Abundance of ticks and tick-borne pathogens in domestic gardens in Belgium, 2020-2022: a citizen science approach

Abstract

Background

Ticks, particularly Ixodes ricinus, are primary vectors for tick-borne diseases in Europe, with private gardens representing an understudied but potentially significant risk habitat. Through a citizen science initiative, we aimed to investigate tick density and pathogen prevalence in domestic gardens across Flanders, Belgium.

Methods

A total of 185 citizen scientists participated in standardized tick dragging and collecting garden data over multiple occasions from 2020 to 2022. Generalized linear mixed effects models were used to analyze tick density and pathogen prevalence.

Results

Ticks were detected in 44.3% of 185 gardens. They were most frequently found in rural gardens (60.2% of rural gardens, 50/83), but also in suburban (28.9%, 11/38) and urban gardens (50.0%, 4/8). Ixodes ricinus was the dominant species (94.7% out of 1162 ticks), with all life stages present. Additionally, few individuals of Ixodes frontalis (5.0%) and Ixodes hexagonus (0.3%) were collected. Ticks were found in a variety of vegetation types, however the highest proportions of successful collections were in fallen leaves (47.4%) and fallow land (46.2%). Nymphal density was positively associated with the number of mammal species observed and an association with vegetation type was found: significantly more nymphs were collected on wild vegetation (with fallow land) and fallen leaves, compared to mown grass, tall grass and flower beds (with vegetable gardens). Pathogen screening revealed that 34.1% of ticks carried at least one pathogen, with Borrelia burgdorferi s.l. most prevalent (19.6%), followed by Rickettsia spp. (11.7%). Co-infections were observed in 6.6% of ticks. Borrelia burgdorferi s.l. prevalence was significantly lower in the presence of dogs and increased with nymphal density. Additionally, a significant interaction was found between life stage and the number of bird species.

Conclusions

The presence of Ixodes ricinus in gardens, coupled with the detection of pathogens at prevalence levels similar to those in natural habitats, suggests that private gardens may serve as overlooked risk areas for tick exposure. This underscores that raising public awareness, combined with effective garden management strategies, may help mitigate the risk of infection with tick-borne diseases. Future research could focus on evaluating the effectiveness of prevention and garden management measures.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-23221-1.

Introduction

Hard ticks (Ixodidae) play a crucial role as vectors of infectious diseases that impact human health. In Europe, Lyme borreliosis is the most prevalent tick-borne disease, caused by a number of genospecies from the bacteria Borrelia burgdorferi (sensu lato) complex [1]. The systematic review of Burn et al. 2023 showed Belgium is among the countries with the highest incidence of Lyme borreliosis in Europe, currently estimated at 103 cases per 100,000 residents. This review was based on studies that reported cases of erythema migrans and/or consultations for tick bites [2]. In addition to Borrelia burgdorferi, ticks can transmit other pathogens such as tick-borne encephalitis virus, Anaplasma phagocytophilum, Borrelia miyamotoi, Neoehrlichia mikurensis, Rickettsia spp., and several Babesia spp., which have also been shown to be pathogenic in humans [3]. While human infections with the latter pathogens are rare in Europe [4], considerable underdiagnosis is suspected due to the difficulties in diagnosis and a lack of awareness among physicians. Most of the diseases caused by these pathogens present with mild and non-characteristic symptoms [5], making it challenging to assess their public health risk and burden accurately. Therefore, gathering information on the geographical tick distribution and the prevalence of tick-borne pathogens in local tick populations remains a crucial step in the reduction of disease burden.

The primary European tick vector is Ixodes ricinus [1], commonly found on the litter and lower layers of forest vegetation. Large tick populations are found when a sufficiently abundant host community is present, consisting of a variety of vertebrate host types differing in body size. Although I. ricinus is a generalist parasite, found feeding on over 300 different vertebrate species [6], each parasitic life stages has its own host preference. The tick’s life cycle includes three active parasitic stages: larva, nymph, and adult female. In contrast, the adult male is non-parasitic. All these stages utilize a sit-and-wait host-finding strategy (known as'questing’), from a vantage point, allowing them to grab onto an animal as it passes by [7]. Larger mammals can host all three parasitic stages—larva, nymph, and adult female—whereas only the immature stages are typically found on smaller mammals and songbirds. Consequently, multiple life stages may feed on the same host species, although they may target different attachment sites on the body of the host [8, 9]. Once attached to the host’s skin, they feed non-stop for a few days before detaching. Subsequently, the engorged larvae and nymphs moult into the next developmental stage (nymphs and adults, respectively). Adult females that succeed to engorge and copulate with males on the mammal (i.e. tick reproduction hosts), will lay approximately 2,000 eggs off-host, from which larvae emerge [10, 11]. Host-searching, moulting and egg survival heavily depend on suitable abiotic conditions, mainly defined by sufficiently high temperatures as well as humidity levels that prevents them from desiccation. Thus, the complex life cycle makes I. ricinus dependent on the combination of habitat structure, abiotic conditions, and the availability of host animals.

Across many European countries, current land use and wildlife management practices appear to contribute to biotic and abiotic conditions that support persistent tick populations, along with high incidences of tick-borne pathogen infections in urban and peri-urban settings [12]. Urban environments are characterized by severe habitat fragmentation where movement of wildlife can be strongly influenced, but sometimes also facilitated, by human infrastructures such as buildings, roads and canals. Many wild vertebrates commonly found in urban and peri-urban areas are species that easily adapt to ecosystems with high human population densities, and many have the potential to serve as tick hosts and pathogen reservoirs [13, 14]. Suitable hosts for ticks, and/or human pathogens found in European (peri-) urban settings include rodents (mice, voles, dormice, squirrels, and rats), shrews, birds, lizards, along with medium-sized and larger mammals like hedgehogs, foxes, roe deer, and wild boars [15]. Also companion animals (dogs and cats) are known tick reproducing hosts and may amplify certain tick-borne pathogens [16, 17].

Ticks typically acquire Lyme disease-causing spirochetes through horizontal host-to-tick transmission, and to a far lesser extent through horizontal tick-to-tick transmission when individuals simultaneously feed close to each other on the same host (also called ‘co-feeding transmission’) [18]. Host species vary in their ability to support different tick life stages and in their capacity to transmit various pathogens, including Borrelia burgdorferi s.l. genospecies [19, 20]. In Europe, at least six different genospecies are considered pathogenic to humans: B. afzelii, B. garinii, B. burgdorferi (sensu stricto), B. bavariensis, B. valaisiana and B. spielmanii [21]. Each of the abovementioned genospecies is linked to certain host types due to differences in host serum sensitivity [22]. For example, B. garinii typically proliferates in avian reservoir hosts, while B. afzelii is associated with mammalian reservoirs hosts [23]. Certain host species, like deer, are incompetent hosts for all Borrelia genospecies [24, 25]. Also in other tick-borne pathogens, host specificity has been observed among genetically contrasting strains. For example, Anaplasma phagocytophilum, which is a rickettsia-like bacterium that can cause granulocytic anaplasmosis in humans, livestock and companion animals [26], consists of four ecotypes that are each associated with different host species. Cattle, red deer, hedgehogs and horses are the main reservoirs of the zoonotic ecotype 1, while roe deer are associated with ecotype 2, rodents with ecotype 3 and birds with ecotype 4 [27–29]. Borrelia miyamotoi is a member of the relapsing fever group of Borrelia spirochaetes that can be hosted by rodents [30, 31]. In contrast to species of the Borrelia burgdorferi s. l. complex, it may also be transmitted vertically through transovarial transmission [32, 33]. Rickettsia helvetica belongs to the spotted fever group, potentially causing cardiac and neurological problems in humans [34, 35]. The rickettsia-like bacterium Neoehrlichia mikurensis is associated with febrile patients [36] and has been found in tissues of wild rodents [30, 37, 38]. Thus, local tick abundances and pathogen prevalence in questing ticks depend in a multifaceted way on the presence of multiple hosts in suitable tick-habitat throughout the ecosystem [15, 39].

Current studies investigating presence and abundance of ticks and transmission dynamics of tick-borne diseases are mostly biased towards nature reserves [40, 41], and those which do focus on (sub-)urban settings are limited to public areas (city parks, suburban forests) [42]. The eco-epidemiology of ticks and their pathogens in gardens, which are the closest potential tick habitats to human homes, is understudied [43, 44]. This despite the high reported incidences of tick bites in gardens; with 47.4% of reported tick bites in Belgium being contracted in gardens [45] and 31% in The Netherlands [46]. Even basic research questions about the biotic and abiotic properties of gardens that explain tick-borne disease risk remain unanswered [41]. Several factors may contribute to these substantial knowledge gaps, including time and budget constraints, and the limited accessibility of private gardens to researchers, often due to privacy considerations. One of the approaches that partly alleviate these limitations is citizen science. Citizen science, an innovative and collaborative approach to scientific research, has gained interest and recognition in recent years. This method empowers the general public to actively participate in scientific data collection, analysis, and problem-solving. By leveraging the collective efforts of citizen scientists, researchers can expand their data collection efforts and enhance spatial and temporal coverage [47]. Furthermore, citizen science promotes public engagement and education, as participants contribute to scientific knowledge and gain a better understanding of issues affecting them [48]. With regard to citizen science projects targeting ticks and tick-borne diseases, the involvement of the participant has mostly been limited to online tick bite registration platforms [45, 46, 49], or sending ticks to a research lab for (molecular) analysis and identification, including accidentally encountered ticks [48, 50, 51], ticks that fed on humans [52, 53] or companion animals [53]. Very few studies have encouraged citizens to actively collect ticks in a standardized fashion in their gardens, which obviously requires a higher level of commitment as well as minimal training of participants [43, 54, 55].

In this study, we aimed to collect data on tick densities and pathogen prevalences in domestic gardens, as well as information on garden characteristics, through a citizen science project called"Teek a Break", with the simultaneous aim of raising public awareness on this topic. By asking citizens to survey their gardens for ticks in a standardized manner and analyzing these ticks in the lab for common tick-borne pathogens, we sought to investigate whether the fundamental principles regarding tick and pathogen ecology apply to this understudied habitat type of domestic gardens. Our hypothesis is that garden characteristics influence the prevalence of ticks and tick-borne diseases.

Methods

Sample collection

From August 2020 to October 2022, ticks were collected by volunteers in their domestic gardens through the “Teek a Break” project in Flanders, Belgium. The volunteers, citizen scientists, were recruited via an outreach campaign in national and regional media and through various communication channels and activities. Comprehensive participation instructions and guidelines were available on a dedicated webpage hosted on the website of the University of Antwerp in Dutch [56]. Participants were instructed to construct their own tick-drag consisting of a piece of white molton or flannel fabric (70 × 70 cm), a 1-m stick to which the fabric was attached, and a 3 m rope for dragging the flag. Ticks were collected by dragging 5-m transects [57] at a slow pace in their garden on multiple occasions and locations of their choice, preferably between 9 AM and 5 PM on non-rainy days when the vegetation was dry. Participants classified vegetation types into categories: wild vegetation, vegetable garden, flower bed, fallen leaves, mown grass, tall grass and fallow land. Each collection was restricted to one vegetation type, but could consist of multiple 5-m transects, and multiple collections in different vegetation types were possible on the same day, but with a minimum one-month interval between collections at the same location. Ticks were collected from the tick-drag by pressing transparent adhesive tape to the tick and transferring it onto a white paper. Ticks from different vegetation types and/or dates were stored on separate sheets of paper. For safety, participants stored the ticks in a freezer at −20 °C for at least one week before mailing them to the lab. For shipment, participants were advised to place the paper containing ticks in a zip-lock bag before enclosing it in an envelope.

