2000—2023年吉林省黑土核心保护区撂荒耕地遥感监测及水土保持效应
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:

国家自然科学基金青年基金项目“全生育期玉米叶面积指数反演的高光谱智能观测模式研究”(41701424); 吉林省科技发展计划项目创新发展战略研究“生态强省视域下吉林省耕地生态安全评价与调控模式研究”(20240701167FG)


Remote sensing monitoring and soil and water conservation effects of abandoned farmland in core black soil protection area of Jilin Province from 2000 to 2023
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    [目的] 分析吉林省黑土核心保护区2000—2023年撂荒耕地时空分布,阐明导致耕地撂荒现象的主要驱动因素,评估撂荒耕地的水土保持效应,为相同区域撂荒耕地遥感监测与水土保持效应研究提供理论依据。[方法] 依托谷歌地球引擎(Google Earth Engine,GEE)平台,利用Landsat长时序影像,结合多维度特征,耦合随机森林及LandTrendr变化检测方法,获取研究区撂荒耕地时空分布范围,并利用地理探测器对其主要驱动因素进行分析。基于AHP-EM-乘积归一化组合赋权法对撂荒耕地水土保持效应进行综合评估。[结果] ①基于RF-LT模型获取研究区撂荒耕地总体精度大于85%,kappa系数不低于0.8,分类精度较高,整体分类效果好。②撂荒耕地面积总体呈现先上升后波动下降的趋势,在2013年达到峰值(459 km2)。在空间分布上,整体呈现先增加再减少、最后局部增加的趋势,撂荒耕地分布较为零散。③夜间灯光强度、地块平均面积和高程对耕地撂荒的解释力较高,q值均超过0.4。坡度与地块平均面积交互作用最为突出,q值为0.997。④研究区5,10,15,20,20 a以上的撂荒耕地水土保持效应综合得分分别为0.388 6,0.477 3,0.801 4,0.604 4,0.443 8分,呈先增加后降低的变化趋势。[结论] 在一定年限内,适度撂荒对水土保持效应具有积极影响,撂荒地的生态多样性和土壤肥力得以提升。

    Abstract:

    [Objective] The spatiotemporal distribution of abandoned farmland in the core black soil protection area of Jilin Province from 2000 to 2023 was analyzed, the main driving factors of farmland abandonment were clarified, and the soil and water conservation effects of abandoned farmland are evaluated, in order to provide a theoretical basis for remote sensing monitoring of abandoned farmland and research on soil and water conservation effects in similar regions. [Methods] Based on the Google Earth Engine(GEE) platform and Landsat long-term time-series imagery, this study integrated multidimensional characteristics and combined random forest with LandTrendr change detection method to map the spatiotemporal distribution of abandoned farmland in the study area. Additionally, the key driving factors were analyzed using the geodetector. Finally, a comprehensive evaluation of the soil and water conservation effects of abandoned farmland was conducted using a combination weighting method based on AHP-EM-product normalization. [Results] ① The RF-LT model achieved an overall accuracy greater than 85% and a Kappa coefficient of at least 0.8, indicating high classification accuracy and good overall performance.② The area of abandoned farmland initially increased and then experienced a fluctuating decline after peaking at 459 km2 in 2013. Spatially, it first increased then decreased, and finally showed local increases, with relatively scattered distribution of abandoned farmland.③ Night-time light intensity, average parcel area, and elevation had high explanatory power for farmland abandonment, with q-values all exceeding 0.4. Furthermore, the interaction between slope and average parcel area was the most significant, with a q-value of 0.997.④ The comprehensive scores of soil and water conservation effects for abandoned farmland in the study area with abandonment durations of 5, 10, 15, 20, and over 20 years were 0.388 6, 0.477 3, 0.801 4, 0.604 4, and 0.443 8, respectively, showing an initial increase followed by a decrease. [Conclusion] The moderate abandonment within a certain period has a positive impact on soil and water conservation, specifically reflected in the improvement of ecological diversity and soil properties.

    参考文献
    相似文献
    引证文献
引用本文

冯馨慧,姜海玲,于海淋,孙玺皓,王静怡,付珈宁.2000—2023年吉林省黑土核心保护区撂荒耕地遥感监测及水土保持效应[J].水土保持通报,2026,46(2):225-236

复制
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2025-09-10
  • 最后修改日期:2025-10-31
  • 录用日期:
  • 在线发布日期: 2026-05-13
  • 出版日期: