Construction of Ecological Conservation and Assessment of Its Impact on Environment

In recent years, environmental problems caused by global warming--- melting glaciers, rising sea levels, frequent wildfires and the death of large coral reefs, moved from the pages of the books to reality. To solve the problem, we develop several mathematical models, using both qualitative and quantitative methods, analyzing the importance of each influencing factor, finally build evaluation model of ecological environment and nature reserve. Firstly, in order to make a comparative analysis of the environmental conditions before and after the restoration of Saihanba, I collect the data of forest coverage rate, water conservation, PM2.5, oxygen release and carbon dioxide absorption of Saihanba in 60 years from 1962 to 2021, establishing Entropy Weight Method and Technique for Order Preference by Similarity to an Ideal Solution model (EWM-TOPSIS). Obtained that water conservation was the index with the largest weight, forest coverage rate is the index with the least weight, and the ecological environment in this area is the best in 2021 and the worst in 1962, indicating that saihanba restoration plays a significant role in improving the environment in this area. Secondly, I analyze the correlation between saihanba and Beijing's ecological environment, and qualitatively judge the degree of impact between the two. Then I take five ecological environment indexes of Saihanba as independent variables and four indexes of Beijing's anti-sandstorm ability as dependent variables of four Multiple Linear Regression models (MlR), which helps me make quantitative judgment, and finally get the conclusion that the improvement of the ecological environment of Saihanba will promote the anti-sandstorm ability of Beijing. After modelling, I also conduct sensitivity analysis to experiment the sensitivity of some parameters included in my models. In the end, I summarize strengths and weaknesses and a brief conclusion is also presented.


Background
For over a half century, builders from the Saihanba Tree Farm dedicated themselves to creating a miracle of turning desert into forests. With practical action, they forge the spirit of Sekhangba and interpret the concept that lucid waters and lush mountains are invaluable assets.The forest coverage rate of Seyhanba Mechanical Forestry Field has increased from 11.4% at the early stage of construction to 82% nowadays, constructing a green barrier for Beijing, Tianjin and Hebei regions. If the forest trees are arranged at a spacing of one meter, they can circle the earth's equator 12 times. (Liu, Y., 2021) However, they have encountered new problems on the road of development. Thus, people in Saihanba now own a higher mission of restoring the ecology.

Literature Review
The recovering of Saihanba Tree Farm from desert to an eco-friendly and green farm is marvelous, which becomes the world's largest artificial forest. Nowadays, Saihanba plays an important role in planting trees to fix the sand and conserve water resources with the perseverance of several young generations (Zou, C.-N., et al., 2021). General Secretary Xi Jinping (2021) propose six principles for building a community of human and natural life based on the Spirit of Saihanba, pointing out that the precious spirit created by Saihanba Mechanical Forest farm contributed Chinese wisdom to solving global ecological problems and promoted the construction of a community of shared future for mankind.
Gao Hainan (2019) points out that Saihanba Forest Farm is representative of the evolution of state-owned forest farms, and its ecological evolution reflects the transformation of the relationship between man and nature from "human-centered theory" to "respect for nature and harmonious coexistence between man and nature". Chen Yuanyuan (2019) takes Saihanba Mechanical Forest Farm as the research object and uses AHP to evaluate the sustainable development status of the forest farm. The result shows that the comprehensive evaluation index is 0.810, which is close to the target value of sustainable forest management. Therefore, the forest farm is considered to be in a state of sustainable development.
The positive impact of Saihanba Forest farm on the ecological environment further promotes the goal of carbon neutrality in China.Carbon neutrality means that enterprises, groups, or individuals measure the total amount of greenhouse gas emissions produced directly or indirectly within a certain period, and offset their carbon dioxide emissions by planting trees, saving energy, and reducing emissions, to achieve "zero" carbon dioxide emissions (Zou, C.-N., et al., 2021). Since the 16th Party Congress, the Party Central Committee and the State Council have proposed new ideas and initiatives to establish and implement the scientific outlook on development, build a harmonious socialist society, build a resource-saving and environment-friendly society, make rivers and lakes rest, and promote the historical transformation of environmental protection, etc. (Ministry of Ecology and Environment, PRC.). Based on this problem, this paper evaluates the impact on the environment after the restoration of Saihanba.
My work mainly includes the following: 1) select the forest coverage rate, the conservation of water, the amount of oxygen release, the amount of carbon dioxide absorption and the concentration of PM2.5 as the environment indicators and then build EWM-TOPSIS Model to evaluate the impact on the environment after the restoration of Saihanba.
2) Based on relevant data of the anti-sandstorm ability of Beijing (i.e., the monthly data of the average air quality index, the proportion of bad weather, the extreme low temperature and the extreme high temperature in 2017-2021.), analyzing the correlation of the ecological environment of Saihanba and the anti-sandstorm ability of Beijing, then establishing Regression Model to estimate the effect of Saihanba to the anti-sandstorm ability of Beijing.

