Contribution of plant traits to the explanation of temporal variations in carbon and water fluxes in semiarid grassland patches.

Aims Accurate prediction of spatiotemporal variations in carbon and water fluxes of heterogeneous landscape is critical to comprehensively address the effects of climate change and vegetation dynamics on landscape and regional carbon and water cycling. Methods A field study was conducted to characterize the seasonal variations in gas fluxes and explore their relationships with abiotic and biotic factors in a small grassland landscape. Daytime carbon and water fluxes including net ecosystem exchange, gross ecosystem productivity, ecosystem respiration and evapotranspiration (ET) were measured for three types of grassland patches over a growing season using the closed chamber method. The key plant trait variables were measured, based on which community weighted mean (CWM) and functional variance (FDvar) were calculated. Important Findings The results showed that the temporal variations in the carbon and water fluxes were regulated by meteorological, soil and community functional variables. Inclusion of the CWM and FDvar of plant trait measures greatly improved the degree of explanation of the predict models. Specific leaf area and leaf δ¹³C content (Lδ¹³C) were the most important trait variables in affecting the variations of the gas fluxes. CWMs indices had greater importance than FDvar indices in predicting the variation of the C fluxes but FDvar indices were more important for ET than C fluxes. Our findings demonstrated that mass ratio hypothesis and the complementary effects hypothesis are not mutually exclusive but have different relative importance for different ecosystem processes. Community functional traits played important roles in predicting the spatiotemporal variations of carbon and water fluxes in semiarid grassland. [ABSTRACT FROM AUTHOR]