Plant phenotypic functional composition effects on soil processes in a semiarid grassland

• Herbivores might influence soil by changing plant composition at phenotypic level.
• Phenotypic functional composition of a keystone species did not affect soil though.
• Soil microbes showed different activity optima as soil dried out.
• Herbivore effects on soil through different mechanisms than phenotypic composition.

Our knowledge of plant functional group effects on ecosystem processes is relatively well established, but we know much less on how changes in plant phenotypic composition affect ecosystem functioning (i.e., phenotypic functional composition). Understanding phenotypic functional composition (PFC) is relevant in plant communities strongly dominated by a few keystone species, since alteration of phenotypic composition of these species might be a mechanisms by which land management practices such as grazing impact on ecosystem functioning. Here, we report results from a field experiment where we manipulated PFC of Bouteloua gracilis (Kunth) Lag. ex Griffiths, a keystone species in the semiarid shortgrass steppe. B. gracilis' PFC was altered by using plant genetic lines which expressed consistently either high or low leaf tissue lignin content (LC), a plant trait known to affect soil biogeochemical processes. High-LC lines came from an area subjected to low grazing intensities, whereas low-LC lines came from an area historically overgrazed. Additionally, those plant genetic lines expressing high LC also expressed high dry matter content and vice versa. We established experimental plots with High (only high-LC plants), Low (only low-LC plants), and mixed (both high- and low- LC plants, BHL) LC genetic lines; and measured selected soil processes after the application of a small water pulse. We hypothesised that those soil processes related to microbial activity and nitrogen cycling would be higher in Low plots in comparison with High and BHL plots. Phenotypic functional composition did not affect most of our soil processes including ammonium and nitrate concentrations (inorganic N), microbial activity, potential ammonium mineralization, and microbial profiles of substrate utilization). Nonetheless, we observed a general response to the water pulse so that soil inorganic N increased, but soil water content and soil microbial activity decreased 48 h after the pulse application. This general response suggests that fractions of the soil microbial community with different soil moisture optima mineralise N-rich substrates. Overall, lack of response to plant phenotypic functional composition suggests that grazing effects on soil biogeochemical processes in the shortgrass steppe are not directly mediated through how grazing affects the phenotypic functional composition of B. gracilis.