Distribution of the Australian Water-rat Hydromys chrysogaster in Victoria: findings from community-based sightings and live-trapping surveys.
The Australian Water-rat or Rakali Hydromys chrysogaster is the largest and arguably most specialised Australian rodent in terms of its tooth structure and other physical features (Watts and Aslin 1981). Weighing as much as 1275 grams (McNally 1960), Water-rats feed mainly in the water on fish, insects, crustaceans, waterbirds, molluscs and (to a lesser extent) frogs and turtles; selected aquatic plants (notably the floating water fern Azolla filiculoides) also can contribute to the diet (Woollard et al. 1978). In addition, Water-rats have been documented to feed opportunistically on terrestrial prey, such as House mice Mus musculus during a mouse plague; cannibalism also can occur (Woollard et al. 1978). The Water-rat's ability to swim efficiently and capture aquatic prey is facilitated by having an elongated, streamlined body with small external ears that can be flattened tightly against the skull, broad hindfeet with partial webbing between the toes, water-repellent fur, a thick and well-furred tail, and a blunt, otter-like muzzle furnished with dense whiskers (Watts and Aslin 1981) (Fig. 1).
Menkhorst (1995) reported that Water-rats are widespread in Victoria, occupying saline environments (such as beaches in Port Phillip Bay) as well as rivers, creeks, irrigation channels and natural and man-made lakes. Though the Water-rat is unable to thermoregulate efficiently at water temperatures below 20 (o) C (Fanning and Dawson 1980), the species occurs up to an altitude of at least 1500 m in Mount Buffalo National Park. Menkhorst (1995) also noted that Water-rats often occur at sites where dense vegetative cover occurs at the water's edge in the form of thick grass, riparian scrub or reed beds. Similarly, results of live-trapping and radio-tracking studies carried out in southwestern Western Australia indicate that Water-rats are most likely to utilise habitats characterised by low-growing dense vegetation on water banks (Speldewinde et al. 2013) or by stable banks and substantial vegetation growing in and near water (Smart et al. 2011). Studies in both Western Australia and Queensland have concluded that Water-rats prefer to forage in relatively shallow water bodies, less than about 2 m deep (Harris 1978; Speldewinde et al. 2013). The species is appropriately characterised as active and mobile: Water-rats have been documented to travel 3.1 km along a stream channel in just 5.5 hours (Gardner and Serena 1995), and at least 3.0 km (though possibly 4.5 km) overnight across dry land to reach a man-made dam (Vernes 1998).
In contrast to most Australian rodents, Water-rats are commonly visible during the day (Watts and Aslin 1981) and are relatively large with distinctively white-tipped tails. This means that opportunistic sightings of the species can be used as a cost-effective technique to help map where Water-rats occur.
The main aim of this study was to summarise where Water-rat sightings were recorded in Victoria in the period from 2000-2017. This information was then used to address the following questions: (1) Does the frequency of Water-rat sightings vary regionally across Victoria? (2) How does regional variation in Water-rat sightings compare with regional variation in Water-rat live-trapping captures? In addition, the types of habitat where Water-rats were observed and factors contributing to Water-rat mortalities are described.
Records of sightings
We recorded pertinent details (locality, date and, in the case of carcasses, the cause of death if this was evident) of 804 first-hand reports of Water-rat sightings occurring in Victoria from 2000-2017 (including 247 sightings reported from 2000-2009 and 557 sightings reported from 2010-2017). The standard criteria for accepting a sighting record normally consisted either of photographic evidence or confirmation that the animal in question had a white-tipped tail and otherwise conformed to the expected size, appearance and behaviour of a Water-rat. In practice, less than 2% of all submitted sighting records (N = 12) were excluded on the grounds that they were of questionable validity or involved a different species being seen (mainly either a smaller rat lacking a white-tipped tail or Common Ringtail Possum Pseudocheirus peregrinus). To augment sample size, 118 Water-rat records obtained in Victoria from 2000 to 2017 and held independently by the Atlas of Living Australia (ALA) were also included in the study (ALA website). We also confirmed that the ALA database contained all relevant records held by the Victorian Bio diversity Atlas as of 30 June 2017.
