中南半岛山区刀耕火种农业时空格局变化及其植被恢复过程研究

Swidden agriculture is an important and disputable land use practice found across Mainland Southeast Asia. In the last two decades, swidden farming has rapidly transformed or is in the process of transforming into other land use systems in many parts of Mainland Southeast Asia. While these changes have been investigated in a number of localities, it remains sparse, sporadic, and scattered throughout the whole morphologic landscape unit of Mainland Southeast Asia. So far, no reliable region-wide map data and classification results of swidden farming have been reported. Scientific community have called on to give top priority to reinforcing accurate and updated assessment of the scale, extent, and pace of the transformation of swidden agriculture at the regional and global level which is critically significant for understanding global climate change. In this proposal, we take the opportunity of free access to Landsat historical imagery (near 40 years), take Montane Mainland Southeast Asia (MMSEA), 300 m elevation and above, as the study area, using the technological tools of Remote Sensing (RS), Global Positioning System (GPS), and Geographic Information System (GIS) to develop the remote sensing detection methods for swidden agriculture based on the satellite-derived parameters, to obtain the spatial distribution information of swidden agriculture from 1990 to 2010, to compile thematic maps at 1: 100, 000 scale for the MMSEA and 1: 50, 000 for typical areas, and to compare the corresponding spatial pattern variation and the rule of territorial differentiation of swidden cultivation during various periods. Then, establish the distribution prediction model for swidden farming by quantifying its relationships with biophysical and socio-economic factors, so as to analyze the suitable distribution extent of this traditional farming practice and make an assessment of land use intensity of swidden cultivation during the past two decades. Finally, assess the impacts of swidden agriculture on the forest landscape pattern across MMSEA with the theories of Landscape Ecology and landscape metrics derived based on the FRAGSTATS software (Version 4.1), and to construct the dynamic curve model of vegetation-water content indices (like, NDVI and LSWI) derived from the interannual time-series Landsat TM/ETM+ data to analyzed and compare the vegetation recovery process of swidden agriculture in different regions and of various types in MMSEA.

刀耕火种农业是中南半岛普遍存在且备受争议的土地利用形式，近20年来已经或正在向其他土地利用类型快速转变。既有相关研究普遍集中在地方尺度，零星而分散，对中南半岛山区刀耕火种农业的时空格局变化与植被恢复过程缺乏清晰认识。鉴于此，项目以Landsat影像共享为契机，基于3S技术建立刀耕火种农业遥感信息识别与提取方法，编制中南半岛山区1990-2000-2010年1：10万和典型地区1：5万刀耕火种农业数字地图；依此，对比研究不同时期刀耕火种农业空间格局变化与地域分布规律，定量揭示刀耕火种农业与自然因素、人类活动的空间依存关系，模拟分析刀耕火种农业适宜范围及其强度特征；最后，利用景观指数评估刀耕火种农业对森林景观的影响，并基于年际时序影像构建植被-水分指数动态曲线，对比研究不同地区/类型刀耕火种农业用地的植被恢复过程及其差异性，以期为中南半岛山区刀耕火种农业可持续利用提供科学依据与决策支持。

刀耕火种农业（Swidden agriculture）是中南半岛乃至整个热带山区广泛分布且备受争议的生计型土地利用类型。自2007年联合国“减少发展中国家毁林和森林退化（REDD）”行动实施以来，针对该传统农业的动态变化及其在森林砍伐与林地退化、水土流失、碳排放与生物多样性等的影响已引起全球科学家关注。然而，目前有关全球刀耕火种农业一些根本性问题，以中南半岛为例，目前究竟有多少地方还存在刀耕火种农业？近30年来每年约有多少新开辟土地面积？每个国家刀耕火种从业人员多少？依旧缺乏必要认识，数据支持严重不足。本项目以中南半岛山区为研究区，以全球Landsat卫星历史数据（TM、ETM+与OLI）共享为契机，基于“3S”技术，围绕刀耕火种农业的遥感信息识别与提取方法研究、刀耕火种农业时空格局变化与地域分布规律研究、刀耕火种农业对森林景观格局的影响及其植被恢复过程研究三大研究内容，完成了7个方面的研究工作。第一、系统综述了20世纪80年代以来东南亚刀耕火种农业的研究进展。第二、基于中南半岛主要土地覆被类型相关植被指数（NDVI、NDMI、MNDWI与NBR）交互验证了ETM+与OLI影像数据具有很好的兼容性。第三、中南半岛1986-2015年所有可获取的TM、ETM+与OLI影像云量覆盖分析显示旱季影像获取量完全能支持刀耕火种农业动态监测。第四、基于TM与ETM+ 8-16天影像显示，位于中南半岛中心位置的老挝北部，旱季刀耕火种农业通常于每年二月中下旬开始砍伐森林，三月风干曝晒，四月集中焚烧。第五、基于OLI首次监测了中南半岛山区2014/2015年最新刀耕火种农业空间格局，新开辟火烧地块与不同休耕期地块面积分别约为3.28×104 km2和16.72×104 km2。第六、基于TM、ETM+与OLI首次遥感重建了中南半岛山区1988-2016年历年新开辟的刀耕火种农业地块数据，多年平均面积为6.08×104 km2，总体呈现减少趋势。第七，基于1988-2016年历年新火烧地块研究刀耕火种农业休耕期变化与植被恢复过程。本项目不仅对揭示区内刀耕火种农业分布特征和预测刀耕火种变化趋势，而且对监测关键地区、典型区域植被恢复过程都具有重要的科学价值和现实意义。