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...