地理研究 ›› 2016, Vol. 35 ›› Issue (10): 1935-1947.doi: 10.11821/dlyj201610012

• 研究论文 • 上一篇    下一篇

农用地整治对产能影响的特征预测及评估:方法与实证

范业婷1(), 金晓斌1,2(), 项晓敏1, 杨绪红1, 黄学锋1, 周寅康1,2   

  1. 1. 南京大学地理与海洋科学学院,南京 210023
    2. 南京大学自然资源研究中心,南京 210023
  • 收稿日期:2016-04-16 修回日期:2016-07-12 出版日期:2016-10-26 发布日期:2016-10-27
  • 作者简介:

    作者简介:范业婷(1990- ),女,安徽和县人,博士研究生,主要从事土地整治与土地利用规划研究。E-mail: dg1527012@smail.nju.edu.cn

  • 基金资助:
    国家科技支撑计划课题(2015BAD06B02)

Prediction and evaluation of characteristic of agricultural productivity change influenced by farmland consolidation: Method and case study

Yeting FAN1(), Xiaobin JIN1,2(), Xiaomin XIANG1, Xuhong YANG1, Xuefeng HUANG1, Yinkang ZHOU1,2   

  1. 1. School of Geographic and Oceanographic Sciences, Nanjing University, Nanjing 210023, China
    2. Natural Resources Research Center, Nanjing University, Nanjing 210023, China
  • Received:2016-04-16 Revised:2016-07-12 Online:2016-10-26 Published:2016-10-27

摘要:

受工程建设和农田生态系统恢复过程的影响,农用地整治后产能效益的发挥存在滞后性,采用实时遥感数据难以对近期产能变化进行有效估算。基于2001-2013年MODIS-NDVI数据,提取表征产能水平、产能波动、产能潜力和复种指数变化的四项特征参数,以2006-2010年验收的农用地整治项目为样本集,采用SVM算法,建立产能变化与潜在影响因素间的联系,对2011-2013年验收项目的产能影响特征进行预测和评估。研究表明:① 在确定已知样本集特征及其影响因素的基础上,对未知样本集各项特征基于SVM模型进行预测的拟合效果可达到全局最优;② “十二五”前期,农用地整治后实现产能水平提升、产能波动减小、产能潜力提高、复种指数增加的项目比例分别为88.18%、66.41%、81.55%和3.59%;③ 农用地整治对产能提升的影响具有显著的区域差异,长江中游平原、江南丘陵、南岭丘陵、粤西贵南等地区产能提升较为显著,松嫩三江平原和辽宁平原丘陵区仍有较大的提升空间。

关键词: 农用地整治, 产能变化, 区域差异, NDVI, SVM

Abstract:

Influenced by the disturbance of farmland consolidation engineering construction and recovery period of farmland ecosystem reestablishment, agricultural productivity improvement caused by farmland consolidation usually has a time lag. This makes recent change of agricultural productivity hard to be effectively assessed by real-time remote sensing data. Based on MODIS-NDVI data (MOD13Q1) from 2001 to 2013, this study proposes four characteristic parameters to demonstrate agricultural productivity change, i.e., the level of agricultural productivity (PL), the variation of agricultural productivity (PV), the potentiality of agricultural productivity (PP) and multiple cropping index (MI). Then, the study takes farmland consolidation projects labeled as completion acceptance from 2006 to 2010 as a sample set, and establishes intrinsic relationship between characteristics of agricultural productivity change and potential influencing factors using a Support Vector Machine (SVM) model. Finally, the rules concluded from SVM model are used to predict the characteristics of agricultural productivity change influenced by farmland consolidation implemented between 2011 and 2013. The results are as follows: (1) On the basis of defining characteristic of agricultural productivity change and potential influencing factors of a sample set, the fitness of feature parameters of the unkown sample set can achieve global optimum by SVM model. (2) During 2001-2013, the proportions of projects that PL, PV, PP and MI have increased after farmland consolidation are 88.18%, 66.41%, 81.55% and 3.59%, respectively. (3) The general characteristic of agricultural productivity improvement influenced by farmland consolidation has significant spatial variation. The regions are of significant improvement concentrated on the middle Yangtze River plain, the hilly region south of the Yangtze River, Nanling hilly region, western Guangdong and southern Guizhou. It is noteworthy that some regions still need to be further improved in agricultural productivity, such as Songnen-Sanjiang plain, plain and hilly region of Liaoning.

Key words: farmland consolidation, agricultural productivity change, spatial variation, NDVI, SVM