Key points are not available for this paper at this time.
We present a predictive, multiscale modeling framework for chemotaxis in porous media. This model results from volume averaging the governing equations for bacterial transport at the microscale and is expressed in terms of effective medium coefficients that are predicted from the solution of the associated closure problems. As a result, the averaged chemotactic velocity is an explicit function of the attractant concentration field and diffusivity, rather than an empirical effective chemotactic sensitivity coefficient. The model was validated by comparing the transverse bacterial concentration profiles with experimental measurements for Escherichia coli HCB1 in a T‐sensor. The averaged chemotactic velocity predicted by the model was found to be within the range of values reported in the literature. Reasonable agreement (approximately 10% mean absolute error) between theory and experiments was found for several flow rates. In order to assess the potential for decreasing the computational demands of the model, the macroscale domain was divided into subdomains for the coupling of bacterial transport to that of the attractant. Sensitivity analysis was performed regarding the number of subdomains chosen, and the results indicate that bacterial transport (as measured by concentration profiles) was not highly affected by this choice. Overall, these results suggest that the predictive, multiscale modeling framework is reliable for modeling chemotaxis in porous media when chemotactic transport is significant compared to convective transport.
Porter et al. (Wed,) studied this question.