In his original description of the species, published in 1870, Robert Swinhoe, the English diplomat and naturalist who first introduced western zoologists to Hydropotes, noted how the now-extinct population on the large riverine islands above Chinkiang (Zhenjiang) on the Yangtze would live among the tall reeds. When the reeds were cut for thatching in the spring they'd depart for the hills, returning in the autumn with their young once the reeds had regrown:
“The rushes are cut down in the spring; and the Deer then swim away to the main shore and retire to the cover of the hills. In autumn, after the floods, when the rushes are again grown, they return with their young and stay the winter through.”
Bedfordshire naturalist Bernard Nau proposed an alternative scenario in 1992, however. Based on the apparent facileness with which water deer colonised the Home Counties, Nau suggested that open country is the primary habitat for this species, not wetlands. In China, he speculates, development of open country by humans may have driven water deer to seek cover wherever available, and reedbeds and swampy ground are less easily developed than countryside and woodland. I know of no evidence in support of Nau's theory and, in their 1998 British Deer Society booklet on the species, Arnold Cooke and Lynne Farrell point out that the distribution of water deer at Woodwalton Fen in Cambridgeshire, coupled with the nature, range and frequency of their vocalisations indicate the species is adapted to dense cover, such as reedbeds, seeming to prefer both wet and unwooded.
On the edge?
Most assessments of the water deer characterise them squarely as an edge species, meaning they tend to be found where habitat transitions from one type to another (e.g. woodland to grassland or reedbed to agricultural fields), preferring areas with tall grass, shrubs and small trees. In their native range, they occur in most coastal reed communities, salt marshes, creeks in low mountains and hills, lowland montane, along rivers and streams and even some river parks in Seoul. During 1982 and '83, Helin Sheng and Houji Lu surveyed water deer habitat requirements on the Zhoushan Archipelago, observing that their ideal habitat was areas where cogon grass (Imperata cylindrica) was abundant, but shrubs were sparse. In the Yangcheng Reserve between March 1999 and September 2000, Endi Zhang and colleagues found that, in spring, deer also showed a strong preference for habitat dominated by Imperata cylindrical as well as king grass (Zoysia macrostachya), common reed (Phragmites autralis), and coastal cat's foot (Aleuropus littoralis). During summer, they preferred habitat with seablites (Suaeda yhaura and Suaeda salsa) and common reed. Cat's foot, seablites, bulrush (Scirpus sp.), and common cordgrass (Spartina angelica) were preferred in autumn.
In England, Cooke and Farrell's studies have found them to be primarily a species of wet carr (i.e. waterlogged wooded terrain) and productive grassland, seemingly avoiding dry carr and being less fond of large tracts of woodland, preferring to use the latter as part of a larger home range for food and shelter/security. Indeed, the most recent (2018) Mammal Society assessment of the species in Britain estimated that nearly half (44%) of the population range is fen, marsh and swamp, with 30% being arable and horticulture, 19% broadleaf woodland, and only 7% unimproved grassland.
Similar assessments to the Mammal Society's in China have yielded much the same results. Huai Wang and Helin Sheng observed that water deer on the Zhoushan Islands preferred tall grass habitats (densities of up to 0.6 deer per ha) over bush areas (only 0.22 deer per ha), while a radio-tracking study conducted around the same time at Poyang Lake by Bing Xiao and Sheng found that nearly 80% of their subjects' home ranges were short grasses, with the remaining 20% tall grasses. The deer spent about 63% of their time in short grass, presumably feeding, and the remaining 37% in long grass resting. More recently, in 2006, Zhang and his team noted that water deer at Yancheng preferred areas within just of half a kilometre (0.3 miles) of water and with vegetation 90-110 cm (35-43 in.) tall, supporting at least 90% vegetation coverage in all three seasons of the study.
A team at Zhejiang Normal University, led by Yixin Bao, studied the seasonal change in habitat use by water deer in the Zhoushan Archipelago during 2008, finding significant differences between spring (March) and autumn (November). In spring the deer chose patchy broadleaf woodland (preferably with <25% tree cover), farmland and hillside meadows, within 200 m (656 ft.) of water. In autumn, the deer preferred south-facing woodland, shrub forest, farmland and the hillside meadows and were within 600m (0.4 mi.) of water. Human disturbance was a contributing factor in habitat selection. In their 2011 paper to Landscape Ecology and Engineering, Baek-Jun Kim and co-workers reported their observations on the distribution and habitat use of water deer on the Korean peninsula between 2000 and 2005. Their analysis showed the deer preferred rural inland areas at elevations below 300 m (984 ft.) and were found up to 1.6 km (1 mile) from water - fewer field signs were found in coastal areas and on islands.
Throughout their native and introduced ranges water deer also inhabit cultivated areas. In China and Korea they are often found in rice paddies, and in South Korea they will feed on arable crops during the winter, particularly vegetables such as soybean sprouts. At Woodwalton, Cooke and Farrell observed that, even in prime wetland, they seem to prefer to have access to adjacent farmland for times of the year when season, food or snow/ice/flooding makes the reedbed less hospitable and in their 1998 booklet note:
“For feeding they prefer sedge-dominated habitat where a greater diversity of food plants grow. Arable fields use include winter wheat, oilseed rape, field beans, carrots, potatoes and, latterly and particularly, set-aside.”
In China, Sheng and Lu's 1984 dataset show that while ideal habitat was cotton grass, the deer also moved out into nearby farm fields to feed. Similarly, a habitat suitability study by Jihyang Jung and colleagues in South Korea, published in Animal Cells and Systems in 2016, determined that they use forest for cover during the daytime and move out into open areas of grassland or arable fields at night to feed. In Buckinghamshire during the summer, I have observed water deer both resting in copses before emerging to feed on pasture and also resting among the grass itself during the daytime, and it is not uncommon to “put one up” while crossing a field.
