Content uploaded by Markus Lindholm
Author content
All content in this area was uploaded by Markus Lindholm on Jan 29, 2014
Content may be subject to copyright.
Aquat. Sci. 68 (2006) 310–337
1015-1621/06/030310-28
DOI 10.1007/s00027-006-0857-y
© Eawag, Dübendorf, 2006
Aquatic Sciences
Overview Article
Species diversity of the Okavango Delta, Botswana
Lars Ramberg1,*, Peter Hancock2, Markus Lindholm3, Thoralf Meyer1, Susan Ringrose1, Jan Sliva4,
Jo Van As5 and Cornelis VanderPost1
1 Harry Oppenheimer Okavango Research Centre, University of Botswana, P. Bag 285, Maun, Botswana
2 Ecotourism Services, Maun, Botswana
3 Department of Biology, University of Oslo, Norway
4 Munich Technical University, Munich, Germany
5 Department of Zoology and Entomology, University of the Free State, Bloemfontein, South Africa
Received: 15 June 2005; revised manuscript accepted: 28 February 2006
Abstract. In the Okavango Delta (about 28,000 km2) the
number of identifi ed species is 1,300 for plants, 71 for
fi sh, 33 for amphibians, 64 for reptiles, 444 for birds, and
122 for mammals. The local occurrence of different spe-
cies of these taxonomic groups in the Okavango Delta is
mainly due to a hydrological gradient from permanent
streams and swamps to seasonal fl oodplains, riparian
woodlands, and dry woodlands. This level of species di-
versity is normal for the southern African region, and all
analyzed aquatic groups are composed of ubiquitous spe-
cies with an additional signifi cant proportion of species
originating from northern, more tropical systems. Cycli-
cal variations in climate over thousands of years have cre-
ated a huge wetland complex in the upper Zambezi and
Okavango Rivers during wet phases. This wetland com-
plex has fragmented into the Okavango Delta and other
large wetlands in Zambia during dry phases. There are no
endemic species in the Okavango Delta while the South-
central African wetland complex is a centre of endemism.
Species diversity of the Okavango Delta is a consequence
of this unique environment, with dynamic shifts in fl ood-
ing patterns that in turn force constant changes in patterns
of plant succession and dependent animals. Temporal
variations in fl ooding also cause accumulation and sud-
den mobilization of nutrients which are readily used by
well adapted plant species. As a consequence, locally high
biological productivity occurs, which in turn results in
high numbers of grazing mammals.
* Corresponding author phone: +267 6861833; fax: +267 6861835;
e-mail: lramberg@orc.ub.bw
Published Online First: October 6, 2006
Introduction
The Okavango Delta in northern Botswana has a large
variety of aquatic, wetland and terrestrial habitats and the
rich typical African bird life and wildlife and its pristine
beauty enthrall anybody visiting it. It is often said that its
biodiversity is “high” but the opposite statement is also
common; that it is “normal” or “low” for the Southern
African region. However, there have been no systematic
attempts to analyze Okavango Delta biodiversity, the fac-
tors that are causing and regulating it, and this is the fi rst
attempt to bring all available facts together.
Physical description
The Okavango River originates on the Angola highlands,
fl ows through the Caprivi strip in Namibia and ends in
the Okavango Delta or during wet periods in the large
Makgadgadi saltpans in the Kalahari (Fig. 1). Due to low
Key words. Biodiversity; species; habitat; Africa; wetlands; Okavango.
Aquat. Sci. Vol. 68, 2006 Overview Article 311
gradients the water that fell as rain in Angola in Novem-
ber arrives in the upper Delta (the Panhandle) around
February–March and moves slowly as a huge wave across
the wetland landscape until it reaches the distal parts in
July. There is, however, another localized wet period
caused by rains occurring in December–March. The Del-
ta thus has two fairly predictable wet periods and is a
typical fl ood-pulsed system usually with one fl ooding a
year. However, during years of heavy rainfall over the
Delta extensive fl ooding can already occur in January
and continue until the second fl ood peak arrives in April–
July. This has happened once during the eight years the
Harry Oppenheimer Okavango Research Centre (HOO-
RC) has been working in the Delta, when seasonal fl ood-
plains which were normally water covered for 3–6 months
were continuously under water for about 12 months in
the year 2000. In the upper permanently fl ooded parts of
the Delta water level variations between and within years
are usually small (Fig. 2), being less than one meter,
while these variations in rivers in the seasonally fl ooded
areas can be 1–2 m. The fl ooding of fl oodplains, which
often are on a lower altitude than the feeding stream, usu-
ally takes the form of an overfl ow of river banks that
function as thresholds. The maximum annual water level
variations in the deepest parts of these seasonally fl ooded
areas can be 3–4 m.
The mean annual infl ow to the Delta is 9.2 × 109 m 3
and rainfall contributes an additional 6 × 109 m 3 (McCa-
rthy and Ellery, 1998). Only 1.6% of this leaves the Delta
as outfl ow through the Thamalakane River (Figs. 2 and
3). It is estimated that the loss through regional ground-
water outfl ow is less than 2% and probably nothing
(Gieske, 1995); consequently 96–98% of total infl ow is
lost through evapotranspiration within the Delta.
