The frankincense tree of Ethiopia : ecology, productivity and population dynamics

Keywords: Boswellian papyrifera, Frankincense tree, matrix model, population dynamics,
population bottleneck, tapping.

Combretum – Terminalia woodlands and Acacia – Commiphora woodlands are the two
dominant vegetation types that cover large parts of the dry land areas in Ethiopia. Several of
their tree and shrub species yield economically valuable products such as gum Arabic,
frankincense and myrrh. Boswellia papyrifera provides the widely traded frankincense that
accounts for >80% of the export revenues that the country is earning from gum and resin
resources. Unfortunately, the Ethiopian dry woodlands and the B. papyrifera populations are
disappearing rapidly due to the combined effects of over-harvesting gums and resins,
overgrazing by livestock, recurrent fires, and excessive wood harvesting. The current lack of
small saplings in the remaining populations of Boswellia suggests that the populations may
not be sustained for the future.
The main objectives of this thesis were to determine diversity and production patterns in B.
papyrifera dominated dry woodlands, to show the regeneration status in various B. papyrifera
populations, and to evaluate the effects of environment, frankincense harvesting, and grazing
on the population dynamics of B. papyrifera. The main research questions were: (1) how do
environmental conditions affect the tree/shrub species richness and production of Ethiopian
dry woodlands? (2) what factors determine the frankincense production by B. papyrifera
trees? (3) how do the vital rates and population dynamics of B. papyrifera vary across
habitats that differ in soil conditions and biotic factors? (4) What are the major bottlenecks in
the life cycle of the trees that hinder the sustainability of the remaining populations? To
address these questions, tree populations were studied in the highlands of Abergelle and the
lowlands of Metema. Metema also has a longer wet season length, higher annual rainfall and
better soil fertility status than Abergelle.
In total 36 and 22 tree and shrub species representing 20 and 9 families were recorded in
Metema and Abergelle woodlands, respectively. The most dominant plant families were
Burseraceae, Fabaceae, Combretaceae and Anacardiaceae. The vegetation at both sites was
dominated by B. papyrifera. The two sites differed in species richness, diversity and
production. Metema, the site with the longer wet season, had a higher species richness,
diversity and production than Abergelle. The productivity of woodlands also increased with a
higher clay content and greater soil depth. Populations structures indeed lacked the saplings,
except for one very isolated population on a steep mountain slope.
The studied frankincense trees produced 41 to 840 gram of frankincense during a year with
seven collection rounds, and 185 to 1826 gram of frankincense during a year with 14
collection rounds. The variation in frankincense production was large across individuals.
Frankincense production increased with tree size, tapping intensity, and tapping frequency.
The increase in production, however, levelled-off beyond a stem diameter of 20 cm, a tapping
intensity of 9 spots, and a tapping intensity of 10 rounds. Growth rate, survival rate and
fruiting probability varied across populations, but were not related to soil conditions or biotic
factors. The growth rates of the 12 Metema populations varied between 0.86 to 0.98,
suggesting that they were all decreasing. Matrix model analyses indicated that the mortality
of adult trees was the major bottleneck for sustainable population growth, and that the lack of
sapling recruitment was a second major bottleneck. These bottlenecks appear both in tapped
and non-tapped stands. Remarkably, tapped stand showed higher growth rates than nontapped
stands, probably because productive stands were selected for harvesting resin.
All results suggest that the remaining populations of B. papyrifera will disappear in the near
future if the current situation continues. Frankincense production is expected to halve in 15-
20 years. Unexpectedly, tapping had no negative effect on vital rates, nor on population
growth rates indicating that other factors are responsible for the decline of the populations.
Adult mortality by insect infestation and windfall, and the negative impact of grazing and fire
on the establishment of saplings need extra attention. Management should be directed
towards releasing two major population bottlenecks (improve sapling regeneration, reduce
adult mortability) to maintain the Boswellia populations and frankincense production in the
future.