Why do savanna trees have flat tops?

Umbrella Thorn, Serengeti: An icon of the savanna?

From sunsets behind a silhouetted acacia (properly Vachellia), to photos of rolling grasslands studded with isolated trees, a savanna landscape is immediately identifiable thanks to the flat-topped tree. But why is this? Why do so many Vachellia and other savanna trees have such a distinctive structure that they have become a virtual icon of the African savanna?

It's an interesting question that was given some answers in a nice paper by Sally Archibald and William Bond who studied one species called the Sweet Thorn (Vachellia karroo) that, rather like some of our Vachellia species in East Africa exhibits a range of different growth forms in different habitats. In the semi-desert of the Karroo, it grows as a medium-sized ball of thorns, whereas in the savanna it has a fairly typical medium-tall  flat-topped acacia look to it and in a forest it's a tall, thin tree. These differences are meditated mainly by genetic differences within the species, but equally could be caused in other species by a variable response to the environment - it's not really important to this discussion and, in fact, much of our discussion could focus on different species if we wanted. As always when we're thinking about what makes the savanna species, we'd be well advised to start with the savanna big four: nutrients, water availability, fire and herbivory.Now, the first two processes have impacts in all biomes, whereas it's the second two that are most distinctive about savanna and where we'll start our discussion.

It's not only Vachellia species that are flat-topped. Here are a
range of species including Balanites aegyptiaca in Serengeti.

Any tree that wants to get established in a fire-dominated biome has to have some special adaptations to cope. Thick bark is an obviously useful trait and has evolved repeatedly by trees that invade savannas globally. But another sensible trait is to grow rapidly tall: trees that are above the 'fire trap' (2-3m tall) aren't burnt back to the roots by a fire, but can resprout from the top. If you've got limited resources to use and live in a place where fires are frequent, it makes sense to put those resources into a single (or at most a few) stems and grow straight up, not branch out sideways. Once you've grown tall enough to escape the impact of fire, you're free to branch sideways. So there's one good reason why in a savanna you might find lots of tall trees with few branches low down then lots more above.

Senegalia mellifera is often in drier areas and is typically a 'ball of thorns'.

Next, consider herbivory. Growing tall isn't easy if you're a young tree and there are lots of browsers around. In such circumstances you'd be well-advised to put a lot of effort into protecting yourself with loads of thorns and grow in a thicket type growth form. By growing wide enough, you might eventually manage to protect some internal branches enough to completely free them from browsing pressure and they could then grow tall if that's useful. So in a browser dominated habitat, branching low down and frequently might be a much better strategy. As we know, in the savanna fire is carried by grass, so in areas where rainfall is low and grass growth small, we'll find a browse dominated system - and it's those areas where William and Sally found ball-shaped thorn trees. (Of course, in wetter areas where there's both fire and herbivory, it's a really struggle for any tree! In such circumstances I guess the certainty that all seedlings will be killed by fire if they're not tall enough, versus the possibility - however small - that a seedling my not be found by a browser would favour a fire-adapted response. Ideally backed up with some serious thorns and chemical defence of saplings!)

Forest on Kilimanjaro - trees here are tall and have
branches to catch the light passing through the top layer

Now, in contrast to a fire and herbivory dominated savanna, in a forest trees are cramped together and their main struggle is to get sufficient light to grow. Consequently, different tree species have evolved two main strategies to survive: they either tolerate low light and grow very, very slowly but live a long time and eventually make it to the top of the canopy and can shade out other less-tolerate species. Or they grow fast and die young, rapidly filling any holes that form in the canopy when a mature tree dies, but eventually being out competed by the slower-growing species. They also need to absolutely maximise any light they can get - even if that light has already passed through the top leaves, they want to catch remaining light lower down too, making the forest floor a rather gloomy place to be. The consequences of this intense light competition are therefore that (a) mature trees tend to be very tall, always struggling to get above their neighbours and (b) they have lots of leaves at all sorts of heights in the canopy, trying to catch every stray bit of light.

I love Vachellia tortilis! Another in Serengeti.

So, in a typical fire-dominated savanna, once you've escaped the fire trap you're free to branch. But it's so rare that a seedling breaks out of both fire and herbivory, that mature trees tend to be at low density in savannas, with canopy cover in woodlands around 20-30%. And under such circumstances there's no going to be any significant competition for light - so there's no need for savanna trees to grow tall, and horizontal growth is enough to gather sufficient light. What's more, even moderately tall trees don't always escape giraffe browsing, but a horizontal growth-form still protects the central branches - in fact, it's common to see large Vachellia tortillis that only grow large thorns on the outer tips of their branches, providing enough defence for protect inner branches without requiring further costs of thorn growth. So, in a fire dominated, browse affected and unlimited light environment, the flat topped tree structure is exactly what you would expect to see evolve!


Main reference:

Archibald, S., & Bond, W. (2003). Growing tall vs growing wide: tree architecture and allometry of Acacia karroo in forest, savanna, and arid environments Oikos, 102 (1), 3-14 DOI: 10.1034/j.1600-0706.2003.12181.x