Identifying the plant traits and patterns of trait distribution in communities that are responsible for biotic regulation of CO₂ uptake–climate responses remains a priority for modeling terrestrial C dynamics. 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. Climatic variables explained about 50% of the variance in temporal trends in GPP despite large variation in CO₂ uptake among seasons, years, and plots of differing composition. GPP was highly correlated with contemporary changes in net radiation (Rn) and precipitation deficit (potential evapotranspiration minus precipitation) but was negatively correlated with precipitation summed over 210 days prior to flux measurements. Plots differed in GPP–Rn and GPP–water (deficit, precipitation) relationships. Accounting for differences in GPP–climate relationships explained an additional 11% of variance in GPP. Plot differences in GPP–Rn and GPP–precipitation slopes were linked to differences in community-level light-use efficiency (GEE*). Plot differences in GPP–deficit slopes were linked to differences in a species abundance-weighted index of specific leaf area (SLA). GEE* and weighted SLA represent vegetation properties that may regulate how CO₂ uptake responds to climatic variation in grasslands.