Yala or Kumana? Sri Lanka's busiest safari park, and the one nobody sends you to

Two national parks in Sri Lanka's dry southeast have leopards in them.

One took 646,704 visitors in 2024.^[1] The other took 35,594.^[2]

Kumana takes about five visitors for every hundred that go to Yala. Ask most people planning a Sri Lankan trip to name the second park and they cannot. Then, in 2025, a study reported that the eastern part of Kumana holds one of the highest leopard densities ever recorded anywhere.^[3]

The obvious article to write from that is "skip Yala, go to Kumana." We are not going to write it, because it would be wrong. But the reason the gap exists is worth your time, and so is the one fact that decides which park you should actually book.

The gap is not because nobody noticed

Kumana was closed for eighteen years.

A ranger posted there from February 1983 gives a first-hand account, published by Sri Lanka's own wildlife ministry, of an attack on the park on 3 March 1985. Staff were withdrawn and transferred to Yala for the following three years.^[4] The park did not reopen until March 2003.^[5]

So Kumana did not fail to catch on. It was gone for a generation. The hotels, the jeep fleets, the drivers, the tour packages and the guidebook pages all grew up during the years it was shut. Yala did not out-compete Kumana. Yala was the only one open.

It is not an undiscovered wilderness either. Kumana was set aside in 1938 as a breeding ground for birds, and became a national park, then called Yala East, in 1969.^[4:1] It has been famous for eighty-eight years. Just not to us.

Why Yala is the right answer for most people

We want to be straight about this, because a lot of writing about Yala treats its popularity as a character flaw. It is popular because it works, and most of the reasons are things you would want.

You can actually get there. Tissamaharama, the town nearly every Yala safari starts from, sits 30 to 45 minutes from the main gate.^[6] From Ella it is 88 km, about an hour and a half. From Mirissa 123 km, an hour and three quarters. From Galle 147 km, just under two hours. From the airport 239 km, just under four.^[7] The Southern Expressway reaches Hambantota, the last significant town before the turning, and it cut the Colombo–Matara run from four hours to about ninety minutes.^[8]

A safari that costs you half a travel day is a different proposition from one that costs you two.

You can do it from the south coast without moving hotels. The Mirissa and Weligama day trip is a real, packaged, commonly sold thing, not a stunt. It is also a 2am to 3.30am pickup, three and a half to four hours in the car each way, and back late afternoon or early evening, and you should book it knowing that rather than discovering it at the hotel desk.^[9]

You can stay however you like. Tissamaharama has guesthouses and hostels at the budget end, and forty minutes away at the park boundary there is a genuine luxury tier. Wild Coast Tented Lodge is a Relais & Châteaux property, and it, Jetwing Yala, Cinnamon Wild and the Mahoora tented camps are all run by Sri Lankan companies rather than foreign chains.^[10] Very few wildlife destinations anywhere offer that spread inside one small town.

There is more there than leopards. Roughly 350 elephants, sloth bears, 215 recorded bird species.^[11] Two working Buddhist pilgrimage sites, Sithulpahuwa and Magul Vihara, sit inside the park boundary and draw an estimated 400,000 pilgrims a year on their own.^[12] Kataragama, just outside, is one of the few religious sites in Sri Lanka venerated by Buddhists, Hindus, Muslims and the Vedda people alike.^[13]

And the leopards really are there. A photo-identification study run from the Yala Leopard Visitor Center identified 137 individual leopards in Block I between 2013 and 2021, from 750 days of field observation.^[14] These are not anonymous animals in a statistical model. They are known, catalogued and recognised on sight. In an exit survey of 936 visitor groups, 34.5% saw a leopard and 97.4% saw an elephant.^[15] On the days the park's two mobile phone towers were switched on, the leopard rate was 51.1%.^[15:1] Those are counted rates from a peer-reviewed study, not an operator's promise, and they are the reason we would send a first-time visitor to Yala without hesitating.

