Explaining variation in productivity requires intraspecific variability in plant height among communities

While recent studies have shown the importance of intraspecific trait variation in the processes of community assembly, we still know little about the contributions of intraspecific trait variability to ecosystem functions. Here, we conducted a functional group removal experiment in an alpine meadow in Qinghai-Tibetan Plateau over 4 years to investigate the relative importance of inter- and intraspecific variability in plant height for productivity. We split total variability in plant height within each of 75 manipulated communities into interspecific variability (TV_inter) and intraspecific variability within a community (ITV_within). Community-weighted mean height among communities was decomposed into fixed community-weighted mean (CWM_fixed) and intraspecific variability among communities (ITV_among). We constructed a series of generalized additive mixed models and piecewise structural equation modeling to determine how trait variability (i.e. TV_inter, ITV_within, CWM_fixed and ITV_among) indirectly mediated the changes in productivity in response to functional group removal. Community productivity was not only affected directly by treatment manipulations, but also increased with both inter- and intraspecific variability (i.e. CWM_fixed and ITV_among) in plant height indirectly. This suggests that both the ‘selection effect’ and a ‘shade-avoidance syndrome’ can incur higher CWM_fixed and ITV_among, and may simultaneously operate to regulate productivity. Our findings provide new evidence that, besides interspecific variability, intraspecific trait variability in plant height also plays a role in maintaining net primary productivity.