论文部分内容阅读
Vegetation Indices (VIs) are often used as input parameters for estimating gross primary productivity (GPP),but the relationships between Vis and GPP are various depending on the terrestrial environments and climate conditions.To optimize GPP models,we need adopt the most suitable Vis as the input parameters.In this study,we examined relationships between Moderate Resolution Imaging Spectrometer (MODIS) derived Vis and tower-measured GPP at four flux tower sites in arid and semi-arid grassland of northern China.Our results showed that 1) Among the nine Vis (RVI,NDVI,SAVI,ARVI,EVI,GRVI,RSR-5,RSR-6,RSR-7),EVI was the best indicator of GPP in the arid and semi-arid grassland.The decreasing trend of correlation coefficients from south sites to north sites showed GPP and EVI has closer relationships in high vegetation coverage of south sites; 2)Considering the seasonal impacts on the correlation of GPP and Vis,GPP had the best correlation with SAVI in spring when the grassland canopy was sparse,but EVI had the best correlation with GPP in summer with the dense coverage of grassland; 3) Drought had a suppress impact on vegetation growth,and GPP has a decrease in drought season,however,EVI failed to capture the fluctuations of GPP through the drought season thought it can catch the trend of GPP; 4) Temperature had more influences on GPP in drought season than precipitation.GPP had a quicker response to precipitation in drought season than that in non-drought season.The lagging effect of precipitation on GPP prolonged 24-d in non-drought season,but only 8-d in drought season.And the period of lagging effect lasted around one month in non-drought season,but around two months in drought season.