Biodiv Sci ›› 2018, Vol. 26 ›› Issue (8): 828-837.  DOI: 10.17520/biods.2018089

Previous Articles     Next Articles

Dynamic change of vegetation cover and productivity of Poyang Lake wetland based on MODIS EVI time series

Linlu Shi1, Yifei Jia1, Aojie Zuo1, Tonghui Ma1, Jialin Lei1, Guangchun Lei1,*(), Li Wen2,*()   

  1. 1 School of Nature Conservation, Beijing Forestry University, Beijing 100083, China
    2 Science Division, Office of Environment and Heritage, Sydney, 2141, Australia
  • Received:2018-03-29 Accepted:2018-08-24 Online:2018-08-20 Published:2018-09-27
  • Contact: Lei Guangchun,Wen Li
  • About author:# Co-first authors


Poyang Lake, the largest freshwater lake in China, is one of two lakes that maintain a natural hydrological link with the Yangtze River. The lake system is critical for biodiversity conservation globally, harboring large number of migratory waterbirds. During the dry season, Poyang Lake fragments in to numerous sub-lakes, and different sub-lakes have different hydrological control and management mode. However, the recent hydrological alternation, presumably caused by the operation of Three Gorge Dam (TGD), is threatening the ecological integrity of the lake system, especially as a wintering ground for waterbirds. A robust investigation of the effects of TGD on vegetation cover and productivity at this critical biodiversity hotspot is therefore timely, and could incorporate recent advances in remote sensing and ecological modelling. In this study, using MODIS EVI (enhanced vegetation index) time series, we investigated the spatiotemporal patterns of growth in the lake for the period of 2000-2014, which includes periods before (2000-2006) and after (2007-2014) TGD was commissioned. Firstly, we extracted 107 16-day MODIS EVI time series (February 2000 to April 2015) for 10 randomly placed transects across the lake. We then applied the adaptive Savitzky-Golay smoothing algorithm to the EVI time series, and extracted four key growth metrics, namely, the starting date of growth season, growth season length, seasonal peak EVI, and productivity index. We found significant associations between the hydrological alternation and changes in vegetation seasonality. First, we found that the vegetation growth characteristics of wetlands under different hydrological control modes showed significant differences. In particular, the vegetation located in the freely connected sub-lakes had a later start of growing season, shorter growing season, lower peak EVI value, and lower primary productivity compared to sub-lakes of other modes. Second, due to the hydrological alteration, growth characteristics of sites in freely connected sub-lakes displayed two cycles per year and differed significantly before and after 2006. The advance of the autumn growing season led to excessive accumulation of biomass, which reduced the palatability of the food of migratory geese. However, this difference does not exist in the sites located in the local controlled sub-lake. Third, free connected and local controlled sub-lakes are both important for the protection of migratory birds of Poyang Lake. It is necessary to protect areas harboring both types of sub-lakes to provide a wider food source for wintering migratory birds. Local hydrology control can, to some extent, slow down the impact of much larger scale hydrological alteration on wetland vegetation growth.

Key words: time series data, remote sensing monitoring, vegetation dynamic changes, TIMESAT, hydrological alternation, floodplain lakes