江苏省生态足迹时间维度变化及其驱动因素分析——基于PLS方法对STIRPAT模型的修正

在分析江苏省生态足迹时间维变化规律的基础上,探讨江苏省生态足迹变化的重要驱动因素。1)1990-2007年,江苏省生态足迹以年均5.36%的速度从7 227.75万hm2升至13 817.76万hm2,而生态承载力以年均0.30%的速度从2 850.98万hm2降至2 703.58万hm2,生态足迹对区域生态赤字变化的贡献率达97.81%,成为江苏省生态赤字增加的主要原因;2)STIRPAT模型分析结果显示,人均GDP及其二次项、第一二产业产值占总产值比重及其二项式是江苏省生态足迹变化的重要驱动因素,总人口及城市化率指标却被排除在外,然而VIF值计算结果显示模型中各驱动因素间多重共线性明显;3)采用PLS方法修正STIRPAT模型,消除因素间的多重共线性问题,显示总人口、人均GDP及其二次项、第一二产业产值占总产值比重及其二次项、城市化率及其二次项都是江苏省生态足迹变化的重要驱动因素,且按VIP重要性排序为城市化率>人均GDP二次项>一二产业产值占总产值比重>人均GDP>城市化率二次项>总人口>一二产业产值占总产值比重二次项。两种方法中人均GDP二次项的系数均为正,表明江苏省生态足迹变化不存在环境EKC曲线的假说。该文进一步明确STIRPAT模型在分析环境压力的驱动因素中存在的缺陷,另一方面验证了基于PLS修正的STIRPAT模型的准确性与可行性。

A Study on Ecological Footprint Time Series and Its Drivers of Jiangsu Province:Using the STIRPAT Model and the PLS Method

Taking Jiangsu Province of China as an example,the change trend of ecological footprint(EF) time series was computed and analyzed and the important drivers of lnEF in Jiangsu Province were researched.The results showed that the EF in Jiangsu Province increased from 72.2775 million hm2 to 138.1776 million hm2 with the annual growth rate of 5.36%,but its ecological carrying capacity(EC) was rather low and was in a state of slow decline from 28.5098 million hm2 to 27.0358 million hm2 with the annual decrease rate of 0.30%,indicating that EF amounted to 97.81% to the contributive rate of ecological deficit(ED).Therefore,the major drivers of the EF′s change were analyzed.According to the simulations with STIRPAT model,the major drivers of Jiangsu′s lnEF were GDP per capita(lnA),quadratic term of GDP per capita((lnA)2),percent of GDP excluded in the service sector(lnN),and quadratic term of percent of GDP excluded in the service sector((lnN)2),but not human population(lnP),percent of urban population(lnM),quadratic term of percent of urban population((lnM)2).Nevertheless,these drivers themselves had strong collinearity by VIF value,which might produce some uncertain impact to the final results.In order to avoid the impact of collinearity,the method of partial least squares(PLS) was used to rectify the former STIRPAT model.The results showed that the major drivers of lnEF were lnP,lnA,(lnA)2,lnN,(lnN)2,lnM and(lnM)2.Compared with the results by the STIRPAT model,which showed that lnP,(lnN)2,lnM and(lnM)2 are the most dominant drivers and the effect of them on EF could almost be ignored.So,the results by PLS method were considered as more reasonable and acceptable.In addition,the results acquired by both methods showed that the curvilinear relationship between economic development and EF or the classical EKC hypothesis didn’t exist in Jiangsu Province.This paper,on one hand,further ensured the shortcoming of STIRPAT model used to analyze the major drivers on EF,on the other hand,confirmed the accuracy and feasibility of PLS method verified the STIRPAT model.