Rewetting of dry soils usually stimulates soil carbon (C) emission, a phenomenon known as the Birch effect. Soil C cycling in drylands, which store approximately one third of terrestrial soil organic C (SOC), is strongly affected by wetting–drying cycles. However, the physical, chemical, and biological mechanisms that link rewetting cycles with dryland soil C cycling have not been comprehensively studied, nor do we understand how these mechanisms interact with each other. Here, we conducted a dryland soil incubation experiment manipulating four factors related to global change (soil moisture content, soil moisture variability, C availability, and prior warming) in a factorial design. The experiment was divided into two periods: a rewetting period consisting of six 14-d wetting–drying cycles; and a recovery period lasting 28 days during which soil moisture content was held stable, allowing for examination of the legacy effects of the wet-dry cycles. Rewetting cycles decreased soil aggregate stability under some conditions, but their effects on soil microbial biomass and fungal communities, soil enzyme activities, soil priming, and soil dissolved C were not significant. We found lower average soil respiration under the wetting–drying treatment than the stable soil moisture treatment, and Birch effects were observed, but only under some conditions. This was probably because moisture variability exacerbated soil microbial metabolic stress, which showed itself as oxygen limitation during the initial precipitation pulse and as water limitation during soil drying. Notably, respiration rates remained low even after moisture fluctuations ceased, suggesting a legacy effect of rewetting cycles on dryland microbial communities. Overall, rewetting inhibited aggregate formation (physical mechanism), and suppressed soil respiration by inducing soil microbial metabolic stress (biological mechanism), ultimately leading to lower soil C loss under rewetting. Our findings indicate that Birch effects are mediated by the magnitude of moisture variability, the availability of C, and the degree of physiological stress microbes experience.

干燥土壤的再湿润通常会刺激土壤碳 (C) 排放，这种现象称为桦木效应。旱地的土壤碳循环，储存了大约三分之一的陆地土壤有机碳（SOC），受到干湿循环的强烈影响。然而，将再润湿循环与旱地土壤碳循环联系起来的物理、化学和生物机制尚未得到全面研究，我们也不了解这些机制如何相互作用。在这里，我们进行了一项旱地土壤孵化实验，在因子设计中操纵与全球变化相关的四个因素（土壤水分含量、土壤水分变异性、碳可用性和先前变暖）。实验分为两个阶段：再润湿期由 6 个 14 天的润湿-干燥循环组成；持续 28 天的恢复期，在此期间土壤水分含量保持稳定，以便检查干湿循环的遗留影响。在某些条件下，再润湿循环会降低土壤团聚体稳定性，但对土壤微生物量和真菌群落、土壤酶活性、土壤引发和土壤溶解碳的影响并不显着。我们发现干湿处理下的平均土壤呼吸低于稳定土壤湿度处理，并且观察到桦木效应，但仅在某些条件下。这可能是因为水分变化加剧了土壤微生物代谢应激，表现为初始降水脉冲期间的氧限制和土壤干燥期间的水限制。值得注意的是，即使在湿度波动停止后，呼吸速率仍然很低，这表明再润湿循环对旱地微生物群落产生了遗留影响。总体而言，再润湿抑制团聚体形成（物理机制），并通过诱导土壤微生物代谢应激（生物机制）抑制土壤呼吸，最终导致再润湿下土壤碳损失降低。我们的研究结果表明，桦木效应是由湿度变化的幅度、碳的可用性以及微生物经历的生理应激程度介导的。