Microbiome assembly in thawing permafrost and its feedbacks to climate

Abstract

The physical and chemical changes that accompany permafrost thaw directly influence the microbial communities that mediate the decomposition of formerly frozen organic matter, leading to uncertainty in permafrost-climate feedbacks. Although changes to microbial metabolism and community structure are documented following thaw, the generality of post-thaw assembly patterns across permafrost soils of the world remains uncertain, limiting our ability to predict biogeochemistry and microbial community responses to climate change. Based on our review of the Arctic microbiome, permafrost microbiology, and community ecology, we propose that Assembly Theory provides a framework to better understand thaw-mediated microbiome changes and the implications for community function and climate feedbacks. This framework posits that the prevalence of deterministic or stochastic processes indicates whether the community is well-suited to thrive in changing environmental conditions. We predict that on a short timescale and following high-disturbance thaw (e.g., thermokarst), stochasticity dominates post-thaw microbiome assembly, suggesting that functional predictions will be aided by detailed information about the microbiome. At a longer timescale and lower-intensity disturbance (e.g., active layer deepening), deterministic processes likely dominate, making environmental parameters sufficient for predicting function. We propose that the contribution of stochastic and deterministic processes to post-thaw microbiome assembly depends on the characteristics of the thaw disturbance, as well as characteristics of the microbial community, such as the ecological and phylogenetic breadth of functional guilds, their functional redundancy, and biotic interactions. These propagate across space and time, potentially providing a means for predicting the microbial forcing of greenhouse gas feedbacks to global climate change.

FIGURE 1

The proposed framework for community assembly processes in a generalized thawing permafrost landscape. In the thawing permafrost landscape through time (a), assembly processes (e.g., drift, diversification, dispersal, selection per Box 2 figure) act on microbial cells from the thawed soil and the overlying active layer, resulting in a post‐thaw community. Within the post‐thaw community (b), acquisition of new members into the species pool is limited by frozen conditions in intact permafrost, where the community is shaped by selection for survival and growth under permafrost conditions. The disturbance of thaw causes immediate disruptions that have a lasting effect on the trajectory of community composition and ecosystem processes. In early thaw stages, the assembly of the post‐thaw microbiome is dominated by dispersal of new members and drift. As time since thaw advances, dispersal and drift continue, while the collective impact of selection by post‐thaw conditions (as depicted by variation in environmental filters over time) builds and genetic diversification occurs (more rapidly, as generation pace increases post‐thaw). Environmental filters may be selective for abiotic factors (like temperature, redox), biotic factors (like interspecies competition for C substrates or predation), or functions. The magnitude of the effect of dispersal, drift, and diversification will depend on time and the intersection of site characteristics and disturbance intensity. The timescales of these processes in natural systems, particularly understudied permafrost systems, is unknown and should be a subject of further research. However, based on studies of transcription (Coolen & Orsi, 2015), community change (e.g., Mackelprang et al., 2011), and assembly (Doherty et al., 2020), it is expected that the immediate effects of the thaw disturbance are realized in minutes to months, and the longer term effects are felt in years to decades. Artwork by Victor O. Leshyk.

FIGURE 2

Conceptual framework connecting thaw disturbance and community characteristics to assembly and the predictability of system function. The relative contribution of stochastic and deterministic assembly processes are expected to be impacted by permafrost thaw disturbance characteristics (e.g., time since thaw and thaw disturbance intensity) and functional guild features (phylogenetic and ecological breadth) discussed in Sections 2, 3, 4. Note that all grey and white wedges are independent and are visual representations of the scale of each disturbance or community characteristic (e.g., when disturbance intensity is high the wedge is big, and when it is low the wedge is small). Phylogenetic breadth refers to functional guild evolutionary cohesiveness (e.g., narrow for anaerobic methane oxidation, broad for aerobic methane oxidation); phylogenetically narrow, for example, indicates that functional guild members (e.g., those members sharing a trait) are phylogenetically clustered (see Section 4.1). Ecological breadth among members of a functional guild refers to how similarly members respond to environmental drivers (i.e., an ecologically narrow guild responds more similarly to a change in abiotic and/or biotic conditions). Notably, these operational concepts depend on the trait and habitat characteristic being considered (e.g., the guild of anaerobic methane oxidizers is phylogenetically narrow and ecologically narrow with respect to oxygen, but not to salinity). See Section 2. The numbers indicate observations in the field or laboratory (references follow), and their placement on the wedges shows the relative contribution of disturbance and functional guild features that affect whether stochasticity or determinism dominate. These characteristics have implications for the predictability of community function (bottom arrow) arising from dominance by stochastic versus deterministic processes, and thus the approaches required to model specific microbiome outputs (such as greenhouse gas emissions). References: 1. McCalley et al., ; 2. Drake et al., ; Spencer et al., ; 3. Allan et al., ; 4. See example below; 5. Knoblauch et al., , Ernakovich et al., ; 6. Siljanen et al., .