Although belowground biomass (BGB) plays an important role in global cycling, the storage of BGB and climatic effects on it are remaining unclear. With data from 49 sites, we aimed to investigate BGB and its climatic controls in alpine shrublands in the Tibetan Plateau. Our study showed that the BGB (both grass-layer and shrub-layer biomass) storage in the alpine shrublands was 67.24 Tg, and the mean BGB density, and shrublands area were 1 567.38 g m^-2 and 4.29×10⁴ km², respectively. Shrub layer had a larger BGB stock and accounted for 66% of total BGB this area, while only 34% was accumulated in the grass layer. BGB of the grass-layer in the Tibetan Plateau shrublands was larger than that of Tibetan alpine grasslands, indicating that shrubland ecosystem played a critical importance role in carbon cycle on the Tibetan Plateau. The BGB in the grass layer and shrub layer demonstrated different correlations with climatic factors. Specifically, the effects from mean annual temperature on shrub-layer BGB were not significant, similarly to the relationship between mean annual precipitation and grass-layer BGB. But shrub-layer BGB had a significantly positive relationship with mean annual precipitation (P<0.05), while grass-layer BGB showed a trend of decrease with increasing mean annual temperature (P<0.05). Consequently, the actual and potential increases of BGB varied due to different increases of mean annual precipitation and temperature among different areas of the Tibetan Plateau. Therefore, in the warmer and wetter scenario, due to contrary relationships from mean annual precipitation and temperature on shrub-layer BGB and grass-layer BGB, it is necessary to conduct a long-term monitoring about dynamic changes to increase the precision of assessment of BGB carbon sequestration in the Tibetan Plateau alpine shrublands.

A total of 147 plots were collected across 49 sites (i.e., each site including three plots) in the Tibetan Plateau shrublands during the period of July and August from 2011 to 2013. The distance was between 5 m and 50 m between any two plots. Each plot of 5 m × 5 m was used for ecological investigation, such as definite dominant species. According to the sampling method for shrubland ecosystem, the sampling size should be not smaller than 1 m × 1 m (Technical Manual Writing Group of Ecosystem Carbon Sequestration Project, 2015). one subplot of 1 m × 1 m was sampled, and the total of all plants in the subplots was dug up to determine BGB. Generally, we dug a pit of 1 m depth, or we dug until the parent material horizon to collect root samples. The BGB comprised biomass including the grass layer and the shrub layer. Grass and shrub roots grow together in each plot, and based on lignify of BGB, we separated them in each plot. The criteria of distinguishing grass and shrub root were whether they are herbs or woody. Specifically, the shrub layer primarily consisted of woody biomass, while grass layer composed by herbs biomass. Meanwhile, combining with AGB, we can easily distinguish grass-layer BGB and shrub-layer BGB. Then, we only retained root sample and cleaned up the other parts (Technical Manual Writing Group Of Ecosystem Carbon Sequestration Project, 2015). The biomass samples were all weighed to the nearest 0.01 g after it was oven-dried to a constant mass at 65°C. The BGB in the grass layer and shrub layer were shown in Table 1.