This repository is a companion piece to the manuscript "Too few, too many, or just right? Optimizing sample sizes for population-level inferences in animal tracking projects". The main workflow presented in the manuscript provides a comprehensive approach for optimizing sample sizes in animal tracking studies, balancing sampling duration, sampling interval, and the number of sampled individuals to ensure robust, unbiased population-level inferences for home range, and speed & distance estimation. By integrating robust, sampling-insensitive analytical methods and accounting for uncertainty and logistical constraints (including fix success, location error, device malfunctions, and individual variation), this workflow guides researchers in designing effective studies or evaluating existing data. The workflow has been fully implemented in the movedesign Shiny application and R package, allowing users to easily test sampling strategies and assess the reliability of space-use and movement metrics, ultimately promoting more rigorous and impactful wildlife research and conservation. Tutorial The workflow described in this manuscript is implemented in the movedesign R package, which uses Shiny to provide an easy-to-use user interface. This application allows you to test different tracking schedules and population sample sizes, considering one or two research questions (home range and/or speed estimation). To install the stable version of movedesign from CRAN: install. packages ("movedesign") To install the latest version of movedesign from GitHub: install. packages ("remotes") remotes: : installgithub ("ecoisilva/movedesign) To launch movedesign, load the library and run the following command in your R console: library (movedesign) movedesign: : runₐpp () A step-by-step tutorial for the Shiny interface is available within the documentation folder (tutorialₚop. html). These workflows can also be run directly in the R console. As an example, you can follow along the tutorial available within the documentation folder (tutorialₚopconsole. html). Directory structure The directory structure below provides all code required to reproduce the data and figures from the manuscript. For guidance on applying the workflow to your own data, refer to the files listed in the Tutorials section above. studydesignₘs/│-- cluster/ # HPC bash files and scripts│-- data/ # Data files│-- documentation/ # Tutorial and supplementary files│-- figures/ # Generated figures│-- outputs/ # Processed outputs│-- R/ # R scripts│-- renv/ # `renv` package management directory│-- renv. lock # Lock file for dependencies│-- global. R # Global settings and objects│-- studydesignₘs. Rproj # RStudio project file Scripts description The 📁 cluster folder stores scripts for HPC jobs: - cluster/createⱼobₛh. R: generates job submission scripts. - cluster/meanₕr. R: runs simulations for mean home range. - cluster/meanₛpeed. R: runs simulations for mean movement speed. - cluster/metaᵣesampled. R: runs meta-analyses and resampling approach. The 📁 R folder contains the other main scripts: - R/plottingcase-study. R: generates output plots for a case study. - R/runningcase-studyₗoocv. R: runs LOOCV analyses. - R/plottingcase-studyₗoocv. R: generates LOOCV output plots for a case study. - R/plottingcase-studyᵣesampled. R: generates resample plots for a case study. - R/plottingₛimsᵣesampled. R: generates resample plots for the simulations. The 📁 R/functions subfolder contains helper functions used throughout the project. License CC BY 4. 0
Anonymous (Tue,) studied this question.