Do fires stop the Serengeti migration?

Dr Kate Parr lighting a controlled fire in the Serengeti Ecosystem

There's been a bit in the East African press recently claiming that Tanzania has been deliberately setting fires in the Serengeti NP to block the migration. The Tanzanian National Park Authority (TANAPA) have, of course, denied this. Reading the articles and press releases there's obviously both some serious ignorance and some seriously bad journalism going on here, and I thought it might be useful to share a few of my observations.

The big burn...

So, I promised more details would come, so here goes with some more background and some of the detailed plans I've got to burn Serengeti. If you like what you see don't forget to support the project over here!

There are few active management activities taking place in most of East Africa's protected areas – a policy that has so far proved fairly effective at managing functioning ecosystems. However, one of the management activities that does happen in many areas, is burning. Fire is generally considered a vital component of savannah ecology, with trees, grasses and animals all adapted to a fire prone ecosystem: in fact, globally, 85% of fires occur in savannah habitats. In most of Tanzania's protected areas, rangers deliberately set fires each year for a number of different purposes, including the encouragement of new grass growth for grazers and the control of bush spread. Recent evidence from a global study suggests that at least for those savannahs occurring in areas with over 1000 mm of rainfall, the forest/savannah edge is often maintained primarily by fire, and that in its absence many savannahs will revert to forest. There is some debate, however, about whether grazing and browsing alone may provide an equivalent process when animals are at particularly high densities, with the result that the benefits of fire in certain areas have been questioned. For example, in some areas of Tarangire National Park with high numbers of migrant animals during the dry season, fire has been totally suppressed for over 30 years with relatively little obvious difference in grass and tree cover between fire-suppressed and more frequently burnt areas: it is argued that in the burnt areas fire simply burns potential animal forage. Similarly, in Kenya, the Kenya Wildlife Service has recently instigated an overall policy of fire suppression except where the fire is caused by natural events (essentially, lightening storms) in an attempt to encourage one view of naturalness. There are, of course, alternatives to these two extremes, where fire is sometimes suppressed and sometimes encouraged. Given that fire is one of very few activities conducted by management in these areas it seems sensible to understand more fully the consequences of these different management decisions.

Traditionally, pastoralist communities in northern Tanzania have set fires towards the end of the dry season, in anticipation of the rains. They make a calculated gamble between burning possibly useful forage if the rains are delayed, and waiting too long and being unable to burn (and encourage the new growth that will come with the rains) if the rain comes before they have time to burn. Obviously, by this time of the year grazing animals will already have reduced fuel loads to relatively low levels in many areas, so late dry season fires will only occur in less grazed areas, with other areas remaining unburnt. Such management was also the norm in many protected areas (including Serengeti) until around 1970, when concern over the regeneration of trees prompted a switch in fire management towards controlled burns in the early dry season (mainly in June). Such fires are generally cooler and patchier, with possibly less impact on woody plants. They also prevent the late season burns that are likely to be hotter and less easy to control with consequent reduced concerns about tourism infrastructure. Recently, managers of the Grumeti Reserves in western Serengeti have attempted to suppress early season fires, only allowing burns later in the dry season (from late July), allowing the migrant wildebeest and zebra to graze unburnt areas and increasing the forage available to these animals to the extent that their residence times within the reserves have tripled. Fires are also set during the short dry season, in February, and grass growth rates are such that some areas of the national park burn twice per year. Observations suggest that areas burnt in February may be preferred by migrant ungulates during June, but beyond these patterns and the immediate, short-term responses of grazing animals, we understand relatively little about how fire impacts the ecology of Serengeti or other East African protected areas in the bimodal rainfall area. Whilst we know a little about the impacts on large mammals and early vegetation responses, we know almost nothing about the impacts on other taxa, or the soil fauna and nutrient flows of the savannah ecosystem.

