The North American beaver Castor canadensis is widely recognized for its ability to modify freshwater habitats and facilitate changes in community composition. However, the seasonal composition of terrestrial wildlife at littoral beaver lodges remains poorly described, even though beaver lodges are distinctive semipermanent features of the terrestrial–aquatic interface and thus important resources for wildlife. Over 17 months, we used camera trapping, weather data, and satellite vegetation data to determine how vertebrate species richness and seasonal changes in community composition are associated with beaver activity and beaver lodges in a temperate artificial pond. Our results indicate clear changes in the composition of beaver lodge visitors across seasons. Moreover, there was not a strong association of species richness with beaver activity, vegetative condition, or weather condition. Littoral beaver lodges are likely important foraging sites for a wide range of taxa throughout the year. Our findings highlight the importance of beaver lodges in facilitating seasonal interactions and variation in species composition. We hope our work can be used as a baseline to investigate the importance of beaver lodges in promoting diversity at the terrestrial–aquatic interface.
The North American beaver Castor canadensis is an important species in freshwater ecosystems. Beavers are considered ecosystem engineers—that is, species that alter ecosystem dynamics by modifying their surrounding habitat (Wright et al. 2002). They construct large dams in temperate and boreal freshwater systems, thereby expanding wetlands, driving nutrient cycling and decomposition dynamics, and altering community composition (Naiman et al. 1988; McKinstry et al. 2001; Rosell et al. 2005; Thompson et al. 2016). The ability of beavers to dam and flood areas that overlap with human settlements has also led to a long history of poor public perception and the development of various beaver removal methods (Peterson and Payne 1986; Huffaker et al. 1992; Bhat et al. 1993; Siemer et al. 2013).
Evidence demonstrates that beavers and beaver lodges facilitate the occupancy of many animals, including waterfowl, macroinvertebrates, amphibians, and small mammals (France 1997; Rosell et al. 2005; Stevens et al. 2007; Ulevičius and Janulaitis 2007; Bromley and Hood 2013; Glabischnig 2015). Aquatic restoration studies have therefore emphasized the importance of beavers in the conservation and preservation of freshwater biodiversity (Gibson and Olden 2014; Law et al. 2017; Willby et al. 2018). Despite this, we still do not sufficiently understand how many species, especially terrestrial vertebrates, might interact and associate with beaver dams and lodges over broad timescales.
Monitoring animal associations can be difficult without invasive or intensive sampling efforts. Camera trapping is a minimally invasive method to study animal presence and activity. Camera traps are efficient and cost-effective tools for analysis of occupancy, density, activity patterns, and habitat associations, among others (O'Connell et al. 2011; Rovero et al. 2013). Although infrared camera trapping cannot be used to study fully aquatic or ectothermic animals, this method allows for long-term monitoring of terrestrial mammals and birds (O'Connell et al. 2011; Rovero et al. 2013). In this study, we used camera traps to monitor vertebrate species richness and beaver activity at littoral beaver lodges in a temperate artificial pond. We additionally collected weather and satellite vegetation data to better understand how these environmental variables are associated with seasonal changes in beaver activity and visitor composition at beaver lodges. Results suggest a potential role of beaver lodges in facilitating interactions between species in temperate freshwater habitats. Overall, our results contribute to an understanding of the importance of beaver lodges in freshwater ecosystem functioning and restoration.
