Free-roaming domestic cat (Felis catus) populations present unique challenges to communities across the United States for various reasons. Reasons for managing free-roaming cat populations may relate to the health and welfare of the cats,1,2 people with whom they come in contact with,3 and wildlife populations that are preyed upon.4–8 In areas with an abundance of free-roaming cats, decisions about implementing population control programs incite controversy.9 Although results of some empirical studies10–12 have shown that trap-neuter-return programs can be used to halt free-roaming cat population growth, others have revealed evidence of stable populations or even population increases when such programs were implemented.13,14 A recent review of free-roaming cat population management literature15 noted that lack of baseline population estimates for comparison following interventions and scarcity of studies on the ecology of cat populations are major barriers to assessing whether population management strategies are effective in various communities. This gap in the literature makes it difficult for communities to decide which management strategy will be effective in their areas.
Several methods have been used to estimate free-roaming cat population sizes, including hands-on methods that involve capturing and marking cats16,17 and hands-off methods that are used to survey cats from a distance.18,19 While hands-on methods such as capture for marking by ear-tip clipping or placement of radio collars (termed mark-recapture) are possible in some areas, less invasive methods are more accessible to a wider variety of communities and may be a more viable option in communities desiring to gather baseline data before beginning a population control program.
The Alliance for Contraception in Dogs and Cats has published guidelines for monitoring populations of free-roaming cats over time to evaluate the effectiveness of population control methods.20 These guidelines include suggestions for field survey protocols, sampling methods, and stratification when sampling diverse habitats. In particular, they suggest rapid surveys performed with the line-transect method (as described herein) combined with more intensive surveys of selected sites, in which researchers attempt to record the presence of every cat in that location. Similarly, in a collaborative effort between the Audubon Society of Portland and the Feral Cat Coalition of Oregon, guidelines were established that allowed volunteers to survey free-roaming cat populations on Hayden Island with the line-transect method.21 These guidelines were meant to standardize collection of baseline data on the specific locations and distributions of cat populations and monitoring of these populations over time.21 These simple, hands-off survey methods allow consistent data collection from volunteers with little to no experience conducting field research. In the line-transect method recommended by these guidelines, researchers survey an area for the presence of cats by walking along predetermined transects (ie, identifiable paths that allow visualization of a defined area in a consistent and repeatable manner) and recording the number of cats sighted by use of mark-resight analysis methods, in which identifying features of cats are noted so that the number of times the same cat is sighted can be quantified in the analysis of the data, if possible. This method has been used in several studies14,19,22–24 of free-roaming cat populations around the world and has the advantage of requiring very little equipment and training. An important consideration for this method is ensuring that sampling occurs at times of the day or night when cats are most active and likely to be seen. A study19 of cats in Brooklyn, New York, found that cat activity seemed to peak at approximately 1:00 am and again at dawn (approx 7:00 am), suggesting that urban cats may have a crepuscular activity pattern and may be affected by local human activity levels.
Recently, trail cameras have been used to monitor and estimate population sizes and document behaviors for a variety of wildlife species.25 The literature regarding the use and limitations of this method in wildlife assessment is extensive. Concerns for communities wishing to establish baseline data for evaluation of cat populations include the importance of camera placement, the risk of camera theft or vandalism,26 the possibility of cameras altering natural behavior of the cats,27 whether individual animals can be identified by this method,28,29 and the need for appropriate study design28 and advanced analysis methods30,31 to improve population estimates from this type of data. Trail cameras have been used to study free-roaming cat populations, especially in remote areas of Australia.32–36 This method requires some skill in the use of cameras and video equipment, but it also allows continuous sampling for several days, limiting researcher field time, and can be useful for recording the presence of cats and wildlife not accustomed to or accepting of human presence. Some wildlife studies33,35,37 that relied on trail cameras for documentation have included the use of lures to increase the likelihood of target species being captured on camera and to increase the duration for which an animal remained in range of the camera. Although hands-off survey methods that include the use of spotlighting,18 track counts,18 line transects,14,19 and trail cameras32 have all been used to estimate free-roaming cat population sizes, to the authors' knowledge, the different approaches have not been directly compared in urban and mixed environments in the United States.
In addition to studying the size and distribution of free-roaming cat populations, communities considering a population control program may also want to know what other factors influence the presence of free-roaming cats in an area. Examples of these factors include the animal species that compete with cats for food,38 the population of prey species,39 and the presence of anthropogenic food sources such as garbage19 and cat food provided outdoors.24 One study19 performed in New York found no association between the number of uncovered trash cans and the number of cats in the area, but results of a study24 performed on a South African university campus suggested that the locations of feral cat feeding stations influence cats' home ranges. If the presence of anthropogenic food resources has an effect on free-roaming cat population size, this information could help communities to plan a more holistic approach to population control.
The purpose of the study reported here was to identify determinants of cat presence and to compare the line-transect and trail-camera methods used to assess free-roaming cat population size in a mixed-urban environment on and around The Ohio State University's campus. Although the study was also intended to collect preliminary information for use in a planning a survey of the local cat population, we did not attempt to estimate the free-roaming cat population size in the study area.
