Background/Question/Methods If pulse and press changes in water availability have distinct effects on carbon (C) fluxes, increased water deficits under climate change may have important but hard-to-predict effects on ecosystem C balance. Long-term climate patterns shape key ecosystem attributes such as the structure of plant communities, the size and activity of soil microbial communities, and the cycling of colimiting nutrients. Through these mechanisms, historic climate patterns can exert legacy effects that alter the response of ecosystems to novel climate regimes. Long-term precipitation manipulations are a powerful tool to understand how climate shapes ecosystems on multiple timescales. We used a long-term irrigation experiment at Konza Prairie Biological Station (Kansas, USA) to assess how historic climates interact with divergent current climate treatments to shape carbon cycling in a tallgrass prairie landscape. Since 1993, irrigation treatments have increased growing-season precipitation by 33% and decreased the frequency of water deficits. In 2017, we reversed some of these treatments and imposed novel drought treatments. We assessed the effects of historic and current climate in driving key grassland carbon inputs and effluxes—aboveground net primary productivity (ANPP) and soil respiration—and resulting effects on soil C pools. Results/Conclusions We found that, surprisingly, the legacy effects of irrigation on ANPP rivaled those of current irrigation and drought in magnitude. While interannual precipitation amounts modulated ecosystem responses to current climate—with much more pronounced effects during dry than wet years—legacy effects were strongly contingent on landscape position and only manifested in lowland plots where decades of irrigation had shifted the community towards high-producing grass species. A history of irrigation in the lowlands decreased the sensitivity of ANPP to experimental drought. Similarly, historic and current climate treatments only affected soil respiration during the natural drought year of 2018. In the lowlands, a history of irrigation conferred resistance to drought, and historic and current irrigation had additive positive effects on soil respiration. Because the long-term climate manipulation had similar effects on both soil C inputs and effluxes, it was not surprising that historic irrigation did not affect total soil C. However, both current and historic climate manipulations affected labile pools of C, which may help explain the observed legacy effects on plant and microbial C cycling. This unique long-term experiment shows how historic climate patterns can influence ecosystem sensitivity to future climate perturbations.
