ALMaSS Rabbit ODdox  1.00
The rabbit model description following ODdox protocol
ALMaSS Rabbit ODdox Documentation



Created by:
Chris J. Topping
Dept. Bioscience,
Aarhus University,
DK-8410,
Roende,
Denmark
11th July 2016

Overview

1. Purpose

The rabbit model has been developed with the intention to determine the resilience of the population to agrochemical usage assuming that the stressors affect mortality or reproduction of rabbits exposed.The aim will be to determine how toxic, and over what area and time, chemicals would have to be applied before significant or lasting rabbit population damage is predicted.

2. State variables and scales

Individuals in the rabbit model are represented by objects of four different classes, and there is a fifth class to represent locations for warrens and if they are occupied the warrens themselves. This means that the individual rabbits are classified as being either Young, Juveniles, adult Males, or adult Females. Each class has a number of attributes and behaviours associated with it, and these attributes may be transferred between classes e.g. as young mature to become juveniles, or juveniles to adult.
All of these classes are based on an ancestor class Rabbit_Base which has some basic functionality common to all rabbits. Additionally, a class Rabbit_Adult exists which has functionality common to adult rabbits. These two classes are simply programming constructs and don't affect the model behaviour, and there are never rabbits in the simulation based on these classes. Note however, that the attributes possessed by Rabbit_Base are inherited by all Rabbit classes, and those possessed by Rabbit_Adult are inherited by Rabbit_Male and Rabbit_Female.

Warrens require special mention. The Rabbit_Warren class handles warren functionality. The warren is a square area of the landscape which is of a standard size and must contain a certain proportion of permanent forage habitat (input parameter RABBIT_WARRENMINPCTFORAGE ). The warren manages its own list of resident rabbits and is responsible for ensuring that the number of rabbits does not rise above its carrying capacity which is determined by the total forage available.
Carrying capacity is determined by the number of burrows possible and the number that are actually present. On creation there is no burrows present and a rabbit must first occupy the warren then dig the burrow before breeding can start. Subsequently the warren will grow as rabbits dig further burrows if the warren is continually occupied. After 2 years it will be assumed to have reached the maximum number of burrows possible.

A class attribute overview is presented below, navigate to the individual entry to obtain further details, or look at 'Process Overview and Scheduling' for functions and behaviours below:

Rabbit Base

Rabbit Young

Rabbit Juvenile

Rabbit Adult

The adult class handles those attributes and functions common to all adults. This is currently only mate and dominance status.

Rabbit Male

Rabbit Female

Rabbit Warren

3. Process Overview and Scheduling


As with all ALMaSS models the rabbit model is based on a state machine which can be described by a state-transition diagram. Rabbits are in a behavioural state and due to decisions or external events move via transitions to another state where they will exhibit different behaviour.

Base States & Functions

Young States & Functions

Juvenile States & Functions

Adult States & Functions

Male States & Functions

Female States & Functions

Warren States & Functions

The warren holds all the information about the cover and forage resources available to the rabbit. In addition it has a list of rabbits that are resident. The number of breeding pairs is controlled by the number of burrows, which itself is controlled by the resources and calculated as a carrying capacity. The actual number of burrows present is also determined by the maintenance of the warren. The warren will achieve its maximum number of burrows if continuously occupied for 2 years. If the warren is abandoned then it will slowly fall into disrepair. The number of burrows possible as a maximum ranges from 3 to 15, fewer than three and the warren is considered unsuitable. Warrens are linked by local networks. This is a method for allowing the rabbits to have some knowledge of their surroundings. Local warrens will be linked and the probability of visiting them during exploration is dependent on their distance from the home warren. Warrens also manage breeding in that a rabbit is currently only allowed to breed if it has both a mate and a burrow. This provides a strong density dependence and prevents juveniles from breeding unless death releases parental burrows.

Rabbit_Population_Manager The rabbit population manager is responsible for handling all lists of rabbits and warrens and scheduling their behaviour. It provides the facilities for output and handles central rabbit functions needed to interact with the landscape. One key function is the pre-processing of the landscape to determine the potential locations for rabbit warrens and the formation of local warren networks allowing rabbits to efficiently locate nearby warrens. This is a very time-consuming step so will typically be done for each landscape and the results stored and used as input later.

Design


4. Design Concepts


4.a Emergence

The breeding season of the rabbits is an emergent property of the weather inputs and the criteria assumed for breeding. This is done by Rabbit_Population_Manager::CheckForRabbitBreedingConditions using monthly mean temperature data as input.
The number and spatial occupancy of warrens (spatial distribution) is emergent from rabbit behaviour interacting with warren networks and the resource distribution in terms of cover and forage.
The rabbit population size is emergent as its age-structure.
The exposure of rabbits to pesticides is emergent dependent upon the pattern of spraying (crops and localities) and the foraging pattern of the rabbits. This is primarily determined by the warren location and the likelihood of a rabbit foraging from a treated area.

4.b Adaptation

Not used.

4.c Fitness

Fitness is measured for the rabbit in terms of whether it has a mate and a burrow, if so it can breed and has a lower mortality rate.

4.d Prediction

The rabbit population manager uses prediction to determine when the rabbit breeding season stops. This is done by looking ahead 90 days into the future weather. This is not prediction in the normal sense of the ODdox, but is an unusual case.

4.e Sensing

Rabbits can sense vacant burrows, other rabbits as potential mates, and can interact with the warren network. The warren network allows them to sense conditions in nearby warrens by travelling there and "looking". They also sense when the breeding season starts and stops.

4.f Interaction

Interactions occur between adult rabbits in finding mates, and between rabbits and the warrens in determining the potential for occupation of warrens and therefore breeding. Weather and environmental conditions influence rabbits both directly (e.g. breeding season timing with temperature, or pesticide induced physiological changes), and indirectly via determining the suitability of warrens and their carrying capacity.

4.g Stochasticity

Stochasticity is used extensively. All rabbits are subject to stochastic mortality, the severity of which is modified by their status and behaviours (e.g. if dispersing). Exploration of local warrens is also stochastic, but modified by distance such that the chance of visiting a nearby warren is distance dependent. Litter size also has a stochastic component.

4.h Collectives

Collectives in the strict sense are not used, although the warren does form a useful administrative unit which can collectively manage some behaviour. Otherwise all rabbits are simulated individually from birth to death.

4.i Observation

Currently the rabbit model uses standard ALMaSS output. This provides a time-indexed population census broken down into young, juvenile, male and females.

5. Initialisation

Uses standard ALMaSS input files, but can be set up to use predefined warren maps. This saves time because calculation of the warren networks takes a long time for big maps. On start-up male and female rabbits are randomly located in the landscape and must initially establish warrens, requiring a burn in time before the simulation can be used for scenario modifications, measurements etc.
All other inputs needed for initialisation are described below.

6. Inputs

Input variables

The upper-case names are the names used in the configuration file. See example below.

In addition to these simple parameter inputs the file RabbitWarrenLocations.txt can be used to prevent calculation of warren locations at run time. See RABBIT_WARRENREWADLOCATIONS above.

7. Interconnections

The rabbit model relies on environmental data primarily from the Landscape and Weather classes. They are also dependent on interactions with the Farm and Crop classes for information on management. Pesticide use and uptake depends on the Pesticide class. The Rabbit_Population_Manager is descended from Population_Manager and performs the role of an auditor in the simulation. Many of the functions and behaviours needed to execute and ALMaSS model are maintained by the Population_Manager.

8. References