Materials and Methods
The study took place at and near The Ohio State University in Columbus, Ohio, with survey experiments performed between May 23 and August 17, 2016. The city of Columbus included a population of > 850,000 people in 2016,40 with > 2 million people residing in the metropolitan Columbus area.41 At the time of the study, 59,000 students were enrolled at the Columbus campus.42 The surveillance area for the study included 6.25 km2 of the campus and 4.36 km2 of the neighborhoods surrounding its eastern edge. Although located within a large urban area, the university campus comprised a mixed-urban landscape that included agricultural land, athletic fields, and wooded areas.
Line-transect method
The study area was divided into zones that were considered easily walkable in < 30 minutes. Each zone consisted of a single predominant land-use category. Zones were classified as on campus or off campus by use of maps available on The Ohio State University's website.a Of 705 designated zones, 542 were on campus and 163 were off campus. Land areas owned by the university but not contiguous with the main campus, including a wetland research park, were not used in the study. Each zone was classified as one of the following land-use categories: urban (pavement and buildings), mowed grass (lawns and athletic fields), wooded (trees), agricultural (fields of crops and pasture for livestock), and unmowed grass (tall grasses and prairies). These land-use categories corresponded to color-coded regions on maps available from the university website.a For off-campus areas, the land-use category was extrapolated visually from the available aerial views and confirmed by direct visual observation. Zones with a mixture of land-use types were classified according to the predominant type; for example, a field with a border of trees was classified as mowed grass. If ≤ 10 zones in a particular land-use category were present, all of the zones were sampled. If > 10 zones were available (urban on and off campus and wooded on campus), a subset of available zones was randomly selected to be sampled once each by the line-transect method. Zones were assigned numbers from 1 to the total number for that category, and then zones were chosen with unique numbers obtained from a random numbers generator.b We selected from available zones until our target sample size of 100 zones was reached. There were no zones classified as agricultural or wooded in the off-campus area (Appendix). Mean total areas of the on-campus and off-campus zones that were sampled were 19,899 m2 (4.9 acres) and 28,994 m2 (7.2 acres), respectively.
Beginning before sunrise at approximately 5:45 am and ending by 8:30 am, 2 researchers (1 author [ECV] and 1 student assistant) performed sampling of 3 to 9 zones in the same general area on each weekday between May 23 and June 15, 2016. Line-transect sampling was conducted at this time of day because results of a previous investigation19 suggested that free-roaming cats are most active around sunrise. The researchers walked through the public areas of each zone (using sidewalks when off campus) at a steady pace, visualized the entire area of the zone, and recorded the presence of any cats sighted within the boundaries of the zone. If a cat was observed, the researchers attempted to obtain a photograph and record the following information: time of the sighting, estimates of cat age (kitten or adult) and body condition (underweight, overweight, or healthy), activity of the cat at the time of sighting, and any pertinent or identifying physical features (eg, markings, apparently clipped ear tips, or presence of a collar). The location of the cat was annotated on a printed map of the zone. Information about the environmental temperature, precipitation, number of people seen in the zone, and any other animals present was also recorded.
If any part of the zone was inaccessible (eg, blocked because of construction), this was noted at the time of data collection. A smartphone applicationc was used to record the exact location, assessment times, and distance walked in each zone.
Trail-camera method
Trail cameras were used to continuously assess for the presence of cats over a 7-day period. Four different models of trail camerasd–g were used on the basis of availability. The cameras used infrared light to record nighttime images. Each camera was set to its highest sensitivity setting and was programmed to respond to motion as quickly as possible, obtaining 1 photograph/triggering event. One camerae responded to motion after 5 seconds, and the other 3 responded at the time motion was detected (with no delay).
Directly after zones were selected for data collection with the line-transect method, a subset of 3 of these zones from each land-use category was randomly selected for trail camera placement on and off campus according to the same method used for the line-transect sampling. If any land-use category had < 3 zones, all zones in that category were sampled. A single trail camera was placed in each of the 23 selected zones (15 on campus [3 zones in each category] and 8 off campus [3, 3, and 2 in the urban, mowed grass, and unmowed grass categories, respectively]) for a 1-week period after the line-transect sampling had been completed. The person placing the camera was aware of the results of the line-transect sampling. The cameras were placed approximately 0.6 m above the ground, secured to a tree or a pole with a cable lock, and angled slightly toward the ground by use of wooden shims (Figure 1). A lure station comprised of a plastic container with drilled holes and affixed to a plywood base was positioned an estimated 3 to 6 m in front of each camera, and the actual distance was subsequently measured. Freshly chopped catnip leaves were placed in the plastic containers at each lure station, catnip oil was sprayed on the outside of the container and on the camera, and dried catnip was sprinkled over the top of the lure station.