Reedbed habitat is lacking in Buckinghamshire and Bedfordshire, and here the deer are attracted to mixed farmland, with sparsely wooded arable land including sloping fields apparently ideal habitat. In both counties I have observed them in reasonable numbers on newly sown spring grasses in the early part of the year. Water deer also seem to spend a lot more time than other deer species “in crops” without apparently eating much of them, particularly during the summer where they may give birth. It may be that cereal crops sufficiently resemble their native reedbeds to be attractive.
The high life
Mountainous habitats seem generally less preferable to water deer. Indeed, topography appears to have a significant impact on habitat selection, at least across some of their range. From May to September 2016, for example, Taw-Kyung Eom and colleagues at the Chung-Ang University investigated the habitat selection by water deer living in the low mountainous area of Maehwasan mountain in South Korea. They discovered that deer prioritised ridge sites with a predominance of understorey cover to rest in over food (despite plentiful shrubs and grasses), while most feeding activity occurred on the forbs and shrubs in the valley. On the mountain slope, deer again chose canopy cover over food availability.
Writing in Terrestrial Ecology and Behaviour in 2019, Eom and his team suggest that the more prominent wind along the ridge helps to reduce biting insects, while the wide views offered by the terrain may help deer detect predators at greater distances, explaining why the ridge was the preferred resting site, despite abundant understorey cover on the slopes and in the valley. The ridge also had more mid- and overstory cover than understory, which the deer might prefer for the shade it offers during the summer. Interestingly, an earlier study by the same group conducted during the winter found no such topographical differences, presumably because food was generally scarce so the deer couldn't be as selective.
At around the same time Eom and his co-workers were carrying out their research, a team led by Kwon Hyuk-Soo at the National Institute of Ecology in Seocheon were conducting a field study of mammal habitat use in and around the Dangjin City area of South Korea. The researchers found that water deer were most frequently associated with “fluctuant and concave topography”; that is to say the edge habitats of mountains. In their 2016 paper to Contemporary Problems of Ecology, Hyuk-Soo and his colleagues suggest that the cover afforded by this topography is necessary to help hide the deer from predators.
Finally, the presence of people and livestock influence Hydropotes' habitat selection. Observations by Arnold Cooke in Cambridgeshire, Sharon Scott in Buckinghamshire and myself in Bedfordshire suggest that water deer generally avoid fields with livestock, although some of the grazing marshes remain a major habitat for them in the south-east, and I have observed deer and cattle using the same fields at Strumpshaw and Buckenham Fen in Norfolk. Indeed, as Dr Cooke has pointed out to me recently:
“The point about grazing marshes being important despite water deer shunning fields with livestock is perhaps due to scale/situation. Outside Woodwalton Fen, water deer don't avoid the farms where cattle and sheep are grazed, but they do tend to avoid a field when livestock are present (usually at quite high densities).”
In Britain I know of no reports of water deer from urban areas, although the first record of the species to the east of the M1 motorway in Bedfordshire was an individual seen at a vehicle test track in Millbrook in 1975. Water deer are very infrequently reported in the gardens of even rural properties. Indeed, on their farm, Sharon and Graham Scott note that the water deer almost never enter their garden, despite probably being the most abundant deer in the area and both muntjac and roe regularly coming in. This does not appear to be the case throughout their range, however, and they seem more receptive to urbanisation in South Korea.
Baek-Jun Kim and his team recorded water deer field signs at 67% of the metropolitan cities (i.e. populations in excess of one million) they surveyed, while in their 2014 paper to the Journal of Wildlife Diseases, Seol-Hee Kim and colleagues note that “water deer are among the most common wildlife to approach farmhouses and livestock barns in Korea”. They don't actually quantify this statement, but they do present their survey data that recorded water deer at apartment complexes and factories in Ulsan City in the south-east of the country. Similarly, the wildlife survey by Hyuk-Soo and co-workers showed water deer in habitat near residential areas, between 200 and 600 metres (220-660 yds) of buildings and Jihyan Jung and his unit observed that deer in South Korea avoided roads, but were seemingly less bothered by nearby urbanisation. I have seen video footage of water deer on industrial estates, and even at logging sites just outside urban areas, in South Korea and, interestingly, Korean naturalist Junha Kim told me that the water deer's Korean name, Gorani, is often used on Korean websites as slang for pedestrians, cyclists, and skateboarders who cause traffic accidents with their careless behaviour.
I know of no direct observations of water deer from inside cities or towns in China. Indeed, the highest abundance of field signs found by Endi Zhang and colleagues during their Yancheng survey were more than a kilometre (0.6 miles) from human settlements and activity, while Yixin Bao and his team observed that in spring deer at Zhoushan were within 100 m (330 ft.) of human disturbance, while in the autumn it was 200 m. A project by researchers at the East China University, however, has succeeded in reintroducing water deer to several parks in the suburbs of Shanghai.
Overall, as a species we can say that water deer certainly seem capable of adapting to a wide range of surroundings, from waterlogged reedbed to suburban parks. Here in England, however, they appear to occupy a smaller subset of these habitats, found primarily in carr, marshland and farmland, seemingly much less fond of woodland, urban and suburban areas. It will be interesting to see, as the water deer's distribution in Britain progresses north and westwards, whether they become accustomed to the higher level elevations of Wales and the Lake District. The extremely wide-ranging habitat use that has been recorded in South Korea might be at least partially a reflection of apparently remarkable population density within the country forcing dispersal into habitats that might not otherwise be actively selected for.