Rewetting without rain: Cryptic controls on dryland decomposition in a hyperarid desert

Background/Question/Methods

In drylands, a large proportion of annual decomposition occurs under rewetting events. Ecosystem models often underestimate respiration in these periods, suggesting that certain mechanisms that influence decomposition in drylands are not taken into account. We aimed to describe the contribution of two such mechanisms: stimulation of decomposition by non-rainfall moisture, and the specialization of microbial traits to unique desert habitats. We examined the decomposition of the desert grass Stipagrostis sabulicola across a precipitation gradient in the hyperarid (< 100mm/year) dune sea of the Namib Desert in southern Africa. Precipitation in the western region of the gradient is almost completely in the form of fog, while in the east, rewetting is induced solely by sporadic rainfall events. We collected S. sabulicola stems across this gradient and described the taxa that were active under a simulated wet-up using BrDU nucleotide labeling and 16S rRNA and ITS Illumina sequencing. We also measured respiration responses of litter to water and humidity in the lab, and described differences in the traits of fungi isolated from S. sabulicolastems. 

Results/Conclusions

In the lab, microbial communities on S. sabulicola litter respired within 5 minutes of being wet, with fungal activity contributing approximately 80%, providing evidence that communities are viable in this environment, despite low and infrequent available water. High humidity also induced respiration on larger pieces of litter. We found compositional legacies from different precipitation types (fog and rain) in both active and total microbial communities. This suggests that different precipitation types selects for unique microbial communities in the long-term, and this altered ‘seed bank’ changes the community that responds during rewetting. Taxa isolated from these communities had different physiological traits that allow them to make use of the predominant form of rewetting (rain or fog) in a given area. Combined, our results suggest that not all rewetting events are created equal; non-rainfall precipitation may be an unaccounted-for control on dryland decomposition, and may select for specific microbial traits and taxa that could influence the relationship between biotic decomposition and environmental variables.