacity and carbon emission intensity, and obtain the prediction results within a certain future time range. Use the Matplotlib and Seaborn libraries in Python to visually present the prediction results, and intuitively display the future changes of carbon sink capacity and carbon emission intensity in the form of line charts and area charts, providing decision-making support for carbon asset management.

[0021] Specifically, the preprocessing of carbon asset basic data is a key step to ensure the accuracy of subsequent analysis. Using the Pandas library can efficiently process large-scale data, such as cleaning the carbon emission data of an enterprise in an ecological park over the years. By deleting duplicate records and filling missing values (such as using the average value or interpolation method), the data quality can be improved. Splitting the data in the time dimension can better capture the change trends of carbon sink capacity and carbon emission intensity. Constructing a time series dataset is the basis for predictive analysis. For example, the carbon sink data of an ecological park in the past three years can be extracted, including indicators such as the annual afforestation area and tree growth, to form a continuous time series. Such a dataset can reflect the change law of carbon sink capacity over time and provide a basis for subsequent modeling. The ARIMA model is suitable for processing data with obvious trends and seasonality, such as the monthly carbon emissions of a factory. Optimizing the model parameters through grid search and cross-validation can improve the prediction accuracy. The visual presentation of the prediction results is crucial for decision-making support. Using Matplotlib to draw a line chart can intuitively display the future trend of carbon emission intensity, while the area chart of Seaborn can better show the cumulative effect of carbon sink capacity. For example, the prediction curve of the carbon sink capacity of an ecological park in the next three years can be drawn, and the expected effects under different afforestation policies can be superimposed and displayed at the same time to help decision-makers choose the optimal plan.

[0022] S103. Use geographic information system technology and combine it with the change trend model to model the spatial distribution of carbon storage and carbon flux, and obtain the spatial distribution model.

[0023] Obtain the geographical information data of the study area, including land use types, vegetation coverage, etc., and construct a geographical information database. At the same time, collect the historical monitoring data of carbon storage and carbon flux in the study area, preprocess the data, identify and remove outliers using methods such as box plots, and reasonably fill in the missing values. Carbon flux refers to the amount of carbon exchange between the ecosystem and the atmosphere per unit time, including carbon absorption (positive flux) and carbon emission (negative flux); spatially associate the geographical coordinates of the carbon storage and carbon flux monitoring points with the geographical information data, and use the spatial join tool in ArcGIS to assign the carbon storage and carbon flux values of each monitoring point to the corresponding geographical location. Use the trend analysis tool in ArcGIS to analyze the change trends of carbon storage and carbon flux in the time dimension. According to the trend analysis results, select an appropriate time series model, such as the ARIMA model, to establish a change trend model for carbon storage and carbon flux. Use the established change trend model to predict the future changes in carbon storage and carbon flux. Use the spatial analysis tool in ArcGIS to spatially match the predicted results of carbon storage and carbon flux changes with the geographical information data to obtain the spatial distribution prediction data of carbon storage and carbon flux in the study area. Adopt the Kriging interpolation method, use the "Geostatistical Wizard" module, select the spherical variogram model, determine the interpolation parameters based on the cross-validation results (the principle of minimizing the root mean square error), and perform interpolation processing on the spatial distribution prediction data of carbon storage and carbon flux to obtain a continuous spatial distribution layer. Evaluate the interpolation accuracy under different parameter combinations and select the optimal parameters for the final interpolation. According to the spatially distributed layer generated by interpolation, use the spatial statistics tool in ArcGIS to calculate the global Moran's I index of carbon storage and carbon flux and evaluate its spatial autocorrelation. Use the Geostatistical Analyst tool in ArcGIS to fit the empirical variogram of carbon storage and carbon flux and analyze its spatial heterogeneity characteristics. Combine the spatial distribution characteristics and the change trend model to construct a spatial distribution model of carbon storage and carbon flux. Use the random forest algorithm to train the spatial distribution model, and use grid search to optimize the hyperparameters of the random forest, including the number of decision trees, the maximum number of features, etc. Evaluate the prediction performance of the spatial distribution model through cross-validation and select the model parameters with the best performance. Use the optimized spatial distribution prediction model to predict the carbon storage and carbon flux in the study area. Set different future scenarios, such as land use change, climate change, etc., and simulate the changes in carbon storage and carbon flux under these scenarios. Evaluate the carbon budget balance under different scenarios to provide a quantitative basis for formulating carbon neutralization strategies.

