Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: Prev Vet Med. 2020 Mar 31;178:104978. doi: 10.1016/j.prevetmed.2020.104978

Movement patterns of free-roaming dogs on heterogeneous urban landscapes: implications for rabies control

Brinkley Raynor 1, Micaela De la Puente-León 2, Andrew Johnson 3, Elvis W Díaz 2, Michael Z Levy 1,2, Sergio E Recuenco 4, Ricardo Castillo-Neyra 1,2,*
PMCID: PMC7234908  NIHMSID: NIHMS1586013  PMID: 32302776

Abstract

In 2015, a case of canine rabies in Arequipa, Peru indicated the re-emergence of rabies virus in the city. Despite mass dog vaccination campaigns across the city and reactive ring vaccination and other control activities around positive cases (e.g. elimination of unowned dogs), the outbreak has spread. Here we explore how the urban landscape of Arequipa affects the movement patterns of free-roaming dogs, the main reservoirs of the rabies virus in the area. We tracked 23 free-roaming dogs using Global Positioning System (GPS) collars. We analyzed the spatio-temporal GPS data using the time-local convex hull method. Dog movement patterns varied across local environments. We found that water channels, an urban feature of Arequipa that are dry most of the year, promote movement. Dogs that used the water channels extensively move on average 7 times further (p=0.002) and 1.2 times more directionally (p=0.027) than dogs that do not use the water channels at all. They were also 1.3 times faster on average, but this difference was not statistically significant (p= 0.197). Our findings suggest that water channels can be used by dogs as ‘highways’ to transverse the city and have the potential to spread disease far beyond the radius of control practices. Control efforts should focus on a robust vaccination campaign attuned to the geography of the city, and not limited to small-scale rings surrounding cases.

Keywords: Canine rabies, temporo-spatial analysis, ecological corridors

Introduction

Since March 2015, hundreds of rabid dogs have been detected in the city of Arequipa in southern Peru, following 15 years of epidemiological silence (Castilalo-Neyra et al., 2017; Perú Ministerio de Salud, 2015). The system of water channels in Arequipa, a unique urban landscape feature, has been associated with the location of rabid dogs, suggesting that the city’s complex environment influences the transmission of the disease (Castillo-Neyra et al., 2017). In addition to the landscape, Arequipa, like much of Latin America, is characterized by complex socio-ecological characteristics that complicate efforts to eliminate rabies.

Dog ownership practices interplay with unique geographical features of Arequipa city to impact both dog ecology and rabies control. In many areas of Latin America, including Arequipa, owned dogs commonly have access to the street without owner supervision (Castillo-Neyra et al., 2019); these dogs receive varying degrees of feed and veterinary care (Acosta-Jamett et al., 2010; Cortez-Aguirre et al., 2018; Flores-Ibarra and Estrella-Valenzuela, 2004; Guilloux et al., 2018; Jackman and Rowan, 2007; Vigilato et al., 2013). Populations of free-roaming dogs without owners (strays) are believed to be small due in part to the high mortality of stray dogs (Belo et al., 2017; Font, 1987).

There is some evidence that Arequipa’s landscape may affect free-roaming dog movement and ecology. A distinctive geographic feature of the city is an extensive system of open water channels that are dry most of the year (Figure 1), and only carry water during the short rainy season between January and March. When dry, these water channels collect garbage and dogs have been observed foraging for food through the waste. These channels have been spatially associated with canine rabies cases (Castillo-Neyra et al., 2017), and we hypothesize that they might be used as ecological corridors by dogs.

Figure 1: Dry water channels.

Figure 1:

Open in a new tab

Water channels traversing the City of Arequipa, Peru are dry most of the year. They accumulate trash and attract free roaming dogs.

The ecology of urban rabies is intimately connected with dog population dynamics and home range, critical concepts in disease ecology (Burt, 1943; Davlin and VonVille, 2012; Slater, 2001). In a seminal paper introducing home range in 1943, Burt describes the concept of home range as “the area traversed by the individual in its normal activities of food gathering, mating, and caring for young” (Burt, 1943). The study of animal home range has advanced considerably with the incorporation of Global Positioning System (GPS) technology and has allowed for sophisticated analyses of temporal and spatial usage data (Lyons et al., 2013).

There is very little literature on the home ranges and movement patterns of free-roaming dogs; most published studies are focused on rural settings (Dürr and Ward, 2014; Hudson et al., 2017; Meek, 1999; Sepúlveda et al., 2015; van Bommel and Johnson, 2014). These studies in rural areas suggest that dogs home ranges are heterogeneous with variation being influenced by dog function, resource availability, biological characteristics and human interactions (Dürr and Ward, 2014; Hudson et al., 2017; Laager et al., 2018; Meek, 1999; Sepúlveda et al., 2015; van Bommel and Johnson, 2014). In Chile, rural dogs had a preference to move using trails and roads, avoiding dense vegetation (Sepúlveda et al., 2015). Very little is known about how the urban landscape affects the space use and movement of free-roaming dogs in cities. A study in Baltimore, Maryland done before the application of GPS technology found that dogs used alleys behind houses to forage and move (Beck, 1973). The larger set of studies on wild animals in urban environments also suggests that animal movement is strongly influenced by landscape features. For instance, cougars preferentially use dry river beds to move in bordering urban areas in California, foxes use ravines in Toronto to move across the city, and bobcats and coyotes use culverts and connected fragments of vegetation to move through southern California (Dickson and Beier, 2007; Rosatte and Allan, 2009; Tigas et al., 2002).

Since March 2015, over 190 rabid dogs have been detected in the city of Arequipa in southern Peru, following 15 years of epidemiological silence (Centro Nacional de Epidemiología, Prevención y Control de Enfermedades, 2017; Centro Nacional de Epidemiología, Prevención y Control de Enfermedades, 2018; Centro Nacional de Epidemiología, Prevención y Control de Enfermedades, 2019; Prevención y Control de Enfermedades Centro Nacional de Epidemiología, 2016). The current response protocol is to conduct 300–500 meter ring containment activities, which include vaccination of owned dogs and culling of unowned, and sometimes owned, free-roaming dogs around each detected case (MINSA, 2017). However, these focalized small-scale interventions are not supported by data or scientific evidence (Rhodes and Anderson, 1997; WHO, 2018). Small, reactive vaccination campaigns are also not recommended when the disease has spread over extended areas (WHO, 2018) and euthanasia is recommended only for rabid dogs or suspected cases. A deep understanding of dog ecology is critical to rabies control (Davlin and VonVille, 2012) and elucidating how dogs move within cities will provide important insights towards this end. In this study we focused on the movements of free-roaming owned dogs. Our objectives were to quantify the variability in dogs’ home ranges, to characterize dogs’ movement in urban areas, and to identify how water channels might affect them.

Material and methods

Study Sites and Population

The city of Arequipa is surrounded by a desert. There are some wild mammals in this desert, but they are few and far between, and no evidence that wildlife species sustain, or could sustain, rabies virus transmission (Castillo-Neyra et al., 2017). Three urban localities, a political city district subdivision, were selected for inclusion in this study (Figure 2). One of the localities is urban and intersected by a water channel; the other two localities are peri-urban, one intersected by a water channel and one approximately a kilometer from the nearest water channel. Houses were selected purposively based on location with respect to the water channels. Dogs were eligible for inclusion if they were at least 1 year old, were apparently healthy at the time of the first visit, and were medium or large dogs (at least 35cm at the withers)(The Kennel Club, 2018). When multiple dogs were found in a house, we chose the first dog seen by the field team. Dogs of both sexes were included. All dogs included in the study had unrestricted access to the street most of the day as stated by their owners. The characteristics of the selected dog population are summarized in Table 1.

Figure 2: Map of study area and location of dogs’ houses.

Figure 2:

Open in a new tab

Dogs were selected for inclusion in the study in both urban and peri-urban locals with houses either directly adjacent to a water channel or approximately a kilometer from a water channel.

Table 1.

Dog characteristics of study population

Urban near water channels (n=8) Periurban near water channels (n=9) Periurban away from water channels (n=6)
Sex [no. (%)] Male 6 (75.0%) 7 (77.8%) 6 (100.0%)
Female 2 (25.0%) 2 (22.2%) 0 (0.0%)
Age [no. (%)] <= 3 years 5 (62.5%) 4 (44.4%) 0 (0.0%)
4–6 years 2 (25.0%) 4 (44.4%) 4 (66.6%)
7–9 years 0 (0.0%) 0 (0.0%) 2 (33.3%)
> 9 years 1 (12.5%) 1 (11.1%) 0 (0.0%)
Open in a new tab

GPS tracking of dogs

To record dog locations and speeds we used IgotU® GT-120 (Mobile Action Technology) GPS loggers that have shown good accuracy in urban environments in Peru (point accuracy of 4.4 m and line accuracy of 10.3 m) (Vazquez-Prokopec et al., 2009) and have been used to track free-roaming domestic animals (Pray et al., 2016). Given the high-speed dogs can achieve, we programmed the GPS receivers to log the dog’s geographical coordinates every 3 minutes. For every dog, one GPS logger was turned on, and placed in a nylon collar on the dog (Figure 3). The GPS device were hidden in the collars and weighed less than 20 grams (Figure 3). After 4 to 7 days, depending on dog and owner availability, we returned to enrolled households. If the dog owner allowed us to continue tracking the dog, we changed the GPS receiver in the collar for another one with fresh batteries or took the collar off if the dog owner requested to drop from the study. Follow-up times per dog varied from 4 days to 4 weeks. We estimated the correlation between observation period duration and home range to evaluate if home range estimates were affected by the variability in observation duration

Figure 3: Dog wearing GPS collar.

Figure 3:

Open in a new tab

Dog fit with GPS collar, IgotU® GT-120 (Mobile Action Technology) GPS loggers. Battery life of the collars lasted 4–7 days while recording GPS coordinates every 3 minutes.

Data cleaning

In cities there is potential for urban structures to affect the accuracy of GPS loggers. To handle this potential issue, we excluded speed points that were higher than 30% the fastest speed recorded for a dog (17.003 m/s) (Denny, 2008). This speed limit (5.101 m/s) is similar to the speed limit used by Dürr and Ward (5.556 m/s) in their study of Australian free-roaming rural dogs (Dürr and Ward, 2014).

Statistical Analysis and Mapping

T-LoCoH

We analyzed dog home ranges and other dog movement metrics using the Time Localized Convex Hulls (T-LoCoH) method. T-LoCoH extends the classically used Local Convex Hull (LoCoH) to include temporal data in the calculations. Briefly, the LoCoH method creates individual convex polygon hulls from a set of selected nearest neighbors around each GPS point without taking account when those nearest neighbors were registered (Getz et al., 2007). T-LoCoH goes beyond the classic LoCoH by using “time-scaled distance (TSD)” instead of geographical distance to create local hulls around GPS points (Lyons et al., 2013). This approach is ideal for modern GPS data that standardly includes a time stamp with the GPS coordinates. The advantage of T-LoCoH to LoCoH and other previous methods is that nearest neighbors are selected based on closeness in both space and time with the weight given to time closeness based on parameters that can be inferred from the data (Lyons et al., 2013).

Parameter estimation

We used the a-method of nearest neighbor selection. The a-method selects neighbors by finding the difference in time scaled distance between the parent point and other points in the set (Lyons et al., 2013). Then the nearest points are added in ascending order of time scaled distance until some value, a, is reached (Lyons et al., 2013).

In order to calculate time scaled distance, the central feature of T-LoCoH, a scaling factor, s, is needed. The s value determines the weight that time is given when selecting nearest neighbors, where s=0 would not take time into account when selecting neighbors, while a s=1 would select neighbors completely based off of time (Lyons et al., 2013). We chose our s value based off temporal cyclicity observed in the field. The free-roaming dogs of Arequipa exhibited distinct morning, afternoon, evening, and night behaviors, so we decided to base our s value off a 6-hour cycle time of interest.

For each dog we estimated an s value such that the number of time-selected and space-selected neighbors were balanced equally for specified observation periods of interest (Lyons and Andy, 2014). Once we calculated an s value for each dog, we calculated the a values. To do this, we selected an a value that did not cause drastic increases or decreases in isopleth area and isopleth perimeter (Lyons and Andy, 2014). A local convex polygon was drawn around each set to create a hull and a combined set of hulls sorted in a specific way is referred to as an isopleth (Lyons et al., 2013).

Spatio-temporal Models

We used the T-LoCoH package in R (Lyons et al., 2013) to create hulls; the distance between points was calculated in time-scaled distance using our selected s value and then points included in each hull were found by summing the distance between closest neighbors until the a value was reached (but not surpassed) (Lyons and Andy, 2014). Isopleths were sorted by both density of points and by eccentricity of the minimal area bounding ellipse eccentricity of each hull (Lyons et al., 2013). Eccentricity is a measure of how directed a movement is; very directed movements having an eccentricity close to 1, while the eccentricity of more random movements is closer to 0 (Lyons et al., 2013). Core home ranges were calculated using the densest 50% isopleth, while extended home range the densest 95% isopleth (Dürr and Ward, 2014; Zehnder et al, 2018).

Mapping

Dog home ranges were mapped using the R package ggmap (Kahle and Wickham, 2015) on top of a background generated from OpenStreetMap data (OpenStreetMap Contributers, 2015).

Time spent in the water channels

We used the proportion of recorded locations within 30 m of the water channels as a proxy for the proportion of time spent within these structures. We compared the time spent in the water channels between dogs living close and far away from the water channels. We eliminated from this analysis the points recorded at the dogs’ house, so that the proportion of time not spent at the water channel represents time spent on the streets or other open areas (e.g. parks).

Statistics

All analyses, maps, and graphs were produced with R (R Core Team, 2017). We compared the core (50% isopleth) and extended (95% isopleth) home ranges, farthest GPS point recorded from the dogs’ homes, and the average eccentricity of hulls between dogs living in 3 different areas and between dogs exhibiting different behaviors in respect to water channel usage. To group dogs based on water channel usage, first a group was made of dogs who had no GPS locations recorded in the water channels. Then the total length utilized (length of water channel in between farthest points in the water channel recorded) was calculated for all remaining dogs. K-means analysis was used to separate dogs into light and heavy water channel users.

Results

We were able to collar 25 dogs and retrieve data from 23 of them. One dog did not have the collar at the second visit, and we decided to remove it from the study. For another, the logger did not record any data, and the dog owner did not want to continue in the study. The sex and age of the 23 dogs are summarized in Table 1. One dog was included in the study because the owner said it was one year old, but later she rectified it was 10 months old. We kept the dog in the study. Spaying and neutering dogs is a very rare practice in the study area and, as expected, none of the dogs tracked in this study was spayed or neutered.

After data cleaning, our total data set of 23 dogs included 74,120 observations. 595 observations were excluded for being faster than our set cut-off speed. We suspect that most points recorded above our cut-off speed were due to GPS error. The excluded points contained some few speeds that could plausibly be attributed to true dog movement.

We observed high variability in home ranges (Figure 4) as well as eccentricity (Figure 5). We also found high variability among what times of the day dogs moved the most (Figure 6). We found important differences when comparing dogs based on their water channel usage. In Figure 4 and 5 we show the home ranges and eccentricity maps of 3 dogs, A, B, and C. These dogs exemplify movement patterns of dogs that either never use the water channels (A), lightly use the water channels (B) or heavily use the water channels (C). Home ranges and eccentricity maps of all dogs are included in the supplementary materials.

Figure 4: Core and extended home range of 3 dogs that exemplify different movement behavior patterns.

Figure 4:

Open in a new tab

Maps of the core (50% isopleth) and extended (95% isopleths) home ranges for dogs with three different behavior patterns. Dog A tends to stay at home or close to the house most of the day and never goes into the water channels. Dog B stays in a small, defined area around her house and goes into the water channel in proximity to her house. Dog C ranges far from his house and has points going along multiple water channels.

Figure 5: Eccentricity of 3 different dogs exemplifying different movement behaviour patterns.

Figure 5:

Open in a new tab

Maps of Dog A, B, and C’s isopleths sorted by eccentricity: the redder the isopleth color, the greater eccentricity values of the isopleths.

Figure 6: Average distance travelled per hour of day by dogs in the city of Arequipa, 2018.

Figure 6:

Open in a new tab

The background colors represent night-time (dark grey), dawn/dusk (light grey), and daytime hours (white). Panel A is a plot of the mean of each individual dogs’ distance travelled for each hour of the day. Panel B represents the mean of the results shown in panel A with the range bars displaying the minimum and maximum.

Dogs that used the water channels more tended to have the largest core and extended home ranges. Core home ranges, based on the 50% density isopleth, ranged from 0.00013 km2 to 0.46 km2. Extended home ranges, based on the 95% density isopleth, ranged from 0.0012 km2 to 3.70 km2. Water channel usage was also associated with moving the longest distance from home (p=0.002) and moving with higher directionality (p=0.027) (Table 2). One dog that regularly used the water channels travelled up to 14 km from its home. The maximum speed achieved by dogs also increased with water channel usage, but there was not statistical association. Interestingly, no significant differences in dog movement patterns were found in dogs grouped by home location; neither distance from the water channels nor the level of development of the neighborhood appeared to affect distance travelled.

Table 2:

Area utilization based on dog’s usage of water channel

Heavy water channel usage (n=4) Light water channel usage (n=13) No water channel usage (n=6) p-value (Kruskal Wallis)
Core home range area (50% isopleth) in km2
Median (IQR) 0.156 (.280) 0.011 (0.015) 0.003 (0.006) 0.010
mean (SD) 0.199 (0.208) 0.013 (0.011) 0.011 (0.019)
Extended home range area (95% isopleth) in km2
Median (IQR) 1.83 (1.83) 0.055 (0.065) 0.017 (0.018) 0.002
Mean (SD) 1.951 (1.447) 0.102 (0.113) 0.020 (0.018)
Longest distance from home
Median (IQR) 3.572 (3.533) 1.355 (1.629) 0.982 (0.291) 0.002
Mean (SD) 6.043 (SD) 2.251 (SD) 0.866 (SD)
Eccentricity
Median (IQR) 0.771 (0.037) 0.618 (0.079) 0.613 (0.057) 0.027
Mean (SD) 0.746 (0.055) 0.624 (0.078) 0.622 (0.037)
Maximum speeds
Median (IQR) 4.994 (0.057) 4.772 (0.511) 3.856 (0.882) 0.197
Mean (SD) 4.971 (0.091) 4.642 (0.495) 3.734 (1.462)
Open in a new tab

Discussion

We found a strong effect of the urban landscape on dog movement in Arequipa, Peru. Dogs that spend more time in the water channels have more linear movements, significantly larger home ranges, and venture further from home than those that spend little or no time in these channels. Our findings suggest that the water channels in Arequipa function as ecological corridors, a component of dog ecology that could affect rabies virus transmission. These corridors might greatly complicate the control of the transmission of rabies, as corroborated by previous studies based on the distance of canine rabies cases from the water channels (Castillo-Neyra et al., 2017). The traditional response to every detected rabid dog is to conduct “ring containment”. The principle of ring containment is adapted from ring vaccination where all potential contacts with a positive case are immunized to create a buffer preventing disease spread (Deen and von Seidlein, 2018). The current response protocol is to conduct ring containment activities, which include vaccination of owned dogs and culling of unowned, and sometimes owned, free-roaming dogs around each detected case, as well as conduct health promotion focused on rabies and find people who might need post-exposure prophylaxis (MINSA, 2017). The national technical guidelines in Peru suggest a 500-m radius around the case. However, this radius varies depending on the availability of resources (e.g. number of vaccinators), the political boundaries, and the geographical features (e.g. the area of containment would not cross a political boundary or a river) (MINSA, 2017). We found that apparently healthy dogs move far beyond the current fixed-radius ring containment on a regular basis. Rabid dogs sometimes exhibit erratic behavior with some records of dogs moving more than 15 km from home and have varied movement patterns (Townsend et al., 2013), therefore, it is unlikely that small-scale ring containment activities would reach all or most dogs that may have come into contact with a rabid individual. It is not feasible to increase the radius of the preventative ring activities when the rabies control teams are frequently understaffed. In a door-to-door survey conducted in the same study area, 25% of owned dogs have unrestricted access to the streets (Castillo-Neyra et al., 2019). Under these conditions, implementing dog-centered small-scale activities that might include culling and injection-delivered vaccination becomes challenging due to difficulty of finding dogs at home and distinguishing between free-roaming owned dogs and strays.

Water channels and other similar structures that function as ecological corridors have implications for appropriate modelling of disease spread (Riley et al., 2015; Tigas et al., 2002; Vogt et al., 2007). Infected animals moving directionally along these corridors have the potential to connect parts of an urban landscape that are not geographically contiguous or even proximate. The increased connectivity created by the city landscape has implications on vaccination goals. It has been reported that small pockets of unvaccinated dogs can sustain rabies transmission (Townsend et al., 2013), and these ecological corridors have the potential to connect otherwise separated sub-optimally vaccinated populations increasing the chances of extending the area of the epidemic (Algeo et al., 2017; Wilschut et al., 2013). Importantly, the water channels also have the potential to harbor rabid dogs without human interference. Outside of the water channels, rabid dogs are killed or reported by their owners or community members when they start displaying symptoms or become disoriented and more likely to be a hit by a car (Hampson et al., 2009). Very few people go into the water channels (Castillo-Neyra et al., 2017), creating the potential for rabid dogs to survive longer and spread more virus than they otherwise would.

By tracking apparently healthy dogs, we have gained insight on the impact of the urban landscape of movement of owned free-roaming dogs, an important reservoir or rabies virus. It is known that rabies can change the movement patterns of dogs (Garg, 2014). However, specific dogs that have movement patterns that put them in contact with many other dogs have a greater risk of virus transmission. Future studies are needed to link ecological corridors and contact rates (Laager et al., 2018). Our study captured the movements of 23 dogs; a larger follow up study is needed to obtain reliable parameters to model rabies transmission in urban landscapes that favor long, directed and fast incursion of dogs or packs of dogs into new areas. The ecology of canine rabies works at different spatial scales and individual dog-movement is just one of them. To fully understand the rabies disease dynamics in Arequipa, further investigation should also include human mediated transportation of possibly rabid dogs from other communities into the city (Bourhy, 2016; Talbi et al., 2010; Brunker, 2012). It is possible that water channels are shared pathways and not part of the dogs’ core areas. Therefore, the water channels are not defended against intruders and allow more fluid movement of dogs, but with less aggressive behavior. However, in a few instances, our team observed segments of the water channels guarded by owned dogs from houses located on the sides of the water channels. Finally, our categorization of water channel usage includes inherent bias with increased water channel usage being related to increased home range. It is possible that a larger study with more followed up individuals would allow the use of synthetic likelihoods (Barbu et al. 2018) to tease apart the effect of water channels on these long, directed “flights”.

Interestingly, we found significant movement at night for over half the dogs tracked. These observations, contradict the paradigm that owned free roaming dogs have a more diurnal pattern while stray dogs exhibit a more nocturnal pattern (Beck, 1973), and advise caution when following guidelines that state that daytime counting methods to estimate the dog population are appropriate for owned dogs (MINSA, 2009). Even though we did not track stray dogs to compare against them, our data does not show any trends of owned dogs moving more during daytime, instead varying significantly by individual. Any strategies to control stray dog populations should not focus on dogs that are out at night (as have been suggested by local authorities) as these may be owned free-roaming dogs, not strays. In addition, assessment of the dog population size might be inaccurate if one of the assumptions of the methods is that most owned dogs are diurnal. In 2019, there were some reports from locals and sightings from our team of feral dogs in the outskirts of the city. The role of this free-roaming dog subpopulation in rabies transmission is uncharacterized and needs further investigation.

Creating a successful dog rabies control program and transforming it into a sustainable prevention program requires planning with knowledge of epidemiological concepts and deep understanding of local populations and local needs (Davlin and VonVille, 2012; Franka et al., 2013). Around the world, dog rabies control programs have demonstrated that regional canine rabies elimination is feasible (Kamoltham et al., 2003; Lembo et al., 2010; Schneider et al., 2011; Vigilato et al., 2013). Particularly, in Latin America the control programs have been very successful reducing the burden of disease significantly (Schneider et al., 2011; Vigilato et al., 2013). Our findings reinforce the importance of focusing on city-wide approaches with vaccination programs that reach both, optimal levels and even coverage across the city. Our study and previous studies of rabid dogs spatially associated with these water channels, suggest that surveillance activities in and around these structures, as well as vaccinating the animals that move along them (most feasibly with an oral vaccine) could be important to re-eliminating rabies virus from the city.

Conclusions

We found that water channels in Arequipa affect the way free-roaming dogs move in an urban landscape. These urban structures serve as ecological corridors facilitating dog’s long distance, linear movement across otherwise unconnected portions of the city. The effect of these corridors on rabies control is unknown, but likely makes control more difficult by harboring symptomatically rabid dogs and facilitating contact over longer distances. Implementing surveillance and control measures in and around water channels, and collecting systematic data on these activities, is critical to find new solutions for halting rabies virus transmission in the city.

Supplementary Material

1

Acknowledgements

Dr. Seth O’Neill and the Center for Global Health, Tumbes — UPCH for lending some of the GPS loggers.

Lina Mollenasa, Liz Leon, Edwin Cardenas, and Julio Meza for their hard work in the field handling and collaring the dogs.

OpenStreetMap contributors for data used to generate map figures. Map data copyrighted OpenStreetMap contributors and available from https://www.openstreetmap.org.

Funding

Funding for this study came from National Institutes of Health 1K01 Al139284 and from Penn Global Engagement Fund. The funders had no role in study design, data collection and analysis, decision to publish or preparation of manuscript.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Ethics

This study was reviewed and approved by the Institutional Animal Care and Use Committee of the Universidad Peruana Cayetano Heredia (approval number 67258).

Competing interests

We declare no competing interests.

References

  1. Acosta-Jamett G, Cleaveland S, Cunningham AA, Bronsvoort BM, 2010. Demography of domestic dogs in rural and urban areas of the coquimbo region of Chile and implications for disease transmission. Prev. Vet. Med 94, 272–281. 10.1016/j.prevetmed.2010.01.002. [DOI] [PubMed] [Google Scholar]
  2. Algeo TP, Slate D, Caron RM, Atwood T, Recuenco S, Ducey MJ, Chipman RB, Palace M, 2017. Modeling raccoon (Procyon lotor) habitat connectivity to identify potential corridors for rabies spread. Trop. Med. Infect. Dis 2, 44 10.3390/tropicalmed2030044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Barbu CM, Sethuraman K, Billig EMW, Levy MZ, 2018. Two-scale dispersal estimation for biological invasions via synthetic likelihood. Ecography. 41 10.1111/ecog.02575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Beck A, 1973. The ecology of stray dogs: A study of free-ranging urban animals. West Lafayette, IN: Purdue University Press. [Google Scholar]
  5. Belo VS, Struchiner CJ, Werneck GL, Teixeira Neto RG, Tonelli GB, de Carvalho Junior CG, Ribeiro RAN, de Silva ES, 2017. Abundance, survival, recruitment and effectiveness of sterilization of free-roaming dogs: A capture and recapture study in brazil. PLoS ONE. 12 e0187233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Burt WH, 1943. Territoriality and home range concepts as applied to mammals. J. of mammalogy 24, 346–352. https://www.jstor.org/stable/1374834?seq=1#metadata_info_tab_contents. [Google Scholar]
  7. Brunker K, Hampson K, Horton D, Biek R, 2012. Integrating the landscape epidemiology and genetics of RNA viruses: Rabies in domestic dogs as a model. Parasit. 139, 1899–1913. doi: 10.1017/S003118201200090X [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Castillo-Neyra R, Brown J, Borrini K, Arevalo C, Levy MZ, Buttenheim A, Hunter G, Becerra V, Behrman J, Paz-Soldan VA, 2017. Barriers to dog rabies vaccination during an urban rabies outbreak: Qualitative findings from Arequipa, Peru. PLoS NTD. 11 e0005460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Castillo-Neyra R, Toledo AM, Arevalo-Nieto C, MacDonald H, De l. P., Naquira-Velarde C, Paz-Soldan VA, Buttenheim AM, Levy MZ 2019. Socio-spatial heterogeneity in participation in mass dog rabies vaccination campaigns, Arequipa, Peru. PLOS NTD. 13 e0007600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Castillo-Neyra R, Zegarra E, Monroy Y, Bernedo FR, Cornejo-Rosello I, Paz-Soldan A, Levy MZ, 2017. Spatial association of canine rabies outbreak and ecological urban corridors, Arequipa, Peru. 2 10.3390/tropicalmed2030038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Centro Nacional de Epidemiología, Prevención y Control de Enfermedades, 2017. Boletín epidemiológico del perú SE 52–2017 No. 52). Lima, Perú: Ministerio de Salud. [Google Scholar]
  12. Centro Nacional de Epidemiología, Prevención y Control de Enfermedades, 2018. Boletín epidemiológico del perú SE 19–2018. Lima, Perú: [Google Scholar]
  13. Centro Nacional de Epidemiología, Prevención y Control de Enfermedades, 2019. Boletín epidemiológico del perú. Lima, Perú: Ministerio de Salud. [Google Scholar]
  14. Cortez-Aguirre G, Jiménez-Coello M, Gutiérrez-Blanco E, Ortega-Pacheco A, 2018. Stray dog population in a city of southern mexico and its impact on the contamination of public areas. Vet. Med. Int 2018, 1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Davlin SL, VonVille HM, 2012. Canine rabies vaccination and domestic dog population characteristics in the developing world: A systematic review. Vaccine, 30, 3492–3502. 10.1016/j.vaccine.2012.03.069 [DOI] [PubMed] [Google Scholar]
  16. Deen J, von Seidlein L, 2018. The case for ring vaccinations with special consideration of oral cholera vaccines. Hum. vaccines & immunotherapeutics 14, 2069–2074. 10.1080/21645515.2018.1462068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Denny MW, 2008. Limits to running speed in dogs, horses and humans. J. Exp. Biol 211, 3836–3849. 10.1242/jeb.024968. [DOI] [PubMed] [Google Scholar]
  18. Dickson BG, Beier P, 2007. Quantifying the influence of topographic position on cougar (Puma concolor) movement in southern California, USA. J. Zoo 271, 270–277. 10.1111/j.1469-7998.2006.00215.x [DOI] [Google Scholar]
  19. Dürr S, Ward MP, 2014. Roaming behaviour and home range estimation of domestic dogs in aboriginal and torres strait islander communities in northern australia using four different methods. Prev. Vet. Med 117, 340–357. 10.1016/j.prevetmed.2014.07.008. [DOI] [PubMed] [Google Scholar]
  20. Flores-Ibarra M, Estrella-Valenzuela G, 2004. Canine ecology and socioeconomic factors associated with dogs unvaccinated against rabies in a Mexican city across the US-Mexico border. Prev. Vet. Med 62, 79–87. 10.1016/j.prevetmed.2003.10.002. [DOI] [PubMed] [Google Scholar]
  21. Font E, 1987. Spacing and social organization: Urban stray dogs revisited. Appl. Anim. Behav. Sci 17, 319–328. 10.1016/0168-1591(87)90155-9 [DOI] [Google Scholar]
  22. Franka R, Smith TG, Dyer JL, Wu X, Niezgoda M, Rupprecht CE, 2013. Current and future tools for global canine rabies elimination. Antivir. Res 100, 220–225. 10.1016/j.antiviral.2013.07.004. [DOI] [PubMed] [Google Scholar]
  23. Garg SR, 2014. Rabies manifestations and diagnosis In Garg SR (Ed.), Rabies in man and animals (pp. 37–49). New Delhi: Springer India; 10.1007/978-81-322-1605-6_4 [DOI] [Google Scholar]
  24. Getz WM, Fortmann-Roe S, Cross PC, Lyons AJ, Ryan SJ, Wilmers CC, 2007. LoCoH: Nonparameteric kernel methods for constructing home ranges and utilization distributions. PLoS ONE, 2, e207 10.1371/journal.pone.0000207 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Guilloux A, Panachão L, Alves A, Zetun C, Cassenote A, Dias RA, 2018. Stray dogs in urban fragments: Relation between population’s perception of their presence and socio-demographic factors. Pesquisa Veterinária Brasileira. 38, 89–93. 10.1590/1678-5150-pvb-4293. [DOI] [Google Scholar]
  26. Hampson K, Dushoff J, Cleaveland S, Haydon D, Kaare M, Packer C, Dobson A, 2009. Transmission dynamics and prospects for the elimination of canine rabies. PLoS Biol. 7 10.1371/journal.pbio.1000053 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hudson EG, Brookes VJ, Dürr S, Ward MP, 2017. Domestic dog roaming patterns in remote northern australian indigenous communities and implications for disease modelling. Prev. Vet. Med 146, 52–60. 10.1016/j.prevetmed.2017.07.010. [DOI] [PubMed] [Google Scholar]
  28. Jackman J, Rowan A, 2007. Free-roaming dogs in developing countries: The benefits of capture, neuter, and return programs In Rowan A, Salem DJ (Eds.), The state of the animals (pp. 55–78). Washington, DC: Humane Society Press. [Google Scholar]
  29. Kahle D, Wickham H, 2015. Ggmap: Spatial visualization with ggplot2. R J. 5, 144–161. [Google Scholar]
  30. Kamoltham T, Singhsa J, Promsaranee U, Sonthon P, Mathean P, Thinyounyong W, 2003. Elimination of human rabies in a canine endemic province in thailand: Five-year programme. Bulletin of the WHO. 81(5), 375–381. [PMC free article] [PubMed] [Google Scholar]
  31. Laager M, Mbilo C, Madaye EA, Naminou A, Léchenne M, Tschopp A, Naïssengar SK, Smieszek T, Zinsstag J Chitnis N, 2018. The importance of dog population contact network structures in rabies transmission. PLoS NTD. 12 10.1371/journal.pntd.0006680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lembo T, Hampson K, Kaare MT, Ernest E, Knobel D, Kazwala RR, Haydon DT, Cleaveland S, 2010. The feasibility of canine rabies elimination in africa: Dispelling doubts with data. PLoS NTD, 4 10.1371/journal.pntd.0000626 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lyons AJ, Turner WC, Getz WM, 2013. Home range plus: A space-time characterization of movement over real landscapes. Mov. Ecol 1, 2 10.1186/2051-3933-1-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lyons A, 2014. T-LoCoH for R tutorial and users guide R-Forge. [Google Scholar]
  35. Meek P, 1999. The movement, roaming behavior and home range of free-roaming domestic dogs, canis lupus familiaris, in coastal new south Wales. Wildl. Res 26, 847–855. 10.1071/wr97101 [DOI] [Google Scholar]
  36. MINSA, Norma técnica de salud para la vigilancia, prevención y control de la rabia humana en el perú, 2017. Lima: Ministerio de Salud. [Google Scholar]
  37. OIE, 2009. Terrestrial animal health standards commission report. https://www.oie.int/standard-setting/terrestrial-code/
  38. OpenStreetMap Contributers, 2015. OpenStreetMapopenstreetmap.org. [Google Scholar]
  39. Perú Ministerio de Salud, 2015. Alertar ante el riesgo de presentación de casos de rabia humana y diseminación de rabia canina. Lima, Peru: Alerta Epidemiológica. [Google Scholar]
  40. Pray IW, Swanson DJ, Ayvar V, Muro C, Moyano LM, Gonzalez AE, Garcia HH, O’Neal SE, 2016. GPS tracking of free-ranging pigs to evaluate ring strategies for the control of cysticercosis/taeniasis in Peru. PLoS NTD, 10 10.1371/journal.pntd.0004591. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Prevención y Control de Enfermedades Centro Nacional de Epidemiología, 2016. Boletín epidemiológico del perú SE 52 No. 52). Lima, Perú: Ministerio de Salud. [Google Scholar]
  42. R Core Team, 2017. R: A language and environment for statistical computingR Foundation for Statistical Computing.Vienna, Austria. [Google Scholar]
  43. Rhodes CJ, Anderson RM, 1997. Epidemic thresholds and vaccination in a lattice model of disease spread. Theor. Pop. Biol 52 10.1006/tpbi.1997.1323. [DOI] [PubMed] [Google Scholar]
  44. Riley S, Eames K, Isham V, Mollison D, Trapman P, 2015. Five challenges for spatial epidemic models. Epid. 10, 68–71. 10.1016/j.epidem.2014.07.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rosatte R, Allan M, 2009. The ecology of red foxes, vulpes vulpes, in metropolitan toronto, ontario: Disease management implications. Can. Field-Nat 123 10.22621/cfn.v123i3.967 [DOI] [Google Scholar]
  46. Schneider MC, Aguilera XP, Barbosa da SJ, Ault SK, Najera P, Martinez J, Requejo R, Nicholls RS, Yadon Z, Silva JC, Leanes LF, Periago MR, 2011. Elimination of neglected diseases in latin america and the caribbean: A mapping of selected diseases. PLoS NTD. 5 10.1371/journal.pntd.0000964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Sepúlveda M, Pelican K, Cross P, Eguren A, Singer R, 2015. Fine-scale movements of rural free-ranging dogs in conservation areas in the temperate rainforest of the coastal range of southern Chile. J. Mam. Bio 80, 290–297. 10.1016/j.mambio.2015.03.001. [DOI] [Google Scholar]
  48. Slater MR, 2001. The role of veterinary epidemiology in the study of free-roaming dogs and cats. Prev. Vet. Med; Plenary Sessions of the International Society of Veterinary. 48, 273–286. [DOI] [PubMed] [Google Scholar]
  49. The Kennel Club, 2018. Dog measuring regulations. Retrieved Aug 6, 2018, from https://www.thekennelclub.org.uk/activities/agility/new-to-agility/agility-measuring/
  50. Talbi C, Lemey P, Suchard MA, Abdelatif E, Elharrak M, Jalal N, Faouzi A, Echevarría JE, Morón SV, Rambaut A, Campiz N, Tatem AJ, Holmes EC, Bourhy H, 2010. Phylodynamics and human-mediated dispersal of a zoonotic virus. PLOS path. 6, 10.1371/journal.ppat.1001166 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Tigas LA, Van Vuren DH, Sauvajot RM, 2002. Behavioral responses of bobcats and coyotes to habitat fragmentation and corridors in an urban environment. 108 10.1016/S0006-3207(02)00120-9. [DOI] [Google Scholar]
  52. Townsend SE, Lembo T, Cleaveland S, Meslin FX, Miranda ME, Putra AA Haydon DT, Hampson K, 2013. Surveillance guidelines for disease elimination: A case study of canine rabies. Comp. Immunol. Micro. Infect. Dis 36, 249–261. 10.1016/j.cimid.2012.10.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Townsend SE, Sumantra IP, Pudjiatmoko, Bagus GN, Brum E, Cleaveland S, Crafter S, Dewi APM, Dharma DMN, Dushoff J, Girardi J, Gunata IK, Hiby EF, Kalalo C, Knobel DL, Mardiana IW, Putra AAG, Schoonman L, Scott-Orr H, Shand M, Sukanadi IW, Suseno PP, Haydon DT, Hampson K, 2013. Designing programs for eliminating canine rabies from islands: Bali, indonesia as a case study. PLoS NTD. 7 10.1371/journal.pntd.0002372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. van Bommel L, Johnson CN, 2014. Where do livestock guardian dogs go? Movement patterns of free-ranging maremma sheepdogs. PLoS ONE, 9, e111444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Vazquez-Prokopec G, Stoddard ST, Paz-Soldan V, Morrison AC, Elder JP, Kochel TJ, Scott TW, Kitron U, 2009. Usefulness of commercially available GPS data-loggers for tracking human movement and exposure to dengue virus. Int. J. Health Geogr 8, 68 10.1186/1476-072X-8-68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Vigilato MAN, Clavijo A, Knobl T, Silva HMT, Cosivi O, Schneider MC, Leanes LF, Belotto AJ, Espinal MA, 2013. Progress towards eliminating canine rabies: Policies and perspectives from Latin America and the Caribbean. Philos. Trans. R. Soc. B: Biol. Sci 368 10.1098/rstb.2012.0143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Vogt P, Riitters K, Iwanowski M, Estreguil C, Kozak J, Soille P, 2007. Mapping landscape corridors. Ecol. Indications 7, 481–488. [Google Scholar]
  58. WHO, Department of control of neglected tropical diseases, 2018. WHO expert consultation on rabies: Third report. Geneva: World Health Organization. [Google Scholar]
  59. Wilschut LI, Addink EA, Heesterbeek H, Heier L, Laudisoit A, Begon M, Davis S, Dubyanskiy VM, Burdelov LA, de Jong SM, 2013. Potential corridors and barriers for plague spread in Central Asia. Int. J. Health Geogr 12, 49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zehnder A, Henley S, Weibel R, 2018. Home ranges of lions in the Kalahari, Botswana exhibit vast sizes and high temporal variability. Zoo, 128, 46–54. 10.1016/j.zool.2018.04.001. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1586013-supplement-1.docx (15.8MB, docx)

RESOURCES