Disturbance-level-dependent post-disturbance succession in a Eurasian steppe.

Anthropogenic disturbances may decrease as we take measures to control them. However, the patterns and mechanisms of post-disturbance ecosystem succession have rarely been studied. Here we reported that disturbance level determined the importance of stochastic relative to deterministic changes in ecosystem components (plant community composition, soil microbial community composition, and soil physicochemical indices), and thus predefined the pattern of post-disturbance ecosystem succession. We proposed a theoretical framework with five disturbance levels corresponding to distinct succession patterns. We conducted a nitrogen addition experiment in a temperate steppe, monitored these ecosystem components during "disturbance" treatment (2010-2014) and post-treatment "succession" (2014-2018). The disturbance level experienced by each component in each treatment was inferred by fitting the observed succession patterns into the theoretical framework. The mean disturbance level of these components was found to increase quadratically with nitrogen addition rate. This was because increasing nitrogen addition reduced the importance of stochastic relative to deterministic changes in these components, and these changes had a quadratic relationship with disturbance level. Overall, our results suggested that by monitoring the importance of stochastic relative to deterministic changes in an ecosystem, we can estimate disturbance levels and predict succession patterns, as well as propose disturbance-level-dependent strategies for post-disturbance restoration.