Microbial eco‐physiological responses to long‐term crude oil contamination in arid soils: Field evidence and <scp>AI</scp> ‐based prediction of qmic and <scp> qCO ₂ </scp>

Abstract Crude oil contamination poses a persistent environmental challenge because of its complex hydrocarbon composition and slow degradation in soil systems. This study evaluated the long‐term effects of petroleum pollution on microbial eco‐physiological indicators in arid soils of Naft Shahr, western Iran. A total of 120 soil samples were collected across a gradient of oil contamination (low, moderate, and high), and microbial indicators, including microbial biomass carbon (MBC), basal respiration (BR), microbial quotient (qmic), and metabolic quotient (qCO 2 ), were quantified. Results showed that MBC and BR significantly increased with oil concentration, indicating microbial adaptation and elevated activity. In contrast, qmic declined from 72.7 to 21.9 mg Cmic g −1 Corg, while qCO 2 rose from 2.33 to 3.45 mg C‐CO 2 g −1 Cmic h −1 , reflecting reduced microbial efficiency and elevated stress under higher pollution levels. Artificial intelligence models, artificial neural network (ANN) and gene expression programming (GEP), were applied to predict qmic and qCO 2 based on soil physicochemical and biological properties. ANN achieved R 2 values up to 0.999 for qCO 2 estimation using biological variables, while GEP also showed high predictive accuracy. Long‐term petroleum exposure triggers both microbial stress and compensatory adaptations, reflecting a complex ecological adjustment.