What "busy" actually means

Yala is crowded, and the crowding is concentrated. Block I is 141 km² of a 979 km² park, about 14%, and it is where nearly everyone goes.^[15:2] In the first half of 2025 the park drew close to 390,000 visitors.^[16]

The vehicle numbers have been argued over for a decade. In 2017 the department let 300 vehicles a day in through the main Palatupana gate, then raised that to 400 after drivers parked their jeeps outside the park in protest, which took the park's limit to 450 alongside the 50 allowed at the Katagamuwa gate. The safari vehicle owners said they wanted their full capacity of 1,000 a day.^[17] Today there are 552 registered jeeps, with around 500 admitted across the morning and afternoon sessions, and a deputy minister has proposed cutting that to 250–300. That is a proposal, not policy.^[16:1]

The best study on this argues that cutting the jeep numbers would not fix the problem:

"Overcrowding at sightings is not proportionate to park visitation per se... reducing the number of vehicles even by half is unlikely to have an impact on overcrowding."^[15:3]

A hundred jeeps spread across a park is not a crowd. A crowd is thirty jeeps at one leopard, and that happens because drivers radio and phone each other, not because the gate let too many in. When the two mobile towers serving the park were switched off, the leopard sighting rate fell from 51.1% to 15.6%.^[15:4] Two thirds of the leopards people see at Yala, they see because somebody phoned. The network that causes the jam is the network that finds the animal.

In that same survey, 52.3% of groups reported overcrowding, and it made no measurable difference to how they rated the day.^[15:5] 77.7% called the experience excellent or good. Under 5% called it poor or terrible.^[15:6] People notice the jeeps. They mostly do not mind.

What the Kumana study says, and what it does not

Researchers estimated about 41 leopards per 100 km² in the studied eastern part of Kumana, not across the whole park.^[3:1] The authors describe that eastern region as having one of the highest recorded leopard densities globally, and that comparison is theirs, not ours.^[3:2]

Three things we will not do with it.

We will not make it park-wide. The study area is a portion of Kumana, and the authors say so themselves.

We will not turn it into a headcount. A modelled density is not a census, and nobody has counted Kumana's leopards individually. Separately, a volunteer project called Kumana Leopards, running since 2019, publishes a field guide reporting 92 adult leopard records. That is a different kind of evidence and worth knowing about.^[18]

And we will not rank the two parks against each other, which takes a section of its own, because everyone else will.

Why you cannot line these two numbers up

Yala Block I: 0.5461 leopards per km², which is 54.6 per 100 km².^[14:1] Eastern Kumana: about 41 per 100 km².^[3:3]

The Yala figure is a minimum confirmed count: 77 leopards the researchers could individually name and photograph in Block I as at 31 March 2021, divided by the area.^[14:2] The Kumana figure is a statistical estimate from camera-trap modelling over a 16-month study.^[3:4] One is a floor built from animals someone has actually met. The other is a projection built from detections. Different years, different methods, different questions.

Comparing them is like comparing two schools' exam results when the pupils sat different exams. You can put the numbers next to each other. It does not mean the higher one is better.

What both studies support is narrower and more useful. Kumana holds a serious leopard population that almost nobody visits, and Yala holds a serious leopard population that a great many people do. Take that and leave the ranking.

Getting to Kumana is the whole story

There is no public road from Yala into Kumana.

They share a river boundary, the Kumbukkan Oya, with Yala's Block II on the far side, and nothing else.^[4:2] The official access route runs the long way around the bottom of the island: Wellawaya, Monaragala, Siyambalanduwa, Lahugala, Pottuvil, Panama, Okanda.^[4:3] From Okanda to Kataragama, beside Tissamaharama, is 164 km and about two hours fifty.^[19]

You do not add Kumana to a Yala trip. You go to Sri Lanka's east coast for it. The realistic base is Arugam Bay, Pottuvil or Panama. Sources put the gate anywhere between 12 and 40 km away and the drive at 45 minutes to an hour and a half, on a road that every source describing it calls poor.^[20] Accommodation is surf-guesthouse standard. There is no equivalent of the Tissamaharama luxury tier.

Visitors complain about the road, an entrance toilet described as not clean, and no restaurants nearby.^[21] The same page rates the park 4.5 out of 5 across 24 reviews, none of them poor or terrible.^[21:1] People like it in spite of the state of it, and we would rather tell you that than sell you somewhere polished.

If you want to book locally rather than through an agency, the operators there are genuinely local. Panama Village is a network of resident families running eight homestays and states plainly that it is not a tour company. Kumana Safari works out of Panama South, at the park's doorstep rather than the surf town, and sells a full day that takes in the Kudumbigala temple, a village lunch and the Urani lagoon alongside the park. Sanka Safaris is one Arugam Bay guide running his own drives, with prices published by group size.

On money: Yala is $25 an adult in tourist season and $20 outside it; Kumana is $12 all year, under the same fee gazette.^[22] Neither is the amount you pay. Both add VAT, a per-vehicle charge and a per-jeep service charge at the gate, and the jeep itself is hired separately on top. Our Yala park page works the full arithmetic through.