Meta-data collection

For online data submission and participant recruitment, we collaborated with the “Mijn Tuinlab” project (“My Garden Lab”) [58]. This online platform, created in 2020 by a consortium of organizations focusing on citizen science in gardens, offered a variety of activities by multiple partners. Upon subscribing to this platform, participants completed a general online questionnaire about their garden's location, composition, and maintenance. Questions included details about garden type (urban, suburban, rural), garden surface area and presence of different types of green such as trees and bushes (yes/no). The choice between garden types was illustrated with an image of an example of each type (urban – “stadstuin”, suburban – “verkavelingstuin”, rural – “landelijke tuin”). The garden surface area was calculated by first having participants draw the boundaries of their parcel on a digital map, after which the building footprint was subtracted from the total parcel area. When subscribing specifically to “Teek a Break”, participants completed an additional short questionnaire regarding the presence of pets (dogs, cats), and observations of wildlife hosts in their garden. Following the completion of this questionnaire, participants gained access to the data input module, where they entered details for each tick collection event: date, time, weather conditions (sunny, cloudy, windy), vegetation type, number of 5-m dragged transects, number of ticks collected, and any accompanying remarks. All questionnaires used in this study are available in supplementary file 2. After submission, participants received a unique code to write on their sheet with collected ticks, enabling us to link the ticks arriving at our lab with the online submitted data.

Based on the answers of the wild life observation questions in the questionnaire, we calculated scores for the number of mammal species (0 to 8, henceforth ‘mammal score’) and the number of bird species (0 to 7, henceforth ‘bird score’) per garden. A score for the degree of greenness (1 to 9) of the garden was calculated based on the answers to the questions regarding the presence of different types of green in the garden (one point for each type present in the garden). Data were assigned to a season based on the collection date: spring (March, April, May), summer (June, July, August), fall (September, October, November) and winter (December, January and February).

As a side aim for this study, we investigated whether tick encounter risk could be partially predicted (accounting for other variables) by a general risk classification for tick bites at municipality level. This classification made by the Flemish government is based on a combination of registered tick bites and environmental indicators. Class 1 stands for “average risk”, class 2 for “moderately elevated risk” and class 3 for “significantly elevated risk” [59].

Tick identification and pathogen screening

Upon arrival at our lab, ticks were stored at −20°C. Ticks were removed from the adhesive tape and identified to life stage and species based on morphological characteristics with a stereo microscope and identification literature [60–62]. All nymphs (322 individuals) and adult ticks (male and female, 162 individuals) received at our lab were screened for the presence of B. burgdorferi s.l., R. helvetica, A. phagocytophilum, N. mikurensis, and B. miyamotoi. DNA was extracted using alkaline lysis [63]. Quantitative PCR (qPCR) was employed to detect B. burgdorferi s.l. [64], B. miyamotoi [65], N. mikurensis [66], and A. phagocytophilum [27], following the protocols described in the respective references. The list of the entire primer sets, mix components and run cycles is provided in supplementary file 3 (Tables D and E). R. helvetica was detected using duplex qPCR, targeting R. helvetica, spotted fever, and typhus group Rickettsia [67, 68]. Sequencing of the variable 5S-23S (rrfA-rrlB) intergenic spacer (IGS) region was performed to specify the genospecies of each B. burgdorferi s.l. positive sample [69]. To determine the ecotype for samples positive for A. phagocytophilum, a fragment of the groEL gene was amplified and sequenced [27, 28].

Statistical analyses

Generalized linear mixed effects models (GLMM) were fitted to test whether I. ricinus density and pathogen prevalence were independent of biotic and abiotic garden characteristics (including the presence of hosts), while correcting for temporal variation, and taking into account the correlation structure of repeated measurements from the same gardens. Two separate models were fitted for tick abundance (counts per collection, 0—∞) as the response variable: one only including data on adult ticks, and another one with only nymphs—the most important developmental stage for human infection. The distribution of model residuals for tick abundances typically shows high overdispersion, the reason why log-link and negative binomial distribution were utilized. Statistical modelling was performed on prevalence of pathogens with at least 10% prevalence (B. burgdorferi. s.l. and Rickettsia spp.) as binary response variables, with logit-link and binomial distribution. The infection outcomes of nymphs and adult ticks were combined in these models, to maximize the dataset for model convergence. Larvae were not considered in the models, given the near absence of vertical transmission in these bacteria (see ‘Introduction’ section) and lower human tick bite incidence [52, 70], resulting in low infection risk. Additionally, larval distribution is expected to be highly clumped, not correctly representing the local tick risk.

In all models, the explanatory variables included season, mammal score, bird score, presence of cats in the garden, presence of dogs in the garden, and location ID was included as a random effect. In the tick abundance models, additional variables included weather conditions, vegetation type, and the logarithm of dragged surface area as an offset. In the pathogen prevalence models, life stage, nymph density (nymphs/m²) and adult tick density (adult ticks/m²) were included as an explanatory variable, complemented by the interactions between life stage and all other explanatory variables. The variable weather conditions was not included in the pathogen models, since there was no effect expected, and the variable vegetation type was not included in these analyses due to convergence problems. All continuous variables were standardized by subtracting the mean and dividing by the standard deviation. Due to the small sample size of collection events on fallow land and in vegetable gardens, these vegetation categories were merged with wild vegetation and flower beds respectively. For weather conditions, we similarly grouped windy with cloudy weather. For the variable season, data collected in winter were left out of all models due to small sample size and convergence problems. The fall season was left out of the pathogen analyses for the same reason.

We conducted statistical analyses using generalized linear mixed effects models (GLMMs) in Rstudio (version 4.4.), employing the package glmmTMB to examine tick density and B. burgdorferi s.l. prevalence in relation to biotic and abiotic garden characteristics. Model diagnostics were evaluated using the DHARMa R-package [71], including the inspection of the residuals and assessing overdispersion and zero-inflation in the model. Automated model selection was performed using the MuMIn package in R, with the dredge function systematically generating and evaluating all possible combinations of the explanatory variables included in our global model. Each candidate model was ranked based on the Akaike Information Criterion (AIC), whereafter the model with the lowest AIC was extracted using the get.models function, and was considered the best model.

To assess the relationships between model residuals and the variables risk score, greenness score and garden surface area, we performed Spearman rank correlation tests on the Pearson residuals (mean per garden) from our generalized linear mixed effects models of tick density and pathogen prevalence. These variables were not included in the models as the risk score might explain variation which is now explained by the other variables, and including greenness score and garden surface area would limit the size of the dataset since this information was not provided for all gardens (sample sizes represented in Table A, supplementary file 1). Garden type was similarly not included in the models, due to the limited sample size of gardens provided with this information and therefore a Kruskal–Wallis test was performed on the Pearson residuals (mean per garden) of the different models and garden type.

