垂直地带性植被物候遥感提取尺度误差估算方法研究

As the most sensitive area of global climate change, study of vertical zonal vegetation phenology in High Mountains can better evaluate and understand the process of regional and global climate change. However, scale error which caused by low spatial resolution pixel heterogeneity, seriously affected the precision of remote sensing vegetation phenology to climate change.In order to clarify the spatial scale error of vegetation phenology extraction by remote sensing, the research object focus on the remote sensing extraction vegetation phenology of vertical landscape zones in Changbai Mountain. We use the multispectral camera to obtain the physiological phenology and remote sensing phenology of the dominant plant in the sample area, and the relationship between physiological phenology and remote sensing phenology are analyzed. Then their spatial scales were up-scaling. Combining with vegetation phenology of satellite remote sensing data acquisition, we discussed the spatial scale error on different spatial resolution remote sensing data in study areas. Finally, we use the spatial interpolation method to obtain the distribution characteristics of landscape scale error for different vegetation phenology, and build a scale error algorithm which suit in the vertical zonal vegetation phenology extraction. This study is significance for scale error of precise correction in the vertical zonal vegetation phenology remote sensing extraction and provides scientific foundation for precise reflection of the spatial-temporal change on the vegetation phenology to the global climate change.

高山地区作为全球气候变化最为敏感区域之一，研究其垂直地带性植被物候可以更好地评估与理解区域乃至全球气候变化过程。然而，由于低空间分辨率像元异质性带来的尺度误差，严重影响遥感提取的植被物候对气候变化响应的精度。本项目针对高山地区垂直景观带植被物候卫星遥感提取尺度误差问题，以垂直地带性典型区—长白山垂直景观带的植被物候为研究对象，采用多光谱相机采样方法，研究物种尺度上植被生理物候和遥感物候之间的内在关系；然后通过升尺度转换，结合多种分辨率遥感数据获取的植被物候，分析样区尺度上不同空间分辨率植被物候遥感提取的尺度误差及变化规律；在此基础上，通过建立空间内插模型，获得区域尺度上植被物候遥感提取尺度误差的分布特征，构建一套适合于高山地区垂直地带性植被物候卫星遥感提取尺度误差估算模式，为植被物候卫星遥感提取尺度误差纠正奠定理论与方法基础，进而为高山地区植被物候对全球气候变化响应研究提供科学依据。

高山地区作为全球气候变化最为敏感区域之一，研究其垂直地带性植被物候可以更好地评估与理解区域乃至全球气候变化过程。然而，由于低空间分辨率像元异质性带来的尺度误差，严重影响遥感提取的植被物候对气候变化响应的精度。本项目针对高山地区垂直景观带植被物 候卫星遥感提取尺度误差问题，以垂直地带性典型区—长白山垂直景观带的植被物候为研究对象，研究物种尺度上植被生理物候和遥感物候之间的内在关系，长白山植被物候空间尺度误差变化模式，以及大长白山地区植被物候变化特征及对气候变化响应，结果如下：1)基于频率域HANTS滤波分析对拟合曲线平滑具有更大优势以及双logistic模型在提取物候精度更高；2）阔叶林展叶期对应NDVI曲线阈值为0.37，针叶林展叶期对应NDVI阈值为0.41；3）植被生长期空间尺度误差随着分辨率增加呈现正相关趋势，相关系数为0.81，与海拔高度无相关关系，且升尺度过程中面积占优法较样条插值法尺度误差相对较小；4）1999-2019年长白山地区植被物候都表现为生长期提前和枯黄期延迟的变化趋势，≥10℃对植被物候决定了植被物候期的变化。以上研究为长白山区植被物候遥感提取尺度误差纠正奠定方法基础，同时也揭示了近20年来长白山植被物候变化特征以及对气候的响应。