Greater responses of flower phenology of Kobresia pygmaea community to precipitation addition than to constant and stepwise warming

There is a debate about unmatched results between manipulative warming using constant warming rates every year (CW) and long-term observations warming affect temperature sensitivity of flowering phenology. This may be because long-term observations represent the actual yearly increase in temperature (i.e. a yearly stepwise warming rate per year, SW) which would differ from CW and their effects would be regulated by precipitation alteration. We conducted a warming experiment with CW (temperature increase by +1 degrees C and sustained this elevated temperature for the duration of the study) and SW (temperature increase by + 0.25 degrees C progressively each year) with precipitation addition in an alpine grassland for four years. Our results showed that neither warming rate affected community flowering phenology. However, precipitation addition advanced onsets of flowering for early-spring flowering (ESF) and mid-summer flowering (MSF) groups, and advanced the end date of flowering for ESF but delayed it for the MSF group. Therefore, flowering duration remained stable for the ESF group and prolonged for the MSF group, and further prolonging the flowering duration of the community. There were no interactions between warming rates and precipitation addition on the community's flowering phenology. A severe drought in a year significantly decreased the maximal number of community flowers in the following year. Therefore, a change in precipitation has a greater effect than warming on the community flowering phenology in the semi-arid alpine grassland.