Cross‐continental comparison of parasite communities in a wide‐ranging carnivore suggests associations with prey diversity and host density

Abstract Parasites are integral to ecosystem functioning yet often overlooked. Improved understanding of host–parasite associations is important, particularly for wide‐ranging species for which host range shifts and climate change could alter host–parasite interactions and their effects on ecosystem function. Among the most widely distributed mammals with diverse diets, gray wolves (Canis lupus) host parasites that are transmitted among canids and via prey species. Wolf–parasite associations may therefore influence the population dynamics and ecological functions of both wolves and their prey. Our goal was to identify large‐scale processes that shape host–parasite interactions across populations, with the wolf as a model organism. By compiling data from various studies, we examined the fecal prevalence of gastrointestinal parasites in six wolf populations from two continents in relation to wolf density, diet diversity, and other ecological conditions. As expected, we found that the fecal prevalence of parasites transmitted directly to wolves via contact with other canids or their excreta was positively associated with wolf density. Contrary to our expectations, the fecal prevalence of parasites transmitted via prey was negatively associated with prey diversity. We also found that parasite communities reflected landscape characteristics and specific prey items available to wolves. Several parasite taxa identified in this study, including hookworms and coccidian protozoans, can cause morbidity and mortality in canids, especially in pups, or in combination with other stressors. The density–prevalence relationship for parasites with simple life cycles may reflect a regulatory role of gastrointestinal parasites on wolf populations. Our result that fecal prevalence of parasites was lower in wolves with more diverse diets could provide insight into the mechanisms by which biodiversity may regulate disease. A diverse suite of predator–prey interactions could regulate the effects of parasitism on prey populations and mitigate the transmission of infectious agents, including zoonoses, spread via trophic interactions.


| INTRODUC TI ON
Parasite-host relationships affect ecological processes in profound ways. As a key example, parasites can regulate host abundance by reducing host survival and fitness (Arneberg et al., 1998;Lafferty et al., 2008;. Parasites that exploit trophic interactions can also affect host abundance via influences on predator-prey dynamics, for example, by affecting foraging decisions by consumers (Hutchings et al., 2006) or by making prey more susceptible to predation (Lafferty et al., 2008;Lefèvre et al., 2009).
Understanding the ecological role of parasites is essential given today's extreme levels of environmental change. For example, environmental stressors can disrupt natural balances between parasites and hosts, leading to altered transmission dynamics and host immune function (Acevedo-Whitehouse & Duffus, 2009;Brearley et al., 2013). Biodiversity loss can also influence parasite-host relationships, for example, through the removal of nonhost species that minimize parasite transmission via dilution (Civitello et al., 2015). Moreover, the loss of parasite diversity from ecosystems may have substantial effects on ecosystem processes (Wood & Johnson, 2015). Collectively, the diverse effects of parasite-host interactions on ecosystem function combined with ongoing change call for further understanding of parasite-host relationships.
Host geographical range, population density, and body size are recognized as general predictors of parasite richness across a broad range of taxa (Kamiya et al., 2014). Parasite communities of gray wolves (Canis lupus) thus present a valuable system for understanding the role of parasites in host regulation and predatorprey dynamics. As top predators with few natural (i.e., nonhuman) predators, infectious disease may play a role in top-down regulation of wolf populations, particularly by causing mortality in pups (Kreeger, 2003;Mech et al., 2008;Mech & Kurtz, 1999;Peterson et al., 1998;Seguel & Gottdenker, 2017). Parasites might also affect host vulnerability to predation (Fenton & Rands, 2006). Such infections include Echinococcus granulosus in moose (Alces alces) (Joly & Messier, 2004), which could influence broader predator-prey relationships by altering the proportion of moose versus alternate prey species in wolf diet. Moreover, wolves occupy one of the broadest ranges among terrestrial mammals, including maritime, plains, and montane regions (Mech & Boitani, 2003), with dietary differences among regional populations (Newsome et al., 2016). This diversity of habitats and diets presents an opportunity to examine general processes (e.g., host density and prey diversity) that influence the prevalence of fecal parasites across populations and differences related to specific geographical regions and habitats.
Ongoing studies of wolf-parasite associations are particularly important as emerging findings from Europe and North America have revealed new infections and spatial changes in parasitism of wolves (Beck et al., 2017;Ćirović et al., 2015;Hermosilla et al., 2017;Molnar et al., 2015Molnar et al., , 2019Pavlović et al., 2018) as well as in other canids (Fuehrer et al., 2016). Given that canids host several zoonotic parasites, changing host-parasite relationships could have major implications for both ecosystem and public health (Cerda et al., 2018;Tasić-Otašević et al., 2016). Moreover, these changes combined with the long-distance dispersal capacity of wolves and other canids also highlight the importance of establishing an early-warning system analyses. Canadian Prairie Study Area: We gratefully acknowledge funding from Parks Canada, Manitoba Conservation's Sustainable Development Innovations Fund, Louisiana Pacific Canada, and World Wildlife Fund Canada. AVS was supported by the Danish Natural Science Research Council (grant 1337-00007), and a senior postdoctoral fellowship from Insubria University in Varese, Italy. For the YNP, PNALM, and MNP study areas, research was supported by the laboratories of animal physiology, parasitology, and ecoethology of the University of Neuchâtel, by the Fondation Gérard Pierre, the Société académique neuchâteloise, and Private donors with wolf density. Contrary to our expectations, the fecal prevalence of parasites transmitted via prey was negatively associated with prey diversity. We also found that parasite communities reflected landscape characteristics and specific prey items available to wolves. 5. Several parasite taxa identified in this study, including hookworms and coccidian protozoans, can cause morbidity and mortality in canids, especially in pups, or in combination with other stressors. The density-prevalence relationship for parasites with simple life cycles may reflect a regulatory role of gastrointestinal parasites on wolf populations. Our result that fecal prevalence of parasites was lower in wolves with more diverse diets could provide insight into the mechanisms by which biodiversity may regulate disease. A diverse suite of predator-prey interactions could regulate the effects of parasitism on prey populations and mitigate the transmission of infectious agents, including zoonoses, spread via trophic interactions.

K E Y W O R D S
Canis lupus, dietary diversity, direct life cycle, indirect life cycle, noninvasive monitoring, population density (i.e., baseline data and ongoing monitoring programs) to identify the emergence and spread of new pathogens (Fuehrer et al., 2016).
Examining patterns in wolf-parasite associations is problematic because study designs and methods used in parasitological research often differ among studies, complicating the potential for inference from multiple studies. Despite these challenges, recent efforts have been made to compare findings across various studies to highlight wide-scale patterns of parasite diversity (e.g., Pappalardo et al., 2020). Our aim with this comparative study, which includes different ecosystems, sampling periods, and parasitological methods, was to investigate associations between predator-prey relationships and parasite communities. Here, we combined results from three studies (Bryan et al., 2012;Molnar et al., 2019;Stronen et al., 2011) and compiled associated data on population metrics, such as wolf density and dietary diversity (Tables 1 and 2). The three studies include data from six wolf populations on two continents.
All three studies employed a similar, noninvasive research approach using fecal samples to identify gastrointestinal parasites over multiple years and seasons.
Fecal samples have the advantage of being noninvasive and relatively inexpensive to collect; however, parasite detection and quantification is complicated by temporal variability in the shedding of larval stages (i.e., the eggs, oocysts, sporocysts, cysts, and larvae of helminths and protozoans) in feces (Branda et al., 2006;Cartwright, 1999). We therefore focused our comparative analysis on differences in fecal prevalence (i.e., proportion of fecal samples in which a parasite's larval stage was detected) rather than parasite richness and intensity, which are particularly sensitive to detection and quantification. Notably, fecal prevalence of larval stages represents the level of environmental contamination and potential for transmission of common parasites among hosts, including wolves and prey species that serve as intermediate hosts. Our overall goal was to identify processes that shape the prevalence of fecal parasites across populations over broad spatial and temporal scales.
Seasonal differences often affect wolf diet and parasite prevalence, which have been found to vary among prey species, parasite species, packs, and years (Bryan et al., 2012;Ciucci et al., 2018Ciucci et al., , 2020Sallows, 2007). Our specific objectives were therefore to examine whether fecal prevalence of parasites was associated with (a) wolf density, which could indicate density-dependent transmission and a potential regulatory role of parasites on wolves; (b) dietary diversity, which could influence the parasite taxa to which wolves are exposed via trophic interactions; or (c) study areas, which represent a proxy of environmental variables that influence wolf-parasite associations.
In addition, we examined potential associations of parasitic prevalence with the number of alternative hosts and the recent history of each wolf population (i.e., historical presence vs. recolonization).
To achieve our first two objectives, we classified parasites into broad categories according to life cycle. Specifically, parasites with direct life cycles are those that require only one host to complete their life cycles (i.e., are transmitted directly among wolves or other suitable hosts). For parasites with direct life cycles, higher wolf densities might facilitate transmission; therefore, we predicted that the fecal prevalence of parasites with direct life cycles would be positively correlated with wolf density. Alternatively, parasites with direct life cycles could be transmitted to wolves via other canids, including domestic dogs (C. l. familiaris). Accordingly, if alternative hosts are primarily responsible for transmission, the fecal prevalence of parasites with direct life cycles might be correlated more strongly with the diversity or abundance of alternative host species than with wolf density. Parasites with indirect life cycles are those that have a larval stage that requires one or more intermediate hosts to complete its life cycle (i.e., are transmitted to wolves via prey). For parasites with indirect life cycles, the relationship between diet diversity and the prevalence of fecal parasites could be even more complex. On one hand, given that wolves are exposed to parasites from many prey species, and that parasite intensity may be limited within a host by competition for resources (e.g., attachment sites in the gastrointestinal tract), the fecal prevalence might be similar across study areas regardless of variation in wolf diet. In this case, we would expect no relationship between the fecal prevalence of parasites with indirect life cycles and dietary diversity. On the other hand, parasitic diversity often correlates with host diversity (Lafferty et al., 2008), so ecosystems with a higher diversity of either definitive hosts, including domestic or wild carnivores, or intermediate hosts (i.e., prey species for definitive hosts) might have higher parasitic diversity. To test these competing hypotheses for parasites with direct and indirect life cycles, we assessed the effect of diet diversity, wolf density, and number of alternative hosts on fecal parasite prevalence. We also included recent history of wolf residency as a predictor in our models, as parasite communities may differ among established and recolonizing wolf populations (Molnar et al., 2019). To address our third objective of examining differences among study areas, we compared parasitic larval stages identified across study areas and interpreted our findings in relation to the life history and ecological requirements of specific parasite groups.

| Study areas and sample collection
Our study areas included two in Europe, one in the United States, and three in Canada (Table 1; Figure 1a  Note: The six areas comprise five different habitat types, where the last two entries represent two areas from similar habitat but with different degrees of protection and human use. 1 W-winter (1 October-31 December), P-spring (1 January-31 March), S-summer (1 April-30 June), and F-fall (1 July-31 September). PNALM was sampled in , MNP in 2005and 2007, YNP in 2007, GBR in 2005, RMNP in 2001-2005, and DMPPF in 2003-2005 Classified according to the Shannon index (H) ( Table 3). 3 Wolf density: (a) Molnar et al. (2015); (b) Darimont and Paquet (2000); and (c) Stronen et al. (2012). DMPPF density is unknown but believed to be like RMNP.
Unknown rodent -  (Table 1) and wolf diet (Table 2). Within each study area, samples were collected from a broad spatial scale (ranging from 5,000 to 60,000 km 2 ) and from multiple wolf social groups. Sampling at each site spanned at least two seasons and occurred over 2-4 different years. Due to differences in the timing and spatial extent of sampling among the six studies, we were unable to include the effects of year and season in our analysis. Moreover, we did not have comparably detailed information on social groups or individual wolf identity in all study areas. Consequently, we pooled parasite data collected from all feces within each study area. Although this limits our ability to draw inferences about spatial and temporal variability in wolf-parasite associations, our analysis nonetheless provides a coarse-scale overview of fecal parasite prevalence across the six study areas (

| Identification of larval stages of parasites
Samples from all study areas were frozen at −20°C before analysis.
Larval stages (i.e., eggs, oocysts, sporocysts, and nematode larvae) of the parasites in wolf feces collected from the three Canadian study areas (RMNP, DMPPF, and GBR) were identified at the University of Saskatchewan using a modified Wisconsin sugar flotation method on 4 g of feces followed by microscopy (Foreyt, 1989(Foreyt, , 2001Stronen et al., 2011). For YNP, MNP, and PNALM, helminth eggs and protozoan cysts were identified using a sodium acetateacetic acid-formaldehyde (SAF) procedure on 1.55 g of feces followed by microscopy (Molnar et al., 2015). Both flotation and SAF methods are well suited to identify the presence of helminth eggs and protozoan cysts in fecal samples; however, the ability of the two procedures to detect parasite stages may differ by taxa. The flotation method uses more fecal material (4 g) than the sedimentation method (1.55 g) and should therefore be more sensitive. The sedimentation method, however, may be better than flotation for detecting heavier larval stages, such as the eggs of trematodes and Taeniid cestodes, which may in part make up for the lower sensitivity related to sample volume for some parasite taxa (Öge et al., 2017;Wolf et al., 2014). We applied correction factors based on published comparisons of parasite detection methods to examine the potential effects of methodological differences on our results (Appendix S1).
Results using the dataset where we corrected for methodological differences in the fecal flotation used in GBR, RMNP, and DMPFF were similar to those obtained using the original dataset (Appendix S1; Figures S1 and S2;

| Data preparation and analysis
To focus on helminth and protozoan endoparasites of wolves, the ectoparasite Demodex spp. and endoparasites Moniezia and Soboliphyme spp.-whose definitive hosts are thought to originate in the prey consumed by the wolves rather than in wolves themselveswere excluded from the comparison. Unidentified nematode larvaewhich may have represented free-living nematodes or nematode parasites of prey-were also excluded due to potential differences in detection and identification among studies combined with a low fecal prevalence (Bryan et al., 2010(Bryan et al., , 2012Molnar et al., 2019). For taxa that had been identified to species or genus in one study, but to genus or order in another, the higher-level classification was used.
Specifically, parasites classified as Isospora by Bryan et al. (2012) and Molnar et al. (2015) were classified as Coccidea to be consistent as a random effect. We included only one predictor per candidate model because we had only one measure of each variable for each of the six sites (i.e., an effective sample size of six). Models were ranked using Akaike information criterion corrected for small sample size (AIC c ). We found that AIC scores were lower or similar when independent variables were included as continuous or categorical predictors; therefore, we included all variables except wolf population history as continuous variables in the final models. We used a simulation to examine potential effects of methodological differences (i.e., fecal flotation vs. sedimentation) on our comparisons of parasites with direct and indirect life cycles among study areas (Appendix S1).
Accordingly, we used published data comparing the two approaches to identify correction factors for each parasitic taxon (Table S1 in Appendix S1). We then randomly added or removed positive fecal samples depending on whether a particular taxon was over-or under-reported using sedimentation compared with flotation. We then recreated figures and reran the statistical analysis to evaluate the effects of methodological differences on our results and interpretations (Figures S1, S2 and Table S2 in Appendix S1

| Parasites with direct life cycles
Comparison of parasites with direct life cycles in wolves showed that the highest fecal prevalence occurred in PNALM and YNP, which also had the highest wolf densities (Figures 2 and 3a). This was followed by the three areas GBR, RMNP, and DMPPF with medium wolf den- 2.5-20.5) more likely to be positive for parasites for each increase in the level of wolf density (i.e., low, medium, and high, ranging from 11.5 to 50 wolves/km 2 ). The high fecal prevalence of parasites with direct life cycles in PNALM was explained primarily by the presence of a single group, Capillaria spp. (80% prevalence), a parasite of the respiratory system, as well as by the hookworm Uncinaria spp. (15% prevalence). In contrast, YNP wolves exhibited primarily the parasite group Toxascaris leonina (15% prevalence), which is usually found in pups but rarely in adults.

| Parasites with indirect life cycles
Contrary to our expectations, comparison of parasites with indirect life cycles showed clear differences in fecal prevalence among areas (Figures 2 and 3b). Most notably, fecal prevalence was lower in areas with higher dietary diversity (

| Parasites associated with specific study areas
Although several parasite taxa were common across study areas (e.g., Taeniids), others were study area specific as associated with different landscapes inhabited by wolves. In particular, the taxa Diphyllobothrium spp. and Metorchis spp. were linked to aquatic environments and consumption of fish and therefore were detected in GBR but not elsewhere. In addition, we observed that Alaria spp., a parasite also linked with aquatic environments, was common in DMPPF (ca. 30% prevalence) and RMNP (ca. 13%) wolves, where freshwater and wetlands are abundant.