In practice, over much of Serengeti, fires occur once or twice per year. If there are species within the ecosystem that are rather more sensitive to fire than others – for example, ground nesting birds, toutoises or various plant species – such frequent fires may have negative impacts on the population. On the other hand, burning in different seasons may affect different species in different ways. Whilst it is generally considered 'a good thing' to burn Serengeti whenever it can burn, maybe reducing this frequency doesn't actually have the negative impacts expected and could, instead, benefit other taxa not usually considered in protected area management plans? How many unburnt seasons does it take before grass quality starts to decline? Would a general switch back to late-season burning be beneficial? The only way to answer these various questions is a detailed, but large-scale study of fire impacts in the Serengeti ecosystem. To date, there have been a few relatively small scale and short-term fire experiments in the ecosystem, there have also been longer-term but equally small scale studies in other African savannah systems. Both have provided insights into the management of fire in these systems, but only in Australia (where herbivory plays a completely different role) have experiments been carried out on a large scale in savannah ecosystems, providing a different range of insights to those of small scale experiments. Large-scale and long-term management trials in an area of bimodal rainfall will provide insights of relevance not only for the Serengeti ecosystem, but also for savannah ecology in general.

So, what am I actually planning to do? It's all fairly applied. My aims are (I hope) clear:

1. To understand the impacts of alternative fire management options on the ecology (both pattern and process) of the Serengeti ecosystem
2. To be able to inform managers on the best way to use fire as a management tool to achieve their stated aims.
3. To understand the importance of fire and grazing in shaping the ecology of the Serengeti ecosystem across the rainfall gradient.

So it's really a very applied experiment: I want to change burning patterns from what could be currently descirbed as 'burn-if-you-can' to a number of alternative strategies that are also perfectly manageable for the ranger in the field:

Frequency treatments:

'Maximal burn': corresponding to a burn-if-you-can treatment as carried out across much of the National Park, with fires set whenever possible.

'Hold-back': corresponding to a management regime that notes when burning would be possible but delays by one or two seasons until fuel loads are higher.

'Minimum burn': corresponding to a management regime that attempts to suppress fire return rates to 4-5 years, a frequency that has been suggested to be suitable for recovery of riverine forest and thickets.

Seasonal treatments:

'Early burn': corresponding to the long dry season management regime across most of the National Park, where fires in the long dry season will be permitted at the end of the rainy season – in northern and western areas of Serengeti (e.g. the western corridor) this will be before the migration arrives.

'Late burn': corresponding to the type of management traditionally carried out by pastoralist communities, with fire suppression during the first part of the long dry season and burning set in anticipation of the rains.

'February burn': corresponding to areas where management will attempt to burn in the short dry season and suppress fires in the long dry season (only northern and western Serengeti).

And where will we be doing this? Well, the study sites are burn management units within the TGT concessions of Maswa Game Reserve, within the Mwiba Holdings private concession around Makau Village and on Grumeti and Ikorongo Game Reserves and the adjoining Wildlife Management Area. These areas include a near complete range of Serengeti ecosystem habitats and rainfall gradients, from dry Acacia – Commiphora woodland and short grass plains in Mwiba Ranch, to Terminalia - Combretum woodlands and tall Themida grasslands in the Ikorongo and Grumeti Game Reserves. In total, the area managed within these three reserves is in excess of 2700km² (270,000ha), making this study easily one of the largest experiments in the world. Burning blocks within each reserve will be assigned to fire treatment, stratifying for size and rainfall across the entire set of study areas.

In addition to the experimental work in the reserves surrounding the Serengeti National Park, we will conduct similar monitoring activities at sites within the National Park, to ensure that the 'maximum burn' treatments within the experimental areas show ecological dynamics similar to those within the National Park.

So, hope that fills you in with lots of details! If you've got this far you're obviously really, really interested so please do go to the project page and support us! Every little bit helps...

Breaking the silence...

Hello everyone, and apologies for being offline for so long! I've been running around like a mad thing - Manyara, Loliondo, Ndutu, Cape Town, you name it, I've been there! I'm now back home and full of ideas for lots of new posts to keep things rolling. Whilst I sort out my photos and 1000s of emails I thought I'd bring to people's attention one thing I'm up to next. Regular readers will know I have a thing about fire in savannas and may even have picked up my plans to do a large-scale burning experiment around the Serengeti NP (I mentioned it here...). Well, the project has been officially approved and will be happening, but we're trying to raise some money to help get started and we're trying out crowdfunding as an experiment. What's crowdfunding you ask? Go and find out here! Using the newly set up website called Petridish, we're trying to get some start up funds in the next few weeks - there's opportunities for anyone to donate online, any amount, and some fun rewards to donors too! There are loads of other projects all looking for funding on the Petridish site, so we need to get some momentum going and then the site will push us harder too. I'll post some of the rationale for the experiment and more details here soon, but please do spread work about this and make a donation, however small, as soon as possible! There's a deadline on donations, and if we don't get the money we've asked for by then all the money goes back to the donors... Launching today, we've still got a long way to go! So, share the word, donate and (if you can bear it) watch me explain the project on the little video!

