Don’t go chasing waterfalls: the phenology of Amblyomma americanum and increased collection of medically relevant ticks further away from hiking trails in the Piedmont of North Carolina, USA

Abstract

Hiking is a popular recreational activity in North Carolina that may expose people to ticks and tick-borne pathogens. However, there is a lack of knowledge on how ticks are distributed on and near trails. Our study tested the hypothesis that ticks are more likely to be encountered when moving further away from trails by measuring differences in relative abundance at various distances. We tested 4 distances (middle of trail, edge of trail, 5 m, 20 m), as well as an on-trail and off-trail grouping. We collected significantly more ticks and ticks were more likely to be collected at our 20 m and 5 m sampling distances rather than directly on or adjacent to trails, and significantly more were collected during off-trail collections than on-trails. When looking only at Amblyomma americanum, post hoc comparisons revealed that significantly more juvenile stages were collected at 5 m and 20 m distances, but not for adults. Our monthly sampling also allowed us to describe the phenology of A. americanum in North Carolina, which is consistent with the phenology of this species in the southeastern United States with adults peaking May–Jun, nymphs Jun–Jul, and larvae in Jul–Aug. These results generally demonstrate the importance of utilizing established trails when hiking to decrease tick-borne disease risk and should be communicated to the public as a recommendation for reducing tick-encounter risk.

Introduction

Tick-borne diseases (TBD) in the United States are an increasing public health threat, particularly in the eastern half of the United States where TBD are more common (Diuk-Wasser et al. 2012). However, counts of TBD are only reported, confirmed cases and there is likely severe underreporting of TBD in the United States (CDC 2018, Schiffman et al. 2018, Madison-Antenucci et al. 2020). Increases in TBD are largely due to changes in landscapes across time, as well as shifts in the global climate (Deshpande et al. 2024). Most TBD transmission is focused in the northeastern and midwestern parts of the United States (Spach et al. 1993), but North Carolina (NC) has also recorded a marked rise in TBD cases (Byrd et al. 2020, NCDHHS 2023). The most frequently encountered tick in NC depends on the region of the state, with Amblyomma americanum (Linnaeus) being the most common tick species in the Piedmont of NC (Apperson et al. 1990, Smith et al. 2010). However, only the Piedmont of NC has been systematically surveyed for ticks as it has the highest population density (Smith et al. 2010, United States Census Bureau n.d). Pathogens transmitted by A. americanum include Ehrlichia spp., Rickettsia spp., Heartland virus, and Bourbon virus, as well as serving as the cause of alpha-gal syndrome in humans (formerly “red meat allergy”) (Childs and Paddock 2003, Crispell et al. 2019, Commins 2020). Increasing evidence of Lyme disease transmission in the mountains (western NC) and coastal plains (eastern NC) suggests that tick bites from Ixodes scapularis (Say) are becoming more common in these parts of the state, which historically has not been the cases due to differences in southern populations in I. scapularis host-seeking behavior (Ginsberg et al. 2021, NCDHHS 2023). The increasing incidence of these diseases demonstrates a need for more research on the biology and ecology of ticks and TBD in NC.

There are many factors that can contribute to the risk of acquiring a tick-borne disease, such as locality and behavior. Understanding which risk factors contribute to TBD exposures can allow us to make clear public health recommendations to the public to reduce TBD exposure risk. People who work or spend time outdoors (e.g., recreational hiking) are likely at a higher risk of encountering ticks and their associated pathogens (Wilson et al. 2023). Recreational hiking often occurs in wooded or forested areas that are preferred habitat for white-tailed deer (Odocoileus virginianus), and subsequently their associated tick ectoparasites (McShea 2012). Despite interest in connecting recreational use of forests and tick exposure in other parts of the United States, studies investigating ticks on recreational spaces and hiking trails in the southeastern United States. are limited (Bhosale et al. 2023). Even fewer of these studies have occurred in NC, which has an abundance of wooded, forested, and mountainous regions and a continuously increasing population of people, potentially bringing more people into contact with ticks through recreation (US Census Bureau 2024). This gap in knowledge further emphasizes the need to understand how the ticks of NC and their hosts utilize hiking trails.

