2.1 New England cottontail and eastern cottontail presence data
We used presence data collected by the New England Cottontail Regional Monitoring Program from 2016 – 2022 in Connecticut. The Regional Monitoring Program was established in 2015 as a collaboration between all states with extant populations of New England cottontail (Connecticut, Maine, Massachusetts, New Hampshire, New York, and Rhode Island), where biologists from each state collected cottontail fecal pellets at designated sites. Observers collected pellets each winter (from November to April) during conditions that maximize pellet detection and quality: snow-covered ground, low temperatures, and 2-4 days after a snowfall or high wind event (Kovach et al. 2003, Brubaker et al. 2014, Whipps et al. 2020). To collect pellets within sites, observers walked parallel transects that were spaced at least 30 m apart. Observers chose the direction of the transects within each site and consistently used the same direction within a site but transect direction could differ between sites. Observers searched up to 15 m on either side of the transect for a pile of cottontail pellets and once a pellet pile was found, observers collected at least one pellet from the pile (hereafter referred to as sample) and placed the sample into a vial. Observers recorded GPS coordinates of each sample collected. Observers then walked at least 30 m before collecting the next sample to ensure that samples were collected from throughout the site (Rittenhouse and Kovach 2020). Following this field protocol, an average of 0–10 samples were collected at each site for each visit.
Following sample collection, observers sent samples to two laboratories (University of Rhode Island and University of New Hampshire) to process samples and identify the species of each sample. Laboratories used fecal mitochondrial DNA analysis with barcoding of diagnostic characters to extract high quality DNA and identify species (Sullivan et al. 2019, Whipps et al. 2020). DNA extraction and species identification occurred in over 99% of the samples (Sullivan et al. 2019), but when DNA was not extracted from the sample, we excluded the sample from the analysis. We cross-referenced species identification with the location of the pellet to create a presence dataset of both New England and eastern cottontail pellets within sites in Connecticut. Observers recorded when a species was absent from a site, not a specific location, so we could not include absence data within the models.
Sites were sampled within defined New England cottontail focus areas, thus, we accounted for spatial autocorrelation using the ‘spThin´ package (Aiello-Lammens et al. 2015) in R 4.3.1 (R Core Team 2023). We used a thinning distance of 10 m to align with the fine spatial resolution of the environmental predictors (10 m cell size). Samples within 10 m of each other were randomly selected for inclusion in the data set. The total occurrences before thinning for New England cottontail was 2086 and eastern cottontail was 4285. After thinning and removing multiple occurrences within one cell, the number of occurrences for New England cottontail was 1735 and eastern cottontail was 3629, a reduction of 16.83% and 15.31%, respectively. The average geographic distance between thinned locations was 49.57 m for New England cottontail locations and 33.81 m for eastern cottontail locations. We used the thinned datasets for all analyses. Since we used data collected in the winter only, models described New England cottontail and eastern cottontail winter distribution and niches. Winter is the time of year where cottontail survival is typically lowest, depending on body condition (Cheeseman et al. 2021), due to limited resources and harsher weather conditions.