With three exceptions, the names and boundaries of river basins used in the current analysis are as defined by Department of Water Resources Victoria (1989). The exceptions are: (1) the Murray River and its minor stream tributaries upstream of Lake Hume were excluded from the Mitta Mitta River basin (and instead were deemed to comprise part of 'Murray East', see below); (2) south-flowing river catchments from the Bemm to the Genoa Rivers were collectively referred to as the 'Far East Gippsland' basin; (3) the Little River was included within the Werribee (rather than Moorabool) River basin (Fig. 2). Water-rat sightings recorded along the Victorian portion of the Murray River or in its anabranches and minor tributaries (including the Gunbower Creek system and
Little Murray River) were grouped according to whether they occurred in or west of Echuca ('Murray West') or east of Echuca ('Murray East'). Sightings recorded in the Gippsland Lakes or other saline coastal localities (including Port Phillip Bay and Western Port) were deemed to be 'Coastal. The 29 river basins, two Murray River segments and coastal sites were then allocated to six geographic regions:
* North-west: Murray West, Campaspe, Loddon, Avoca, Wimmera-Avon, Mallee
* North-east: Murray East, Mitta Mitta, Kiewa, Ovens, Broken, Goulburn
* Melbourne: Bunyip, Yarra, Maribyrnong, Werribee
* South-west: Moorabool, Barwon, Lake Corangamite, Otway Coast, Hopkins, Portland Coast, Glenelg, Millicent
* South-east: Far East Gippsland, Snowy, Tambo, Mitchell, Thomson, Latrobe, South Gippsland
* Coastal: Gippsland Lakes, all other coastal sites (including Port Phillip Bay and Western Port)
Data relating to the incidental captures of Hydromys in fyke nets deployed by Australian Platypus Conservancy (APC) staff for Platypus Ornithorhynchus anatinus survey or monitoring purposes were summarised for the period from January 2000 through to June 2017. At each site, a pair of nets (one facing upstream and one facing downstream) were set in the afternoon following methods outlined in Serena and Williams (2012) and monitored at intervals of 2-4 hours through the entire following night. A Water-rat was deemed to have been captured if an animal was encountered in a net or if one or more holes were recorded of the type characteristically appearing when a Water-rat chews its way through netting to escape. In practice, it was rare for only a single hole to appear, presumably because a Water-rat typically tries to snip through netting at several points before settling down to create a gap large enough to accommodate its entire body. Capture frequency for a given water body was calculated as the number of live-trapping sites where a Water-rat and/or holes were recorded divided by the number of site-nights of live-trapping effort (one site-night = two fyke nets set to sample a site overnight). Estimates of capture frequency based on holes were potentially subject to underestimation insofar as two or more individuals may have created holes at a given site. However, given that there is no reason to believe that the amount of bias differed consistently between regions, this index presumably provides a reasonably valid basis for comparing capture frequencies between areas.
Mean (or average) capture frequency is reported plus or minus SEM (standard error of the mean). The hypotheses that mean Water-rat capture frequency in western Victoria differs from that in the rest ofthe state and mean capture frequency in south-eastern Victoria differs respectively from those in north-eastern Victoria and the Melbourne area were tested statistically using two-sample t-tests (with significance set at 0.05) after applying the arcsine-transformation to the frequency data to correct for possible non-normality due to their proportional nature (Zar 1984).
Regional distribution of sightings
Water-rats were reportedly seen in the period from 2000-2017 in all Victorian river basins apart from the Lake Corangamite basin (which is dominated by more or less saline lakes and their small inflowing streams: Department of Water Resources Victoria 1989) and the very dry Mallee and Millicent basins in far western Victoria. The absence of sightings in the Lake Corangamite, Mallee and Millicent basins may also have been influenced by the relatively low numbers of persons living in each of these areas. Based on the combined APC and ALA databases, the highest number of regional sightings was reported by those living in the densely populated Melbourne area (22% of all records), followed by north-western Victoria (20%), coastal habitats including the Gippsland Lakes (17%), north-eastern and south-western Victoria (16% each) and lastly south-eastern Victoria (9%) (Fig. 3).
Distribution of sightings in north-western Victoria
The largest proportion of sighting records for north-western Victoria originated in the Loddon River basin (41%), followed by the Murray River (23%) and Campaspe River basin (16%). In the Loddon basin, Water-rats were reportedly seen upstream of Cairn Curran, Laanecoorie and Tullaroop Reservoirs in the Loddon River proper and Jim Crow, Creswick, Tullaroop/Deep and Bet Bet Creeks. Sightings also were recorded at Laanecoorie Reservoir and particularly at Lake Daylesford, which was the source of 12% of Water-rat records for this basin. Downstream of Laanecoorie Reservoir, multiple sightings were reported in the Loddon River at Bridgewater, along Serpentine Creek, in the Dingee irrigation district, and along Bendigo Creek; 13% of records in the Loddon basin occurred at Lake Weeroona in Bendigo township.
In and near the Murray River, Water-rat sightings reportedly occurred from Echuca to as far downstream as Mildura. The species appeared to be particularly widespread and abundant in the Gunbower Creek system, with eight reports received for Gunbower Creek and over 30 reports received for associated lagoons and irrigation channels. However, a number of Gunbower residents also mentioned that the frequency of Water-rat sightings has dropped noticeably in recent years after local irrigation channels were lined with thick plastic sheeting in a bid to reduce seepage. Water-rats also were seen in several lakes on the Murray floodplain, including Reedy Lake and Kangaroo Lake, with multiple sightings made at Lake Charm and Little Lake Boort.
In the Campaspe River basin, Water-rats were reportedly seen along the Campaspe River (at sites distributed from Woodend downstream to Echuca) and the Coliban River (from Malms-bury downstream to a site near Redesdale) and in Lake Eppalock. One report described the species as being particularly common in the Campaspe River near Elmore, with at least 20 sightings made annually by the informant while fishing. Although only five persons reported seeing Water-rats in the Avoca River basin, these sightings were widely distributed along the Avoca River (at Avoca township, Archdale, Coonooer Bridge, Charlton, and below Quambatook). In the Wimmera basin, Water-rats were recorded at locations along the Wimmera River from Elmhurst downstream to Dimboola. In addition, sightings were reported at Mount William Creek (two records), Concongella Creek (one record) and at Lake Lonsdale (three records) and Lake Fyans (two records). No Water-rat sightings were reported for the Avon-Richardson River system.