Photograph showing typical configuration of a trail camera secured to a tree and placement of a catnip lure station in a study to evaluate 2 field research methods of surveying free-roaming cats (Felis catus; line-transect [visual surveillance] vs trail camera [motion-triggered still photography]) and identify factors potentially associated with the presence of such cats in the mixed-urban environment of a large university campus. Lures were not used with the line-transect method.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Photograph showing typical configuration of a trail camera secured to a tree and placement of a catnip lure station in a study to evaluate 2 field research methods of surveying free-roaming cats (Felis catus; line-transect [visual surveillance] vs trail camera [motion-triggered still photography]) and identify factors potentially associated with the presence of such cats in the mixed-urban environment of a large university campus. Lures were not used with the line-transect method.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Photograph showing typical configuration of a trail camera secured to a tree and placement of a catnip lure station in a study to evaluate 2 field research methods of surveying free-roaming cats (Felis catus; line-transect [visual surveillance] vs trail camera [motion-triggered still photography]) and identify factors potentially associated with the presence of such cats in the mixed-urban environment of a large university campus. Lures were not used with the line-transect method.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Camera stations were assessed each morning. If cameras were frequently triggered by movement of inanimate objects, they were repositioned to a different angle on the same post during the evaluation week. Lure stations were resprayed with catnip oil and sprinkled with dried catnip daily. Fresh catnip was replaced halfway through the week.
Time, date, and species present were recorded for all images. If an animal was photographed in multiple concurrent images, only the first image was used to record data on the spreadsheet. If ≥ 10 minutes passed between sightings of the same animal, these were recorded as multiple events. When multiple animals were photographed in a single frame, they were each recorded as separate sightings. Any images of persons were promptly deleted to protect their privacy.
The maximum distance of a cat from a specific trail camera at the time of imaging was estimated by bringing a printed photograph of the 1 cat observed at the greatest distance from the camera and using that print to approximate its location at the time of image capture. The distance between the camera and approximate location of the cat was then measured.
The study followed guidelines set by the American Society of Mammalogists.43 Because the study included the use of lures that could potentially alter the natural behavior of free-roaming cats, this part of the study protocol was submitted to The Ohio State University Institutional Animal Care and Use Committee and was approved (No. 2016A00000036). Permission to place trail cameras was gained from campus administrators and private landowners. No landowners refused the request to place trail cameras on their land.
Resource locations
Cross-sectional surveys for resources that could influence the presence of cats, such as cat feeding stations, open dumpsters containing food waste, and overturned trash cans, were recorded for each zone by a researcher (ECV) using a global positioning system device.h That researcher walked through each of the 705 zones over the course of 4 weeks and saved the location of each resource on the device. To reduce the potential for errors in data entry, resource locations were downloaded directly from the device and plotted on a digital map of the university campus by means of mapping software.i The total number of resources per area was calculated by dividing the number of resources observed on campus by the total area of zones on campus (6.25 km2) and off campus (4.36 km2).
Statistical analysis
Statistical softwarej,k was used for all data analysis, with values of P < 0.05 considered significant. Because imaging of animals in motion or by use of infrared lighting with the trail-camera method impaired the ability to identify individual cats, the number of separate sightings, rather than numbers of individual animals, was used for analyses. Because the distributions of numbers of cat sightings recorded by trail cameras per site were skewed (mean, 4; median, 0; skewness, 2.43), 3 Kruskal-Wallis rank sum tests were run. First, counts of sightings per imaging period (ie, observation week) were compared between sites grouped as on-campus versus off-campus locations; second, counts were compared among the 5 land-use categories (ie, habitat types for the cats); and third, counts were compared among the 8 combinations of habitat type and on- versus off-campus location. A Spearman correlation was used to test for relationships between the number of cat sightings recorded by a single trail camera and the number of resources recorded in that zone plus all adjacent zones (ie, zones sharing borders with the selected zone because cats would not be expected to remain in a given zone). A Spearman correlation was also used to determine whether the number of cat sightings in a zone was associated with the number of sightings for other species in that zone.
The time of day when each cat was photographed with a trail camera was categorized as day (6:00 am to 8:59 pm) or night (9:00 pm to 5:59 am). Day and night classification was determined on the basis of typical sunrise and sunset times during the summer in central Ohio. A rate ratio calculationl was used to compare the number of cat sightings per observation hour recorded on the trail cameras during the day versus night, and the median unbiased method was used to estimate the P value. The rate-ratio calculation method was also used to compare the number of cat sightings per hour with the trail-camera versus line-transect methods. For this method, the number of cat sightings was input into the software with the total time spent observing by each method (either the total time that cameras were set in place or the total time researchers were actively walking on transects).
Results
Line-transect method
The mean length of transects was 0.40 km (median, 0.30 km; range, 0.12 to 1.6 km); the distribution of distances walked per transect was summarized (Figure 2). A total of 6 free-roaming cat sightings were made with the line-transect method for 100 zones sampled (Figure 3). Five of the 6 cat sightings occurred off campus, in 4 zones categorized as urban land use areas or habitats. One sighting was on campus in an urban zone. Four of the 6 sightings involved cats that were at a dumpster or were interacting with trash at the time of observation.