[0024] Specifically, the construction of a geographic information database is the basis for carbon storage and carbon flux analysis. Taking a certain forest ecosystem as an example, data such as land use type, vegetation coverage, and terrain can be collected. Through a method combining remote sensing image interpretation and field surveys, a detailed land use classification map of the study area can be obtained, such as coniferous forest, broad-leaved forest, shrub forest, etc. Vegetation coverage can be calculated using the Normalized Difference Vegetation Index (NDVI). These data provide spatial background information for subsequent analysis. The preprocessing of historical monitoring data of carbon storage and carbon flux is crucial for ensuring the reliability of the analysis results. Taking the monitoring data of an ecological park over three years as an example, box plots are first used to identify outliers. Suppose in a certain summer, due to extreme weather events, the carbon flux shows an abnormally high value. These data points will appear as outliers in the box plot and need to be removed or corrected. For missing values, appropriate filling methods can be selected according to the data characteristics. For example, for carbon flux data with obvious seasonality, the historical average value of the same season can be used for filling. Spatial correlation analysis can combine carbon storage and carbon flux data with geographical location information. For example, associate the carbon storage data of forest plots with information such as the land use type and altitude where they are located. This step lays the foundation for subsequent spatial analysis and enables exploring the relationship between carbon storage and environmental factors. Time series analysis helps to reveal the long-term change trends of carbon storage and carbon flux. Taking a certain forest ecosystem as an example, the seasonal fluctuations and long-term growth trends of carbon storage can be captured through the ARIMA model. The model may show that as the forest ages, the carbon storage shows an increasing trend year by year, but the growth rate gradually slows down. This trend analysis provides a basis for predicting future carbon sink potential. Spatial interpolation is a key technology for extending discrete monitoring point data to the entire study area. Taking Kriging interpolation as an example, based on the carbon storage data of known monitoring points, the carbon storage distribution of the entire forest area can be estimated. By comparing the interpolation results of different variogram models (such as spherical model, exponential model), the optimal model can be selected. The interpolation results may show that the carbon storage is higher in areas with higher altitude and gentler slope, and lower in the marginal areas with more human disturbances. Spatial autocorrelation analysis can reveal the spatial distribution patterns of carbon storage and carbon flux. The global Moran's I index can quantify the spatial aggregation degree of the entire study area. For example, if the Moran's I index of carbon storage is significantly positive, it indicates that high carbon storage areas tend to cluster together, which may be related to specific terrain or vegetation types. The construction of spatial distribution models comprehensively considers geographical factors and time trends. Taking the random forest algorithm as an example, land use type, vegetation coverage, altitude, etc. can be used as predictive variables, and carbon storage as the target variable for modeling. By optimizing parameters through grid search, such as setting the number of decision trees to 500 and the maximum number of features to the square root of the total number of variables, the prediction accuracy of the model can be improved. Scenario analysis provides a scientific basis for formulating carbon neutrality strategies. For example, the change of carbon storage in the study area in the next 3 years under different afforestation policies can be simulated.The results may show that under the scenario of active afforestation, the regional carbon storage may increase by 30%, while it only increases by 10% under the scenario of maintaining the status quo. These quantitative results provide important references for decision-makers to weigh the effects of different policies.

[0025] S104. Extract key variables from the spatial distribution model and use the multiple regression analysis method to quantify the effects of different factors on carbon assets.