TA B L E 3
Diversity indices of prey items in the diet of wolves in each study area, calculated using prey species and proportions listed in Table 2 F I G U R E 2 Percentage of wolf parasites with direct and indirect life cycles per study area in relation to estimated wolf density (individuals per 1,000 km 2 ) and dietary diversity (see Table 2  where the assemblages C and D (G. canis) are hosted by dogs and other canid species.

| Parasites with direct life cycles
For parasites with direct transmission, our findings most strongly support a density-dependent relationship between fecal prevalence of parasites and wolf density across populations. A similar  relationship has been demonstrated in nematodes across 19 mammalian species (Arneberg et al., 1998) and relates to increased transmission probabilities as host density increases. Pup mortality and the prevalence of canine distemper virus were both higher in an area of higher wolf density (Almberg et al., 2009), and the relationship between population density and risk of exposure was supported by a broad-scale analysis of pathogens in North American wolves . Although gastrointestinal parasites often have minimal effects on host populations, some directly transmitted parasites in this comparative study, including hookworms in the genera Uncinaria and Toxocara and Coccidian protozoans, can cause morbidity or mortality in canids (e.g., Kreeger, 2003;Mech & Kurtz, 1999;Seguel & Gottdenker, 2017), especially in pups or in combination with other stressors (Hudson et al., 2006). If mortality due to these parasites with direct life cycles increases with higher density within populations, then our findings imply that gastrointestinal parasites might play a role in regulating wolf population size. tive analysis of gastrointestinal parasites in reptiles found that the detection of nematode eggs was 13.6% higher using flotation than sedimentation (Wolf et al., 2014). However, our simulation indicated that the methodological differences did not have any major effects on the findings for the taxa we examined. Moreover, because we found the highest prevalence of nematodes in samples analyzed with SAF sedimentation-the technique expected to produce more conservative results for nematodes-our results suggest that these nematodes were indeed more prevalent in wolf feces collected from PNALM and YNP compared with the other study areas. Such a pattern aligns with the relationship we observed between higher wolf density and higher prevalence of parasites with direct life cycles.