Here's the link to read about the project, see the pictures and video and make a donation. Thanks!

What is the savanna biome?

Savannas around the world are often open woodland like this Serengeti pic.

Remember a few months ago I reported on some talks I'd enjoyed at the ATBC/SCA-Africa meeting here in Arusha? Well, the research behind one the the talks I enjoyed most was published late last year and I think it really helps explain what makes the savannah biome (which is probably why it was published in Science - great job Carla!).  However, before going into too much detail about this research, I think it's important to make sure we know exactly what I mean by a biome in the first place...

When I talk of the savanna biome (or, indeed, the savannah biome, since I am English) I'm not referring to a habitat like grassland, or Acacia woodland; nor am I referring to a specific ecosystem like the Serengeti Ecosystem. Rather, I'm refering to the set of habitats that make up the savannah biome globally - the collection of grassland and woodland types that all have the same main processes operating on and in them, the savannah big four are, of course, nutrients, water availablity, herivory and fire. They might (and often do) contain remarkably different species, but the same processes are at work and, to remarkable degree, they show largely similar vegetation forms and structures. A biome may therefore be thought of as a set of habitats that share similar ecological processes wherever they occur across the world. By contrast, I tend to define an ecosystem as a single geographical area (like the Serengeti Ecosystem), within which nutrients are cycled with relatively little input or output from neighbouring ecosystems. Hope that's clear...

Fire is crucial to maintain savanna, particularly in wetter areas. Tarangire July 11

So, if we're going to understand what the savanna biome is, we need to look at the processes that are the dominant forces within it. If we can understand these, then we can predict where savannahs will be found throughout the world - and we can also understand what might happen to savannahs if we humans mess around with these processes through, for example, the effects of climate change. And this is exactly what Carla Staver and her co-authors have done in their paper (which I'm afraid is probably hiding behing the paywall here). They were particularly interested in discovering what determines the boundary between forest and favanna biomes (the boundary with desert is a much more obviously rainfall driven boundary), and decided that the most likely factors were rainfall and fire. Whilst that might well be true, there's plenty of evidence that herbivory is a major player too (at least here in Africa) and nutrients have been proposed as major players too (though there's also plenty of evidence to suggest this really isn't the main thing). However, they didn't look at these - herbivory they mention but exclude from their analysis with the very reasonable excuse that we just don't have a good idea how much herbivory there is around the world, but I think it's a bit of a shame they didn't have a go at nutrients too. Still, enough of what wasn't looked at - what did they find?

The Serengeti Story, part 1: history

So I guess this is the post I've been putting off longest. Not because it's not interesting, but because I know I'm going to forget some crucial component. But I'm just back again from a fantastic trip (thanks to all the guys at Dunia!) and decided it's definitely time to bite the bullet. However, it's going to be a long story, and I'm going to split it in two sections so I don't spend all night here (and so I stand a chance of remembering what I've forgotten before I consider the story told!). If you want more details on any of these things the essential references are the excellent series of very technical books edited by Tony Sinclair and colleagues you can get from Amazon. I've cut and pasted a few of the graphs from 'Serengeti III' into this post, hopefully 'fair use' for education...

I always start telling the Serengeti story with a bit of history, since it helps us understand how scientists have uncovered some of these things. There's no really obvious beginning to the story, but let's start with something we've already discussed on Safari Ecology - the introduction of Rinderpest to Africa in 1887. As we saw in that post, this had a massive impact on wildlife throughout Africa, the disease reaching Cape Town by 1897. The Serengeti migration was decimated, and when it was finally erradicated from the wildebeest population in 1963, there were still only around 250,000 wildebeest (see the plot below).
As you can see, once rinderpest was erradicated the wildebeest population exploded, reaching it's current total of somewhere betwen 1.2 and 1.4 million in about 1977, and this is the huge change that has let us understand so much of what happens in Serengeti.