[0026] According to the spatial distribution model, obtain multi-dimensional data related to carbon assets, such as environment, geography, and climate, and construct a dataset of carbon asset influencing factors. Conduct an exploratory analysis of the dataset to understand the distribution characteristics, missing values, and outliers of the data, and perform necessary data cleaning and preprocessing. Use methods such as mean imputation or KNN imputation to handle missing values, use methods such as box plots or Z-scores to identify and handle outliers, and standardize the data at the same time. Analyze the correlation between each influencing factor and the carbon asset volume using the Pearson correlation coefficient, and screen out the key variables that are significantly correlated with the carbon asset volume according to the absolute value of the correlation coefficient and the significance level. For the selected key variables, check the multicollinearity problem between variables through the variance inflation factor (VIF). When the VIF value is greater than 10, use principal component analysis to extract comprehensive variables. Construct a multiple linear regression model, with the carbon asset volume as the dependent variable and the key variables as the independent variables, and establish a regression equation. Use the least squares method to estimate the parameters of the regression model, obtain the regression model coefficients, and conduct a significance test through the t-test to determine the influence degree of each key variable. Conduct a diagnostic analysis of the regression model to test the goodness of fit, residual distribution, and heteroscedasticity problems of the model, and judge the effectiveness and reliability of the model. At the same time, use 10-fold cross-validation to evaluate the prediction ability and generalization performance of the model through the root mean square error (RMSE) and the coefficient of determination (R²). On the basis of the regression analysis, consider the influence of spatial autocorrelation, and test the spatial autocorrelation of the residuals through the Moran's I index. When the Moran's I value is significantly greater than 0, it indicates that there is positive spatial autocorrelation in the residuals, and the spatial lag model is selected; when the Moran's I value is significantly less than 0, it indicates that there is negative spatial autocorrelation in the residuals, and the spatial error model is selected. Modify the estimated values of the regression model coefficients according to the test results. Use the modified regression model coefficients as weights, combine the actual data of each region, and calculate the quantitative evaluation results of carbon assets in each region through weighted summation, and use ArcGIS to generate a spatial distribution map of carbon assets to visually display the spatial distribution characteristics of carbon assets.

[0027] Specifically, constructing a dataset of carbon asset influencing factors is a key step in carbon asset assessment. Taking a certain forest ecosystem as an example, multi-dimensional data including vegetation coverage, average annual temperature, annual precipitation, soil type, terrain slope, etc. can be collected. In the data exploration stage, it may be found through a histogram that the average annual temperature follows a normal distribution, while the vegetation coverage follows a right-skewed distribution. For the missing precipitation data, the average value of adjacent meteorological stations can be used for filling. When using a box plot to identify outliers, it may be found that the carbon storage of some sample points is extremely high. After data preprocessing, a correlation analysis is carried out. Suppose the Pearson correlation coefficient shows that the correlation coefficient between vegetation coverage and carbon asset quantity is 0.85, the average annual temperature is -0.62, and the annual precipitation is 0.58, all of which are significant at the 0.05 significance level. This indicates that vegetation coverage has a strong positive impact on carbon asset quantity, while an increase in temperature may lead to a decrease in carbon asset quantity. In the multi-collinearity check, if it is found that the VIF values of soil organic matter content and soil type both exceed 10, the two variables can be considered to be combined into a "soil fertility index" through principal component analysis. When constructing a multiple linear regression model, suppose the obtained equation is: carbon asset quantity = 2.5 * vegetation coverage - 1.8 * average annual temperature + 1.2 * annual precipitation + 0.8 * soil fertility index + 0.5 * terrain slope + constant term. Through the t-test, it is found that the coefficients of all variables except terrain slope are significant at the 0.01 level. This means that in this ecosystem, the impact of terrain slope on carbon asset quantity is relatively small. Model diagnosis shows that the residuals follow a normal distribution, but the Q-Q plot shows slight deviations at high and low values, indicating that there may be certain biases in the model's prediction in extreme cases. The results of 10-fold cross-validation show that the average RMSE of the model is 0.35 and the R² is 0.78, indicating that the model has good prediction ability and generalization performance. Considering the spatial characteristics of carbon assets, a spatial autocorrelation test is carried out. Suppose the Moran's I value is 0.32 and the p-value is less than 0.01, indicating significant positive spatial autocorrelation. This means that high-carbon asset areas tend to cluster, which may be related to specific geographical features. Based on this, a spatial lag model is selected for correction to obtain the corrected coefficients considering spatial effects. Finally, using the corrected model coefficients and combining the actual data of each region, the quantitative assessment results of carbon assets are calculated. The spatial distribution map of carbon assets generated by ArcGIS may show that high-carbon asset areas are mainly concentrated in areas with high forest vegetation coverage, while carbon assets are relatively low in areas concentrated with grasslands. Such visualization results not only intuitively display the spatial distribution characteristics of carbon assets but also provide an important basis for formulating regional carbon neutralization strategies.

