| The emissions of carbon dioxide and other greenhouse gases has aroused great attention at home and abroad.Fujian province is one of the important coastal provinces.It is of great significance to research on the characteristics and influencing factors of carbon emissions of energy consumption on different scales in Fujian Province.At present,most studies on carbon emissions of energy consumption are based on regional energy consumption statistics at home and abroad.However,raster-level studies can not be realized only by using regional statistics data.Nighttime light data can be used to simulate raster-level carbon emissions data of energy consumption.Therefore,this paper introduces night light data,and matchs nighttime light data with energy statistics data to study the characteristics and influencing factors of carbon emissions of energy consumption on provincial,municipal,county and raster levels from 2012to 2019 in Fujian province.Firstly,the basic research data are preprocessed.Secondly,the carbon emissions of energy consumption are rasterized based on nighttime light data.Then,the characteristics of carbon emissions of energy consumption are analyzed based on carbon emissions data and the results of rasterization.Finally,based on geographically and convolutional neural network temporally weighted regression,the influencing factors of carbon emissions of energy consumption are studied on the county level.The specific research contents and results are as follows:(1)Nighttime lighting data,energy consumption data and impact factors data from 2012to 2019 were preprocessed.This paper reduces the saturation effect of nighttime light image based on the POI and vegetation modified nighttime light urban index.Based on the statistical data of energy consumption,the carbon emissions of energy consumption of all regions of Fujian Province are calculated by IPCC.This paper determines the collinearity between the impact factors using variance factor,characteristic value indicators,etc,and the results show that GDP,urbanization rate,the proportion of secondary industry,the proportion of the tertiary industry,the population does not exist collinearity,and they can be used for regression model.(2)The carbon emissions data of energy consumption in Fujian province from 2012 to 2019are rasterized based on desaturated night light data.Firstly,a simulation model of carbon emissions of energy consumption was constructed,and the goodness of fit of the simulation model is greater than 0.8.Then,carbon emissions of energy consumption are rasterized based on the simulation model,and the statistical results of rasterization of cities and countries are counted.It is found that the goodness of fit of linear relation model between the simulated value and the real value at the county level is greater than 0.8.and the average relative error between the simulated value and the real value at the city level is 17%and 11.4%,respectively.This shows that the model simulation effect is good;Finally,the error correction model is established and the final rasterization result is obtained.(3)Based on the carbon emissions data of energy consumption and rasterization results from 2012 to 2019,the characteristics of carbon emissions of energy consumption in Fujian province are analyzed by the bar graph,spatial distribution map,slope tendency value,Moran’s I index and local _iG index,and gravity center migration model.The results show that the carbon emissions of energy consumption in fujian province decrease from 2.27 million tons in2012 to 2.17 million tons in 2019.And the intensity of carbon emissions of energy consumption decreased from 1.15 tons/thousand yuan in 2012 to 0.51 tons/thousand yuan in 2019.From2012 to 2019,the intensity of carbon emissions of energy consumption of nine cities in Fujian province decreased year by year.In 2012 and 2019,the high value of carbon emissions of energy consumption is distributed in the southeast of Fujian province,while the low value is distributed in the central and northwest of Fujian province.The change type of carbon emissions of energy consumption in most counties of Fujian province is basically unchanged,accounting for 87.34%of the area.In 2012,2015,2017 and 2019,the carbon emissions of energy consumption showed more low-low clustering in space,and the distribution region remained unchanged,mainly in the northwest of Fujian province.The high-high clustering is mainly distributed in Yongchun,Nan’an,Fengze and Shishi.The high-low clustering only appeared in Fuan in 2012.The low-high clustering is mainly distributed in Lisheng,Jinmen and Pingtan.The carbon emissions of energy consumption in Fujian province mainly transition from low value to high value along the southeast direction.From 2012 to 2019,the centre of gravity of carbon emissions of energy consumption is distributed in Yongchun,and moves 6.43km to the southeast.(4)The impact of influencing factors on carbon emissions of energy consumption varies with time and space.namely,there is spatio-temporal heterogeneity.In order to solve the problem that the traditional geographically and temporally weighted regression model is difficult to fully detect the spatio-temporal heterogeneity,this paper introduces the convolutional neural network and builds the geographically and convolutional neural network temporally weighted regression model.By designing a spatio-temporal weighted convolutional neural network and optimizing it using dropout algorithm and batch normalization algorithm,the model achieves an accurate solution to the spatio-temporal weight matrix,and then fully explores the spatio-temporal variation of the impact of factors on carbon emissions of energy consumption.The result shows that the performance evaluation indexes RSS,AICc and R~2 of the built model are better than those of the geographically and temporally weighted regression model.This shows that the built model in this paper has higher precision,better performance and stronger stability.The results of regression analysis show that the promoting effect of population factors on carbon emissions of energy consumption is weak in the northwest and strong in the southeast,and the range of intensity increases with time.The effect of GDP on carbon emissions of energy consumption is inhibiting in most counties of Fujian Province.The influence of urbanization rate on the carbon emissions of energy consumption in coastal areas of Fujian Province is promoting in inland areas and inhibiting in coastal areas,and the range of inhibition increases with time.The role of the proportion of secondary industry in promoting carbon emissions of energy consumption is weak in the northwest and strong in the southeast,but the range of intensity decreases with time.The proportion of tertiary industry has a restraining effect on the carbon emissions of energy consumption in most areas of Fujian Province,and a promoting effect in a few areas of northwest,and the range of promote decreases with time. |
PDF Full Text Request