The Mara wildlife "crash" - how real is it?

The BBC's 2 June 2011 story about a “crash” in wildlife numbers in the Maasai Mara region has been widely picked up, in most cases without further comment. The research was first reported in 2009 and has now been published in the Journal of Zoology. In African conservation, news is rarely wonderful, but is the BBC story a case of hyping broad-brush data into a shock-horror news report?

The sense of impending disaster is certainly not what you get from visiting the Mara and staying in touch with travellers, conservationists and operators on the ground, so this post is my short – and doubtless flawed – investigation into the data behind the story. 

The research is based on 49 aerial surveys conducted between 1977 and 2009, visually counting various species – buffalo, eland, elephant, giraffe, Grant’s gazelle, impala, kongoni, ostrich, Thomson’s gazelle, topi, warthog, waterbuck, wildebeest and zebra – as well as domestic livestock, across calibrated areas. Observers sat in the back of a light plane flying 100m above the ground and making sound recordings as they counted the animals they could see between two marker rods projecting from the plane. Groups of ten or more animals were photographed to be counted later. Digital photography must make that a lot easier, but even so it sounds very imprecise. 

The paper was published on 20 May 2011 in the Journal of Zoology and I’ve been looking at the raw data available for free download as "supporting info" at the journal website, at as well as at the paper itself.  

Swings and slides

One would expect to see a decline. We know there’s been a decline, as wildlife habitats have been hammered by human population pressure and livestock herds. The so-called “dispersal area” around the unfenced reserve has become a much more heavily populated region, with tracks, fences and settlements all inhibiting wildlife behaviour and free movement. 


What is surprising, however, is how enormously variable the figures are, with estimates of the population of the various species swinging up and down wildly, sometimes by several hundred percent in the space of a month. For example:

Koiyaki

Buffalo: biggest swing from 40,553 in Jul 1979 to 9131 in Aug 1979 to 39,835 in Sep 1979

Elephant: biggest swing from 125 in Mar 1979 to 2407 in Apr 1979 

Giraffe: biggest upswing from 1766 in Nov 1986 to 3771 in April 1987; biggest downswing from 3568 in Dec 1983 to 2704 in Jan 1984

Topi: biggest downswing from 16,178 in May 1997 to 5641 in Nov 2000; biggest upswing from 7498 in Sep 2005 to 17,004 in May 2007

Waterbuck: biggest swing from 1188 in May 1986 to 0 in Aug 1986 to 856 in Nov 1986 

Oloololo escarpment

These huge swings must indicate serious flaws in the estimation system, or a remarkable ability of species to bounce back in a short space of time or the possibility of major wildlife movements in and out of the areas being surveyed. The relative ease of counting different species depending on their numbers, size and habits, must also be a factor: bush-loving waterbuck are likely to be harder to count than large elephant and giraffe. 

Whatever the causes, the huge swings are not discussed in the paper.

The researchers say the populations of many species have declined to “a third or less of their former abundance”. The BBC went further, reporting “During the wet season, when there is no migration, resident wildebeest in the reserve have all but disappeared, falling by 97%”. In fact, the researchers only undertook two wet season surveys in the last fourteen years (May 2005 and May 2007) and wet season surveys in April 1979 and May 1982 also recorded zeros for wildebeest inside the reserve.

Data gaps 

The problem is we’re not comparing like with like: the results of surveys every month throughout 1979 and every few months from 1983–86, are being compared with a few scattered snapshots taken since the late 1990s.  Only two counts were conducted between the counts of May 1997 and May 2005, both in November (2000 and 2002) and since then there have only been four counts – September 2005, May 2007, November 2008 and October 2009. In recent years, there have been no counts in the peak migration months of July (the last and only one was in 1979) and August (counts took place in 1979, 1983, 1984, 1986, 1990, 1991, 1992 and 1996).

The problem with the figures is that they are so hard to interpret. Even the researchers acknowledge that while “the basic cause of the wildlife population declines seems to be the expanding human population in the ranches along with livestock influences spreading into the nominally protected area. . . the Mara ranches now support higher densities of some wildlife species than the protected area.” 

The BBC story refers to a “wildlife crash”. If this happened, it took place in the 1980s, judging by  the research charts, with only a gradual decline since then, or even an increase in some cases. Where the BBC’s story captions a photo of a buffalo with the words “African buffalo are all but gone”, that is not what the research shows. Instead it indicates a buffalo population in the reserve, based on the seven surveys undertaken in the last decade, of between 5000 and 13,000 animals, which doesn’t sound a bad number for an area of 1530 square kilometres.