Histogram of distances walked per transect by researchers performing visual surveillance for free-roaming cats by use of the line-transect method in on-campus (dark gray bars) and off-campus (light gray bars) regions.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Histogram of distances walked per transect by researchers performing visual surveillance for free-roaming cats by use of the line-transect method in on-campus (dark gray bars) and off-campus (light gray bars) regions.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Histogram of distances walked per transect by researchers performing visual surveillance for free-roaming cats by use of the line-transect method in on-campus (dark gray bars) and off-campus (light gray bars) regions.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the study site divided into 705 zones (542 and 163 zones on and off campus, respectively). The 100 zones selected for the surveillance of cats by use of the line-transect method and locations where cats were observed by this method are shown. Each point (red and black) depicts 1 cat sighting. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the study site divided into 705 zones (542 and 163 zones on and off campus, respectively). The 100 zones selected for the surveillance of cats by use of the line-transect method and locations where cats were observed by this method are shown. Each point (red and black) depicts 1 cat sighting. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the study site divided into 705 zones (542 and 163 zones on and off campus, respectively). The 100 zones selected for the surveillance of cats by use of the line-transect method and locations where cats were observed by this method are shown. Each point (red and black) depicts 1 cat sighting. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Trail-camera method
The mean distance of lures from trail cameras was 4.8 m (median, 4.4 m; range, 2.3 to 7.6 m). The greatest distance from which a cat was detected by a given camera ranged from 3.5 to 15.8 m; in each case, the cat at the greatest distance from the camera was behind the lure station (ie, farther away from the camera than the lure station). No animals of any species were detected by cameras in 3 zones. Of those 3, 2 were in on-campus mowed-grass areas and 1 was in an on-campus urban area. With the trail-camera method, 92 cat sightings were recorded for 23 zones sampled (Figure 4). Cats were photographed in 9 different zones (5 on campus and 4 off campus). Only 1 lure station was tampered with on 1 night during the study. No cameras or lure stations were stolen, so complete data were recorded except for 9 hours when the affected lure station was displaced.
Map of the same zones as in Figure 3 with the 23 zones selected for evaluation by trail cameras depicted in blue. The number of cat sightings (ie, still images of cats) obtained with this method is reflected by increasing color intensity. A total of 92 sightings were recorded by the cameras. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the same zones as in Figure 3 with the 23 zones selected for evaluation by trail cameras depicted in blue. The number of cat sightings (ie, still images of cats) obtained with this method is reflected by increasing color intensity. A total of 92 sightings were recorded by the cameras. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the same zones as in Figure 3 with the 23 zones selected for evaluation by trail cameras depicted in blue. The number of cat sightings (ie, still images of cats) obtained with this method is reflected by increasing color intensity. A total of 92 sightings were recorded by the cameras. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Comparison of field research methods
The trail-camera method recorded presence of cats in 8 zones where no cats were detected with the line-transect method. Controlling for the longer time of viewing by the trail cameras (3,855 hours vs 11.2 hours for the line-transect surveys), the rate ratio analysis revealed that line transects had 23.0 times as many cat sightings/h (95% confidence interval, 8.89 to 48.26; P < 0.001) as did the trail-camera method.
Determinants of cat presence
On the basis of trail camera data, significantly (P = 0.004) more cat sightings were recorded off campus (n = 89) than on campus (3). Most sightings were recorded in urban (86/92 [93%]) rather than unmowed (2 [2%]), mowed grass (2 [2%]), agricultural (1 [1%]), or wooded (1 [1%]) zones, but overall, habitat type and the combination of habitat type and on- or off-campus location were not significantly (P = 0.34 and 0.054, respectively) associated with the number of sightings.
More food resources were recorded off campus (44.5 resources/km2) than on campus (4.80 resources/km2; Figure 5). Resources found were exclusively dumpsters and trash cans overflowing with food waste. No signs of residents feeding cats were observed, although this activity was reported anecdotally. There was a significant (P = 0.004) positive correlation (r = 0.58; 95% confidence interval, 0.19 to 0.81) between the number of cat sightings with a trail camera and the number of food resources in the area (ie, the selected and adjacent zones).
Map of the same zones as in Figure 3 showing the locations of food resources identified by a researcher and mapped by use of a global positioning system device. All identified food resources consisted of open dumpsters or trash cans containing food waste. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the same zones as in Figure 3 showing the locations of food resources identified by a researcher and mapped by use of a global positioning system device. All identified food resources consisted of open dumpsters or trash cans containing food waste. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Map of the same zones as in Figure 3 showing the locations of food resources identified by a researcher and mapped by use of a global positioning system device. All identified food resources consisted of open dumpsters or trash cans containing food waste. Used with permission. Copyright © 2017 Esri, ArcGIS Online, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo and the GIS User Community. All rights reserved.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Other species photographed by trail cameras were dogs (Canis familiaris), opossums (Didelphis virginiana), gray squirrels (Sciurus carolinensis), raccoons (Procyon lotor), geese (Canada geese [Branta canadensis] and domestic hybrids), foxes (Vulpes vulpes), groundhogs (Marmota monax), birds (various songbirds and perching birds), cottontail rabbits (Sylvilagus floridanus), and white-tailed deer (Odocoileus virginianus). The number of sightings for individual species was not significantly correlated with the number of cat sightings within a zone (Table 1).
Results of Spearman correlation analysis for association between the number of cat sightings and number of sightings of other animals by use of trail cameras in a study to evaluate 2 field research methods for surveying free-roaming cats (Felis catus; line transect vs trail camera) and identify factors potentially associated with the presence of such cats in a mixed-urban environment.
| Species | No. of sightings | r | P value |
|---|---|---|---|
| Dog (Canis familiaris) | 33 | 0.42 | 0.059 |
| Opossum (Didelphis virginiana) | 5 | 0.27 | 0.25 |
| Gray squirrel (Sciurus carolinensis) | 3 | 0.14 | 0.53 |
| Raccoon (Procyon lotor) | 40 | 0.060 | 0.80 |
| Goose (Branta canadensis and domestic hybrids) | 14 | −0.19 | 0.42 |
| Fox (Vulpes vulpes) | 6 | −0.27 | 0.24 |
| Groundhog (Marmota monax) | 16 | −0.34 | 0.13 |
| Bird (various species) | 27 | −0.36 | 0.11 |
| Cottontail rabbit (Sylvilagus floridanus) | 63 | −0.37 | 0.095 |
| White-tailed deer (Odocoileus virginianus) | 23 | −0.42 | 0.059 |
Fifty-eight cat sightings were recorded by trail cameras between 9:00 pm and 5:59 am, compared with 34 sightings between 6:00 am and 8:59 pm. Rate ratio analysis with the median-unbiased estimation method indicated that the number of sightings per trap-hour during the night was 2.86 times the number during the day (95% confidence interval, 1.88 to 4.41; P < 0.001). There was a peak in cat sightings around 11 pm and again near 5 am (Figure 6).
Number of cat sightings by use of trail cameras during daytime and nighttime hours.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Number of cat sightings by use of trail cameras during daytime and nighttime hours.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Number of cat sightings by use of trail cameras during daytime and nighttime hours.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.745
Discussion
In the mixed-urban environment of the present study, the trail-camera method was preferable to the line-transect method for surveying free-roaming cats. Not only did the trail-camera method yield higher counts of cats, but cats were identified in 9 of 23 (39%) zones sampled versus 5 of 100 (5%) with the line-transect method. The trail-camera method documented the presence of cats in all land-use (or habitat) types, whereas cats were detected only in urban habitats with the line-transect method. This finding may have been attributable to the greater temporal coverage that the trail cameras provided (given that the rate ratio [accounting for time differences between methods] was 23.0 in favor of the line-transect method). Subjectively, it did not appear that cats were unintentionally frightened by researchers performing surveillance with the line transect method, and the time spent actively looking for cats in zones may have contributed to the finding that this method resulted in a greater number of sightings per observation hour than the trail-camera method. The trail-camera method was also less labor intensive, potentially making it more feasible for communities wishing to collect baseline data on free-roaming cat populations, depending on the cost of the cameras used and availability of reliable volunteers. Although theft and vandalism of trail cameras were reported as a substantial problem in a similar study,26 tampering with equipment (involving 1 lure station) occurred only once in the present study. Placement of the trail cameras after the line-transect surveys had been completed in the same zones might have introduced bias in the selection of camera sites. We placed the trail cameras by attaching them to existing poles or trees in the zone and positioning them in a location where the lure station could be placed at an appropriate distance from the camera and away from any roads or walkways. Because there were typically very few locations in a zone that met these criteria, and because only 1 zone where a cat was detected with the line-transect method was sampled by the trail-camera method, we believed that bias was not a substantial problem in the study. However, in future studies comparing such methods, it would be preferable to have 2 groups of researchers applying different methods and sharing results only after all data collection is concluded. In addition, owing to resource limitations, 4 different models of trail cameras were used in this study, but as different cameras were not deployed in the same sites, we did not directly compare the efficacy of these cameras. It would be advisable in future studies to use the same model of trail camera for each site.
The trail-camera surveillance included placement of lure stations to draw in nearby cats that otherwise might not have passed in front of the camera, and this may have mitigated potential placement bias to some extent. Multiple types of olfactory lures have been used to attract domestic cats to trail cameras, but catnip has been shown to be effective in increasing the amount of time that cats spend in front of the camera.44 Catnip is a preferable lure over items such as tuna or cat food because it can more specifically target felids over other carnivores.44 Pictures taken by the trail cameras in this study showed cats investigating the lures, rubbing their faces on them, rolling on the ground nearby, and sometimes staying for > 10 minutes. We did not include trail cameras without lures or line transects with lures for comparison, which may have added a bias in the comparison between the line-transect and trail-camera methods. In the line-transect method, the entire zone sometimes could not be visualized because of thick vegetation, privacy fencing, and other obstructions, so the use of lures might improve population estimates with either method. However, cats' reactions to catnip are not universal and may depend on factors such as genetics, demeanor, and stage of the reproductive cycle.45 Thus, although catnip has been shown to be a good attractant for cats,44 it has the potential to introduce some bias in sampling and does not guarantee viewing of all cats in a zone.
Although the line-transect surveillance was conducted just after sunrise, results of the trail-camera surveillance indicated that there was greater free-roaming cat activity in our study site during the night than during the day (P < 0.001), especially around 11 pm and 5 am. These differed slightly from the peaks at 1 am and 7 am reported for a study19 in New York City, with differences possibly attributable to factors such as the time of year or timing of human activity. Researchers in our study also noted while walking on transect paths that building and road construction projects adjacent to transect sites had usually started before they arrived. For future studies, it is important to realize that free-roaming cats may have different activity patterns in different cities, so it may be difficult to predict at what time of day the line-transect method may be most effective. The trail-camera method allows researchers to survey for the presence of cats at all times of the day and night without specific knowledge of regional cat activity patterns and might be useful prior to implementing line-transect methods, which may allow for a more detailed assessment of populations.