[0028] S105. Obtain carbon asset monitoring data, evaluate the data quality, and if there are outliers, use a data cleaning algorithm for correction.

[0029] Obtain carbon asset monitoring data through a data interface and store the obtained data in a MySQL database. The carbon asset monitoring data includes, but is not limited to: natural carbon sink data: vegetation coverage (based on Landsat-8 remote sensing images, resolution 30 meters, quarterly update), soil organic carbon content (laboratory determination, annual sampling); anthropogenic emission data: electricity consumption in the park (real-time monitoring by smart meters), transportation fuel consumption (GPS mileage statistics, monthly summary). For the stored carbon asset monitoring data, use the Pandas library in Python for data preprocessing, including converting string-type data to numerical type and filling missing values with the mean to ensure the consistency and integrity of the data format. According to the preset data quality assessment rules, use the Numpy library in Python to assess the quality of the preprocessed carbon asset monitoring data. Adopt the 3σ principle, calculate the mean and standard deviation of the data, and determine whether each data point is within the range of the mean ± 3 times the standard deviation. Data points outside the range are regarded as outliers. For outliers outside the range, different processing methods are selected according to the degree of abnormality. For extreme outliers that deviate significantly from the normal range, directly delete them; for minor outliers, replace them with the median to reduce the impact of outliers on the overall data distribution. While processing outliers, use the Z-score normalization method to calculate the Z-score value of each data point, that is, the difference between the data point and the mean divided by the standard deviation. Data points with a Z-score value exceeding the preset threshold (such as ±3) are marked as outliers. For data points marked as outliers, replace them with the median to ensure the continuity and integrity of the data. Compare the processed carbon asset monitoring data with the original data, and use the Scipy library in Python to calculate statistical indicators such as the mean, median, and standard deviation before and after data processing. By comparing the data distribution before and after processing, evaluate the effectiveness of the outlier processing and normalization methods. According to the evaluation results, adjust the threshold parameter of Z-score normalization to optimize the accuracy and recall rate of outlier detection and processing, and improve the robustness and adaptability of the data cleaning algorithm. Apply the optimized data cleaning algorithm to the carbon asset monitoring data again to obtain high-quality cleaned data.

