热带地理 ›› 2020, Vol. 40 ›› Issue (5): 881-892.doi: 10.13284/j.cnki.rddl.003268

所属专题: 红树林研究

• 论文 • 上一篇    下一篇

基于Google Earth Engine19862018年广东红树林年际变化遥感分析

王子予1(), 刘凯1,2(), 彭力恒1, 曹晶晶1,2, 孙映雪1, 钱雨昕1, 史舒悦1   

  1. 1.中山大学 地理科学与规划学院//广东省公共安全与灾害工程技术研究中心//广东省城市化与地理环境空间模拟重点实验室,广州 510275
    2.南方海洋科学与工程广东省实验室(珠海),广东 珠海 519000
  • 收稿日期:2019-12-25 修回日期:2020-07-15 出版日期:2020-09-28 发布日期:2020-10-10
  • 通讯作者: 刘凯 E-mail:wangzy38@mail2.sysu.edu.cn;liuk6@mail.sysu.edu.cn
  • 作者简介:王子予(1998-),女,吉林四平人,硕士研究生,主要从事红树林遥感,(E-mail)wangzy38@mail2.sysu.edu.cn
  • 基金资助:
    国家重点研发计划项目(2018YFB0505500);广东省自然科学基金项目(2016A030313261)

Analysis of Mangrove Annual Changes in Guangdong Province during 19862018 Based on Google Earth Engine

Ziyu Wang1(), Kai Liu1,2(), Liheng Peng1, Jingjing Cao1,2, Yingxue Sun1, Yuxin Qian1, Shuyue Shi1   

  1. 1.School of Geography and Planning, Sun Yat-sen University//Provincial Engineering Research Center for Public Security and Disaster// Guangdong Key Laboratory for Urbanization and GeoSimulation, Guangzhou 510275, China
    2.Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519000, China
  • Received:2019-12-25 Revised:2020-07-15 Online:2020-09-28 Published:2020-10-10
  • Contact: Kai Liu E-mail:wangzy38@mail2.sysu.edu.cn;liuk6@mail.sysu.edu.cn

摘要:

以广东沿海红树林为研究对象,结合谷歌地球引擎(GEE)云计算平台,以1986—2018年32期3 359景Landsat系列卫星遥感影像为数据源,采用随机森林(RF)方法提取1986—2018年广东省红树林面积,比较全省沿海城市红树林年际时空变化特征,并从景观斑块角度分析广东省红树林斑块演变特征。结果表明:1)1986—2018年红树林遥感分类总体精度均高于90%,广东省沿岸红树林面积总体呈先减少后增加的趋势,且其在2014年后变化幅度逐渐减小。2)从各沿海城市来看,红树林共分布在14个市内,其中湛江和阳江是红树林面积分布最大的2个城市;各市红树林面积变化可分为先减后增、波动增加和无明显变化3类。3)1986—2018年广东省红树林斑块数量总体呈减少趋势,但斑块平均面积(MPS)呈上升趋势,红树林破碎化程度减轻。获取年际红树林面积分布信息和空间结构变化趋势,可为红树林合理开发与保护提供数据和参考,服务于红树林生态恢复和精细化管理。

关键词: 红树林, 年际变化监测, Google Earth Engine, 随机森林, 广东省

Abstract:

Mangroves have critical ecological functions and social and economic value, and are an important target for protection in the coastal wetland ecosystem. By monitoring long-term dynamic changes in mangrove ecosystems, the overall change process can be systematically and accurately recorded, providing data support and a basis for decision-making on scientific protection and effective management of the ecosystem. This study focuses on mangrove forests in coastal areas of Guangdong Province. A map of the mangrove forest, from 1986 to 2018, was made using Landsat remote sensing image based on Google Earth Engine (GEE),which is a cloud computing platform. The Random Forest (RF) method was used to extract mangrove trees from 32 periods from 1986 to 2018, in Guangdong Province. The interannual variation in mangrove characteristics in coastal cities of Guangdong province were compared. In addition, the evolution characteristics of mangrove patches in Guangdong province were analyzed. The results show that 1) The computing capacity and massive data of the GEE cloud platform provide data support for analyzing the inter-annual evolution of mangroves in Guangdong province, which greatly improves the computing efficiency. From 1986 to 2018, the overall classification accuracy of mangrove remote sensing was higher than 90%, with high classification accuracy and reliable results. In general, the coastal mangrove area of Guangdong province first decreased and then increased, and the range of change gradually declined after 2014, remaining at about 11 000 hm2. Mangrove forests are unevenly distributed in the province and occur mainly in the west. 2) Concerning coastal cities, mangroves are distributed in 14 cities, among which Zhanjiang and Yangjiang have the largest mangrove area, which is about 70% of the mangrove area of Guangdong province. The mangrove area changes in each city fall under three categories: decreasing first and increasing later, increasing fluctuation, and no obvious change. 3) From 1986 to 2018, the overall number of patches in mangrove forests in Guangdong province showed a decreasing trend, but the average patch area (MPS) showedan increasing trend, and mangrove fragmentation was reduced. In 2018, the mangrove MPS was 4.11 hm2 in Guangdong province, and the total number of patches was 2 782. From 1986 to 2018, when the change trend of MPS in mangrove forests in Guangdong province was opposite to that of patch quantity, the changes of patches were mainly expansion and fragmentation. When MPS change trend was consistent with the change trend of plaque number, the increase and decrease in the change of plaque were dominant. Information on annual mangrove area distribution and structural changes can provide more detailed data and reference for the rational development and protection of mangroves and support ecological restoration and finely tuned mangrove management.

Key words: mangrove, inter-annual change monitoring, Google Earth Engine (GEE), Random Forest, Guangdong Province

中图分类号: 

  • TP79