Diet of fallow deer suggests potential for invasion of novel habitats in Tasmania

Introduction

Large herbivores have strong effects on the structure and composition of vegetation in many ecosystems (Burkepile and Parker 2017). Information on the diets of large herbivores is essential to understand and manage their impacts on vegetation (Holechek et al. 1982). This information is especially important in forecasting the diversity of habitats that might be occupied by invasive large herbivores and anticipating the potential impacts of those species on native biodiversity (Clavel et al. 2011; Slatyer et al. 2013). A flexible diet should allow herbivores to invade a wide range of habitats and to have significant impacts on vegetation in novel habitats (Lurgi et al. 2014).

Deer are among the most successful of invasive large herbivores worldwide (Genovesi 2005; Chakanya et al. 2016). The European Fallow deer (Dama dama dama) originally occurred in a small part of Anatolia (Masseti et al. 2008; Wilson and Mittermeier 2011). It is now one of the most widely distributed cervids globally (Chapman and Chapman 1980), having established populations in much of western and southern Europe (Genovesi 2005) as well as North America, New Zealand, South Africa and Australia (Chapman and Chapman 1980; Chakanya et al. 2016; Esattore et al. 2022). The preferred habitat of fallow deer is a mosaic of grassland and forest with a grassy understorey (Chapman and Chapman 1975). The species is categorised as an intermediate feeder, consuming both grass and browse (Chapman and Chapman 1975; Kerridge and Bullock 1991; Spitzer et al. 2020). Although its food preferences may be wide (Esattore et al. 2022), it generally shows a higher preference for grasses (Chapman and Chapman 1975; Hofmann 1989), especially during the growing season (Spitzer et al. 2020).

Fallow deer were introduced to Tasmania (and to other parts of Australia) in the early nineteenth century (Bentley 1998; Hall and Gill 2005). The initial introductions of deer in Tasmania were to grassy woodlands in the low-elevation, warm and dry Midlands of Tasmania. These habitats resemble the preferred habitats of fallow deer elsewhere in their global geographic range. Initially, the Tasmanian population grew slowly. By the 1970s, its size was estimated to be ~8000 animals (Chapman and Chapman 1975). Since the mid-1980s, however, the population has been increasing at 11.5% per year and extending its distribution such that it currently occurs over 27% of Tasmania (Cunningham et al. 2022), including areas with high conservation value such as the Tasmanian Wilderness World Heritage Area (TWWHA). A 2019 aerial survey estimated the population in the eastern part of the state to be ~53 000 (Lethbridge et al. 2020). Cunningham et al. (2022) suggest that because the survey was conducted late in the 2019 hunting season, during which approximately 30 000 deer were killed (Lethbridge et al. 2020), the population size was at least ~83 000 at its 2019 seasonal peak.

Very little is known about the diet of fallow deer in Tasmania, or elsewhere in Australia. Rumen-content analysis of a population in New South Wales, Australia found that grass ranged from 99% of diet composition in agricultural areas to 66% in more densely wooded areas, the balance of the diet consisting of browse (Calleja 2001). Davis et al. (2023) also found that fallow deer ranged from a grazing to a browsing diet that included eucalypts. In Tasmania, Duncan (1992) used histological analysis of faecal material to estimate dietary overlap of eastern grey kangaroos (Macropus giganteus) and fallow deer in the Midlands of Tasmania. Both species consumed grasses as the main component of their diet. However, kangaroos had a higher proportion of grasses in their diet than fallow deer (61% versus 46%), but fallow deer ate a greater diversity of plant species overall, including a higher proportion of shrub species.

In the absence of management intervention, such as the 2023 aerial deer cull in the TWWHA (Powell 2023), fallow deer are predicted to further expand their abundance and distribution in Tasmania; there are large areas of the island with apparently suitable climate and habitat but which are not yet occupied (Potts et al. 2015; Cunningham et al. 2022). The portions of Tasmania that are not currently occupied by fallow deer include areas with high conservation value, which support many vegetation types that differ from the grassy woodlands of the Midlands. In ungulates in general, behaviours related to habitat selection are largely driven by satisfying dietary needs (Spitzer et al. 2020). Therefore, the degree to which the Tasmanian fallow deer population will be able to increase its distribution will depend partly on its ability to broaden its dietary niche so that it can occupy a wider range of habitats, especially those that differ from the grassy woodlands in which deer currently occur.

In this study, we aimed to describe variation in diet of fallow deer in Tasmania by comparing diets of deer in grassy woodlands in the dry lowlands of Tasmania with those of deer in highland woodlands and forest with a cooler and wetter climate. Our grassy woodlands study area is typical of areas in which fallow deer have been well established in Tasmania since the nineteenth century (Bentley 1957) and represents habitat that has been generally considered to be preferred by the species (Chapman and Chapman 1975). The cooler highland study areas have more varied and dense vegetation with less grass and a higher representation and diversity of woody species. These study areas are typical of environments into which fallow deer have spread more recently (Wapstra 1973).

We expected that in the lowland grassy woodlands, deer would eat mostly grasses and forbs; our study was designed to test the extent to which the diet differed in contrasting high-elevation habitats, especially by including a higher diversity of plant species and increased browsing on woody vegetation.

We described diet by metagenomic analysis of faecal material. The metagenomic approach was chosen to maximise information on the taxonomic composition of the diet, identified to the finest possible level, and to identify taxa that might be strongly affected by herbivory in the highlands. Metagenomic analysis provides more highly resolved information on herbivore diets than traditional methods such as histological examination of faecal material (Pompanon et al. 2012), and has been used successfully in previous analysis of faecal material in deer (Czernik et al. 2013; Nakahara et al. 2015; Erickson et al. 2017).

Materials and methods

Study area

We investigated diets of fallow deer at four study areas in Tasmania, one (Beaufront) in the low-lying Midlands and three (Five Rivers, Silver Plains and Arthurs Lake) in highland areas in central Tasmania (Fig. 1). Beaufront is a private grazing property. Five Rivers and Silver Plains are reserves owned and managed by the Tasmanian Land Conservancy. Arthurs Lake is within the Great Lake Conservation Area and managed by the Tasmanian Parks and Wildlife Service. Beaufront is at an elevation of ~ 300 m ASL (above sea level), and its climate is characterised by low rainfall (mean annual 489 mm) and moderate temperatures (mean annual daily minima and maxima 5.6°C and 17.7°C respectively). The three highland study areas are 800–1000 m ASL, with climates that are wetter (mean annual rainfall 840–940 mm) and cooler (mean daily minima 1.6–3.4°C, maxima 10.3–13.9°C).

Fig. 1.

Location of study areas (black circles) used for faecal pellet collection. Black line depicts the Tasmanian Wilderness World Heritage Area. Red line depicts the distribution of fallow deer in central Tasmania from Cunningham et al. (2022). Satellite populations are not shown on map.

The vegetation on Beaufront is a mixture of open grasslands (used to graze sheep and cattle) and remnant open eucalypt woodland. We sampled in an area of the property that was predominantly open woodland over unimproved pasture dominated by native species. The canopy tree species were widely spaced white gum (Eucalyptus viminalis) and black peppermint (E. amygdalina) over a sparse and patchy mid-storey of silver wattle (Acacia dealbata), silver banksia (Banksia marginata) and black she-oak (Allocasuarina littoralis). The ground level vegetation was dominated by native grasses (kangaroo grass (Themeda triandra), tussock grasses (Poa spp.), spear grasses (Austrostipa spp.) and wallaby grasses (Rytidosperma spp.)), some introduced grasses and a high diversity of forbs, both native and introduced. Dense patches of sagg (Lomandra longifolia) and bracken fern (Pteridium esculentum) also occurred in places. The vegetation communities as defined by Kitchener and Harris (2013) were Eucalyptus viminalis grassy forest and woodland (DVG) within a mosaic of lowland grassland complex (GCL) and Lowland Poa labillardierei grassland (GPL).

The three high-elevation study areas consisted of forested areas dominated by gum-topped stringybark (E. delegatensis), cabbage gum (E. pauciflora) and swamp peppermint (E. rodwayi), growing over a dense mid-storey of shrubs dominated by small-fruit hakea (Hakea microcarpa). These forested areas were variously interspersed in some places by patches of alpine and sub-alpine heath, most commonly dominated by mountain pinkberry (Leptecophylla parvifolia), as well as coral heath (Epacris gunnii) and lowland richea (Richea procera), and in other places by tussock grasslands dominated by Poa and Rytidosperma species. These grasslands also contained a variety of both native (Cranesbill species (Geranium spp.), raspwort species (Gonocarpus spp.), button weeds (Leptinella spp.)) and introduced forbs (Hypochaeris radicata, Leontodon saxatilis). The vegetation communities represented in these high-elevation study areas were Eucalyptus delegatensis dry forest and woodland (DDE), Eucalyptus rodwayi forest and woodland (DRO), Eucalyptus pauciflora forest and woodland on dolerite (DPD) and Eucalyptus pauciflora forest and woodland not on dolerite (DPO), interspersed with Highland Poa grassland (GPH), Highland grassy sedgeland (MGH), Eastern alpine sedgeland (HSE) and Eastern alpine heathland (HHE) (Kitchener and Harris 2013).

Results

Fallow deer consumed a wide range of plant species across all study areas. This included representatives of bryophytes (5% of samples), ferns (7.5%), orchids (2.5%) forbs (87.5%), grasses (30%), sedges (35%), rushes (17.5%), shrubs (50%) and trees (70%). Aquatic plant taxa were present in 10% of samples. Both introduced and native plant species were eaten. In all, 52 genera from 30 families were detected. Forty-four taxa were identified to species level (Supplementary Table S1). The most common taxa were Asteraceae at the family level (37.5% of all scats), Eucalyptus (60%) at the genus level and Gonocarpus tetragynus (50%) at the species level. The shrub small-fruit Hakea (Hakea microcarpa) was detected in 15% of faecal pellet samples (Supplementary Table S1).

The composition of the diet varied strongly between the low-elevation Midlands study area and the highland study areas, which were variable but overlapping with one another in the NMDS ordination space (Fig. 2). The only significant dissimilarity among study areas in ordination scores were between Beaufront and the other three study areas (ANOSIM, Beaufront versus Arthurs Lake R = 0.4404, P < 0.001; Beaufront versus Five Rivers R = 0.3300, P < 0.001; Beaufront versus Silver Plains, R = 0.3593, P < 0.001). The taxa most responsible for causing the differences in positions in the ordination space were the genera Oxalis, Cerastium and Eucalyptus, and the Family Asteraceae (Fig. 2). The first ordination axis was characterised by a shift from high representation of forbs (Asteraceae, Oxalis and Cerastium) at one end, to trees (Eucalyptus) and subshrubs (Gonocarpus) at the other (Fig. 2). The three Central Plateau study areas had significantly more scats with the tree genus Eucalyptus and the subshrub Gonocarpus tetragynus (Table 1).

Fig. 2.

Non-metric multidimensional scaling (NMDS) ordination plot of community dissimilarities based on the plant species found in deer faecal pellets (2D stress value = 0.214). The shorter the distance between two samples, the higher the similarity between the scats. Significant (P < 0.05) vectors of maximum correlation are shown.

Fig. 3 summarises diet composition in the four study areas in terms of major plant life forms. This shows that at Beaufront, herbs were most commonly recorded in scats, and trees were relatively uncommon; trees and shrubs were more commonly recorded in scats at the highland sites, especially at Silver Plains and Arthurs Lake, and at one highland site (Arthurs Lake) grass was not detected (Fig. 3).

Fig. 3.

Histogram of most common plant lifeform categories represented in the diet of fallow deer across four study areas in central Tasmania.

Discussion

Most studies of fallow deer diet have been conducted in Europe, followed by North and South America (Esattore et al. 2022). Our results provide replication at a finer taxonomic resolution for the small number of published Australian studies (Duncan 1992; Calleja 2001; Davis et al. 2023). We confirm that fallow deer has a wide dietary niche, consuming plants from a broad range of taxa and growth forms. We also show that there are large differences in diets of deer occupying contrasting habitats in Tasmania. This versatility of feeding ecology should mean that fallow deer will be able to occupy a broad range of vegetation types in Tasmania, shifting their diets locally in response to availability of different kinds of plants. Because of this versatility, deer appear capable of substantially extending their current distribution in Tasmania to large areas of the north and the south-west of the island (Cunningham et al. 2022). The broad dietary niche of fallow deer also means they could have significant impacts on vegetation communities in those regions.

The taxonomically wide diet of deer in Tasmania is consistent with a global literature that demonstrates that, although typically a grazer (Chapman and Chapman 1975; Hofmann 1989), fallow deer are versatile in their dietary choices (Chapman and Chapman 1975; Jackson 1977). The species can shift its diet in response to availability of different kinds of forage in different habitats or seasons (Spitzer et al. 2020; Esattore et al. 2022). For example, in the Czech Republic, under heavy snow cover, fallow deer switch from a diet of brambles and grasses to one of mostly low-quality spruce needles (Kamler and Homolka 2011). The differences in diet that we observed reflect availability according to vegetation structure. In the open grassy woodlands of the low-elevation woodlands, the scats of fallow deer contained predominantly forb and grass taxa, but in more densely wooded highland habitats with less grass and more heath or shrub vegetation in the understorey, the scats had a high proportion of shrub and tree taxa. We studied diet at the end of the winter, when plant growth (and therefore forage availability) was likely to be at an annual low point, especially for grasses and herbs. It is therefore likely that feeding on grasses would have increased during the warmer months, as has been observed in Italy (Bruno and Apollonio 1991).

