Comparison of numerical simulations of reactive transport and chemostat-like models

TL;DR

This paper compares reactive transport models and chemostat-like models in microbial ecology, highlighting numerical accuracy issues near non-hyperbolic equilibria and emphasizing cautious interpretation of simulation results.

Contribution

It evaluates the accuracy of reactive transport models like MIN3P in simulating microbial transformations in chemostat conditions, revealing limitations near certain equilibria.

Findings

Numerical accuracy depends on system state, especially near non-hyperbolic equilibria.

MIN3P's discretization may cause inaccuracies in advective transport simulations.

Caution is advised when interpreting model predictions in complex microbial systems.

Abstract

The objective of the paper is to evaluate the ability of reactive transport models and their numerical implementations (such as MIN3P) to simulate simple microbial transformations in conditions of chemostat or gradostat models, that are popular in microbial ecology and waste treatment ecosystems. To make this comparison, we first consider an abstract ecosystem composed of a single limiting resource and a single microbial species that are carried by advection. In a second stage, we consider another microbial species in competition for the same limiting resource. Comparing the numerical solutions of the two models, we found that the numerical accuracy of simulations of advective transport models performed with MIN3P depends on the evolution of the concentrations of the microbial species: when the state of the system is close to a non-hyperbolic equilibrium, we observe a numerical inaccuracy that may be due to the discretization method used in numerical approximations of reactive transport equations. Therefore, one has to be cautious about the predictions given by the models.