Distribution of sightings in north-eastern Victoria
The largest proportion of Water-rat sighting records for north-eastern Victoria originated in the Goulburn River basin (45%), followed by the Murray River (23%) and Ovens River basin (17%). The frequency of sightings in the Goulburn system generally increased in the downstream direction, with 13% of records in this basin obtained upstream of Lake Eildon (in the Goulburn, Howqua, Jamieson, Big and Taponga Rivers and Mansfield's Ford Creek), 38% obtained from Nagambie to Lake Eildon (in the Goulburn, Rubicon, Acheron and Yea Rivers and Kilmore and King Parrot Creeks) and 48% occurring downstream of Nagambie. A large percentage of the Water-rat sightings in the last group comprised sightings for either Victoria Lake Park in Shepparton (18% of all sightings for the Goulburn basin) or channels in the Shepparton-Tatura irrigation district (15% of sightings). In addition, sightings reportedly occurred downstream of Nagambie in the Goulburn River, Lake Bartlett (in Tatura), Craig Muir Lake (in Mooroopna) and Pranjip, Faithfull and Honeysuckle Creeks. A number of persons living near Shepparton or Tatura reported that the practice of lining irrigation channels with plastic had resulted in a marked local reduction in Water-rat sightings.
Along the Murray River, 26% of Water-rat records were for locations distributed from Hume Dam to as far upstream as the Murray Gorge in Kosciuszko National Park, 68% for sites located from below Hume Dam to Lake Mulwala, and 8% for sites located between Lake Mulwala and Echuca. Water-rats were recorded upstream of Hume Dam along the Murray River and in Thowgla, Corryong, Nariel and Koetong Creeks. Downstream of Hume Dam, the species was seen in Lake Mulwala and a number of lagoons, backwaters or anabranches of the Murray River (Brown's Lagoon, Horseshoe Lagoon, Wonga Wetlands and Wodonga Creek at Albury-Wodonga, Torgannah Lagoon near Koonoomoo, Lake Moodemere near Wahgunyah). Water-rats were also seen in Wodonga in Sumsion Gardens Lake and several small creeks (Fell Timber Creek, Huon Creek, House Creek), and in Barmah National Park.
Elsewhere in north-eastern Victoria, Water-rats were recorded along the Ovens River at sites distributed from Harrietville to Whorouly, and also at Lake Catani (Mount Buffalo National Park), in the King River upstream of Lake William Hovell (on the Wabonga Plateau) and in Hurdle Creek (at Carboor Upper and near the King River confluence), Tea Garden Creek (near Milawa), Sheep Station and Silver Creeks (near Beechworth) and One Mile Creek (near Wangaratta). In the Broken River basin, Water-rats were most often recorded at Benalla (in the Broken River, Holland Creek and Lake Benalla), but also in Broken Creek (at sites located from Goorambat downstream to Nathalia), Nine Mile Creek and Sandy Creek. In the Kiewa River basin, the species was seen at Pretty Valley dam (near Falls Creek) and Sandy Creek Upper Reservoir, in Kinchington Creek, and in the Kiewa River near Tawonga and at sites along its east and west branches. In the Mitta Mitta River basin, Water-rats reportedly occurred upstream of Lake Dartmouth in the Big, Gibbo and Dart Rivers and Livingstone Creek. The species was also seen in the Mitta Mitta River near Lake Hume.
Distribution of sightings in the greater Melbourne region
Across the greater Melbourne area, the largest proportion of Water-rat sightings occurred in the Yarra River basin (67%), followed by the Werribee River basin (22%). In the Yarra catchment, the frequency of sightings generally increased in the downstream direction, with 7% of records obtained along the river and streams located upstream of Yarra Glen, 45% occurring between Yarra Glen and Dights Falls, and 48% occurring downstream of Dights Falls. Three or more sightings were reported for eleven named water bodies: the Yarra River (22% of all sightings for the basin), Gardiners Creek (15%), Moonee Ponds Creek (15%), Merri Creek (9%), Toorourrong Reservoir (4%), Olinda Creek (mainly near Lilydale Lake, 4%), Mullum Mullum Creek (4%), Diamond Creek (4%), Darebin Creek (2%), Edgars Creek (2%) and Albert Park Lake (2%). Other notable records included one sighting made in the Fitzroy Gardens and one at Richmond train station, where a Water-rat was observed scavenging for food dropped by commuters.
In the Werribee River basin, 31% of sightings occurred along the Werribee River at sites distributed from Ballan to just upstream of Port Phillip Bay. In addition, Water-rats were recorded along Skeleton Creek at Point Cook, Hoppers Crossing and Tarneit (20% of all sightings for the basin), at man-made lakes associated with housing developments at Point Cook (20%), in ponds at the Werribee Sewage Treatment Plant (13%), along Kororoit Creek (7%) and along the Little River (4%). Single sightings were also recorded at Cherry Lake in Altona and along Djerriwarrh Creek. In the Maribyrnong River basin, Water-rat sightings were reported at sites distributed along the Maribyrnong River from Keilor East downstream to Yarraville, along Jacksons Creek in Organ Pipes National Park and Emu Bottom Wetlands, and in Riddells Creek at Wybejong Park.
In Melbourne's south-eastern outskirts, Water-rat sightings were recorded in Cardinia Creek, Monbulk Creek at Belgrave, the Elwood Canal and on the Mornington Peninsula (at Kananook Creek, Paterson River, Devilbend Creek and Bald Hill Creek). Water-rats also appear to be well-established at many coastal locations around Port Phillip Bay (see below).