In addition, the trail-camera method may prove useful in that it allows researchers to simultaneously collect data on other wildlife species while studying free-roaming cats, even those that are not active at the same time of day as cats. This information may be useful from a public-health perspective, as it can provide information about populations of urban wildlife species such as raccoons that may be sharing the environment with domestic animals and people. Further research into interactions between free-roaming cat populations and populations of urban wildlife may be useful. Although various wildlife species were detected with trail cameras in the present study, we did not find any significant associations between the number of cat sightings and sightings of these other species.
The most important limitation of the trail-camera method in the present study was the inability to identify each individual cat. The infrared cameras took only black-and-white photographs during the night, making it difficult to individually identify each cat unless it had distinctive markings. Infrared cameras were chosen in this study instead of cameras that use a white flash because the flash may frighten animals and reduce sightings,46 but this choice came at the cost of better cat identification. Cats also moved quickly in front of the cameras, producing a blurry image that made identification difficult. If each individual cat could be identified from trail-camera photographs, mark-resight methods could be used to make a more accurate estimate of actual population sizes in an area.29 Although individual animal identification could be made by use of the line-transect method, so few cats were seen by investigators using that method in the present study that mark-resight methods were still not considered useful. Population-based calculations could not be performed with either method in this study, and this limited the data to relative abundance estimates. Still, the relative differences in sightings among locations were considered accurate and can be useful for general study planning purposes in terms of which sites have more free-roaming cat activity. In addition, future studies in the same location can repeat the trail camera experiments to assess longitudinal changes in cat sightings over time, and this information may help to inform management strategies and assessment of the effectiveness of population control programs. This quality of data is not useful for establishing population estimates, but can aid planning of more detailed studies.
An important finding of the present study was that the number of cat sightings in a given zone was correlated with the density of food resources in the same zone and adjacent areas. Although results of another study24 revealed a correlation between cat feeding stations and the presence of cats, to our knowledge, the present study was the first that found a correlation between anthropogenic food waste and free-roaming cat presence. In support of this finding, 4 of 6 sightings by researchers using the line-transect method were of cats at dumpsters or interacting with trash. This information is important to note for communities implementing free-roaming cat population control programs, as it suggests that actions as simple as placing lids on dumpsters and removing food waste regularly might reduce the presence of free-roaming cats. Although the inclusion of resources in areas adjacent to the survey zones was meant to reflect a typical home range for urban cats,24 the home range for cats in substantially different land-use categories or habitats (eg, agricultural, wooded, or unmowed grass zones) might have been underestimated.36,38 Although there were subjectively fewer cat sightings and fewer resources found in nonurban zone categories, and we identified no significant association between land-use categories and the number of cat sightings, the presence of trash at a greater distance may influence the presence of cats in these zones. This study included only a cross-sectional survey of potential food resources, so further investigation into availability of resources throughout the week (eg, trash pick-up schedules) and at different times of the year (eg, presence or absence of students and dates of parties) as well as formal documentation of cats using trash or small mammals around trash as food might provide more detail regarding specific resources that promote the presence of free-roaming cats. Future studies could also investigate whether implementing waste management programs in a community impacts the carrying capacity for free-roaming cats or whether the food waste is an indicator of other unobserved variables influencing cat presence, such as human presence and supplemental feeding.
The present study was conducted on and near a university campus when undergraduate students were typically not present (May through August). As such, the data may not have represented the conditions present in this location for the remaining months of the year. By conducting the survey during the summer, we attempted to remove many sources of variation in the cat populations associated with the presence of students, so these results would more likely be generalizable to a mixed urban environment with a more consistent population. Additional studies are recommended to examine seasonal variation in population dynamics and the potential impact of transient student populations in this region.
Overall, the results of this study suggested that the trail-camera method, including the use of lure stations, is a useful tool for surveillance of free-roaming cats in a mixed-urban environment that can provide information about relative abundance or activities of such cats without requiring trapping and manipulation of the animals. Although the quality of data obtained by this method in the present study was not sufficient for population estimates, the information can be of use in designing such studies going forward.
Acknowledgments
Supported in part by an NIH T35 Training Grant awarded to Ms. Vincent.
The authors declare that there were no conflicts of interest.
Presented in abstract and poster form at the Merial-NIH National Veterinary Scholars Symposium, Columbus, Ohio, July 2016, and The Ohio State University College of Veterinary Medicine's Advances in Veterinary Medicine Research Day, Columbus, Ohio, April 2017.
The authors thank Christina Voise for technical assistance with online mapping tools used in the study and Stan Gehrt for use of field equipment and for technical advice.
Footnotes
The Ohio State University Planning and Real Estate website. GIS maps. Available at gismaps.osu.edu. Accessed Jul 13, 2017.
RANDBETWEEN, Microsoft Excel 2007, Microsoft Corp, Redmond, Wash.
Map My Walk for iPhone, 2016. MapMyFitness Inc, Austin, Tex.
HC600 Hyperfire, Reconyx Inc, Holman, Wis.
