| 摘要: |
| 基于2000和2007年TM影像和2015年OLI影像，借助于地学信息图谱理论，分析东北山地丘陵地带农林交错区实施退耕还林15年间的土地利用变化，在此基础上利用CA-Markov模型对2023年研究区的土地利用类型进行预测和模拟。结果表明：1）不同研究时期内，研究区林地面积比例均过65%，耕地面积均占总面积20%以上；2）稳定型图谱所占比例最大，散布于整个研究区，其中"林地-林地-林地"是其最大演变类型，"旱地-旱地-旱地"次之；"旱地-林地-林地"是前期变化型的最大变化类型，"草地-草地-林地"是后期变化型的主要变化模式，"林地-旱地-林地"是反复变化型面积最大的变化模式，持续变化型所占面积最小；3）运用CA-Markov预测的2015年土地利用类型，其精度为0.94，模型可用，由此模拟了2023年土地利用数量和空间分布，结果显示土地利用结构仍以林地为主，林地-旱地是其最主要的转移轨迹。平衡生态环境保护、建设用地需求与耕地保护三者之间的矛盾是研究区土地利用的焦点。 |
| 关键词: CA-Markov模型 农林交错区 地学信息图谱 土地利用转移矩阵 预测 |
| DOI：10.11841/j.issn.1007-4333.2017.09.18 |
| 投稿时间：2016-09-18 |
| 基金项目:黑龙江省教育厅科学技术研究项目（12541037）；东北农业大学青年才俊项目（14QC30） |
|
| Dynamic changes and prediction of land use in farm-forest ecotone based on CA-Markov model |
|
GAO Fengjie, HAN Wenwen, SHAN Peiming, JU Tienan, WU Xiao, ZHANG Bingfang, HUANG Shanlin
|
| (College of Resources and Environment, Northeast Agricultural University, Harbin 150030, China) |
| Abstract: |
| By taking Landsat TM images in 2000a and 2007a and Landsat OLI images in 2015a as main analyzing data,the LUCC under the "Grain for Green Project" of a farm-forest ecotone located in a mountain foothill area in Northeast China,including changes in number by Land-Use Transfer Matrix and changes in space by Geo-information Tupu,was discussed.The land use pattern in 2023a was forecasted by CA-Markov model.The results showed that:1) The proportion of forest in the study area was all over 65%,and the cultivated land accounted for more than 20% of the total area;2) The proportion of stable change atlas was the largest and scattered throughout the whole study area,among which the forestland→forestland→forestland pattern was the first evolution type followed by the dry land→dry land→dry land type;dry land→forest→forest was the main type of early change atlas,grass→grass→forest was the main type of late change atlas,and forest→dry land→forest was the main type of repeated change atlas;continuous change atlas was the smallest change type in the area.3) By comparing the forecasting land-use map by CA-Markov with the present land-use map by remote sensing interpretation in 2015a,the Kappa coefficient was 0.94,indicating that the CA-Markov model was applicable in the study area.So the CA-Markov model was applied to forecast the land-use in 2023,showing that forest was still the main land-use type and forest→dry land was the change trace.In conclusion,balance between the ecological environment protection,construction land supply and cultivated land conservation was the focus of land use study. |
| Key words: CA-Markov model farm-forest ecotone geo-information tupu land-Use transfer matrix forecast |
