Study Site 1 – Hantana Range: Dunumadalawa Forest Reserve
A.Kittle & A.Watson – The Leopard Project aalanka@sltnet.lk

Contents:

Acknowledgments
Summary & Recommendations

Background
Study Site
Methodology
Discussion
Summary

References:

Appendix I – Photographs of Resident Female & Cub
Appendix II – Connector Map

Acknowledgements:

We would like to thank Mr. Cecil Dharmasena for first putting us in touch with the relevant personnel. We are greatly indebted to Mr. Hemantha Ekanayake for his open permissions to work within the forest, his support of our work and providing us with a base station throughout our study period.

We would like to thank the Mayor of Kandy, Mr. Kesara Senanayake and the Chief Engineer, Water works Dept. for giving us the approval to work within the forest reserve.

Thanks to the Wildlife Department - Mr. Herath, Director General, and Mr. Ratnayake Additional Director, Research and Training for support of our work.

We would like to thank all the dedicated staff at Dunumadalawa for their continued record keeping and time spent with us in the forest.

SUMMARY

Study site:

• Dunumadulawa is approximately 5 square kilometers and is bordered by connecting tea, pine forest and villages.
• It is a secondary growth forest approximately 70-100 years and is mid-country wet zone forest.
• It forms the terminal end of the Hantana-Galaha-Pedro-Boga valley-Peak Wilderness landscape chain. A study of the extent of actual connective forest is still underway.

Leopard Population:

• Expected population number is low; actual numbers to date are in accordance- one resident female, possible one adjoining resident female and one visiting male.
• A variety of prey species (utilized as prey) specific to this mid-country wet zone.
• Territoriality following expected patterns with usual associated behaviours.
• Range utilization is also apparent with movement within and outside of the study area occurring. Evidence also shows visiting animals may pass through the forest utilizing it as a movement zone between other wilderness areas.

Recommendation:

• Incorporation of these forest patches and connecting corridors into the larger framework of wilderness areas under protection.
• Strengthened protection of the forest against poaching of prey species, tree felling and the poaching of leopard itself.
• Realization of the importance that border tea estates and pine forests play in the movement of leopard and other wildlife between forest areas. Incorporation of these land use areas into a connecting frame work with other existing mid-country, sub-montane and montane wilderness areas.
• Decreasing any human/animal conflict that may arise by educating the border residents of these forest areas. (i.e. informing them that domestic cattle and dogs must be kept within their premises and must not be allowed to wander into the forest especially during evening and night time when the leopard prowls/hunts most).
• Raising the awareness of local residents to the importance of the leopard in the overall ecosystem. Also reducing the perception that the leopard is a dangerous animal.

(We are attempting to deal with these issues via our current work). Translocation is strongly NOT recommended

Background:

Fieldwork was conducted at the Dunumadalawa (Wakarawatte) Forest Reserve in the central hills of Sri Lanka from October 2003 to March 2004. This was the first of the three “upcountry” sites identified for study at which full-scale fieldwork was conducted. This forest reserve is a watershed for the town of Kandy and falls under the auspices of the Kandy Municipal Council’s Waterworks department. There are two reservoirs within the reserve – Dunumadalawa (or Dorwin) and Roseneath. This site is within the Kandy city limits making it an interesting focal point for leopard research as the existence of large carnivores in close proximity to urban centres underlines the notion that humans and wildlife can co-exist in such settings. This area was plantation estate land over 70 years ago (the original forest was cut down decades prior to this) and is now regenerating into healthy forest habitat, displaying the ease with which a non-forested area can regenerate into a wilderness suitable for wildlife. It is hoped that the eventual and overall results of our project will underline the importance that such altered patch forests play in the overall ecology of these regions. The continued protection for terminal/connected forests such as Dunumadalawa cannot be overstated.

Study Site:

History:

Originally know as the Walker Estate (or “Wakara-watte”), consisting of the Dorwin and Roseneath estates, this area was released from plantation type land use in the early 1900’s at which time coffee, tea and cocoa were all grown. Subsequently it has been allowed to naturally regenerate with some plant species being reforested within the last 5 years.