The Migration

And what about the Great Migration? The BBC reports “huge numbers of wildebeest no longer pass through the region on their epic migration”. Well maybe not as huge, but the awe-inspiring migration of hundreds of thousands of wildebeest still takes place every year. We just cannot be sure what the numbers are because the researchers haven’t conducted enough surveys, “owing to financial constraints” as they admit. Only two surveys have been conducted in the dry season since 1996 (one in September 2005 and one in October 2009). 

The last time the researchers conducted a survey in July or August – peak wildebeest migration months – was August 1996, when there were 334,488 wildebeest in the Mara region by their estimate. Previous figures show migration numbers can go up as well as down (Aug 1992 494,027; Aug 1991 278,986; Aug 1990 262,947; Aug 1986 228,723; Aug 1984 216,918; Aug 1983 491,882, August 1979 520,456). We have only one July survey in the whole study, from July 1979 (1,171,684). 

There’s some complicated statistical analysis in the paper about B-splines and automatic smoothers, which I don’t begin to understand, but the fact that the study is unable to compare like with like by producing estimated populations from the same months over a number of years, rather than a cluster of surveys from the 1977 to 1987 and then occasional counts up until 2009, is surely a serious shortcoming. There were thirty counts in the first 10 years  (1977–87), and only six in the last ten (1999–2009), a period which has also included serious droughts in 2005–6 and 2008–9.

Without the missing population estimates, in a constantly shifting eco-system like the Mara, we’re looking at the equivalent of a redacted intelligence report, with huge gaps of blacked-out data.

I’m guessing many statisticians would look at the way the data has been analysed and agree that it passes the reliability tests. I’m sure the paper has had the usual peer reviews and its mathematical models have been well tested elsewhere. On the other hand, the Mara eco-system is virtually unique, so finding other research studies to corroborate this one will have been hard. And indeed most of the references are from the Mara-Serengeti region.

I am certainly not advocating complacency, but the sheer paucity of data over the last decade, especially an absence of surveys conducted during the peak migration season, makes me wonder if there’s not something more nuanced going on here than the headlines are reporting. 

Community eco-tourism

The current lower numbers are still huge by comparison with most other wildlife areas in Africa, but the larger numbers of wildlife from the early decades after independence have gone forever, as Kenya’s population has quadrupled. With the land in the Mara region sub-divided and sold off and human population still growing, the old eco-system, in which the Maasai and the herds mingled with the wildlife, is beyond being challenged. Local people no longer tolerate lions and hyenas near their homes, or buffalo where their children are walking to school or elephants raiding their corn. The answer, in an imperfect world, is wildlife conservancies for the wildlife and ranching and settlement areas for people.

Ol Kinyei Conservancy

The theory being enthusiastically endorsed – and it seems confirmed – by conservationists and eco-tourism champions, is that wildlife populations in the twenty-first century can be sustained by changing land use, involving local communities leasing their lands to wildlife conservation managers and established eco-tourism operators that limit visitors to one bed per 700 acres (three beds per square kilometre) and no more than 12 tents in a camp. 

An earlier BBC report on the Mara, filmed in 2009 (below), starts with a set of alarming figures and quotes Joesph Ogutu, the paper’s chief researcher, warning about extinction, but then goes on to reach more positive conclusions about the future of the community conservancies.  These now occupy much of the area that the research paper describes as “ranches”. Here, wildlife numbers, far from being in precipitate decline, are level or growing sustainably. If you’re planning a visit, think of going to a small camp in one of these well-managed sanctuaries, rather than to the National Reserve. 

The current status of the Mara

In 2011, highly endangered wild dogs are back (not mentioned by the research paper, but described as “all but disappeared” by the BBC), both in the reserve itself and the neighbouring conservancies. The future for predators is always uncertain, but it would be good to have regular estimates of their populations: the Mara region has substantial numbers of lion, leopard and cheetah, but their numbers are small compared with the grazers.

It’s the Narok council-run Maasai Mara National Reserve, entered via Talek, Sekenani or Ololaimutiek gates, that has the biggest problems with human incursion, cattle and goat grazing, and poaching. By contrast, the private conservancies are policed by their community members and shareholders who have everything to gain by keeping their herds to other areas. The Mara Triangle (aka Mara Conservancy) in particular, has been highly successful in keeping poaching to a minimum, working in partnership with its neighbours across the border in Serengeti.

The future might not look rosy, but it's not the bleak and desperate pictured painted by the BBC report, either. As long as the conservancies thrive and the reserve management is kept under scrutiny, the reduced wildlife populations will thrive, the visitors will come and local communities will increasingly adjust to a future with better prospects and fewer cattle. 

I'd be really interested to get feedback on these issues from specialists and visitors alike. Please comment below or email me - see my profile.