Abstract: As the second largest carbon pool of the world, terrestrial ecosystems have always been a hot research field with respect to the responses of their carbon budget to climate change. Most studies mainly focus on the effects of temperature, humidity, and CO₂ concentration on carbon cycle of terrestrial ecosystems, but the studies on the effects of desertification and degradation of grasslands on carbon sequestration processes of vegetation are relatively insufficient. In this study, we used remote sensing data with a spatial resolution of 30 m×30 m, to decipher grassland desertification and degradation data with a reference of the national standard GB/T 24255-2009. We simulated gross primary productivity (GPP), net primary productivity (NPP) and net ecosystem productivity (NEP) with a remote sensing model of light energy utilization. We analyzed the temporal and spatial variations of vegetation carbon sequestration and the role of grassland desertification and degradation in Gannan. The grassland mainly presented a state of no desertification or light desertification, and was accompanied by small dispersed areas of land degradation with varying degrees. Among them, grassland desertification and degradation were relatively serious in Maqu County. Grassland desertification and degradation areas accounted for 0.03% and 2.50% of the total study area, respectively. The GPP, NPP, and NEP of grassland in Gannan were characterized by monthly increasing in the first half of the year and decreasing in the second half of the year. The spatial pattern was characterized by the distribution of high in the east and low in the west, high in the north and low in the south. Maqu was the county with the largest annual carbon sequestration of grassland (691.28 t C·a^-1). Among all the counties, only Zhouqu County was a carbon source, with a NEP of -88.64 t C·a^-1. GPP, NPP, and NEP showed a significant decreasing trend along with extremely severe, severe, moderate, light, and no desertification, while grassland degradation did not affect GPP, NPP, and NEP in Gannan. Our results provide basis for revealing the dynamic mechanism of terrestrial ecosystem carbon cycle and data support for the formulation of carbon emission reduction policies in Gannan.

Key words: grassland desertification, grassland degradation, C-FIX, Landsat, net ecosystem productivity.