A repository that contains the data and code for:

Larson, R.N., M. Fidino, and H. A. Sander. Urban deer mouse abundance is more strongly correlated with predator occupancy than land cover. Landscape and Urban Planning. (In review)

This README file includes information on the various scripts and datasets used for this analysis. Not every data source is saved in this repository (e.g., GIS data). The manuscript contains citations for where to access the geospatial data.

./Mouse_Model.R: R script that cleans and processes the data & runs the population model for the deer mice

./Predator_Models.R: R script that cleans and processes the data, runs the multi-species occupancy model, and performs post-model calculations (e.g., average occupancy probability) for the predators (coyote, red fox, domestic cats, mink)

./graphing.R: R script that graphs the mouse model results (Figures 2 & 3 in the manuscript, and Supplemental Figures 1 & 2)

There is 1 subfolder and 4 files in this folder

./data/PERO.csv: The counts of deer mice captured for each night of trapping

Data in this file take the following format. Cells that contain an 'NA' instead of a number mean traps were not set at that site for those nights (i.e., the site was not sampled):

./data/predatorOccu.csv: Detection histories of the predaors included in the predator model

Data in this file take the following format:

./data/siteCovs_Pred.csv: The site covariates for each of the 41 camera trap locations. All covariates were measured from a 500m buffer surrounding the camera's location.

./data/siteCovs_Rod.csv: The site covariates for each of the 45 small mammal trapping plots.

There are 3 files in this subfolder

./data/obsVars/jDate.csv: The Julian date of each night of small mammal trapping

./data/obsVars/Moon.csv: The moon illumination (proportion full) on each night of trapping

./data/obsVars/Effort.csv: The number of available traps (i.e., traps that remained undisturbed) on each night of trapping

All files follow this format, with the jDate file as an example. Just sub out 'jDate' for 'moon' with the moon illumination data and 'effort' for the trap effort data:

There are 6 files in this folder, all utility functions that automate or declutter some of the R code

.functions/logit2prob.R: Function script to convert a logit value to a probability

./functions/pred_inits.R: Script that stores the initial values to pass to the multi-species predator occupancy model.

./functions/rod_inits.R: Script that stores the initial values to pass to the mouse abundance model.

.functions/split_mcmc.R: Function to split a model's MCMC matrix into a list of named objects, one for every parameter. Makes graphing results much easier. Credit for this code goes to @mfidino (see his blog post here)

.functions/wide_to_stacked.R: Function to convert a wide-format data frame (i.e., one site per row with one column for each observation) into a stacked format data frame (one observation per row, with sites/seasons/etc. "stacked" on top of each other). Modified from code in a vignette for the umbs R package.

.functions/wideObs_to_stacked.R: Function script to convert observation-level covariate data from a wide format to a 'stacked' format. Different from wide_to_stacked.R in that it does not add a 'Species' column to the resulting dataframe. Modified from code in a vignette for the umbs R package.

There are 2 files in this folder, each with code to pass to JAGS to run the models.

./jags/mouseModel.R: The Bayesian population abundance model code to pass to JAGS.

./jags/predatorModel.R: The Bayesian multispecies occupancy model code to pass to JAGS.

There is 1 subfolder and 1 file in this folder. The contents of this folder are for creating the contagion index values for each site (see the manuscript for more details).

./landscapes/connectivity.R: R script for calculating habitat fragmentation (i.e., contagion index) for each trapping plot

This subfolder contains .tif files of the land cover of each small mammal trapping plot. There are 45 files in here, so I'm not going to list them all, but they are named after their site number: e.g., 1.tif, 124.tif, etc.