Distribution of sightings in south-western Victoria
The greatest proportion of Water-rat sighting records for south-western Victoria originated in the Barwon River basin (40%), followed by the Glenelg River (30%) and the Otway Coast and Hopkins River basins (each 12%). In the Barwon system, 68% of all reported sightings were made at Lake Wendouree in Ballarat. Water-rats were also seen along the Barwon River from Birregurra to as far downstream as Reedy Lake Game Reserve (24% of all sightings), with single reports of sightings received for Waurn Ponds Creek, West Barwon Reservoir, Goslings Creek, Warrambine Creek and Winter Creek. Immediately to the north-east of the Barwon catchment, Water-rats were recorded in the Moorabool River at Fyansford, in the west branch of the Moorabool River upstream of Lal Lal Reservoir, and in Hovell Creek near Lara.
In the Glenelg River basin, Water-rat sightings were distributed along the Glenelg River from Cherrypool (i.e. upstream of Rocklands Reservoir) to as far downstream as Lower Glenelg National Park. However, nearly two-thirds of all sightings originated in the Wannon River system near Hamilton, notably at Lake Hamilton (the source of 30% of sightings for the basin) and the Hamilton Botanic Gardens Lake (11% of sightings). Elsewhere in south-western Victoria, sightings were recorded in the Hopkins River basin along the Hopkins River from Ellerslie downstream to Allansford and also at sites on Brucknell Creek, Mount Emu Creek, Burrumbeet Creek, Bo Peep Creek, Fiery Creek and the Merri River at Warrnambool and Grassmere. The species was also seen in the Ot-way Coast basin at Limestone Creek (a tributary of the Curdies River), along the Gellibrand River and several of its tributaries (Love Creek, Cole Creek, Chapple Creek, Skinner Creek, Latrobe Creek), and along the Barham, Wye, St George and Erskine Rivers. In the Portland Coast basin, Water-rats reportedly occurred along the Moyne River at Hawkesdale West, at the confluence of the Shaw and Eumeralla Rivers in Lake Yambuk, and at Fawthrop Lagoon in Portland township.
No reliable Water-rat sightings were received from the Lake Corangamite basin in the current study period extending from 2000-2017. However, one person reported that the species was seen reasonably frequently at Barongarook Creek in Colac in the 1970s. The most recent Water-rat records held by the Atlas of Living Australia for this basin include two sightings made at Floating Islands Reserve in the 1990s, a Water-rat carcass found on the shores of Lake Milangil in 1994 and one sighting of a live animal made at Lake Corangamite in 1996.
Distribution of sightings in south-eastern Victoria
In south-eastern Victoria, the largest proportion of Water-rat sightings was associated with the Thomson River basin (28%), followed by the Mitchell River basin (23%) and Latrobe River basin (18%). In the Thomson River system, sightings were recorded in the Thomson, Macalister and Avon Rivers as well as the Aberfeldy and Perry Rivers, Valencia Creek, irrigation channels near Nambrok and Lake Guthridge in Sale. In the Mitchell River system, Water-rats were observed in the Mitchell River's lower reaches near Bairnsdale and Eagle Point, in MacLeod Morass and in Clifton Creek at Wy Yung, and also in the upper Mitchell catchment at sites along the Dargo and Crooked Rivers. In the Latrobe River system, sightings were made in the Latrobe River near Sale and in Lake Narracan. Upstream of Lake Narracan, Water-rats were recorded in the Loch River, Sandy Creek (near Willow Grove), Narracan Creek and Bear Creek; in the Morwell River subcatchment, sightings occurred at Billys Creek, Waterhole Creek, Kernot Lake in Morwell township, the Hazelwood cooling pondage near Churchill and the Morwell River Wetlands.
Elsewhere in south-eastern Victoria, fewer than 10 sightings were obtained respectively for the Tambo River, Snowy River, South Gippsland and Far East Gippsland basins. In the Tambo catchment, Water-rats were recorded in the lower reaches of both the Tambo River (in and downstream of Swan Reach) and the Nicholson River (in and downstream of Sarsfield). In the Snowy River basin, sightings were reported in the system's lowest reaches at Marlo and also farther upstream in the Buchan River and in the Little River gorge. In the South Gippsland basin, Water-rats were observed along the Powlett River, Ayr Creek (near Inverloch), an unnamed creek on Wilsons Promontory, Fish Creek, the Tarwin River (at Tarwin Lower and Mirboo North) and Monkey Creek in the Merriman Creek system. Lastly, in the Far East Gippsland basin, sightings were recorded along the Goolengook, Cann, Thurra and Genoa Rivers.
Distribution of sightings in coastal Victoria
In Victorian coastal habitats, the largest proportion of Water-rat sightings originated in the Gippsland Lakes (59%), followed by sites in Port Phillip Bay (37%). In addition, five coastal records originated in eastern Victoria (including three sightings made near Mallacoota, one near Cape Conran and one in Western Port) and one record originated near Portland in western Victoria. None of these records appeared to involve an ocean beach that is routinely associated with large waves. Instead, habitats were described as being relatively sheltered, e.g. a flooded area behind beach dunes at Mallacoota or the San Remo back beach in Western Port.
More than 90% of records for the Gippsland Lakes originated in Lakes King and Victoria, which broadly merge with each other and receive inflows from the Tambo, Nicholson and Mitchell Rivers. Other Water-rat sightings for the Gippsland Lakes were distributed fairly evenly between Lake Wellington at the system's extreme western end (5%) and Lake Tyers at its extreme eastern end (3%).
Although Water-rats were seen around the entire perimeter of Port Phillip Bay, more than half of all records for this water body (52%) originated around the northern end of the bay, at sites contiguous with the Yarra, Maribyrnong and Werribee River basins. In addition, 22% of Water-rat sightings originated at the bay's eastern end (at sites contiguous with the Bunyip River basin along the edge of the Mornington Peninsula) and 26% of sightings originated at its western end (at sites contiguous with the Moorabool and Barwon River basins, especially in Corio Bay near Geelong).
