Explaining temporal variation in above-ground productivity in a mesic grassland: the role of climate and flowering

1. Annual above-ground net primary productivity (ANPP) in mesic grasslands is known to be highly temporally variable. While yearly precipitation or average yearly temperature can explain some of this temporal variability, much of the variation in ANPP remains unexplained. 2. Here we address the heretofore unexplained variation in 25 years of productivity data from a mesic grassland at Konza Prairie (north-eastern Kansas) by examining the effects of precipitation nd temperature during periods relevant to the phenology and growth cycle of the dominant C-4 grasses and the flowering stalk production of these species. We assessed both the direct effects and indirect effects via flowering of phenologically relevant climate periods on ANPP using structural equation modelling (SEM). 3. We found ANPP to be strongly positively influenced by flowering stalk production of the dominant C-4 grasses, precipitation during periods relevant to vegetative growth (15 April-14 July) and flowering stalk elongation (15 July-14 August) of the dominant grasses, and fire. In addition, flowering stalk production was negatively influenced by high temperatures during the flowering stalk elongation period, which therefore resulted in a negative indirect effect on ANPP. We found little evidence for the effects of the previous year's total annual precipitation or mean annual temperature on ANPP. 4. By including flowering stalk production and separating climate variables into phenologically relevant periods we were able to increase the percentage of observed variance in ANPP explained by six models, relating to different topographic positions and burn regimes, from an average of 22% to 48%, with the best model explaining 61% of variation in ANPP. 5. Synthesis. The link between climatic periods relevant to the phenology and growth of dominant C-4 grasses, flowering stalk production of these grasses and ANPP shown here improves our ability to predict productivity in mesic grasslands, an ecologically and economically important ecosystem.