The thing that actually decides it

The answer is months, not leopards.

Yala's dry window runs roughly February to July, and Blocks 1 and 2 are closed from 14 September to 15 October 2026 for drought.^[23] Kumana's best months are April to July, when tens of thousands of birds arrive at the Kumana Villu swamp to breed. It is an east-coast park on the opposite monsoon, and the wet, difficult period there runs from around October into the new year.^[4:4][24]

The two parks sit on opposite sides of an island with opposite weather, so they are barely rivals. The useful question is not which park, it is what month you are here.

So which one

Book Yala if it is your first safari in Sri Lanka, or you are travelling with children or anyone who will not enjoy four hours on a broken road, or you are on the south or west coast and do not want to restructure the trip, or you want the option of a proper hotel afterwards. You will share it. You will also very probably see a leopard, and most people who share it come away glad.

Book Kumana if you are on the east coast already, or you are travelling between April and July, or you are there for birds, which is what the park has been known for since 1938, or you have done Yala and the jeeps spoiled it for you. Go knowing the road is bad and the facilities are basic.

Book both only if you have the days for it. They are three hours apart by the only road that connects them.

One last thing about the 41 per 100 km². It is a reason to take Kumana seriously. It is not a promise that you will see a leopard, and no park anywhere can give you that.

Sources

---

1. Yala's 2024 visitor figures, 646,704 total and 353,646 foreign, from the Department of Wildlife Conservation Performance Report 2024 (via parliament.lk), as recorded in our own park data. Read 21/08/26. ↩︎

2. Kumana's 2024 visitor figures, 35,594 total and 10,980 foreign, from the Sri Lanka Tourism Development Authority Year in Review 2024 park table, as recorded in our own park data. Read 27/08/26. The two parks' figures come from two different bodies' 2024 tables, so read the gap between them as a difference in scale rather than as a like-for-like count. ↩︎

3. Rodrigo, P., Anawarathna, C., Weerasinghe, H., Kumara, A., Gunathilake, C., Sigera, S. & Mahaulpatha, D. (2025), "Occupancy, population density, and activity patterns of endangered Sri Lankan leopard (Panthera pardus kotiya) in Kumana National Park, Sri Lanka", Journal of Asia-Pacific Biodiversity, DOI 10.1016/j.japb.2025.01.008. Density figure from Table 4; study-area description p. 3; the "highest recorded globally" wording is the authors' own, from the abstract. Author-uploaded full text. Read 27/08/26. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

4. Ministry of Wildlife and Forest Conservation, Kumana National Park. Source for the 3 March 1985 attack and the transfer of staff to Yala for the following three years, the 1938 declaration as a bird breeding ground, the 1969 national park designation, the official access road, the Kumbukkan Oya boundary with Yala Block II, and the April–July breeding season at the Kumana swamp. Read 28/08/26. On the founding date the sources disagree: this ministry account says it became a national park in 1969 and expanded to 17,863 hectares in 1970, while Wikipedia gives 20 January 1970. We follow the ministry. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

5. March 2003 reopening: Wikipedia, corroborated independently by the Panama-based operator Kumana Safari. Both read 28/08/26. ↩︎

6. Tissamaharama at 30–45 minutes from the gate, cross-checked on two operator guides: Exclusive Yala Safari and Girl on a Zebra. Read 28/08/26. ↩︎

7. Road distances and driving times via Rome2Rio, each route opened individually: Ella, Mirissa, Galle, Colombo Airport. Read 28/08/26. We could not find one page giving both a distance and a time for Colombo city itself, so no figure for that is published here. ↩︎

8. Southern Expressway (E01), read 28/08/26. ↩︎

9. South coast day-trip timings from Girl on a Zebra, which gives a 2am departure, three and a half to four hours' driving each way and a return in the late afternoon or early evening, and a first-hand TripAdvisor account of a 3.30am Mirissa pickup and a two-hour drive. Neither source gives a total length for the day, so we do not print one. Both read 28/08/26, rechecked 29/08/26. ↩︎

10. Wild Coast Tented Lodge (Resplendent Ceylon, Colombo; Relais & Châteaux), Jetwing Yala, Cinnamon Wild Yala, Mahoora (Eco Team, Colombo, operating tented camps since 1998). All read 28/08/26. Hostel and guesthouse pricing in Tissamaharama comes from aggregator listings we could not open directly, so no figure is published here. ↩︎

11. Species counts from Wikipedia and the official Sri Lanka Tourism page; the ~350 elephant estimate is from the Sri Lankan operator guide Best of Lanka. All read 28/08/26. Sources differ on whether six or seven of the bird species are endemic; six appears more often, so we do not print a figure. ↩︎