Habitat use and factors contributing to mortality
Water-rats were recorded in eight broadly defined habitat types. The animals were most commonly seen in or next to rivers (26% of all records) or creeks (25%), followed by natural and man-made lakes and reservoirs (18.5%) and coastal habitats (17%). Though fewer persons reported sightings from wetlands and morasses (7%) or irrigation channels (4%), population density was often described as being high in both habitat types with numerous sightings made over time. Finally, 2% of sightings reportedly occurred in river estuaries, and just 0.5% at sites lacking substantial surface water in the immediate vicinity.
Cause of death was identified in 40 Water-rat mortalities described in reports dating from 2001-2017. Of these, the factor most often resulting in death was use of enclosed yabby traps or craypots in which Water-rats drowned (17 records, 42.5% of incidents). An additional 17.5% of mortalities were due to Water-rats drowning in other types of traps or nets set to capture fish or crustaceans, including one incident in which a dead Water-rat was recovered from a licensed eel-net. Predation was implicated in 15% of mortality records, including three observed attacks by pet cats and one by a pet dog, with Foxes Vulpes vulpes deemed to be responsible for two other deaths. Finally, five deaths (12.5%) occurred when Water-rats were hit by motor vehicles, one animal died after being shot with a small calibre rifle, one died after consuming poisoned rodent bait, one died after becoming trapped inside a disconnected water tank, one apparently suffocated when a large pile of weeds in which it was sheltering was burnt, and one was found tangled in fencing wire (and presumed to have drowned) after major flooding. Although no cases were reported in which a mortality was specifically attributed to drought, several persons reported that Water-rat sightings ceased when surface water vanished locally in extremely dry periods. In the best documented case, Water-rats were seen at Lake Wendouree in Ballarat as water depth dropped progressively due to drought from 2004 to 2006, but disappeared when the lake dried out entirely in early 2007. The first sighting recorded after the lake refilled in 2010 occurred in late 2013; up to four Water-rats were observed in the lake at the same time by June 2015, with up to six animals seen concurrently by November 2016.
Results of live-trapping studies Water-rat capture frequencies varied by roughly an order of magnitude across water bodies sampled in both regional Victoria (Table 1) and the greater Melbourne area (Table 2). Even within a given river basin, a two- to five-fold difference was recorded in capture frequencies among different named waterways. The highest frequency of Water-rat captures (one or more animals present at 90% of sites) was recorded along Mount Emu Creek in and near Skipton and along the Maribyrnong River between Keilor North and Brimbank Park. The lowest capture frequencies (one or more animals present at [less than or equal to] 10% of sites) were recorded mainly in waterways located in south-eastern Victoria (Buchan River, Wentworth River, Valencia Creek, Aberfeldy River) but also the Taponga River in Lake Eildon National Park (northeastern Victoria) and the two streams flowing into Toorourrong Reservoir in the Plenty River catchment (north of Melbourne).
When analysed regionally, Water-rat capture frequency was highest in south-western Victoria (mean [+ or -] SEM = 0.58 [+ or -] 0.12, n = 4), followed by north-western Victoria (0.45 [+ or -] 0.07, n = 5), the Melbourne area (0.31 [+ or -] 0.05, n = 17), north-eastern Victoria (0.30 [+ or -] 0.15, n = 3) and south-eastern Victoria (0.11 [+ or -] 0.06, n = 5). The combined mean capture frequency for the western half of Victoria significantly exceeded the combined mean capture frequency for the eastern half of Victoria and Melbourne (t = 3.208, P = 0.003). Within the latter group, the mean capture frequency for south-eastern Victoria differed significantly from the mean capture frequency for the Melbourne region (t = 2.930, P = 0.008). There was no significant difference between mean capture frequencies for north-eastern and south-eastern Victoria (t = 1.722, P = 0.136), though this finding remains equivocal due to low sample size, particularly in the case of north-eastern Victoria. A comparison of regional variation in Water-rat capture frequencies as summarised above with regional variation in Water-rat sightings as portrayed in Fig. 3 supports the conclusion that sightings were under-reported in western Victoria compared to the rest of the state.
This study indicates that the Water-rat is appropriately classified as a habitat generalist in aquatic ecosystems. Slightly more than half of all Water-rat sightings occurred along freshwater rivers and streams, with around one quarter of sightings originating in more or less static water bodies, including ponds, lakes, lagoons, wetlands and morasses. Most of the remaining sightings occurred in saline coastal habitats, mainly around the perimeter of protected bays or in river estuaries.
The findings also confirm that Water-rats can adapt successfully to the human transformation of natural landscapes. Numerous sightings originated in man-made impoundments, including major water storages (e.g. Lake Mulwala, Lake Eppalock, Laanecoorie Reservoir, Lake Lonsdale, Lake Fyans) and smaller lakes (e.g. Victoria Lake in Shepparton, Lake Weeroona in Bendigo, Lake Wendouree in Ballarat, and Lakes Daylesford and Hamilton). In the greater Melbourne area, nearly half of all Water-rat sightings in the Yarra basin were recorded in the highly urbanised habitats found downstream of Dights Falls; one-third of sightings made in the Werribee basin originated either in ponds at the Werribee Sewage Treatment Plant or artificial lakes developed in conjunction with housing estates at Point Cook. Large numbers of sightings were also associated with irrigation channels in parts of northern Victoria (e.g. in the Gunbower area and the Dingee and Shepparton-Tatura irrigation districts). Similarly, McNally (1960) and Olsen (1980) reported that substantial Water-rat populations were established in irrigation districts located respectively in the Rochester-Echuca area in Victoria and near Griffith, New South Wales.