A number of surrounding villages utilize the forest for firewood and fruit collection, which is discouraged by the departmental staff. Its immediate borders are Pinus plantations to the southeast at Matinapatana which then run into home-garden villages and a small, private tea estate; the Tea Research Institute and Hantane tea estate to the west which is the beginning of extensive tea cultivation leading to Heeresagala, Bowalwatte and the Hantane hills in the southwest; Kandy town to the north; and Ampitiya town and other villages on the east (see site Map). It appears to be the northern terminus of what could be a vitally important corridor of forest and agricultural estate land that runs via the Hantane range south through Galaha to the Piduratalagala reserve and beyond via Hakgala all the way to Horton Plains, the Boga valley and Peak Wilderness (see Forest Connector Map/Appendix.II). This fragile, contiguous corridor is potentially of vital importance for the country as it encompasses a major portion of the central hill country inclusive of representative low, middle and high altitude wilderness areas.

Habitat:

The Dunumadalawa forest reserve is approximately 5 square kilometers (480 hectares) of secondary growth forest, having once been the site of coffee, tea and cocoa plantations. It is characterized by high canopy mixed forest dominated by several species including Albizzia and Jak remaining from the estate days. The abundant Jak appears to provide an important continuous food source for the toque macaques (Macaca sinica aurifrons), barking deer (Cervus muntjak) and wild boar (Sus scrofa) that inhabit the forest. Large-leaf mahogany (Sweitinia macrophylla), Ceylon almond (Canarium zeylanicum) and various Ficus species are also conspicuous. Tea, coffee and cocoa plants grow wild and in some places make up the overwhelming majority of the understory. In other areas the ironwood tree (Mesua ferrea) dominates. The elevation is between 550 and 980 meters above sea level. There are distinct rainy seasons during the Northeast monsoon from October to January and the southwest monsoon in May/June with inter-monsoonal rains being important in March/April and August. These features classify the region as mid-country wet zone forest. There are also numerous patches of Iluk grassland scattered throughout the reserve.

Fauna:

In addition to the barking deer, wild boar and macaques the area has a large porcupine (Hystrix indica) population. Other mammalian species include golden palm civets (Paradoxurus zeylonesis), Indian civets (Viverricula indica), mouse deer (Tragulus meminna), ruddy mongoose (Herpesthes smithii) fishing cat (Felis viverrina) and of course, leopard (Panthera pardus kotiya).

Numerous reptilian species were observed including the highly endangered star tortoise (Geochelone elegans). Forest geckos and a multitude of lizards are abundant as are amphibians such as frogs, toads and skinks. This has been investigated more thoroughly in the December 2001 Loris by Somaweera et. al.

A large number of birds representing a wide variety of species frequent the forest reserve. During the migrant season from September to March the forest is especially dense with avifauna. Casual observations over the course of our six-month study noted the existence of a minimum of 53 species.

Methodology:

A Global Positioning Systems (GPS) map of the road and trail system and the topography within the forest reserve was created. Links with other land areas (forested and non-forested) were determined and forest corridors also noted. An effort was made to thoroughly investigate the actual extent of connectivity between forest patches for the entire region from Dunumadalawa to Peak Wilderness; due to logistical and time constraints this has only been partially achieved to date.

As the forest floor has high leaf litter content making it nearly impossible to track animals without alteration of substrate, 15 -20 PIPS (Pugmark Impression Pads) were utilized throughout the site in order to monitor the movements of leopards as well as other animals (primarily prey species). (A PIP is a 2-3 meter wide strip of fine sand deposited on roads or pathways within which the spoor of wildlife can be detected). Each morning the trail system was walked and each PIP checked. All recognizable spoor were recorded before the PIP was cleaned with a soft brush in preparation for the next day. All leopard pugmarks found within the PIPs (or elsewhere) were traced on a glass sheet and retraced later for further analysis in the method of Singh (1999).

We searched all trails and roads as and when we walked them, for any spoor. This included pugmarks, scrapes and evidence of spraying. These latter two are communication signs utilized by leopards. They both indicate residence (see Kittle and Watson, 2002).

Infra red camera traps (Trailmaster 1550s) were utilized for visual identification of individual leopards. This leopard population is not at all habituated and therefore is much less visible than previously studied protected area, arid zone populations (see Kittle and Watson, 2002). As a result our reliance on camera trap images was more marked. Traps were set across roads, trails or game paths, which we determined to be utilized by leopards. They were set only during the night and taken down at dawn in order to avoid theft or tampering by visitors to the reserve or villagers.

