1. Plant responses to major environmental drivers like precipitation can influence important aspects of carbon (C) cycling like soil CO₂ efflux (J_CO2). These responses may be predicted by two independent classes of drivers: plant size—larger plants respire more and produce a larger quantity of labile C, and plant economics—plants possessing more acquisitive plant economics strategies (i.e., high metabolic rate and tissue nutrient content) produce higher-quality tissue that respires rapidly and decomposes quickly.

2. At two sites in central Texas, USA with similar climates and differing soil characteristics, we examined the response of eight Panicum virgatum genotypes to three annual precipitation levels defined by the driest, average, and wettest years from each site’s precipitation history. We evaluated the individual and joint influence of plant genotypes and precipitation on J_CO2 and traits related to plant economics and plant size. We then used confirmatory path analysis to evaluate whether effects of precipitation on J_CO2 were in part related to effects of precipitation on plant economics traits or size (‘mediated’ effects).

3. These genotypes exhibited variation in plant economics traits and aboveground net primary productivity (ANPP), an aboveground measure of plant size. Increasing precipitation increased J_CO2 and ANPP more than plant economics traits. At both sites, ANPP was the single best predictor of J_CO2. Moreover, the sites differed in the ways that plant size and plant economics traits combined with precipitation to influence J_CO2. At the Austin site, the positive effect of precipitation on J_CO2 was mediated primarily by ANPP, offset by a smaller effect of leaf nitrogen content; no direct precipitation effect was detected. At the Temple site, increasing precipitation had positive direct and ANPP-mediated effects on J_CO2. This suggests that greater water limitation at Austin may strengthen the links between plant size and J_CO2.

4. Synthesis Estimates of C cycling can be improved by accounting for mediation of precipitation effects on J_CO2 by plant economics traits and plant size in resource-limited environments.

Data file supporting the manuscript, "Plant biomass, not plant economics traits, determines responses of soil CO₂ efflux to precipitation in the C₄ grass Panicum virgatum". plot_global is a unique identifier for each plant; site is the location of the experiment (Austin or Temple); block is a spatial block; sub_block is a plot, the level at which precipitation treatments were applied; treatment is a precipitation treatment with three levels (low, mean, high); genotype is one of eight genotypes used in this study (see Table S2 for more information); precip_cm is the annual precipitation applied to individual plots (cm); precip_dm_centered is the mean-centered annual precipitation applied to a plot (dm); efflux_avg is the soil CO₂ efflux, calculated as an area under the curve of seasonal measurements; soiltemp_summer is the soil temperature measured during the July (Austin) or August (Temple) soil efflux measurement (max soil temperature); soilmois_avg is the average soil moisture (%) from monthly measurements; annp is aboveground net primary productivity (kg plant ^-1); leaf_n is % leaf N; ldmc is leaf dry matter content (mg g^-1); bnpp is belowground net primary productivity (g) measured from root ingrowth cores; root_cn is root C:N.