Our study took place at Rice Creek Field Station, Oswego, New York. The field station, initially surrounded by agricultural fields and young evergreen plantations, began operation in 1966 (Weeks 1988). The area now consists of approximately 121 ha of preserved forests, fields, ponds, and streams (State University of New York at Oswego, 2020a). A large, irregularly shaped artificial pond (Rice Pond, established 1966) is controlled by an earthen dam and spillway and connects to Lake Ontario 2.4 km downstream via Rice Creek (Weeks 1987, 1988). It is also fed by Rice Creek, which originates 13 km upstream. The creek and its watershed have a long history of periodic flooding, causing relatively high sedimentation toward the head of the pond (Weeks 1988). The pond covers approximately 9.2 ha and averages between 0.6 and 0.76 m in depth, and the maximum depth can reach approximately 2.29 m closer to the spillway (Weeks 1988). It contains patchy and dense aquatic vegetation, including milfoil Myriophyllum sp., coontail Ceratophyllum demersum, pondweed Potamogeton sp., and duckweed Lemna sp. (Weeks 1988). Algae during the dry season can further cover large portions of the pond. The shallow perimeter is colonized by common reed Phragmites australis, common cattail Typha latifolia, pickerelweed Pontederia cordata, arrowhead Sagittaria sp., arrow arum Peltandra virginica, bulrush Schoenoplectus acutus, sedges Carex albicans, and pond lily Nuphar microphylla. Adjacent terrestrial vegetation includes silky dogwood Cornus amomum, red osier dogwood Cornus sericea, multiflora rose Rosa multiflora, speckled alder Alnus incana, maple Acer saccharum, elm Ulmus americana, and white ash Fraxinus americana (Weeks 1988; I. Rasik, personal observation).
To determine presence and activity of wildlife at beaver lodges, we positioned a high-definition infrared camera (Reconyx Hyperfire™, model HC500) at two littoral beaver lodges present at Rice Pond. The cameras have an infrared flash range of up to 18 m and a 0.2-s trigger speed. We programmed each camera to obtain a series of three photos at 1-s intervals when triggered, each with date and time stamps. We fastened the cameras to metal stakes and positioned them above the water level, approximately 2 m away from each beaver lodge. To minimize disturbance, we retrieved camera trap data approximately once per month. We monitored the first beaver lodge (lodge 1) for a total of 10 mo (from October 2016 to May 2017, from August 2017 to September 2017). The second beaver lodge (lodge 2) was monitored continuously from October 2016 to February 2018, for a total of 17 mo. Because of differences in sampling effort, we analyzed the data from each beaver lodge separately.
Species richness at beaver lodges
We used camera trap detections of local wildlife to estimate species richness at beaver lodges. To determine the potential effects of seasonal weather on species richness, we averaged air temperature, humidity, and water temperature for each month by using the Rice Creek Field Station weather station (Weather Station Data Archive, Davis Instruments). We selected Temp Out, Out Hum, and Soil Temp 1 for our variables (the lattermost sensor is at pond bottom). We recorded values for each variable automatically on a minute-by-minute basis, 24 h a day (data available at State University of New York at Oswego, 2020b). We then computed multiple regressions in R (lm function from the Stats package 3.6.2; R Core Team 2019) to evaluate the relationship between species richness and weather variables.
To understand the relationship between seasonal variation in vegetative condition and species richness, we calculated normalized differential vegetation indices (NDVIs) to approximate the monthly vegetative condition surrounding Rice Pond. We downloaded Landsat 7 Thematic Mapper+ multispectral imagery at a resolution of 30 m from the U.S. Geological Survey Global Visualization and EarthExplorer remote sensing data archives (U.S. Geological Survey 2020a, 2020b). We chose this resolution based on the size of the pond and the extension of the surrounding forested area. We imported red light (λ = 0.63–0.69 μm) and near infrared (NIR; λ = 0.76–0.90 μm) bands into the geographic information system software ENVI Classic (Harris Spatial Solutions 5.3). We then calculated NDVI with the formula NDVI = (NIR − red)/(red + NIR) (Rouse et al. 1974). NDVI values for each cell range from −1 to 1: positive values indicate healthy green foliage; values close to 0 indicate soil, snow, or clouds; and negative values indicate the presence of water (Neigh et al. 2008). For each month of satellite imagery, we recorded five NDVI values around the forested vicinity of Rice Pond and then averaged these values to approximate NDVI. We calculated average NDVI for the following dates: 5 October 2016, 19 February 2017, 23 March 2017, 24 April 2017, 17 May 2017, 11 June 2017, 21 August 2017, 22 September 2017, 17 October 2017, 4 December 2017, 28 January 2018, and 13 February 2018. Using these data, we computed a linear regression in R (lm function from the Stats package 3.6.2; R Core Team 2019) to analyze the relationship between richness and NDVI.
