生物多样性 ›› 2015, Vol. 23 ›› Issue (5): 575-582.DOI: 10.17520/biods.2015122

所属专题: 中国自然观察专题

• • 上一篇    下一篇

基于Global Forest Watch观察2000-2013年间中国森林变化

王昊1,2,*(), 吕植1,2, 顾垒1,2, 闻丞1,2   

  1. 1 北京大学生命科学学院自然保护与社会发展研究中心, 北京 100871
    2 山水自然保护中心, 北京 100871
  • 收稿日期:2015-05-08 接受日期:2015-08-17 出版日期:2015-09-20 发布日期:2015-10-12
  • 通讯作者: 王昊

Observations of China’s forest change (2000-2013) based on Global Forest Watch dataset

Hao Wang1,2,*(), Zhi Lü1,2, Lei Gu1,2, Cheng Wen1,2   

  1. 1 Center for Nature and Society, School of Life Sciences, Peking University, Beijing 100871
    2 Shanshui Conservation Center, Beijing 100871
  • Received:2015-05-08 Accepted:2015-08-17 Online:2015-09-20 Published:2015-10-12
  • Contact: Wang Hao

摘要:

了解森林的变化, 是评价生物多样性变化趋势和制订有效保护行动的基础, 也是公共政策及广大公众普遍关心的议题, 然而国内一直都没有可供空间分析的森林分布及变化的公开数据。Hansen等于2013年在Global Forest Watch网站上发布了基于Landsat数据的30 m分辨率的全球森林变化数据集, 第一次提供给研究者一套高分辨率的公开数据。我们解读了2000-2013年的数据, 并与已有的国家森林清查、ChinaCover、GlobeLand30等数据集进行比较。结果表明, Global Forest Watch数据质量可靠, 其与GlobeLand30数据集的森林分布有较高的空间重合度, 后者森林类别的生产者精度为87%, 用户精度为89%。在大于20%的树木覆盖度下, 估算出2000年森林面积为1,780,472 km2, 与同期国家森林清查数据1,749,092 km2相近, 解读结果表明2000-2010年间森林减少面积在37,551-42,031 km2之间, 而同期其他几套数据显示森林面积增量为2,370-433,810 km2。森林面积年减少量最大值出现在2008年, 以该年为拐点, 森林每年减少面积的变化趋势在此之前为显著增加, 在此之后为显著降低, 林地变化最活跃的区域在广西、广东、福建、江西和云南等南方省份。到2013年底, 占保护区总面积约60%以上的407个国家级保护区仅覆盖了森林面积的5.03%, 对森林保护的面积覆盖有限, 保护区对其内部的森林保护有成效, 区内森林面积净减1.39%, 低于全国平均水平(3.46%), 相当于少毁林1,856 km2, 但仍有1,200 km2的净减。

关键词: Global Forest Watch, 数据质量, 森林变化, 保护区, 中国

Abstract:

Knowing the status and changes in forests is essential for evaluating biodiversity dynamics and making effective conservation action plans. It is also essential knowledge for public awareness and policy. However, before Hansen and his colleagues published the Global Forest Watch (GFW) data online in 2013, there were no datasets of forest distribution and changes publicly available in China. GFW data was produced using Landsat satellite images, grants free access to the general public to study forest changes at 30 m resolutions, and spans a period from 2000 to 2013. We interpreted this dataset and compared results with other datasets, including the National Forest Inventory, ChinaCover and GlobeLand30. Results show that GFW dataset has reliable accuracy, as its forest distribution highly overlaps with GlobeLand30, which has a producer accuracy of 87%, and a user accuracy of 89% within the forest category. With similar forest definitions of tree cover >20%, the total forest area is 1,780,472 km2 in the year 2000, similar to the National Forest Inventory’s forest area of 1,749,092 km2. While other datasets show forest increases between 2,370-433,810 km2 during 2000-2010, GFW identified a net forest loss of 37,551-42,031 km2 in China. 2008 is a turning point of forest loss, as the area of annual forest loss has a rising trend before, and a declining trend afterwards. The southern provinces of Guangxi, Guangdong, Fujian, Jiangxi and Yunnan have the most active forest changes. Forest coverage from nature reserves is insufficient, as by the end of 2013, 407 national level nature reserves, occupying more than 60% of China’s total nature reserve area, covered only 5.03% of total forest area. However, nature reserves have positive outcomes, with reduced forest loss (1.39%), compared to the national average rate (3.46%). This is equivalent to reducing deforestation by 1,856 km2, but there is still an overall net loss of 1,200 km2 inside nature reserves.

Key words: Global Forest Watch, data quality, forest change, nature reserve, China