No formal sample size calculation was performed prior to the study, as the number of participating gardens could not be predicted. Participation was open and voluntary. To estimate post-hoc the required sample size for a given precision when observing a proportion, the following formula is used: n = (Z^2 * p * (1-p))/d^2, where: n = required sample size, Z = Z-score corresponding to the desired confidence level (e.g., 1.96 for 95% confidence), p = estimated or observed proportion and d = desired margin of error (half the width of the confidence interval).

Results

A total of 185 gardens were included in the dataset, spanning across all provinces of Flanders. Among the gardens where garden type data was available, 64.3% of gardens were classified as rural (n = 83), 29.5% as suburban (n = 38) and 6.2% as urban gardens (n = 8). The median surface of the gardens was 1,138 m², with a minimum of 58 m² and a maximum of 64,311 m². Within these gardens, a total of 569 tick collection events were executed, with a median of 2 collections (IQR = 2) and a maximum of 28 collections per garden. Ticks were found in 82 of the 185 gardens (44.3%). The post-hoc sample size calculation, with a critical value of 1.96, an estimated proportion of 0.44 and a desired margin of error of 0.075, showed a required sample size of approximately 168 gardens. Among the gardens where garden type data was available, ticks were found in 50 out of 83 rural gardens (60.2%), 11 out of 38 suburban gardens (28.9%), and 4 out of 8 urban gardens (50%). For each of the vegetation types, we have received at least one tick record, except for vegetable gardens. Note that there was large variation in sampling effort (see fig. B in supplementary file 1), ranging from 4088.0 m² on mown grass to only 84.0 m² on vegetable gardens. The highest proportion of successful collections was on fallen leaves (47.4% of collections), and lowest on mown grass (18.3%) (see Table B in supplementary file 1). A total of 1,255 ticks were collected, with a maximum of 162 ticks in a single collection The data are skewed towards 0, with an overall median of 0 ticks per collection (IQR = 1). Excluding collections with 0 ticks, the median increases to 2 (IQR = 4). When looking at tick abundance per square meter, the maximum was 46.3 ticks/m², the median 0 ticks/m² (IQR = 0) and the median without zeros included 0.3 ticks/m² (IQR = 0.5).

1,162 ticks arrived in our lab for tick identification and pathogen screening, meaning that less than 10% of ticks were lost due to accidental loss of ticks during storage at home (e.g. freezer malfunctioning, accidental disposal of samples), or loss in the mail.

Tick abundance and prevalence: distribution over species and life stages

The percentages of ticks per species and life stage are shown in Table 1. The majority belonged to the generalist sheep tick Ixodes ricinus (94.7%), with very low numbers of the bird-specialized tick Ixodes frontalis (58 individuals, 5.0%) and the hedgehog tick Ixodes hexagonus (3 individuals, 0.3%) being collected. Most ticks were larvae (674 individuals; 58.0%) followed by nymphs (325 individuals; 27.9%).

Table 1.

The distribution of the 1,162 collected ticks by species and life stages, and across the 66 gardens where ticks were present, in percentages with total numbers (n) in parentheses

		Larvae % (n)	Nymphs % (n)	Adult females % (n)	Adult males % (n)	Total % (n)
I. ricinus	Ticks	54.0 (628)	27.0 (314)	5.5 (64)	8.2 (95)	94.7 (1101)
	Gardens	31.8 (21)	72.7 (48)	37.9 (25)	30.3 (20)	92.4 (61)
I. frontalis	Ticks	4.0 (46)	0.9 (11)	0.1 (1)	0.0 (0)	5.0 (58)
	Gardens	6.1 (4)	12.1 (8)	1.5 (1)	3.0 (2)	18.2 (12)
I. hexagonus	Ticks	0.0 (0)	0.0 (0)	0.3 (3)	0.0 (0)	0.3 (3)
	Gardens	0.0 (0)	0.0 (0)	1.5 (1)	1.5 (1)	1.5 (1)
Total	Ticks	58.0 (674)	27.9 (325)	5.9 (68)	8.2 (95)	100 (1162)
	Gardens	34.9% (23)	75.8% (50)	39.4% (26)	31.82% (25)	100% (66)

Percentages for ticks are calculated based on the number shown in parentheses relative to the total number of ticks collected in the study (1, 162). Percentages for gardens are based on the number of gardens in which a particular tick life stage was found, divided by the total number of gardens sampled (66). Percentages of gardens do not sum to 100% and numbers do not sum up to the total number of gardens, as co-occurrence of different tick species and/or life stages was commonly observed within a garden

When looking at the garden level (Table 1) more than 75,8% of gardens with ticks reported one or more nymphs, compared to only 34.9% for larvae. Given gardens with larvae present, 87.5% of larvae collections occurred in rural gardens (98.9% of all larvae). Regarding the vegetation types, 31.4% of larvae collections happened on wild vegetation (56.9% of larvae), followed by mown grass (25.7% of collections, 23.9% of larvae) and fallen leaves (22.9% of collections, 16.9% of larvae). In 38 gardens more than one life stage of ticks were collected, of which five gardens contained all life stages. In 92.4% of the gardens where ticks were present, I. ricinus was found, followed by 18.2% where I. frontalis was collected. I. hexagonus was found in only one garden (Table 1). In eight gardens more than one tick species was found, in all cases with I. ricinus and I. frontalis.

Ecological models: Tick density

We refer to Table 2 for test-statistics of each parameter estimate belonging to the respective biotic, abiotic and temporal covariates in the models. Adult tick density differed significantly between seasons with higher tick density in spring (P < 0.001) and summer (P = 0.005), compared to fall (Table 2, Fig. 1A). None of the other variables explained variation in adult tick density in a statistically significant way. A very similar association was shown between season and density of nymphs (Table 2, Fig. 1A). Nymphal density additionally showed a significant increase with the number of mammal species (P = 0.001, range scores 0—8) and an association with vegetation type: significantly more ticks were collected on wild vegetation (with fallow land) and fallen leaves, compared to mown grass, tall grass and flower beds (with vegetable gardens) (Table 2 and contrasts in table C in supplementary file 1, Fig. 1B). No significant correlations were found between tick abundances and the presence of pets and birds in the garden nor with the weather conditions during the tick collection.