The Water-rat's adaptability is presumably founded in part on its catholic diet and flexible feeding habits. Animals are known to forage both during the day and at night (Watts and Aslin 1981; Gardner and Serena 1995), and to feed on aquatic species ranging in size from insects and spiders to sizable fish and adult waterbirds (including ducks, grebes, Eurasian Coot Fulica atra, Western Purple Swamp Hen Porphyrio porphyrio and Short-tailed Shearwater Puffinus tenuirostris) (Woollard et al. 1978). Furthermore, introduced fish species such as Goldfish Carassius auratus, Redfin Perch Perca fluviatilis and Mosquito Fish Gambusia affinis are readily consumed and may actually be eaten in preference to indigenous fish (Woollard et al. 1978). Although most of their food is obtained in the water, Water-rats are capable of foraging on land and have even been observed searching inside hollow trees in presumed pursuit of roosting bats (Woollard et al. 1978). Water-rats also have been documented to feed on fish netted by commercial fishermen and offal generated by fish canneries or recreational anglers, to raid poultry runs, and to consume human food waste and reasonably fresh carrion (Woollard et al. 1978; Smales 1984; this study).
Nets or traps set for crustaceans or fish were responsible for 56% of Victorian Platypus mortalities reported from the 1980s to 2009 where the cause of death could be reliably assigned, with 13% of mortalities due to predation by raptors, dogs or foxes (Serena and Williams 2010). Similarly, 60% of the Water-rat deaths described in the current study were caused by animals drowning in enclosed crustacean traps or fish nets (including a commercial eel net), with 15% of deaths due to observed predation by pet cats and dogs or presumed predation by foxes. Water-rat deaths in Victorian eel nets have been documented previously by Department of Primary Industries (2008): four Hydromys reportedly drowned as bycatch in licensed eel nets set in Lake Wellington in the mid-2000s, representing an estimated mortality rate of one Water-rat per 42.5 net-days of eel-netting activity. In Western Australia, Trocini et al. (2015) reported that 43% of the 30 Water-rat deaths described in a community-based survey were caused by animals drowning in enclosed traps set for Hairy Marron Cherax tenuimanus, with predators and motor vehicles respectively accounting for 7% and 20% of mortalities.
The remaining Water-rat mortalities reported in the current study were largely due to animals being hit by motor vehicles (12.5%) or flooding/misadventure (7.5%), but also included one animal that was shot and one that died after ingesting poisoned bait distributed near a river. Improved public awareness of the fact that the Water-rat is a protected native species would presumably help to reduce the number of animals that are killed deliberately, including those that provoke human ire by consuming ornamental goldfish or using moored boats as handy but messy feeding sites.
Although no formal studies have been conducted to date to monitor how Water-rats respond to drought, population size is predicted to drop when aquatic habitats contract or disappear, as supported anecdotally in this study. Along with having less to eat in such circumstances, Water-rats are presumably more likely to be killed by a predator if they spend more time on dry land, particularly if vegetation cover is sparse. In addition, female Water-rats become reproductively senescent by the age of 3.5 years, even when surviving in captivity for up to 6 years (Olsen 1982). Accordingly, protracted drought may plausibly result in populations disappearing if reproduction fails widely in three consecutive dry years.
As suggested by observations recorded at Lake Wendouree, it may take a number of years for a previously sizable Water-rat population to recover after a severe drought concludes. In places where the species has disappeared, the length of the recovery period presumably will be dictated in part by the time required by migrants to find and colonise the vacant habitat. This in turn will be influenced by the distribution of reliable drought refuges that continue to support the species. Population recovery time will also be limited by the reproductive rate of new migrants, although this can be quite rapid in favourable circumstances: female Water-rats may first breed when less than six months old and are capable of producing more than one litter annually (Olsen 1982), with litters of 4-5 juveniles often born in the wild (McNally 1960).
Presuming that the frequency of Water-rats entering nets mirrors the species' relative abundance, Water-rats typically occur in higher numbers in western than eastern Victoria, with the sparsest populations most often found in the state's south-eastern quadrant. The occurrence of low numbers of Hydromys in southeastern Victoria has been reported also by Smales (1984), who succeeded in capturing the species in just two of seven rivers or creeks in the Latrobe, Thomson and Tambo River basins where wire mesh cage traps baited with fish were deployed. One factor that potentially may contribute to reduced Water-rat numbers in south-eastern Victoria as compared to the rest of the state is predation by Long-finned Eels Anguilla reinhardtii, which are restricted to coastal catchments east of Wilsons Promontory and can weigh up to nearly 17 kg (Cadwallader and Backhouse 1983). It is therefore conceivable that they could prey successfully on Water-rats, particularly small juveniles, though corroborating data (e.g. in the form of Water-rat remains recovered from eel stomachs) are currently not available.
