12. PROCESSES IN BACTERIAL AND INANIMATE POOLS
12.1. Bacterial processes
sBacteria can be represented explicitly or implicitly in Atlantis. Not all early ecosystem models included bacteria, and PPBIM did not explicitly model bacteria. Below is an excerpt of the PPBIM report justifying the decision not to include bacteria:
Reasons for excluding bacteria in PPBIM (from Muray and Parslow 1997)
“Heterotrophic bacteria have been included explicitly in a number of previous models, most notable Fasham et al. (1990). Bacteria in aquatic ecosystems play a number of important roles. The most important role is to attack, ingest and remineralise both particulate and dissolved organic matter, releasing dissolved inorganic nutrients. This role is represented implicitly in the model by assigning a breakdown rate to the detrital pools. Given the other uncertainties associated with the nature of detritus, and the lack of information on bacterial biomass, growth rates and loss rates in Port Phillip Bay, there is little reason to believe that an explicit model of bacterial dynamics would increase the accuracy of predicted detrital breakdown rates.
Bacteria can also serve as a food source for microzooplankton, and thereby divert organic matter which would otherwise be "lost" back to higher trophic levels. This alternative food source may have some impact on microzooplankton dynamics, and their interaction with small phytoplankton. The small phytoplankton – microzooplankton interaction is important under oligotrophic conditions, but much less important under the higher nutrient loadings of concern in Port Phillip Bay. Moreover, owing to the inefficiency of the microbial loop, relatively little bacterial production finds it way to higher trophic levels (Ducklow 1991). Recent studies have suggested that a large fraction of water-column bacterial production is not grazed but is killed by viruses instead (Weinbauer and Peduzzi 1995) leading to a further reduction in utilisation of bacterial production by higher trophic levels (Murray and Eldridge 1994).”
Atlantis explicitly models aerobic bacteria in the water column and sediments, identified as pelagic bacteria (PB) and benthic bacteria (BB) respectively. This means that functional groups can feed on bacteria rather than on detritus. The representation of bacteria used in Atlantis does a much better job than old consumption-based representations, which treated bacteria as just another consumer. By making bacterial biomass dependent on the volume of detritus the bacterial processes are dynamically represented in a more rigorous way.
A precursor to Atlantis also included anaerobic bacteria. At present these are only implicitly represented in Atlantis, but a dynamic form will be added in a future release.
Key bacterial processes
The growth of PB and BB bacteria is dynamically dependent on the concentration of labile and refractory detritus (DL and DR), where DL is typically considered a primary food source. Bacteria uptake both DL and DR and produce DR, DON and NH3. Bacterial wastes do not contribute to DL but convert detritus into nutrients as part of the nitrification-denitrification and remineralisation processes.
Mortality of bacteria is caused by predation (grazing) by different consumers, as well as non-predation mortality due to oxygen limitation, acidification (optional), extra mortality (optional) and linear mortality
Pelagic bacteria processes are run by the routine Pelagic_Bacteria_Processes() in atGroupProcess.c
The growth of pelagic bacteria is modelled as
\(G_{BP} = mum \cdot B \cdot \left( 1 - \frac{B}{XPB_{DL} \cdot DL + XPB_{DR} \cdot DR} \right)^{kBP}\)
where mum is the maximum growth rate (mum_PB), XPBDL and XPBDR is the maximum ratio of the bacteria to DL and DR biomass (XPB_DL and XPB_DR parameters), B is the biomass of the pelagic bacteria, DL and DR is the biomass of labile and refractory detritus and k_PB is the constant determining the exponent of growth (1- linear, 3-drop off before maximum, 5- drop off at maximum). The growth of sediment bacteria (BB) is modelled in the same way, except that mum_BB and k_BB parameters are used.
There is an option to allow for stimulation of bacterial growth by bioirrigation and bioturbation. This option is turned setting flagbactstim=1 and will enhance growth depending on the biomass of infaunal organisms and porosity of the sediment (see chapter 4).
Mortality (non-predation) of pelagic and sediment bacteria is modelled as
\(M_{BP} = \ \left( \left( mL + \left( 1 - \delta_{O2} \right) \cdot mO + mA + mS \right) \cdot B_{BP} \right) \cdot mortsc\)
where mL is linear mortality, mA is the acidification mortality, mS is the extra mortality due to acidification effects, mO is the oxygen limitation mortality, where (1-δO2) is the oxygen limitation scalar, BPB is the biomass of pelagic bacteria (or BB is sediment bacteria are modelled) and mortsc is an forced external mortality scalar.
The uptake of detritus by pelagic bacteria is the modelled as
\(\frac{d(DL)}{dt} = \frac{Gr_{BP} \cdot PB_{DL}}{B_{BP} \cdot E_{DL,BP}}\)
where EDL, PB is the assimilation efficiency on labile detritus (EDL_BP for pelagic bacteria and EDL_BB for benthic bacteria) and PB_DL is the current (box specific) partitioning of PB to DL (vs DR).
Uptake of refractory detritus DR is modelled in the same way except that PB_DR is used and assimilation efficiencies on refractory detritus is used (EDR_BP and EDR_BB).
12.2 Biogeochemical processes
Details on the fluxes in the inanimate pools have been described in Fulton et al. 2004a and more recently in Link et al. 2011 Atlantis-NEUS model report. The table of inanimate pools tracked in the water column and sediments is given in chapter 6.1.