Table 2.

Associations of biotic and abiotic characteristics of the gardens with the adult and nymph tick density by generalized linear mixed-effects models

Explanatory variables		Adult density				Nymph density
		Estimate	SE	z-value	p-value	Estimate	SE	z-value	p-value
Intercept		−9.2059	1.3814	−6.664	< 0.001	−7.1183	0.8687	−8.194	< 0.001a
Season	Fall*
	Spring	3.496	0.987	3.544	< 0.001	2.397	0.571	4.198	< 0.001
	Summer	2.907	1.047	2.777	0.005	2.750	0.619	4.445	< 0.001
Vegetation type	Fallen leaves*
	Flower bed + vegetable garden	−0.882	1.060	−0.832	0.405	−1.536	0.638	−2.408	0.016b
	Wild vegetation + fallow land	0.736	0.773	0.953	0.341	0.142	0.449	0.317	0.751a
	Mown grass	−0.787	0.762	−1.032	0.302	−1.496	0.454	−3.297	0.001b
	Tall grass	−0.181	0.854	−0.212	0.832	−1.745	0.594	−2.937	0.003b
Mammal score	Range: 0—8	0.287	0.148	1.942	0.052	0.412	0.126	3.261	0.001
Dog(s)	Present/absent					−0.807	0.510	−1.581	0.114
Random effect		Variance		st. dev		Variance		st. dev
Location ID		2.811		1.677		3.541		1.882

Significant p-values (< 0.05) are indicated in bold. *Reference category. Vegetation types with index ^asignificantly differ from vegetation types with index ^b, while types that share the same index do not differ. Number of collection events = 515, number of gardens = 171

Fig. 1.

Variation in tick density (logarithm of number of nymphs and adult ticks per m²) in relation to: A Season, B Vegetation type, C Garden type

After extraction of the model residuals for the smaller set of gardens for which greenness score, garden type (Fig. 1C) and garden surface could be registered (see “Statistical Methods”), we found that no additional variation could be significantly explained by these variables for the adult tick- and nymph density model residuals and for the B. burgdorferi s.l. prevalence model residuals (table D supplementary file 1). However, we did find a significant positive association between the Rickettsia spp. prevalence model residuals and garden surface area (rho: 0.407, P = 0.019). No significant correlation was found between the model residuals and the general tick risk score that was constructed by the Belgian government.

Pathogen prevalences: distributions over tick species and developmental stages

Tick level data: Out of the 484 screened ticks (322 nymphs and 162 adult ticks), 34.1% tested positive for at least one pathogen. Looking at I. ricinus (311 nymphs and 158 adults), the most prevalent pathogen was Borrelia burgdorferi s.l. (19.6% in both adults and nymphs), followed by Rickettsia spp. (17.1% and 9.0% in adults and nymphs, respectively) (Table 3). Among the I. ricinus ticks infected with B. burgdorferi s.l., mammalian Borrelia genospecies (B. afzelii + B. burgdorferi sensu stricto + B. spielmanii) were the most prevalent, with B. afzelii as the main genospecies (47.0% and 72.5% of the B. burgdorferi s.l. positive adults and nymphs respectively). Avian genospecies (B. garinii + B. valaisiana) were less prevalent, with B. valaisiana the most prevalent (23.5% in adults and 7.8% in nymphs, Table 4). All of the I. ricinus samples positive for A. phagocytophilum represented ecotype 1. The prevalences of B. miyamotoi and N. mikurensis in I. ricinus ticks were respectively 5.7% and 3.8% in adult ticks and 2.9% and 1.6% in nymphs (Table 3). No pathogens were detected in I. hexagonus (only three adults screened). In I. frontalis (one adult and 11 nymphs), A. phagocytophilum was the most prevalent pathogen (27.3% in nymphs) with mostly ecotype 1 but also ecotype 2 detected in one nymph, followed by B. burgdorferi s.l. and Rickettsia spp. (both 9.1% in nymphs). No B. burgdorferi s.l. genospecies could be identified in I. frontalis (Table 3). Co-infections (i.e. more than one pathogenic infection in the same tick individual) were found in 6.6% of the ticks. B. burgdorferi s.l. x Rickettsia spp. turned out to be the most common combination (6 nymphs and 4 adults). Two nymphs were co-infected with three pathogens: B. burgdorferi s.l. x A. phagocytophilum x B. miyamotoi.

Table 3.

The percentages (and numbers) of ticks tested positive for different pathogens per species and life stage

	B.burgdorferi s.l. % (n positive/n tested)			B. miyamotoi % (n positive/n tested)			A. phagocytophilum % (n positive/n tested)			N. mikurensis % (n positive/n tested)			Rickettsia spp. % (n positive/n tested)
	I. ricinus	I. frontalis	I. hexagonus	I. ricinus	I. frontalis	I. hexagonus	I. ricinus	I. frontalis	I. hexagonus	I. ricinus	I. frontalis	I. hexagonus	I. ricinus	I. frontalis	I. hexagonus
Adults (n = 162)	19.6 (31/158)	0.0 (0/1)	0.0 (0/3)	5.7 (9/158)	0.0 (0/1)	0.0 (0/3)	6.3 (10/158)	0.0 (0/1)	0.0 (0/3)	3.8 (6/158)	0.0 (0/1)	0.0 (0/3)	17.1 (27/158)	0.0 (0/1)	0.0 (0/3)
Nymphs (n = 322)	19.6 (61/311)	9.1 (1/11)	NA	2.9 (9/311)	0.0 (0/11)	NA	2.6 (8/311)	27.3 (3/11)	NA	1.6 (5/311)	0.0 (0/11)	NA	9.0 (28/311)	9.1 (1/11)	NA
Total (n = 484)	19.6 (92/469)	8.3 (1/12)	0.0 (0/3)	3.8 (18/469)	0.0 (0/12)	0.0 (0/3)	3.8 (18/469)	25.0 (3/12)	0.0 (0/3)	2.4 (11/469)	0.0 (0/12)	0.0 (0/3)	11.7 (55/469)	8.3 (1/12)	0.0 (0/3)

Table 4.

