Primary productivity is regulated by climatic factors, but also depends on which plant species are present and, more specifically, on how plant traits that control photosynthesis and its response to the environment are distributed in the community. We used remotely-sensed estimates of gross primary productivity (GPP) from plots planted to different combinations of perennial grassland species in order to determine links between traits and GPP-climate relationships. We had two specific goals. The first was to identify primary climatic drivers of GPP and the temporal scale of driver variation that was most highly correlated with GPP. Our second goal was to determine whether among-plot variation in GPP-climate relationships could be explained by variation in either the means or variances in values of plant traits presumed to influence GPP (functional diversity; FD) or by a community-level index of photosynthetic light use efficiency (GEE*). GPP was estimated over three years from measurements of the Normalized Difference Vegetation Index for 64 field plots (1 m x 1m) in central Texas, USA planted as 9-species mixtures of all native or all exotic species.
Climatic variables explained >50% of the variance in temporal trends in GPP of grassland communities despite large variation in CO2 uptake among seasons, years, and communities of differing composition. GPP was negatively correlated with contemporary increases in air temperature (AT) and precipitation deficit (potential evapotranspiration minus precipitation) and with precipitation summed over both 15 days and 210 days prior to flux measurements. GPP varied by as much as a factor of three among communities of differing composition. Communities differed in GPP-AT and GPP-water (deficit, precipitation) relationships. Accounting for community-specific differences in GPP-climate relationships explained an additional 12% of variance in GPP. Community differences in GPP-water (deficit, precipitation) slopes were linked to differences in community-level light use efficiency (GEE*). Community differences in GPP-AT slopes were negatively correlated to a species abundance-weighted index of specific leaf N (SLN). Weighted SLN, in turn, was positively correlated with the C3 fraction of community biomass. GEE* and weighted SLN represent community-level vegetation properties that may regulate how CO2 uptake responds to climatic variation in grasslands.