Most of the world's drinking water is derived from groundwater. Worldwide, the proportion of drinking water from groundwater averages 50% (The United Nations - World Water Development, 2009). The production of drinking water from groundwater is possible because of ecosystem services such as the degradation (breakdown) of organic compounds (natural attenuation; Herzyk et al., 2017; Meckenstock et al., 2015; Schmidt et al., 2017). However, large parts of groundwater are contaminated or vulnerable to the effects of human intervention (Griebler et al., 2010). The main threats appear to be a rise or fall in temperature due to the use of geothermal energy (Griebler et al., 2016) and contamination with organic and inorganic substances from industrial and accidental pollution (Schmoll et al., 2006).
The entire groundwater food web, which not only consists of prokaryotes (i.e. single-celled organisms without a cell nucleus), but also of protozoa ("pre-animals") and metazoa (multicellular animals, i.e. fauna), can be involved in ecosystem services that enable drinking water production. However, the contribution of each individual food web component and the effects of possible anthropogenic stressors on each of them, are largely unknown. In order to ensure drinking water production from groundwater also in the future, the contribution of the food web to the ecosystem service must be known. In the following sections, an overview over the groundwater food web and over modeling techniques is given with which the degradation of organic substances can be modeled.
Groundwater food web
Groundwater is home to a multitude of organisms from almost all phyla (Wilkens et al., 2000). Microorganisms as producers form the basis in the food web. The food web is cut off in the groundwater - there are no real predators (Gibert and Louis Deharveng, 2002). Although the resources are low, there are at least three food web levels in many locations: producers (in the absence of light mostly chemolithotrophic, i.e. they get their energy from the conversion of chemical compounds) and decomposers (they break down dead organic substances, i.e. detritus), 1st order consumers who feed on producers and decomposers by grazing on them, and 2nd order consumers / predators who feed on 1st order consumers (larger invertebrates that feed on protozoa, nematodes and microcrustaceans; classification according to Griebler and Mösslacher, 2003).
Modelling the groundwater reactive transport
Microorganisms are known to occur patchily in groundwater, in micro colonies of up to 100 cells (Harvey et al, 1984). Such microcolonies have been confirmed in those laboratory setting, where resources were set to be limiting (Iltis et al, 2011). Early modeling studies have confirmed that it makes a difference whether microorganisms are modelled as heterogeneously distributed on a microscale (Baveye and Valocchi, 1989; Molz et al., 1986). However, the mass transfer restrictions on the microscale resulting from the parabolic flow profile on the pore scale have largely not been taken into account so far. Where they were considered, the mass transfer constraints were used as the effective rate expressions. However, the parameters for these expressions can hardly be determined (Thullner et al., 2007). The description of the high-resolution pore scale has improved the understanding of the relationship between the geometry of the pore space (Schmidt et al., 2018).
Other approaches for effectively describing the boundaries of the system, but neglecting microbial growth, include homogenization techniques (Jikov et al., 1995) or (heterogeneous) multiscale modeling (Weinan et al., 2005) as well as the concept of rough reactive walls (Veran et al., 2009). However, they focus on the heterogeneous surface structure and not on the heterogeneous distribution of reactivity. The advantage of such an approach is that it can be formulated in mathematically closed equations, i.e. it can be solved analytically. Such an approach is appropriate when our knowledge of average properties is sufficient. These average properties apply, for example, when the geometric dimensions are stationary, i.e. they do not develop over time and space.
But such approaches do not take into account a heterogeneous distribution of the active microorganisms on a microscale. There are currently no effective calculation approaches for such cases. The influence of colony formation on the bioavailability and thus the biological degradation of a substrate is thus still largely unknown (Schmidt et al., 2017).
However, our first results show that computationally, the colonial growth of microorganisms actually reduces pore-scale breakdown by up to an order of magnitude (Schmidt et al., 2018). In this approach, we represented the growth patterns of the organism using individual-based modeling (IbM; e.g. Lardon et al., 2011). IbM introduces random variability for those parameters, for which we cannot describe every single factor that affects the respective parameter, e.g. for the actions and reactions of the organism.
Modelling ecosystem scale degradation
Baas Becking (1934) postulated that “everything is everywhere”, so that any physiologically possible reaction from a microorganism can be expected anywhere in an aquifer. However, the organism densities in groundwater are low. Thus "everywhere" could be true if the temporal and spatial scale is chosen large enough. Given a certain pollution and the aim of using the aquifer as a drinking water resource as quickly as possible after a contamination, the extent to which “everywhere” applies may be too great. Instead, whether a new contaminant is degraded may depend on a single cell and whether it finds the adequate conditions to divide and multiply into large biomasses in precisely those areas of the aquifer that are appropriate for degradation within reasonable periods of time. For this reason, the individual-based model is suitable for investigating the question of how bioremediation, i.e. the steering of biota with the aim of increasing degradation, can be carried out in aquifer areas.
Individual-based bacterial models such as iDynoMiCS (Lardon et al., 2011) or LAMMPS (Jayathilake et al., 2017) were used to examine questions about plasmid transfer, cell aging, motility that leads to capped biofilms, etc. Hellweger et al., 2016). They have demonstrated the advantages of such models where community averages do not apply.
There are a number of individual-based models for invertebrates, such as those from Rinke and Petzoldt (2008). However, none of them have focused on the groundwater fauna or groundwater conditions. The unpublished Proto(zoan)Sim (Jan U. Kreft & Heinrich Eisenmann; personal communication) was intended to extend the individual-based model of bacterial growth BacSim, a precursor to iDynoMiCS (Lardon et al., 2011). Like iDynoMiCS, however, it was only programmed for batch mode, while flow and transport are required for the groundwater simulations. Plus, the protozoa didn't grow by themselves - they were a ball that removed the parts of the biofilm they touched, and these balls moved along the biofilm when they couldn't find enough resources. The more complex individual-based protozoan model for marine ecosystems by Seymour et al. (2009) was limited to 1D movement towards chemical messengers. Holyoak et al. (2000) used an individual-based model for protozoa to examine the predation of one protozoal taxon to another, where the predation led to prey extinction. To the best of my knowledge, however, there is still no protozoan IbM for groundwater. Such an IbM is now being developed and will be presented here and in the specialists’ media.
In the first step, the model from Schmidt et al. (2018) was programmed in such a way that the time-dependent development can be viewed, while the previous approach referred to a pseudo-equilibrium. The video shows how two cells, of which the upstream "shadows" the downstream, grow when acetate is introduced continuously. The same parameters were used as before, and 150 hours were modeled.
Dr. Susanne Schmidt (Dipl.-Biol.)
http://www.sidata.eu
Acetate degradation at groundwater conditions: two starting cells in the 1mm long, 0.2mm wide (only lower half shown) pore grow and break down the continuously flowing acetate. The right colony is in the shadow of the left and therefore grows less quickly and degrades less acetate.