Prevalences of different genospecies in 68 I. ricinus nymphs and adult ticks positive for B. burgdorferi s.l., with total numbers in parentheses

	B. afzelii % (n)	B.b.s.s % (n)	B. garinii % (n)	B. spielmanii % (n)	B. valaisiana % (n)
Adults (n = 17)	47.0 (8)	17.7 (3)	0.0 (0)	11.8 (2)	23.5 (4)
Nymphs (n = 51)	72.5 (37)	2.0 (1)	5.9 (3)	11.8 (6)	7.8 (4)
Total (n = 68)	66.1 (45)	5.9 (4)	4.4 (3)	11.8 (8)	11.8 (8)

Garden level data: When looking at the level of the garden, 77.2% of the 57 gardens with nymphs and/or adult ticks contained at least one infected tick. This was 56.1% of the gardens for B. burgdorferi s.l., 21.1% for B. miyamotoi, 22.8% for A. phagocytophilum, 12.3% for N. mikurensis, and 42.1% for Rickettsia spp. When looking at the genospecies of B. burgdorferi s.l. at garden level (26 B. burgdorferi s.l.-positive gardens where genospecies sequencing was successful), in 69.2% of the gardens B. afzelii was found, followed by B. valaisiana (26.9%), B. burgdorferi s.s. (15.4%) and B. garinii and B. spielmanii (both in 11.5% of the B. burgdorferi s.l. positive gardens). In 33.3% of gardens with nymphs and/or adult tick, a tick with a co-infection was present. In 40.4% of the 57 gardens multiple pathogens were detected in the entire collections of ticks for that garden, with the most frequent combination being B. burgdorferi s.l. x Rickettsia spp. in 8.8% of gardens. A total of five different pathogens was found in 5.3% of the 57 gardens.

Ecological models: pathogen prevalences

B. burgdorferi s.l. prevalence: Borrelia burgdorferi s.l. prevalence decreased statistically significantly with the presence of dogs (P = 0.019) and increased with nymph density (P = 0.015). Additionally, a significant interaction was found between life stage and the number of bird species, showing that prevalence in nymphs increased with the number of birds (P = 0.003; fig. C supplementary file 1). Other covariates did not explain variation in B. burgdorferi s.l. prevalence (All P-values > 0.05, Table 5).

Table 5.

Associations of biotic and abiotic characteristics of the gardens with the prevalence of B. burgdorferi s.l. and Rickettsia spp. in nymphs and adult ticks by the generalized linear mixed-effects models

Explanatory variables		Borrelia burgdorferi s.l.				Rickettsia spp.
		Estimate	SE	z-value	p-value	Estimate	SE	z-value	p-value
Intercept		−0.9346	0.338	−2.763	0.006	−1.899	0.364	−5.216	< 0.001
Season	Spring*
	Summer	0.3397	0.636	0.534	0.593	0.767	0.425	1.805	0.071
Mammal score	Range: 0—8	−0.2808	0.174	−1.61	0.107
Bird score	Range: 0—7	−0.3275	0.220	−1.491	0.136
Dog(s)	Present/absent	−1.2303	0.526	−2.341	0.019
Stage	Adult*
	Nymph	−0.1265	0.392	−0.322	0.747	−0.916	0.429	−2.135	0.033
Nymph density	Nymphs/m2	0.3318	0.137	2.431	0.015
Interactions
Stage: Season	Spring*
	Summer	−1.1068	0.732	−1.512	0.131
Stage: Bird score	Adult*
	Nymph	0.95	0.317	2.998	0.003
Random effect		Variance		st. dev		Variance		st. dev
Location ID		1.683E-09		4.10E-05		0.292		0.541

Significant p-values (< 0.05) are indicated in bold. *Reference category. Estimates are presented in an untransformed log/logit scale. Number of ticks = 328, number of gardens = 44

Rickettsia spp. prevalence: The model investigating the Rickettsia spp. prevalence showed a slight significant correlation with life stage of the ticks: there tends to be a higher prevalence in adult ticks compared to nymphs (P = 0.033, Table 5). No significant correlations were found between Rickettsia spp. prevalence and the presence of any animals in the garden and with the season of the tick collection.

Discussion

This study, using data from the citizen science initiative "Teek a Break", provides new insights into the prevalence and ecological characteristics of tick populations in domestic gardens in Flanders, Belgium. The study's findings underscore the importance of residential areas as potential tick habitats, challenging traditional views that tick populations are mainly confined to natural and semi-natural environments. Our results suggest that even intensely managed spaces like private gardens, a high human contact risk habitat, can harbor substantial numbers of ticks. These ticks carry a variety of pathogens with prevalences similar to those in public greenspaces, thereby representing an overlooked public health risk. Similar results were found in another study in France [43], suggesting that strongly human-modified landscapes may still support tick populations if certain ecological conditions are met.

The presence of ticks in nearly half of surveyed gardens reveals that private gardens, particularly in rural areas, offer suitable habitats for tick species such as I. ricinus. With a post-hoc sample size calculation of 168 gardens, and 185 gardens included in our study, our sample size was deemed sufficient to provide a reasonably precise estimate of the proportion of gardens with ticks.

Nonetheless, several limitations must be acknowledged. First, the data are affected by sampling bias: rural gardens were overrepresented (67%), while urban gardens were underrepresented (5%) compared to Flemish garden statistics (52% rural, 21% urban, [72]). Second, even though participants were encouraged to submit results if no ticks were found, there was evidence that those who did not encounter ticks were less inclined to participate or submit data. This has likely led to an overestimation of overall tick prevalence across Flemish gardens.

Additional limitations arise from the use of citizen scientists for data collection, which limits control over data quality [73]. In some cases, essential information was missing (e.g. ticks were submitted without contextual data), and it remains uncertain to what extent participants fully adhered to the prescribed methodology. For example, it is unclear whether the correct type and dimensions of flagging fabric were used, or whether vegetation types—such as short versus tall grass, or fallow land versus wild vegetation—were classified accurately. Future research could address these issues by including illustrative photographs of the vegetation types in the instructional manual to improve consistency and accuracy in classification. The reliability of reported bird and mammal observations is uncertain, which could be solved in future projects by providing standardized bird monitoring protocols, possibly combined with camera traps. Despite considerable efforts to educate participants on the identification of different tick life stages, it remains possible that some ticks—particularly larvae—were not recognized due to their small size, potentially leading to underreporting.