[0030] Specifically, the raw data obtained through the data interface may include various types, such as carbon dioxide concentration, vegetation coverage rate, soil carbon content, etc. After these data are stored in the MySQL database, systematic preprocessing is required. For example, convert the string "45.2%" to the floating-point number 0.452 to ensure the consistency of data types. For missing data, such as the data gaps caused by equipment failures at some monitoring stations, the mean value of the historical data of that station can be used for filling to ensure the continuity of the data. Data quality assessment is an important link to ensure the reliability of the analysis results. When using the 3σ principle for outlier detection, assume that the mean carbon dioxide concentration at a certain monitoring point is 400 ppm and the standard deviation is 20 ppm. Then, the data outside the range of 340 ppm to 460 ppm will be regarded as outliers. For data that deviates significantly, such as the suddenly appearing 900 ppm, it may be caused by equipment failures or human interference and should be directly deleted. For slightly deviated data, such as 430 ppm, the median can be used for replacement to reduce the impact on the overall distribution. The Z-score normalization method can further refine the judgment of outliers. For example, after normalizing the data of all monitoring points, the data point with a Z-score value of 2.8, although within the 3σ range, may still be regarded as a potential outlier. By adjusting the Z-score threshold, the strictness of outlier detection and the amount of data retained can be balanced. This method is particularly suitable for situations where there are large differences in the data ranges between different monitoring points, such as comparing the carbon absorption capacities of forests and grasslands. The effectiveness evaluation of data processing is crucial for optimizing the algorithm. Assume that the mean carbon dioxide concentration in the raw data is 410 ppm and the standard deviation is 25 ppm. After processing, the mean becomes 405 ppm and the standard deviation drops to 20 ppm. This change indicates that the impact of outliers has been effectively reduced and the data distribution is more concentrated. However, if the processed mean deviates significantly from the raw data, such as dropping to 380 ppm, it may mean overprocessing and the algorithm parameters need to be readjusted. The optimized data cleaning algorithm should be able to adapt to different types of carbon asset monitoring data. For example, for the vegetation coverage rate data with obvious seasonal changes, the algorithm can adjust the outlier judgment criteria according to different seasons. During the growing season, larger numerical fluctuations are allowed; while during the dormant season, stricter criteria are adopted. This dynamic adjustment can improve the adaptability of the algorithm and ensure that the cleaned data not only retains the real environmental change information but also removes unreasonable outliers. Through this series of data processing and optimization steps, the finally obtained high-quality data will provide a reliable basis for carbon asset evaluation. These cleaned data can not only more accurately reflect the actual carbon emissions and absorption situations but also provide strong support for formulating carbon neutrality strategies. For example, by analyzing the processed data, the areas with the highest carbon absorption efficiency can be identified, providing a scientific basis for enhancing the carbon sink capacity of ecological parks.Meanwhile, these data can also be used to build a prediction model to help decision-makers anticipate potential peak carbon emissions in advance and formulate more targeted emission reduction measures.

[0031] S106. Based on the corrected carbon asset data, use the random forest model to predict the carbon sink capacity and carbon emission intensity, and generate a dynamic assessment result.

[0032] Obtain the original dataset from the carbon asset-related data. Use the mean filling method to handle missing values and the box plot method to identify and remove outliers. Perform min-max standardization on the processed data to obtain the cleaned carbon asset dataset. Extract the features related to carbon sink capacity and carbon emission intensity from the cleaned dataset, and use the information gain ratio method to screen out the key feature subset that has a greater impact on the prediction result. Use the key feature subset as the input to build a random forest regression model, set the number of trees to 100, and the minimum number of samples at the node to 5. Optimize the model hyperparameters through grid search. Use 5-fold cross-validation to train and evaluate the model, record the mean absolute error and mean squared error as the model performance indicators. Apply the trained model to the new carbon asset data to predict the carbon sink capacity and carbon emission intensity, and compare the prediction results with the actual values to calculate the R² value to evaluate the model prediction accuracy. According to the prediction results, combined with the dynamic changes of carbon assets, use the weighted average method to calculate the comprehensive score of carbon assets and divide the grades. Use Matplotlib to draw the change curves of carbon sink capacity and carbon emission intensity, generate a carbon asset assessment report, and display the comprehensive score and grade distribution in a heat map.