Currently, there are few control methods for tick populations which makes it challenging to reduce TBD risk in nature. Some studies have evaluated various “tick traps”, but none have been widely adopted to-date (Brown et al. 2020, Yans et al. 2022). There are currently no vaccines for TBD approved for use in humans in the United States, and tick-bite prevention remains the best method of preventing TBD transmission (Piesman and Eisen 2008, Eisen 2021, Abbas et al. 2023). Potential prevention measures include appropriate clothing and arthropod repellents (e.g. DEET), as well as changes to human behavior that reduce the risk of TBD exposure (e.g. post-activity tick checks). Therefore, additional research should be done on risk factors that contribute to tick exposure and behaviors that can be implemented to reduce TBD risk and subsequently communicated to the public.

In this study, we utilized ongoing active surveillance data to test whether we were more likely to encounter ticks on or off the trails to make general recommendations about methods of prevention of human exposure to ticks in NC recreational areas. We predicted that we would both encounter more ticks and have increased odds of encountering any tick at further distances from the middle of the trails. The results of this study may inform public health recommendations to reduce the risk of recreational hiking exposure to ticks and their associated pathogens.

Materials and Methods

Study Sites

We sampled multiple counties in the Piedmont of North Carolina as part of an ongoing active surveillance project. We selected sites that were wooded coniferous, deciduous, or mixed forest that had established recreational trails that were accessible for tick dragging. All trails were accessible by foot, and trails paved with asphalt or concrete were excluded. We chose sections that avoided obstacles that would interrupt area coverage such as lakes, rivers, ravines, heavy underbrush, and intersections with roads and other trails. All sites were accessible for drag sampling and appropriate permissions and permits were obtained. This yielded a total of 15 sites in 7 counties in North Carolina: Montgomery, Johnston, Chatham, Orange, Durham, Wake, and Vance with a total of 25 trail sections sampled (Fig. 1, Supplemental Table S1). Trail sections were sampling areas within a site and some sites had more than one trail section. Sites with more than one trail section are reported in Supplemental Table S1. Trail sections were standardized, and the same amount of effort was conducted for each sampling distance for each date (125 m²). We were only able to conduct routine monthly sampling throughout the duration of our study (Oct 2020–2022) for 16 of these trail sections, and the remainder were temporarily sampled (Fig. 1, Supplementary Table S1). The trail width of sites varied from 0.75 m to 2 m wide.

Tick Sampling Effort

We used cloth dragging to collect ticks from the environment as an effective and standard method of tick sampling (Mays et al. 2016, CDC 2020). Drag cloths were made of 1 m² flannel, cotton cloths attached to a wooden dowel of 3.18 cm diameter with two-1 m sections of rope attached to a 0.3 m PVC pipe as a handle, following CDC guidelines (CDC 2020). Each dragging event consisted of four 125 m transects that were parallel to the trail at 4 distances: middle of trail, edge of trail, 5 m off trail, and 20 m off trail (hereafter “sampling distances”). When sampling trail edges, we pulled one edge of the drag along the natural edge of the trail, and not overlapping onto the trail itself. Because of their proximity to each other, we did not sample middle and edge sections at the same spot along the trail to avoid disrupting the surrounding environment. Instead, we sampled the middle of the trail first for 125 m, then the trail edge for the next 125 m. We inspected the front and back of drag cloths for any tick life stages every 25 m, and we removed nymphal and adult ticks and placed them directly into microcentrifuge tubes with 95% ethanol. We manually removed larval ticks with forceps if few were collected and placed them in microcentrifuge tubes with 95% ethanol, or removed them from drag cloths with a lint roller and stored the sheets at −20°C. All ticks we collected were separated by site, trail section, date, and species. In addition, we collected ticks crawling on the collector, although did not record this separately from ticks on the drag cloths.

Tick Processing

Ticks we collected were returned to the laboratory for identification (species, life stage, and sex) using morphological keys (Clifford et al. 1961, Keirans and Litwak 1989, Durden and Keirans 1996, Dubie et al. 2017, Egizi et al. 2019, Nadolny et al. 2021). Identifications deemed uncertain by our primary identifier were identified again by a secondary observer. Ixodes spp. larvae were only identified as genus. All ticks were stored at −20°C for future molecular screening of pathogens. Results of pathogen screening will be reported in a later publication.

Statistical Analyses

We conducted all data manipulation and analyses using R statistical software v4.3.0 (R Foundation for Statistical Computing, Vienna, Austria). To visualize the phenology of ticks, we only used data from between Oct 2020 and Oct 2022 from 16 trail sections across 4 counties (Chatham (n = 8), Orange (n = 3), Durham (n = 4), Vance (n = 1)) for which we had complete, monthly sampling. Figures for the phenology of D. variabilis, I. scapularis (nymphs and adults), and Ixodes spp. larvae are provided in the supplementary materials (Supplementary Figs. 1 and 2).

To evaluate our hypotheses that distance from established trails influences tick counts and encounter probability, we utilized generalized linear mixed models (GLMM). We analyzed tick counts using a negative binomial GLMM (log link function, hereafter “negative binomial model”, glmmTMB package)(Brooks et al. 2017) and encounter probability (presence/absence) using a logistic GLMM (logit link function, hereafter “logistic model”, lme4 package) (Bates et al. 2015). We used collection year, life stage, species, and sampling distance (i.e., middle of trail, edge of trail, 5 m from trail, and 20 m from trail) as explanatory variables in our models. Collection site and transect nested within collection site were the tested random variables for our models. We utilized AIC to determine the best fit GLMM, and selected final models with the lowest AIC. We repeated the analyses for all ticks (Supplementary Table S2), and then models specifically for A. americanum (Supplementary Table S3), D. variabilis (Supplementary Table S4), and Ixodes spp. (Supplementary Table S5). In addition, we tested models that included on-trail (middle and edge, combined) and off-trail (5 m and 20m, combined) as variables predicting that the off-trail variable would have significantly more and a higher odds of encountering ticks due to distance from the trail (Supplementary Results, Supplementary Tables S2–6).

In addition to our GLMM analyses, we utilized a post hoc analysis of tick count data to identify what effect sampling distance had on each life stage for each species. The first was the Kruskal–Wallis rank sum test to identify if there was a significant difference among sampling distance groups for each life stage. If significant, we followed this test with the Wilcoxon rank-sum test (Holm–Bonferroni adjustment for multiple pairwise comparisons) to identify which sampling distances were significantly different from each other for each life stage. These analyses were completed for A. americanum, D. variabilis adults (due to sample size constraints), and Ixodes spp., however we are presenting only A. americanum data as they represented the majority of ticks collected in our study, with other species having sample sizes too small for meaningful comparisons with this analysis. However, we do provide figures for D. variabilis and Ixodes spp. in the supplementary materials (Supplementary Fig. 3 and 4).

Results

Tick Collections

In total, we collected 23,567 ticks from our 25 trail sections in the Piedmont of NC from Jun 2020 to Oct 2022 reflecting 474 unique collection events and 262,975 m² dragging (Table 1). Species represented in these collections included: A. americanum Linnaeus, Amblyomma maculatum Koch, Dermacentor variabilis Say, Ixodes scapularis Say, Ixodes affinis Neumann, Ixodes brunneus Koch, Haemaphysalis leporispalustris Packard, Haemaphysalis juxtakochi Cooley, and Rhipicephalus sanguineus Latreille. We collected A. americanum most frequently, representing 86.5% of all collected ticks (20,390 A. americanum, 3,177 other species; Table 1). We collected fewer ticks of other species with 594 Ixodes spp. (including 10 adults, 34 nymphal I. scapularis and many larvae we could not positively identify to species), 80 D. variabilis (66 adults, 14 larvae), 8 Haemaphysalis spp., and 2 larval R. sanguineus.

Amblyomma americanum Phenology

Due to the low collection frequency of all other species, we were only able to robustly demonstrate the phenology of A. americanum ticks at our sites. Adult A. americanum generally began to appear in the early spring (March) but could be collected as early as February (Fig. 2). Adults peaked in host-seeking activity in the months of May and Jun and began to decrease until Aug (occasionally being collected in September). Nymphal A. americanum followed a similar pattern as adults but were delayed by approximately one month (Fig. 2). Nymphs appeared in March, peaked between May and Jul, and decreased activity into the fall, but were collected as late as Oct and Nov. Larval phenology peaked in the late summer (Jul-Aug) (Fig. 2) and larval ticks were collected as late as Nov in some years. In 2022, larval A. americanum were collected in Apr, which is outside of their expected phenology in NC.

Sampling Distance Analysis

All species.

Results from all negative binomial and logistic models can be found in Table 2. When considering all collected ticks (i.e., total ticks regardless of species), our best-fit negative binomial model included both sampling distance and life stage with a significant interaction between them (Supplementary Table S2), which means that the counts of ticks at different trail distances were different between life-stages. Our variable comparisons revealed that significantly more larvae were collected at all distances when compared to adults in the middle of trails, but nymphs did not statistically differ at any distance. The collection year was not a significant predictor of the number of ticks collected and was removed from our best fit negative binomial model. Our logistic model for all collected ticks was fit with all variables including collection year and were included in the final model (Supplementary Table S2), while the interaction did not significantly improve fit and was not included. Specific comparisons demonstrated significantly higher odds of encountering ticks at 5 m and 20 m distances when compared to the middle of the trail, but not for the edge of trail distance. Additionally, ticks were significantly less likely to be collected in the first year when compared to years 2 and 3.

Amblyomma americanum.

Our best fit A. americanum negative binomial model also included sampling distance and life stage as explanatory variables, with an interaction between them (Supplementary Table S3). Larval and nymph tick counts at all distances were significantly greater than the middle of the trail, except for nymphs on the edge of the trail. Our best fit A. americanum-only logistic model was similar to the logistic model of all collected ticks (Supplementary Table S3), and we observed significantly greater odds of collecting ticks at 5 m and 20 m distances when compared to the middle of the trail, but not for the edge of trail. Collection year 1 was also significantly less likely to collect ticks than year 3, and nymphs were significantly more likely to be collected than adults. Year 2 and larval collection odds were not significantly different from the reference categories.

Dermacentor variabilis.

Our best fit D. variabilis negative binomial model included sampling distance, collection year, and life stage with no interactions (Supplementary Table S4). We collected significantly more ticks at all other sampling distances when compared to the middle of the trail (Table 2). Our year 2 collections were also significantly more than year 1, and we collected significantly fewer larval D. variabilis than adults. Our year 3 collections were not statistically different from the reference category, and because we did not collect any nymphs, we could not compare this life stage in either of our negative binomial and logistic models. Our best fit logistic model for D. variabilis did not include year but included all other variables with no interactions (Supplementary Table S4). This species had higher odds of being collected at all distances when compared to the middle of the trail, however only edge of trail and 5 m were statistically significant, and larval ticks had a significantly lower odds of collection than adults.

Ixodes spp.

Our best fit Ixodes spp. negative binomial model (Supplementary Table S5) was similar to the negative binomial model for D. variabilis. We collected significantly more ticks at all other sampling distances when compared to the middle of trail collections. We collected significantly more ticks in year 2 than in year 1, and collected significantly more larvae than adults. Our year 3 and nymph collections were not significantly different from the reference categories. Our best fit Ixodes logistic model (Supplementary Table S5) was also the same as the D. variabilis logistic model, however, our variable comparisons demonstrated the opposite result as our D. variabilis model with significantly higher odds of collecting ticks at the 20 m distance when compared to the middle of the trail, but no significant difference when compared to edge of trail and 5 m collections. Like our Ixodes spp. negative binomial model, year 2 had significantly higher odds to collect ticks than year 1, but year 3 was not significantly different from year 1. Additionally, larvae had significantly lower odds of collection than adults, but nymphs did not significantly differ from adults.

Life Stage Post hoc Analysis

The mean number of ticks we collected at each sampling distance for each life stage is presented in Table 3 and all pairwise comparisons are presented as lettering in Fig. 3. For adult A. americanum, we observed a significant difference in tick numbers among sampling distances (Kruskal–Wallis, df = 3, x² = 10.025, P = 0.018), and when we followed with the post hoc analysis, the 20 m sampling distance was significantly higher than the edge of trail (Wilcoxon Rank Sum, P = 0.009) and all other sampling distances did not different significantly. For nymphal A. americanum, we observed a significant difference among sampling distances (Kruskal–Wallis, df = 3, x² = 14.201, P = 0.003). Both edge (Wilcoxon Rank Sum, P = 0.006) and middle of trail (Wilcoxon Rank Sum, P = 0.013) sampling distances collected significantly fewer ticks than the 20 m sampling distance, and the 5 m sampling distance collections were greater than edge of trail, but this difference was marginally insignificant (Wilcoxon Rank Sum, P = 0.062). When we considered only larval A. americanum, we observed a significant difference among sampling distances (Kruskal–Wallis, df = 3, x² = 18.552, P < 0.001), and the 20 m sampling distance was significantly greater than all other sampling distances (middle, P < 0.001; edge, P = 0.029; 5 m, P = 0.035).

Discussion

Overall, our findings support our hypothesis that by moving further away from established recreational trails, the risk of tick exposure increases. For both analysis types (i.e., negative binomial and logistic), we observed a seemingly linear trend in the number of ticks collected or odds of collection, with 20 m sampling distances demonstrating the highest risk. Our findings were consistent across most species that we tested except for the probability of encountering D. variabilis adults, which demonstrated no significant difference in odds of collection between the middle of trail and 20 m sampling distances. However, the odds of collection at 20 m were still more than three times as likely as in the middle of trail. These results demonstrate the necessity to take precautions against tick bites in the broad sense given that these patterns were consistent across all species collected. Our post hoc life-stage analysis for A. americanum generally supported the results of our models, and each observed life stage demonstrated a different pattern in relation to sampling distance. Interestingly, adults were most frequently collected on the middle of trails and at 20 m, but less frequently on the edge of trail and 5 m sampling distances. This contrasts with our hypothesis that more ticks should be collected at further distances and should be studied further to fully understand the distribution of adult A. americanum on recreational trails. A possible explanation is that adults are more desiccation resistant and can more aggressively host-seek on trails, which smaller life stages cannot (Leal et al. 2020). Prior studies have demonstrated seasonal differences in questing activity of A. americanum (Mangan et al. 2022), but few have evaluated differences between host-seeking behavior of A. americanum life stages. Alternatively, they may have simply fallen off a host on the trail and were able to survive due to their desiccation resistance. Nymphs of this species demonstrated a distribution at our sampling distances that we predicted, with 5 m and 20 m distances (i.e., off-trail) collecting more than our middle and edge of trail distances (i.e., on-trail). The distribution of larvae at our sampling distances significantly favored the 20 m sampling distance, with all other distances collecting fewer numbers of ticks. This could be that larvae are more prone to desiccation and require the relative humidity of distances further from the trail (Leal et al. 2020) or that adults preferentially select oviposition microhabitats that are further from trails, for which we have found no literature. Overall, we observed differences among life stages of the same species, and future studies should evaluate the life stage-specific distribution patterns that we observed.

In this study, A. americanum was the only species collected in sufficient numbers to be able to describe the phenology of this species in North Carolina. The phenology we observed was consistent with prior observations in the southeastern United States (Morris et al. 2022). Adults were most abundant during the late spring and early summer (May and Jun), and gradually began to decline. However, our study collected adults as early as February. Continued, long-term surveillance should be completed to identify if A. americanum is regularly appearing earlier in the year, which would suggest that a change in climate is shifting the phenology of this species. A warmer climate may extend the host-seeking period of this tick species and increases the window for pathogen transmission by this tick species, which is a significant public health risk (Gilbert et al. 2014, Ogden et al. 2021). Nymphal A. americanum generally follow the same trend as adults except their populations peak about a month later than adults and decline into the early fall. Larval A. americanum peaked in the hottest months of Jul and Aug and declined into the fall. These results, combined with tick distributions at distances from trails, allow us to make clear and accurate recommendations to people who enjoy recreational hiking so that they may mitigate the risk of encountering a tick and TBD. For example, because A. americanum is the most common tick in the Piedmont North Carolina (Smith et al. 2010), we could advocate taking extra precautions during the peak phenological season for this species (i.e., late spring to summer).