Now, by now we should all know the 'Big 4' of savannah ecology, so it shouldn't come as a surprise that such a huge change in herbivory had a massive impact on the ecology of Serengeti, perhaps most obviously on the amount of another of the big 4 - fire. The figure below shows very clearly how the rise in numbers of wildebeest reduced the amount of fire in those northern woodland areas (essentially the woods from Seronera north).

This is clearly down to the very simple fact that wildebeest eat grass and grass is what carries fire through the savannah - more wildebeest means less grass which means less fire. And a change in the fire regime, of course, will alter the ecology too. So introduction around 1890 and then erradication of rinderpest in 1963 led to a massive change in both grazing pressure and fire frequency. It's not surprising, therefore, that massive changes occurred in Serengeti during the 1900s, most obviously the change in woodland cover. If you dig through old photos of the Serengeti / Mara area you can find some fantastic images of change. Tony Sinclair did it and came up with this beauty from 1944, that he then returned to in 1983 and took the subsequent photo (I've borrowed them from his talk available online here).

 It's pretty obvious that the woodlands vanished sometime between these two photos were taken and more detailed work suggested a rapid decline in woodland cover from about 1945 to 1980 - just the sort of delay you might expect from the increase in fire around the turn of the 1900th Century, given that fire doesn't kill savannah trees above 2m tall, so any established trees would gradually die of old age some time later.

Interestingly, as a direct consequenc of the decline in trees the national park authorities changed their fire management strategy in the 1970s from late burns at the end of the dry season and in anticipation of the rains, to one of early burns which tend to be cooler and rather less damaging to tree seedlings. At the same time, of course, the wildebeest population was recovering and the fire was declining in frequency as a consequence, so this change was probably less necessary than it seemed at the time (though everyone at TANAPA has since forgotten that the current fire strategy is a relatively new one, of course!). And as you might expect, more recently the trees have returned. Again, Tony Sinclair has some fantastic series of photos of these changes too, this from relatively close to Seronera:

(There's a whole lot more of these sorts of photos available on the web if you search for Tony's various talks.) And so the woodlands returned to Serengeti, as a consequence of the return of wildebeest and subsequent decline of fire. [It's interesting too, that savannahs globally are getting woodier, so there's a chance that this change is also related to global change too, not simply a local Serengeti effect - we might return to this in the future...]

But the story's not quite complete yet, as there's a neat twist at the end involving elephants. During  the 1970s and 1980s there was massive and nearly uncontrolled poaching of elephants throughout Serengeti, ending abruptly with the band on ivory trading in 1989. It's had a massive impact on elephant numbers in Serengeti:

At the same time, however, across the border in Kenya poaching remained under tight control, with no such dramatic change in elephant numbers. Such large herbivores can have a massive impact on the vegetation and the story in Serengeti is a particularly interesting one - Elephants walking across grassy plains often 'weed' out the tree seedlings instead of eathing grass. In woodlands they tend to leave the seedlings and concentrate on adult trees. So if there are lots of elephants it can be rather hard to turn grasslands into woodlands, even if the fire frequency is reduced. The difference between Kenya, where elephant numbers remained high throughout the period, and Tanzania, where they crashed at just the same time the fires declined, is stark. And elephants being rather clever animals, they knew where the border was and they were safe. So here's one last picture of Tony's from northern Serengeti / Mara, where the international border is clearly defined by woodlands.

Amazing to see the impacts of elephants so clearly, but also amazing to see how two different habitats (grassland and woodland) within the savannah biome can be stable under exactly the same environmental conditions - these days elephants are common both sides of the border and yet the woodlands remain in Tanzania, thanks to the different way elephants behave in grasslands from woodlands. So the history lesson ends with an important lesson about how important the initial conditions are to how a savannah looks - to turn a grassland to a woodland you need to reduce fire frequency (which can be done by increasing herbivory), but you also need to at least temporarily exclude elephants. All very complicated...

So, that's the history lesson and the broad overview of some population changes as a whole. The next post will continue the Serengeti Story by, I hope, explaining what we know about the migration and the regional differences across the ecosystem today. Hopefully it won't take so long to create either!