[0033] Specifically, first, obtain the original dataset from multiple sources, such as satellite remote sensing data, ground monitoring point data, etc. For missing values, the mean filling method is adopted. For example, if the carbon sink data of a monitoring point is missing in June 2023, the average value of the same period in the past 5 years of this monitoring point can be used for filling. For outliers, the box plot method is used to identify and remove them. Suppose there is an extremely high outlier far beyond the normal range in the carbon emission intensity data of an ecological park, which may be caused by equipment failure and should be removed. Data standardization is an important step to ensure the comparability of indicators with different dimensions. The min-max standardization method is adopted to map each indicator value to the 0-1 interval. For example, the carbon sink capacity is converted from the original 0-500 tons / hectare / year to a standardized value of 0-1. This process helps the stability and accuracy of subsequent model training. Feature selection is the key to improving the model efficiency. The information gain ratio method is used to screen key features, such as indicators with greater influence on the carbon sink capacity, such as vegetation coverage rate, soil organic matter content, etc. This can not only reduce the model complexity but also improve the prediction accuracy. The construction of the random forest regression model is the core of predicting carbon asset changes. Set 100 decision trees, with the minimum number of samples per node being 5, and optimize hyperparameters such as the maximum depth of the tree, the number of features, etc. through grid search. Use 5-fold cross-validation to evaluate the model performance and record the mean absolute error and mean squared error. For example, when the model predicts the carbon sink capacity of a certain area, the mean absolute error is 0.5 tons / hectare / year, and the mean squared error is 0.3, indicating that the model has good prediction ability. When the model is applied to new data, it can predict the future carbon sink capacity and carbon emission intensity. For example, the predicted carbon emission intensity of the ecological park in 2025 is 4.2 tons of CO2 equivalent. Compared with the actual value of 4.3 tons, the R-squared value reaches 0.95, showing high accuracy. Based on the prediction results and combined with historical data, the weighted average method is used to calculate the comprehensive score of carbon assets. For example, the carbon sink capacity score of a forest area is 85, the carbon storage stability score is 90, and the comprehensive score is 87.5, which can be classified as Class A carbon assets. Visualization is an effective means to intuitively display the evaluation results. Use Matplotlib to draw the change curves of carbon sink capacity and carbon emission intensity to clearly present the trends. The heat map can show the distribution of carbon asset scores in different regions, helping decision-makers quickly identify key areas of concern.

[0034] Although the present invention has been described in detail with general descriptions and specific embodiments above, based on the present invention, some modifications or improvements can be made, which are obvious to those skilled in the art. Therefore, these modifications or improvements made without departing from the spirit of the present invention all fall within the scope of protection required by the present invention.

Claims

1. A carbon neutrality assessment method for an ecological park, characterized in that, The method includes: Obtaining basic data of various carbon assets in the ecological park, including the carbon sequestration capacity of natural ecosystems and carbon emission data of human activities; According to the basic carbon asset data, using the time series analysis method, modeling the changing trends of carbon sequestration capacity and carbon emission intensity to obtain a changing trend model; Using geographic information system technology, combined with the changing trend model, modeling the spatial distribution of carbon storage and carbon flux to obtain a spatial distribution model; Extracting key variables from the spatial distribution model and using the multiple regression analysis method to quantify the effects of different factors on carbon assets; Obtaining carbon asset monitoring data, evaluating the data quality, and if there are outliers, using data cleaning algorithms for correction; According to the corrected carbon asset data, using the random forest model to predict the carbon sequestration capacity and carbon emission intensity, and generating a dynamic assessment result.

