ISEC workshop: Combining different data types in a single model

This repo contains information and materials for the ISEC workshop 'Combining different data types in a single model' 29th June 2018. See the (draft) program for the day here. The rest of this page contains information and materials about the three workshop tutorials.

---

Rahel Sollman - Combining spatial capture-recapture, telemetry, and point count data

Rahel will talk about how to integrate abundance and individual location data into spatial capture-recapture models to estimate animal abundance. You can download Rahel's slides here. This paper by Linden et al. is another good reference. Rahel will walk us through how to carry out these analyses using the oSCR package (Github).

installation

You can download the most recent version of oSCR from GitHub, using the devtools package (which is on CRAN), and you'll also need to manually install the car package from CRAN.

# install devtools if you don't already have it installed, with:
# install.packages("devtools")
devtools::install_github("jaroyle/oSCR")
install.packages("car")

If you're a Windows user and this the first time you've tried installing a package from GitHub, you may need to install the RTools software (note: this is not an R package), which provides the tools you need to build packages from source code.

Rahel's example also requires the scrbook package for some plotting functions. scrbook is hosted on the developers' own website, and you should also be able to install it with the following commmands:

ext <- ifelse(.Platform$OS.type == "windows", ".zip", ".tar.gz")
scrbook_url <- paste0("https://sites.google.com/site/spatialcapturerecapture/scrbook-r-package/scrbook_0.28-2",
                      ext)
install.packages(scrbook_url, repos = NULL)

examples

Rahel's example will use synthetic data representing species counts from a spatial capture-recapture design, combined with data on the positions of individuals from telemetry. You can download the code for the example analysis here (you can either copy-paste the code into a file, or go to save page as ... in your browser)

---

Brett McClintock - Combining animal movement and environmental data with hidden Markov models

Brett will be talking about using Hidden Markov Models to combine movement data with other data streams to infer behavioural states of animals. You can download Brett's slides here. Brett and Theo Michelot developed the momentuHMM package (CRAN, GitHub) to make it easier to do these types of analyses. A paper introducing momentuHMM is freely available at Methods in Ecology and Evolution (pdf).

installation

To get started, you will need to load R and install the momentuHMM package from CRAN.

install.packages("momentuHMM")

To run the main example, you'll also need to install the setRNG package from CRAN.

install.packages("setRNG")

examples

The main example Brett will work through is an analysis of Northern Fur Seal foraging trips with position uncertainty. You can download the code for the Northern Fur Seal example analysis here (you can either copy-paste the code into a file, or go to save page as ... in your browser)

There are a couple of more advanced example analyses that you can download and play with, and that Brett might have time to talk about:

• Modelling Loggerhead Turtle movements relative to ocean currents with position uncertainty (link)
• Modelling Northern Fulmar movements relative to fishing vessels (link)

---

Bob O’Hara - Combining presence-only, presence-absence and abundance data with point-process models

Bob will talk about combining different types of distribution data together in a single point process model of a species' distribution. Bob has developed a package PointedSDMS (GitHub) to fit these models using the INLA inference software.

installation

You can install PointedSDMs from GitHub:

devtools::install_github("oharar/PointedSDMs")

You'll also need to install INLA from the INLA team's own repository, using the following code:

install.packages("INLA", repos = "https://inla.r-inla-download.org/R/stable"))

You'll need the following packages from CRAN to run the example too:

install.packages(c("spatstat", "RColorBrewer", "mapview"))

examples

Bob will work through a species distribution modelling analysis of New Zealand's Solitary Tinamou, using distribution data from multiple sources. This analysis is provided as a vignette in the package, and you can download the code for the Solitary Tinamou analysis here