Prediction of Carbon Emissions from Natural and Human Systems Based on Correlation Analysis and Logistic Models

. With the increasing trend of global warming and the growing emphasis on carbon reduction programs at home and abroad, there is an urgent need to propose solutions to prevent the greenhouse effect. In order to solve such problems, exploring and predicting the trend of carbon dioxide emissions is essential to better establish targeted carbon reduction programs. By analyzing the relationship among carbon emissions, human and natural systems, the relationship between factors such as economic level status, resident population, road passenger traffic, and vegetation carbon sequestration capacity and carbon emissions is studied in five characteristic regions of China, and the predicted results are verified by carbon emission release.


Introduction
Carbon dioxide emissions are a major driver of climate change, and as the global warming trend increases dramatically, effective solutions to prevent the greenhouse effect are being sought and developed [1] . As the world's largest energy consumer, the Chinese government has announced that by 2020, CO2 emission intensity per unit of GDP will be reduced by 40%-45% compared to 2005. The Copenhagen Accord envisages that global warming should be limited to less than 2°C and that a reduction of 40%-70% is needed by 2050 compared to 2010 [2] .

Research ideas and key points
This study argues that human and natural systems and the related information derived from them can assist us in predicting carbon emissions. The study builds on and improves a system dynamics-based study of carbon emission relationships by Professor Yunpeng Ling [3] , which presents causality diagrams for the population subsystem, the economic system, and the environmental system, models the system dynamics of CO2 emissions, and proposes a causal circuit diagram between each factor.
Based on this article, this study not only considers the effects of traffic factors and vegetation carbon sequestration on CO2 emissions [4] , but also studies the mathematical relationships between the variables in a more detailed way with linear and nonlinear relationships, and describes them more specifically and carefully. The specific factors considered and their research ideas are shown in the Fig.1.

Selection of study area
Among the data of China Statistical Yearbook 2021 and the evaluation of geographic resources, economic conditions and other factors of each Chinese province and city in the article of Prof. L Ma [5] , we selected the representative five administrative regions, with Beijing Changping as an example, as the target regions of the study. The reasons for the selection and the regions are shown in Table.1. In the process, it is very difficult to grasp the relationship between variables, so we look for the research method from macro to micro, and complete the research according to mathematics and algorithms. The relationship between the economic level, population size, total carbon emission, road passenger volume, and urban energy efficiency is obtained, and at the end a comparison of the simulation results is made to prove that the information obtained by the model can accurately reflect the change pattern of the system.

Research Data Source
Our data comes from papers, the China Statistical Yearbook, and cedata's website, as shown in Table.2.

Methodology
The research objectives of regional carbon emission prediction modeling are divided into two categories: human systems and natural systems. Many studies have shown that population growth on a global scale contributes between 20% and 60% to the growth of CO2 emissions. Population affects the industrialization level, energy consumption ratio, and transportation energy efficiency of a region. Energy consumption carbon emissions and transportation carbon emissions become an important part of carbon emissions within human systems [6] , and therefore human systems have an extremely important impact on regional carbon emissions.
The natural system has a strong limiting effect on CO2 emission because of its green plants' strong air purifying and carbon sequestering ability. Due to the specificity of natural geography, its application in the substitution of primary energy sources can largely reduce the original energy consumption (especially the emissions caused by thermal power and fossil energy capacity), so the analysis for the regional establishment of carbon sequestration capacity and new energy potential can play a guiding role in the future projection of regional carbon emissions.

Correlation analysis of human systems on carbon emission projections
Analyzing several variables with correlation in human and transportation systems, so as to measure the closeness [7] of the two variable factors, which include the relationship between the state of the economic level, the number of resident population, the amount of road passenger traffic, and the carbon emissions, which have some connection between them,as shown in Fig.2-3.
In the formula, represents the economic level status, 2 represents carbon emissions, represents the number of resident population, and 、 are variable parameters. A similar relationship exists between road passenger traffic, urban energy efficiency and carbon emissions as described above.

Fig.4 People Predict of Logistic Model
After studying the mathematical relationships between the variables, we derived parametric results and predicted the number of permanent residents through a logistic model [8] , as shown in Fig.4. We predicted the carbon emission amount for the next 5 years accordingly, as shown in Fig.5.

Impact of natural factors on carbon emission projections
In the ecological cycle, the carbon sequestration capacity of vegetation has a negative effect on carbon emissions, and therefore its accounting becomes an essential part of carbon emission prediction. According to the population competition model and carbon sequestration model, the capacity of vegetation to sequester carbon in the region can be proposed by the following equation.Where equation (4) applies to single vegetation types and (5) applies to multi-vegetation types.
represents the future number of k species in the region, represents the current species variety, represents the total amount of carbon sequestered by plants in the region, Treekind represents the vegetation type, and represents the amount of CO2 that can be sequestered under that vegetation type (in years).
Considering the carbon sequestration capacity of vegetation is a more reasonable calculation conclusion for the whole carbon emission prediction model, and it can be seen from the Fig.6 that the trend of this conclusion is smoother and the calculation and fitting ability is stronger by considering human, transportation and nature.

Result
The number of resident population in the past ten years is used as the data set, and the Logistic model is used to predict the population in the next five years for the five provinces in subsection 2.1 [9] . The conclusions are as shown in Table.3. The comparative conclusions of the projected future CO2 emissions using the human system, transportation system relationship are shown in Table.4, due to the limitation to show only the projection results for 2020-2022. The conclusions of the CO2 emission projections after considering the vegetation carbon sequestration correction are shown in Table.

Conclusion
Prediction of carbon emissions from natural and human systems based on correlation analysis and logistic models was conducted to determine the relationship between carbon emissions and economic status, resident population, energy efficiency, transportation, and regional vegetation among five representative cities in China. The relationship between carbon emissions from natural and human systems based on correlation analysis and logistic models was determined for five representative cities in China, with respect to economic level, resident population, energy efficiency, traffic volume, and regional vegetation, and the results were very good. After the study, it can be concluded that the algorithm can effectively predict the carbon emissions in the region when the human factors, transportation factors and natural factors are determined [10] . This study provides a new method and idea for carbon emission prediction. Both correlation analysis and logistic model are commonly used forecasting models, and this study combines both of them to better predict carbon emissions. The method can be applied not only to natural and human systems, but also to carbon emission prediction in other fields.
The prediction model proposed in this study can not only provide decision support for government departments, but also provide guidance for enterprises and the public on carbon emission control. In conclusion, the study has important theoretical and practical significance, and provides new methods and ideas for carbon emission prediction and control.
Although we have derived mathematical models for carbon emissions in different provinces, we found that the parameters of the models are different for different regions. Therefore, in the future we need to explore the correspondence between regions and parameters, which may be determined by the development characteristics of the regions themselves, after which we can make our findings more complete.
In conclusion, it is expected that the study will be further developed and make greater contributions to carbon emission control and environmental protection.