Xinjiang is the largest arid and semi-arid region in China, where landscapes are highly vulnerable to both fire and human disturbance. Yet the long-term mechanisms linking fire regimes, climate, and human activities in this region remain poorly understood. Here, we use a sediment core from Bosten Lake, China's largest inland freshwater lake, to reconstruct fire and vegetation changes over the past 2000 years based on black carbon and terrestrial pollen, and apply regression-based anomaly detection to examine fire regimes. During 1–1400 AD, fires were likely primarily controlled by fuel availability. Black carbon and pollen were strongly correlated, and fire activity increased with vegetation biomass. During 1400–1800 AD, despite a wetter climate that would naturally increase fuel, rapid population growth, land reclamation, and slash-and-burn farming likely contributed to vegetation destruction and a sharp rise in fires, marked by a clear decoupling of black carbon and pollen. After 1800 AD, fire frequency did not increase with continued population growth, reflecting a landscape pushed past its recovery threshold. Vegetation remained severely depleted, fuel loads stayed low for an extended period, and even persistent anthropogenic ignitions could not generate large fires, trapping the system in a long-term degraded state. Our record shows that large-scale human disturbance likely began at 1400 AD. This suggests a fundamental shift from likely fuel-controlled fires to increased human influence on ignition, followed by a collapse into a fuel-limited regime after the landscape crossed a recovery threshold. Critically, excessive human disturbance in arid regions can decouple fire frequency from population growth and lock the system into prolonged ecological degradation. Under ongoing climate change and intensifying human activities, these findings underscore the need to keep land-use intensity below the historical threshold and to strengthen fire monitoring as warming-wetting and afforestation increase fuel loads.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: