FoxNet User Guide V1.0.0 Fox Net

User Manual:

Open the PDF directly: View PDF .
Page Count: 45

Download
Open PDF In Browser	View PDF

FoxNet User Guide V1.0.0
Bronwyn Hradsky, Quantitative and Applied Ecology, University of Melbourne

Contents
1 About FoxNet
1.1 Overview . . . . . .
1.2 Components . . . . .
1.3 Temporal and spatial
1.4 Process overview and
1.5 Design concepts . . .
1.6 Initialisation . . . .

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

3
3
3
5
5
7
8

FoxNet
. . . . .
. . . . .
. . . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

10
10
10
11
11

3 Building a simple scenario from scratch
3.1 A fox population in a homogenous landscape
3.2 Monitoring your model outputs . . . . . . . .
3.3 Experiment . . . . . . . . . . . . . . . . . . .
3.4 Adding a baiting program . . . . . . . . . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

12
12
14
14
15

. . . . . . .
. . . . . . .
scales . . .
scheduling
. . . . . . .
. . . . . . .

2 Getting started
2.1 Installing NetLogo and
2.2 The user interface . .
2.3 The code . . . . . . .
2.4 Model demonstration .

opening
. . . . .
. . . . .
. . . . .

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

4 An example customisation
18
4.1 Glenelg - a built-in scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.2 Building the Glenelg scenario from scratch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5 Building a new customised scenario
5.1 A customised landscape layer . . . .
5.2 Customised region(s) of interest . . .
5.3 Customised bait layout . . . . . . . .
5.4 Customised survey transect . . . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

25
25
26
26
26

6 Running batch scripts
27
6.1 Within NetLogo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
6.2 From R using the RNetLogo package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7 Implementation verification
31
7.1 Effect of productivity on fox-family territories . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
7.2 Effect of productivity on fox density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7.3 Effect of baiting on fox mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
8 Submodels
34
8.1 Submodels used during model processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8.2 Submodels specific to model set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
9 Appendices
40
9.1 Customisable model parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
9.2 Week Conversion Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
10 References

45
1

Notation
The names of agent-sets are italicised. An agent-set is a group of the same entities within FoxNet.
Submodel names are in bold.
Parameter names are shown in code font.
“Input values” for a parameter are shown in “quotes”. “Input values” can be varied to customise the FoxNet
modelling framework to a scenario and explore key sensitivities.

This project is licensed under the GNU General Public License v3.0 - see the LICENSE.md file for details.
To cite this User Guide
Hradsky B, Kelly L, Robley A, Wintle B (2019) FoxNet: an individual-based modelling framework to support
red fox management Journal of Applied Ecology.
Software source
Data available via Zenodo http://doi.org/10.5281/zenodo.2572045 (Hradsky et al. 2019b). A current version
of FoxNet and the User Guide is also maintained at this location.
For further information or help with customising model code
Email hradskyb@unimelb.edu.au or bhradsky@gmail.com
The development of FoxNet was funded by the
• Australian Government’s National Environmental Science Program through the Threatened Species
Recovery Hub
• Victorian Government (Department of Environment, Land, Water and Planning)
• Parks Victoria
2

1

About FoxNet

1.1

Overview

FoxNet is a customisable, individual-based modelling framework for running spatially-explicit red fox
population models at a landscape-scale. It can be used to predict red fox population density, age structure,
composition and responses to management across an entire landscape or within customised region(s) of
interest.

1.2

Components

FoxNet has four main types of agent: habitat-cells, foxes, fox-families and bait-stations. Another type of
short-lived agent (vacancy) is used to execute territory-checking processes efficiently.
Habitat-cells are the squares of habitat that define the spatial resolution, configuration and productivity of a
model landscape in FoxNet. Within a habitat-cell, productivity and fox access to bait-stations is homogenous.
The area represented by each habitat-cell can be specified using cell-dimension, and is usually in the order
of 0.01 km2 to provide a compromise between computational efficiency and intra-home-range variation (fox
home-range size varies between <0.5 and >9 km2 ; Trewhella et al. 1988). Each habitat-cell keeps track of
the following parameters:
• habitat-type: an integer that denotes the type of habitat (e.g. “0” = ocean, “1” = forest, “2” =
farmland). If the model landscape is generated within FoxNet, there will only be one habitat-type.
However, if the model landscape is imported as an ascii raster layer, the uninhabitable-raster-value,
second-habitat-raster-value and hab2:hab1 inputs can be used to specify, respectively, a raster
integer value that denotes habitat-cells that are not available to foxes (such as ocean or lakes), a raster
integer value that denotes a secondary type of habitat (e.g. farmland if the primary habitat type is
forest), and the productivity of the second habitat-type relative to the primary habitat-type. The
productivity of the habitat-cells determines how many habitat-cells a fox-family needs to acquire to
establish a territory (see below).
• available-to-foxes: “true” or “false”. This parameter defaults to “true” unless the habitat-cell’s
habitat-type is the same as the uninhabitable-raster-value for an imported landscape. In this case,
the habitat-cell becomes unavailable to foxes and is not included in the available-landscape-size.
This allows to you to specify areas that foxes cannot access, such as oceans, lakes, predator-proof fenced
areas etc.
• part-of-region-of-interest: “true” or “false”. This parameter defaults to “false”. When “true”, the
habitat-cell becomes part of the region-of-interest where the fox population is monitored (e.g. a
central square or a nature reserve). The region-of-interest can be the entire landscape if desired.
• part-of-region-of-interest2: “true” or “false”. This parameter defaults to “false”. When “true”,
the habitat-cell becomes part of a second region of interest (region-of-interest2) where the fox
population is also monitored (e.g. a second nature reserve).
• true-color: the intrinsic colour of the habitat-cell when unoccupied by a fox-family. This defaults to
black for the primary habitat type, but will be a different colour if the habitat-cell is the secondary
habitat type, is not available to foxes, or is part of the region(s)-of-interest.
• true-productivity: the intrinsic amount of food available to a fox in the habitat-cell during each
time-step (in grams time-step-1 ). The true-productivity of habitat-cells in the primary habitat
type is calculated from the size of an average fox home-range (input by the user) and the daily food
requirements of an adult fox (378 g/day; Lockie 1959); an approach similar to Carter et al. (2015)
but which facilitates scenario customisation. True-productivity in the secondary habitat-type (if
applicable) depends on the hab2:hab1 ratio. See the set-landscape-productivity submodel for more
information.
• current-productivity: the current amount of food available to a fox in the habitat-cell during each
time-step (in grams time-step-1 ). The current-productivity of each habitat-cell determines how many
3

habitat-cells foxes need to acquire to establish a territory - see the fox-families-check-territories
and foxes-disperse submodels. By default, a habitat-cell’s current-productivity is the same as
its true-productivity, but it can be varied spatially or temporally through manual coding. Future
extensions will use this feature to incorporate disturbance events such as fire that have short-term
effects on red fox habitat selection.
• fox-family-owner: the identity of the fox-family that currently owns the habitat-cell (set to “nobody”
when the cell is unowned).
• cell-relative-productivity: set to 0 by default. However, if the habitat-cell is owned by a fox-family,
it is calculated as the current-productivity of the habitat-cell weighted by the inverse of its distance
from the centre of the fox-family’s territory. This value is updated by the fox-family.
• cell-relative-use: set to 0 by default. However, if the habitat-cell is owned by a fox-family,
it is calculated as the intensity with which the habitat-cell is used by the fox-family, i.e., the
current-productivity of the habitat-cell divided by the total productivity of the fox-family’s territory. For example, in a homogeneous landscape with a territory-size of 100 ha and 1-ha habitat-cells,
cell-relative use would be 0.01. However, if the territory-size was 500 ha, cell-relative-use
would be 0.002. cell-relative-use is used to scale the exposure of foxes to bait-stations with
territory-size and habitat-cell productivity, and to derive the density of foxes (a monitoring output).
• cell-relative-use-foxes: set to 0 by default. However, if the habitat-cell is owned by a fox-family, it
is calculated as cell-relative-use multiplied by the number of foxes in the fox-family. This is used to
calculate the density of foxes. For example, if 4 foxes share the 500-ha territory and cell-relative-use
is 0.002, cell-relative-use-foxes is 0.222 x 4 = 0.008.
• checked-already: “no” (default) or “yes”: indicates whether the fox-family owner has already considered discarding this habitat-cell. Used to speed up the update-territory submodel.
Foxes are mobile individuals whose behaviour is determined by their age and status, and the time of year
(Larivière & Pasitschniak-Arts 1996). Each “alpha” fox is a member of a fox-family . Foxes have the following
characteristics:
• age: in weeks. Foxes are born at age “0”.
• sex: “female” or “male”.
• status: “cub”, “subordinate”, “disperser” or “alpha”, depending on the fox’s age and territory-holding
status.
• natal-id: the fox-family that the fox was born into.
• natal-cell: the habitat-cell where the fox was born.
• family-id: the fox-family that the fox currently belongs to.
• my-dispersal-distance: the distance (km) that the fox intends to move from its natal-cell (drawn
from a random-exponential distribution which is influenced by the fox’s sex and size of its natal territory)
- see the foxes-disperse submodel.
• distance-from-natal: the current distance of the fox from its natal-cell (in kilometres).
• my-dispersal-duration: the length of time the fox has been attempting to join or establish a new
territory (in weeks).
• failed-territory-id: a list of this fox’s territories which have failed.
Fox-families are used to establish and update the territories of their family-members (foxes within a family
share a semi-exclusive territory; Sargeant 1972). A fox-family must contain at least one “alpha” fox, and may
also include the alpha’s mate, cubs and subordinate offspring. Fox-families have the following characteristics:
• family-members - an agent-set comprising the “alpha female” and/or “alpha male” fox, and any of
their “cub” or “subordinate” offspring that have not yet dispersed.

4

• my-territory - an agent-set comprising the habitat-cells held as territory by the fox-family.
• territory-productivity: the total productivity of the fox-family’s territory.
• vacancy-score: the sum of the relative-productivity values of the vacancy agents surrounding
the territory at the end of the last time-step. Fox-families only try to improve their territory if the
vacancy-score this time-step is different - this helps the model run more efficiently. The vacancy-score
will change if the current-productivity of the habitat-cells has changed, if new habitat-cells have
become available through the death of a fox-family, or if location of the fox-family has changed. See
the fox-families-check-territories submodel.
Bait-stations are static agents that mark the locations of baiting sites (this is more efficient than using the
habitat-cells to record this). Bait-stations track:
• bait-present - whether a bait is currently available (“true” or “false”).
• Pr-death-bait-scaled - the likelihood of the bait affecting a fox whose territory overlaps it.
Vacancies are temporary static agents. They are briefly created and used by fox-families to (a) identify any
habitat-cells that are adjacent to their territory and not owned by another fox-family, and (b) determine
the relative-productivity of each of these available habitat-cells (i.e. the current-productivity of the
habitat-cell weighted by the inverse of its distance from the centre of the fox-family’s territory). The sum of
these values becomes the fox-family’s vacancy-score for that time-step. Vacancies are removed at the end
of each fox-family’s territory-checking procedure and are simply used to help the model run efficiently.

1.3

Temporal and spatial scales

FoxNet progresses in time-steps (or ticks) of 1, 2 or 4 weeks, depending on the weeks-per-timestep setting.
There are 52 weeks per year (and therefore 13 ‘months’, each of 4-weeks duration). A series of processes occur
consecutively each time step, and key seasonal events are linked to week-of-year, making the framework
customisable to northern- and southern-hemisphere scenarios.
The landscape can either be generated within FoxNet (as a square with the area specified by the
landscape-size input parameter), or uploaded as raster layer, with each cell of the raster corresponding to
a habitat-cell in the model. By default, the landscape doesn’t wrap (i.e. boundaries are non-permeable).

1.4

Process overview and scheduling

Each timestep, FoxNet works through a series of processes in order. Because key seasonal events are linked
to week-of-year, the framework is customisable to northern- and southern-hemisphere scenarios. There is no
hierarchy among agents of the same type (i.e. the order in which agents conduct each process is random and
varies each time-step).
The key processes are as follows:
(1) Year and week counters are updated (by 1, 2 or 4 weeks, as required). Key seasonal events are linked
to week-of-year, making FoxNet customisable to northern- and southern-hemisphere scenarios.
(2) The age of each fox is updated by the appropriate number of weeks. “Cub” foxes become “subordinates”
if they have reached the age-of-independence. If it is the dispersal season, “subordinate male”
and “subordinate female” foxes have user-specified probabilities of becoming “dispersers”. See the
update-fox-age-and-status submodel.
(3) Natal succession occurs within fox-families. That is, if a fox-family is missing an “alpha” fox, one of
the family’s “subordinates” that is the right sex and at least 24 weeks old becomes the “alpha” (as
described by Baker et al. 1998).

5

(4) Fox-families check their territories. Fox-families seek to maximise their territory-productivity
within a compact area by acquiring or replacing habitat-cells. This enables the fox-family to take over
unoccupied productive habitat-cells (Sargeant 1972), and respond to changes in resource availability
(Bino et al. 2010; Hradsky et al. 2017b). If the territory-productivity of fox-family’s territory
is less than an adult fox’s minimum food requirements (<295 g/day; Winstanley et al. 2003), the
fox-family fails, causing all adult foxes in the family to become “dispersers” and any “cubs” to die. See
the fox-families-check-territories submodel.
(5) If applicable, baits are laid at bait-stations and the cost of the baiting session is calculated. Foxes
belonging to a fox-family with a territory that overlaps an active bait-station are at risk of dying. Risk
scales directly with the number of bait-stations and bait efficacy, and inversely with territory size and
the number of foxes in the familiy. Each bait-station only affects one fox each time-step. See the

6

bait-if-applicable submodel.
(6) Foxes who have just become “dispersers” leave their natal fox-family and move a random distance
generated from a sex-specific exponential distribution, scaled by their territory-size (Trewhella et al.
1988). “Disperser” foxes then explore the area within their territory-perception-radius (set to
an area three times the radius of an average home-range; Soulsbury et al. 2011) where they (1) are
exposed to any active bait-stations; (2) attempt to join a fox-family that lacks an “alpha” fox of the
appropriate sex; (3) try to establish a new fox-family. If unsuccessful, they remain a “disperser” until
the next time-step. See the foxes-disperse submodel.
(7) If it is the breeding season, fox-families that contain an “alpha” male and an “alpha” female fox breed,
producing a Poisson-distributed number of “cub” foxes. If an “alpha” fox is absent, all family-members
become dispersers and attempt to join other nearby fox-families, promoting the persistence of the
population at low densities. See the fox-families-breed submodel.
(8) Provided that include-fox-mortality is “on”, stochastic background mortality of foxes occurs, based
on their age. See the foxes-die submodel.
(9) “Cub” foxes belonging to fox-families without any adults die, reflecting their dependence on foodprovision (Baker et al. 1998). This allows baiting to affect reproductive success. Defunct fox-families
(those that have no family members) are removed from the model.
(10) Model outputs are updated (and plotted if specified). Outputs can include age structure, population structure, dispersal distances, density or number of neighbours of foxes within the region-of-interest(s),
the number of foxes with territories overlapping a monitoring transect, and/or bait-take rates. See the
update-monitors submodel.
(11) If baits were deployed at bait-stations at step 5, any baits that were not eaten by foxes are removed to
mimic the removal of baits by managers or the degradation of the poison to non-toxic levels (Saunders
et al. 2000).
(12) The time-step (tick) counter increases by 1.
(13) The model checks if any foxes are alive. If all foxes are dead, the model stops.
Detailed descriptions of the submodels underlying these processes can be found in Chapter 8

1.5
1.5.1

Design concepts
Basic principles

The size of a fox-family’s territory depends on the productivity of the habitat-cells, and can be updated in
response to changes in the productivity or availability of these habitat-cell.
Foxes with established territories are only exposed to bait-stations within their territory. “Disperser” foxes
are exposed to bait-stations within their territory-perception-radius.
1.5.2

Emergence

The density of foxes, the distribution of ages and dispersal-distances within the population, and the configuration of fox-family territories emerge over time from demographic processes, territory dynamics and the
spatial configuration of different habitat-types.
Bait-take rates result from the configuration of active bait-stations relative to fox-family territories and the
relative-use of habitat-cells, fox density, and the likelihood of a fox consuming a bait.
1.5.3

Adaptation

Fox-families adapt the size and position of their territory to the productivity of the habitat-cells and the
presence of surrounding fox-families.

7

1.5.4

Objectives

Each fox-family aims to maximise the productivity of its territory (up to 110% of an adult fox’s daily food
requirements) within as compact an area as possible.
“Disperser” foxes seek to join an existing fox-family as an “alpha”, or establish their own fox-family.
Foxes seek to breed, and so will disband their fox-family if it lacks an “alpha” of the opposite sex during the
breeding period.
1.5.5

Sensing

When they are dispersing, foxes can sense the location of fox-family territories and unoccupied habitat-cells
within their territory-perception-radius.
Fox-families can sense the productivity of the habitat-cells within their territory. They can also sense the
productivity and availability of the habitat-cells surrounding their territory, via the vacancy agents.
1.5.6

Interaction

Fox-families indirectly compete for habitat-cells, as each habitat-cell can only be used by one fox-family. The
fox-families use vacancy agents to determine which habitat-cells surrounding their territory are currently
available. This effectively limits the carrying-capacity of the landscape, as only foxes within fox-families can
breed. See Effect of productivity on fox density.
1.5.7

Stochasticity

Stochasticity occurs throughout the model to represent natural variation, including the dispersal locations of
foxes, fox sex and survival rates, and fox-family territory-acquisition and litter size.
1.5.8

Collectives

The region-of-interest(s) is nominated by the user, and defines the area(s) in which foxes are monitored.
Foxes form family-groups (represented by a fox-family agent) which share a territory. The status of each fox
is influenced by the sex and status of its other family-members.
The survival of a “cub” fox depends on the presence of at least one adult member in its fox-family until it
reaches the age-of-independence.
1.5.9

Observation

At the end of each time-step, data can be collected on the foxes, fox-families and bait-stations. The user
specifies which data to collect and plot. See the update-monitors submodel.

1.6

Initialisation

FoxNet initialises by setting the input parameters as per the model interface, and then running through the
following processes in order:
1. A random-seed is set and recorded, so that the model run can be reproduced exactly if required.
2. set-current-directory. The current directory is set to the working-directory so that any spatial
layers can be easily imported and model outputs will be saved in the appropriate folder.
3. check-for-errors. This checks for inconsistencies in the input parameters. If an error is found, the
model returns a detailed error message and stops. See the check-for-errors submodel.
4. calculate-conversion-factors. This calculates the factors for converting between input units and the
model (e.g. km2 to habitat-cells).

8

5. set-fox-parameters. This sets a variety of parameters which are not available on the interface and
derives others, such as the duration of the dispersal season. See the set-fox-parameters submodel.
6. create-world. This generates a landscape of an appropriate size, either by generating it within the
model or importing a raster layer. See the create-world submodel.
7. identify-region-of-interest. This identifies the part of the landscape in which fox populations
will be monitored. This may be the entire landscape or a subset group of habitat-cells. See the
identify-region-of-interest submodel.
8. set-landscape-productivity. This calculates the distribution of productivity values across the habitatcells using the home-range-area and kernel-percent inputs. See the set-landscape-productivity
submodel.
9. set-up-bait-stations. This sets up the bait-stations across the landscape (if applicable). See the
set-up-bait-stations submodel.
10. set-up-survey-transect. This sets up a survey transect for monitoring the number of foxes, if
applicable.
11. set-up-foxes. This creates an initial number of foxes at the specified density. Each fox immediately
disperses and tries to join an existing fox-family or establish a new fox-family. See the set-up-foxes
submodel.
12. update-monitors. This calculates the model ouputs and plots them if applicable. See the updatemonitors submodel.
13. Finally, the model resets the time-step counter.

9

2

Getting started

2.1

Installing NetLogo and opening FoxNet

FoxNet was built in NetLogo (Wilensky 1999) version 6.0.2 - an open-source modelling environment, downloadable from https://ccl.northwestern.edu/netlogo/. FoxNet is saved in within the FoxNet folder: FoxNet_model
/ FoxNet.nlogo. You can open it from within NetLogo or by double-clicking the file name.
For an introduction to using NetLogo see https://ccl.northwestern.edu/netlogo/docs/, particularly the three
tutorials listed in the LH menu.

2.2

The user interface

When you first open FoxNet, you should see the graphic user interface. The interface has four key sections:
1. Inputs (green) - these boxes and sliders allow you to set model parameters and customise FoxNet to
your scenario.
2. Buttons (mauve) - these are used to control FoxNet:
• setup will parameterise FoxNet with whatever settings you have chosen in the inputs.
• go will run FoxNet for one timestep.
• go (continuous) will run FoxNet forever (or at least until you click the button again) - this is shown
by the circling arrows.
• territory-demo will provide a demonstration of key model processes - see Model demonstration.
• basic-scenario will setup FoxNet with a simple example scenario - see Building a simple scenario from
scratch.
• glenelg-scenario will setup FoxNet with a complex scenario to demonstrate all the bells and whistles
(see An example customisation). You must set the working-directory to the location of the FoxNet
folder to use this option.
3. World-view (large black square) - this depicts some aspects of the model so that you can watch what is
happening as you run FoxNet.
4. Outputs (beige boxes and plots) - these enable you to track key model outputs.

10

At the top of the screen, there are additional controls that affect how FoxNet displays, including a slider that
controls the model speed, a tick-box that specifies whether you want to watch updates in the world-view or
not, and a drop-down option for whether updates should be shown continuously or just at the end of each
time-step.
Under the Settings button, you can adjust the number of pixels per cell in the world-view and the frame rate.
TIP: To make the world-view larger/smaller you can increase/decrease the number of pixels per
cell. If the world-view is too large, you may encounter memory constraints.

At the bottom of the screen is the command center where you can issue direct instructions to the model.
Information on what is happening during the territory-demo is also shown here.

2.3

The code

The underlying model code can be viewed by clicking across to the Code tab. Submodels are provided in
separate files, which can be viewed by clicking on the Included Files dropdown box, within the Code tab.
Ignore this for now - you don’t need to view or edit the code to build or run a model in FoxNet.

2.4

Model demonstration

FoxNet comes with a build-in demonstration of several key processes, including territory formation, fox
responses to changes in habitat productivity, mate-seeking and reproduction.
Click on territory-demo (mauve button). A description of what is happening will appear in the Command
Center bar at the bottom of the screen. You can expand this bar if you want to see more than one line of
writing at once.
TIP: To stop the demonstration (or any FoxNet model) part-way through, click Tools > Halt.

11

3

Building a simple scenario from scratch

3.1

A fox population in a homogenous landscape

Let’s start by building a simple landscape. Using the Interface tab, either:
• Set the following inputs manually by moving the sliders and entering the values, and then click setup, OR
• Click the mauve basic-scenario button (which will automatically set up the model with the same
input parameters)
Parameter
Landscape setup
‘weeks-per-timestep‘
‘cell-dimension (m)‘
‘landscape-source‘
‘landscape-size (km^2^)‘
‘region-size (km^2^)‘
Fox parameters
‘initial-fox-density (foxes km^-2^) ‘
‘range-calculation‘
‘home-range-area (km^2^)‘
‘kernel-percent (%)‘
‘fox-mortality‘
‘less1y-survival (propn.)‘
‘from1yto2y-survival (propn.)‘
‘from2yto3y-survival (propn.)‘
‘more3y-survival (propn.)‘
‘breeding-season (week)‘
‘number-of-cubs‘
‘propn-cubs-female (propn.)‘
‘age-at-independence (weeks)‘
‘dispersal-season-begins (week)‘
‘dispersal-season-ends (week)‘
‘female-dispersers (propn.)‘
‘male-dispersers (propn.)‘
Baiting parameters
‘bait-layout‘
Monitoring parameters
‘plot? ‘
‘density‘

Value
1
100
generate
400
110
0
1 kernel, 1 mean
[0.454] **make sure to include the brackets
[95] **make sure to include the brackets
on
0.48
0.54
0.53
0.51
13
4.72
0.5
12
37
9
0.378
0.758
none
"on"
"on"

These parameters are based on a fox population in Bristol, United Kingdom, with survival data from
Devenish-Nelson et al. (2013), and breeding and dispersal data from Trewhella and Harris (1988). See
Hradsky et al. (2019a) Table S1 for a full list of citations.
Whichever approach you chose, you should now see a black landscape with a central grey square in the
world-view. The entire landscape is 400 km2 . The central grey square is your ‘region-of-interest’ where the
fox population is monitored, and is 110.25 km2 1 . These values are shown in the beige output boxes.
1 Because FoxNet is generating a square lnadscape with 1 ha pixels (as specified by cell-dimension), it has had to adjust the
size of the region-of-interest slightly

12

Now change initial-fox-density to “0.5” foxes km-2 by sliding or clicking the slider to the right.
TIP: You can also alter the value of a slider by right-clicking the slider, selecting “edit” and then
typing in a new value in the appropriate box.
Click setup again. A scattering of white and yellow triangles should appear - these are “female” and “male”
foxes without territories (i.e. their status is “disperser”). Each fox should attempt to establish or join a
circular brown territory. A fox with a territory acquires “alpha” status, becomes red (“female”) or blue
(“male”), and sits beneath an orange-coloured fox-family in the centre of its territory. A “female” and “male”
fox can share a territory.

TIP: If you would like to slow this process down so that you can see what is happening, try sliding
the speed toggle in the grey header bar to about 25% speed and click setup again.
If you’d like to see the foxes more clearly, right-click the world-view > Edit, and change the patch-size to 2
pixels.
Now click go. You might see a fox or two move, but not much else will change. However, the beige year and
week of year output boxes at the top of the screen will now show year 1, week 1.
Click go (continuous). As time progresses, foxes breed, disperse, establish new territories and die.
You can observe the density of foxes and fox-families within your 110 km2 region-of-interest by looking

13

at the plot to the right of the world-view. For example, by Year 5, week 13, the density of foxes has just
spiked with the fifth breeding season, because the cub-birth-season is currently set to week 13 of the year
(late March - see the Week Conversion Table). The density of fox-families remains relatively stable:

TIP: To export the fox density data, right-click the plot and select “export”. End the filename
with “.csv”.
Click setup and go (continuous) again. Each time you run the model with the same settings, you will get
slightly different results due to stochastic variation.

3.2

Monitoring your model outputs

The basic-scenario defaults to plotting the density of foxes and fox-families within the region-of-interest.
However, you can monitor other aspects of the fox population within the region-of-interest using the
switches under MONITORING on the FoxNet interface. Options include:
• “age-structure” - the proportion of foxes in each age class from 0 to 8+ years.
• “bait-consumption” - the proportion of baits that have been removed by foxes per time-step.
• “count-neighbours” - the mean (min, max) number of territories directly abutting each fox-family’s
territory.
• “density” - the density of foxes and fox-families.
• “dispersal distances” - the distance moved by foxes that have attempted to disperse.
• “family-density” - the density of fox-families.
• “popn-structure” - the proportion of foxes in each demographic group.
• “foxes-on-transect - the number of foxes who have territories that overlap the survey-transect. This
only works for landscapes that have been imported as a raster and where a survey transect shapefile
has been uploaded.
• “home range size” - not provided in this version of FoxNet.
You can choose whether or not to plot summary values using the plot? button. Plotting makes the model
run more slowly.

3.3

Experiment

Have a play with the input parameters. For example, try making landscape-size larger, initial-fox-density
higher, home-range-area bigger or switch fox-mortality to “off”.
TIP: Huge landscapes with high densities of foxes may take a very long time to setup and run.
To cancel a model at any point, select Tools > halt.
Note that all week-related inputs must be consistent with weeks-per-timestep. FoxNet checks this during
model set-up and will return an error if, for example, weeks-per-timestep is “4” but cub-birth-season is
set to week “6”.

14

3.4

Adding a baiting program

Return to the original settings by clicking basic-scenario.
Then alter the baiting parameters as follows to apply a uniform grid of bait-stations across your model
landscape:
Parameter
Baiting parameters
‘bait-layout‘
‘bait-density (km^-2^)‘
‘bait-frequency‘
‘Pr-die-if-exposed-100ha (propn.)‘
‘commence-baiting-year‘
‘commence-baiting-week‘

Value
grid
1
4-weeks
0.3
3
13

Click setup to see the layout of bait-stations.
The current setting will have established a 1 km-2 grid of 400 bait-stations across your landscape.

If initial-fox-density is > “0” and you click “go”, baits will be deployed at bait-stations every 4 weeks.
Baits are deployed at bait-stations from the start of the model, but under the current settings only become
poisonous from Year 3 week 13 - this allows the population to begin to establish without baiting, and enables
you to quantify uptake of non-toxic baits (i.e. free-feeding).
TIP: Baits are only deployed at bait-stations if the initial-fox-density is greater than 0.
Setting Pr-die-if-exposed-100ha to “0.3” means that one fox with a 100-ha territory has a 30% chance of
consuming a bait if one bait is deployed on its territory (and dying if the bait is toxic). The risk to a fox
scales with the number of baits and inversely with the territory size and number of foxes in its fox-family.
See Effect of baiting on fox mortality
Try altering bait-density to “4.0” baits km-2 or changing bait-layout to “random-scatter”, and clicking
setup again.

15

Look at the effects of baiting on the population by changing the settings to a “grid” of baits with “2” baits
km-2 , setting initial-fox-density to “0.5” km-2 , turning “on” the bait-consumption monitor, and then
clicking setup and go (continuous).
Fox density declines rapidly in Year 3 after the baits become toxic, in this case, going extinct by Year 8 week
21. The model stops when all foxes are dead. Bait-take rates fluctuate with fox density.

TIP: By default, baits only remain at bait-stations for one time-step. You can change the duration
of a time-step to “1”, “2” or “4” weeks. Other durations require a customised code. See foxnet /
foxnet_model / foxnet_carnarvon_custombait.nlogo for an example.
Baits can be deployed at bait-stations at custom intervals (e.g. quarterly, only in winter or once per year), by
changing bait-frequency to “custom*" and adding the relevant week numbers to custom-bait-weeks.
TIP: Make sure that the week inputs are surrounded by square-brackets, are consistent with the
weeks-per-timestep parameter and do not include commas between values.

