With the global warming

land ecosystem productivity presents a substantial enhanced trend

which is primarily attributed to the prolonged growing season of terrestrial vegetation. However

the impact of summer greatest growing magnitude on vegetation productivity remains unclear. This study aims to explore the different impacts of growing season length (LOS) and magnitude (MAG) on the long-term trends and interannual variability of vegetation productivity

based on long time series of satellite images. Firstly

the GIMMS NDVI3g data was applied to derive the key phenology parameters

including start and end of growing season (SOS

EOS)

summer growing peak. Summer vegetation productivity (Summary of Vegetation Index

VIsum) was obtained by integrating the area under growth curve. The long-term trends and interannual variability of SOS

EOS and VIsum was explored at pixel and land cover ecosystem level. Special attention was paid to the impacts of LOS and MAG on VIsum changes. The relative important method was used to quantify contribution of LOS and MAG to VIsum. The results indicate the overall vegetation productivity for Northeastern China do not behave a clear trend

but LOS has a decrease trend and MAG show an increase trend. VIsum presents a consistent fluctuation pattern with LOS

but changes asynchronously with MAG. There is a cycle of 10-years length for the trends and variability of LOS

MAG and VIsum. For the spatial distribution at pixel level

VIsum tends to increase in needle forest of northern and southern part

and the western grassland. The spatial distribution of LOS trend shows an opposite pattern with that of MAG. LOS is becoming shortened in the middle cropland and western grassland (81.5% of vegetated area). While MAG is enhanced in this ecosystem (16.5% of vegetated area). This reflects shortened LOS induces the increase of MAG. Across various land types

LOS contribute mostly to the long-term trends and interannual variability of VIsum (75%). LOS plays a key role in northern needle and eastern broad-leaved forest

MAG accounts for 27% of relative importance. The carbon flux data show a similar change pattern with satellite NDVI in the calculation of phenology parameters and vegetation productivity. Although LOS mainly control the trend and interannual variability

MAG may play a control role in the future as it keeps the increase trend.