Evaluation of the extrapolation capabilities of a kinetic biogeochemical model of hydrogen reactivity in deep aquifers by a sensitivity analysis

Integrating dihydrogen (H 2 ) into energy mixes requires assessing its viability in underground gas storages (UGS) like deep aquifers. The primary challenge is H 2 consumption by autochthonous microorganisms, leading to site-dependent issues like souring and petrophysical changes. We performed an innovative approach combining an experimental simulation of a deep aquifer in a high-pressure reactor using native rock and formation water, which was then modeled using a biogeochemical model. Results showed significant acetate production (5.15 mmol kg −1 ), as the site's low temperature favored acetogen development. The model exhibited strong extrapolation capabilities. Using a previous experiment's dataset, adjusting only three parameters of the microbial kinetic model allowed to accurately describe the physicochemical evolution, including H 2 concentration. Finally, a sensitivity analysis identified three key parameters driving fitting accuracy: B max MB , μ max FPB , and the methanogen inhibition constant by sulfate reducers ( k S O 4 2 − M B ).