Burning Mwiba

Back from my trip now, I thought I'd make a few posts based on the things we've seen on safari. As I've been writing up my thoughts about one place already today, I thought I'd use that as the basis for this first post.

Buffalo are a major grazer in Mwiba, note the relatively short grass.

Mwiba Game Ranch is a new private game reserve within the Serengeti ecosystem. I've been before when I put together a bird list for the area. This time it was a trip to see the place in the dry season with a view to including it within my big Serengeti Fire experiment. For those who know Serengeti , Mwiba is squeezed into the corner between NCA and Maswa GR, right down in the south of the ecosystem. (NB, we usually define the Serengeti/Mara ecosystem as the area that encompasses the wildebeest movements - Mwiba includes some of the calving grounds, particularly important during drier years.) This puts it right in the driest region of the ecosystem, with around 400mm of rain per year and as you'd expect at that end of the gradient it's largely Acacia-Commiphora woodland, though there's a surprising number of nice Albizia in there too. It's also interesting because it's got a number of interesting mammals not found or not easy to see in the rest of Serengeti - we saw both Greater Kudu and Roan Antelope again this trip. Anyway, I was there to talk fire, but knowing that we're in a low rainfall part of the ecosystem is important, because water availablity is one of the big four drivers in the savannah (fire, grazing/browsing and nutrients being the other three, of which we'll visit two more shortly). Low rainfall means low productivity - the grass even on the highest nutrient soil never grows tall and thick like in other parts of the Serengeti, but what does grow tends to be nutrient rich annuals, so pretty good grazing, even if it isn't plentiful.

Zebra are the other big grazer - the grass here has already been grazed a bit

Rich grass means plenty of game, with the main dry-season grazers being large populations of buffalo and zebra thanks to the numerous perrenial springs around the ranch. Already, only half way through the dry season the grass in the areas around the waterholes and by the denser thickets is heavily grazed - by October it seems unlikely there'll be much left at all as the grazing impact spreads further from the water points. A lot of the area is pretty dense bush though, with some good thickets in places along the (seasonal) rivers.

Nearby areas with many cattle are already completely denuded, what will they do until the rains come?

 In recent years this area (as with most of Serengeti) has been subject to an early burning management regime - fires being set as soon as the vegetation starts to dry in June. Fires are important in the savannah for a number of reasons we've gone into elsewhere, but the two most important issues to bear in mind here are bush control - there were lots of seedlings in the grassy areas of Mwiba that have been prevented from forming to thick bush by regular burning - and grazing management, maintaining and encouraging new growth of nutritious grass. The early burning policy that has become the norm in Serengeti and most other Tanzanian protected areas ensures fires are controlled and generally are rather cooler than fires set at the end of the dry season when the fuel is drier.  But here in the drier areas with lower water availability it also means the fires burn regularly - you can be fairly sure there's enough fuel to burn early in the dry season, whilst later on all those animals will have eaten so much there might not be anything left to burn. That's good if you want regular firest to control bush encroachment, but isn't so good if you happen to be a buffalo wanting to eat during the dry season and all the nice grass gets burnt at the start. In fact, the ranch manager is of the impression that this year, when for the first time in a long time no fires have been set during the early season, there are lots more animals on the ranch than last year thanks to the availability of unburnt grass.

There's a sand river that forms a firebreak between the grass and the bush - frequent early burns have removed thicket vegetation from the upwind side of the river.

Mwiba's springs attract a lot of wildlife (and reflect sunset)

So, what to do? Maybe late burns will be possible at a lower frequency - give the land a few years to build up sufficient fuel reserves left over at the end of the dry season to allow a fire to take with regular enough frequency to control bush encroachment. Or maybe early burns are the only option to allow fires in these low rainfall areas to carry - though they probably don't need setting each year, balancing fire options against forage loss. The only real way of finding out will be through a big experiment, of course, which is exactly what I intend to do! Watch this space for the answers...

Fire!

I've been thinking a lot about fire recently. Burning is, of course, Tanzania's national pass-time in this season, and is one of very few active management activities occuring in protected areas, and we found a nice fire to admite in Tarangire on our safari a couple of weeks ago.