16

For example, the following will deploy baits in weeks 6, 19, 31 and 45 each year (i.e. the start of February,
May, August and November):
Parameter
Baiting parameters
‘bait-frequency‘
‘custom-bait-weeks‘

Value
custom*
[6 19 31 45]

To calculate the cumulative annual cost of a baiting regime (displayed in a beige box to the right of the
world-view), you can specify the price per bait deployed, the number of person-days it takes to deploy all
baits in a single bait deployment, the number of kilometres travelled per bait deployment, and the cost of
travel per kilometre. For example, if the parameters as set as follows. . .
Parameter
Baiting parameters
‘price-per-bait ($)‘
‘person-days-per-baiting-round (days)‘
‘cost-per-person-day ($)‘
‘km-per-baiting-round (km)‘
‘cost-per-km-travel ($)‘

Value
2.00
3.00
250.00
420.00
0.67

. . . a single bait deployment costs $2599, and the cost of deploying baits every 4 weeks for a year is $33,792.
TIP: The FoxNet interface constrains you to a baiting schedule that remains constant across years.
To vary the location of bait sites or baiting frequency between years, you will need to write a
customised baiting schedule. For an examplecode, open foxnet_carnarvon_custombait.nlogo,
click across to the Code tab, then Included Files and select bait_routines_carnarvon.nls.
Please get in contact with model developers if you need help with writing a customised schedule.

17

4

An example customisation

FoxNet models can be customised to real-world management landscapes using GIS layers that describe the landscape size and configuration of different habitat types, the area and location of the region(s)-of-interest,
survey-transect(s), and/or the layout of bait-stations.
Example spatial layers are provided so that you can explore the suite of customisation features in FoxNet
before developing your own.

4.1

Glenelg - a built-in scenario

To explore the built-in customised scenario:
1) Set the working-directory to wherever you have saved the FoxNet folder, for example:
C:/Users/hradsky/foxnet.
Take careful note of the direction of the slashes if you are using Windows.
2) Click Glenelg-scenario
The world-view should now be much larger, and show an imported landscape layer with three different types
of habitat-cells: forest (black), farmland (dark grey) and ocean (white, unavailable to foxes). Fox home
ranges are three times smaller in the farmland than the forest because of differences in the productivity of
the habitat-cells.
Within the landscape, there are two regions-of-interest for monitoring the fox population. The first
is mid-grey (to the south), and includes bait-stations (white squares) and a survey-transect (red line) for
monitoring the number of foxes. The second region-of-interest is lighter grey (to the north).

18

If you’d like to see the landscape more clearly, set initial-fox-density to 0 and click setup again.

3) Set initial-fox-density back to 0.5, and explore the effects of some of the other parameters . After
you change the parameters, you must click set-up again. For example, make the productivity of the
farmland the same as the productivity of the forest by changing the hab2:hab1 slider to “1.00”.
TIP: Don’t forget that you can make the model run more quickly by sliding the toggle at the top
of the screen to the right.

4.2

Building the Glenelg scenario from scratch

Now let’s work through each customisation process to build the Glenelg scenario from scratch
Once you are comfortable with how the Glenelg customisation scenario works, you can follow the same
processes to customise FoxNet to your own management scenario. Additional information on how to do this
is provided in the next chapter.
4.2.1

A customised landscape layer

Let’s start by importing an ascii raster layer to create a customised landscape of south-western Victoria,
Australia. This landscape is rectangular and contains three types of habitat-cell: forest, farmland and ocean.
Begin by clicking “basic-scenario” to reset all your parameters.
Alter the number of pixels per cell but clicking on the worldview > Edit. . . > Patch size > 0.5.
Then adjust the following parameters and click “setup”:

19

Parameter
‘working-directory‘
Landscape setup
‘cell-dimension (m)‘
‘landscape-source‘
‘landscape-raster (.asc)‘
‘uninhabitable-raster-value ‘
‘second-habitat-raster-value‘
‘hab2:hab1‘
Fox parameters
‘initial-fox-density (km^-2^)‘
Baiting parameters
‘bait-layout‘

Value
Location of FoxNet folder, e.g. C:/Users/hradsky/foxnet
100
import raster
gis_layers/glenelg/mtclay_landscape.asc
2
0
1.00 x
0
none

The worldview should now look like this:

Forest (the primary habitat type, which has a value of “1” in the raster) is shown in black, farmland (the
second habitat type, with raster value ‘0’) is shown in dark grey, and the ocean (specified by raster value ‘2’)
appears white.
TIP. You can check these values by right-clicking a habitat-cell, selecting inspect patch x y and
then seeing the integer shown next to habitat-type.
The region-of-interest has still been generated within the model and so forms a 200 km2 lighter-shaded
block in the centre of the landscape.

20

At the top of the interface, the output box actual landscape size shows the area habitable by foxes
(4970.79 km2 ).
The total size of the landscape including the ocean is 6260.46 km2 . To find this out, type count patches *
cells-to-km2 into the command center at the bottom of the screen.
4.2.2

Add some foxes

Provided the intial-fox-density is large enough that the population doesn’t go extinct through stochastic
processes, the fox density will eventually stabilise at the landscape’s carrying capacity (see Effect of productivity
on fox density ).
The productivity of the landscape is set by specifying an average home-range-area for foxes in this landscape
(in km2 ), and the percentage of the home range kernel that this comprises.
For example, foxes in the forested regions of south-western Victoria have an average 95% kernel of 2.14 km2 .
Specify the following parameters, then click setup again.
Parameter
Fox parameters
‘initial-fox-density (km^-2^)‘
‘home-range-area (km^2^)‘
‘kernel-percent (%)‘
‘less1y-survival‘
‘from1yto2y-survival‘
‘from2yto3y-survival‘
‘more3y-survival‘
‘breeding-season‘
‘number-of-cubs‘
‘propn-cubs-female‘
‘age-at-independence‘
‘dispersal-season-begins‘
‘dispersal-season-ends‘
‘female-dispersers‘
‘male-dispersers‘

21

Value
0.5
[2.14]
[95]
0.39
0.65
0.92
0.18
37
3.2
0.5
12
9
21
0.7
0.999

If forest and farmland habitats have equal productivity, a landscape with an initial-fox-density of 0.5 km-2
looks like this after setup:

If different habitat types support different densities of foxes (Sálek, Drahnikova & Tkadlec 2015), you can
specify this using the hab2:hab1 input slider. For example, farmland (the second habitat type) might be
three times more productive than forest (the primary habitat type), and so the slider would be set to “3.00”.
This causes foxes to select for farmland over forest. Territories entirely within the farmland are a third of the
size of those entirely within the forest (0.75 km2 vs. 2.26 km2 , 100% kernel), and the landscape’s carrying
capacity increases.

TIP To find out the size of the smallest and largest fox-family territories, type min [count
my-territory] of fox-families * cells-to-km2
or
max [count my-territory] of
fox-families * cells-to-km2 into the command center at the bottom of the screen

22

4.2.3

Add two monitoring regions

Rather than monitoring foxes in a central, square region-of-interest, we can monitor fox density in one or two
customised regions, in this case: Mt Clay Nature Reserve (where fox baiting occurs) and Annya State Forest
(where baiting does not occur). These regions are specified as separate polygon shapefiles:
Parameter
Value
Landscape setup
‘region-shp‘
gis_layers/glenelg/mtclay_region.shp
‘region2-shp‘ gis_layers/glenelg/annya_region.shp
The first region-of-interest is Mt Clay Nature Reserve (48.29 km2 , as shown in the output box).
The second region of interest (region-of-interest2) is Annya State Forest. The size of this region is not
automatically displayed but can be shown by typing “region-of-interest2-size” into the observer input pane at
the bottom of the screen. The value is in km2 .

4.2.4

Add customised bait sites

To import a customised layout of bait-stations, use a shapefile with a point for each bait-station location.
Parameter
Baiting parameters
‘bait-layout‘
‘bait-layout-shp‘
‘bait-frequency‘
‘Pr-die-if-exposed-100-ha‘

Value
custom
gis_layers/glenelg/mtclay_baits.shp
4-weeks
0.3

In this case the bait-stations are entirely within the first region-of-interest, however, this isn’t required bait-stations can cover the entire landscape if preferred.

