Fig. 3 Mean relative growing season biomass of the functional groups - (a) perennial rhizome grasses, (b) perennial bunchgrasses,(c) perennial forbs, (d) shrubs and semi-shrubs, (e) annuals and biennials - and (f) the dominant species at the study site between 1981 and 2011 (n = 10, with standard deviation). Asterisks indicate significant changes in relative biomass (*P<0.05).
Having analyzed all possible mean temperature and accumulative precipitation distribution patterns, we found that the community biomass was positively correlated with P1-9 and P5-9, but negatively correlated with T4-8 and T8 (Table 1). The relative biomass of functional group PR was negatively correlated with P9, but positively correlated with P1-6, P4-6 and T7. The relative biomass of functional group PB was negatively correlated with P1-6and P5-6 in the early growing season. The relative biomass of PF was positively correlated with P6-9, but negatively correlated with T6-8. P9 had positive and negative relationships with the relative biomass of functional group SS and dominant species, respectively. In addition, the relative biomass of dominant species was negatively correlated with P8-9 (Table 1).
Table 1 Climate factors retained in the stepwise regression of community biomass, functional groups relative biomass, species richness, community intra-annual stability, functional groups biomass intra-annual stability