Statistical Learning and Uncommon Soil Microbiota Explain Biogeochemical Responses after Wildfire.

Wildfires are a perennial event globally, and the biogeochemical underpinnings of soil responses at relevant spatial and temporal scales are unclear. Soil biogeochemical processes regulate plant growth and nutrient losses that affect water quality, yet the response of soil after variable intensity fire is difficult to explain and predict. To address this issue, we examined two wildfires in Colorado, United States, across the first and second postfire years and leveraged statistical learning (SL) to predict and explain biogeochemical responses. We found that SL predicts biogeochemical responses in soil after wildfire with surprising accuracy. Of the 13 biogeochemical analytes analyzed in this study, 9 are best explained with a hybrid microbiome + biogeochemical SL model. Biogeochemical-only models best explain 3 features, and 1 feature is explained equally well with the hybrid and biogeochemical-only models. In some cases, microbiome-only SL models are also effective (such as predicting NH). Whenever a microbiome component is employed, selected features always involve uncommon soil microbiota (i.e., the "rare biosphere" [existing at <1% mean relative abundance]). Here, we demonstrate that SL paired with DNA sequence and biogeochemical data predicts environmental features in postfire soils, although this approach could likely be applied to any biogeochemical system. Soil biogeochemical processes are critical to plant growth and water quality and are substantially disturbed by wildfire. However, soil responses to fire are difficult to predict. To address this issue, we developed a large environmental data set that tracks postfire changes in soil and used statistical learning (SL) to build models that exploit complex data to make predictions about biogeochemical responses. Here, we show that SL depends upon uncommon microbiota in soil (the "rare biosphere") to make surprisingly accurate predictions about soil biogeochemical responses to wildfire. Using SL to explain variation in a natively chaotic environmental system is mechanism independent. Likely, the approach that we describe for combining SL with microbiome and biogeochemical parameters has practical applications across a range of issues in the environmental sciences where predicting responses would be useful.