23

4.2.5

Add a survey transect

We can also calculate the number of foxes that have territories which overlap a linear transect, if for example,
we intend to collect fox scats and identify individuals using DNA analysis. The transect can be imported as a
line shapefile.
Parameter
Landscape setup
‘survey-transect‘

Value
gis_layers/glenelg/mtclay_transect.shp

initial-fox-density has been set to 0 to make the transect easier to see.

24

5

Building a new customised scenario

5.1
5.1.1

A customised landscape layer
Preparing your ascii file

The underlying landscape in FoxNet can be customised using a raster in ascii format (.asc). This allows you
to depict, for example, a rectangular landscape, different habitat types that support different densities of
foxes, and/or areas that are inaccessible to foxes such as oceans, lakes or fenced reserves.
You can prepare your ascii file in any spatial software you like (e.g. ArcGIS, QGIS or R).
The usual procedure involves drawing a rectangle for your landscape, dividing it into polygons that represent
each habitat type, and editing the attribute table so that each habitat type has a different number (e.g. 1 =
ocean, 2 = forest, 3 = farmland). You then convert the polygon to a raster, and the raster to an ascii.
The size of your landscape should reflect the likely dispersal distance of foxes into your management zone.
A simple calculator can be found in foxnet/user_guide/Calculator_dispersaldistance.xlsx. You can
use this to estimate the proportion of potential dispersers that would be captured by a given buffer size,
based on the relationship between home range size and dispersal distance derived by Trewhella et al. (1988).
Take care with your projection file!
The .asc file must be accompanied by a projection file (.prj), which has the same filename and is saved in the
same folder. Your spatial software will automatically generate this. However,for NetLogo to accept it, the
.prj file must show the projection in one continuous line, like this:

PROJCS["WGS_1984_UTM_Zone_54S",GEOGCS["GCS_WGS_1984",DATUM["D_WGS_1984",SPHEROID["WGS_1984",6378137
You can open the .prj file in Notepad (or similar) to check.
If you have generated your ascii layer in ArcGIS, the projection layer might be in the wrong format, and
instead look like this:
Projection
Zone
Datum
Spheroid
Units
Zunits
Yshift
Parameters

UTM
54
WGS84
WGS84
METERS
NO
10000000.0

This will not work! The easiest way to fix this problem is to open the .prj file for a shapefile that uses the
same projection as your raster, and copy and paste the contents over the .prj information for your ascii.

A list of acceptable projection coordinate systems can be found at  BehaviourSpace.
You can then setup a new experiment or run an existing script.

Information on how to use BehaviourSpace can be found at 156 weeks), and is given by:
age.specif ic.annual.probability.of.survivalweeks.per.tick/52
8.1.7

update-monitors

The parameters that are calculated depend on the monitoring options selected on the interface - see Monitoring
your model outputs and Customisable model parameters for descriptions of each parameter. Outputs are
plotted if plot? is set to “on”.
8.1.8

move-to-centroid

To move to the centre of its territory, the fox-family creates a list of the x and y coordinates of each of the
habitat-cells in its territory, and moves to the habitat-cell with the mean x and y coordinates. It then asks
its family-members to move to this location. Finally, it updates the cell-relative-productivity of the
habitat-cells in its territory and the relative-productivity of any vacancies to:
1
× current.productivity × 100
distance.f rom.f ox.f amily
8.1.9

update-territory

With each repeat, the fox-family checks whether the productivity of its territory is below the average
requirements of an adult fox. If so, it will try to add new habitat-cells to its territory. This involves
acquiring-the-best-available-vacancy by:
(1) Identifying the vacancy that has the highest relative-productivity and adding the habitat-cell that
the vacancy is sitting on to the fox-family’s territory.
(2) Asking the neighbours of this habitat-cell that aren’t currently owned by a fox-family or don’t already
have a vacancy on them, to create a vacancy agent.
(3) Updating the productivity score for the fox-family’s territory.
(4) Moving to the centre of the fox-family’s territory and updating the relative-productivity of the
habitat-cells and vacancies (see move-to-centroid).
36

If the productivity of its territory is already adequate, the fox-family won’t expand its territory. Rather, it
will try to swap-poor-territory-for-better. This involves identifying which habitat-cell in its territory has
the lowest cell-relative-productivity. If the habitat-cell has already been checked, the routine aborts
(this speeds up the model). If not:
(5) The habitat-cell’s checked-already parameter is updated to “true”.
(6) The vacancy agent on the unoccupied habitat-cell neighbouring the fox-family’s territory with the
highest relative-productivity is identified.
(7) If the best vacancy has a higher relative-productivity than the worst territory cell (and isn’t a
neighbour of that cell), the fox-family will:
• Remove the worst habitat-cell from its territory.
• Update its territory-productivity score.
• Add the best vacancy cell to its territory, following the acquiring-the-best-available-vacancy
routine described above.
8.1.10

territory-fail

A fox-family will fail if the total productivity of its territory is less than the minimum required to sustain
an adult fox’s metabolism. In this case, the fox-family-owner of the habitat-cells in its territory is set
to “nobody”, any “cub” foxes die and the status of all other family-members is reset to “disperser”. The
fox-family and all associated vacancies are removed from the model.

8.2
8.2.1

Submodels specific to model set-up
check-for-errors

This checks for inconsistencies in the input parameters. The model will return a detailed error message and
stops if:
• bait-frequency is set to “weekly” but the weeks-per-timestep input is greater than “1”.
• bait-frequency is set to “fortnightly but the weeks-per-timestep input is greater than”2“.
• bait-frequency is set to “custom” but the custom-bait-weeks values are incompatible with the
weeks-per-timestep input (i.e. if (value - 1) MOD weeks-per-timestep != 0).
• the dispersal-season-begins or dispersal-season-ends values are incompatible the weeks-per-timestep
input.
• dispersal-season-begins and dispersal-season-ends are the same value.
• the cub-birth-season value is incompatible with the weeks-per-timestep input.
• you have asked the model to monitor the number of foxes on a transect without uploading a transect
• you have imported a shapefile for region2-shp without importing a shapefile for region-shp
• you are trying to use a customised bait layout in a model-generated (rather than imported) landscape.
8.2.2

set-fox-parameters

This sets a variety of parameters which are not available on the interface, including the minimum amount of
food required for short-term sustenance of an adult fox’s metabolic rate - 295 g day-1 (Winstanley et al. 2003)
and a fox’s average daily food requirements - 378 g day-1 (Lockie 1959), see also (Sargeant 1978; Winstanley
et al. 2003).
It also calculates several derived parameters:

37

• maximum-territory-update-area: The maximum area of territory that can be updated each time-step
- half the area of a home range: hr.100perc × 0.5
• territory-perception-radius: the distance
that a fox can perceive adjacent fox-families. Three
p
times the radius of a home range: 3 × hr.100perc)/π
• dispersal-duration: The duration of the dispersal season:
If dispersal.season.ends > dispersal.season.begins (i.e.southern hemisphere scenarios):
dispersal.duration = dispersal.season.ends−dispersal.season.begins
weeks.per.tick
Else (i.e. northern hemisphere scenarios): dispersal.duration =
8.2.3

dispersal.season.ends+(52−dispersal.season.begins)
weeks.per.tick

create-world

This generates a landscape of an appropriate size.
If landscape-source = “generate”, the landscape is generated by the model as a square with an area as
close to world-size as possible.
If landscape-source = “input raster”, the landscape is imported at a raster layer (ascii file), and will
automatically re-size to the appropriate dimensions. In the latter case, the habitat-type of each habitat-cell
will be determined by the raster value.
Habitat-types that are unavailable to all agents (e.g. ocean) can be specified using the uninhabitable-raster-value.
These areas are not included in the landscape size output.
For both options, the spatial resolution is set by cell-dimension; this must correspond to the size of the
raster cells if applicable.
8.2.4

identify-region-of-interest

This identifies the part of the landscape in which fox populations will be monitored. This may be the entire
landscape, or a smaller area to reduce any potential edge effects. The region-of-interest can be specified
using the region-size input (which creates a central square of appropriate size), or imported as a shapefile.
This changes the part-of-region-of-interest value of habitat-cells to “true”.
8.2.5

set-landscape-productivity

This determines the true-productivity values for the habitat-cells across the landscape using the input
data on fox home-range-area. The productivity of each habitat-cell in the primary habitat-type is calculated
from the home-range-area and kernel-percent of an average fox for that location (input by the user) and
the daily food requirements of an adult fox (378 g/day; Lockie 1959); an approach similar to Carter et al.
(2015). The calculations are based on the assumption that the intensity of use of an area by a fox corresponds
to the amount of food it obtains from the area. For example, if an average adult red fox requires ~370 g of
fresh food a day (Lockie 1959; Winstanley et al. 2003) and 100 % of its range covers 100 ha, 1-ha habitat-cell
would be supplying 3.7 g food day-1 on average.
If range-calculation is set to “1 kernel, 1 mean”, a single estimate of home-range-area and
kernel-percent (e.g. mean 95% MCP) is used to create a homogeneous landscape, with the productivity of
each cell during each time-step being:
adult.f ox.timestep.f ood × proportion.of.range
home.range.area × area.of.habitat.cell
If range-calculation is set to “1 kernel, min and max”, minimum and maximum estimates of
home-range-area (and one kernel-percent, e.g. mean 95% MCP) are used to create a heterogeneous
landscape, with the productivity of each cell during each time-step being a random value drawn between the
minimum and maximum productivity values, each calculated as above.

38

Whichever approach is used, the productivity of the second habitat type is specified relative to this primary
habitat type, according to the hab2:hab1 ratio.
8.2.6

set-up-bait-stations

The locations of bait-stations can be specified by choosing a bait-layout: “none”, “regular grid”, “random
scatter” or “custom”.
If “regular grid” or “random scatter” is chosen, bait-stations are generated by the model at the specified
density. If “custom”, the locations for the bait-stations are imported as a shapefile. The bait-present
attribute of all bait-stations is set to “false” (i.e. no bait is present).
8.2.7

set-up-foxes

An initial number of foxes are created at the specified density, with ages ranging between 36 and 166 weeks,
and a sex ratio in accordance with the propn-cubs-female input. Each fox is moved to a random location
and its status is set to “disperser”. All foxes then disperse (see the foxes-disperse submodel).

39

9

Appendices

9.1

Customisable model parameters

Parameter
‘working-directory‘

Unit
—

Explanation
Location of FoxNet folder, e.g.
‘C:/Users/hradsky/FoxNet‘

Landscape setup
‘weeks-per-timestep‘

weeks

Number of weeks for each model
timestep ("1", "2" or "4")
Distance along one edge of a
*habitat-cell*
Whether the landscape is
generated within the model or
imported as a raster
The size of the landscape, if it is
generated within the model
The size of the region-of-interest
(where the *fox* population is
monitored), if it is generated
within the model
The raster (ascii) file describing
the landscape layout (e.g.
‘GIS_layers/Glenelg/
mtclay_landscape.asc‘), if the
landscape is imported
The value used in the raster layer
to specify *habitat-cells* that
can’t be inhabited by *foxes*
(e.g. ocean)
The value used in the raster layer
to specify *habitat-cells* that are
a secondary type of habitat (e.g.
farmland rather than forest)
The productivity of the
secondary habitat type relative
to the primary habitat type
A shapefile (polygon) describing
a region-of-interest (where *fox*
populations are monitored) if the
landscape is imported
A shapefile (polygon) describing
a second region-of-interest (where
*fox* populations are monitored)
if the landscape is imported –
optional.
A shapefile (line) describing a
transect for surveying *fox*
populations
A shapefile (line) describing a
second transect for surveying
*fox* populations

‘cell-dimension‘

m

‘landscape-source‘

“generate” or “import raster”

‘landscape-size‘

km^2^

‘region-size‘

km^2^

‘landscape-raster‘

.asc file

‘uninhabitable-raster-value‘

integer

‘second-habitat-raster-value‘

integer

‘hab2:hab1‘

ratio

‘region-shp‘

.shp file

‘region2-shp‘

.shp file

‘survey-transect-shp‘

.shp file

‘survey-transect2-shp‘

.shp file

Fox parameters
‘initial-fox-density‘

individuals km^-2^

40

The number of *foxes* per
square kilometre when the model
is initiated

(continued)
Parameter
‘range-calculation‘

Unit
‘ “1 kernel, 1 mean” ‘ or “1
kernel, min and max”

‘home-range-area‘

km^2^

‘kernel-percent‘

%

‘fox-mortality‘

"true"/"false"

‘less1-survival‘

propn.

‘from1yto2y-survival‘

propn.

‘from2yto3y-survival‘

propn.

‘more3y-survival‘

propn.

‘cub-birth-season‘

week of year

‘number-of-cubs‘

cubs fox-family^-1^

‘propn-cubs-female‘

propn

‘age-at-independence‘

weeks

‘dispersal-season-begins‘
‘dispersal-season-ends‘
‘female-dispersers‘
‘male-dispersers‘
Baiting parameters
‘bait-layout‘

week of year
week of year
propn.
propn.
“none”, “grid”, “random-scatter”,
“custom”

‘bait-density‘

baits-stations km^-2^

‘bait-layout-shp‘

.shp file

‘bait-frequency‘

“weekly*”, “fortnightly*,
“4-weeks” or “custom*”

‘custom-bait-weeks‘

week(s) of year

41

Explanation
The method used for calculating
the productivity of the landscape
from the fox home range data
The area of an average fox home
range in this landscape
The proportion of the home
range kernel included in the %
area (e.g. 90 for a 90% MCP)
Whether natural mortality of
*foxes* occurs. Usually set to
"on".
Annual survival rate for *foxes*
< 1 year old
Annual survival rate for *foxes* 1
- 2 years old
Annual survival rate for *foxes* 2
- 3 years old
Annual survival rate for *foxes*
> 3 years old
Time of year when *fox* cubs are
born
Average number of cubs born to
a *fox-family*
Proportion of cubs that are
female when born
Age of cubs when they can
survive the death of all adults in
their *fox-family*
Start of dispersal season
End of dispersal season
Female dispersal rate
Male dispersal rate
The layout of *bait-stations*. A
shapefile must be imported for
the "custom" option.
The density of *bait-stations*.
This only affects the “grid” and
“random-scatter” ‘bait-layout‘
options
A shapefile (points) describing
the locations of the
*bait-stations*, if ‘bait-layout‘ is
“custom”
The frequency with which baits
are laid at baitstations. Check
that * options are compatible
with ‘weeks-per-timestep‘.
Weeks-of-year when baits will be
deployed at *bait-stations* if a
“custom*” bait-frequency is
chosen

(continued)
Parameter
‘Pr-die-if-exposed-100ha‘

Unit
index

‘commence-baiting-year‘

year

‘commence-baiting-week‘

week of year

‘price-per-bait‘

$
days

Explanation
The efficacy of the poison baits
when they are deployed at
*bait-stations*
The first year when baits at
*bait-stations* will be poisonous
(until then, baits are deployed
and eaten but don’t kill *foxes*)
The first week of the first year
when baits at *bait-stations* will
be poisonous
The cost of a single bait
The number of person-days it
takes to deploy a bait at each
*bait-station*
The cost per person-day while
deploying baits
Total distance travelled to deploy
a bait at each *bait-station*
The per-kilometre cost of travel
while deploying baits