Lots of smoke, and some big flames as a fire goes through bushes, Tarangire June 2011

But again, I've been pretty surprised by how little guides know about fire in savannas. I'm currently trying to put togher a big experiment that wll see us burning lots of areas around the Serengeti National Park (it burns every year anyway, all I want to do is tinker with the management in a controlled way to understand the impacts). As part of thinking about this I organised a meeting at the ATBC-SCB confere of lots of people interested in both the Serengeti ecosystem, and of those interested in fire throughout the savanna world. I also tracked down a number of people in South Africa when I was there in May to get their thoughts. As part of all this thinking, I've put together a document that pulls together lots of ideas, so I thought I'd share some of the introduction to that here:

Fire is generally considered a vital component of savannah ecology, with trees, grasses and animals all adapted to a fire prone ecosystem: in fact, globally, 85% of fires occur in savannah habitats. In most of Tanzania's protected areas, rangers deliberately set fires each year for a number of different purposes, including the encouragement of new grass growth for grazers and the control of bush spread. Recent evidence from a global study suggests that at least for those savannahs occurring in areas with over 1000mm of rainfall, the forest/savannah edge is often maintained primarily by fire, and that in it's absence many savannahs will revert to forest. There is some debate, however, about how much animals may take over this role when at particularly high densities and the benefits of fire in certain areas have been questioned. For example, in some areas of Tarangire National Park with high numbers of migrant animals during the dry season, fire has been totally suppressed for over 30 years with relatively little obvious difference between fire-suppressed and more frequently burnt areas: it is argued that in these areas fire simply burns potential animal forage. Similarly, in Kenya, the Kenya Wildlife Service has recently instigated an overall policy of fire suppression except where the fire is caused by natural events (essentially, lightening storms) in an attempt to encourage one view of naturalness. There are, of course, alternatives to these two extremes, where fire is sometimes suppressed and sometimes encouraged. Given that fire is one of very few activities under active management in these areas it seems sensible to understand more fully the consequences of these different management decisions.

Note the fire burns the grass, bigger bushes are untouched.

Traditionally, pastoralist communities in northern Tanzania have set fires towards the end of the dry season, in anticipation of the rains: they make a calculated gamble between burning possibly useful forage if the rains are delayed, and waiting too long and being unable to burn (and encourage the new growth that will come with the rains) if the rain comes before they have time to burn. Obviously, by this time of the year grazing animals will already have reduced fuel loads to relatively low levels in many areas, so late dry season fires will only occur in less grazed areas, with other areas remaining unburnt in any one year. Such management was also the norm in many protected areas (including Serengeti) until around 1970, when concern over the regeneration of trees prompted a switch in fire management towards controlled burns in the early dry season (mainly in June). Such fires are generally cooler and patchier, with possibly less impact on woody plants. They also prevent the late season burns that are likely to be hotter and less easy to control with consequent reduced concerns about tourism infrastructure. Recently, managers of the Grumeti Reserves in western Serengeti have attempted to suppress early season fires until later in the dry season, allowing the migrant wildebeest and zebra to graze unburnt areas and increasing the forage available to these animals to the extent that their residence times within the reserves have tripled. Fires are also set during the short dry season, in February, and grass growth rates are such that some areas of the national park burn twice per year. Observations suggest that areas burnt in February may be preferred by migrant ungulates during June, but beyond these patterns and the immediate, short-term responses of grazing animals, we understand relatively little about how fire impacts the ecology of Serengeti or other East African protected areas in the bimodal rainfall area. Whilst we know a little about the impacts on large mammals and early vegetation responses, we know almost nothing about the impacts on other taxa, or the soil fauna and nutrient flows of the savannah ecosystem.

In practice, over much of Serengeti, fires occur once or twice per year. If there are species within the ecosystem that are rather more sensitive to fire than others – for example, ground nesting birds, or various plant species – such frequent fires may have negative impacts on the population. On the other hand, burning in different seasons may affect different species in different ways. Whilst it is generally considered 'a good thing' to burn Serengeti whenever it can burn, maybe reducing this frequency doesn't actually have the negative impacts expected of it and could, instead, benefit other taxa not usually considered in protected area management plans? Quite how many unburnt seasons does it take before grass quality starts to decline? Would a general switch back to late-season burning be beneficial?

