Microbial representation#
The microbial groups we include are represented as separate carbon pools. These pools can produce enzymes which drive many of the key processes in the soil model. Most of the parameters related to the activity of each microbial group can be configured separately. This includes their uptake rates for different resources, their turnover rates, their stoichiometric ratios, their thermal responses, and details of the enzymes they produce. The processes that microbial groups are involved in and the different functional groups we represent in the model will now be explained in more detail.
Microbial processes#
Microbial activity is one of the biggest drivers of soil organic matter turnover, so it is important to capture the underlying processes in detail. In our model, microbes take up carbon, nitrogen and phosphorus (in organic and inorganic forms). This is then used to synthesis new biomass, which can be used to replace lost biomass (from cell death and cellular maintenance), to grow, to produce externally secreted enzymes or (in the case of fungi) to produce reproductive bodies.
Nutrient uptake and growth#
In the model microbes can take up the following resources: carbon in the form of LMWC, inorganic nitrogen in the form of ammonium and nitrate, and inorganic labile phosphorus. When microbes take up carbon they also take up organic nitrogen and phosphorus with it, so organic nutrient uptake is not tracked as a separate process. The maximum uptake rate for each resource is found using
where \(k_{i,j}\) is the rate constant for uptake of resource \(j\) by microbial group \(i\), \(C_i\) is the abundance of microbial group \(i\), \(R_j\) is the availability of resource \(j\), \(K_{i,j}\) is the saturation constant for uptake of resource \(j\) by microbial group \(i\), \(f_{T,r}\) is a factor capturing the impact of temperature on the uptake rate, \(f_{T,s}\) is a factor capturing the impact of temperature on the concentration at which the uptake saturates, \(f_W\) is a factor capturing the impact of soil moisture on the uptake rate, and \(f_{p}\) is a factor capturing the impact of soil pH on the uptake rate. These factors are all defined in the soil-abiotic environment links documentation page.
The maximum rate that carbon can be incorporated into biomass is then found by multiplying the carbon use efficiency by the maximum rate of LMWC uptake. This will only be the growth rate when the microbial group is carbon limited (rather than nitrogen or phosphorus limited). Each microbial group has a stoichiometry for new biomass, which is an weighted average of their cellular biomass stoichiometric ratios and the stoichiometric ratios of the extracellular enzymes that they produce. This is average is weighted by the fraction of new biomass allocation to growth vs each type of enzyme production. The limitation that each nutrient places on carbon assimilation is then found by multiplying the stoichiometric ratio for each nutrient by the sum of the maximum uptake rates for all forms of the nutrient (organic and inorganic). The actual growth rate is then found (following Liebig’s law of the minimum) as
where \(G_C\) is the maximum growth rate possible due to the availability of carbon, \(G_N\) is the maximum growth rate (in carbon terms) due to the availability of nitrogen, and \(G_P\) is the maximum growth rate (in carbon terms) due to the availability of phosphorus.
Organic nutrients are preferentially taken up, this is because organic matter will always be needed for growth as it contains all three potential limiting factors. Any nutrient demand not met by organic uptake is met by the uptake of inorganic nutrients. In the case of nitrogen, this inorganic uptake is split between ammonium and nitrate based on the ratio of their maximum uptake rates. It’s important to note that while the maximum rate of organic nutrient uptake is determined by the product of maximum rate of LMWC uptake and the stoichiometry of the LMWC pool, the actual uptake rate can be higher than the product of the actual rate of LMWC uptake and the pool stoichiometry. This represents LMWC being taken up, nutrients being extracted, and then surplus carbon being exuded.
Microbes can also acquire an excess of nutrients through organic matter uptake and need to exude them. However, there are multiple forms in which these nutrients can be exuded as. We assume that the fraction that gets returned in organic form can be estimated based on the degree of carbon limitation as
where \(m_{i,c}\) is the maximum uptake rate of carbon for microbial group \(i\) and \(\epsilon_i\) is the carbon use efficiency for microbial group \(i\). There isn’t anything mechanistic to base the split between ammonium and nitrate on, so instead we introduce a (configurable) ammonium mineralisation proportion constant to determine how nitrogen mineralisation is split between the two.