| Parasites with indirect life cycles
In contrast with our original prediction, the overall fecal prevalence of parasites with indirect life cycles differed among study areas in relation to dietary diversity. Notably, the fecal prevalence of parasites with indirect life cycles was lowest in Europe, where wolves consume a wide variety of prey that includes wild and domestic species. By contrast, parasite prevalence was higher in the three North American study areas where wolves specialize on fewer species, with the highest parasite fecal prevalence occurring in YNP where elk comprise >95% of wolves' winter diet. The difference in fecal prevalence between continents was driven largely by Sarcocystis spp., which was lower in wolves from Europe compared with the three North American study areas. Although Sarcocystis spp. infects prey species consumed by wolves in Europe and North America, different species of Sarcocystis differ in host specificity, especially in their definitive hosts. Wolves with more diverse diets could potentially be exposed to fewer wolf-specific Sarcocystis species, although this possibility warrants further research. Differences in the fecal prevalence of Sarcosystis spp. sporocysts and Taeniid eggs in wolf feces could also reflect different prevalence among prey species.
Differences in the fecal prevalence of parasites with indirect life cycles may have been influenced by methodological differences in coprology among studies. For example, Wolf et al. (2014) found that the flotation technique had better detection capabilities for protozoan oocysts than SAF sedimentation. Therefore, the methodology could have influenced the apparent higher prevalence of Sarcocystis in GBR, RMNP, and DMPPF compared with PNALM and MNP. However, the simulation results suggested that methodological differences did not have a major effect on the detection of parasites with indirect life cycles. Additionally, the prevalence of Sarcocystis spp. in YNP was also higher than in the European study areas PNALM and MNP, all of which used SAF sedimentation. The trematode Alaria spp. was most common in the North American study areas RMNP and DMPPF examined with flotation, whereas trematodes have been reported to be more detectable with SAF (Wolf et al., 2014). For the cestode Diphyllobothrium spp. and the Taeniid family, we have not found published records of comparative studies.
However, both the highest prevalence and the lowest prevalence were detected with SAF. For Diphyllobothrium spp., it is possible that we have missed detection in some of our study areas, although our results are consistent with the feeding ecology of coastal wolves.
Although unexpected, a negative relationship between dietary diversity and the fecal prevalence of parasites with indirect life cycles is plausible and could occur via multiple mechanisms. For example, diverse diets could reduce wolf exposure to parasitic species that occur in any single, dominant prey species, thereby reducing overall infection rates. Exposure to different parasites from multiple prey species could also reduce overall infection rates via interspecific competition among parasites in the digestive tract of wolves or other final hosts (Mideo, 2009). Regardless of the mechanism, our results appear consistent with the findings of Friesen and Roth (2016), who suggested that wolves' use of alternate prey such as beavers (Castor canadensis)-which do not serve as alternate host to many wolf parasites-may limit cestode intensity in wolves and ultimately reduce parasite density in local ungulates.
Although differences in wolf density, in the number of alternative hosts, and in wolf population history among study areas did not emerge as top predictors of the fecal prevalence of parasites with indirect life cycles, these variables may have contributed to the observed results in some study areas. For example, the high fecal prevalence of Taeniid tapeworms in wolves from YNP might reflect the recent re-introduction of wolves and associated changes in parasitehost communities (Molnar et al., 2019). Interestingly, Taeniid fecal prevalence was lower in wolves from MNP with a similarly recent history of recolonization, suggesting either that recolonization does not influence parasite fecal prevalence, that the process of recolonization by wolves and their parasites differed in these two populations (e.g., re-introduction of dewormed wolves in YNP vs. natural recolonization of wolves with their own parasite communities in MNP), or that other differences between the two areas-such as the higher diversity of prey species in MNP-influence Taeniid fecal prevalence. Finally, numerous biological and ecological variables, including combinations of variables (e.g., high wolf density and high prey density), could also influence the fecal prevalence of parasites with indirect life cycles and could be examined in further studies.
The relationship between dietary diversity and fecal prevalence of parasites warrants further study in part because of the potential implications for the long-term health of predator-prey relationships and ecosystem function. Although gastrointestinal parasites may have low pathogenicity in healthy, unstressed definitive hosts, the effects of parasites may differ when combined with stressors such as food shortages or human disturbance, including contaminants (Marcogliese & Pietrock, 2011). Therefore, minimizing parasite burden by ensuring wolves have access to alternative prey may be beneficial to wolves. In addition to affecting wolf health, lower prevalence of parasites in wolves with higher diet diversity may reduce parasite prevalence in prey species by decreasing the number of parasite larval stages spread via wolf feces (Joly & Messier, 2004).

| Environmental influences on the distribution of specific parasitic taxa
Our study revealed differences related to specific prey items for which access differs sharply among the six study areas. The most striking example was that samples from coastal marine habitats yielded Diphyllobothrium spp. and Metorchis spp., known to be associated with the consumption of marine resources (Bryan et al., 2012).
Diphyllobothrium spp. has also been reported in wolves from the Canadian Northwest Territories (Choquette et al., 1973;Schurer et al., 2016) and from Croatia (Hermosilla et al., 2017) where wolves presumably had access to marine or freshwater fish. Moreover, there are reports of wolves catching and consuming freshwater fish in continental areas far from the ocean. These include the Northwest Territories were Kuyt (1969) noted the use of lake trout (Salvelinus namaycush), northern pike (Esox lucius) and potentially other species as important for wolf diet in areas without access to caribou (Rangifer tarandus) during the denning season, and in Minnesota where wolves were recently reported to have consumed northern pike and possibly other fish species (Gable et al., 2018). This association between wolves and aquatic diets seems ancient (Meachen et al., 2020). Although it is unknown whether such behavior is common and shared with other wolves in the region, wolves can learn to hunt species to which they have never been exposed previously (Smith et al., 2000).
Detecting rare prey items in wolves, however, can be challenging, especially as dietary analysis of wolf feces typically focuses on mammalian prey items. As a key example, Alaria spp. was detected in feces from wolves in the Prairie-Boreal region and in Italy (PNALM).
Alaria typically uses species such as frogs and tadpoles (Anura), and snakes (Serpentes) as intermediate hosts, and might therefore represent ingestion of these prey items by wolves. In Europe, however, Alaria alata has been identified in tissues of wild boar, representing a potential source of infection for wolves (Bagrade et al., 2009;Strokowska et al., 2020

