Our sampled farms were small , had conducted some form of clearing for production space, and three had constructed some form of fence or barrier around their crop. Nonetheless, specific land use practices and production philosophies differed between farms . We cannot disclose farm locations, as per our research agreement for access. Monitored farms were clustered within each watershed: one farm in Slate Creek, five in Lower Deer Creek, and two in Lower East Fork Illinois River; however, most farms were also located near other nearby cannabis farms that were not directly monitored in this study. We placed unbaited motion sensitive cameras on cannabis farms as well as in random locations up to 1.5 km from the monitored farms. This is an expansion on previous camera research that only assessed on-site wildlife at these same farms . We placed cameras approximately 0.5 m off the ground to capture animals squirrel-sized and larger. We set cameras to take bursts of 2 photos, with a quiet period of 15 seconds. To guide the placement of cameras, we overlaid the area surrounding each cannabis farm cluster with a 50 x 50 m grid and then selected a random sample of at least one quarter of grid cells . We selected a 50 x 50 m grid size because we wanted to be able to detect fine scale space use responses of wildlife. The random sample was stratified by vegetation openness and distance to cannabis farm in all watersheds, and additionally by distance to clear cut in the Slate Creek watershed, such that cameras were placed in proportion to the landscape attributes and a distance gradient was achieved. When a selected site was inaccessible, we selected a new one that also met the same stratification criteria. We rotated 15-20 cameras through the sampled grid cells, hydroponic rack system ensuring each camera was deployed for at least one round of two week duration. Because of rotations and field constraints, all cannabis sites were not monitored at the same time or for the same length of time .

Altogether, we monitored a total of 149 camera stations for a combined 4,664 trap nights. We then used a team of researchers trained to identify species found in the study area to sort photos by hand, grouping by species.To assess the local space use response of wildlife to cannabis production, we used single-season, hierarchical single and multi-species occupancy models. Our approach is a departure from the typical use of these models to estimate occupancy in that we knowingly violated multiple assumptions of occupancy models: first, because cameras were spaced relatively close together compared to the home range of species included in the study, we have likely violated the assumption of independent cameras; second, as a result of the aforementioned spacing as well as sampling across two years , we likely violated the model’s assumption of geographic and demographic closure . We have done our best to account for these violations in our use of regional fixed effects, as well as our narrow interval of replication . However, given our interest was in space use associations and not estimates of occupancy, we believe the violations are a minimal issue. This use of occupancy models is not particularly unusual, as the use of occupancy modeling to assess space use is becoming more common in wildlife response studies, and even traditional uses of occupancy modeling are influenced by wildlife space use . With the closure assumption violated, the occupancy probability estimate represents the likelihood that the animal occupied the site at any point during the study period, while the detection probability represents a combination of the probability that the species is detected and the intensity of use of the site within its larger range . This interpretation is common in camera trapping studies , but we proceed while being careful to acknowledge where appropriate that any covariate’s influence on detection probability is a combination of its effect on detection and the intensity with which an animal uses a given space. In addition, we have taken care to include variables in the detection process to account for what we anticipate to be the largest sources of variation in detectability, so that the other variables should primarily reflect space use intensity. We therefore interpret occupancy for the models as space use rather than true occupancy .

We operationalize detection as a combination of intensity of use, and camera detectability or error . For the single species occupancy models, occupancy and detection varied by species . Recall that for our models, we are interpreting occupancy as space use, and detection as a combination of detectability and space use intensity . Deer and tree squirrel occupancy probability increased with distance from cannabis farms, indicating potential avoidance. Domestic dogs, as expected, decreased in predicted occupancy with distance to cannabis farms. Interestingly, gray fox and ground squirrel occupancy probability also decreased with distance from cannabis farms, indicating that these species may be more likely to be found on and around cannabis farms . Six species had a meaningful detection response to cannabis farms . As expected, bobcat and ground squirrel detection probability increased with distance from cannabis farms, indicating that they may use areas further from cannabis farms more intensively. For ground squirrels, this implies that although they are more likely to be found closer to cannabis farms, they may use the spaces farther from farms more intensively. Again as expected, domestic dog detection probability decreased with distance from cannabis farms, confirming that they spend most of their time on and surrounding cannabis farms. Surprisingly however, deer, jackrabbit, and striped skunk detection also decreased with distance from cannabis farms. More frequent detections on occupied cannabis farms implies that these species may also be using the space on and surrounding cannabis farms more intensively . The other model covariates aside from cannabis also varied by species . For a majority of species, at least one regional intercept was meaningfully associated with occupancy probability. Elevation predicted occupancy for coyotes and striped skunks, and forest proportion predicted occupancy for jackrabbits, tree squirrels, and ground squirrels. Distance to highways was the only occupancy covariate that was not credibly non-zero for any species. As for detection, all covariates were meaningful for at least some species. The covariates for detectability, camera type and camera view, were credibly non-zero for four species all together. There was evidence for seasonal effects, with date and date2 meaningfully predicting detection for a majority of species. The activity indices had meaningful, and somewhat surprising results. Coyotes, bobcats, and tree squirrel detection was negatively associated with human activity, and ground squirrel detection was negatively associated with dog activity. However, coyote, gray fox, and jackrabbit detection probabilities were all positively associated with dog activity.For the multi-species occupancy models, almost no population-level parameters were meaningful .

No group meaningfully responded to cannabis in either detection or occupancy processes. No covariates were meaningful for occupancy or detection at the population level, aside from omnivore detection intercept. However, there was more variation at the species level . For the species that also had single species model results, the MSOM results largely matched, with occasional changes in credibility. For instance, for the deer SSOM, date and date2 were not credibly non-zero, but in the MSOM they were, even though the actual estimated values were similar in both . Despite the lack of population-level associations, some groups did have common responses to cannabis at the species level. For example, the occupancy probability for all ground bird species was credibly positive, increasing with distance from cannabis farms, which implies possible spatial avoidance of cannabis farms . For all ground bird species and both herbivore species, detection probability credibly decreased with increasing distance from cannabis farms, which may imply that these groups use areas around farms more intensively . Domestic species largely responded as predicted at the species level: cat and dog occupancy decreased with distance to cannabis, and dog and horse detection decreased with distance from cannabis . The other groupings were more mixed. Carnivores largely did not respond meaningfully to cannabis in either detection or occupancy . Omnivores had slightly more sensitivity, with three out of seven species responding meaningfully to cannabis in either occupancy or detection . For small mammals, tree squirrels and ground squirrels had opposite occupancy responses, rolling tables grow and only ground squirrels had a credibly non-zero detection response . This study assessed wildlife space use responses to active small-scale outdoor cannabis farms on private land. Our work provides a timely baseline for understanding potential wildlife community consequences from an emerging land use frontier. Our application of occupancy modeling to space use responses has yielded two main conclusions: 1) even at small scales, rural cannabis farming can affect local wildlife space use; 2) patterns of animal space use responses are species-specific, but there may be common patterns for herbivores, ground birds, and some mesopredators in how they use spaces near to cannabis farms. These results have implications for the cannabis industry and small farm strategies for conservation. Eight out of ten species modeled individually had a meaningful response to distance from cannabis farms, either in occupancy or detection. Although the population-level means were not meaningful, at an individual level, 13 out of 24 of the species included in multi-species models had a meaningful response to distance from cannabis farms, either in occupancy or detection. Our hypothesis that a majority of species would avoid farms was not supported, since the strength and direction of effects were species-specific. However, the results imply a general ability for cannabis farming to affect local wildlife space use. The relationships between occupancy and detection probabilities and distance to cannabis also indicate that there could be threshold effects relatively close to farms where the slope of the relationship is steeper , though further steps would be needed to confirm this relationship. These results are in contrast with research from the western US on vineyards and avocado production that indicates the ability of some wildlife to use farmed land in seeming preference over surrounding land uses . However, these other studies were conducted in areas where the agricultural land formed a corridor through more human-dominated land covers, which is the inverse of the landscape studied here. Our results are similar to studies on agroforestry systems with annual and perennial croplands, where there may be differential responses to agricultural land use and potential for filtering responses . Compared to the other covariates in the models, distance to cannabis farms meaningfully affected more species than any other single covariate other than the intercepts, or Date and Date2 . It was particularly surprising that wildlife responded to the physical land use of cannabis farms even more than human or dog activity, given that in other systems their space use intensity often responds more to human activity than human footprint , and is often negatively affected by the presence of dogs . This implies that cannabis farms may combine multiple potential sources of disturbance that wildlife may react to, and/or that the physical modifications for cannabis farms on their own are enough to trigger wildlife responses. More research is needed to disentangle some of the potential mechanistic pathways by which cannabis farms may affect wildlife. Overall, space use responses to cannabis were species-specific, confirming our alternative hypothesis for individual responses. While functional- or diet-group patterns are not as clear in this case as in other study systems , a few general patterns may be emerging, specifically in regard to herbivores/ground birds, and mesopredators. Our approach of using an occupancy modeling framework to assess wildlife space use associations was useful to identify some of these emerging patterns, because it allowed us to look at space use, separately from inferences on space use intensity . This is important because it helps capture different types of responses: attraction and deterrence, as well as potential behavioral shifts in activity patterns . For example, this helped identify opposing occupancy and detection responses from some herbivore and ground bird species. For medium to large herbivores and ground birds , occupancy credibly increased with distance from cannabis farms, while detection credibly decreased. This is the inverse of our alternative hypothesis that species using cannabis farms would decrease their activity intensity near to cannabis and suggests that while these species may generally avoid cannabis farms in space , the few areas that they do use, they may use more intensively.