Towards an understanding of the factors controlling bacterial diversity and activity in semi-passive Fe- and As-oxidizing bioreactors treating arsenic-rich acid mine drainage.

Semi-passive bioreactors based on iron and arsenic oxidation and coprecipitation are promising for the treatment of As-rich acid mine drainages. However, their performance in the field remains variable and unpredictable. Two bioreactors filled with distinct biomass carriers (plastic or a mix of wood and pozzolana) were monitored during 1 year. We characterized the dynamic of the bacterial communities in these bioreactors, and explored the influence of environmental and operational drivers on their diversity and activity. Bacterial diversity was analyzed by 16S rRNA gene metabarcoding. The aioA genes and transcripts were quantified by qPCR and RT-qPCR. Bacterial communities were dominated by several iron-oxidizing genera. Shifts in the communities were attributed to operational and physiochemical parameters including the nature of the biomass carrier, the water pH, temperature, arsenic, and iron concentrations. The bioreactor filled with wood and pozzolana showed a better resilience to disturbances, related to a higher bacterial alpha diversity. We evidenced for the first time aioA expression in a treatment system, associated with the presence of active Thiomonas spp. This confirmed the contribution of biological arsenite oxidation to arsenic removal. The resilience and the functional redundancy of the communities developed in the bioreactors conferred robustness and stability to the treatment systems.

Hybrid recursive regularized lattice Boltzmann simulation of humid air with application to meteorological flows.

An extended version of the hybrid recursive regularized lattice-Boltzmann model which incorporates external force is developed to simulate humid air flows with phase change mechanisms under the Boussinesq approximation. Mass and momentum conservation equations are solved by a regularized lattice Boltzmann approach well suited for high Reynolds number flows, whereas the energy and humidity related equations are solved by a finite volume approach. Two options are investigated to account for cloud formation in atmospheric flow simulations. The first option considers a single conservation equation for total water and an appropriate invariant variable of temperature. In the other approach, liquid and vapor are considered via two separated equations, and phase transition is accounted for via a relaxation procedure. The obtained models are then systematically validated on four well-established benchmark problems including a double diffusive Rayleigh Bénard convection of humid air, two- and three-dimensional thermal moist rising bubble under convective atmospheric environment, as well as a shallow cumulus convection in the framework of large-eddy simulation.