| Giardia
We also found differences in the fecal prevalence and composition of Giardia spp. between the Coastal and Prairie study areas.
Interestingly, prevalence of Giardia spp. was three and six times higher in RMNP and DMPPF, respectively, compared with the GBR.
Within the GBR, Giardia spp. prevalence was higher in samples collected from islands compared with those from the mainland (Bryan et al., 2012). Collectively, these findings suggest that landscape characteristics, such as the abundance of and access to freshwater, could play an important role in the distribution and prevalence of Giardia spp. in wolves, the frequency of transmission being facilitated by water (Thompson & Ash, 2016). Although the number of genotyped samples from each study area was modest, we found interesting differences in the composition of Giardia assemblages. Assemblages with zoonotic potential (A and/or B; reviewed in Thompson & Ash, 2016) were detected in both Pacific Coastal and Prairie regions, appearing to be broadly distributed in wolves and other wildlife in western Canada (Schurer et al., 2016;Thompson & Ash, 2016).
By contrast, assemblages thought to be specific to canids (C and D) were only detected in samples from the Prairies (Thompson & Ash, 2016). Future surveys could investigate possible signs of change in Giardia prevalence or assemblage composition, such as geographical expansion of canid-specific assemblages. Giardia infection might be a particular concern for transmission between wild and domestic species, which could have important zoonotic potential even at low prevalence, especially where wolf territories bring them into close contact with domestic animals and humans (Hermosilla et al., 2017).

| Diet and season
Importantly, seasonal differences often affect wolf diet and parasite prevalence, which have been found to vary among prey species, parasite species, packs, and year also in our study areas (Bryan et al., 2012;Ciucci et al., 2020;Sallows, 2007

| Coprological methods
Obtaining consistent results on parasite presence given irregular excretion of parasites can be difficult (Branda et al., 2006;Cartwright, 1999), even in heavily infected individuals (Marti & Koella, 1993). Moreover, direct comparison examining the same fecal samples with flotation and SAF methods has demonstrated that there can be significant differences in detection, where some taxa are better detected by SAF and others by flotation (Wolf et al., 2014). We were unable to perform direct comparison of the original samples from our earlier studies, and some parasite taxa may have been missed or under-reported because of methodological differences or because of sampling was done during a time when there was less shedding of parasite eggs. Instead, we used a simulation approach to adjust for potential biased introduced by methodological differences among study areas. Although the simulation was based on published literature whose methods, host species, and parasite taxa differed somewhat from ours (Appendix S1), the results suggested that our main findings would not be greatly affected by methodological differences and could be useful in future monitoring and research.

| CON CLUS IONS
Tracking parasite-host dynamics and their interaction with environmental conditions can contribute to the maintenance of healthy ecosystems in a time of rapid environmental change. A priority for monitoring programs is to investigate parasite transmission between wild and domestic species (Beck et al., 2017;Ciucci et al., 2020;Clark et al., 2018;Fuehrer et al., 2016;, particularly in response to environmental change (Acevedo-Whitehouse & Duffus, 2009;Scheffers et al., 2016). Further research could help clarify the extent to which naturally transmitted parasites (i.e., from natural wolf prey) versus anthropogenically transmitted parasites (transmitted from feeding on or contact with domestic animals or human refuse) influence the health of wolves and other carnivores. Recent findings also highlight the importance of monitoring recolonizing wolf populations, as wolf-specific parasites could increase in prevalence following the return of wolves (Lesniak et al., 2018). Parasites might thus contribute to re-establishing ecosystems toward a more naturalized state in ways not typically observed or reported by humans. Notably, the link we found between higher dietary diversity and reduced prevalence of trophic transmission (i.e., indirect life cycle) of parasites in wolves provides a novel line of evidence by which biodiversity may maintain ecosystem health (Civitello et al., 2015). Moreover, the positive relationship between wolf density and the prevalence of parasites with direct transmission provides support for a regulatory role of parasites in ecosystems. Combined, these findings highlight the importance of multihost-parasite associations in the maintenance of ecosystem health and underscore the importance of conservation and restoration strategies that promote not only species but also a diverse suite of trophic interactions and host-parasite associations.

ACK N OWLED G M ENTS
We are grateful to B. Wagner and R.C.A. Thompson

CO N FLI C T O F I NTE R E S T
No conflict of interest.