Agents and elements

Here is my initial brainstorming for agents and elements that could play a role in my model (if it ends up being agent-based):

Adult Terns:

necessary actions
fish
feed young
feed self
feed each other
guard nest
idle

optional* actions
incubate
copulate
court
fight

necessary properties
nest (identify which nest)
fish (boolean, or possibly more descriptive about fish type/size)
location

optional* properties
chicks (optional because this could be identified by looking at which chicks have the same nest as this tern)

Adult Gulls:

necessary actions
fish
hunt for terns
feed self
feed young
idle

optional* actions
none

necessary properties
none

optional* properties
nest

Tern Chicks:

necessary actions
feed
idle
wander

optional* actions
beg

necessary properties
nest
fitness/health

optional* properties
none

Nest Scrapes:

necessary properties
none

optional* properties
location
occupied (boolean)

Colony:

necessary properties
nests
gulls
terns

optional* properties
terrain (grass, etc.)

*optional: may or may not be included in the model because I am not sure if it is necessary.

A duel of applicable modeling strategies

After reading the introduction to modeling in Modeling and Simulation Fundamentals (J. Sokolowski, C. Banks 2010) I believe I have whittled the choices in modeling strategies down to data-based modeling or agent-based modeling. The former would be the math formula based solution, likely modeling survivorship as a single quantity (treating the tern chicks as one population rather than many individuals). The latter would be the individual decision based solution, modeling survivorship of each tern chick individually and allowing statistics to provide an overall measure of survivorship.

I am still not sure which method I am leaning towards. I would really like to attempt this problem from an agent-based angle but I am not sure how feasible that is. My concern with the agent-based strategy is that I may either not be able to figure out the right distribution of variables to obtain data similar to the observed survivorship in past years OR I might not have adequate support for my choices of what I model on an individual scale, resulting in a model that others deem useless in actually predicting future survivorship.