Divergent responses of ecosystem water-use efficiency to extreme seasonal droughts in Southwest China

The recurrent drought extremes have resulted in deleterious impacts on ecological security. Despite that many attempts have been made to explore ecosystem responses to different severities of droughts, a deep understanding of how ecosystem water-use efficiency (WUE) responds to extreme seasonal droughts is critical for predicting the trends under future climate change, especially in the ecologically-fragile karst ecosystem across Southwest China. This study systematically examined the spatio-temporal variations of ecosystem WUE over the karst and non-karst areas, as well as their divergent responses to different seasonal droughts. Our findings revealed the apparent increase in drought frequency, duration, and severity in Southwest China during the past four decades. Meanwhile, spring and summer drought events were the prevailing drought types. Compared with the non-karst area, multi-year mean WUE in the karst area was relatively lower, whereas the area exhibiting significant increase in WUE (p < 0.01) accounted for 39.3% and 22.3%, respectively. However, the effects of drought on ecosystem WUE varied in different seasons with more severe consequence in the karst ecosystem. During the early stage of autumn-spring drought in 2009/2010, ecosystem WUE was apparently larger than the baseline condition with the difference turning to be negative anomalies during the peak period, whereas the effect of summer drought in 2011 led to negative anomalies nearly throughout the duration. Further analysis revealed that the anomalies in evapotranspiration acted a prominent role in altering WUE at the onset of both droughts, while ecosystem WUE was mainly determined by the sensitivity of gross primary production during the later stage. All analyses are beneficial for expecting the coupling relationship between global carbon and water cycles to future climate change, particularly as droughts are projected to increase in terms of frequency and severity.