中国农业科学 ›› 2018, Vol. 51 ›› Issue (22): 4316-4327.doi: 10.3864/j.issn.0578-1752.2018.22.010

• 土壤肥料·节水灌溉·农业生态环境 • 上一篇    下一篇

基于模型集成的中国耕地非农化影响因素及其时空特征研究

崔许锋(),马云梦,张光宏()   

  1. 中南财经政法大学工商管理学院,武汉 430073
  • 收稿日期:2018-06-14 接受日期:2018-10-11 出版日期:2018-11-16 发布日期:2018-11-16
  • 基金资助:
    国家社会科学基金(16BGL154)

The Factors of Farmland Conversion and Its Temporal and Spatial Characteristics: An Integrated Model

CUI XuFeng(),MA YunMeng,ZHANG GuangHong()   

  1. School of Business Administration, Zhongnan University of Economics and Law, Wuhan 430073
  • Received:2018-06-14 Accepted:2018-10-11 Online:2018-11-16 Published:2018-11-16

摘要:

【目的】揭示耕地非农化影响因素作用的时空特征,为耕地资源保护和利用政策制定提供决策支撑。【方法】研究采用2006—2015年耕地非农化的面板数据,通过构建“一般回归模型-面板模型-地理加权回归模型-时空加权回归模型”模型集成(简称OPGT),对耕地非农化影响因素进行计量分析。【结果】一般回归模型、地理加权回归模型(GWR)和时空加权回归模型(GTWR)估计结果显示,城镇人口增长、固定资产投资、经济发展水平、耕地资源禀赋和产业结构变量均通过显著性检验;耕地非农化莫兰指数(Moran’s I)为0.740,并且通过1%水平上显著性检验,表明耕地非农化具有显著的空间正相关性;采用一般回归模型、GWR、GTWR模型估计,方程拟合优度分别为0.689、0.785、0.858,加入时空权重信息的GWR和GTWR模型方程解释能力有显著提升;GWR和GTWR模型方程结果显示,耕地非农化影响因素弹性系数存在时空非平稳特征;空间分析显示,城镇人口增长和耕地资源禀赋对耕地非农化影响在经向上呈现出由西向东递减的状态,在纬向上呈现出倒“U”型状态,固定资产投资与经济发展水平对耕地非农化的影响程度在经向上呈现出由西向东递增的特征,在纬向上呈现出“U”型特征,产业结构对耕地非农化的影响程度在经向上由西向东递增,在纬向上由北向南递减;时序分析显示,城镇人口增长、固定资产与经济发展水平投资系数呈现减小的趋势,耕地资源禀赋系数有所增大,产业结构系数在部分省域有所降低。【结论】(1)OPGT是一个有机整体,各部分相互检验、互为补充,可以更加细致的刻画因素的时空作用;(2)耕地非农化因素总体作用强度方面,弹性系数最大的是产业结构,其次为经济发展水平、固定资产投资和耕地资源禀赋,最小为城镇人口增长;(3)空间特征方面,城镇人口增长和耕地资源禀赋总体呈现出由西向东递减的趋势,而固定资产投资、经济发展水平和产业结构呈现出由西向东递增的趋势;(4)时序演变特征方面,城镇人口增长、固定资产与经济发展水平投资对耕地非农化的影响作用呈现下降趋势,耕地资源禀赋与耕地非农化关联性趋于增强,产业结构的影响虽在部分省域有所降低,但其整体影响程度仍然相对较高。

关键词: 耕地非农化, 模型集成, 影响因素, 时空特征

Abstract:

【Objective】The purpose of this paper was to reveal the temporal and spatial characteristics of the factors affecting farmland conversion, and to provide decision-making information support for policy making for the protection and utilization of farmland.【Method】Based on the panel data of farmland conversion from 2006 to 2015, an integrated model of "ordinary regression model-panel model-geographically weighted regression-geographically and temporally weighted regression" (abbreviately named “OPGT” ) was established to analyze the factors of farmland conversion【Result】The ordinary regression model, GWR and GTWR model results showed that urban population growth, fixed asset investment, economy, arable and industrial structure variables all passed the significance test; Moran's I of farmland conversion was 0.740, and passed significance test at the 1% level. The results showed that there was a significant positive spatial correlation of farmland conversion. Ordinary regression model, GWR and GTWR models were used to estimate the equations, and the fit goodness of the equations were 0.689, 0.785 and 0.858, respectively. The interpretation ability of GWR and GTWR models was improved significantly under the condition of adding spatio-temporal weight information. The results of GWR and GTWR models showed that the elastic coefficients of factors were spatio-temporal non-stationary. The results of spatial analysis showed that the influence of urban population growth and farmland resource endowment on farmland conversion was declining from west to east in longitude direction, and reversed U-shaped curve in latitude direction. The influence of fixed assets investment and level of economic development was increasing from west to east in longitude direction, and U-shaped curve in latitude direction. The influence of industrial structure was increasing from west to east in longitude direction, and declining from north to south in latitude direction. From the perspective of temporal evolution, the coefficients of urban population growth, fixed assets investment and level of economic development had a downward trend, while coefficients of farmland resource endowment tended to increase. Coefficients of industrial structure had been reduced in some provinces.【Conclusion】(1) OPGT was an organic whole, each part was mutually tested and complementary, which could describe the spatio-temporal effect of factors in more detail. (2) In terms of the overall action intensity of the factors, the largest elastic coefficient was industrial structure, followed by level of economic development, fixed asset investment and farmland resource endowment, and the smallest was urban population growth. (3) In terms of the spatial characteristics of factor intensities, the influence of urban population growth and farmland resource endowment on farmland conversion was declining from Western China to Eastern China, while fixed assets investment, level of economic development and industrial structure increasing. (4) From the perspective of temporal evolution, the influence of urban population growth, fixed assets investment and level of economic development on farmland conversion had a downward trend. The relationship between farmland resource endowment and farmland conversion tended to strengthen. Although the influence of industrial structure had been reduced in some provinces, its degree of overall influence was still relatively high.

Key words: farmland conversion, integrated model, factors, temporal and spatial characteristics