Because participation was open to everyone and participants had complete freedom in choosing when and where to flag, certain garden types, seasons, vegetation types and weather conditions were underrepresented in the dataset. This occasionally has led to grouping categories during the analysis. Depending on the research objective, future studies might benefit from stricter protocols—for example, requesting standardized flagging efforts across specific months, vegetation types, or weather conditions, or pre-selecting gardens (e.g. based on garden type or location) to ensure more balanced sampling.

All life stages of I. ricinus and two other tick species were present in gardens, indicating that multiple garden- and vegetation types provide the minimal biotic and abiotic requirements for tick survival, moulting, hatching, as well as questing (see “introduction”). While I. ricinus is considered to be a host generalist [74], are I. frontalis and I. hexagonus host specialists. Ixodes hexagonus, an endophilic tick known to infest hedgehogs or foxes and living in their borrows [65], was only found in one garden. Ixodes frontalis is an exophilic tick infesting terrestrial bird hosts and is associated to the vegetation underneath branches on which birds roost and forage. It shows questing behavior, similar to I. ricinus’, which is likely the reason why it has been sampled in multiple gardens [75]. Questioning of the garden owners about the specific collection site revealed that these individuals were collected underneath or in the proximity of trees and bushes. Ticks were found on almost all sampled vegetation types. We observed notable variation in tick densities across vegetation types within gardens, which aligns with the idea that specific microhabitats may support tick populations more effectively than others [76]. For example, fallen leaves and wild vegetation hosted the highest nymph densities, likely due to their moist and shaded environments, which are ideal for preventing desiccation and supporting questing behaviors. These findings support prior research indicating that ticks are more likely to thrive in environments that provide higher humidity and cooler temperatures, as such conditions reduce water loss, which is critical for tick survival during questing [43, 44, 76–78]. However, another study investigating tick-borne encephalitis risk, found no association with removal of fallen leaves [79]. Finding the ticks in (micro)climatic unfavorable locations may indicate that I. ricinus can detach from a host (more or less) randomly [80]. Interestingly, short grass, which is typically considered a low-risk area for ticks, showed comparable tick density to long grass. This is in contrast with findings of other studies, stating that mowing is an effective method to reduce tick abundance [81, 82] and tick-borne encephalitis risk [79]. Our data suggest that tick densities are influenced more by moisture-retaining ground cover like leaf litter than by grass height alone. This is in line with findings from other studies in gardens, predicting the highest probability of questing ticks on transects with hedges or groundcover in gardens [44] and stating that the likelihood of finding a nymph was nearly three times higher in transects shaded by vegetation compared to those in open areas, with no relationship between nymph occurrence and grass height [43]. This challenges the efficacy of traditional garden management practices that focus on mowing as a means of reducing tick presence, recommended by public health guidelines [81–84], pointing instead to the importance of managing specific vegetation types such as the removal of fallen leaves or the maintenance of wild vegetation.

Our study highlighted the link between mammalian host presence and nymphal density, the primary life stage driving human infections with Borrelia burgdrferi s.l. The positive correlation between mammal diversity and nymphal density underscores the role of mammals in sustaining tick populations in gardens, with implications for public health. This is consistent with findings in another study in gardens, finding higher tick densities in gardens with observations of deer [43], and by studies in natural ecosystems where host diversity, particularly of mammals, supports higher tick densities by providing a consistent blood meal source for all life stages [85, 86]. However, we found no significant association between tick density and the presence of domestic pets (e.g., dogs and cats), which is in line with findings of another study stating the risk of encountering ticks did not appear to be associated with the presence of a dog [43]. Domestic animals, such as dogs, can influence tick abundance in multiple ways. On one hand, they may support tick populations by bringing ticks into the garden after walks in forests or parks, and by hosting ticks already present in the garden. On the other hand, they can reduce tick populations—either directly, through the use of acaricides that kill feeding ticks, or indirectly, by preventing wild host species from entering the garden. [43, 87, 88]. If the garden is not located in proximity to natural greenspaces, the influx of ticks by wild hosts will be low. And if the amount of small host species in the garden is limited, the life cycle of ticks is less likely to be completed, despite the presence of domestic animals, as mainly the adult female ticks feed on medium to large domestic animals like dogs [17, 89].

Important was the relationship between garden type (rural, suburban, urban) and tick occurrence. Although the majority of ticks were collected in rural gardens, ticks were also found in suburban and urban gardens. This presence across garden types implies that gardens close to dense residential areas can contribute to exposure risk and support tick life cycles, even when isolated from larger, contiguous natural areas. Larvae in particular were more often found in rural settings. The majority of larvae were found in a small number of gardens, while nymphs and adults were more homogeneously distributed. Adult ticks use propagation hosts as for instance roe deer, which are more abundant in rural regions [90]. Urban regions therefore may represent ecological sinks to ticks. Similarly, recent research emphasizes that both local garden features and the broader landscape significantly influence tick presence in residential areas [43, 44]. The Belgian risk score did not explain additional variation in tick density. However, without the correction for other explanatory variables, risk score was significantly correlated with tick densities, underlining its general use for risk assessments.

A significant finding of this study is the high prevalence and diversity of tick-borne pathogens in ticks collected in gardens, which presents a potential health hazard for residents. We identified B. burgdorferi s.l. as the most common pathogen in gardens (around 19% in both adult ticks and nymphs), which is in line with results from forests (17% in nymphs [91]) and parks (almost 18% in nymphs and 33% in adult ticks [42]) in Flanders. The prevalences of the other pathogens were also comparable to those observed in forested and semi-natural environments [42, 91]. This similarity in pathogen prevalences between garden- and natural ticks and the variety of tick-borne pathogens present in gardens suggests that host communities transmitting more than one pathogenic agent are present in gardens, increasing the risk of disease transmission to humans and pets within residential areas.