Enzyme production#
Microbial groups in the model can produce extra-cellular enzymes that drive soil processes. Which enzymes are produced and the allocation to production of each type is controlled by the parameterisation of the specific microbial functional groups. Enzyme classes are differentiated by the substrate they break down (at present this can be POM or MAOM) and by whether they were produced by bacteria or fungi. The reason for splitting enzymes by source is that, due to their substantially larger genomes, fungi can produce significantly more complex enzymes, which we feel is an important difference to capture. Much like microbial functional groups, the keys parameters of each enzyme class are individually configurable.
In the model, microbial groups allocate a (configurable) fraction of their synthesis of new biomass to the production of each type of enzyme that they produce. This comes from the synthesis of new biomass rather than net change in biomass so that enzymes are actually produced when microbial populations are at steady-state. The abundance of extra-cellular enzymes in the soil can also decline due to denaturation, with these denatured enzymes being added to the soil necromass pool.
Microbial functional groups#
We include a total of four microbial functional groups in the model: bacteria, saprotrophic fungi, arbuscular mycorrhizal fungi and ectomycorrhizal fungi. We differentiate between bacteria and fungi because it’s a major taxonomic division and so should be expected to show meaningfully different responses to environmental change. We treat saprotrophic and mycorrhizal fungi separately because they acquire carbon in a very different manner. Finally, we split mycorrhizal fungi into the two most common mycorrhizal types (ecto and arbuscular) because ectomycorrhizal fungi commit substantially more resources to the production of enzymes. There are more microbial groups that we could have included explicitly in the model (e.g. nitrogen fixing bacteria, nitrifying bacteria, etc). However, for these groups there tends to be a real lack of field data on the abundance of these groups and so we don’t include them.
Details of processes that the model includes that are not shared between all microbial functional groups are given below.
Fungal fruiting#
Fungal groups allocate a (configurable) fixed proportion of their synthesis of new biomass to the production of fungal fruiting bodies. The stoichiometric ratios of these bodies are lower than for soil fungal biomass, so producing them takes a proportionally greater uptake of nutrients.
These fruiting bodies are made available for animal consumption (this process is tracked with the animal model). Fungal fruiting bodies that aren’t consumed decay back into the soil in a labile organic form, i.e. as LMWC.
Mycorrhiza#
Mycorrhizal fungi differ from both bacteria and saprotrophic fungi in that they cannot use the forms of carbon that can be taken up from the soil to grow. Instead, they are entirely dependent on their symbiotic plant partners for the carbon they need to grow. Because of this mycorrhizal fungi are assumed to preferentially take up inorganic forms of nutrients. They can take up organic matter if the availability of inorganic matter is insufficient, but this is used solely for the acquisition of nutrients and the carbon contained in this organic matter is returned to the soil as LMWC. If the plant supply of carbon exceeds the mycorrhizal fungi’s ability to use it then the excess is exuded into the soil (again as LMWC).
In return for the carbon they depend on, mycorrhizal fungi supply nutrients to the plants. We assume that mycorrhiza supply a fixed (but configurable) fraction of the nutrients that they uptake to their plant partners. This is a fairly strong assumption, as in reality mycorrhizal fungi should vary the amount of nutrients they supply to plants based on the amount of carbon received, not just soil nutrient availability. We hope that by using this assumption we can treat the outcome of the (incredibly) complex eco-evolutionary involved in plant-mycorrhizal nutrient trading as a model parameter, to be determined based on real world data.
At present, the soil model runs after the plants model (when running the standard set of models). This means that the nutrients the plants receive were effectively supplied on the previous time step. In many ways this lag is biologically realistic, as plants do show delayed responses to changes in soil nutrient conditions. However it does mean that an estimate of the nutrient supplied by the mycorrhizal fungi on the zeroth timestep is needed in order to run the first plant model time step. This estimate is made based on the initial soil conditions under the assumption that mycorrhizal fungi take up nutrients at the maximum possible rate. This is almost certainly an overestimate, but one that only really impact the initial time step of the simulation.
We include two different mycorrhizal groups in the soil model, ectomycorrhizal fungi and arbuscular mycorrhizal fungi. The only difference between these groups are how they are parametrised, with arbuscular mycorrhizal fungi investing far less in the production of extracellular enzymes.