2. The carbon neutrality assessment method for the ecological park according to claim 1, characterized in that, The obtaining of basic data of various carbon assets in the ecological park, including the carbon sequestration capacity of natural ecosystems and carbon emission data of human activities, includes: Obtaining the type and area data of natural ecosystems in the ecological park, and using the object-oriented classification method to interpret the remote sensing images of the ecological park to obtain the spatial distribution vector data of forest, grassland, and wetland natural ecosystems; For the forest ecosystem, using the allometric growth model, and calculating the biomass of the forest ecosystem according to the input tree diameter at breast height and tree height data; For grassland and wetland ecosystems, based on the statistical relationship between the normalized difference vegetation index and biomass, establishing a regression model to estimate the biomass of grassland and wetland ecosystems; According to the biomass estimation results, combined with the preset carbon content conversion coefficient, calculating the carbon storage and annual carbon sequestration of various natural ecosystems; Obtaining the type and scale data of human activities such as energy consumption, transportation, and waste treatment in the ecological park, and establishing a human activity carbon emission inventory through field investigations and data collection; Using the emission factor method, and calculating the carbon emissions of various human activities according to the energy consumption and the preset emission factors; Integrating the carbon sequestration data of natural ecosystems and the carbon emission data of human activities, analyzing the carbon budget balance in the ecological park, and obtaining the carbon budget balance analysis result of the park.

3. The carbon neutrality assessment method for the ecological park according to claim 1, characterized in that, The using of the time series analysis method according to the basic carbon asset data to model the changing trends of carbon sequestration capacity and carbon emission intensity to obtain a changing trend model includes: Obtaining the basic carbon asset data, which includes carbon sequestration capacity data and carbon emission intensity data; Preprocessing the basic carbon asset data to obtain the preprocessed carbon asset data; The preprocessing includes data cleaning, deduplication, and missing value filling; Slicing the preprocessed carbon asset data according to the time dimension, extracting the carbon sequestration capacity and carbon emission intensity indicators, and constructing a time series data set; Using the ARIMA model to conduct modeling analysis on the time series data set to obtain a time series prediction model; Optimizing the hyperparameters of the time series prediction model through grid search and cross-validation to obtain a changing trend prediction model; Using the changing trend prediction model to predict the future changing trends of carbon sequestration capacity and carbon emission intensity to obtain prediction results within a certain future time range; Visualize the prediction results to obtain the future changes in carbon sequestration capacity and carbon emission intensity.

4. The carbon neutrality assessment method for the ecological park according to claim 1 or 3, characterized in that Using geographic information system technology and combining with a change trend model, model the spatial distribution of carbon storage and carbon flux to obtain a spatial distribution model, including: Obtain the geographic information data of the study area, including land use type and vegetation coverage, and construct a geographic information database; Obtain the historical monitoring data of carbon storage and carbon flux in the study area, preprocess the data, use the box plot method to identify and remove outliers, and reasonably fill in the missing values; Spatially associate the geographic coordinates of the carbon storage and carbon flux monitoring points with the geographic information data; Establish a time series prediction model according to the temporal dimension change trend of carbon storage and carbon flux; Use the time series prediction model to predict the future changes in carbon storage and carbon flux; Spatially match the predicted carbon storage and carbon flux change results with the geographic information data; Adopt the Kriging interpolation method to interpolate the spatial distribution prediction data of carbon storage and carbon flux to obtain a continuous spatial distribution layer; Calculate the global Moran's I index of carbon storage and carbon flux to evaluate their spatial autocorrelation; Fit the empirical variogram of carbon storage and carbon flux to analyze their spatial heterogeneity characteristics; Construct a spatial distribution model of carbon storage and carbon flux, and use the random forest algorithm for training and optimization; Use the optimized spatial distribution model to predict the carbon storage and carbon flux in the study area; Set different future scenarios, simulate the changes in carbon storage and carbon flux under the scenarios, evaluate the carbon budget balance, and provide a quantitative basis for formulating carbon neutralization strategies.

