Decoupling geogenic and bioclimatic drivers of global microbial niches: A machine learning framework for the genus Rubrobacter

Understanding the macroecological distribution of extremophiles requires separating bioclimatic variables from geological drivers. We present a biogeochemical and Machine Learning (ML) framework mapping the global abundance distribution of the radiation-resistant genus Rubrobacter across 76 soil metagenomes. By combining satellite remote sensing with isotope geochemistry, we developed an Aridity-Alkalinity Index (AAI) to isolate deep geogenic tracers. Models show that Rubrobacter is constrained by a dual pressure: bioclimatic coupling with arid terrains (βAAI = +2.24) and geogenic dependence on unradiogenic marine carbonates (βSr = −0.42; R2 = 0.922). Integrating Rubrobacter Form IE RuBisCO kinetics into a Soil Model for Enhanced Weathering (SMEW) reveals that late-century climate forcing (SSP5–8.5) will unlock winter catalytic windows, accelerating carbon fixation by four orders of magnitude. Fluvial carbon partitioning bifurcates between arid local mineral precipitation and high-efficiency dissolved bicarbonate export along Mesozoic margins (88.2 %), establishing a scalable paradigm for geogenic-microbial climate feedbacks.