Lots of birds were making use of the fire to catch fleeing insects, including this lilac-breasted roller - both animals and plants in the savannah are adapted to fires.

There's lots more to fire in savanna too that's of interest to guides and their clients (like what do the animals actually do?!), so we'll revisit the topic. But maybe the ideas here will help some understand why I'm so interested in fire. It's not just that I want a bonfire big enough to see on Google Earth. Honest!

Savannah Ecology

Most East African safaris spend a lot of time in the savannah biome. Forests and coastal areas are also popular, but the savannah is where the safari focussed and a basic understanding of the ecology of this biome will make a visit much more interesting. You can read more about savannahs and the savanna biome here on Wikipedia, of course, and there's a large team making sure that post is up to date. But I like to break into the subject rather differently so will do my own thing here, with future posts picking up the threads we identfy here.

Let's start by defining the savannah biome. Note first that I'm trying to be careful to talk of a biome here, not simply a habitat - the savannah biome is made up of many different habitats from grasslands and woodlands, to kopjes and swamps. Each of these habitats (and others) play an important role in the savannah biome and we'll visit them individually in future posts. In fact, the biome is defined as a grass dominated system - the grasslands are obviously part of the savannah ecosystem, but the woodlands and other habitats also have an understory dominated by grasses. The two photos above show typical grassland savannah from Kruger NP (South Africa) in the top (plus White Rhino) and an Acacia woodland (plue Oryx) with thick grassy understory in Tarangire NP (Tanzania). Other savannahs might looks less familiar to East African safari types - check the nice shot of a Guinea savanna in West Africa here, and the interesting savannah woodlands of Australia here. All savannahs, as all grass dominated ecosystems.

Right, definition out of the way it's time to introduce the Big Four of the savannah (sorry, moved on from the original three, but still can't make five!) - the four processes that shape the savannah biome globally. With an understanding of each of these, you can start to understand savannah ecology and begin to guess at what drives the patterns you see in this biome.

Firstly, there's climate and particularly water availability. Temperature and rainfall/precipitation combine to define the earth's major biomes - to get savannah, you need to be warm and fairly dry. Too wet and you'll end up with a forest of one type or another, to dry and you'll head rapidly towards desert. In fact, globally the savannah biome tends to dominate in tropical areas with rainfall above about 400mm, and below something between 1400 and 1650mm. Within this range, depending on how the other big processes combine, you'll probably get savannah habitats of one form or another - though how they look depends exactly where you are on the rainfall gradient. And, of course, understanding seasonal rainfall patterns are vital to understanding the seasonal movements of wildlife.

Three processes in one! Wildebeest near Naabi in Serengeti are gathering in the rainy season when surface water is drinkable to graze the nutrient rich grasslands of the short-grass plains. The impact of so many grazers is extreme!

Secondly, there's the impact of animals - grazing and browsing in particular. Savannahs are often full of animals - it's why people come to visit after all! And the impact of all those animals is not to be underestimated - take them away and the savannah can change dramatically from grasslands to woodlands and even (depending on the other processes) forest. Animal impacts can be seen all over the savannah and again, we'll visit these issues in subsequent posts.

Thirdly, there's fire. Savannahs burn and always have done so - today, many fires are deliberately set as part of the management, but people have probably been burning savannahs as long as there have been people around and before that lightning would have set fires naturally - probably about every 3-6 years we think. This is an ecosystem that has evolved with a constant presence of fire, the trees regrow, the grass regrows and (most) of the animals are perfectly capable of escaping fires by running or hiding in holes, etc. But fire frequency and intensity can certainly shape the savannah and it's a vitally important process to understand.

Finally, there are nutrients. Many savannahs are found on ancient and highly nutrient poor soils where every little patch of nutrients will be highly valued by something. Other areas are on recent volcanic and nutrient-rich soils, providing ideal grazing opportunities and different niches for vegetation types. Where nutrients are found (and how they get moved about) dramatically shapes the ecology of the savannah biome from the small scale of termite mounds to the larger scale of soil types, determining seasonal patterns of movement for animals and many of the habitat differences found from place to place.

And that's it! Future posts will develop all these issues further, but it's a great start in savannah ecology to have in mind the processes that shape the biome before we look too far at each one.