Key points are not available for this paper at this time.
Soil respiration, the main ecosystem process that produces carbon dioxide into the atmosphere, is sensitive to extreme climatic events. The immediate, usually negative effect of droughts on soil respiration has often been observed, but the recovery of soil respiration following drought is rarely documented. Soil respiration can be reduced beyond the drought year if drought-induced changes suppress soil activity. Alternatively, reduction in soil respiration may be overcompensated in the subsequent years due to increased substrate input and soil moisture, resulting from plant dieback during drought. In addition, post-drought weather patterns may also affect the recovery of soil respiration. In a full-factorial grassland experiment, we combined an extreme (5 months) summer drought in 2014 with four levels of post-drought precipitation regimes, including severe (2 months) droughts, moderate (1 month) droughts, ambient weather, and water addition (four large rain events) in summers of 2015 and 2016. We measured soil respiration monthly between May and November, from 2013 to 2016. The extreme drought had an immediate strong negative effect, decreasing soil respiration by 50.8% in 2014 compared to the control plots, and it had a negative legacy effect in 2015 (14.5% reduction), but not in 2016. This legacy effect was unaffected by the post-drought precipitation regime. Moderate drought decreased soil respiration by 12.1% and 18.6%, while severe drought decreased soil respiration by 18.3% and 27.3% in 2015 and 2016, respectively, while water addition had no effect. Since soil water content in extreme drought plots recovered by 2015, we hypothesize that changes in soil biota and reduced root activity are responsible for extreme drought's long-term negative effects. Overall, our results highlight that extreme droughts may have negative effects on soil respiration well beyond the event, and thus the full effect on carbon cycling may be much larger than what is estimated solely based on the immediate effects.
Lellei‐Kovács et al. (Sat,) studied this question.