The Influence of Heterogeneous Landscapes on Banded Mongoose (Mungos Mungo)

Home , Apidae

The Influence of Heterogeneous Landscapes on Banded Mongoose (Mungos mungo)

Behavior in Northern Botswana: Inferences about Infectious Disease Transmission

Carol Anne Nichols

Thesis submitted to the faculty of the Virginia Polytechnic Institute and State University

in partial fulfillment of the requirements for the degree of

Master of Science

Fisheries and Wildlife

Kathleen Alexander

Emmanuel Frimpong

Jeff Walters

April 27, 2018

Blacksburg, VA

Keywords: Olfactory communication, disease transmission, life history, den resources,

tuberculosis, urbanization

The Influence of Heterogeneous Landscapes on Banded Mongoose (Mungos mungo)

Behavior in Northern Botswana: Inferences about Infectious Disease Transmission

Carol Anne Nichols

ABSTRACT

Infectious disease transmission is driven by a complex suite of drivers with behavior and landscape dynamics contributing to epidemics across host-pathogen systems. However, our understanding of the interaction between landscape, behavior, and infectious disease remains limited. In the banded mongoose (Mungos mungo), a novel tuberculosis pathogen, Mycobacterium mungi, has emerged in Northern Botswana that is transmitted through olfactory communication behaviors. Using this host-pathogen system, this thesis explores the influence of various land use areas along the human-wildlife interface on animal behavior, and ultimately, pathogen transmission potential. Using behavior data from remote sensing camera traps, a generalized linear mixed model identified vigilance behavior, land use, and their interaction as important factors in predicting olfactory behavior. Cluster and Classification and Regression Tree (CART) analysis of active den sites (n= 308, across 23 troops) identified the important characteristics of dens across land use areas. In human-modified environments, man-made den sites persisted longer than did natural dens which became unsuitable through environmental processes (e.g., collapse). We also document the occurrence of nighttime activity for this species, perceived to be strictly diurnal. These data provide information critical to the development of robust computational models and underscore the importance of both landscape and behavior in accurately predicting and managing infectious disease outbreaks.

The Influence of Heterogeneous Landscapes on Banded Mongoose (Mungos mungo)

Behavior in Northern Botswana: Inferences about Infectious Disease Transmission

Carol Anne Nichols

ABSTRACT (Public)

The overall objective of this thesis was to identify the manner in which changing landscapes may influence disease transmission. Differences in host behavior across the landscape can influence pathogen transmission. Banded mongooses (Mungos mungo) are infected with a tuberculosis pathogen (Mycobacterium mungi) that is transmitted through olfactory secretions used for communication. We studied how olfactory communication is influenced by land type (i.e., lodge, national park, residential, undeveloped, and urban) and vigilance behavior. The vigilance behaviors that allow mongooses to avoid predators and competitors had a positive influence on olfaction behavior in some land use areas, but a negative influence on others. We also examined the characteristics of den sites used by banded mongooses in relation to different land use areas. Den sites in human-mediated land use areas were often man-made, remained in use for much longer than natural den sites, and were used by multiple troops. These findings suggest that landscape can be a driving force influencing behavior, and subsequently disease transmission. Understanding the influence of anthropogenically modified landscapes on disease transmission will be key to managing wildlife species.

Acknowledgments

I would like to express my appreciation to Dr. Kathleen Alexander who continuously provided encouragement and knowledge throughout this learning process. Without her expertise, this work would not have been possible. I would also like to thank Dr. Jeff

Walters and Dr. Emmanuel Frimpong for always being available for questions and providing insightful comments on earlier drafts.

I thank Dr. Mark Vandewalle for his involvement with the field work portion of this project and providing logistical support. I am grateful to Dr. Claire Sanderson for taking the time to provide valuable comments on this thesis.

I am thankful to Stephanie Vandewalle, Sean Hill, and Mike Heneghan for going out of their way to help me with trapping events and data collection. Previous mongoose researchers in this study site also provided helpful pointers. I am extremely grateful to

Emily Schlake, Madalyn Fox, Hannah Sylvester, Trish Deshmukh, and Kelton Verble who devoted a lot of time and effort to assist with the behavioral data.

Thanks are finally due to my parents, Mike and Kim Nichols, who always encourage me to follow my passion. They have provided support throughout every aspect of this work.

Thank you.

Table of Contents

Introduction: The role of wildlife behavior in infectious disease emergence ...... 1

Emerging Infectious Disease ...... 1

The Effects of Human-Modified Landscapes on Wildlife Behavior ...... 2

References ...... 6

Variation in olfactory behavior across the landscape: inferences about pathogen shedding in wild banded mongoose (Mungos mungo) populations ...... 8

Abstract ...... 8

Introduction ...... 9

Methods ...... 12

Study Site and Species ...... 12

Behavioral Data ...... 14

Observation Time ...... 19

Data Analysis ...... 19

Results ...... 22

Olfactory Communication Behavior ...... 23

Discussion ...... 25

Acknowledgments ...... 31

References ...... 31

Characteristics of banded mongoose (Mungos mungo) communal den sites across the human-wildlife interface in Northern Botswana ...... 35

Abstract ...... 35

Introduction ...... 36

Methods ...... 38

Study Site ...... 38

Banded Mongoose ...... 39

Den Occupancy Assessment ...... 44

Statistical Analysis ...... 44

Exploratory data analysis: Cluster and CART: ...... 44

Results ...... 45

Den site persistence ...... 47

Cluster and classification and regression trees ...... 47

Discussion ...... 49

Cluster and CART analysis of den sites ...... 52

Acknowledgments ...... 55

References ...... 56

Creeping in the night: what might ecologists be missing? ...... 60

Abstract ...... 61

Introduction ...... 62

Methods ...... 64

Field methods ...... 64

Photo classifications and moon phase ...... 67

Data analysis ...... 67

Results ...... 68

Discussion ...... 68

Acknowledgments ...... 71

References ...... 72

Conclusions ...... 80

Appendix ...... 82

vi Introduction: The role of wildlife behavior in infectious disease emergence

Emerging Infectious Disease

Over the last 40 years, the emergence of zoonotic diseases has increased (Jones et al.

2008), signaling the escalating importance of understanding infectious disease dynamics at the human-wildlife-environmental interface. Disease emergence events are often coupled with ecological perturbations that affect the environment, host, and pathogen, changing the manner in which pathogen communities connect with susceptible hosts

(Schrag and Wiener 1995, Plowright et al. 2008). Seemingly minor changes in landscape features or host exposures can have significant influences on disease emergence, as for example, the role of agricultural intensification in the emergence of the deadly bat-borne zoonosis, Nipah Virus (Capelle and Morand 2017). Computational models of infectious disease dynamics often neglect elements of these interactions (Alexander et al. 2012), which are likely fundamental to disease emergence processes in disturbed landscapes.

Landscape change can interact with and influence wildlife behavior, and therefore, infectious disease dynamics. For example, natural disturbances (e.g. fragmentation, extreme weather events) can modify community structure, contact networks, and pathogen transmission dynamics (Poteet 2006, Watson et al. 2007). Increases in host density around limited or non-seasonal human-associated resources can increase contact and transmission rates and create opportunities for cross-species transmission (Becker et al. 2015). Changes in stress and pollution along the urban-rural gradient can also affect

host susceptibility to infectious disease and the outcome of pathogen exposure events

(Bradley and Altizer 2007).

The Effects of Human-Modified Landscapes on Wildlife Behavior

The potential for human-modified ecosystems to influence wildlife host behavior and disease transmission dynamics is especially complex as changes in behavior can influence both host exposure and susceptibility (Hawley and Altizer 2011). Studies have principally focused on two overarching questions in this area of research: 1) how does social behavior lead to variability in pathogen susceptibility and exposure between individuals, and 2) how might behavior increase the potential for an individual to act as a super-spreader or, more generally, increase transmission heterogeneities among individuals? Here, super-spreader refers to individuals that contribute disproportionately to pathogen transmission (Lloyd-Smith et al. 2005, Woolhouse and Gowtage-Sequeria

2006). The research focus here has been directed at understanding heterogeneity at the individual level (Paull et al. 2012). However, there is a limited understanding of the role that heterogeneous landscapes play in shaping host behavior, and consequently disease dynamics, representing an area of research ripe for investigation.

To explore these dynamics and test related hypotheses, I use a long-term study population of banded mongooses (Mungos mungo) in Northern Botswana. This population is infected with an emerging Mycobacterium tuberculosis complex pathogen,

Mycobacterium mungi (Alexander et al. 2016). After the onset of clinical symptoms, diseased mongooses usually die within 2-10 months (Alexander et al. 2016). Mortality associated with M. mungi ranges from 10 to 15% per year in Chobe District, Northern

Botswana (Fairbanks 2013b). Unique to tuberculosis pathogens, this strain is transmitted through scent-marking behaviors used in olfactory communication (See Figure 1 from

(Alexander et al. 2016)). When an infected individual deposits a signal (i.e. a scent mark), the uninfected individual that investigates the signal is exposed to the pathogen, and can then become infected. This is the first report of a tuberculosis pathogen being transmitted through olfactory communication behaviors (Alexander et al. 2016).

Figure 1: Banded mongooses become infected with Mycobacterium mungi through secretions used in olfactory communication behavior. (A) Banded mongooses get injuries from daily interactions with conspecifics and/or the environment. (B) When engaging in olfactory communication behaviors, contact with infected anal gland secretions allows the pathogen to invade the mongoose host through these injuries. (C) Pathogen invasion results in disease (Alexander et al. 2016).

M. mungi occupies lineage 6 of the M. tuberculosis Complex (MTC), which is comprised of the West African human-associated pathogens (M. Africanum subtypes) and those infecting African mammals (See Figure 2 from (Alexander et al. 2016)). The only known

outbreaks of M. mungi in banded mongooses have occurred in Northern Botswana and northwestern Zimbabwe. Other wildlife species infected by wildlife-associated lineage six pathogens include the rock hyrax (Procavia capensis; dassie bacillus) and meerkat

(Suricata suricatta; M. suricattae), but there is no range overlap between infected populations. Though the outbreaks in banded mongooses, rock hyraxes, and meerkats are in Southern Africa, the only known outbreaks of M. africanum and chimpanzee bacillus are in West Africa.

Figure 2: Phylogenetic relationships among Mycobacterium tuberculosis Complex

(MTC) pathogens and reservoir species. Range overlap for reservoir hosts of lineage 6

MTC and known outbreaks of wildlife-associated lineage six MTC are provided

(Alexander et al. 2016).

Infected banded mongoose populations in Chobe District, Northern Botswana inhabit both the urban areas of Kasane and Kazungula and the rural area of Chobe National Park.

In urban areas, banded mongooses utilize resources from humans. Previous work

identified increased levels of aggression between banded mongooses when foraging in garbage compared to foraging naturally. The increase in aggressive behavior led to more injuries, increasing pathogen exposure and transmission of M. mungi (Fairbanks 2013b).

In addition to the increased availability of food resources in the study site, the results from past work show that the study population also uses anthropogenic resources for denning when they are available (Laver 2013a). Further, the banded mongoose troops that live in close association with humans, and therefore utilize anthropogenic resources

(i.e. food, water, and den shelter), have smaller overall home ranges, overall core ranges, and overall space use dispersion compared to troops that inhabit more remote areas without human influences (Laver and Alexander 2017). These results show that anthropogenic resources in human-modified areas are important to mongooses and should be examined to better understand the conditions that influence host behaviors across the landscape. This host-pathogen system provides the opportunity to explore how social behavior and landscape features may interact to modify transmission probabilities across complex landscapes and is the focus of the following three chapters in this thesis

References

Alexander, K. A., P. N. Laver, A. L. Michel, M. Williams, P. D. van Helden, R. M. Warren, and N. C. G. van Pittius. 2010. Novel Mycobacterium tuberculosis complex pathogen, M. mungi. Emerging infectious diseases 16:1296. Alexander, K. A., B. L. Lewis, M. Marathe, S. Eubank, and J. K. Blackburn. 2012. Modeling of wildlife-associated zoonoses: applications and caveats. Vector-Borne and Zoonotic Diseases 12:1005-1018. Alexander, K. A., C. E. Sanderson, M. H. Larsen, S. Robbe-Austerman, M. C. Williams, and M. V. Palmer. 2016. Emerging Tuberculosis Pathogen Hijacks Social Communication Behavior in the Group-Living Banded Mongoose (Mungos mungo). mBio 7:e00281-00216. Becker, D. J., D. G. Streicker, and S. Altizer. 2015. Linking anthropogenic resources to wildlife–pathogen dynamics: a review and meta‐analysis. Ecology letters 18:483-495.

Bradley, C. A., and S. Altizer. 2007. Urbanization and the ecology of wildlife diseases. Trends in Ecology & Evolution 22:95-102. Capelle, J., and S. Morand. 2017. Wildlife and emerging diseases in Southeast Asia. Biodiversity Conservation in Southeast Asia: Challenges in a Changing Environment. Fairbanks, B. M. 2013. Bidirectional interactions between behavior and disease in banded mongooses (Mungos mungo) infected with Mycobacterium mungi. Hawley, D. M., and S. M. Altizer. 2011. Disease ecology meets ecological immunology: understanding the links between organismal immunity and infection dynamics in natural populations. Functional Ecology 25:48-60. Jones, K. E., N. G. Patel, M. A. Levy, A. Storeygard, D. Balk, J. L. Gittleman, and P. Daszak. 2008. Global trends in emerging infectious diseases. Nature 451:990-993. Lloyd-Smith, J. O., S. J. Schreiber, P. E. Kopp, and W. M. Getz. 2005. Superspreading and the effect of individual variation on disease emergence. Nature 438:355-359. Paull, S. H., S. Song, K. M. McClure, L. C. Sackett, A. M. Kilpatrick, and P. T. Johnson. 2012. From superspreaders to disease hotspots: linking transmission across hosts and space. Frontiers in Ecology and the Environment 10:75-82. Plowright, R. K., S. H. Sokolow, M. E. Gorman, P. Daszak, and J. E. Foley. 2008. Causal inference in disease ecology: investigating ecological drivers of disease emergence. Frontiers in Ecology and the Environment 6:420-429. Poteet, M. F. 2006. Shifting roles of abiotic and biotic regulation of a multi-host parasite following disturbance. Disease ecology. Oxford University Press, New York, New York, USA:135-153. Schrag, S. J., and P. Wiener. 1995. Emerging infectious disease: what are the relative roles of ecology and evolution? Trends in Ecology & Evolution 10:319-324. Watson, J. T., M. Gayer, and M. A. Connolly. 2007. Epidemics after natural disasters. Emerging infectious diseases 13:1. Woolhouse, M. E., and S. Gowtage-Sequeria. Host range and emerging and reemerging pathogens. 2006.