Assumptions and Justifications
Aiming to realize the discussion of the employment of college graduate, we could comprehensive consideration which needs appropriate assumptions for better simplified and understood. The assumptions are as follows.
1. Carbon emissions only refer to carbon dioxide.
In addition to carbon dioxide and carbon monoxide, the carbon-containing gases in the air are ignored because of their low concentrations.

Hypothesis of linear regression model:
The residual of the constructed regression model follows a normal distribution with zero mean and homoskedasticity.
The independent variables are exogenous explanatory variables, and there is no perfect multicollinearity between independent variables. The residuals are not correlated with the independent variables 3. It is assumed that air quality and water quality are defined by the same standard across all regions and countries.
4. The number of divisions in each nature reserve is proportional to the size of the whole reserve.
In reality, the number of subdivisions in each nature reserve is restricted by local administrative divisions, and it is normal for them to vary in size. But to simplify models, we assume that the regions are evenly divided.
5. Species are evenly distributed in the certain nature reserve, and all types of vegetation have the same carbon fixation capacity, so the size is proportional to the carbon fixation capacity.

Entropy Weight Method (EWM)
Entropy is a concept in information theory, which is the measure of uncertainty. The greater the amount of information, the smaller the uncertainty, the smaller the entropy, and vice versa. Accordingly, the entropy values can be used to measure the forest coverage rate, water conservation, PM2.5, oxygen release, and carbon dioxide absorption. The smaller the entropy value, the greater the dispersion degree of the indicator, and the greater the influence (i.e., weight) of the indicator on the local ecological environment.

Techni
Step 1: Da and norma For positiv absorption For negativ The matrix Step 2   The weight calculation results of the entropy weight method indicate that the weight of forest coverage rate is 11.149%, the weight of oxygen release is 16.249%, the weight of water conservation is 33.973%, the weight of carbon dioxide absorption is 16.249%, and the weight of urban PM2.5 concentration is 22.381%, where the maximum value of index weight is water conservation (33.973%) while the minimum is the forest coverage rate (11.149%). Positive and negative ideal solutions (non-distance) respectively represent the maximum or minimum values of the evaluation index (i.e., optimal and inferior solutions). Using equations () and (), the optimal and inferior matrix vectors can be determined, i.e., the optimal matrix vector Z+ = (0.29, 0.26, 0.54, 0.26, 0.37); the inferior matrix vector Z-= (0.00000122, 5.8e-7, 0.00002195, 4e-7, 7.7e-7) Using equations () and (), we can obtain the positive ideal solution distance D+ and negative ideal solution distance D-. The smallest value of D+ in 2021 shows that it is the closest to the optimal solution; the smallest value of D-in 1962 indicates that it is the closest to the worst solution. Using equation (), we are capable of calculating the integrate score C value for each year. The C value in 1962 was the smallest and ranked the last; while in 2021 it is the largest and ranks the first.

Results Analysis
According to the calculation results of the entropy weighting method, the degree of influence of different indicators on the ecological environment can be obtained. The weight of water conservation is the largest, illustrating that the current state of the water environment has the greatest impact on the ecological environment; while from the perspective of economic evaluation of forest resources, the weight of forest coverage rate is the smallest, indicating that the current state of forest environment does not have much impact on the ecological environment. After establishing the importance of different indicators, the ecological environment of different years is considered as the evaluation object and analyzed by TOPSIS. From the ranking results, it is evident that the environmental conditions were the worst in 1962 and the best in 2021. It shows that after the restoration of Saihanba, the ecological environment gradually improved, so the restoration of Saihanba plays an essential role in improving the environmental conditions.