‘person-days-per-baiting-round‘
‘cost-per-person-day‘

$

‘km-per-baiting-round‘

km

‘cost-per-km-travel‘

$

Monitoring Parameters
‘plot‘

"true"/"false"

‘age-structure‘

"true"/"false"

‘count-neighbours‘

"true"/"false"

‘density‘

"true"/"false"

‘dispersal-distances‘

"true"/"false"

‘family-density‘

"true"/"false"

‘foxes-on-transect‘

"true"/"false"

‘popn-structure‘

"true"/"false"

‘range-size‘

"true"/"false"

Whether you want monitored
variables to be shown in the
beige plot(s)
The number of *foxes* in each
age class within the
‘region-of-interest‘
The number of neighbouring
territories for each *fox-family*
within the ‘region-of-interest‘
The density of *foxes*, various
classes of *foxes*, and
*fox-families* within the
‘region-of-interest‘ (and
‘region-of-interest2‘ if applicable)
The distance that *foxes* within
the ‘region-of-interest‘ have
dispersed (excludes cubs &
individuals that have not
attempted to disperse)
The density of *fox-families*
within the ‘region-of-interest‘
The number of *foxes* (excluding
cubs) who have a territory that
overlaps ‘survey-transect‘ (and
‘survey-transect2‘, if applicable)
The number of *foxes* in various
status/sex classes within the
‘region-of-interest‘
Not currently activated

42

43

9.2

Week Conversion Table
week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week
Week

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52

start
1-Jan
8-Jan
15-Jan
22-Jan
29-Jan
5-Feb
12-Feb
19-Feb
26-Feb
5-Mar
12-Mar
19-Mar
26-Mar
2-Apr
9-Apr
16-Apr
23-Apr
30-Apr
7-May
14-May
21-May
28-May
4-Jun
11-Jun
18-Jun
25-Jun
2-Jul
9-Jul
16-Jul
23-Jul
30-Jul
6-Aug
13-Aug
20-Aug
27-Aug
3-Sep
10-Sep
17-Sep
24-Sep
1-Oct
8-Oct
15-Oct
22-Oct
29-Oct
5-Nov
12-Nov
19-Nov
26-Nov
3-Dec
10-Dec
17-Dec
24-Dec
44

end
7-Jan
14-Jan
21-Jan
28-Jan
4-Feb
11-Feb
18-Feb
25-Feb
4-Mar
11-Mar
18-Mar
25-Mar
1-Apr
8-Apr
15-Apr
22-Apr
29-Apr
6-May
13-May
20-May
27-May
3-Jun
10-Jun
17-Jun
24-Jun
1-Jul
8-Jul
15-Jul
22-Jul
29-Jul
5-Aug
12-Aug
19-Aug
26-Aug
2-Sep
9-Sep
16-Sep
23-Sep
30-Sep
7-Oct
14-Oct
21-Oct
28-Oct
4-Nov
11-Nov
18-Nov
25-Nov
2-Dec
9-Dec
16-Dec
23-Dec
30-Dec

10

References

Baker, P.J., Robertson, C.P.J., Funk, S.M. & Harris, S. (1998) Potential fitness benefits of group living in the
red fox, Vulpes vulpes. Animal Behaviour, 56, 1411-1424.
Bino, G., Dolev, A., Yosha, D., Guter, A., King, R., Saltz, D. & Kark, S. (2010) Abrupt spatial and numerical
responses of overabundant foxes to a reduction in anthropogenic resources. Journal of Applied Ecology, 47,
1262-1271.
Carter, N., Levin, S. et al. (2015) Modeling tiger population and territory dynamics using an agent-based
approach. Ecological Modelling, 312, 347-362.
Devenish-Nelson, E.S., Harris, S., Soulsbury, C.D., Richards, S.A. & Stephens, P.A. (2013) Demography of a
carnivore, the red fox, Vulpes vulpes: what have we learnt from 70 years of published studies? Oikos, 122,
705-716.
Hradsky, B.A., Kelly, L.T., Robley, A. & Wintle, B.A. (2019a) FoxNet: an individual-based model framework
to support management of an invasive predator, the red fox. Journal of Applied Ecology.
Hradsky, B.A., Kelly, L.T., Robley, A. & Wintle, B.A. (2019b) Data from: FoxNet: an individual-based
model framework to support management of an invasive predator, the red fox. Zenodo. http://doi.org/10.
5281/zenodo.2572045.
Hradsky, B.A., Mildwaters, C.A., Ritchie, E.G., Christie, F. & Di Stefano, J. (2017a) Responses of invasive
predators and native prey to a prescribed forest fire. Journal of Mammalogy, 98, 835-847.
Hradsky, B.A., Robley, A., Alexander, R., Ritchie, E.G., York, A. & Di Stefano, J. (2017b) Human-modified
habitats facilitate forest-dwelling populations of an invasive predator, Vulpes vulpes. Scientific Reports, 7,
12291.
Larivière, S. & Pasitschniak-Arts, M. (1996) Vulpes vulpes. Mammalian Species, 537, 1-11.
Lockie, J.D. (1959) The estimation of the food of foxes. The Journal of Wildlife Management, 23, 224-227.
Pech, R.P., Sinclair, A., Newsome, A. & Catling, P. (1992) Limits to predator regulation of rabbits in
Australia: evidence from predator-removal experiments. Oecologia, 89, 102-112.
Šálek, M., Drahnikova, L. & Tkadlec, E. (2015) Changes in home range sizes and population densities of
carnivore species along the natural to urban habitat gradient.* Mammal Review*, 45, 1-14.
Sargeant, A.B. (1972) Red fox spatial characteristics in relation to waterfowl predation. Journal of Wildlife
Management, 36, 225-&.
Saunders, G., McLeod, S. & Kay, B. (2000) Degradation of sodium monofluoroacetate (1080) in buried fox
baits. Wildlife Research, 27, 129-135.
Soulsbury, C.D., Iossa, G., Baker, P.J., White, P.C. & Harris, S. (2011) Behavioral and spatial analysis of
extraterritorial movements in red foxes (Vulpes vulpes). Journal of Mammalogy, 92, 190-199.
Thiele, J.C. (2014) R Marries NetLogo: Introduction to the RNetLogo Package. Journal of Statistical
Software, 58, 1-41.
Trewhella, W. & Harris, S. (1988) A simulation model of the pattern of dispersal in urban fox (Vulpes vulpes)
populations and its application for rabies control. Journal of Applied Ecology, 435-450.
Trewhella, W.J., Harris, S. & McAllister, F.E. (1988) Dispersal distance, home-range size and population
density in the red fox (Vulpes vulpes) - a quantitative analysis. Journal of Applied Ecology, 25, 423-434.
Wilensky, U. (1999) Netlogo. Center for Connected Learning and Computer-Based Modeling, Northwestern
University, Evanston, IL.
Winstanley, R.K., Buttemer, W.A. & Saunders, G. (2003) Field metabolic rate and body water turnover of
the red fox Vulpes vulpes in Australia. Mammal Review, 33, 295-301.

45



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.5
Linearized                      : No
Page Count                      : 45
Page Mode                       : UseOutlines
Author                          : Bronwyn Hradsky, Quantitative and Applied Ecology, University of Melbourne
Title                           : FoxNet User Guide V1.0.0
Subject                         : 
Creator                         : LaTeX with hyperref
Producer                        : pdfTeX-1.40.19
Create Date                     : 2019:02:20 11:12:39+11:00
Modify Date                     : 2019:02:20 11:12:39+11:00
Trapped                         : False
PTEX Fullbanner                 : This is MiKTeX-pdfTeX 2.9.6839 (1.40.19)

EXIF Metadata provided by EXIF.tools