Borrelia burgdorferi s.l. was the most prevalent pathogen in I.ricinus ticks, however, this pathogen was not detected in I. hexagonus ticks and only in one I. frontalis nymph. Important to consider is the small sample size of these species in our study, as other studies do find B. burgdorferi s.l. in these species [64, 92, 93]. Both I. hexagonus [94] and I. frontalis [95] are competent vectors for certain Borrelia genospecies, not just carriers. Since they share hosts with I. ricinus, they indirectly affect the tick-borne pathogen risks to humans. The prevalence of B. burgdorferi s.l. showed no correlation with any of the garden characteristics, except for the presence of dogs. While the presence of dogs could potentially influence the presence of wild hosts in gardens, we would rather expect an influence on the tick density. Dogs are used as a sentinel for human Lyme borreliosis [96]. Serological studies showed that exposure of dogs to B. burgdorferi is similar to or even higher than the geographical distribution of reports of Lyme borreliosis in humans [97, 98]. Although Borrelia can be transmitted from dogs to ticks, dogs are not considered as important reservoir hosts [97], and dogs can get clinical manifestations of Lyme borreliosis and can subsequently be treated for the infection [99]. Therefore a hypothesis could be that dogs being present in a garden as a potential host to ticks, without being an important reservoir for Borrelia, results in a decrease in the prevalence of Borrelia in the garden. However, dogs in Belgium are often treated with acaricides, questioning the importance of dogs as hosts to ticks in a garden.

The prevalence of B. burgdorferi s.l. in nymphs increased with the number of bird species present, despite the fact that mammalian Borrelia genospecies (B. afzelii + B. burgdorferi s.s. + B. spielmanii) were more prevalent in nymphs than avian genospecies (B. garinii + B. valaisiana). One possible explanation is that bird species richness, as reported by garden owners, serves as a proxy for overall biodiversity. Birds are generally easier to observe than small mammals, which are often nocturnal and elusive. As a result, the number of observed bird species may reflect both the actual avian presence and broader ecological richness. The majority of ticks positive for A. phagocytophilum represented ecotype I. However, ecotype II, generally associated with roe deer [27], was detected in one I. frontalis nymph, a species considered to feed on birds [93]. There seem to be no records of ecotype II in birds or of I. frontalis infesting roe deer. A lower prevalence of Rickettsia spp. was found in nymphs, compared to adult ticks, which is in line with results of other studies [100, 101]. This may reflect differences in host preferences or transmission efficiency between tick life stages. N. mikurensis and B. Miyamotoi, the causative agent for relapsing fever, are present in seven and twelve Flemish gardens respectively. Little is known about the infectiousness and reservoir hosts of these pathogens [65, 30, 66].

Co-infections in individual ticks were also observed (over 6%), with B. burgdorferi. s.l. x Rickettsia spp. being the most common combination. This is a relatively high prevalence, compared to around 2% in Flemish parks [42]. Tick can acquire co-infections by feeding on one host individual infected with multiple pathogens or by feeding on multiple hosts throughout the life cycle [102, 103], and co-infections can influence transmission dynamics [102].

The presence of ticks and their pathogens in domestic gardens presents an important public health challenge, as these spaces are frequently used for recreation and household activities. The proximity of tick habitats to homes increases the likelihood of exposure, particularly for children and pets, who are often closer to the ground and spend extended periods outdoors. In contrast, a previous study on the use of protective practices against tick bites showed that people are less likely to use these practices in gardens, compared to natural greenspaces [104]. Our findings on habitat-specific risk suggest that the combination of exposure (tick density) and hazard (pathogen prevalence) makes some gardens and vegetation types higher-risk environments for tick-borne disease transmission. Exposure is also influenced by the behavior of garden visitors [105]. On the one hand, some parts of the garden are used more than others, for instance garden visitors might be less likely to spend time in fallen leaves and wild vegetation, compared to the lawn. While on the other hand some activities result in more contact with the vegetation, e.g. sun bathing in the grass compared to playing in the grass [106].

An infection with tick-borne pathogens can cause severe health problems if not diagnosed and treated adequately. Our findings have several implications for managing tick exposure in private gardens. Firstly, managing leaf litter and reducing dense, shaded vegetation may help lower tick densities by reducing ideal questing habitats. These actions could be considered in areas of the garden where garden users spend most of their time, while other parts of the garden could be managed with a focus on biodiversity. Additionally, educational campaigns that inform the public about tick presence in various garden types could encourage more protective behaviors, such as wearing long clothing and performing a tick body check, even in settings typically perceived as low-risk [104].

Conclusion

This study illustrates the potential of citizen science to expand data collection and knowledge on tick populations in domestic environments. We reveal gardens as a habitat for tick populations, emphasizing their potential contribution to the risk of tick-borne disease transmission to both humans and animals. The presence of ticks and the detection of pathogens in these environments suggest the need for further research into the epidemiological implications, including human and animal exposure to tick-borne diseases. Future studies could focus on monitoring tick exposure and infection rates in local humans, domestic animals and local animal populations, to better assess public health risks.

While our findings do not directly assess the impact of garden management strategies or public education, they underscore the importance of considering gardens in broader tick burden management efforts. By improving our understanding of tick ecology in private gardens, this research provides a foundation for future studies that can inform targeted management practices and risk mitigation strategies for homeowners and policymakers.

Authors’ contributions

KR, EM, DH and MVG contributed to the study conceptualization and design. Material preparation and data collection were performed by KR and citizen scientists. Pathogen analyses were performed by MF and HS. Data were analyzed by KR, DH and ML and interpreted by KR, DH and EM. The first draft of the manuscript was written by KR and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This study was funded by the Interreg North Sea Region programme, NorthTick. The funding body had no role in the design of the study, collection, analysis, interpretation of data nor in writing the manuscript.

Data availability

The data used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Study participation was voluntary. All participants provided informed consent and agreed with the terms and conditions of the My Gardenlab platform before entering the study. No medical or personal data were collected. According to our national guidelines (FWO), ethical approval is not obligatory for studies involving voluntary participants when no medical procedures or collection of medical data is performed, as is the case in our study. Therefore no ethical approval was requested. All information is kept confidential and anonymous, in compliance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.