Moultrie M80 Gamespy, EBSCO Industries, Birmingham, Ala.
Trophy camera, Bushnell Outdoor Products, Overland Park, Kan.
Spypoint Force-12 ultra-compact trail camera, GG Telecom, Swanton, Vt.
eTrex 30, Garmin Ltd, Schaffhausen, Switzerland.
ArcMap, version 10.2.2, Esri, Redlands, Calif.
R Studio, version 0.99.902, RStudio Inc, Boston, Mass.
R, version 3.1.3, R Foundation for Statistical Computing, Vienna, Austria. Available at: www.R-project.org. Accessed Jun 26, 2016.
Epitools: Epidemiology Tools. R package version 0.5-7. Available at: CRAN.R-project.org/package=epitools. Accessed Jan 30, 2018.
References
1. Wallace JL, Levy JK. Population characteristics of feral cats admitted to seven trap-neuter-return programs in the United States. J Feline Med Surg 2006;8:279–284.
2. Nutter FB, Dubey JP, Levine JF, et al. Seroprevalences of antibodies against Bartonella henselae and Toxoplasma gondii and fecal shedding of Cryptosporidium spp, Giardia spp, and Toxocara cati in feral and pet domestic cats. J Am Vet Med Assoc 2004;225:1394–1398.
3. Gerhold RW, Jessup DA. Zoonotic diseases associated with free-roaming cats. Zoonoses Public Health 2013;60:189–195.
4. Bonnaud E, Medina FM, Vidal E, et al. The diet of feral cats on islands: a review and a call for more studies. Biol Invasions 2011;13:581–603.
5. Pearre S, Maass R. Trends in the prey size-based trophic niches of feral and house cats Felis catus L. Mammal Rev 1998;28:125–139.
6. Woods M, Mcdonald RA, Phen STE, et al. Predation of wildlife by domestic cats Felis catus in Great Britain. Mammal Rev 2003;33:174–188.
7. Driscoll CA, Menotti-Raymond M, Roca AL, et al. The Near Eastern origin of cat domestication. Science 2007;317:519–523.
8. Loss SR, Will T, Marra PP. The impact of free-ranging domestic cats on wildlife of the United States. Nat Commun 2013;4:1396.
9. Slater MR. Understanding issues and solutions for unowned, free-roaming cat populations. J Am Vet Med Assoc 2004;225:1350–1354.
10. Levy JK, Gale DW, Gale LA. Evaluation of the effect of a long-term trap-neuter-return and adoption program on a free-roaming cat population. J Am Vet Med Assoc 2003;222:42–46.
11. Levy JK, Isaza NM, Scott KC. Effect of high-impact targeted trap-neuter-return and adoption of community cats on cat intake to a shelter. Vet J 2014;201:269–274.
12. Natoli E, Maragliano L, Cariola G, et al. Management of feral domestic cats in the urban environment of Rome (Italy). Prev Vet Med 2006;77:180–185.
13. Gunther I, Finkler H, Terkel J. Demographic differences between urban feeding groups of neutered and sexually intact free-roaming cats following a trap-neuter-return procedure. J Am Vet Med Assoc 2011;238:1134–1140.
14. Kilgour RJ, Magle SB, Slater M, et al. Estimating free-roaming cat populations and the effects of one year trap-neuter-return management effort in a highly urban area. Urban Ecosyst 2017;20:207–216.
15. Boone JD. Better trap–neuter–return for free-roaming cats: Using models and monitoring to improve population management. J Feline Med Surg 2015;17:800–807.
16. Calhoon RE, Haspel C. Urban cat populations compared by season, subhabitat and supplemental feeding. J Anim Ecol 1989;58:321–328.
17. Warner RE. Demography and movements of free-ranging domestic cats in rural Illinois. J Wildl Manage 1985;49:340–346.
18. Forsyth DM, Robley AJ, Reddiex B. Review of methods used to estimate the abundance of feral cats. Arthur Rylah Institute for Environmental Research, Department of Sustainability and Environment, Melbourne, Australia, 2005. Available at: www.environment.gov.au/system/files/resources/20855c98-c575–4130–9a23-c54200526a2a/files/feral-cats-review.pdf. Accessed Jan 21, 2018.
19. Hapsel C, Calhoon RE. Activity patterns of free-ranging cats in Brooklyn, New York. J Mammal 1993;74:1–8.
20. Boone JD, Slater M. A generalized population monitoring program to inform the management of free-roaming cats. Alliance for Contraception in Dogs and Cats, November 29, 2014 revision. Available at: www.acc-d.org/docs/default-source/think-tanks/frc-monitoring-revised-nov-2014.pdf. Accessed Jan 30, 2018.
21. Audubon Society of Portland. Hayden Island free-roaming cat survey protocol. Available at: audubonportland.org/files/research/cat-protocol/at_download/file. Accessed Mar 15, 2017.
22. Lessa ICM, Bergallo HG. Modelling the population control of the domestic cat: an example from an island in Brazil. Braz J Biol 2012;72:445–452.
23. Say L, Gaillard J-M, Pontier D. Spatio-temporal variation in cat population density in a sub-Antarctic environment. Polar Biol 2002;25:90–95.
