Model Description

The model runs on a daily iteration interval and uses daily rainfall data to calculate vegetation growth and its allocation to plant parts, the selection and intake of these by animals, the animals' consequent energy and protein balances, body growth, reproduction and mortality.

It therefore simulates animal population dynamics mechanistically, i.e., by coupling them to vegetation biomass dynamics.

The model can be run without animal sales, in which case the ecological carrying capacity of the system is expressed (as the size of the animal population maintained by intake-dependent birth and mortality, without offtake), or with a range of animal sales policies.

The SimSAGS modelling software started as non-spatial representations of savanna dynamics and animal physiology developed by Illius et al. (1996). The mathematical relationships used in the animal component of this model were described in Illius et al. (1998) and Illius & Gordon (1999). The latter also contains an outline of the plant component plus results showing that the model gives reasonable predictions of the relationship between the long-term mean ecological carrying capacity and mean annual rainfall (Fritz and Duncan, 1994).

Module Inputs Outputs Source literature
Climate Daily rainfall, wind speed, atmospheric pressure, radiation, temperature, relative humidity Walker & Langridge 1996
Soil Soil depth, cracking, fertility, root distribution Daily run off, evaporation, infiltration, transpiration
Vegetation Daily growth (from transpiration) Daily biomass fluxes Poupon 1976, Rutherford 1984, Dye & Walker 1987
Animal Daily plant parts abundance and bite sizes. Daily diet selection, intake, biomass fluxes, mortality, reproduction, milk yield Illius et al. 1998, Illius & Gordon 1999, Illius et al. 2000
The components modelling the soil moisture balance and growth of vegetation were based a non-spatial model written by Walker & Langridge (1996). These models used inputs of daily weather conditions (e.g., rainfall, wind speed, atmospheric pressure, radiation, temperature and relative humidity) and soil/plant properties (e.g., depth, fertility and root distribution), and predicted the changes in soil moisture as a function of losses to deep drainage, evaporation and transpiration. Transpiration was translated into daily growth on a per unit area basis and this was partitioned according to the balance of woody plants and grasses.

The phenology and allometric relations between the plant parts of these components (Poupon 1976, Rutherford 1984, Dye & Walker 1987) was used to predict the daily growth of green leaf, stem and seed (grasses) and green leaf, twig, wood and fruit (trees). Trees were assumed to have the same rain-use efficiency (the relationship between net carbon assimilation and transpiration) as grasses, in the absence of clear evidence to the contrary. Literature estimates of tissue senescence, decomposition and invertebrate herbivory were included in the prediction of tissue flow from net photosynthesis through to loss from the system. The state variables were, for grasses: carbohydrate stores, green leaf, dead leaf, green stem plus seed, dead stem, fallen seed; and for trees: carbohydrate stores, green leaf, fallen leaf, current season's twig, wood, fruit, fallen fruit.

Selection of these plant parts and their intake rates were calculated on the assumption that, each day, each species will select the diet that allows maximum daily energy intake, net of the energy costs of foraging. Daily intake for each vegetation component (grass: green and dead leaf and stem; browse: green leaf, shoot, fallen leaf, fallen fruit) was calculated according to the equations of Spalinger & Hobbs (1992), which used the abundance and potential bite size of these components. Selection between grass components was calculated from incisor breadth (based on Illius & Gordon 1987) and a limit, imposed by mouth size, on the ability to select the highest-digestibility component while rejecting those of lower digestibility. Daily potential intake, when abundance is not limiting, was calculated from equations summarising the digesta kinetics model of Illius & Gordon (1991, 1992) which showed good agreement between predicted intake of tropical grasses and that observed in a range of ruminant species. Actual intake was the lesser of that calculated subject to the constraints of food abundance, digestive capacity or ability to deposit protein and fat in animals of each age, sex and reproductive status.

Reproduction in females was determined by animal state (conception could take place if animals had >50% of the maximum fat mass for mature females of the species; pregnancy costs and lactation yield are calculated from body condition and nutrient intake). Mean body fat in each age class, sex and reproductive status was obtained daily from the calculated energy balance. Mortality occurs for animals on reaching zero fat mass.

Non-spatial structures for the plant and animal modules are given in Derry (1998). Spatial aspects of the model include topography, soil nutrient distribution, surface water dynamics, plant distribution, drinking water location and animal spatial foraging behaviour, all of which are addressed by Derry (2004).

Details about the technical interpretation of ecological components in the modelling software are documented in a section on Ecological Architecture.

XHTML 1.0 6sm Get PFE