occstanhm: Hierarchical occupancy models with correlated error structure

Authors: Richard A. Erickson, Charles J. Labuzzetta

Point of contact: Richard A. Erickson ([email protected])

Repository Type: R packing calling Stan models

Year of Origin: 2024 (original publication)

Year of Version: 2024

Version: 2.0.0

Digital Object Identifier (DOI): 10.5066/P13FNZOF

USGS Information Product Data System (IPDS) no.: IP-170402 (internal agency tracking)

---

Suggested Citation:

Erickson, RA, and Labuzzetta, CJ. 2024. occstanhm: Hierarchical occupancy models with correlated error structure. U.S. Geological Survey software release. Reston, Va. https://doi.org/10.5066/10.5066/P13FNZOF.

Authors' ORCID nos.:

• Richard A. Erickson, 0000-0003-4649-482X
• Charles J. Labuzzetta, 0000-0002-6027-0120

---

---

This repository contains a R package with 2-level and 3-level occupancy models. The package also contains a set of tutorials designed to help users understand how to use and code occupancy models in Stan. The models include a statistical hierarchy that allows for multi-species modeling and estimating correlations among species. The models may also be used as a more general "random-effect" type occupancy model.

The models are written in Stan and called through R. The cmdstanr package is used, rather than the more common rstan package, because features used in Stan were not supported by rstan at the time of development. Additionally, cmdstanr allows for quicker performance than rstan.

Installation

This code requires the cmdstanr package to run in R. Please look up the official documentation for install directions. We also include a Dockerfile for people who prefer to use Docker.

Once you have cmdstanr installed and setup, this package may be installed using this code:

if (!require("remotes")) install.packages("remotes")
remotes::install_gitlab('umesc/quant-ecology/occstanhm@main', host='code.usgs.gov')

You may wish to lockdown a specific version by changing main to the version (e.g., v2.0). As of May 2024, the build_vignettes option appears to not be working. Vigenttes may be built by cloning the repository using git and then installing locally.

Where to get started

After installing the program, please consult the vignettes to learn more. The Introduction_overview provides a starting place and describes a suggested learning path through the tutorials.

Repository Files

This repository contains the code for an R package using RStan. This repository contains the standard R repository files (see the official R Documentation Writing R Extensions accessed May 2024 or the online book, R Pakcages (2e) for an descriptions of these files). In addition to the R Package source files, this repository contains the following files:

This repository file contains the following files and folder:

• README.md is this file.
• LICENSE.md is the Official USGS License.
• code.json is the code metadata.
• CONTRIBUTING.md describes how to contribute to this project.
• DISCLAIMER.md is the standard USGS disclaimer.
• .gitignore is a file telling git which files to not track.
• docker_files contains the Dockerfile to use this code.

User skill level

This package expects a user to understand Bayesian Statistics and occupancy models. Users seeking to adapt code would also benfit from understanding programming in Stan.

Acknowledgments

This research was funded by the USGS Biological Threats and Invasive Species Research Program and the US Fish and Wildlife Service. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.