12. Sithulpahuwa and Magul Vihara, and the ~400,000 annual pilgrim estimate: Wikipedia, read 28/08/26. ↩︎

13. Kataragama temple, read 28/08/26. ↩︎

14. Wattegedera, M., Silva, D., Siriwardena, R. & Piyasena, M. (2022), "A Multi-Point Identification Approach for the Recognition of Individual Leopards (Panthera pardus kotiya)", Animals 12(5):660, PubMed Central. 137 individuals identified 2013–2021; minimum confirmed Block I population of 77 as at 31/03/21, giving 0.5461 per km², up from 0.217 per km² in 2017. Read 28/08/26. Note: this paper's methods text gives Block I as "141,000 hectares" while its own results table gives 141 km². The table figure matches two independent sources and is the one used here. ↩︎ ↩︎ ↩︎

15. "Is restricting mobile communication a solution to overcrowding? A test from Yala National Park, Sri Lanka", Ceylon Journal of Science 53(2): 243–251 (2024), open access. Exit survey of 936 visitor groups. Source for Block I at 141 km² of 979 km², the 34.5% of groups who saw a leopard across the whole survey, the 51.1% and 15.6% leopard sighting rates on days the mobile towers were on and off, the 97.4% elephant rate, the 77.7% excellent-or-good satisfaction rate, the 52.3% overcrowding report with no significant effect on satisfaction, and the quoted passage on vehicle numbers. Read 28/08/26. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

16. Mongabay, "In Sri Lanka, animals pay the price for overcrowding and speeding jeeps", 17/04/26. Source for 552 registered jeeps, ~500 admitted per session, the deputy environment minister's proposed 250–300 cut, and the visitor figure for January to June 2025, which Mongabay takes from the Sri Lanka Tourism Development Authority's 2025 year in review. That figure is for Yala as a whole, not for Block I. Read 28/08/26, rechecked 29/08/26. ↩︎ ↩︎

17. Sunday Times Sri Lanka, "Yala safaris: Battle over wilderness", 05/11/17. Source for the 300 vehicles a day at the Palatupana gate raised to 400, the 50 at the Katagamuwa gate, the 450 limit the wildlife director general said would be strictly implemented, and the safari vehicle owners' demand for their full capacity of 1,000 a day. Read 28/08/26, rechecked 29/08/26. ↩︎

18. Kumana Leopards and its field guide, reporting 92 adult leopard records. Volunteer project launched 2019. Read 27/08/26. ↩︎

19. Okanda to Kataragama, 164 km and about 2h50m: Rome2Rio, read 28/08/26. ↩︎

20. Distances and times differ by source. Amazing Lanka puts the park 12 km south of Arugam Bay and gives Pottuvil to the Okanda entrance as 32 km and an hour. Sri Lanka Spirit gives 35 to 40 km and 45 to 60 minutes by jeep. A TripAdvisor reviewer gives an hour to an hour and a half. On the road itself: Sri Lanka Spirit calls it surfaced but deteriorating in sections, the operator Kumana Safari calls the tracks basic and rough, and TripAdvisor reviewers call the condition terrible and say the Panama road is hard to drive in a car. Amazing Lanka gives distances and does not describe the road. All read 28/08/26, TripAdvisor and Sri Lanka Spirit rechecked 29/08/26. ↩︎

21. Kumana National Park on TripAdvisor, rated 4.5 out of 5 from 24 reviews when we read it: 14 excellent, 7 good, 3 average, none poor or terrible. Source for the road complaints and for the reviewer who reports that there are no restaurants nearby and that the only toilet at the park entrance is not clean. Read 28/08/26, rechecked 29/08/26. ↩︎ ↩︎

22. Gazette Extraordinary 2292/11, read from a preserved copy of the Department of Government Printing PDF. Category I parks including Yala carry a tourist-season foreign adult rate of $25 and a regular-period rate of $20; Kumana is Category II at $12. VAT, a per-vehicle charge and a per-jeep service charge are added at the gate. ↩︎

23. Ada Derana, "Blocks 1 and 2 of Yala National Park temporarily closed", 26/08/26, announcing closure from 14 September to 15 October for drought. Read 28/08/26. ↩︎

24. The wet period on the east coast is given differently across sources, running from October to somewhere between January and April; we could not find a single authoritative month-by-month figure, so we describe it as October into the new year rather than printing a boundary month. ↩︎