Tradeoffs of fungal and bacterial residues mediate soil carbon dynamics under persistent drought in subtropical evergreen forests

Global climate change has greatly accelerated hydrological processes, causing significant increases in frequency and intensity of drought events, which may have a great impact on soil carbon (C) dynamics. Microbial residues are frequently represented as the important constituents involved in soil C formation and stability. However, how persistent drought influences microbial residues to regulate soil C cycling remains elusive, especially in forest ecosystems. In this study, we investigated drought effects on microbial (fungal and bacterial) residues and soil C dynamics by a 7-year filed experiment (2013-2020) with a 70 % rainfall reduction in a subtropical evergreen forest of Eastern China. Soil samples (0-10 cm) were collected in 2014, 2015, 2016, and 2020 with drought duration of 1, 2, 3, and 7 years, and soil moisture, soil temperature, soil organic carbon (SOC) and soil nitrogen (N) were measured. The contents of microbial (fungal and bacterial) residues were quantified by amino sugar biomarkers. Our results showed that fungal residues significantly decreased with increasing drought duration, which were mainly induced by the decline of soil N and soil moisture, causing a decrease in soil C under the 7 -year persistent drought. In contrast, bacterial residues increased with drought duration due to fast adaption of bacteria to drought and aggregate protection. The retention of bacterial residues might be preserved as a "nutrient reservoir " to cope with the long-term drought in the near future. These results suggest that persistent drought-induced tradeoffs between fungal and bacterial residues mediate soil C dynamics. Fungal residues were the essential prerequisite for soil C sequestration in response to drought, while bacterial residues may mainly regulate soil C cycling under the prolonged drought. Therefore, microbial residues are imperative to understand the responses of microbial-derived C to drought, which could be integrated into terrestrial C models to accurately predict soil C dynamics in unpredictable climate regimes in forest ecosystems.