Correlation and Regression Analysis
Select the five indicators of forest coverage rate, water conservation, PM2.5 concentration, oxygen release, and carbon dioxide absorption to measure the ecological environment of Saihanba. Besides, choose the four indicators of extremely high temperature, extreme low temperature, average air quality index, and percentage of severe weather in Beijing to measure the ability of Beijing to withstand sandstorms.
The Pearson product-moment correlation coefficient between the five indicators of the ecological environment of the Saihanba Forest and the four indicators of the ability of Beijing to resist sandstorms is firstly established, which can be used to measure the strength of the linear correlation between two variables X and Y, whose values range between -1 and +1. The magnitude of the PMCC is used to roughly determine the extent to which the ecological environment of the Saihanba Tree Farm contributes to the protection of Beijing against sandstorms.
Next, establish four multiple regression models with five ecological indicators of Saihanba as independent variables and four indicators of Beijing's sandstorm resistance as dependent variables. Linear regression is a statistical analysis method that uses regression analysis in mathematical statistics to determine the quantitative relationships between variables that are dependent on each other.
Step 1: The PMCC between two variables is defined as the quotient of the covariance of these two variables and the standard deviation of both. Here we use the sample correlation coefficient r because only part of years of data from Saihanba and Beijing were extracted for analysis, and the variable Sample correlation coefficient: Step 2: Suppose the four linear regression model of the random variable y (extreme low temperature, extreme high temperature, average air quality index, proportion of bad weather) and the general variable 1 x , 2 x ,..., p x (forest coverage rate, conservation of water, concentration of PM2.5, the amount of oxygen release, the amount of carbon dioxide absorption) where 0 1 , ,..., p    is p +1 unknown parameters, 0  are called the regression constant, and 1 ,..., p   are called the regression coefficient.  is the random error term, which is required to satisfy the following two assumptions.
It is regarded as a mathematical model to evaluate the ability of the Saihanba to resist sandstorms in Beijing, and then quantitatively to acquire the regression relationship between them.  is obvious tha e average air q prove the air q e extremely hi increase of wa nsidering that t be done in the reduce the f PM2.5 conce ing. That is, t ssibility of sand

Sensitiv
In the sec multivaria independe introduce is necessar

Signifi
The signif

Goodness-of-Fit Test of the Regression Equation
The goodness of fit is used to test how well the regression equation fits the sample observations. In multivariable linear regression, the sample coefficient of determination is defined as The value of the sample coefficient of determination is within the interval 0,1     . The closer it is to 1, the better the regression fits; the closer it is to 0, the worse the regression fits.

Multicollinearity Test of the Regression Equation
When the independent variables are related as above, multi-collinearity is presented between the independent variables. The variance expansion factor method is applied to test the model for multi-collinearity. A central normalization is done for the independent variables and noted as Regard its leading diagonal element VIF  c jj as the variance inflation factor VIF   of the independent variable x j , where When VIF  10, it means that there is a serious multi-collinearity between the independent variable x j and the rest of the independent variables, and the multi-collinearity may affect the least square estimates excessively.

Conclusion
In this paper, I use various analytical methods to evaluate different kinds of factors that impact on the ecological environment. Firstly, I gather data of forest coverage rate, water conservation, PM2.5, oxygen release and carbon dioxide absorption of Saihanba in 60 years from 1962 to 2021 and build the EWM-TOPSIS model based on the data. From the model, it is obvious to figure out that water conservation is the indicator with the largest weight, and the restoration of Saihanba hugely influence on the improvement of the environment. Then in the Multivariable Linear model, I apply 4 indicators of Beijing's ability of resisting sandstorms as dependent variables. From the results, it has been concluded that Saihanba plays a vital part in resisting sandstorms in Beijing.
1. When evaluating sandstorm resistance (Model II) and extending carbon emissions from one region to the whole country (Model III), indicators with great impact on ecological environment in the first model are used, which unify measurement standards and facilitate horizontal comparison. In addition, when establishing nature reserves, the analysis is carried out by combining objective indexes and subjective policies, so that the analysis is closer to reality under the premise of science. Pearson correlation analysis was carried out before regression analysis when evaluating the relationship between the anti-sand ability of Beijing and the ecological environment of Saihanba, which can clearly judge the linear relationship between the next two indicators excluding the influence of other variables and supplement the regression analysis. However, in the establishment of regression relationship, four indicators representing sand resistance capacity were used as dependent variables in the four regression models, and then the estimated values of these four indicators were integrated to evaluate sand resistance capacity. The default weight of the four indicators was the same, which was not rigorous enough. And some regression models fail to pass the significance test, indicating that they may be nonlinear regression or have a problem of small sample size. Although we prefer the calculation to be as accurate as possible, owing to the long number of decimal places in the data we used, we had to round off some of the decimals, which may introduce errors into the final result.