北疆地区草地TI-NDVI与NDVImax时空异质性评价

Spatio-temporal heterogeneity evaluation of grassland TI-NDVI and NDVImax in northern Xinjiang

选择草地类型丰富多样的北疆作为研究区。基于MODIS NDVI数据,获取时间累积归一化植被指数TINDVI和年最大NDVI(NDVImax),借助GIS空间分析技术、变异系数(CV)及Mann-Kendall非参数趋势检验等方法,对2000—2019年北疆地区草地覆盖动态变化进行了分析,并探究了TI-NDVI和NDVImax对草地时空异质性的表达能力的比较优势。结果表明:(1)用NDVImax和TI-NDVI表征的北疆草地呈现明显海拔分异。TI-NDVI总体随NDVImax的增大而增大,但NDVImax或TI-NDVI相同的区域,其TI-NDVI或NDVImax却存在较大差异。(2)2000—2019年北疆地区草地TI-NDVI和NDVImax总体呈显著增加趋势(P<0.01),但草地TI-NDVI变化的空间分异与NDVImax明显不同,全区17.55%的草地TI-NDVI变化趋势与NDVImax变化趋势相反。尤其阿尔泰山与伊犁河谷,高覆盖草地分布区的NDVImax与TI-NDVI均呈相反变化趋势。(3)在时间和空间维度上,北疆山区高覆盖草地TI-NDVI的CV均高于NDVImax。TI-NDVI对高覆盖草地的时空异质性反映更敏感,能在一定程度上削弱草地动态评价中NDVI光饱和缺陷的影响。

Grassland change is an important component of global change, which has attracted considerable attention. The temporal and spatial heterogeneity of grassland dynamics is the main concern in evaluating grassland dynamics. Northern Xinjiang, which is characterized with diverse grassland types, was selected as the research area. In this study, we calculated the time-integrated normalized vegetation index(TI-NDVI) and annual maximum NDVI(NDVImax) on the basis of the MODIS NDVI data. Using spatial analysis technology of GIS,mathematical statistical methods of coefficient of variation(CV), and Mann-Kendall non-parametric statistics, the dynamic changes of grassland in northern Xinjiang were analyzed from 2000 to 2019, and the comparative advantages of TI-NDVI and NDVImax in expressing the temporal and spatial heterogeneity of grassland were explored. Results indicated that(1) the grasslands in northern Xinjiang, characterized by NDVImax and TINDVI, showed evident altitudinal differentiation. In general, the TI-NDVI increased with the increase of NDVImax.However, the areas with the same NDVImax or TI-NDVI showed great differences in TI-NDVI or NDVImax.(2)From 2000 to 2019, the grassland TI-NDVI and NDVImax in the northern Xinjiang showed a significant increasing trend(P<0.01), but the spatial differentiation of the changing trends of TI-NDVI was different from that of NDVImax. 17.55% of the grassland in northern Xinjiang showed opposite changing trends in TI-NDVI and NDVImax. For Altai Mountains and the mountains around Ili Valley, which are characterized with grassland of high coverage, the NDVImax and TI-NDVI showed opposite changing trends.(3) The CVof TI-NDVI was higher than NDVImax in temporal and spatial dimensions in grassland with high coverage in northern Xinjiang.Furthermore, TI-NDVI was more sensitive to the temporal and spatial heterogeneity of high-coverage grassland,which can weaken the influence of saturation defect of NDVI in grassland dynamic evaluation to a certain extent.