5. The carbon neutrality assessment method for the ecological park according to claim 1, characterized in that, Extract key variables from the spatial distribution model and use the multiple regression analysis method to quantify the effects of different factors on carbon assets, including: Obtain multi-dimensional data on the environment, geography, and climate related to carbon assets, and construct a dataset of carbon asset influencing factors; Conduct an exploratory analysis of the dataset, use the mean filling or KNN filling method to handle missing values, use the box plot or Z-score method to identify and process outliers, and standardize the data at the same time; Analyze the correlation between each influencing factor and the carbon asset quantity using the Pearson correlation coefficient, and screen out the key variables that are significantly correlated with the carbon asset quantity; For the key variables, check the multicollinearity problem between variables through the variance inflation factor. When the VIF value is greater than 10, use the principal component analysis to extract the comprehensive variables; Construct a multiple linear regression model, use the carbon asset quantity as the dependent variable and the key variables as the independent variables, use the least squares method to estimate the parameters of the regression model, and determine the influence degree of each key variable through the t-test; Conduct a diagnostic analysis of the regression model, use 10-fold cross-validation, and evaluate the prediction ability and generalization performance of the model through the root mean square error and the determination coefficient; Test the spatial autocorrelation of the residuals through the Moran index. When the Moran index value is greater than 0, use the spatial lag model; When the Moran index value is less than 0, use the spatial error model, and correct the estimated value of the regression model coefficient according to the test results. Using the coefficients of the corrected regression model as weights, combined with the actual data of each region, the quantitative evaluation results of carbon assets in each region are calculated through weighted summation, and a spatial distribution map of carbon assets is generated to display the spatial distribution characteristics of carbon assets.

6. The carbon neutrality assessment method for the ecological park according to claim 1, characterized in that The acquisition of carbon asset monitoring data, the evaluation of data quality, and if there are outliers, data cleaning algorithms are used for correction, including: Acquire carbon asset monitoring data and store the carbon asset monitoring data in a database; Preprocess the carbon asset monitoring data, convert the data of string type to numerical type, and use the mean to fill in the missing values; According to the preset data quality evaluation rules, evaluate the quality of the preprocessed carbon asset monitoring data; Adopt the 3σ principle, calculate the mean and standard deviation of the carbon asset monitoring data, and judge whether each data point is within the range of mean ± 3 times the standard deviation. The data points outside the range are regarded as outliers; According to the degree of abnormality of the outliers, different processing methods are adopted. For the extreme outliers that are significantly deviated from the normal range, they are directly deleted; For mild outliers, the median is used for replacement; Calculate the Z-score value of each data point, mark the data points with Z-score values exceeding the preset threshold as outliers, and use the median for replacement; Compare the processed carbon asset monitoring data with the original data, calculate the statistical indicators before and after data processing, adjust the threshold parameters of Z-score normalization, and optimize the accuracy and recall rate of outlier detection and processing; Apply the optimized data cleaning algorithm to the carbon asset monitoring data to obtain high-quality cleaned data.

7. The carbon neutrality assessment method for the ecological park according to claim 6, characterized in that, According to the corrected carbon asset data, using the random forest model, predict the carbon sequestration capacity and carbon emission intensity, and generate dynamic evaluation results, including: Acquire carbon asset-related data, and for the missing values in the carbon asset-related data, use the mean filling method for processing; For the outliers in the carbon asset-related data, use the box plot method for identification and remove the outliers to obtain the processed carbon asset data; Perform min-max standardization processing on the processed carbon asset data to obtain a cleaned carbon asset data set; Extract the features related to carbon sequestration capacity and carbon emission intensity from the cleaned carbon asset data set, and use the information gain ratio method to screen the features to obtain a key feature subset; Using the key feature subset as the input, construct a random forest regression model. The number of trees in the random forest regression model is 100, and the minimum number of samples at the nodes is 5. Optimize the hyperparameters of the random forest regression model through the grid search method; Train and evaluate the forest regression model to obtain the mean absolute error and mean square error; Apply the trained random forest regression model to new carbon asset data, predict the carbon sequestration capacity and carbon emission intensity, compare the prediction results with the actual values, and calculate the R² value; According to the prediction results, combined with the dynamic changes of carbon assets, use the weighted average method to calculate the comprehensive score of carbon assets and divide the grades; Use Matplotlib to draw the change curves of carbon sequestration capacity and carbon emission intensity, generate a carbon asset evaluation report, and display the comprehensive score and grade distribution in a heat map.

Images