Beaver activity and species richness
To determine beaver activity patterns and the temporal relationship between beaver activity and species richness, we used all beaver photos from the duration of the study. To describe the relationship between environmental variables and beaver activity, we again used average monthly values for air temperature, humidity, and water temperature and computed a multiple linear regression in R (lm function from the Stats package 3.6.2; R Core Team 2019). To evaluate the relationship between seasonal variation in vegetative condition and beaver activity, we performed a simple linear regression in R between NDVI and detection count. We log transformed beaver detection counts because monthly detection counts were zero inflated. Finally, we used a linear regression in R to test for the effect of beaver detection count on species richness.
Species richness at beaver lodges
We captured 11,944 photos from the two beaver lodges (lodge 1: 5,845 photos; lodge 2: 6,099 photos). Of the total photo captures, 1,425 were of beavers (lodge 1: 135 photos; lodge 2: 1,290 photos). The local beaver population at Rice Pond was approximately 6 ± 2 individuals at the start of the study.
Monthly detection counts for all species at both lodge 1 and lodge 2 are available in Data S1 (Supplemental Material). Overall, we detected 16 species visiting lodge 1 (Table 1) and 26 species visiting lodge 2 (Table 2). At lodge 1, Canada goose Branta candensis was the most prevalent species, followed by beaver and the North American raccoon Procyon lotor. At lodge 2, beaver was the most prevalent species, followed by raccoon and red fox Vulpes vulpes (Figure 1). Most species were locally common; however, there were several species that are regionally rare, such as the fisher Martes pennanti and the North American river otter Lontra canadensis. To our knowledge, the North American river otter has not been previously documented at the study site. We also detected a red-tailed hawk Buteo jamaicensis at lodge 2. During autumn, the most common species at lodge 1 include muskrat Ondatra zibethicus and cottontail rabbit Sylvilagus floridanus. During autumn at lodge 2, the most common species were beaver and raccoon. The number of visits by mammalian predators, including several mustelid and canid species, generally increased during winter. In spring, detections increased for Canada goose, beaver, common grackle Quiscalus quiscula, great blue heron Ardea herodias, and white-tailed deer Odocoileus virginianus at both beaver lodges, among others. The most common species observed during summer at lodge 2 were raccoon and common grackle, and several other migratory bird species also appeared at this time.
Although species composition varied seasonally, we did not find a significant relationship between species richness and air temperature, humidity, or water temperature at lodge 1 (F3,6 = 0.9823; P = 0.4614; R2 = −0.005925) or at lodge 2 (F3,13 = 0.1661; P = 0.9173; R2 = −0.1853). We also did not find a relationship between species richness and average NDVI at lodge 1 (F1,5 = 3.42; P = 0.1237; R2 = 0.2874) or at lodge 2 (F1,11 = 0.01304; P = 0.9112; R2 = −0.08962). We provide summarized data for species richness and environmental variables in Data S2 (Supplemental Material).