The relatively large size and distinctive appearance of the Water-rat, and the fact that it is partly diurnal, means that there is considerable potential for describing its distribution (and monitoring its population status) by collecting observational data. The occurrence of this species in a diverse range of fresh and saline water bodies, including those found in urbanised settings, also means that it is well placed to help command community support for protecting and improving aquatic habitats, particularly in modified landscapes. By the same token, there is an onus on management agencies both to be aware of the role of modified habitats in supporting Water-rats and to proceed with appropriate care when substantially altering those habitats. For example, anecdotal evidence obtained in this study suggests that lining irrigation channels with plastic as a water-saving measure can have devastating consequences for associated Water-rat populations. Further research on this issue is warranted, both to assess the actual impact on Water-rat numbers and develop strategies for addressing immediate or longer-term adverse outcomes.
This study was supported by The Wettenhall Environment Trust's Small Grants program. We are particularly grateful to the Trust's Executive Director, Beth Mellick, for her advice. We also thank the many other persons, too numerous to list individually, who contributed sighting records or helped to host and promote the community talks and other activities that were integral to the project's success. Lastly, the role of the Atlas of Living Australia in recording and sharing details of Water-rat sightings is gratefully acknowledged. Live-trapping activities were authorised by Victorian Wildlife Research Permits 10000390, 10000929, 10001440, 10001899, 10003545, 10004761, 10006077, 10006959 and 10008195 and Fisheries Permit RP 553.
Atlas of Living Australia website at http://bie.ala.org.au/species/Hydromys+chrysogaster. Accessed on 30 June 2017.
Cadwallader PL and Backhouse GN (1983) A Guide to the Freshwater Fish of Victoria. (Victorian Government Printer: North Melbourne) Department of Primary Industries (2008) Fisheries Status Report 2008. (Department of Primary Industries: Melbourne)
Department of Water Resources (1989) Water Victoria: A Resource Handbook. (Victorian Government Printer: North Melbourne)
Fanning FD and Dawson TJ (1980) Body temperature variability in the Australian water-rat, Hydromys chrysogaster, in air and in water. Australian Journal of Zoology 28, 229-238.
Gardner JL and Serena M (1995) Observations on activity patterns, population and den characteristics of the water rat Hydromys chrysogaster (Muridae: Hydromyinae) along Badger Creek, Victoria. Australian Mammalogy 18, 71-75.
Harris WF (1978) An ecological study of the Australian water-rat (Hydromys chrysogaster Geoffroy) in southeast Queensland. (Unpublished MSc thesis, University of Queensland)
McNally J (1960) The biology of the water rat Hydromys chrysogaster Geoffroy (Muridae: Hydromyinae) in Victoria. Australian Journal of Zoology 8, 170-180.
Menkhorst PW (1995) Mammals of Victoria. (Oxford University Press: South Melbourne)
Olsen PD (1980) Seasonal and maturational pelage changes, and injuries, in the eastern water rat, Hydromys chrysogaster, at Griffith, N.S.W. Australian Wildlife Research 7, 217-233.
Olsen PD (1982) Reproductive biology and development of the water rat, Hydromys chrysogaster, in captivity. Australian Wildlife Research 9, 39-53.
Serena M and Williams G (2010) Factors contributing to Platypus mortality in Victoria. The Victorian Naturalist 127, 178-183.
Serena M and Williams GA (2012) Effect of sex and age on temporal variation in the frequency and direction of platypus (Ornithorhynchus anatinus) captures in fyke nets. Australian Mammalogy 34, 75-82.
Smales LR (1984) A survey of Hydromys chrysogaster, the Australian water rat in central Gippsland. The Victorian Naturalist 101, 115-118.
Smart C, Speldewinde P and Mills H (2011) Influence of habitat characteristics on the distribution of the water-rat (Hydromys chrysogaster) in the greater Perth region, Western Australia. Journal of the Royal Society of Western Australia. 94, 533-539.
Speldewinde PC, Close P, Weybury M and Comer S (2013) Habitat preference of the Australian water rat (Hydromys chrysogaster) in a coastal wetland and stream, Two Peoples Bay, south-western Australia. Australian Mammalogy 35, 188-94.
Trocini S, Barrett G, Howard K and Ramalho C (2015) Rakali Community Survey 2014-2015. Report by WWF-Australia and Western Australia Department of Parks and Wildlife. (WWF-Australia: Perth)
Vernes K (1998) Observation of a long-range overland movement event by an adult common water rat, Hydromys chrysogaster. Australian Mammalogy 20, 409-410.
Watts CHS and Aslin HJ (1981) The Rodents of Australia. (Angus & Robertson: Sydney)
Woollard P, Vestjens WJM and Maclean L (1978) The ecology of the eastern water rat Hydromys chrysogaster at Griffith, NSW: food and feeding habits. Australian Wildlife Research 5, 59-73.