Much of Tasmania is within the range of climate tolerance of fallow deer, with 56% of Tasmania estimated as highly suitable (Cunningham et al. 2022). However, it is possible that deer would have greater difficulty surviving in the oligotrophic environments of western Tasmania, where grazing animals are extremely sparse. Fallow deer in Tasmania thus have the potential to have negative effects on many plant species and distinctive vegetation communities. For example, several species of ancient endemic conifers, such as pencil pine (Athrotaxis cupressoides) occur in the Tasmanian highlands. Athrotaxis cupressoides is a slow growing species (Worth et al. 2017) and has been historically reduced by browsing by marsupial herbivores interacting with fire (Kirkpatrick and Bridle 2007; Holz et al. 2015). The grassy ecosystems of eastern and central Tasmania contain a high proportion of threatened taxa, including orchids, daisies, grasses and sedges (Quarmby et al. 2023). All of these types of plants were represented in the diet of Tasmanian deer.

The fact that Eucalyptus species were commonly eaten by fallow deer is concerning because Eucalypt woodlands are in decline across much of the lowland range of fallow deer. Eucalypt woodlands in the Tasmanian Midlands are already threatened (Kirkpatrick et al. 1995; Cowell et al. 2013) and in decline (Neyland 1996; Prior et al. 2013). Attempts to restore woodland patches in the Midlands (Davidson et al. 2021) are severely hampered by deer browsing. For example, Bailey et al. (2015) observed that 45% of Eucalyptus tenuiramis trees used for a restoration planting project in Tasmania were damaged by fallow deer after 3.25 years, and they cited damage from fallow deer as a significant impediment to cost-effective restoration.

In the higher elevation study areas of the Central Plateau, Eucalyptus species may experience even higher browsing pressure from fallow deer. Impacts on some species are potentially significant. For example, the Tasmanian endemic Eucalyptus gunnii has suffered increased dieback over the last decades, attributed to increased droughts associated with a warming climate (Calder and Kirkpatrick 2008). However, Calder and Kirkpatrick (2008) demonstrate that the most severe dieback is also related to heavy possum browsing pressure. The endangered E. gunnii subsp. divaricata is restricted to the eastern Central Plateau, in proximity to the Silver Plains and Arthurs Lake study areas used for sampling in the present paper, where our results show high prevalence of browsing on Eucalyptus species.

Of the taxa discovered in the diet, 14 were introduced species. Most of these species are common pasture weeds e.g. Alopecurus pratensis, Cirsium palustre, Leontodon saxatilis. The Midlands study area had the highest proportion of introduced species in the diet. This is most likely to be related to the higher level of disturbance in the landscape than in the higher elevation study areas. Plant germinants emerging from fallow deer scat are 57% greater per scat than those emerging from kangaroo scat (Claridge et al. 2016). Considering the wide dietary preferences demonstrated in this paper, it is likely that deer will act as weed vectors as they continue to colonise currently unoccupied environments.

Fallow deer have been shown to consume ferns, e.g. Dryopteris filix-mas and Pteridium sp. (Jackson 1977; Kerridge and Bullock 1991), although research that has quantified the proportions of plant species in the diet suggests that ferns are only consumed in low amounts, such as <1% (Davis et al. 2023), 5% (Jackson 1977) and 10% (Trabalza Marinucci et al. 2010). We report an even lower proportion, with each of the three species of fern detected being found in one deer scat each.

In New Zealand, at a similar latitude to temperate Australia, Nugent (1990) concluded that browsing by fallow deer in beech forests would inhibit regeneration of subcanopy tree species when densities of deer exceed 10 deer per km². Fallow deer have a preference for young seedlings and coppiced trees in Australia (Hall and Gill 2005). Given the New Zealand and Australian observations, it is likely that browsing by fallow deer could limit recruitment of woody plants in some Tasmanian plant communities, especially given a potential for the populations to grow (Potts et al. 2015; Cunningham et al. 2022; Botterill-James et al. 2023). Fallow deer currently occur at a mean density thought to be 2.7 deer per km² within their core range in Tasmania (Lethbridge et al. 2020), compared with levels on one site on the Australian Mainland of 39.3 fallow and red deer per km² (Bengsen et al. 2023). In Budj Bim National Park, Victoria, the density of fallow deer has been estimated as 17.6 deer/km² (Woodford et al. 2023).

Conclusions

Our study confirms that fallow deer in Tasmania have a varied diet. Therefore, it is unlikely that the distribution of certain vegetation types, or availability of a limited range of plant species, will constrain the distribution of fallow deer in Tasmania. The capacity of fallow deer to browse as well as graze indicates a need for further research to identify potential impacts on woody vegetation. At high densities of deer, these impacts could include decline or disappearance of Tasmanian endemic plant taxa, further decline of grassy woodlands and significant changes to the structure of other vegetation communities.

Supplementary material

Supplementary material is available online.