Fig 1.
A 1:2000 scale map of the 2.12 km2 study area: The UC Santa Cruz campus.
Sites remained constant across Trapping Periods 1 and 2. Cameras placed at initial food sources (labeled) are marked with a triangle, while cameras placed at each sequential site (75 m and 150 m toward natural habitat, respectively) are marked with a circle. Site 1 was located in pure redwood habitat, Sites 2 and 3 were located in mixed redwood habitats, and Sites 4 and 5 were located in mixed oak habitats.
Fig 2.
Examples of the trash vessels in the study.
Depicted on the left is an open compost heap, and depicted on the right is an open trash can and a metal trash can with a swing lid.
Fig 3.
An example of rodent classification based on ear: body ratios (Peromyscus left, Neotoma right). Solid lines denote ear measurements, while dotted lines denote body measurements. All body measurements were taken from the tip of the snout to the base of the tail.
Fig 4.
Distribution of rodent presence or absence in pre- and post- COVID lockdown conditions.
Light gray circles indicate the presence of invasive Rattus, yellow circles indicate the presence of Peromyscus, blue circles indicate the presence of Neotoma, and dark gray circles indicate the absence of any rodent detections.
Fig 5.
Total site occupancy capture rates in Neotoma, Peromyscus, and invasive rats between trapping Periods 1 and 2.
“Successful capture rate” indicates the proportion of sites where the animal was captured. Invasive rat activity decreased between periods (ρ = 0.33 successful capture rate during Trapping Period 1, ρ = 0 successful capture rate during Trapping Period 2).
Table 1.
Estimated proportion of total sites occupied by invasive Rattus, followed by the detection probability.
All results are back transformed linear combinations and fitted with a 95% confidence interval.
Table 2.
Occupancy model results using habitat as a covariate that affects occupancy, detection probability, or both (pre-COVID conditions only).
We tested three models: constant detection and variable occupancy, variable detection and constant occupancy, and variable detection and variable occupancy as a function of habitat.
Table 3.
Top three performing mixed linear regression models for the influence of human supplementation on invasive rat activity, including AIC values and the difference between each model and the lowest AIC (ΔAIC).
All models included trap night and site as random effects.
Table 4.
Top three performing mixed linear regression models for the influence of human supplementation on native small mammal activity, including AIC values and the difference between each model and the lowest AIC (ΔAIC).
All models included trap night and site as random effects.