北半球中高纬植被聚集指数遥感产品长时间序列变化特征的案例研究  

Long-term temporal characteristics of vegetation clumping index remote sensing products at mid/high latitudes in the Northern Hemisphere-a case study

作　　者：高歌 焦子锑 王晨霞 尹思阳 郭静 佟一冬 李志龙 谭哲友 陈偲喆 GAO Ge;JIAO Ziti;WANG Chenxia;YIN Siyang;GUO Jing;TONG Yidong;LI Zhilong;TAN Zheyou;CHEN Sizhe(Institute of Remote Sensing Science and Engineering,Faculty of Geographical Science,Beijing Normal University,Beijing,China;State Key Laboratory of Remote Sensing Science,Beijing Normal University,Beijing,China)

机构地区：[1]北京师范大学地理科学学部遥感科学与工程研究院,北京师范大学遥感科学国家重点实验室,北京

出　　处：《北京师范大学学报（自然科学版）》2024年第3期386-396,共11页Journal of Beijing Normal University(Natural Science)

基　　金：国家自然科学基金资助项目(42090013,41971288)。

摘　　要：植被聚集指数(clumping index,CI(Ω))是表征植被冠层聚集程度的重要结构参数,由于其定量化研究起步较晚,导致对CI季相变化特征的研究不充分,结论争议较大.为此,本文基于长时间序列的MODIS CI产品,从北半球中高纬度植被物候特征敏感区,在13个国际地圈-生物圈计划(IGBP)类型中,优选了84个高质量的代表性像元,开展典型像元CI季相变化特征的案例研究.以归一化植被指数(normalized difference vegetation index,NDVI)作为对比,提出改进的动态阈值法,结合离散Fourier变换方法,分别提取不同地类的生长季开始时间(start of season,SOS)及生长季结束时间(end of season,EOS),最终建立北半球中高纬度各地类生长季与休眠期的经验Ω.结果表明:CI具有较为明显的物候变化规律及季节变化特征,甚至能够识别出耕地的一年两熟迹象,但相对于NDVI相对稳定的季相变化特征,大部分地类的CI表现出较大的变化和不确定性,其中,SOS和EOS多分别在第100和第300天左右变化,生长季则多维持在200d左右;提取物候特征参数的最佳阈值随提取时期、地物类别的变化而变化,其中提取SOS和EOS的最佳阈值多集中在40%~80%和80%~90%;经验Ω呈现出针叶林的聚集效应最强,耕地的聚集效应最弱的特征.本研究对于揭示不同地类CI季相特征及相关应用研究提供了有用的证据和参考.Clumping index(CI)is a vital structural parameter characterizing non-random spatial distribution of vegetation canopy leaves.Quantitative research on CI started relatively late compared to other vegetation structural parameters.This is especially true for the study of seasonal variations in CI,resulting in significant controversies.In this work,we conducted a case study on seasonal characteristics of typical CI pixels in mid/high latitudes vegetation phenology-sensitive regions in the Northern Hemisphere,using recently developed long-term MODIS CI products developed by us.We selected 84 high-quality representative pixels from 13 land cover types designated by the International Geosphere-Biosphere Program(IGBP).We applied an enhanced dynamic threshold method in conjunction with discrete Fourier transform to extract start of season(SOS)and end of season(EOS)for various land cover types,using normalized difference vegetation index(NDVI)as reference.We established empiricalΩvalues for both growing season and dormant period across various landcover types.CI exhibits distinct phenological and seasonal variations,to even detect biannual cropping signs in cropland.In comparison with the relatively stable seasonal variations of NDVI,CI exhibits somewhat large variability and uncertainty in most land cover types.SOS and EOS vary within approximately 100 and 300 days,respectively;the growing season typically spans around 200 days.The optimal threshold values for extracting phenological parameters vary dependent on extraction period and landcover type.The thresholds for SOS vary within the range from 40%-80%,but for EOS,from 80%-90%.EmpiricalΩvalues show the strongest clumping effect in coniferous forests and the weakest in croplands.

关 键 词：聚集指数 时间序列 植被物候 MODIS 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]