Beaver activity and species richness
Monthly detection counts of beaver were highest during spring and autumn; detection counts decreased during winter and summer (Figure 2). We did not see any signs of activity at lodge 1 during summer 2017 so we did not deploy a camera; however, weekly observations appeared to indicate that the beaver lodge had been abandoned and was not reoccupied before the end of the study. Although the predictive power of vegetative condition was not strong, beaver detection tended to be higher at low NDVI values at lodge 1 (n = 3; F1,1 = 98.04; P = 0.06408; R2 = 0.9798; Figure S1, Supplemental Material). At lodge 2, beaver detection count had a tendency to be positively correlated with NDVI (n = 10; F1,8 = 4.063; P = 0.07857; R2 = 0.2539; Figure 3). Detection counts were not explained by monthly variation in air temperature, humidity, or water temperature at lodge 1 (F3,2 = 0.52; P = 0.7099; R2 = −0.4045) or at lodge 2 (F3,8 = 0.2327; P = 0.8711; R2 = −0.2646). Furthermore, we found no significant relationship between beaver detection count and species richness at lodge 1 (F1,4 = 1.369; P = 0.307; R2 = 0.06866) or at lodge 2 (F1,10 = 0.1443; P = 0.712; R2 = −0.08435).
The North American beaver is an ecosystem engineer that often facilitates long-term changes in community composition. Although this topic has received considerable interest over the past few decades, we still lack thorough descriptions of how species interact with beavers and beaver lodges in freshwater habitats. In our study, we used camera trapping to quantify seasonal variation in the composition of terrestrial visitors at littoral beaver lodges in a temperate artificial pond. We found that this variation was not directly associated with beaver activity, abiotic environmental variables, or seasonal variation in vegetative condition.
In highly seasonal environments, where food availability varies on a spatiotemporal scale, beaver lodges are persistent natural sources of accumulated woody debris and other organic matter. Therefore, beaver lodges may act as semipermanent foraging resources for many animal taxa, including those that may be regionally rare or uncommon. At our study site, mammalian predators were commonly associated with beaver lodges during winter, whereas larger avian predators were more commonly observed during summer and autumn. Beaver lodges can also serve as refuge for many small mammals, such as mice, shrews, and voles (Ulevičius and Janulaitis 2007; Glabischnig 2015). In addition, several exothermic species that cannot be detected with infrared camera trapping were observed at beaver lodges, including painted turtles Chrysemys picta, green frogs Lithobates clamitans, and northern water snakes Nerodia sipedon. Although it was not within the scope of this study to determine the abundance or richness of invertebrate, amphibian, or fish species, these organisms are known to associate with deadwood in freshwater systems, and particularly, littoral beaver lodges (McLachlan 1970; Nilsen and Larimore 1973; France 1977; Moring et al. 1986). Thus, the presence of prey species at beaver lodges has the potential to attract a wide variety of mammalian, reptilian, and avian predators, thereby helping to explain diversity at littoral beaver lodges year-round.
At our study site, beaver detection count appeared to be more closely associated with seasonal variation in vegetative condition, compared with other abiotic environmental variables (air temperature, humidity, and water temperature). Although beaver detection counts showed considerable variation in relation to NDVI across beaver lodges, we generally had more detections during spring and autumn compared with summer and winter. Interestingly, detection counts were not directly proportional to vegetative condition (NDVI), which was greatest during spring and summer. Beavers forage more intensively for woody vegetation during autumn as they create a food cache to use throughout winter (Aleksiuk and Cowan 1969; Vorel et al. 2015). During spring (i.e., growing season) in the northeastern United States, beavers also increase their metabolic and physical activity to expand their foraging range and obtain additional resources (Aleksiuk and Cowan 1969). Increased snowmelt and water flow at this time further enhance the need for beaver lodge and dam maintenance (Eimers et al. 2008). Thus, resource availability (i.e., vegetative condition) and other biotic interactions (e.g., predator–prey or population density dynamics) appear to be the primary drivers of seasonal foraging and activity patterns of beavers (Svendsen 1980; Sun et al. 2000; DeStefano et al. 2006; Gallant et al. 2016).
Although beavers at our study site exhibited seasonal variation in activity, species richness at beaver lodges was not clearly associated with beaver activity. This indicates that community-level biotic interactions are more important for understanding seasonal variation in species richness and composition at beaver lodges. Thus, it is possible that beaver lodges help maintain local species richness in seasonal environments; however, to support this hypothesis, a larger sampling effort of species richness in relation to the surrounding habitat is required.
