Comparison of time-integrated NDVI and annual maximum NDVI for assessing grassland dynamics
Junjie Yan, Guangpeng Zhang, Hongbo Ling, Feifei Han
Abstract
The saturation of NDVI makes the annual maximum NDVI(NDVImax) underestimate vegetation with high coverage. Time-Integrated NDVI(TI-NDVI) accumulates NDVIs of vegetation at different growing stages in one growing season. TI-NDVI might alleviate the saturation effect of NDVImax when vegetation dynamics are assessed at annual time scale. Taking grassland in Ili Valley at the west end of Tianshan Mountain as the study area, this paper compared NDVImax and TI-NDVI systematically for assessing grassland dynamics and also their correlations with precipitation and temperature. Data of time-series Moderate Resolution Imaging Spectroradiometer(MODIS) NDVI were used and methods of Mann-Kendall trend test, Sen's slope estimator test, coefficient of variation(CV) and partial correlation coefficient were employed. We found that there were 28.69% grassland in Ili Valley where the TI-NDVI and NDVImax showed different changing trends over the years of 2000–2019. TI-NDVI not only revealed more about the spatial and temporal heterogeneities of grassland with high coverage than NDVImax at annual time scale, but also captured more about the close relationship between grassland dynamics and precipitation or temperature in Ili Valley.