SEARCH

SEARCH BY CITATION

Keywords:

  • carbon cycle;
  • climate change;
  • mesocosm;
  • photosynthesis;
  • productivity;
  • rainout shelter;
  • soil moisture;
  • soil respiration;
  • tallgrass prairie

Abstract

Climate change is causing measurable changes in rainfall patterns, and will likely cause increases in extreme rainfall events, with uncertain implications for key processes in ecosystem function and carbon cycling. We examined how variation in rainfall total quantity (Q), the interval between rainfall events (I), and individual event size (SE) affected soil water content (SWC) and three aspects of ecosystem function: leaf photosynthetic carbon gain (inline image), aboveground net primary productivity (ANPP), and soil respiration (inline image). We utilized rainout shelter-covered mesocosms (2.6 m3) containing assemblages of tallgrass prairie grasses and forbs. These were hand watered with 16 I×Q treatment combinations, using event sizes from 4 to 53 mm. Increasing Q by 250% (400–1000 mm yr−1) increased mean soil moisture and all three processes as expected, but only by 20–55% (P≤0.004), suggesting diminishing returns in ecosystem function as Q increased. Increasing I (from 3 to 15 days between rainfall inputs) caused both positive (inline image) and negative (inline image) changes in ecosystem processes (20–70%, P≤0.01), within and across levels of Q, indicating that I strongly influenced the effects of Q, and shifted the system towards increased net carbon uptake. Variation in SE at shorter I produced greater response in soil moisture and ecosystem processes than did variation in SE at longer I, suggesting greater stability in ecosystem function at longer I and a priming effect at shorter I. Significant differences in ANPP and inline image between treatments differing in I and Q but sharing the same SE showed that the prevailing pattern of rainfall influenced the responses to a given event size. Grassland ecosystem responses to extreme rainfall patterns expected with climate change are, therefore, likely to be variable, depending on how I, Q, and SE combine, but will likely result in changes in ecosystem carbon cycling.