Here, we described seasonal variation in vertebrate species richness and beaver activity at littoral beaver lodges in a temperate artificial pond. Littoral beaver lodges, as semipermanent fixtures of the freshwater landscape, have an important role as a major foraging resource for both small prey and large predators throughout the year. Although it largely remains unknown to what extent littoral beaver lodges effectively attract local wildlife to the terrestrial–aquatic interface and drive biotic interactions, it is possible that these processes help to explain temporal species richness patterns in highly seasonal habitats. Moving forward, it is important to understand how beavers influence both early stages of artificial pond colonization as well as their effects on long-term community composition dynamics. We hope that our study will help in creating comprehensive beaver management protocols and improving freshwater habitat conservation plans.
Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.
Data S1. Detection counts per month, in number of days, for terrestrial vertebrate species that were observed at either of two littoral North American beaver Castor canadensis lodges between October 2016 and February 2018 in Rice Pond, Oswego, New York.
Found at DOI: https://doi.org/10.3996/092019-JFWM-078.S1 (19 KB XLSX).
Data S2. Species richness and North American beaver Castor canadensis detection counts per beaver lodge as well as weather and vegetative condition data per month between October 2016 and February 2018 in Rice Pond, Oswego, New York.
Found at DOI: https://doi.org/10.3996/092019-JFWM-078.S2 (14 KB XLSX).
Figure S1. Relationship between log-transformed monthly detection count of the North American beaver Castor canadensis and vegetative condition (normalized differential vegetation index) at a littoral beaver lodge in Rice Pond, Oswego, New York. The beaver lodge (lodge 1) was monitored via camera trap from October 2016 to May 2017 and from August 2017 to September 2017. We used Landsat 7 Thematic Mapper+ satellite imagery to quantify vegetative condition in the surrounding habitat. Vegetative condition may have an effect on log-transformed beaver detection count, although these variables would appear inversely related with limited data (F1,1 = 98.04; P = 0.06408; R2 = 0.9798).
Found at DOI: https://doi.org/10.3996/092019-JFWM-078.S3 (7 KB PDF).
Reference S1.Rouse JW, Haas RH, Schell JA, Deering DW. 1974. Monitoring vegetation systems in the Great Plains with ERTS. Pages 301–317 in Freden SC, Mercanti EP, Becker, MA, editors. Proceedings of the third earth resources technology satellite-1 symposium, Vol. 3. Washington, D.C.: National Aeronautics and Space Administration.
Found at DOI: https://doi.org/10.3996/092019-JFWM-078.S4 (16.47 MB PDF).
Reference S2.Weeks JA. 1987. Natural areas of Oswego County. Rice Creek Field Station Special Publication 2:1–36.
Found at DOI: https://doi.org/10.3996/092019-JFWM-078.S5 (1.61 MB XLSX).
Reference S3.Weeks JA. 1988. Guidelines for environmental management at Rice Creek Field Station. Bulletin No. 6. Rice Creek Field Station Bulletin 6:1–69.
Found at DOI: https://doi.org/10.3996/092019-JFWM-078.S6 (6.87 MB XLSX).
We thank the Rice Creek Small Grants Program and Rochester Academy of Science for providing funding and Rice Creek Field Station for equipment support throughout this project. We also thank Cayla Turner, Tenaja Smith-Butler, Matthew Upright, and Jerry Mahar for positive attitudes and valued assistance in the field. Finally, we thank the reviewers and the Associate Editor for comments and suggestions during the review process.
Any use of trade, product, website, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
Citation: Razik I, Sagot M. 2020. Vertebrate species richness at littoral beaver lodges in a temperate artificial pond. Journal of Fish and Wildlife Management 11(2):422-429; e1944-687X. https://doi.org/10.3996/092019-JFWM-078
The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.