logo


Simulation

The Simulation Process

The simulation process involves importing your data and modifying the default model to fully parameterize your model, checking your model, running simulations and inspecting your model outputs.
You may then want to proceed to analysis and interpretation of your model outputs, however the right-hand looping arrow in the diagram above acknowledges that the simulation process can be iterative, in that it may require a few test simulations before you are entirely satisfied with your model data and other parameters, before you run your final simulations to supply you with robust outputs for analysis.

The SimSAGS Descision Support System is Windows-based its design is orientated about the simulation process.



So, each window is designed to facilitate testing your data and importing it into your model, running your simulations and viewing and collecting your model output. You can find out all about each of the windows in the Windows section of this guide, but you are most likely to use the GIS Control, Site Area Mask and Water Locator windows to import your spatial data. The Parameter View is useful to entering your non-spatial data.

Obtaining output from your model is dealt with in the Output Section of this guide and analysis of your output is dealt with in the Analysis Section of this guide. The rest of this section deals with the details of the processes that make up each iteration of your simulation within the modelling software.



Basically, this is what happens:

the status panels and information displays are updated. These allow you to keep track of each of these processes, like the ones on the Main Control Panel and the Animal Populations window. You can request different levels of information and also track progress in a Trace file via your Preferences.

Next, your simulation timings are updated. View these in the Simulation Timings window.

Your rainfall data for the entire year has already been read into the model before starting the daily iterations, so the rainfall is retrieved for just the current day and stored for later use.

Diet choices for the herbivores are predicted based on available forage distributions and animal body-size relations (allometry) with foraging behaviour. Once it is known what they will eat, the modelling software calculates how much they will eat and stores it for later. You can read more about these calculations in Illius et al. (1998), Illius & Gordon (1999) and Illius et al. (2000).

Animal fat gain or loss is calculated based on the previous day's food intake and stored until later.

Today's intake is then levied on the plant parts selected for consumption during the diet selection process. Vegetation biomass is reduced by the predicted amounts.

Animal capacity to survive and reproduce is now predicted based on fat mass calculation made earlier during the animal metabolism process. Animals have to maintain some fat mass to survive and females need to have at least 50% of their maximum possible fat mass for their age before they can become pregnant. Gestation and lactation periods are calculated based on animal allometric relations, and the net tally of births and deaths defines the daily dynamics of the animal populations. You can read more about these calculations in Illius et al. (1998), Illius & Gordon (1999) and Illius et al. (2000).

PHEWS model can be integrated for household-level herbivore management decisions, but more practical at present are the various management rules that can be used and which are described in Illius et al. (1998).

The changes calculated in the herbivore population dynamics and management processes are now made to the herbivore populations.

At last, the day's rainfall is fed in to the soil water balance process to be redistributed according to landscape topography before the actual amount seeping into the soil ready for plant uptake can be calculated. This process is an adaptation of the model described in Walker & Langridge (1996).

The amount of growth resulting from plant uptake of soil water is predicted and then allocated to the parts of each plant type. The phenology and allometric relations between the plant parts of these components are based on the ones described in Poupon (1976), Rutherford (1984) and Dye & Walker (1987).

Lastly, any fire damage is predicted for all of the plant types in your model whether intended (prescribed) or accidental (natural), before finally updating any outputs registered through the Graphics Control and Output Selector.



REM
registered
XHTML 1.0 6sm Get PFE