In our study, we did not measure trail usage by humans or other potential mammal hosts, so we cannot directly connect tick presence or abundance to either host distribution or quantify human risk directly. Nevertheless, human and mammal use of trails may help explain the patterns we observed, and we suggest collecting mammal and human trail use data in the future. Mammals may potentially avoid (or intentionally utilize) trails created by humans, and the distribution of ticks from these hosts may reflect that. For example, Mols et al. found that deer droppings and host-seeking Ixodes ricinus were significantly less abundant closer to recreational trails, demonstrating that the deer carrying these ticks prefer to avoid areas frequented by humans (Mols et al. 2022). However, not all mammal hosts may respond the same to recreational trails and human presence. Kays et al. (2016) found that predators (e.g., coyotes, Canis latrans) frequently selected for human-made trails, many of which may serve as hosts for immature stages of ticks (Kays et al., 2016). Studies combining human and mammal behavior with the density of ticks will provide a better idea of why ticks are not as abundant directly and adjacent to trails.

We observed significant increases in risk of tick exposure as we move away from trails, however, it is not clear what factors are contributing to these increases. It is known that relative humidity of microhabitats is important for the survival of ticks with all life stages burrowing into the soil to avoid desiccation, so a lack of moisture on trails could contribute to the lack of ticks on trails (Burtis et al. 2019). Additionally, the presence of abundant leaf litter at further distances may provide additional ideal habitat for ticks. It is not clear how our edge and 20 m sampling distances differ though, given that both distances contain leaf litter. Our study did not measure the relative humidity of leaf litter at our distances, so future studies should measure this to better understand why ticks exhibit these patterns of distribution.

While our study design was robust with 2 complete years of phenology across multiple sampling sites in the Piedmont of NC, other regions (i.e., Coastal Plains and Mountains) may have differing tick species compositions. The mountainous region of NC has an elevated incidence of Lyme disease (NCDHHS 2023), which suggests that I. scapularis is more common in this part of the state and differs from our current study region where we collected only 44 adults and nymphs of this species across 2 years of sampling. Tick species other than A. americanum may exhibit differing distributions on trails in different regions of NC, but our data suggest that these patterns are similar for D. variabilis and Ixodes spp., and similar recommendations could be applied to reduce exposure risk. We also observed variation in tick abundance in our multiple years of sampling, however, we could not draw conclusions from only 2 years of sampling. Multi-year studies are needed to assess how tick populations change across time in the Piedmont of NC, especially given that northern populations of I. scapularis may be expanding into NC and transmitting The Lyme disease pathogen to residents (Eisen and Eisen 2023).

The overall significant differences in tick distribution with respect to trails suggests general recommendations to citizens that enjoy recreational hiking: stay on established paths to avoid tick-exposures, especially from May to Aug. Given the increase in TBD in North Carolina, as well as increased human population density, these recommendations should be publicized through informational campaigns, (NCDHHS 2023, United States Census Bureau n.d).

Funding

Surveillance of tick populations was supported by a grant from the North Carolina Department of Health and Human Services (Award Number: 00044671 “Tick Surveillance in Eastern North Carolina) through the Centers for Disease Control Enhancing Laboratory Capacity Funds. This work is also supported by the United States Department of Agriculture Multistate Project NE-1943, “Biology, Ecology & Management of Emerging Disease Vectors.”

Acknowledgments

We would like to thank multiple undergraduate workers for their assistance throughout the sampling period including Carly Ward, Aiden Winters, Zach Benfield, Kiran Khan, Amy Dinh, Chloe Jonas, Farouk Osman, and Shea Phillips. The authors declare that they have no conflicts of interest to disclose.

Author Contributions

Dayvion Adams (Data curation [Supporting], Formal analysis [Lead], Writing—original draft [Lead], Writing—review & editing [Equal]), Anastasia Figurskey (Data curation [Lead], Formal analysis [Supporting], Investigation [Equal], Methodology [Lead], Supervision [Equal], Writing—review & editing [Equal]), Alexis Barbarin (Conceptualization [Equal], Funding acquisition [Equal], Writing—review & editing [Equal]), and Michael Reiskind (Conceptualization [Lead], Data curation [Equal], Formal analysis [Supporting], Funding acquisition [Lead], Investigation [Lead], Methodology [Equal], Project administration [Lead], Supervision [Equal], Writing—review & editing [Equal])

References

Author notes