Informal prey censuses were conducted as a daily routine along all trails utilizing a variation of Hirst’s 1969 strip census method. This was intended to provide a general idea of prey abundance.

Surveys of adjoining households regarding knowledge of leopard existence in the area, conflict situation and forest usage were also conducted. These surveys were conducted via interviews with local residents and water-works department staff who have been working in the area.

Results:Results:

Population:

Table 1: Population numbers including residents and non-residents

During the course of the study period we recorded a total of 24 sets of leopard pugmarks, 15 of which were deemed of good enough quality to trace. Of the total traced, 13 were of female or young animals while the remaining 2 were of males. On 6 occasions we encountered the pugmarks of a female and cub together.

Due to budget constraints our total camera tap nights during this initial period was low- a total of 53 camera trap nights during which 50 photographic images were recorded.

Table 2: Camera trap image results

One (1) sighting was made during the study. This was of a female and her 6-8 month old cub at 07:00 on a remote forest path (Appendix I).

Eleven (11) signs were documented during the study. These were divided into the following categories:

Scrapes with spray – 4
Scrapes without detectable spray - 2
Spray directly on tree – 3
Evidence of prey being dragged – 2

Half of the scrapes with spray and all occasions of spray directly on the tree were associated with female pugmarks.

Home Range:

The resident female appeared to occupy the entire 5 square kilometers of the study site. Frequency of use shows that this was her core area. The Male has a larger home range with the study site encompassing a section of his home range. The cub’s natal area is that of the mothers range.

Feeding Ecology:

A total of twelve (12) fecal samples were collected during the study period. Macroscopic hair and fragment analysis was conducted in the field. (Microscopic analysis is in process).A total of twelve (12) fecal samples were collected during the study period. Macroscopic hair and fragment analysis was conducted in the field. (Microscopic analysis is in process).

Fig. 2 Fecal Sample analysis results

A total of six (6) carcasses of animals killed by leopards were located during the study period. Of these three (3) were the remains of dogs, two (2) were the remains of porcupine and one (1) was the remains of deer.

Discussion

Population:

There is a minimum of two (2) adult resident leopards living in the Dunumadalawa forest reserve. The adult female is reproductive and currently has at least one cub who was approximately 7 to 9 months old at the end of the study period. Evidence in the form of persistent scrape and scent marking suggests that she spends the majority of her time within the Forest reserve. The male leopard, in contrast, appears to incorporate the Dunumadalawa Forest reserve into a larger territory as only occasional evidence of his habitation was documented. Perhaps the most obvious indication of the existence of a male leopard in the study area was the fact that the female was with cub. This fits in with the land tenure system of leopards whereby males occupy much larger home ranges than females (Mizutani and Jewell, 1998). The former tend to have home ranges that overlap the whole or partial home ranges of a number of females (Baily, 1993). It is possible that another adult female is resident in the upper area of Dunumadalawa but this was never verified with certainty. That this forest reserve is connected to other areas is apparent in that the male seems to only occasionally enter the reserve. In addition, there have been three (3) leopards poached from the vicinity of the reserve over the last three years. It appears that this poaching pressure is compensated for by the influx of new animals, another indication of connection between forest areas.

Home range:

In order to reach truly definitive conclusions regarding the home range size of leopards in this habitat a longer-term study is required. However, based on the frequency of use it appears that the resident female with her cub spends the majority of her time in and around this small forest reserve. It is possible that her home range extends somewhat beyond the bounds of the reserve but it is certain that the majority of her range, and perhaps all of it, falls within the reserve. The male appears to have a much larger home range, the extent of which remains unknown. That the female is with cub might also reduce the frequency with which the male utilizes this particular section of his home-range as he knows that she is presently not in reproductive state.In order to reach truly definitive conclusions regarding the home range size of leopards in this habitat a longer-term study is required. However, based on the frequency of use it appears that the resident female with her cub spends the majority of her time in and around this small forest reserve. It is possible that her home range extends somewhat beyond the bounds of the reserve but it is certain that the majority of her range, and perhaps all of it, falls within the reserve. The male appears to have a much larger home range, the extent of which remains unknown. That the female is with cub might also reduce the frequency with which the male utilizes this particular section of his home-range as he knows that she is presently not in reproductive state.

