Simulated Survey Of Megafauna At Offshore Windfarms ✈️🐬

The package sismow was developed as part of the OWFSOMM project.
The objective of this project is to demonstrate the technical relevance
of aerial digital monitoring while ensuring their comparability with
observer-based visual monitoring.

The goal of sismow is to simulate different datasets of aerial
monitoring. To do this, sismow allows to simulate:

1. density maps
2. sampling transects
3. observation probabilities

With the different functions of this package, it is possible to simulate
digital and visual aerial monitoring for different conditions (more or
less abundant species, more or less homogeneously distributed in space,
more or less discreet…). This allows then to calculate abundances and
distributions under different conditions and to calculate
intercalibration parameters between the two methods.

Installation

You can install the development version of sismow from
GitHub with:

# install.packages("devtools")
#devtools::install_github("maudqueroue/sismow")
library(sismow)

Population Abundance Emulator

Simulate density

First, the function simulate_density allows to create map with
different spatial variation of density.

Examples :

• homogeneous: same density througout the studied area.

map <- simulate_density(shape_obj = shape_courseulles,
                        density_type = "uniform")

• gradient : density decreasing from a hotspot that could be placed
  on different directions (gradient_direction : Center, North, East…).
  The amplitude and wavelength of the hotspot created can be
  modulated by the user.

map <- simulate_density(shape_obj = shape_courseulles,
                        density_type = "gradient",
                        gradient_direction = "NW",
                        amplitude = 200,
                        wavelength = 40000)

• Random: random density. Several hotspots nb_hotspots are
  randomly created with amplitude and wavelength randomly chosen. The
  maximum value of amplitude and wavelength can be modulated by the
  user.

map <- simulate_density(shape_obj = shape_courseulles,
                        density_type = "random",
                        amplitude = 200,
                        wavelength = 10000,
                        nb_hotspots = 15)

Simulate individuals

The simulate_ind function allows to simulate an approximate number of
individuals/groups with an inhomogenous Poisson point process according
to the densities provided in the map object.

Examples :

• Simulation of N = 200 individuals with mean_group_size = 1:

ind <- simulate_ind(map_obj = map, 
                    mean_group_size = 1, 
                    N = 200)

• Simulation of N = 100 groups of mean_group_size 5 individuals

ind <- simulate_ind(map_obj = map, 
                    mean_group_size = 5, 
                    N = 100)

Sampling Design Emulator

Simulate transects

The function simulate_transects allows to simulate transects with
different conditions:

• different types of survey design such as parallel, zigzag, crossed
  zigzag or random transects
• total transect length (approximately)
• angle of the transects
• segmentation of the transects - segments length
  (approximately)

Examples:

• parallel transects design with an approximate total line length of
  400km:

transects <- simulate_transects(shape_obj = shape_courseulles,
                             design = "systematic",
                             line_length = 400000)

• zigzag transects design with an approximate total line length of
  600km that are segmentized with a `length of approximatively
  2000m per segment.

transects <- simulate_transects(shape_obj = shape_courseulles,
                             design = "eszigzag",
                             line_length = 600000,
                             design_angle = 90,
                             segmentize = TRUE,
                             seg_length = 2000)

Observation Data Emulator

Simulate observations

The function simulate_obs allows to simulate datasets of observations
according to:

1. The individuals/groups simulated on the density map.
2. The transect design simulated.
3. A detection probability.

To determine the detection probability, the user can choose:

• The form of the detection function : uniform or half-normal
• The maximum distance of observation (truncation)
• The probability of observation at distance 0 (g_zero)
• The effective strip width (esw) for half-normal detection function

Examples:

• Half normal detection probability equal to 1 at 0 distance
  with a effective strip half width of 180m.

obs <- simulate_obs(ind_obj = ind,
                    transect_obj = transects,
                    key = "hn",
                    g_zero = 1,
                    esw = 180)

• Uniform detection probability equal to 1 below distance
  200m .

obs <- simulate_obs(ind_obj = ind,
                    transect_obj = transects,
                    key = "unif",
                    g_zero = 1,
                    truncation = 200)