7 Variation in olfactory behavior across the landscape: inferences about pathogen shedding in wild banded mongoose (Mungos mungo) populations

Abstract

Infectious disease transmission is driven by a complex suite of drivers with behavior and landscape dynamics contributing importantly to epidemic patterns across host-pathogen systems. However, our understanding of the interaction between landscape, behavior, and infectious diseases remains limited. In the banded mongoose (Mungos mungo), a novel tuberculosis pathogen, Mycobacterium mungi, has emerged in Northern Botswana and is transmitted through olfactory communication behaviors. Using this host-pathogen system, we assess the influence of landscape use and mongoose behaviors on the frequency of olfactory communication observations. We use remote sensing camera traps at den sites to eliminate observer influence across human-modified and urban landscapes

(n=18 troops, 18,229 detections of banded mongooses from 7,497 photographs). We identified a significant influence of landscape, vigilance behavior, and the interaction between the two on the frequency of olfactory communication behaviors that are associated with disease transmission. Vigilance across the landscape is multifaceted, associated with territorial defense and predator avoidance and can be a stimulant for olfactory communication in some land use areas and an inhibitor in others. While vigilance in lodge areas increased olfactory behaviors, in natural and urban landscapes with putatively higher predator risk, vigilance had a negative impact. Our results

underscore the complexity of the interaction between landscape and behavior and the importance this may have on pathogen transmission dynamics.

Introduction

Human-mediated landscape change is increasingly linked to zoonotic disease emergence

(Jones et al. 2008). Outbreaks of disease have occurred, however, across a wide array of landscape types and host-pathogen systems, challenging our ability to develop a full mechanistic understanding of the process (Brearley et al. 2012). Urbanization can have complex impacts on the environment and wildlife hosts with changes in stress and pollution exposures occurring along the urban-rural gradient, affecting host susceptibility to infectious diseases (Bradley and Altizer 2007). High concentrations of resources in modified habitats can influence pathogen exposure through increased contact due to host aggregation or indirectly through changes in host behavior (Becker and Hall 2014,

Becker et al. 2015, Flint et al. 2016). For example, banded mongoose troops in Northern

Botswana were more aggressive when foraging in garbage than when they were in other habitats. The elevation in aggressive behavior led to more injuries, increasing pathogen invasion potential for the emerging pathogen Mycobacterium mungi (Flint et al. 2016).

The foraging plasticity of foxes (Vulpes vulpes L.) in urban areas has also been linked to lower infection levels of Echinococcus multilocularis, as foxes primarily consume waste instead of rodents, the intermediate host for this parasite. Understanding comparative behavior dynamics across landscape type is central to determining how these divergent areas may influence host ecology.

The changes brought about by human population expansion (i.e. urbanization) can alter the biology of hosts, pathogens, and vectors in human-modified landscapes (Bradley and

Altizer 2007, Gottdenker et al. 2014). The rate of urbanization is accelerating (Ramalho and Hobbs 2012, DeSA 2013) with urban ecosystems becoming increasingly common

(Magle et al. 2012). The process of urbanization is one of extreme ecological change influencing ecosystem processes as well as wildlife community structure. Wildlife behavior is identified as one of the first things to change in human-modified environments (Plotkin 1988, Lowry et al. 2013) and behavior is increasingly recognized as a critical variable influencing infectious disease dynamics (Dizney and Dearing 2013).

While these urbanized areas were once considered to be suitable for only a few species, we are now observing use by a wide array of wildlife species (Ditchkoff et al. 2006), with consequent implications to zoonotic disease transmission in these landscapes. These novel, anthropogenically-influenced landscapes tend to be subject to sustained processes of ecological change (i.e. ongoing urbanization and development) that can modify the fitness of wildlife species through complex and interdependent mechanisms (Brearley et al. 2012, Russo and Ancillotto 2015). Understanding these potential effects of urbanization on wildlife populations has been identified as central to conservation planning (Angeloni et al. 2008) and increasingly public and animal health (McKinney

2002, Wu et al. 2017).

Communication is often seen as a key behavior central to individual and group fitness in many social species, which can be influenced strongly by environment (Hutton and

McGraw 2016). Animals show an enormous diversity in whether, when, where, and how

they communicate, ranging from auditory signals, like the roar of a lion to broadcast ownership of a territory and stay in contact with other members in the social group

(Grinnell and McComb 2001), to visual signals, like the brightly colored plumage of male cardinals to help females assess the health of potential mates (Bradburry and

Vehrencamp 2011). A common mode of communication is through olfaction, where individuals deposit scents with chemical cues in the environment to communicate with conspecifics in various social contexts (Arakawa et al. 2008). Olfactory communication is common among mammalian species and plays a crucial role in spatial organization and sexual and other social behaviors (Stoddart 1976, Brown and Macdonald 1985).

Information such as the depositor’s identity (Boogert et al. 2006), reproductive status

(Washabaugh and Snowdon 1998), competitive abilities (Rich and Hurst 1999), and even health status (Kavaliers and Colwell 1995, Penn et al. 1998, Klein et al. 1999, Willis and

Poulin 2000, Zala et al. 2004) can be communicated to conspecifics. Numerous mammalian species have successfully inhabited urban ecosystems, but limited attention has been directed at understanding how transformed landscapes may influence this key behavior (Hutton and McGraw 2016). We know even less about the manner in which changes in communication may change infectious disease transmission.

In Northern Botswana, an emerging Mycobacterium tuberculosis pathogen, M. mungi infects the highly social banded mongoose (Mungos mungo). Pathogen transmission occurs through olfactory communication (anal gland scent marks and urine (Alexander et al. 2016). In this species, olfactory behaviors are utilized in both inter- and intra-group communication (Jordan et al. 2010) and play a crucial role in territorial defense (Rood

1975, Müller and Manser 2007) and reproduction (Jordan et al. 2010, Jordan et al. 2011a,

Jordan et al. 2011b, Jordan et al. 2011c). We use this host-pathogen system to investigate the influence of landscape change on olfactory communication and disease transmission.

Methods

Study Site and Species We studied 15 troops ranging from eight to 50 individuals. The study population in

Northern Botswana occupies the town of Kasane and urban-transiting Kazungula and

Chobe National Park. Banded mongoose troops occur across land use type with territories that can span different land use designations (See study site map in Chapter 3).

Banded mongooses are small (<2kg), diurnal mesocarnivores that live in social groups

(Rood 1975). Troops are territorial and occupy home ranges that spatially decrease with increasing association with humans (Laver and Alexander 2017). Home ranges of adjacent troops regularly overlap in these higher density environments. Home ranges are defended against intrusions and boundaries can be demarcated by feces, urine, and other scent marks (Rood 1975, Müller and Manser 2007). Intertroop conflicts do occur and can result in mortality, but can also lead to intergroup group breeding events (Jordan et al.

2010). Mongooses are communal breeders with a low reproductive skew with females of the same troop going into estrous simultaneously (Cant 2000). Lactating females allosuckle pups and pack members cooperate in raising the young at communal dens with some adults staying behind to guard pups while others forage (Cant 2003). Once the pups are old enough to accompany adults on foraging bouts, a pack member escorts them

(Cant 2003). In our study site, pups emerge from dens throughout the wet season

(November-March), but first matings of the breeding season occur in August (Laver and

Alexander 2017).

There are several documented predators to banded mongooses such as birds of prey, particularly martial eagles (Polemaetus bellicosus) (Rood 1983), marabou storks

(Leptoptilos crumenifer) (Otali and Gilchrist 2004), African lions (Panthera leo) (Rood

1975), African rock pythons (Python sebae)(Rood 1975, Otali and Gilchrist 2004, Laver et al. 2012), warthogs (Phacochoerus africanus), and monitor lizards (Varanus exanthematicus) (Otali and Gilchrist 2004). Banded mongoose group members cooperatively gather to inspect predator cues (Müller and Manser 2007, Furrer and

Manser 2009) and bunch together to attack or mob predators and competitors (Rood

1975). They also emit alarm calls and pack members rush to assist individuals calling in distress (Furrer and Manser 2009).

In the human-dominated landscape, mortality is more commonly associated with vehicles, human conflict, and predation by domestic animals than in other areas (Laver and Alexander 2017). Human responses to banded mongooses can be quite divergent, from feeding mongooses to direct persecution when banded mongooses are perceived as pests (e.g. hot oil being poured into active den sites, stoning). Banded mongoose will mob domestic dogs that threaten the group, often resulting in fatal injuries for the banded mongoose (11%, n=20/188 deaths recorded from 2008-2016)

The study population is infected with a novel strain of tuberculosis, Mycobacterium mungi, and diseased individuals usually die within three months (Alexander et al. 2010).

Sick individuals are identifiable through observation of clinical symptoms of infection including anorexia, hunched body posture, matted fur, excessive watering of the eye, sneezing, nasal enlargement and/or deviation of the nasal planum accompanied with a build-up of mucus, enlarged testicles, and lethargy. Sick individuals are not evicted from the troop, but they often lag behind during group movements (Alexander et al. 2016).

Behavioral Data

Radio tracking and camera traps

Study banded mongooses were followed with the use of radio telemetry. Within each troop, one or two banded mongooses were trapped and fitted with radio collars as previously described (Alexander et al. 2010). The 15 collared troops were tracked five days a week, alternating seven troops on the first day and the other eight troops the following day.

A challenging problem in the study of wildlife behavior is obtaining observational data that is not influenced by the observer. Banded mongoose troops range across land type and respond strongly to the observation of humans in certain land types (i.e. urban, residential, national park) or in places they do not expect to see people (e.g., behind certain buildings or in hedgerows). Fleeing responses are common making behavioral studies challenging. In order to overcome these limitations, we deployed camera traps at den sites. This allowed us to track mongooses and place the camera trap before they came

out of the den, allowing us to capture data across troops and land use in a study location that could be common to all study troops across land type.

We placed cameras at den sites of 15 banded mongoose study troops (January 2016 to

March 2017). As banded mongooses change dens every two to three days, remote sensing cameras were placed at the identified den sites of the troops tracked that day and moved when the troop vacated the site. If the den was not located during the day, we used homing telemetry to track towards the troop and locate mongoose at night while the troop was in the den. If the den was within a gated area, or an area considered dangerous to enter at night, we returned the next morning to attempt to locate the den and place the camera before banded mongooses emerged from the den. In the event the animals were out before the camera could be placed, the camera would be deployed later in the day and data collection commencing the next uninterrupted den emergence event.

Cameras were placed using naturally occurring objects (e.g. trees, shrubs, fences), as well as self-made mounts to obtain a favorable angle and distance between the camera and the area of interest. Den sites were classified by land area as follows: 1) lodge areas, including landscaped areas that were watered daily, guest rooms, kitchens, and staff housing, 2) residential areas, including private homes and local village residential plots,

3) urban areas, such as the hospital and shopping areas, 4) undeveloped areas not within the protected area of the national park or forest reserves, but still susceptible to some human or vehicle traffic, and 5) the national park, which was the protected area of Chobe

National Park.

Animal research activities were conducted through the use of an approved protocol from the Virginia Tech Institutional Animal Care and Use Committee (IACUC #16-217) and under permit from the Botswana Ministry of Environment, Natural Resources

Conservation and Tourism (EWT 8/36/4 XXVI).

Behavioral observations from photographs

For each photograph the date and time, the number of individuals within each photograph, and their behavioral state (i.e., olfactory, marking, non-vigilant, and vigilant) were tallied (See Table 1 for ethogram). Both the marking and olfactory behaviors contribute to olfactory communication. The marking behaviors included: (1) anal scent marks which are deposited by dragging secretions from the anal gland across a surface;

(2) cheek marks or wipes; (3) urination or defecation; (4) urination followed by stomping the back limbs coined token dance urination; and (4) feces that are deposited at specific marking sites called latrines (Müller and Manser 2007, Jordan et al. 2010, Fairbanks et al.

2014). The olfaction behaviors crucial for investigating chemical cues included: (1) smelling the ground or an object; (2) smelling the location where a scent was deposited; and (3) smelling another mongoose. All other behaviors were grouped into vigilant and non-vigilant behaviors (Table 1).

Table 1. Ethogram for banded mongooses (Mungos mungo). The behaviors were grouped into four categories: (1) non-vigilant, (2) olfactory behaviors, (3) marking behaviors, and

(4) vigilant behaviors.

Category Behavior Description Non- Forage Digging for food item, manipulating food item, chewing vigilant Running Trotting or running

Walking Walking

Play (alone) Manipulating non-food object, usually interspersed with bouncing and play calls

Contact garbage or Contact with garbage or human food, primarily while human food foraging in it

Emerge Emerging from den or other structure

Excavate Digging at den site. Differs from digging for food in that both front paws are usually used synchronously for excavating, the depth is usually greater than for foraging, and no food items are eaten.

Social play Rough and tumble contact or chasing interspersed with bouncing and play call

Aggression/defense Agonistic contact with at least one other mongoose. Distinguished from play by aggressive sounds, lack of bouncing, and more forceful motions that may cause injury (e.g. forceful scratching and biting rather than pawing and mouthing)

Copulate Mating

Allofeed Feeding another mongoose, typically a pup

Jerky greeting A greeting where two animals jerk their heads back and forth while approaching each other, culminating in bodily contact with continued jerking of the head

Clasp One mongoose clasping another with front paws, similar to a mating position, but without intromission. Often performed during play.

Beg Performed by pups and sub adults. Following and sometimes touching another mongoose while emitting begging calls.

Lead An individual incites the group to move from the current location with contact calls that are louder and faster than usual while walking or trotting away from the current location.

Follow Walking or trotting in a direction that the leading animal is moving. Usually accompanied by mimicking the leading call.

Lying down Lying down with very little movement, whether asleep or awake.

Sit or stand still Sitting or standing with little to no movement. Head is below horizontal

Groom Scratching, gnawing, or licking oneself. (Counted as resting behavior because generally interspersed with lying down or sitting during periods when the troop is resting)

Allogroom Licking or gnawing another mongoose. Typically performed during a troop resting periods.

Social rest Lying while in physical contact with at least one other mongoose

Social sit or stand Sitting or standing while in physical contact at least one other mongoose

Olfactory Smell Smelling the ground or an object Behaviors Scent smell Smelling an area that is known to be scent marked Social smell Smelling another mongoose Marking Defecate/urinate Excretion Behaviors Token dance Urination involving the stereotypical stomping of ground urination with hind limbs

Latrine Defecation Feces are deposited in token amounts at specific sites called latrines

Social scent mark Rubbing another mongoose with anal glands

Scent mark Rubbing ground or object with anal glands

Over-marking One individual places its scent mark directly over the scent mark of another individual

Cheek mark Rubbing the side of the ‘face’ on surfaces and along the ground

Vigilance Vigilance Head is horizontal or above horizontal but not oriented to Behaviors another mongoose.

Head bobbing may occur. Gaze may be direct toward a perceived threat. All four legs are on the ground.

Bipedal Vigilance Head is horizontal or above horizontal but not oriented to another mongoose Head bobbing may occur. Gaze may be direct toward a perceived threat. Standing on back two legs.

Scanning Head is horizontal or above horizontal and scans the landscape.

Flee Trotting or running from a perceived threat

Alarm call Any of several calls used when a perceived threat is present. Often accompanied by vigilance or fleeing.

Vigilance while Exhibits vigilance behaviors while chewing chewing

Observation Time The length of camera observation periods varied by trap event and den sites. Each camera trap event was defined as the start time in which the first photo was taken to the end time of the series. If there were no photos for 10 minutes, the next photo would be considered a new event (Flint et al. 2016). For the instances that a single photograph was collected, it was given the lowest time interval designation of 10 seconds, since that is the lapse time for the remote sensing camera traps to capture another photograph.

Data Analysis Olfactory communication behavior was analyzed using generalized linear mixed models in the glmmADMB package in RStudio Version 1.1.383. The top model was selected using the Akaike’s Information Criterion adjusted for small sample sizes (AICc)

(Burnham and Anderson 2003). The den sites were used as the random effects (1|Den).

We used a negative binomial distribution with a log link. The fixed effects were set as land use, the presence of pups, season (wet and dry), and vigilance (Table 2). When significance of the top model was indicated (α ≤ 0.05), the Tukey-Kramer adjustment in the Multcomp package was used to compare least-square mean estimates to determine differences.

The number of mongooses, observation minutes, and photographs obtained during an observation event varied between and within den sites. To standardize data and account for the unequal observations, the offset was calculated by the equation

where p is the number of photos, t is the observation time in minutes for that observational period, and b is the number of banded mongooses observed across all photos in that observational period.

Table 2: Variables used in candidate models describing olfactory communication behavior with justification and supporting literature. Candidate models represent banded mongoose olfactory communication behavior in Northern Botswana.

Results

From January 2016 to March 2017, we sampled 281 den nights from 18 troops leading to

1,815 observation periods (15 radio-collared troops, three opportunistically sampled troops). There were 18,229 detections of banded mongooses from 7,497 photographs.

Land use, vigilance, and an interaction between these variables explained variation in olfactory communication behavior across den sites and troops (Table 3). There were no competing models. Sick mongooses were captured in photos (n=98). However, across photos, we did not observe olfactory behavior in these sick individuals.

Table 3: Modeling rankings for banded mongoose olfactory communication behavior in

Northern Botswana, with k (number of parameters), Akaike’s information criteria (AIC) value, Akaike’s information criteria, corrected for small sample size (AICc), difference in

AICc value between best supported model and ith model (ΔAICc), wi (model weight), and ERi (evidence ratio).

Olfactory Communication Behavior Vigilant behavior was an informative parameter for olfactory communication behavior

(Table 4; t=5.66, p =<0.001) and had a positive correlation with olfactory behavior. The top model also had an interaction between vigilance and land use areas indicating that the relationship with vigilance and olfactory communication was dependent on land use area.

The coefficient for vigilance in the national park (-6.75) and urban (-0.60) land use areas had a negative association with olfactory communication, whereas vigilance at the lodges 23

(0.84) had a positive association (Vigilance: National Park t=-4.03, p =< 0.001 and

Vigilance: Urban t= -3.80, p =< 0.001). The interaction of vigilance in the residential areas also had a negative association, but it was not significantly different than zero.

Table 4: Parameter estimates of our best supported model for banded mongoose olfactory communication in Northern Botswana.

Olfactory communication varied by land use areas with the highest in the urban land use areas. The estimate for the undeveloped land area was significantly different than zero and negatively associated with olfactory communication (t= -2.88, p=0.004). The least square means of undeveloped land use area was significantly less than the lodges, the national park, and the urban land use areas, but not the residential land use area (Figure

Table 5).

Table 5: The mean differences, standard errors, p-values, and 95% confidence intervals for the differences in banded mongoose olfactory communication across the land use areas in Northern Botswana.

Discussion

Individual behavior can influence infectious disease transmission and understanding how this might be modified across a landscape is of fundamental importance, particularly when these behaviors themselves provide a transmission route for infectious disease. We identified the importance of the landscape and vigilance behavior in shaping the olfactory communication behaviors that may influence disease transmission of M. mungi. Olfactory communication was significantly lower in the undeveloped landscape than in the national park, lodges, and urban areas. However, in land use areas where olfactory communication was higher, vigilance was a stimulant in some land use areas (i.e., lodges) and an 25

inhibitor in others (i.e., national park, urban). Olfactory communication is a common form of communication for mammalian species and is involved in numerous mechanisms from territoriality to sexual reproduction (Brown and Macdonald 1985). The interaction between behaviors, here, vigilance, land use, and olfaction was bidirectional, providing insight into the complex interaction between land type and behavior.

Vigilance behaviors are fundamental to fitness across species, providing information on predation risk (Beauchamp 2015). Nothing has more unforgiving impacts to overall fitness than being killed by a predator (Lima and Dill 1990). Vigilance allows for early detection of predators and varies in accordance with type and abundance of predators

(Laundré et al. 2010). Where larger predators may invoke a fleeing behavior, others such as domestic dogs can elicit mobbing and aggressive behaviors (Lima and Dill 1990). In order to avoid predation, individuals must assess the predation risk in an environment and decide whether to invest in anti-predator behaviors or other fitness-promoting activities

(e.g. foraging, acquiring a mate) (Lima and Dill 1990). While most studies assess the trade-offs associated with vigilance and feeding behavior (Lima and Dill 1990), this trade-off likely extends to other behaviors such as olfaction communication.

In addition to the threat of predation, vigilance behaviors are also associated with territorial defense (Beauchamp 2015). Troops respond more aggressively to known neighboring troops than unknown troops, which is likely related to competition for resources as well as breeding opportunities (Müller and Manser 2007). Previous research in our study site shows that mongooses living in association with humans have smaller

overall home ranges, overall core ranges, and overall space use dispersion (Laver and

Alexander 2017). This decrease in space use is likely the result of non-seasonal resource provisioning and denning opportunities associated with human waste and buildings in transformed landscapes. The groups living in association with humans den in man-made structures, feed from refuge sites, and drink from anthropogenic water (i.e gray-water, sewage, and lawn sprinklers) (Laver and Alexander 2017). In addition to using man-made den sites, anthropogenic den sites such as drains and buildings remain intact much longer and are used by multiple troops, but not simultaneously (Chapter 3). Though olfactory communication behaviors were highest in the urban, lodge, and national park land use areas, the concentration of troops around human-mediated resources in the urban and lodge land use areas likely results in more encounters with neighboring troops, where between troop exposure to scent marks may occur since those behaviors are associated with territorial defense (Rood 1975, Müller and Manser 2007). During inter-pack encounters, while some individuals rush towards the interaction with the other group, others mark or counter-mark objects surrounding the event and intra-group copulations are sometimes observed (Cant et al. 2002, Müller and Manser 2007).

Similar to the urban land use areas, the lodges also have a constant human presence (i.e. full-time staff and tourists), but the Botswana tourism industry is based on wildlife viewing opportunities, so the humans in lodge environments rarely pose a threat to banded mongooses. Further, the lodges typically have fences to deter megafauna populations, but other animals (e.g. banded mongooses, warthogs, vervet monkeys

(Chlorocebus pygerythrus), Chobe bush buck (tragelaphus scriptus ornatus)) regularly

utilize these areas (See animal and human presence across landuse in Table 6). These fenced environments preserve riparian forests in the region providing mature, closed canopies with an open cultivated lawn understory that likely allows the detection of predators from afar and a delay in flight response. Flushing behavior is the distance between an observer and animal when it flees and is used to determine tolerance of predators (Petrelli et al. 2017). If the mongooses are able to see threats from afar, they might invest in olfactory communication behaviors. Though the banded mongooses in the urban areas still face the threat of competition, because of the infrastructure in this area there is a likely a shorter line of sight. Similarly, the national park land use areas has tall grasses, likely reducing the line of sight, and therefore, the ability to detect predators from afar. Considering that there was a negative association between vigilance and olfactory communication in the urban and national park land use areas, the mongooses here might initiate flight behaviors once a threat is detected.

Table 6: The mean animal and human presence per camera trap day across the land use areas in Northern Botswana. These values were derived from 74 remote sensing camera traps placed at active banded mongoose den sites. For more information about the species captured, see Supplementary Table 1 in Appendix.

Previous research in our study system identified reduced activity, decline in alert behaviors, and increased resting behaviors among clinically ill mongoose (Fairbanks et al. 2014). In this study, we did not observe any clinically ill mongooses engaging in olfactory behaviors. This may have important implications to pathogen transmission dynamics suggesting that a window of pathogen transmission may be bounded by the interaction between pathogen shedding and the clinical impacts of disease hindering the ability to undertake normal behaviors. Here, behavior may have a bidirectional affect.

Pathogen shedding is directed through olfactory behavior but sickness behavior might 29

lead to containment of contagion at the individual level with the reduction or cessation of pathogen shedding. Scent marking behaviors can be costly and may be reduced when energy investments are required elsewhere, such as immunity (Gosling et al. 2000). For example, male house mice (Mus musculus domesticus) reduced scent marking during experimental infections with Salmonella enterica C5TS (Zala et al. 2004). In Uganda, banded mongooses that had higher levels of infection with the protozoan parasites,

Isospora, and the nematode worm, Toxocara, scent marked at lower rates than did mongooses with lower infection levels. (Mitchell et al. 2017). Counter-marking of scent marks from the opposite sex among individuals with high parasite loads were also reduced (Mitchell et al. 2017). It is unclear if M. mungi has a latency period and how pathogen shedding occurs over the course of disease. Further research is needed to evaluate the influence of clinical disease on pathogen shedding among infected individuals and the influence this may have on epidemic behavior and pathogen spread.

Landscape-behavior interactions may influence the transmission of other infectious diseases, as for example, leptospirosis (Leptospira sp), which is transmitted through urine and infects 45% of sampled banded mongooses in our study site (Jobbins and Alexander

2015). While there is great interest in developing our understanding of the role of individual behavior in shaping disease transmission dynamics, our work demonstrates the fundamental influence that the landscape may have on these interactions. Complex landscapes may organically create heterogeneity in pathogen transmission dynamics between individuals across land type.

Acknowledgments

We are grateful to numerous CARACAL team members for their assistance in the field and reviewing photographs. We also thank Mark Vandewalle for logistical and field assistance. This project was supported by the National Science Foundation Ecology and

Evolution of Infectious Diseases (#1518663) and Morris Animal Foundation (#D14ZO-

083).

References

Jones, K.E., Patel, N.G., Levy, M.A., Storeygard, A., Balk, D., Gittleman, J.L. et al. 2008 Global trends in emerging infectious diseases. Nature, 451 (7181), 990-993. Brearley, G., McAlpine, C., Bell, S. and Bradley, A. 2012 Influence of urban edges on stress in an arboreal mammal: a case study of squirrel gliders in southeast Queensland, Australia. Landsc. Ecol., 27 (10), 1407-1419. Bradley, C.A. and Altizer, S. 2007 Urbanization and the ecology of wildlife diseases. Trends in ecology & evolution, 22 (2), 95-102. Flint, B.F., Hawley, D.M. and Alexander, K.A. 2016 Do not feed the wildlife: associations between garbage use, aggression, and disease in banded mongooses (Mungos mungo). Ecology and evolution, 6 (16), 5932-5939. Becker, D.J., Streicker, D.G. and Altizer, S. 2015 Linking anthropogenic resources to wildlife–pathogen dynamics: a review and meta‐analysis. Ecology letters, 18 (5), 483-495. Becker, D.J. and Hall, R.J. 2014 Too much of a good thing: resource provisioning alters infectious disease dynamics in wildlife. Biology letters, 10 (7), 20140309. Gottdenker, N.L., Streicker, D.G., Faust, C.L. and Carroll, C. 2014 Anthropogenic land use change and infectious diseases: a review of the evidence. Ecohealth, 11 (4), 619-632. Ramalho, C.E. and Hobbs, R.J. 2012 Time for a change: dynamic urban ecology. Trends in ecology & evolution, 27 (3), 179-188. DeSA, U. 2013 World population prospects: the 2012 revision. Population division of the department of economic and social affairs of the United Nations Secretariat, New York. Magle, S.B., Hunt, V.M., Vernon, M. and Crooks, K.R. 2012 Urban wildlife research: past, present, and future. Biological Conservation, 155, 23-32. Plotkin, H.C. 1988 The role of behavior in evolution. MIT press. Lowry, H., Lill, A. and Wong, B. 2013 Behavioural responses of wildlife to urban environments. Biological Reviews, 88 (3), 537-549. Dizney, L. and Dearing, M.D. 2013 The role of behavioural heterogeneity on infection patterns: implications for pathogen transmission. Animal Behaviour, 86 (5), 911- 916.

Ditchkoff, S.S., Saalfeld, S.T. and Gibson, C.J. 2006 Animal behavior in urban ecosystems: modifications due to human-induced stress. Urban ecosystems, 9 (1), 5-12. Russo, D. and Ancillotto, L. 2015 Sensitivity of bats to urbanization: a review. Mammalian Biology-Zeitschrift für Säugetierkunde, 80 (3), 205-212. Angeloni, L., Schlaepfer, M.A., Lawler, J.J. and Crooks, K.R. 2008 A reassessment of the interface between conservation and behaviour. Animal Behaviour, 75 (2), 731- 737. McKinney, M.L. 2002 Urbanization, Biodiversity, and Conservation The impacts of urbanization on native species are poorly studied, but educating a highly urbanized human population about these impacts can greatly improve species conservation in all ecosystems. BioScience, 52 (10), 883-890. Wu, T., Perrings, C., Kinzig, A., Collins, J.P., Minteer, B.A. and Daszak, P. 2017 Economic growth, urbanization, globalization, and the risks of emerging infectious diseases in China: a review. Ambio, 46 (1), 18-29. Hutton, P. and McGraw, K.J. 2016 Urban Impacts on oxidative balance and animal signals. Frontiers in Ecology and Evolution, 4, 54. Grinnell, J. and McComb, K. 2001 Roaring and social communication in African lions: the limitations imposed by listeners. Animal Behaviour, 62 (1), 93-98. Bradburry, J.W. and Vehrencamp, S.L. 2011 Signals and Communication. In Principles of Animal Communication, Sinauer Associates, Inc., Publishers, Sunderland, Massachusetts, pp. 1-16. Arakawa, H., Blanchard, D.C., Arakawa, K., Dunlap, C. and Blanchard, R.J. 2008 Scent marking behavior as an odorant communication in mice. Neuroscience & Biobehavioral Reviews, 32 (7), 1236-1248. Brown, R.E. and Macdonald, D.W. 1985 Social odours in mammals. Clarendon Press. Stoddart, D.M. 1976 Mammalian odours and pheromones. Edward Arnold London. Boogert, N.J., Hofstede, F.E. and Monge, I.A. 2006 The use of food source scent marks by the stingless bee Trigona corvina (Hymenoptera: Apidae): the importance of the depositor’s identity. Apidologie, 37 (3), 366-375. Washabaugh, K. and Snowdon, C.T. 1998 Chemical communication of reproductive status in female cotton‐top tamarins (Saguinus oedipusoedipus). American Journal of Primatology, 45 (4), 337-349. Rich, T.J. and Hurst, J.L. 1999 The competing countermarks hypothesis: reliable assessment of competitive ability by potential mates. Animal behaviour, 58 (5), 1027-1037. Zala, S.M., Potts, W.K. and Penn, D.J. 2004 Scent-marking displays provide honest signals of health and infection. Behavioral Ecology, 15 (2), 338-344. Kavaliers, M. and Colwell, D.D. Discrimination by female mice between the odours of parasitized and non-parasitized males. The Royal Society, pp. 31-35. Klein, S.L., Gamble, H.R. and Nelson, R.J. 1999 Trichinella spiralis infection in voles alters female odor preference but not partner preference. Behavioral Ecology and Sociobiology, 45 (5), 323-329. Penn, D., Schneider, G., White, K., Slev, P. and Potts, W. 1998 Influenza infection neutralizes the attractiveness of male odour to female mice (Mus musculus). Ethology, 104 (8), 685-694.

Willis, C. and Poulin, R. 2000 Preference of female rats for the odours of non-parasitised males: the smell of good genes? Folia Parasitologica, 47 (1), 6-10. Alexander, K.A., Sanderson, C.E., Larsen, M.H., Robbe-Austerman, S., Williams, M.C. and Palmer, M.V. 2016 Emerging Tuberculosis Pathogen Hijacks Social Communication Behavior in the Group-Living Banded Mongoose (Mungos mungo). mBio, 7 (3), e00281-00216. Jordan, N.R., Mwanguhya, F., Kyabulima, S., Ruedi, P. and Cant, M.A. 2010 Scent marking within and between groups of wild banded mongooses. Journal of Zoology, 280 (1), 72-83. Müller, C.A. and Manser, M.B. 2007 ‘Nasty neighbours’ rather than ‘dear enemies’ in a social carnivore. Proceedings of the Royal Society B: Biological Sciences, 274 (1612), 959-965. Rood, J.P. 1975 Population dynamics and food habits of the banded mongoose (Mungos mungo). African Journal of Ecology, 13 (2), 89-111. Jordan, N.R., Manser, M.B., Mwanguhya, F., Kyabulima, S., Rüedi, P. and Cant, M.A. 2011 Scent marking in wild banded mongooses: 1. Sex-specific scents and overmarking. Animal Behaviour, 81 (1), 31-42. Jordan, N.R., Mwanguhya, F., Furrer, R.D., Kyabulima, S., Rüedi, P. and Cant, M.A. 2011 Scent marking in wild banded mongooses: 2. Intrasexual overmarking and competition between males. Animal Behaviour, 81 (1), 43-50. Jordan, N.R., Mwanguhya, F., Kyabulima, S., Rüedi, P., Hodge, S.J. and Cant, M.A. 2011 Scent marking in wild banded mongooses: 3. Intrasexual overmarking in females. Animal Behaviour, 81 (1), 51-60. Laver, P.N. and Alexander, K.A. 2017 How living in association with humans, and seasonality interact to reverse traditional animal space use predictions. Movement Ecology. Cant, M.A. 2000 Social control of reproduction in banded mongooses. Animal Behaviour, 59 (1), 147-158. Cant, M.A. 2003 Patterns of helping effort in co-operatively breeding banded mongooses (Mungos mungo). Journal of Zoology, 259 (2), 115-121. Rood, J.P. 1983 Banded mongoose rescues pack member from eagle. Animal Behaviour, 31 (4), 1261-1262. Otali, E. and Gilchrist, J.S. 2004 The effects of refuse feeding on body condition, reproduction, and survival of banded mongooses. J. Mammal., 85 (3), 491-497. Laver, P.N., Ganswindt, A., Ganswindt, S.B. and Alexander, K.A. 2012 Non-invasive monitoring of glucocorticoid metabolites in banded mongooses (Mungos mungo) in response to physiological and biological challenges. General and Comparative Endocrinology, 179 (2), 178-183. Furrer, R.D. and Manser, M.B. 2009 Banded mongoose recruitment calls convey information about risk and not stimulus type. Animal Behaviour, 78 (1), 195-201. Alexander, K.A., Laver, P.N., Michel, A.L., Williams, M., van Helden, P.D., Warren, R.M. et al. 2010 Novel Mycobacterium tuberculosis complex pathogen, M. mungi. Emerging infectious diseases, 16 (8), 1296. Burnham, K.P. and Anderson, D.R. 2003 Model selection and multimodel inference: a practical information-theoretic approach. Springer Science & Business Media.

Jobbins, S.E. and Alexander, K.A. 2015 Evidence of Leptospira sp. infection among a diversity of African wildlife species: beyond the usual suspects. Transactions of The Royal Society of Tropical Medicine and Hygiene, 109 (5), 349-351. Beauchamp, G. 2015 Animal vigilance: monitoring competitors and predators. Academic Press, Oxford. Lima, S.L. and Dill, L.M. 1990 Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology, 68 (4), 619-640. Laundré, J.W., Hernández, L. and Ripple, W.J. 2010 The landscape of fear: ecological implications of being afraid. Open Ecology Journal, 3, 1-7. Cant, M.A., Otali, E. and Mwanguhya, F. 2002 Fighting and mating between groups in a cooperatively breeding mammal, the banded mongoose. Ethology, 108 (6), 541- 555. Fairbanks, B.M., Hawley, D.M. and Alexander, K.A. 2014 No evidence for avoidance of visibly diseased conspecifics in the highly social banded mongoose (Mungos mungo). Behavioral Ecology and Sociobiology, 1-11. Gosling, L.M., Roberts, S.C., Thornton, E. and Andrew, M. 2000 Life history costs of olfactory status signalling in mice. Behavioral Ecology and Sociobiology, 48 (4), 328-332. Mitchell, J., Cant, M.A., Vitikainen, E.I. and Nichols, H.J. 2017 Smelling fit: scent marking exposes parasitic infection status in the banded mongoose. Current Zoology, 63 (3), 237-247.

34 Characteristics of banded mongoose (Mungos mungo) communal den sites across the human-wildlife interface in Northern Botswana

Abstract

Denning behavior is a critical life history attribute for many mammalian species and can be impacted by urbanization, affecting species reproductive success and survival in these landscapes. Banded mongooses (Mungos mungo) in Northern Botswana are infected with a novel Mycobacterium tuberculosis complex pathogen, M. mungi that is transmitted through olfactory communication networks and infected scent marks that allow the pathogen to move through the population through both direct and environmental transmission routes. Here, we studied characteristics of active banded mongoose den sites across 24 troops and 308 den sites in Northern Botswana (348 den nights from 2008-2010 and 281 den nights from 2016-2017). Dens were located across the human-wildlife interface (national park, lodge, urban, residential, undeveloped). Cluster and

Classification and Regression Trees (CART) analysis of the den sites identified important differences in den site characteristics across the land area designation. Habitat type was the most important factor for den sites located at lodges, associated with the presence of gallery forest and teak woodlands. Distance to the nearest tar road appeared to be the most important variable separating the dens in the National Park from those outside of lodge environments in human transformed landscapes. Den type was the most predictive of den sites located in human-modified environments. Sites that had the longest use were dominantly anthropogenic in nature (94% ± 6%, n= 67, used in 2008-2010 and 2016-

2017). Natural dens, in contrast, appeared to be more vulnerable to destruction and

shorter-lived. Only anthropogenic dens were used by troops other than the resident troop

(6%, n=19/308). These space-sharing behaviors in a territorial species may have important influences on pathogen transmission dynamics allowing pathogen exposure and transmission potential to increase between territorial groups. Anthropogenic land areas appear to exert an important influence on denning behaviors in the banded mongoose, which is information beneficial for researchers and managers as numerous species readily adapt to novel opportunities provided by human-land transformation.

Introduction

The majority of mammalian species use some form of shelter on a daily basis (Reichman and Smith 1990). These can vary widely from simple bedding near coniferous trees during the night (e.g., white-tailed deer, Odocoileus virginianus (Armstrong et al. 1983)), to the more intricate construction of lodges by beavers (Castor canadensis) (Collen and

Gibson 2000). Burrowing mammals require access to in-ground shelter and the availability of these sites can have important influences on species reproductive success and survival (Reichman and Smith 1990). Dens act as a place to reside when these species are not engaging in other fitness-promoting activities (e.g. foraging, searching for mates) and also offer critical protection from harsh, fluctuating environments, and predation (Reichman and Smith 1990). Denning behavior is a critical element of a species natural history (Reichman and Smith 1990) and a central element of a species conservation management.

Den sites can be a limiting resource in a landscape and may be used by more than one social group, exerting critical influences on contact networks and intra-and inter-specific

pathogen transmission potential for vector, environmentally and directly transmitted pathogens. Particular landscape features then provide a connection through space sharing

(active or delayed) among infected and susceptible hosts, elements that may change pathogen transmission potential across a population and landscape. For example, den sharing was identified as a primary risk for Mycobacterium bovis infection in brushtail possums (Trichosurus vulpecula) (Corner et al. 2003). Subadult rabbits (Oryctolagus) move between warrens of different groups and therefore play a role in transmission of rabbit hemorrhagic disease and myxomatosis (Jennings and Mutze 2018).

Banded mongooses are small, burrowing species, which are group living and den communally (Rood 1975, Cant 2000). This population is infected with a novel

Mycobacterium tuberculosis complex pathogen, M. mungi that is transmitted through olfactory communication networks and infected scent marks that allow the pathogen to move through the population through both direct and environmental transmission routes

(Alexander et al. 2016). For banded mongooses, natural dens tend to consist of excavated holes in the ground and termite mounds, but troops may also make use of man-made den sites (Rood 1975, Laver and Alexander 2017). Den use is central to banded mongoose breeding success and survival across landscape type. Multiple females will produce litters at the same time and the young will stay in the natal den for about four to five weeks, with one or two adults staying to guard the young while the rest of the troop forages

(Rood 1975, Cant 2003). Pups emerge from the den and start to accompany the rest of the troop on foraging bouts at around four weeks of age, with both sexes helping to provision the pups for up to eight weeks after emergence (Cant 2000). After foraging bouts, troop

members often reconvene around the den site entrance to rest, groom, scent mark, and nurse the pups, while others engage in play behaviors (camera trap observations). Troops relocate den sites every two to three days (Hiscocks and Perrin 1991), but have been known to stay at the same den for longer periods (54 days in Rwenzori National Park,

Uganda (Rood 1975) and over two months in Queens Elizabeth National Park, Uganda

(Neal 1970)). Banded mongooses will often reuse den sites within their home range, with our study troops in Northern Botswana averaging 30 unique dens sites (Laver and

Alexander 2017). Once a troop leaves one of their known den sites, they return to the same den after a median of 106 days.

Understanding the impacts of anthropogenic change on banded mongoose ecology is central to their management and can provide insight into the influence of land transformation on wildlife behavior, space use, and disease transmission. We evaluate banded mongoose den site use and attributes across anthropogenic and natural landscapes and discuss implications to disease transmission in changing landscapes.

Methods

Study Site Our study population is located in Chobe District, Northern Botswana (17.828°S and

25.163°E; See Figure 1). The human population for this area was estimated to be 13,141 in 2011 (2011 Botswana Population Census (Botswana Government 2015)). The urban center of Kasane and the urban-transitioning village of Kazungula are experiencing significant growth and development related to a growing tourism market. The urban center of Kasane has a variety of hotels, lodges and camps, and an international airport.

Expansion and development has resulted in a loss of forest land, with around 3,060 hectares of the Kasane Forest Reserve degazetted for residential and developmental purposes (Lepetu et al. 2010).

Figure 1: The study site in Chobe District, Northern Botswana with active banded mongoose (Mungos mungo) den sites.

Banded Mongoose We obtained den site characteristics from study troops across land use type (n=24, 2008-

2010 and 2016-2017, see below). Due to group fusion, extinction events, and other influences, seven troops observed from 2008-2010 are not included in the 2016-2017 data capture period. Briefly, within each troop, one or two banded mongooses were trapped and fitted with radio collars, allowing troops to be located through the use of radio telemetry. All methods were carried out in compliance with Virginia Tech IACUC

procedures (IACUC 13-164-FIW) and under a permit from the Botswana Ministry of

Environment, Natural Resources Conservation and Tourism (EWT 8/36/4 XXVI). Radio collared troops were tracked five days a week, alternating seven troops on the first day and the other eight troops the following, resulting in each troop being observed two to three times a week (see Table 1). Once the den site was located, the GPS location, mongoose troop, and den site characteristics were recorded and a remote sensing camera was placed at the site (2016-2017 only).

Table 1: Study banded mongoose (Mungos mungo) troops used multiple land types within their home ranges in Chobe District, Northern Botswana.

*Indicates the troops that were tracked during the 2008-2010 data period only. 41

Den sites were characterized by landscape features and mongoose troop characteristics that were comprised of ten variables: land use, habitat, GIS land cover classification building density, den type, level of land transformation, seasonal use, mortality density, multiple troop use, and distance to road (Table 2). Land use was designated into five different categories related to the nature of human influence on landscape structure: urban, undeveloped/unprotected land areas, residential, lodge, and the national park.

Within each land class, the dominant habitat type and cover classes were identified as previously described (See Table 2 for classifications (Child 1968, Weare and Yalala

1971, Mosugelo et al. 2002, Wolf 2009, Alexander and Vandewalle 2014, Fox et al.

2017)). Digitized data on building locations (2008, n = 5,641) was used to create a point density surface using the density function in ArcGIS Spatial Analyst. The density gradient raster values were classified into four categories with one being the lowest and four the highest. This procedure calculates the density of point features around each output raster cell. A circular neighborhood was defined around each raster cell center and the number of points that fall within the neighborhood were totaled and divided by the area of the neighborhood. To generate the distance between the den sites and the nearest tar road, we used remote imagery for the area to digitize roads and used the proximity analysis tool to obtain a distance measurement. Den structure was characterized by the type as natural (termite mounds, bushes, hollow logs) or anthropogenic (slash piles, buildings, building materials, concrete paths, drains, natural holes in the ground, rocks and scrap heaps).

Table 2: The den site variables used to characterize banded mongoose den sites in

Northern Botswana.

Mongoose group size was excluded from the analysis as it varied considerably over the study period by troop and was often unavailable by den night. Dens sites were characterized by use (single or multiple troops) and by season (dry season, April-

October; wet season, November-March, or both). Mortality data were incorporated in a similar manner as the building data above with the density gradient raster value classified as low, medium, and high based on the standard deviation. This classification method places class breaks around the mean based on the standard deviation.

Den Occupancy Assessment To investigate the number of days a troop would occupy a den before relocating to a new den site, we limited our evaluation to the data obtained from camera traps (2016-2017) as this data set provided more robust information on nightly den use. Den occupancy data was derived from camera traps as previously described in Chapter 2. Briefly, once a den site was located, camera traps were placed in an optimal position to capture photographs of banded mongooses at the den site. In the event that a troop relocated in the middle of the night, if the departure was after midnight then that den night was still included in the total den nights for that denning event.

Statistical Analysis All statistical analyses were conducted in the open source integrated programming environment RStudio Version 1.1.383. Fisher’s exact tests were used to test the differences between anthropogenic and natural den structures and the period of their use across study periods.

Exploratory data analysis: Cluster and CART: Exploratory data analysis provides an opportunity to uncover underlying structure in a data set and identify the most influential variables. Cluster analysis is a type of exploratory data mining that attempts to group data in such a way that objects in the same group (i.e. a cluster) are more similar to each other than other groups (clusters). The package Cluster and specifically the Daisy algorithm were used for hierarchical cluster analysis and creation of cluster dendrograms of den site variables (Table 2). The Gower’s distance dissimilarity method was used as the linkage approach (Gower 1971). This linkage method is able to handle different data types in the analysis (nominal, ordinal, asymmetric binary, and interval scaled). We assessed several distance measures (i.e. 44

ward, single, average, complete, centroid, and mcquitty) and compared goodness of the dendrogram fit with the dissimilarity matrix using both the Gower Distance Assessment and Cophenitic Correlation (Borcard et al. 2011). The average distance measure provided the best fit and the dendrogram was pruned using a clustering heat map (Pacheco 2015).

Further, using the sil_width function, the entire clustering was displayed by combining the silhouettes (silhouette represents a cluster) into a single plot and the average silhouette width provided a means to select the appropriate number of clusters

(Rousseeuw 1987). We used the rafalib package in R and the myplclust function to plot the final dendrogram.

Recursive partitioning is a statistical approach for multivariable analysis used to create a decision tree that splits objects into similar groups based on dichotomous independent variables (Lemon et al. 2003). For this, we used classification and regression tree

(CART) analysis, a nonparametric decision tree method to evaluate associations between land use and banded mongoose den sites. We used the rpart Package and the “class” method to create the decision trees (Plaat et al.). Land use was set as the dependent variable and we selected the lowest xerror (estimate of the cross-validation prediction error) and the corresponding conditional probability (cp) in the rpart object to identify the optimal level for pruning the resulting tree. The function as.party in the partykit package in R were used to plot the final decision tree.

Results

Den site characteristics across the land use areas are provided in Table 3. The den types were dominantly anthropogenic in nature in the human-modified land use areas (i.e.

buildings and scrap heaps) and natural structures in the remote national park (i.e. termite mounds). The lodge land use areas provided pristine environments with intact gallery forests. The urban, residential, and national park land use areas were slightly degraded and the undeveloped areas were heavily degraded. Degraded riparian forests were dominant in the urban, park, and undeveloped areas. Scrub savanna was prominent in residential areas. Land cover for designated land areas is provided in Table 3. Mortality was high in all of the land use areas with the exception of the National Park, however, surveillance bias likely influenced this observation. Buildings were the highest in the urban and residential areas and lowest in the National Park.

Table 3: Den site characteristics for study banded mongoose (Mungos mungo) troops in

Northern Botswana.

We also found that 6% (n=308) of mongoose den sites were used by multiple troops. All of these dens were anthropogenic in nature with these dens more commonly found in the urban and lodge land use designations. Communal dens were used on average for one to three days and this pattern remained across the land use areas, between the wet and dry seasons, and when pups were at the dens (see den occupancy, Table 2). When pups are 46

still too young to emerge from the den, adult mongooses carry the pups to the next den site (camera trap observations).

Den site persistence A total of 308 den sites were identified in the study site, of which 139 were in use during

2008-2010 (Laver 2013b), and 67 den sites were used repeatedly throughout both time periods. Of those dens still in use by the 2016-2017 study period, the greater proportion were anthropogenic in nature (94%, n= 67, p < 0.0001 by Fisher's exact test).

Cluster and classification and regression trees The average distance metric provided the best fit to the dissimilarity matrix on both the

Gower Distance Assessment and Cophenetic Correlation. The cluster analysis of den site characteristics identified four primary clusters (Figure 2). Den sites clustered according to land type with the exception of the residential den sites, which grouped with the urban dens.

Figure 2: Dendrogram of cluster analysis of den site data collected for study population with clusters noted in color. The analysis identified four clusters: (1) national park, (2) urban, (3) undeveloped, and (4) lodge areas. The residential dens group with the urban dens.

CART analysis included all 308 den sites and identified three splits in the classification tree (cp = 0.02) and was pruned accordingly (Figure 3). The classification tree indicated that habitat, distance from tarred roads, and the type of den were the key variables predicting banded mongoose den site by land use designation (Figure 3).

Figure 3: Decision tree from the CART analysis of cluster assignments.

Discussion

In our study, banded mongooses appear to be exceptionally plastic in their use of communal den sites across land type, adapting to opportunities to maximize fitness needs according to the land area and its characteristics (Table 3). Anthropogenic den structures were commonly observed in human-modified landscapes (e.g. buildings, scrap heaps, cement paths). In more natural settings, troops often utilized holes in the ground, termite mounds, or snags, but these structures tended to degrade and disappear over the study period.

A subset of den sites was regularly used by multiple troops, all of which were anthropogenic in nature. These dens sites were dominantly located in the urban and lodge land use areas. Banded mongooses are territorial and are known to defend their home ranges vigorously, where intergroup encounters can result in mortalities (Cant et al. 2002,

Otali and Gilchrist 2004). Thus, these findings are particularly interesting and suggest that anthropogenic landscapes may relax or reduce territorial defense or drive troops to risk aggressive encounters in order to enter and stay within another troops territory to access associated resource opportunities. Den sharing can have important implications for pathogen transmission in this species allowing infected/susceptible individuals to have indirect contact with environments that have potentially high levels of pathogen related to marking behavior. The frequency of marking behaviors varies across the landscape but is higher in the urban and lodge settings (Chapter 2). We do not currently know how long

M. mungi might survive in the environment, but M. bovis in environmental samples was detected four to 21 months after possible contamination (Young et al. 2005). The anal gland secretions fundamental in the transmission of M. mungi are highly lipid in nature and depending on the den site, can be protected from desiccation for a considerable period of time (Alexander et al. 2016). Deposition within the den would provide further protection from pathogen degrading environmental forces (wind, rain, sunlight). Previous studies in this system found that banded mongoose space use was strongly associated with the degree of overlap with humans (Laver and Alexander 2017); as association with humans increased, home range, core range, and space use dispersion decreased. Resource acquisition (i.e. food, water) from the subsidized resources around the human settlements might motivate different troops to utilize the same den sites in the urban and lodge areas, amplifying disease transmission potential.

Across land type, den sites were consistently used for one to three nights before mongoose relocated to another den. Relocating dens sites may decrease infestations by

ectoparasites, but it is unlikely that build up occurs within one to three days (Boydston et al. 2006). Other social, diurnal Herpestidae species are known to relocate dens along a similar time frame; dwarf mongooses (Helogale parvula) predominantly use a den for a single night, moving to another den the following day (Hiscocks 1999). Meerkats

(Suricata suricatta) change burrows every two to three days, except when young are at the dens (Clutton-Brock and Manser 2016). This is a short relocation time compared to other larger communally denning species and may be related to forage limitations for these smaller fossorial species. For example, clans of spotted hyenas (Crocuta crocuta) can use a communal den for over a month (Périquet et al. 2016). However, aardwolves

(Proteles cristatus), an insectivorous Hyaenidae, relocates dens several times within a month (Kotze et al. 2012).

Urban landscapes present novel predation risks and we recorded mortalities in our study troops from a variety of human-mediated sources including vehicles, human conflict, and predation by domestic animals. Direct eradication efforts from humans have been observed when banded mongooses are perceived as pests (e.g., hot oil being poured into active den sites, razor-wire placed in the entrance to den sites, shooting with a weapon, and stoning). Research in Uganda also reported humans throwing rocks and adult mongooses being killed by human inflicted burns (Otali and Gilchrist 2004). Domestic dogs in our system commonly pursued and engage in aggressive interactions with study banded mongoose troops, often resulting in mongoose fatalities (11%, n=188 deaths recorded from 2008-2016).

Cluster and CART analysis of den sites Cluster analysis identified land use type (national park, lodge, urban, and the undeveloped areas) as the primary variable grouping communal den sites. This was not true for den sites in residential areas, however, these sites appeared to share features common across human modified land categories (Figure 3). CART analysis identified three primary nodes: habitat, distance to the nearest tar road, and den type as the most important predictors of den sites by land type (Figure 3).

The presence of gallery forests/riparian woodlands was most predictive of dens at lodge facilities as compared to dens sites located in other land use types. Loss of riparian ecosystems has been an in issue of increasing concern across Africa and the globe with pressures that include invasive species (Holmes et al 2005), wood extraction, fire, and increasingly in Southern Africa, complex interactions between elephant damage, lack of recruitment, and declining water tables (Skarpe et al. 2004, Alexander and Vandewalle

2014, Nichols et al. 2017). Banded mongooses are not known to have an ecological requirement for gallery forests and this association may simply be a characteristic of lodge sites rather than a habitat feature important to banded mongoose fitness. These riparian woodlands provide critical habitat to other species such as the Chobe bushbuck

(Tragelaphus scriptus ornatus) that has declined since the 1970s but populations still persist in preserved habitat found at lodge sites (Addy 2016). While gallery forests are under threat across the globe (Palmer et al. 2000), certain human transformed landscapes may provide greater protection for these plant communities allowing them to persist where they might otherwise be lost.

The distance from the tar road was the variable most predictive of den sites located within the national park where infrastructural development is limited and only one tar road traverses the region. Wildlife viewing from safari vehicles is common, but tracks are not paved. The mongooses in this system den along the floodplains, which further increases their distance from tar roads. Subsequently, the wildlife within this land use area experiences the least exposure to human influences.

For undeveloped and urban land use areas, den site type (e.g. bush, hole in the ground, building material) appeared to be the most important variable differentiating these land areas. This tendency likely follows den site opportunities where natural den sites would be more frequently found in the undeveloped land type and anthropogenic opportunities provided in the urban landscape (See Figure 4 anthropogenic and natural den site example).

Figure 4: Photograph of an anthropogenic and a natural den site utilized by banded mongooses (Mungos mungo) in Northern Botswana. The photo on the left (A) is a drain used by a study troop in the urban land use area. The photo on the right (B) is a termite mound used by a study troop in an undeveloped area.

As human populations increase worldwide, the resulting habitat loss and degradation often force animals to live in close proximity to humans (Woodroffe 2000). Banded mongooses appear to be able to access suitable den sites across landscape type and display behavioral flexibility in den site use, successfully taking advantage of anthropogenically derived opportunities where these arise (Herr et al. 2010). The ability to adapt to urban landscapes is variable among vertebrate species and predictive of urban establishment. Some species, “urban avoiders” (Blair 2001), cannot adapt and persist only in more natural settings. These species deliberately avoid areas near humans and select shelter or den sites far from human influences (Theuerkauf et al. 2003, Elfström et al. 2008). In contrast, “urban adapters” (Blair 2001) are able to live in close human proximity exploiting non-seasonal food resources (e.g. waste, cultivated plants) and other

anthropogenic landscape attributes (McKinney 2002). Urban adapters display considerable behavioral plasticity, exploiting man-made structures for denning (Hwang et al. 2007, Herr et al. 2010), many in close proximity to humans (Hwang et al. 2007,

Baldwin and Bender 2008). Banded mongooses should be considered urban adapters, as they are able to take advantage of novel urban denning opportunities.

Though there are numerous species that thrive in urban environments, the wildlife in these areas face novel selection pressures (Ditchkoff et al. 2006). The banded mongooses in human-mediated areas have access to non-seasonal resources, but drawbacks such as increased pathogen transmission associated with injury while foraging and invasion of the novel pathogen M. mungi is presented in these environments (Flint et al. 2016). Here, sharing of dens sites may also increase pathogen transmission in these same landscapes.

As urbanization pressures continue, it is clear that land use can have an important influence on denning behavior in those species that can adapt to urban environments. The use of these landscapes may have important implications for disease transmission and pathogen persistence, and potentially, human and animal health.

Acknowledgments

We are grateful to Dr. Mark Vandewalle for logistical assistance and Dr. Claire

Sanderson for comments on earlier drafts. We would like to thank the dedicated

CARACAL team members and volunteers that contributed to data collection.

This project was supported by the National Science Foundation Ecology and Evolution of

Infectious Diseases (#1518663) and Morris Animal Foundation (#D14ZO- 083).

References

Addy, J. E. 2016. Impact of elephant induced vegetation change on the status of the Bushbuck (tragelaphus scriptus ornatus) along the chobe river in Northen Botswana. Alexander, K., and M. Vandewalle. 2014. Guns, Ivory, and Disease: Past Influences on the Present Status of Botswana's Elephants and their Habitats. in Elephants and savanna woodland ecosystems: a study from Chobe National Park, Botswana. John Wiley & Sons. Alexander, K. A., C. E. Sanderson, M. H. Larsen, S. Robbe-Austerman, M. C. Williams, and M. V. Palmer. 2016. Emerging Tuberculosis Pathogen Hijacks Social Communication Behavior in the Group-Living Banded Mongoose (Mungos mungo). mBio 7:e00281-00216. Armstrong, E., D. Euler, and G. Racey. 1983. Winter bed-site selection by white-tailed deer in central Ontario. The Journal of Wildlife Management 47:880-884. Baldwin, R. A., and L. C. Bender. 2008. Den-site characteristics of black bears in Rocky Mountain National Park, Colorado. Journal of Wildlife Management 72:1717- 1724. Blair, R. B. 2001. Birds and butterflies along urban gradients in two ecoregions of the United States: is urbanization creating a homogeneous fauna? Pages 33-56 in Biotic homogenization. Springer. Borcard, D., F. Gillet, and P. Legendre. 2011. Numerical ecology with R. Springer Science & Business Media. Boydston, E. E., K. M. Kapheim, and K. E. Holekamp. 2006. Patterns of den occupation by the spotted hyaena (Crocuta crocuta). African Journal of Ecology 44:77-86. Cant, M. A. 2000. Social control of reproduction in banded mongooses. Animal behaviour 59:147-158. Cant, M. A. 2003. Patterns of helping effort in co-operatively breeding banded mongooses (Mungos mungo). Journal of Zoology 259:115-121. Cant, M. A., E. Otali, and F. Mwanguhya. 2002. Fighting and mating between groups in a cooperatively breeding mammal, the banded mongoose. Ethology 108:541-555. Child, G. F. 1968. An ecological survey of northeastern Botswana. Food and Agriculture Organization of the United Nations. Clutton-Brock, T., and M. Manser. 2016. Meerkats: cooperative breeding in the Kalahari. Cooperative breeding in vertebrates (eds WD Koenig, JL Dickinson):294-317. Collen, P., and R. Gibson. 2000. The general ecology of beavers (Castor spp.), as related to their influence on stream ecosystems and riparian habitats, and the subsequent effects on fish–a review. Reviews in fish biology and fisheries 10:439-461. Corner, L., D. Pfeiffer, and R. Morris. 2003. Social-network analysis of Mycobacterium bovis transmission among captive brushtail possums (Trichosurus vulpecula). Preventive veterinary medicine 59:147-167.

Ditchkoff, S. S., S. T. Saalfeld, and C. J. Gibson. 2006. Animal behavior in urban ecosystems: modifications due to human-induced stress. Urban Ecosystems 9:5- 12. Elfström, M., J. E. Swenson, and J. P. Ball. 2008. Selection of denning habitats by Scandinavian brown bears Ursus arctos. Wildlife Biology 14:176-187. Flint, B. F., D. M. Hawley, and K. A. Alexander. 2016. Do not feed the wildlife: associations between garbage use, aggression, and disease in banded mongooses (Mungos mungo). Ecology and evolution 6:5932-5939. Fox, J. T., M. E. Vandewalle, and K. A. Alexander. 2017. Land Cover Change in Northern Botswana: The Influence of Climate, Fire, and Elephants on Semi-Arid Savanna Woodlands. Land 6:73. Gower, J. C. 1971. A general coefficient of similarity and some of its properties. Biometrics:857-871. Herr, J., L. Schley, E. Engel, and T. J. Roper. 2010. Den preferences and denning behaviour in urban stone martens (Martes foina). Mammalian Biology-Zeitschrift für Säugetierkunde 75:138-145. Hiscocks, K. 1999. The impact of an increasing elephant population on the woody vegetation in southern Sabi Sand Wildtuin, South Africa. Koedoe 42:47-55. Hiscocks, K., and M. Perrin. 1991. Den selection and use by dwarf mongooses and banded mongooses in South Africa. S. AFR. J. WILDL. RES./S.-AFR. TYDSKR. NATUURNAVORS. 21:119-122. Hwang, Y. T., S. Lariviere, and F. Messier. 2007. Local-and landscape-level den selection of striped skunks on the Canadian prairies. Canadian Journal of Zoology 85:33-39. Jennings, S., and G. Mutze. 2018. Multiple warren use by subadult wild rabbits, Oryctolagus cuniculus, and its implications for disease transmission. Wildlife Research 44:582-586. Kotze, R., N. C. Bennett, E. Z. Cameron, J. L. de Vries, D. G. Marneweck, C. W. Pirk, and F. Dalerum. 2012. Temporal patterns of den use suggest polygamous mating patterns in an obligate monogamous mammal. Animal behaviour 84:1573-1578. Laver, P. N. 2013. The foraging ecology of banded mongooses (Mungos mungo): epidemiological and human-wildlife conflict implications. Virginia Polytechnic Institute and State University, Blacksburg, VA. Laver, P. N., and K. A. Alexander. 2017. How living in association with humans, and seasonality interact to reverse traditional animal space use predictions. Movement Ecology. Laver, P. N., A. Ganswindt, S. B. Ganswindt, and K. A. Alexander. 2012. Non-invasive monitoring of glucocorticoid metabolites in banded mongooses (Mungos mungo) in response to physiological and biological challenges. General and Comparative Endocrinology 179:178-183. Lemon, S. C., J. Roy, M. A. Clark, P. D. Friedmann, and W. Rakowski. 2003. Classification and regression tree analysis in public health: methodological review and comparison with logistic regression. Annals of behavioral medicine 26:172- 181.

Lepetu, J., J. Alavalapati, and P. K. Nair. 2010. Forest dependency and its implication for protected areas management: A case study from Kasane Forest Reserve, Botswana. McKinney, M. L. 2002. Urbanization, Biodiversity, and Conservation The impacts of urbanization on native species are poorly studied, but educating a highly urbanized human population about these impacts can greatly improve species conservation in all ecosystems. BioScience 52:883-890. Mosugelo, D. K., S. R. Moe, S. Ringrose, and C. Nellemann. 2002. Vegetation changes during a 36‐year period in northern Chobe National Park, Botswana. African Journal of Ecology 40:232-240. Neal, E. 1970. The banded mongoose, Mungos mungo Gmelin. African Journal of Ecology 8:63-71. Nichols, C. A., M. E. Vandewalle, and K. A. Alexander. 2017. Emerging threats to dryland forest resources: elephants and fire are only part of the story. Forestry: An International Journal of Forest Research 90:473-484. Otali, E., and J. S. Gilchrist. 2004. The effects of refuse feeding on body condition, reproduction, and survival of banded mongooses. Journal of Mammalogy 85:491- 497. Pacheco, E. R. 2015. Unsupervised Learning with R. Packt Publishing Ltd. Palmer, M. A., A. P. Covich, S. Lake, P. Biro, J. J. Brooks, J. Cole, C. Dahm, J. Gibert, W. Goedkoop, and K. Martens. 2000. Linkages between Aquatic Sediment Biota and Life Above Sediments as Potential Drivers of Biodiversity and Ecological Processes: A disruption or intensification of the direct and indirect chemical, physical, or biological interactions between aquatic sediment biota and biota living above the sediments may accelerate biodiversity loss and contribute to the degradation of aquatic and riparian habitats. AIBS Bulletin 50:1062-1075. Périquet, S., C. Mapendere, E. Revilla, J. Banda, D. Macdonald, A. Loveridge, and H. Fritz. 2016. A potential role for interference competition with lions in den selection and attendance by spotted hyaenas. Mammalian Biology-Zeitschrift für Säugetierkunde 81:227-234. Plaat, S. A., E. Deckers, G. Degenaars, I. Martorelli, and J. van Kollenburg. Exploratory Data Analysis on Multivariate Data. Reichman, O., and S. C. Smith. 1990. Burrows and burrowing behavior by mammals. Current mammalogy 2:197-244. Rood, J. P. 1975. Population dynamics and food habits of the banded mongoose (Mungos mungo). African Journal of Ecology 13:89-111. Rousseeuw, P. J. 1987. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics 20:53-65. Skarpe, C., P. A. Aarrestad, H. P. Andreassen, S. S. Dhillion, T. Dimakatso, J. T. du Toit, D. J. Halley, H. Hytteborn, S. Makhabu, and M. Mari. 2004. The return of the giants: ecological effects of an increasing elephant population. AMBIO: A Journal of the Human Environment 33:276-282. Theuerkauf, J., S. Rouys, and W. Jedrzejewski. 2003. Selection of den, rendezvous, and resting sites by wolves in the Bialowieza Forest, Poland. Canadian Journal of Zoology 81:163-167.

Weare, P., and A. Yalala. 1971. Provisional vegetation map of Botswana. Botswana Notes and Records 3:1-1. Wolf, A. 2009. Preliminary assessment of the effect of high elephant density on ecosystem components (grass, trees, and large mammals) on the chobe riverfront in Northern Botswana. University of Florida. Woodroffe, R. Predators and people: using human densities to interpret declines of large carnivores. Cambridge University Press, 2000. Yang, H., A. Viña, Y. Tang, J. Zhang, F. Wang, Z. Zhao, and J. Liu. 2017. Range-wide evaluation of wildlife habitat change: A demonstration using Giant Pandas. Biological Conservation 213:203-209. Young, J. S., E. Gormley, and E. M. Wellington. 2005. Molecular detection of Mycobacterium bovis and Mycobacterium bovis BCG (Pasteur) in soil. Applied and environmental microbiology 71:1946-195

59 Creeping in the night: what might ecologists be missing?

C.A. Nichols12¶ and K.A. Alexander12¶*

1 Department of Fish and Wildlife Conservation, Virginia Tech, Blacksburg, VA, United

States of America

2 Centre for Conservation of African Resources, Animals, Communities, and Land use

(CARACAL), Kasane, Botswana

*Corresponding Author email: [email protected] (KA)

¶ These authors contributed equally to this work.

This manuscript was formatted for and submitted to PLOS ONE

Abstract

Wildlife activity patterns tend to be defined by terms such as diurnal and nocturnal that might not fully depict the complexity of a species’ life history strategy and behavior in a given system. These activity pattern categories often influence the methodological approaches employed, including the temporal period of study (daylight or nighttime). We evaluated banded mongoose (Mungos mungo) behavior in Northern Botswana through the use of remote sensing cameras at active den sites in order to characterize early morning fear behavior for this diurnal species. Our approach, however, provided the facility to capture unexpected nocturnal activity in a species that had otherwise only been studied during daylight hours. Camera traps were deployed 215 trap days (24 hour data capture period) at den sites, capturing 5,472 photos over all events. Nocturnal activity was identified in 3% of trap days at study den sites. There was no association between the occurrence of nocturnal activity and light availability from lunar phases (Fisher’s exact test, n=215, p = 0.638) and thus, increased moonlight was not identified as a factor influencing nocturnal space use. Nearly a third of nocturnal events were characterized as fear behaviors (fleeing the den, relocation events, n=2/7). Drivers of nocturnal activity in mongoose are uncertain but captured events have important implications for fear behavior and space use during this activity period. Our findings highlight the need for ecological studies to more explicitly address and evaluate the potential for temporal variability in activity periods. Modifying our approach and embracing variation in wildlife activity patterns might provide new insights into the interaction between ecological phenomenon and species biology that spans the diurnal – nocturnal spectrum.

Introduction

The categorization of wildlife activity tends to be constrained to traditional day-night niches, strictly classifying species as diurnal or nocturnal (Tattersall 1987, Curtis and

Rasmussen 2002, Hill 2006). Less frequently, the term crepuscular is used, since many diurnal species are active at dawn and dusk as well as in full daylight (Ashby 1972).

Identification of more distinct adaptations to either a diurnal or nocturnal life history strategy (e.g. larger orbits in nocturnal primate species improving night vision (Kay and

Kirk 2000)) may contribute to the tendency to strictly view species’ activity patterns as a binary trait, rather than extending across a spectrum of temporal activity types (Tattersall

1987, Bearder et al. 2006). The characterization of a species to an activity pattern may translate into a methodological approach where evaluation extending beyond the dominant activity period is never attempted.

This more simplistic approach, however, was questioned in the 1970s with the characterization of the activity patterns of the presumed diurnal Mayotte lemur (Lemur fluvus mayottensis). This species was observed to apportion it’s behavior equally between day and night periods (Tattersall 1979). As a consequence, the term cathemerality was developed to capture this temporal activity type, adding to the previously established, and mutually exclusive categories of nocturnal and diurnal. As a behavioral strategy, cathemerality is thought to allow considerable advantages, identifying the ability for the species to engage in diurnal and/or nocturnal activity periods in response to varying ecological conditions (Hill 2006). Factors such as temperature, access to food resources, and predation risk are all thought to be important in promoting cathemerality (Hill 2006).

At least eight species of Madagascar lemur exhibit cathemeral activity patterns (Schaik and Kappeler 1996, Wright 1999, Curtis and Rasmussen 2002). This new term was initially met with skepticism and even rejected by a reviewer in the peer review process as it was considered to be “unnecessary new jargon” (Tattersall 1987). This initial rejection illustrates the commitment to the notion that animal activity cycles can be viewed as if they are a binary trait, rather than extending across a spectrum of temporal activity types.

For the animals that restrict activity to day or night periods, the rare occurrences of behavior ongoing in the other phases are considered unusual circumstances. Since these events are considered unusual, the additional time and effort required to gain an understanding of these occurrences are often disregarded. Regardless, it is evident that expansion of activity into the night occurs in a number of species that are classified as diurnal (e.g. over 70% of nocturnally vocalizing avian species were diurnal (La 2012), nine West Indian diurnal lizards and one species of diurnal snake extending their activity into the night around lights (Henderson and Powell 2001)). A prominent hypothesis for diurnal species expanding activity into the night is that nighttime activity occurs due to elevated light levels. Singing at night was typical of unmated northern mockingbirds

(Mimus polyglottos) and regularly occurred during full moon periods and less frequently during other lunar phases (Derrickson 1988). For carnivores, more recent studies reveal that nocturnal activity may be more pronounced than previously thought for traditionally perceived diurnal species such as the African wild dog (Lycaon pictus (Cozzi et al. 2012,

Rasmussen and Macdonald 2012)) and cheetahs (Acinonyx jubatus (Cozzi et al. 2012)),

with nocturnal activity for both species comprising roughly 25% of their overall activity budget (Cozzi et al. 2012). African wild dogs and cheetahs diurnal activities are thought to arise as a temporal avoidance of stronger competitors, the lions (Panthera leo) and hyenas (Crocuta crocuta), which are active at night (Cozzi et al. 2012). However, nocturnal activity of both wild dogs and cheetahs fluctuated with the lunar cycle, constrained to the brighter nights, suggesting that brighter moon phases might offset or reduce potential costs of encountering stronger competitors and predators (Cozzi et al.

2012).

We used camera traps to systematically study fear behavior among banded mongooses

(Mungos mungo), a diurnal mesocarnivore (Rood 1975), at den sites in Northern

Botswana. Systematic studies of banded mongoose ecology and behavior had only been undertaken during daylight hours (e.g. (Rood 1975, Cant et al. 2002, Jordan et al. 2010,

Jordan et al. 2011a, Jordan et al. 2011b, Jordan et al. 2011c, Fairbanks 2013a)). Here, we report nocturnal activity for this presumed exclusively diurnal species. We aim to test whether the detection of nocturnal activity patterns can be attributed to brighter lunar phases and discuss our findings in the light of our current approach to activity pattern classification and methodological approaches for ecological studies of wildlife.

Methods

Field methods Remote sensing camera traps were deployed from January 4 to November 9, 2016, in order to study banded mongoose behavior at the den sites of 17 troops located in our long-term study site in Northern Botswana (Table 1). The study area covered the urban

areas of Kasane and Kazungula (defined as town e.g., tourist lodges, residential areas, camp kitchens, farm land, and garbage sites) as well as natural habitats (defined as park e.g., forest reserves and Chobe National Park). Animals were tracked through the deployment of radio collars as previously described (Laver 2013b). Each radio collared troop (n=15) was tracked five days a week alternating eight one day and seven the next.

Once the den site was located, camera traps were mounted on nearby fence posts, trees, poles, and occasionally man-made stake mounts to obtain the optimal position for photographing movement and behavior at the den site. While tracking the marked troops, we also opportunistically placed cameras at den sites (n=2) that were identified but the troops had not been fitted with a radio collar. A trap day was defined as a 24-hour monitoring period of banded mongoose activity at the den site.

Table 1. The banded mongoose (M. mungo) troops sampled from January 4 - November

9, 2016. The individual study troops are listed with the general land class (i.e. park or town) that they inhabited throughout the study period. The occurrence of nighttime activity is documented for the troops that engaged in nighttime behavior along with the month that the event took place.

The study was conducted under approval and in accordance with the guidelines of the

Virginia Tech Institutional Care and Use Committee (16-217-FIW) and under permit of the Botswana Government.

Photo classifications and moon phase The trap events were categorized as “day” or “night” according to the time stamp on the photographs. The hours used to classify trap events were based on astronomical twilight.

Considering Botswana’s tropical location, sunrise and sunset do not drastically change throughout the year. The days are slightly shorter in the winter months from May to

August. During the winter, the “day” length was 0500-2000, and accordingly, the “night” photos include the times between 2000 and 0500. During the summer, the “day” photos were lengthened to the hours between 0400 and 2100, and the “night” photos include the times between 2100 and 0400. A trap event that contained photographs during the nighttime hours is categorized as a “night” trap event. The trap events were then aligned to the concomitant lunar phase classified into four categories (i.e. full, waning, waxing, and new).

Data analysis

The proportion of nighttime activity was calculated as the number of trap days where nighttime activity was identified over the total number of trap days. Due to the rare occurrence of nocturnal activity, our approach is limited to evaluating nights with activity regardless of some nights incorporating multiple observations from different troops.

Fisher’s exact test was used in RStudio (Version 0.99.484) to evaluate associations between the trap nights with nighttime activity and variables of interest. Moonlight was

available during the waxing, waning, and full moon phases, so they were grouped as the bright nights. The new moon phases were considered the dark nights.

Results

From January to November 2016, nighttime activity was identified for 3% of the trap days (SD= 18%, n=7/215) and across five of the 11 months of the study. The nocturnal activity occurred between 2000 and 2400 hours. The nighttime activity was observed in

29% of the troops (n =5/17 study troops, Table 1). There was no association between land classification of the den site (town or park) and the occurrence of night activity (Fisher’s exact test p=1). However, the limited number of park troops might have influenced our ability to detect an effect.

Banded mongoose nighttime activity was seen across all moon phases with no significant difference between the nights with no light (i.e. the new moon) and the phases (i.e. wane, wax full) that did provide moonlight (Fisher’s exact test p=0.638).

Discussion

Like other social mongooses, banded mongooses are considered diurnal species (Veron et al. 2004), and across studies ecological and behavioral investigations have been restricted to daylight periods (Rood 1975, Cant et al. 2002, Jordan et al. 2010, Jordan et al. 2011a,

Jordan et al. 2011b, Jordan et al. 2011c, Fairbanks 2013a). However, this limited temporal approach may fail to capture data critical to understanding the ecology and temporal nature of space use for a species. Here, we identify important nocturnal behavior and movement in banded mongooses. Unlike a number of other species, nighttime activity was seen across both the brighter and darker moon phases and thus, did

not appear to be influenced by sufficient light. Our evaluation of fear behavior at den sites identified both vigilant and non-vigilant behavioral states (See Chapter 2 for

Ethogram) across both diurnal and nocturnal activity periods with nocturnal vigilance behavior restricted to fleeing and den relocation. The occurrence of vigilance behavior followed by fleeing and den relocation was unique to nocturnal observations. Lack of association with moon phase suggests other important ecological factors or stimuli are encountered at night and influence nocturnal space use, and denning behavior and site fidelity.

Nighttime predation forces diurnal species to engage in anti-predation behaviors. For example, snake predations on avian nests results in females leaving the nest during the night (Reidy et al. 2009). In addition to avoiding predation, there are other fitness advantages to diurnal species engaging in nighttime movements. Although yellow breasted chats (Icterina virens) are diurnal, both males and females are active and foray outside of their territories at night. Fertile females foraying at night to engage in extrapair copulations potentially allowing them to go undetected by their social mate (Ward et al.

2014). Brown bears (Ursus arctos) are primarily diurnal but forage equally between daylight and darkness during salmon migrations with capture success increasing during the night (Klinka and Reimchen 2002).

Photos from the cameras did not provide information as to what caused nocturnal vigilant behavior. Fleeing behavior had previously only been reported in this species in association with intergroup aggression events (Thompson et al. 2017). When collecting

the cameras for one of the nighttime evacuation events, the residents of the house that the den site was located under reported that they had been woken up in the middle of the night by a disturbance involving the mongoose. Another researcher studying the same population found that a study troop at a lodge relocated dens during the night and a staff member reported that lions entered the property that night. These vigilant responses may hinder our ability to understand the influence of moonlight on nocturnal behavior for this normally diurnal species, but further research may provide better insight.

For the nighttime photographs that banded mongooses were not responding to a threat, there are one or two individuals engaging in activity instead of the entire troop. These active individuals tend to be walking towards an active den site, and in one situation, an individual appears to be investigating the den site. There is no evidence of emerging from the den, so it is unclear if these individuals belong to the troop residing in the den. It is possible that these individuals could be from a neighboring troop or dispersing individuals. The advantages of these nocturnal movements remains unclear.

Clearly, nocturnal and diurnal data are necessary to fully understand den selection and fidelity across a landscape. Restricting data collection for diurnal species to expected daytime activity periods might hinder or limit our ability to gain further insight of the biology and ecology of wildlife species. Limiting observations may also result in erroneous conclusions when restricting data collection to a particular activity time. For example, non-invasive fecal glucocorticoids are used to evaluate stress responses in wildlife (Keay et al. 2006) and results are often interpreted in conjunction with behavioral

data collected during the same diurnal time period of collection or the diurnal period the day before the collection (Wielebnowski and Watters 2007). Here, stress levels are interpreted and aligned with observed daytime movements and behaviors when, in fact, the stress response may reflect the occurrence of unobserved nocturnal events. Failure to account for variation in activity periods might not only undermine study objectives, but also the veracity of our understanding of a species’ ecology and behavior.

Considering that heterogeneity is a fundamental underlying phenomenon of ecology

(Scheiner and Willig 2011), temporal variation in activity patterns should be explicitly accounted for in methodological approaches. Our observations of banded mongooses illustrate the importance of including observations during low activity times when studying a wildlife species. Initially, ecology assumed spatial homogeneity for convenience and simplicity, with heterogeneity an unwelcome but necessary complication (Pickett and Cadenasso 1995). Study approaches and theoretical frameworks in the field of landscape ecology that have embraced heterogeneity and complexity have provided new and critical insights into key ecological processes (e.g. community dynamics or succession, the role of edges, patch dynamics)(Pickett and

Cadenasso 1995). A similar inclusion of heterogeneity into our consideration of activity periods and associated behaviors is called for and will likely lead to a new era of ecological discovery.

Acknowledgments

We thank the Government of Botswana’s Ministry of Environment Wildlife and Tourism, and the Department of Wildlife and National Parks for their support for this project, and

the residents and business owners in Kasane and Kazungula for permitting research on their properties. We are especially thankful to Dr. Mark Vandewalle for administrative and logistical support. We are grateful to CARACAL staff for their hard work locating active den sites and the volunteers at Virginia Tech and CARACAL that assisted with reviewing photographs. We thank Dr. Claire Sanderson and Kelton Verble for comments on earlier drafts.

References

Addy, J. E. 2016. Impact of elephant induced vegetation change on the status of the Bushbuck (tragelaphus scriptus ornatus) along the chobe river in Northen Botswana. Alexander, K., and M. Vandewalle. 2014. Guns, Ivory, and Disease: Past Influences on the Present Status of Botswana's Elephants and their Habitats. in Elephants and savanna woodland ecosystems: a study from Chobe National Park, Botswana. John Wiley & Sons. Alexander, K. A., P. N. Laver, A. L. Michel, M. Williams, P. D. van Helden, R. M. Warren, and N. C. G. van Pittius. 2010. Novel Mycobacterium tuberculosis complex pathogen, M. mungi. Emerging infectious diseases 16:1296. Alexander, K. A., B. L. Lewis, M. Marathe, S. Eubank, and J. K. Blackburn. 2012. Modeling of wildlife-associated zoonoses: applications and caveats. Vector-Borne and Zoonotic Diseases 12:1005-1018. Alexander, K. A., C. E. Sanderson, M. H. Larsen, S. Robbe-Austerman, M. C. Williams, and M. V. Palmer. 2016. Emerging Tuberculosis Pathogen Hijacks Social Communication Behavior in the Group-Living Banded Mongoose (Mungos mungo). mBio 7:e00281-00216. Angeloni, L., M. A. Schlaepfer, J. J. Lawler, and K. R. Crooks. 2008. A reassessment of the interface between conservation and behaviour. Animal behaviour 75:731-737. Arakawa, H., D. C. Blanchard, K. Arakawa, C. Dunlap, and R. J. Blanchard. 2008. Scent marking behavior as an odorant communication in mice. Neuroscience & Biobehavioral Reviews 32:1236-1248. Armstrong, E., D. Euler, and G. Racey. 1983. Winter bed-site selection by white-tailed deer in central Ontario. The Journal of Wildlife Management 47:880-884. Ashby, K. 1972. Patterns of daily activity in mammals. Mammal Review 1:171-185. Baldwin, R. A., and L. C. Bender. 2008. Den-site characteristics of black bears in Rocky Mountain National Park, Colorado. Journal of Wildlife Management 72:1717- 1724. Bearder, S., K. Nekaris, and D. Curtis. 2006. A re-evaluation of the role of vision in the activity and communication of nocturnal primates. Folia Primatologica 77:50-71. Beauchamp, G. 2015. Animal vigilance: monitoring competitors and predators. in Academic Press, Oxford.

Becker, D. J., and R. J. Hall. 2014. Too much of a good thing: resource provisioning alters infectious disease dynamics in wildlife. Biology letters 10:20140309. Becker, D. J., D. G. Streicker, and S. Altizer. 2015. Linking anthropogenic resources to wildlife–pathogen dynamics: a review and meta‐analysis. Ecology letters 18:483-495. Blair, R. B. 2001. Birds and butterflies along urban gradients in two ecoregions of the United States: is urbanization creating a homogeneous fauna? Pages 33-56 in Biotic homogenization. Springer. Boogert, N. J., F. E. Hofstede, and I. A. Monge. 2006. The use of food source scent marks by the stingless bee Trigona corvina (Hymenoptera: Apidae): the importance of the depositor’s identity. Apidologie 37:366-375. Borcard, D., F. Gillet, and P. Legendre. 2011. Numerical ecology with R. Springer Science & Business Media. Boydston, E. E., K. M. Kapheim, and K. E. Holekamp. 2006. Patterns of den occupation by the spotted hyaena (Crocuta crocuta). African Journal of Ecology 44:77-86. Bradburry, J. W., and S. L. Vehrencamp. 2011. Signals and Communication. Pages 1-16 in Principles of Animal Communication. Sinauer Associates, Inc., Publishers, Sunderland, Massachusetts. Bradley, C. A., and S. Altizer. 2007. Urbanization and the ecology of wildlife diseases. Trends in Ecology & Evolution 22:95-102. Brearley, G., C. McAlpine, S. Bell, and A. Bradley. 2012. Influence of urban edges on stress in an arboreal mammal: a case study of squirrel gliders in southeast Queensland, Australia. Landscape Ecology 27:1407-1419. Brown, R. E., and D. W. Macdonald. 1985. Social odours in mammals. Clarendon Press. Burnham, K. P., and D. R. Anderson. 2003. Model selection and multimodel inference: a practical information-theoretic approach. Springer Science & Business Media. Cant, M. A. 2000. Social control of reproduction in banded mongooses. Animal behaviour 59:147-158. Cant, M. A. 2003. Patterns of helping effort in co-operatively breeding banded mongooses (Mungos mungo). Journal of Zoology 259:115-121. Cant, M. A., E. Otali, and F. Mwanguhya. 2002. Fighting and mating between groups in a cooperatively breeding mammal, the banded mongoose. Ethology 108:541-555. Capelle, J., and S. Morand. 2017. Wildlife and emerging diseases in Southeast Asia. Biodiversity Conservation in Southeast Asia: Challenges in a Changing Environment. Child, G. F. 1968. An ecological survey of northeastern Botswana. Food and Agriculture Organization of the United Nations. Clutton-Brock, T., and M. Manser. 2016. Meerkats: cooperative breeding in the Kalahari. Cooperative breeding in vertebrates (eds WD Koenig, JL Dickinson):294-317. Collen, P., and R. Gibson. 2000. The general ecology of beavers (Castor spp.), as related to their influence on stream ecosystems and riparian habitats, and the subsequent effects on fish–a review. Reviews in fish biology and fisheries 10:439-461. Corner, L., D. Pfeiffer, and R. Morris. 2003. Social-network analysis of Mycobacterium bovis transmission among captive brushtail possums (Trichosurus vulpecula). Preventive veterinary medicine 59:147-167.

Cozzi, G., F. Broekhuis, J. W. McNutt, L. A. Turnbull, D. W. Macdonald, and B. Schmid. 2012. Fear of the dark or dinner by moonlight? Reduced temporal partitioning among Africa's large carnivores. Ecology 93:2590-2599. Curtis, D. J., and M. A. Rasmussen. 2002. Cathemerality in lemurs. Evolutionary Anthropology: Issues, News, and Reviews 11:83-86. Derrickson, K. C. 1988. Variation in Repertoire Presentation in Northern Mockingbirds. The Condor 90:592-606. DeSA, U. 2013. World population prospects: the 2012 revision. Population division of the department of economic and social affairs of the United Nations Secretariat, New York. Ditchkoff, S. S., S. T. Saalfeld, and C. J. Gibson. 2006. Animal behavior in urban ecosystems: modifications due to human-induced stress. Urban Ecosystems 9:5- 12. Dizney, L., and M. D. Dearing. 2013. The role of behavioural heterogeneity on infection patterns: implications for pathogen transmission. Animal behaviour 86:911-916. Elfström, M., J. E. Swenson, and J. P. Ball. 2008. Selection of denning habitats by Scandinavian brown bears Ursus arctos. Wildlife Biology 14:176-187. Fairbanks, B. M. 2013a. Bidirectional interactions between behavior and disease in banded mongoose (Mungos mungo) infected with Mycobacterium mungi. Virginia Polytechnic Institute and State University, Blacksburg, VA. Fairbanks, B. M. 2013b. Bidirectional interactions between behavior and disease in banded mongooses (Mungos mungo) infected with Mycobacterium mungi. Fairbanks, B. M., D. M. Hawley, and K. A. Alexander. 2014. No evidence for avoidance of visibly diseased conspecifics in the highly social banded mongoose (Mungos mungo). Behavioral Ecology and Sociobiology:1-11. Flint, B. F., D. M. Hawley, and K. A. Alexander. 2016. Do not feed the wildlife: associations between garbage use, aggression, and disease in banded mongooses (Mungos mungo). Ecology and evolution 6:5932-5939. Fox, J. T., M. E. Vandewalle, and K. A. Alexander. 2017. Land Cover Change in Northern Botswana: The Influence of Climate, Fire, and Elephants on Semi-Arid Savanna Woodlands. Land 6:73. Furrer, R. D., and M. B. Manser. 2009. Banded mongoose recruitment calls convey information about risk and not stimulus type. Animal behaviour 78:195-201. Gosling, L. M., S. C. Roberts, E. Thornton, and M. Andrew. 2000. Life history costs of olfactory status signalling in mice. Behavioral Ecology and Sociobiology 48:328- 332. Gottdenker, N. L., D. G. Streicker, C. L. Faust, and C. Carroll. 2014. Anthropogenic land use change and infectious diseases: a review of the evidence. Ecohealth 11:619- 632. Government, B. 2015. Population and Housing Survey Census 2011. in Analytical Report. Gaborone, Botswana. Gower, J. C. 1971. A general coefficient of similarity and some of its properties. Biometrics:857-871. Grinnell, J., and K. McComb. 2001. Roaring and social communication in African lions: the limitations imposed by listeners. Animal behaviour 62:93-98.

Hawley, D. M., and S. M. Altizer. 2011. Disease ecology meets ecological immunology: understanding the links between organismal immunity and infection dynamics in natural populations. Functional Ecology 25:48-60. Henderson, R. W., and R. Powell. 2001. Responses by the West Indian herpetofauna to human-influenced resources. Caribbean Journal of Science 37:41-54. Herr, J., L. Schley, E. Engel, and T. J. Roper. 2010. Den preferences and denning behaviour in urban stone martens (Martes foina). Mammalian Biology-Zeitschrift für Säugetierkunde 75:138-145. Hill, R. 2006. Why be diurnal? Or, why not be cathemeral? Folia Primatologica 77:72-86. Hiscocks, K. 1999. The impact of an increasing elephant population on the woody vegetation in southern Sabi Sand Wildtuin, South Africa. Koedoe 42:47-55. Hiscocks, K., and M. Perrin. 1991. Den selection and use by dwarf mongooses and banded mongooses in South Africa. S. AFR. J. WILDL. RES./S.-AFR. TYDSKR. NATUURNAVORS. 21:119-122. Hutton, P., and K. J. McGraw. 2016. Urban Impacts on oxidative balance and animal signals. Frontiers in Ecology and Evolution 4:54. Hwang, Y. T., S. Lariviere, and F. Messier. 2007. Local-and landscape-level den selection of striped skunks on the Canadian prairies. Canadian Journal of Zoology 85:33-39. Jennings, S., and G. Mutze. 2018. Multiple warren use by subadult wild rabbits, Oryctolagus cuniculus, and its implications for disease transmission. Wildlife Research 44:582-586. Jobbins, S. E., and K. A. Alexander. 2015. Evidence of Leptospira sp. infection among a diversity of African wildlife species: beyond the usual suspects. Transactions of The Royal Society of Tropical Medicine and Hygiene 109:349-351. Jones, K. E., N. G. Patel, M. A. Levy, A. Storeygard, D. Balk, J. L. Gittleman, and P. Daszak. 2008. Global trends in emerging infectious diseases. Nature 451:990-993. Jordan, N. R., M. B. Manser, F. Mwanguhya, S. Kyabulima, P. Rüedi, and M. A. Cant. 2011a. Scent marking in wild banded mongooses: 1. Sex-specific scents and overmarking. Animal behaviour 81:31-42. Jordan, N. R., F. Mwanguhya, R. D. Furrer, S. Kyabulima, P. Rüedi, and M. A. Cant. 2011b. Scent marking in wild banded mongooses: 2. Intrasexual overmarking and competition between males. Animal behaviour 81:43-50. Jordan, N. R., F. Mwanguhya, S. Kyabulima, P. Ruedi, and M. A. Cant. 2010. Scent marking within and between groups of wild banded mongooses. Journal of Zoology 280:72-83. Jordan, N. R., F. Mwanguhya, S. Kyabulima, P. Rüedi, S. J. Hodge, and M. A. Cant. 2011c. Scent marking in wild banded mongooses: 3. Intrasexual overmarking in females. Animal behaviour 81:51-60. Kavaliers, M., and D. D. Colwell. Discrimination by female mice between the odours of parasitized and non-parasitized males. The Royal Society, 1995. Kay, R. F., and E. C. Kirk. 2000. Osteological evidence for the evolution of activity pattern and visual acuity in primates. American Journal of Physical Anthropology 113:235-262.

Keay, J. M., J. Singh, M. C. Gaunt, and T. Kaur. 2006. Fecal glucocorticoids and their metabolites as indicators of stress in various mammalian species: a literature review. Journal of Zoo and Wildlife Medicine 37:234-244. Klein, S. L., H. R. Gamble, and R. J. Nelson. 1999. Trichinella spiralis infection in voles alters female odor preference but not partner preference. Behavioral Ecology and Sociobiology 45:323-329. Klinka, D., and T. Reimchen. 2002. Nocturnal and diurnal foraging behaviour of brown bears (Ursus arctos) on a salmon stream in coastal British Columbia. Canadian Journal of Zoology 80:1317-1322. Kotze, R., N. C. Bennett, E. Z. Cameron, J. L. de Vries, D. G. Marneweck, C. W. Pirk, and F. Dalerum. 2012. Temporal patterns of den use suggest polygamous mating patterns in an obligate monogamous mammal. Animal behaviour 84:1573-1578. La, V. T. 2012. Diurnal and nocturnal birds vocalize at night: a review. The Condor 114:245-257. Laundré, J. W., L. Hernández, and W. J. Ripple. 2010. The landscape of fear: ecological implications of being afraid. Open Ecology Journal 3:1-7. Laver, P. N. 2013a. The foraging ecology of banded mongooses (Mungos mungo): epidemiological and human-wildlife conflict implications. Laver, P. N. 2013b. The foraging ecology of banded mongooses (Mungos mungo): epidemiological and human-wildlife conflict implications. Virginia Polytechnic Institute and State University, Blacksburg, VA. Laver, P. N., and K. A. Alexander. 2017. How living in association with humans, and seasonality interact to reverse traditional animal space use predictions. Movement Ecology. Laver, P. N., A. Ganswindt, S. B. Ganswindt, and K. A. Alexander. 2012. Non-invasive monitoring of glucocorticoid metabolites in banded mongooses (Mungos mungo) in response to physiological and biological challenges. General and Comparative Endocrinology 179:178-183. Lemon, S. C., J. Roy, M. A. Clark, P. D. Friedmann, and W. Rakowski. 2003. Classification and regression tree analysis in public health: methodological review and comparison with logistic regression. Annals of behavioral medicine 26:172- 181. Lepetu, J., J. Alavalapati, and P. K. Nair. 2010. Forest dependency and its implication for protected areas management: A case study from Kasane Forest Reserve, Botswana. Lima, S. L., and L. M. Dill. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68:619-640. Lloyd-Smith, J. O., S. J. Schreiber, P. E. Kopp, and W. M. Getz. 2005. Superspreading and the effect of individual variation on disease emergence. Nature 438:355-359. Lowry, H., A. Lill, and B. Wong. 2013. Behavioural responses of wildlife to urban environments. Biological Reviews 88:537-549. Magle, S. B., V. M. Hunt, M. Vernon, and K. R. Crooks. 2012. Urban wildlife research: past, present, and future. Biological Conservation 155:23-32. McKinney, M. L. 2002. Urbanization, Biodiversity, and Conservation The impacts of urbanization on native species are poorly studied, but educating a highly

urbanized human population about these impacts can greatly improve species conservation in all ecosystems. BioScience 52:883-890. Mitchell, J., M. A. Cant, E. I. Vitikainen, and H. J. Nichols. 2017. Smelling fit: scent marking exposes parasitic infection status in the banded mongoose. Current Zoology 63:237-247. Mosugelo, D. K., S. R. Moe, S. Ringrose, and C. Nellemann. 2002. Vegetation changes during a 36‐year period in northern Chobe National Park, Botswana. African Journal of Ecology 40:232-240. Müller, C. A., and M. B. Manser. 2007. ‘Nasty neighbours’ rather than ‘dear enemies’ in a social carnivore. Proceedings of the Royal Society B: Biological Sciences 274:959-965. Neal, E. 1970. The banded mongoose, Mungos mungo Gmelin. African Journal of Ecology 8:63-71. Nichols, C. A., M. E. Vandewalle, and K. A. Alexander. 2017. Emerging threats to dryland forest resources: elephants and fire are only part of the story. Forestry: An International Journal of Forest Research 90:473-484. Otali, E., and J. S. Gilchrist. 2004. The effects of refuse feeding on body condition, reproduction, and survival of banded mongooses. Journal of Mammalogy 85:491- 497. Pacheco, E. R. 2015. Unsupervised Learning with R. Packt Publishing Ltd. Palmer, M. A., A. P. Covich, S. Lake, P. Biro, J. J. Brooks, J. Cole, C. Dahm, J. Gibert, W. Goedkoop, and K. Martens. 2000. Linkages between Aquatic Sediment Biota and Life Above Sediments as Potential Drivers of Biodiversity and Ecological Processes: A disruption or intensification of the direct and indirect chemical, physical, or biological interactions between aquatic sediment biota and biota living above the sediments may accelerate biodiversity loss and contribute to the degradation of aquatic and riparian habitats. AIBS Bulletin 50:1062-1075. Paull, S. H., S. Song, K. M. McClure, L. C. Sackett, A. M. Kilpatrick, and P. T. Johnson. 2012. From superspreaders to disease hotspots: linking transmission across hosts and space. Frontiers in Ecology and the Environment 10:75-82. Penn, D., G. Schneider, K. White, P. Slev, and W. Potts. 1998. Influenza infection neutralizes the attractiveness of male odour to female mice (Mus musculus). Ethology 104:685-694. Périquet, S., C. Mapendere, E. Revilla, J. Banda, D. Macdonald, A. Loveridge, and H. Fritz. 2016. A potential role for interference competition with lions in den selection and attendance by spotted hyaenas. Mammalian Biology-Zeitschrift für Säugetierkunde 81:227-234. Petrelli, A. R., M. J. Levenhagen, R. Wardle, J. R. Barber, and C. D. Francis. 2017. First to Flush: The Effects of Ambient Noise on Songbird Flight Initiation Distances and Implications for Human Experiences with Nature. Frontiers in Ecology and Evolution 5:67. Pickett, S. T., and M. L. Cadenasso. 1995. Landscape ecology: spatial heterogeneity in ecological systems. Science 269:331. Plaat, S. A., E. Deckers, G. Degenaars, I. Martorelli, and J. van Kollenburg. Exploratory Data Analysis on Multivariate Data. Plotkin, H. C. 1988. The role of behavior in evolution. MIT press.

Plowright, R. K., S. H. Sokolow, M. E. Gorman, P. Daszak, and J. E. Foley. 2008. Causal inference in disease ecology: investigating ecological drivers of disease emergence. Frontiers in Ecology and the Environment 6:420-429. Poteet, M. F. 2006. Shifting roles of abiotic and biotic regulation of a multi-host parasite following disturbance. Disease ecology. Oxford University Press, New York, New York, USA:135-153. Ramalho, C. E., and R. J. Hobbs. 2012. Time for a change: dynamic urban ecology. Trends in Ecology & Evolution 27:179-188. Rasmussen, G., and D. Macdonald. 2012. Masking of the zeitgeber: African wild dogs mitigate persecution by balancing time. Journal of Zoology 286:232-242. Reichman, O., and S. C. Smith. 1990. Burrows and burrowing behavior by mammals. Current mammalogy 2:197-244. Reidy, J. L., M. M. Stake, and F. R. Thompson III. 2009. Nocturnal predation of females on nests: an important source of mortality for Golden-cheeked Warblers? The Wilson Journal of Ornithology 121:416-421. Rich, T. J., and J. L. Hurst. 1999. The competing countermarks hypothesis: reliable assessment of competitive ability by potential mates. Animal behaviour 58:1027- 1037. Rood, J. P. 1975. Population dynamics and food habits of the banded mongoose (Mungos mungo). African Journal of Ecology 13:89-111. Rood, J. P. 1983. Banded mongoose rescues pack member from eagle. Animal behaviour 31:1261-1262. Rousseeuw, P. J. 1987. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics 20:53-65. Russo, D., and L. Ancillotto. 2015. Sensitivity of bats to urbanization: a review. Mammalian Biology-Zeitschrift für Säugetierkunde 80:205-212. Schaik, C. P., and P. M. Kappeler. 1996. The social systems of gregarious lemurs: lack of convergence with anthropoids due to evolutionary disequilibrium? Ethology 102:915-941. Scheiner, S. M., and M. R. Willig. 2011. The theory of ecology. University of Chicago Press. Schrag, S. J., and P. Wiener. 1995. Emerging infectious disease: what are the relative roles of ecology and evolution? Trends in Ecology & Evolution 10:319-324. Skarpe, C., P. A. Aarrestad, H. P. Andreassen, S. S. Dhillion, T. Dimakatso, J. T. du Toit, D. J. Halley, H. Hytteborn, S. Makhabu, and M. Mari. 2004. The return of the giants: ecological effects of an increasing elephant population. AMBIO: A Journal of the Human Environment 33:276-282. Stoddart, D. M. 1976. Mammalian odours and pheromones. Edward Arnold London. Tattersall, I. 1979. Patterns of activity in the Mayotte lemur, Lemur fulvus mayottensis. Journal of Mammalogy 60:314-323. _____. 1987. Cathemeral activity in primates: a definition. Folia Primatologica 49:200- 202. Theuerkauf, J., S. Rouys, and W. Jedrzejewski. 2003. Selection of den, rendezvous, and resting sites by wolves in the Bialowieza Forest, Poland. Canadian Journal of Zoology 81:163-167.

Thompson, F. J., H. H. Marshall, E. I. Vitikainen, and M. A. Cant. 2017. Causes and consequences of intergroup conflict in cooperative banded mongooses. Animal behaviour 126:31-40. Veron, G., M. Colyn, A. E. Dunham, P. Taylor, and P. Gaubert. 2004. Molecular systematics and origin of sociality in mongooses (Herpestidae, Carnivora). Molecular phylogenetics and evolution 30:582-598. Ward, M. P., M. Alessi, T. J. Benson, and S. J. Chiavacci. 2014. The active nightlife of diurnal birds: extraterritorial forays and nocturnal activity patterns. Animal behaviour 88:175-184. Washabaugh, K., and C. T. Snowdon. 1998. Chemical communication of reproductive status in female cotton‐top tamarins (Saguinus oedipusoedipus). American Journal of Primatology 45:337-349. Watson, J. T., M. Gayer, and M. A. Connolly. 2007. Epidemics after natural disasters. Emerging infectious diseases 13:1. Weare, P., and A. Yalala. 1971. Provisional vegetation map of Botswana. Botswana Notes and Records 3:1-1. Wielebnowski, N., and J. Watters. 2007. Applying fecal endocrine monitoring to conservation and behavior studies of wild mammals: important considerations and preliminary tests. Israel journal of ecology & evolution 53:439-460. Willis, C., and R. Poulin. 2000. Preference of female rats for the odours of non- parasitised males: the smell of good genes? Folia Parasitologica 47:6-10. Wolf, A. 2009. Preliminary assessment of the effect of high elephant density on ecosystem components (grass, trees, and large mammals) on the chobe riverfront in Northern Botswana. University of Florida. Woodroffe, R. Predators and people: using human densities to interpret declines of large carnivores. Cambridge University Press, 2000. Woolhouse, M. E., and S. Gowtage-Sequeria. Host range and emerging and reemerging pathogens. 2006. Wright, P. C. 1999. Lemur traits and Madagascar ecology: coping with an island environment. American Journal of Physical Anthropology 110:31-72. Wu, T., C. Perrings, A. Kinzig, J. P. Collins, B. A. Minteer, and P. Daszak. 2017. Economic growth, urbanization, globalization, and the risks of emerging infectious diseases in China: a review. Ambio 46:18-29. Yang, H., A. Viña, Y. Tang, J. Zhang, F. Wang, Z. Zhao, and J. Liu. 2017. Range-wide evaluation of wildlife habitat change: A demonstration using Giant Pandas. Biological Conservation 213:203-209. Young, J. S., E. Gormley, and E. M. Wellington. 2005. Molecular detection of Mycobacterium bovis and Mycobacterium bovis BCG (Pasteur) in soil. Applied and environmental microbiology 71:1946-1952. Zala, S. M., W. K. Potts, and D. J. Penn. 2004. Scent-marking displays provide honest signals of health and infection. Behavioral Ecology 15:338-344.

79 Conclusions

We identified interesting implications for disease pathogen transmission across landscape type in Northern Botswana. The assessment of den site characteristics along the human- wildlife interface exemplified the ability of banded mongooses to locate and utilize human-subsidized resources for denning needs. The anthropogenic den sites used by mongooses remain intact for much longer periods of time than natural dens. We also identified locations where den sharing is ongoing in the human-mediated landscapes.

Both the lack of natural regeneration processes and therefore, continued use for anthropogenic dens and den sharing likely have implications for pathogen transmission both within and between troops.

We showed that the landscape influences wildlife behaviors in complex ways that can influence pathogen transmission. Both the land use areas, vigilance behaviors, and the interaction between the two explained the variation in the olfactory communication behaviors that transmit M. mungi. Vigilance appears to be a stimulant for olfactory communication when the threat is competition related, but a reducer when the threat is predation related. However, there are situations when predation risks are inspected or mobbed (e.g. snakes). Vigilance is a multifaceted behavior and its influence on olfactory communication across the landscape reflects this. The intense competition between neighboring troops of mongooses in areas with human-subsidized resources likely promotes between-troop transmission.

Our use of remote sensing camera traps allowed us to study nighttime behavior, which had not been previously documented for this species. The categorization of a species as diurnal or nocturnal usually limits behavior observations to the hours of dominant activity periods. However, though rare, these events can have substantial influences on ecological interactions and should be accounted for in methodologies, especially when applicable to study questions and objectives.

As numerous species begin to utilize anthropogenic landscapes, these fundamental changes in behavior and life history attributes should be taken into consideration for conservation and management purposes. We have shown the implications of behavioral changes across the landscape on disease transmission dynamics, but it is likely that additional interactions are affected as well. Future studies should focus on refining our understanding of sickness behavior on pathogen shedding. Additional focus on the period of pathogen shedding will be important as we seek to refine computational models and outbreak prediction capacity.

81 Appendix

Supplementary Table 1: The wildlife species captured in camera trap photographs at banded mongoose den sites in Northern Botswana.