Spatiotemporal temperature trajectories underpin predictive modeling of microbial succession in single-batch...

Spontaneous solid-state fermentation (SSF) efficiency and product quality depend on microbial consortia regulation under dynamic temperature regimes, yet how spatially heterogeneous temperature trajectories are linked to community assembly, volatile profiles, and abundance modeling remains unclear. Using high-temperature daqu SSF as a model, this study integrates spatial multi-omics, machine learning, and physiological characterizations to characterize microbial responses across a 35-65 °C gradient. Results revealed pronounced phylogenetic clustering and niche partitioning along the observed thermal gradient, consistent with temperature-associated structuring of microbial communities into temperature-sensitive and temperature-resistant groups. Microbial succession showed a marked temperature-associated transition around 45 °C within the sampled batch. Lower thermal exposure was associated with the enrichment of temperature-sensitive Lactobacillaceae, whereas higher thermal exposure was associated with temperature-resistant taxa and pyrazine-related volatile profiles. In internal five-fold cross-validation within the single-batch terminal dataset, the Extra Trees regression model showed moderate-to-high fitting performance for selected core bacterial genera (R² = 0.75-0.97). By linking intra-batch thermal trajectories with microbial abundance and volatile profiles, this study provides an exploratory framework for understanding spatial thermal heterogeneity in solid-state fermentation systems.