Precipitation Determines the Spatial Variability of Vegetation and Topsoil Organic Carbon Densities of Alpine Grasslands in the Qinghai-Tibetan Plateau, China

Alpine grassland soils accumulate massive stocks of organic carbon and function as important carbon sinks on the Qinghai-Tibetan Plateau. Substantial uncertainties prevent a full understanding of ecosystems’ carbon pools and their responses to environmental factors, mainly because of limited observations and inconsistent up-scaling algorithms. This study compiled data since 2000 for 422 alpine grassland sites in Qinghai Province, China, to investigate vegetation and organic carbon densities in the topsoil (at 0–30 cm depth) and their spatial variations. The site-averaged below-ground biomass carbon density (BOD) and topsoil organic carbon density (SOD) were both highest in alpine meadows with values of 0.43 ± 0.34 (Mean ± S.D.) and 12.52 ± 5.74 kg C/m2, respectively. They are about five times the lowest corresponding values of alpine desert steppes. The above-ground biomass carbon density (AOD) is not significantly different between alpine steppes and alpine desert steppes, and averaged 32.66 ± 22.02 g C/m2, around twice that for alpine meadows. Boosted regression tree models and a structural equation model consistently show that mean annual precipitation, rather than mean annual air temperature, was the predominant factor influencing the spatial variability of site-level AOD, BOD, and SOD across the alpine grasslands. The boosted regression tree models, integrated with spatial datasets of topographic attributes, mean annual air temperature and precipitation, and mean annual maximal normalized difference vegetation index, yielded area-averaged AOD, BOD, and SOD values of 22.67 ± 4.48 g C/m2, 0.37 ± 0.074 kg C/m2, and 9.53 ± 4.48 kg C/m2, respectively. Modeling results indicate that ecosystem carbon densities increase from northwest to southeast, mainly following the spatial patterns of vegetation greenness and precipitation. The size of the total terrestrial ecosystem carbon pool in Qinghai province is estimated to be 3.65 Pg C, of which 96.02% is stored in topsoil, 3.75% in below-ground biomass, and 0.23% in above-ground biomass. Alpine meadows and alpine steppes account for 70.9% and 15.6% of the total carbon stocks of Qinghai province, respectively. Our results provide improved estimates of carbon pools and underscore the predominance of precipitation in determining spatial variability in the carbon stocks of alpine grassland ecosystems. Our findings will help elucidate the feedback mechanisms between carbon storage and climate changes in alpine grassland.