Zar JH (1984) Biostatistical analysis, 2nd Edition (Prentice Hall: Englewood Cliffs)
Geoff Williams (1) and Melody Serena (1,2)
(1) Australian Platypus Conservancy, PO Box 22, Wiseleigh, Victoria 3885
(2) Corresponding author Email: firstname.lastname@example.org
Received 8 February 2018; accepted 16 April 2018
Table 1. Estimated Hydromys capture frequency in Platypus survey nets set outside the greater Melbourne area (along water bodies sampled for [greater than or equal to] 9 site-nights) from 2000-2017. See Methods for definition of a site-night and description of how capture frequency was calculated. Nets were set in every year of hyphenated trapping periods. Locations = Malmsbury area (Coliban R), Vaughan Springs area (Loddon R), Grampians National Park to Laharum (Mackenzie R), headwaters to Wimmera R (Mount Cole Ck), Elmhurst to Glenorchy (Wimmera R), Omeo area (Livingstone Ck), Lake Rowan to Katamatite (Boosey Ck), Lake Eildon National Park (Taponga R, White Ck), Morrisons to Batesford (Moorabool R), Birregurra to Fyansford (Barwon R), Skipton area (Mount Emu Ck), Buchan area (Buchan R), Wentworth State Forest (Wentworth R, Pheasant Ck), Glencairn area (Barkly R, Skenes Ck), Valencia Creek State Forest (Valencia Ck), Thomson State Forest (Aberfeldy R, Donnelly Ck). River basin Water body Trapping period N of sites N of (years) sampled site-nights NW Victoria Campaspe Coliban R 2001, 2010 5 9 Loddon Loddon R 2002 7 12 Wimmera Mackenzie R 2000-02, 2004-06 13 58 Wimmera Mount Cole Ck 2000-01 8 15 Wimmera Wimmera R, 2000-05 25 97 Nowhere Ck NE Victoria Mitta Mitta Livingstone Ck 2012-17 10 39 Broken Boosey Ck 2003 16 19 Goulburn Taponga R, 2000-02 15 47 White Ck SW Victoria Moorabool Moorabool R 2003-06 20 49 Barwon Barwon R 2001-03 25 64 Otway Coast Curdies R 2002 11 14 Hopkins Mount Emu Ck 2003-04 5 9 SE Victoria Snowy Buchan R 2009-10, 2012-13, 8 23 2017 Mitchell Wentworth R, 2008 5 9 Pheasant Ck Thomson Barkly R, 2008 5 10 Skenes Ck Thomson Valencia Ck 2008 5 10 Thomson Aberfeldy R, 2008 5 10 Donnelly Ck River basin Capture frequency NW Victoria Campaspe 0.6 Loddon 0.25 Wimmera 0.3 Wimmera 0.5 Wimmera 0.6 NE Victoria Mitta Mitta 0.6 Broken 0.2 Goulburn 0.1 SW Victoria Moorabool 0.4 Barwon 0.6 Otway Coast 0.4 Hopkins 0.9 SE Victoria Snowy 0.04 Mitchell 0.1 Thomson 0.2 Thomson 0.1 Thomson 0.1 Table 2. Estimated Hydromys capture frequency in Platypus survey nets set in the greater Melbourne area (along water bodies sampled for [greater than or equal to] 9 site-nights) from 2000-2017. See Methods for definition of a site-night and description of how capture frequency was calculated. Nets were set in every year of hyphenated trapping periods. Locations = upstream of Longwarry North (Tarago R, Labertouche Ck), Bunyip Weir to Iona (Bunyip R), The Basin (Upper Dandenong Ck, Dobsons Ck), Belgrave to Scoresby (Monbulk Ck, Corhanwarrabul Ck), Three Bridges to Yarra Junction (Little Yarra R), Monbulk area (Woori Yallock Ck, Sassafras Ck, Emerald Ck), Mt Evelyn to Lilydale Lake (Olinda Ck), Strathewen/Kinglake National Park to Arthurs Creek (Arthurs Ck, Running Ck), Hurstbridge to Eltham (Diamond Ck), Donvale to Templestowe (Mullum Mullum Ck), upstream of Toorourrong Reservoir (Plenty R, Jacks Ck), Lower Plenty to Greensborough (lower Plenty R), Keilor North to Brimbank Park (Maribyrnong R), Darraweit Guim area (Deep Ck), Werribee area (Werribee R). River basin Water body Trapping period N of sites (years) sampled Bunyip Tarago R, 2002-05 8 Labertouche Ck Bunyip Bunyip R 2000-01 10 Bunyip Cardinia Ck 2003-13 14 Bunyip Upper Dandenong 2001-04 9 Ck, Dobsons Ck Bunyip Monbulk/Corhanwarrabul 2000-07 24 Cks Yarra Little Yarra R 2000-02 10 Yarra W. Yallock Ck, Sassafras 2001 9 Ck, Emerald Ck Yarra Olinda Ck 2000-07 13 Yarra Running Ck, Arthurs Ck 2000-04 14 Yarra Diamond Ck 2000-07 12 Yarra Mullum Mullum Ck 2000-07 10 Yarra Plenty R (E branch), 2000, 2002, 2004-06 6 Jacks Ck Yarra Plenty R (gorge) 2000, 2002-03 12 Yarra Plenty R (lower) 2000-04 6 Maribyrnong Maribyrnong R 2001, 2003 9 Maribyrnong Deep Ck 2001-06 7 Werribee Werribee R 2000 9 River basin No f Capture site-nights frequency Bunyip 33 0.45 Bunyip 13 0.2 Bunyip 113 0.11 Bunyip 43 0.2 Bunyip 235 0.12 Yarra 22 0.3 Yarra 9 0.4 Yarra 126 0.28 Yarra 84 0.2 Yarra 95 0.44 Yarra 118 0.34 Yarra 33 0.1 Yarra 15 0.2 Yarra 37 0.5 Maribyrnong 14 0.9 Maribyrnong 52 0.3 Werribee 12 0.3
|Printer friendly Cite/link Email Feedback|
|Author:||Williams, Geoff; Serena, Melody|
|Publication:||The Victorian Naturalist|
|Date:||Jun 1, 2018|
|Previous Article:||Records of the Common Wombat Vombatus ursinus (Shaw, 1800) in the Pilliga Forest, northern inland New South Wales.|
|Next Article:||A glimpse into the anatomy of Xanthorrhoea glauca subsp. angustifolia.|