Feeding Ecology:

It is disturbing that carcasses of dogs were discovered during the study period however this number is slightly misleading as, unlike wild prey, dogs have owners who are actively interested in locating their missing animal. Therefore searches were made when a known dog went missing. Searches of this kind resulted in all three, dog carcasses being discovered. It is almost certain that the number of dogs preyed upon was much less than that of wild game. This was supported by the scat samples macroscopically analyzed, which indicated a reliance on porcupine, deer and pig. It is possible that some of the unknown remains are from dogs, but equally possible that they are from barking deer, hares, mouse deer, civets etc. It is interesting to note that the porcupine appears to play an important role in the feeding ecology of leopards at the Dunumadalawa forest reserve as a full 50% of the scat samples indicated porcupine remains.

Summary:

We recognize that a great deal more data needs to be collected in order to fully understand the leopard population and its attributes in the Hantane area; however this is an important pilot study, which addresses a number of meaningful questions. Of primary importance is the confirmed fact that leopards are residing full time in the Dunumadalawa forest reserve and not merely visiting this forest patch from time to time. Not only are they residing, they are raising cubs. This means that the forest reserve cannot be seen as marginal forest or a buffer. Furthermore, the human-leopard conflict, manifest in the form of leopards killing dogs and livestock is an issue, however the forest reserve does hold a wide variety of prey and it appears from the fecal sampling that the overwhelming majority of leopard prey is wild game and not domestic pets or livestock. From discussions with boundary landowners, reserve staff and villagers it is apparent that the issue of leopards taking dogs has been in existence for many years; it appears to be cyclical, perhaps occurring more during the existence of cubs which can occur approximately every 2 years.. While we heard about the issue from many different people, the number of actual incidents were relatively few but widely reported.

The Dunumadalawa forest reserve is a very important wilderness area, made more so by the verified existence of a resident leopard population - an endangered species listed under CITES Appendix II throughout its Asian range. That it is connected to other vibrant forest patches via low quality habitat such as tea estates and Pinus plantations is of vital importance for the future of the faunal inhabitants of this forest. Only one National Park - Horton Plains - exists in the entire Central Hill zone of Sri Lanka, while this ecological zone of sub-montane, wet to moist evergreen forests exhibits a remarkably high level of biodiversity and endemism. Already the land use history of estate plantations has decimated the original forest in this area, which underlines the importance of making a concerted effort to protect the few remaining forest patches and perhaps regenerate additional ones within this important zone. Our final recommendation and suggestion is that this area be given higher protection levels and that with more study (forthcoming under our on going project work) the connecting forested areas of Hantana through Galaha to Piduratalagala be considered for Reserve, Sanctuary or National Park status and that other land use areas (eg. Estate lands such as the Hantane tea estate, Uda Pussellawa tea estate) be conserved as they are and protected from increased development.

References:

Baily, T. N., 1993. The African leopard: ecology and behaviour of a solitary felid, Columbia University Press, New York.

Hirst S. M., 1969. Road-strip census techniques for wild ungulates in African woodland. Journal of Wildlife Management 33:40-48.

Kittle A. M. and A. C. Watson, 2002. Home range, demography and behaviour of the Sri Lankan leopard (Panthera pardus kotiya). Department of Wildlife Conservation report.

Mizutani F. and P. A. Jewell, 1998. Home-range and movements of leopards (Panthera pardus) on a livestock ranch in Kenya. Journal of Zoology, London 244: 269-286.

Singh L. A. K., 1999. Tracking tigers: Guidelines for estimating wild tiger populations using the Pugmark Technique (Revised Edition). WWF Tiger Conservation Programme, New Delhi.

APPENDIX I:

Camera trap photo of resident female leopard at Dunumadalawa forest reserve, Kandy. Taken on November 27, 2003.

Photo of 6-8 month old leopard cub taken at 7:00am on March
2004 at the Dunumadalawa Forest

Appendix 11:

Central Hills of Sri Lanka showing the area between Kandy and Peak Wilderness.
Tea estates, forest patches and villages make up this valuable wildlife region.
Connections within this landscape (red arrows) are patchy but of vital importance to the future existence of viable leopard populations in the Hills.
(Taken from Survey Department 1:250 000 Sri Lanka map, 1992 Edition)