24. Tennent J, Downs CT. Abundance and home ranges of feral cats in an urban conservancy where there is supplemental feeding: a case study from South Africa. Afr Zool 2008;43:218–229.
25. Caravaggi A, Banks PB, Burton AC, et al. A review of camera trapping for conservation behaviour research. Remote Sens Ecol Conserv 2017;3:109–122.
26. Meek PD, Ballard GA, Falzon G. The higher you go the less you will know: placing camera traps high to avoid theft will affect detection. Remote Sens Ecol Conserv 2016;2:204–211.
27. Meek P, Ballard G, Fleming P, et al. Are we getting the full picture? Animal responses to camera traps and implications for predator studies. Ecol Evol 2016;6:3216–3225.
28. Foster RJ, Harmsen BJ. A critique of density estimation from camera-trap data. J Wildl Manage 2012;76:224–236.
29. Elizondo EC, Loss SR. Using trail cameras to estimate free-ranging domestic cat abundance in urban areas. Wildl Biol 2016;22:246–252.
30. Hofmeester TR, Rowcliffe JM, Jansen PA. A simple method for estimating the effective detection distance of camera traps. Remote Sens Ecol Conserv 2017;3:81–89.
31. Sollmann R, Tôrres NM, Furtado MM, et al. Combining camera-trapping and noninvasive genetic data in a spatial capture–recapture framework improves density estimates for the jaguar. Biol Conserv 2013;167:242–247.
32. Lazenby BT, Mooney NJ, Dickman CR. Effects of low-level culling of feral cats in open populations: a case study from the forests of southern Tasmania. Wildl Res 2014;41:407–420.
33. Stokeld D, Frank ASK, Hill B, et al. Multiple cameras required to reliably detect feral cats in northern Australian tropical savanna: an evaluation of sampling design when using camera traps. Wildl Res 2015;42:642–649.
34. Lazenby BT, Dickman CR. Patterns of detection and capture are associated with cohabiting predators and prey. PLoS One 2013;8:e59846.
35. Bengsen A, Butler J, Masters P. Estimating and indexing feral cat population abundances using camera traps. Wildl Res 2011;38:732–739.
36. McGregor HW, Legge S, Potts J, et al. Density and home range of feral cats in north-western Australia. Wildl Res 2015;42:223–231.
37. Garrote G, Gil-Sanchez JM, McCain EB, et al. The effect of attractant lures in camera trapping: a case study of population estimates for the Iberian lynx (Lynx pardinus). Eur J Wildl Res 2012;58:881–884.
38. Gehrt SD, Wilson EC, Brown JL, et al. Population ecology of free-roaming cats and interference competition by coyotes in urban parks. PLoS One 2013;8:e75718.
39. Horn JA, Mateus-Pinilla N, Warner RE, et al. Home range, habitat use, and activity patterns of free-roaming domestic cats. J Wildl Manage 2011;75:1177–1185.
40. United States Census Bureau website. Available at: www.census.gov/quickfacts/table/PST045215/3918000. Accessed Nov 13, 2016.
41. United States Census Bureau website. Metropolitan and micropolitan statistical areas. Available at: factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk. Accessed Jan 21, 2018.
42. The Ohio State University website. Statistical summary. Available at: www.osu.edu/osutoday/stuinfo.php. Accessed Aug 1, 2016.
43. Sikes RS, Gannon WL, Animal Care and Use Committee of the American Society of Mammologists. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. J Mammal 2011;92:235–253.
44. Clapperton BK, Eason CT, Weston RJ, et al. Development and testing of attractants for feral cats, Felis catus L. Wildl Res 1994;21:389–399.
45. Tucker AO, Tucker SS. Catnip and the catnip response. Econ Bot 1988;42:214–231.
46. Glen AS, Cockburn S, Nichols M, et al. Optimising camera traps for monitoring small mammals. PLoS One 2013;8:e67940.
Appendix
Mean and range values for area (m2) of 100 sampled zones categorized by land-use type in a study to evaluate 2 field research methods for surveillance of free-roaming cats (Felis catus) and identify factors potentially associated with the presence of such cats in the mixed-urban environment of a large university campus.
| On campus | Off campus | |||||
|---|---|---|---|---|---|---|
| Land-use category | No. of zones | Mean | Range | No. of zones | Mean | Range |
| Mowed grass | 10 | 12,726 | 5,687–19,723 | 5 | 30,067 | 12,545–80,929 |
| Urban | 30 | 16,158 | 3,857–38,844 | 25 | 27,656 | 10,479–41,050 |
| Agricultural | 7 | 19,907 | 5,831–45,840 | 0 | — | — |
| Wooded | 12 | 27,604 | 5,413–6,493 | 0 | — | — |
| Unmowed grass | 9 | 30,064 | 7,314–82,481 | 2 | 43,026 | 29,121–56,931 |
| All | 68 | 19,899 | 3,857–82,481 | 32 | 28,994 | 10,479–80,929 |
All 100 zones were surveyed for the presence of free-roaming cats by use of the line-transect method; 23 zones (6, 6, 3, 3, and 5 from the mowed grass, urban, agricultural, wooded, and unmowed grass categories, respectively) were subsequently surveyed by use of the trail-camera method.
— = Not applicable.