Figure 1. The location of the Okavango River Delta on the Southern African sub-continent with adjacent major rivers and wetlands. Note
that the Okavango River basin is located almost entirely within the Kalahari (sand) Basin and that the river ends in a large salt pan, the
Makgadikgadi. The river basin is shared by three states.
Figure 2. A. Mean annual rainfall in the Okavango Delta. B. Water
level variations during two contrasting years, (dry 1998 and wet
2001) in the upper permanently fl ooded area, and in the Boro River
in the seasonally fl ooded area. C. Infl ow and outfl ow pattern.
312 L. Ramberg et al. Okavango biodiversity
The Delta has a typical continental climate with the
highest daily maximum temperature of 34–35 °C in Oc-
tober and the lowest of 25 °C in July. For this month the
mean minimum temperature during the night is however
only 8 °C (Mendelsohn and el Obeid, 2004), frosts are
locally common and thin ice forms occasionally on the
shallow waters in the Delta. The mean annual rainfall is
460 mm in the south and 490 mm in the northern part of
the Delta, and the evapotranspiration rate is about
1,500 mm (Dincer et al., 1987). As in other areas close to
the tropics of Cancer and Capricorn the variations in rain-
fall between years are very large (Nicholson, 2000).
The Okavango Delta is located between a series of
fault lines (Fig. 3) that form the southwestern extension
of the East African Rift system (Cooke, 1980; Modisi et
al., 2000). Seismic activity in this area started about 2.5
million years ago, which is also the approximate age of
the Delta (Tiercelin, 2003). The area between the faults
appears to have been subsiding and tilting through time
causing the infl owing river to be un-confi ned laterally
causing it to branch out in a number of dispersed distribu-
taries (Fig. 3). Sediment transport in these has caused the
deposition of an alluvial fan that is slightly conical and
has a low gradient, 0.00036 along the main direction of
fl ow (McCarthy et al., 1997).
The extent of alluvial deposition has varied widely
over geological time and the palaeo-delta was two to
three times larger than at present. This is caused by large
swings in climate from very wet periods to dry ones (Ty-
son et al., 2002). The development of the present delta-
fan appears to have been preceded by an extensive dry
period with east-west sand dune formation, whose re-
worked sediment has contributed largely to the delta
sands (Stokes et al., 1997; Ringrose et al., 2002). The
actual size of the Okavango Delta is very much a matter
of defi nition and large differences in size can be noted
between authors. Gumbricht et al. (2004) used images
taken during the last 30 years and gave as a summary for
this period the total area fl ooded at least every decade to
14,000 km2, of which 9,000 km2 is actual wetland, the
rest being islands. The total areas of the “Okavango wet-
land” are given as panhandle: 820 km2, permanent
swamp: 2,500 km2, seasonal swamp: 3,300 km2 and oc-
casional swamp (fl ooded at least each 10th year) to
7,100 km2. Permanently dry areas not included in this
classifi cation forming islands or peninsulas are about
4,000 to 10,000 km2. By including these areas and areas
that have been fl ooded during historical times (1850-
present) (Fig. 3), the total area of the Okavango Delta
thus defi ned is 28,000 km2.
The major swings in Delta size appear to have been
due to a combination of climatic changes and tectonic
shifts brought about by faultline reactivation. Evidence
from old fl oodplains suggest that early alluvial fans may
have been fed from the Kwando and now ephemeral
rivers to the west, while later fans accumulated from a
pre-Okavango system and were deposited along a north
easterly and south easterly trends in the the Magkadig-
kadi-Okavango-Zambezi (MOZ) basin (Ringrose et al.,
2002). A system resembling the present Okavango prob-
ably developed over the last 40,000 years, as a result of
movement along the Thamalakane and Kunyere faults
(Ringrose et al., 2005). Intervening fl uctuating years of
high and low rainfalls were prevalent but diffi cult to track
because of “anti-phase” conditions in the Delta (Hunts-
man-Mapila, 2005). A wet period appears to have been
prevalent during the regionally dry Late Glacial Maxi-
mum (Partridge et al., 1997; 1999). Periods of high fl ow
interspersed with low fl ow conditions through the Oka-
vango system 13,000–14,000 years ago, and intermittent
seasonal fl ow occurred 6,300 years ago. This was proba-
bly the last time when all the major distributaries were
fully fl ooded. By this time the major pattern of rivers,
large peninsulas and islands was formed thereby provid-
ing geomorphic controls for the landscape diversity
present today (Anderson et al., 2003).
The geological history of the Zambezi and Okavango
rivers, recently reviewed by Timberlake and Childes
(2004), explains much of present biodiversity. There is
both geological and biological evidence that the Okavan-
go River and Delta has been directly hydrologically con-
nected to the upper Zambezi and to the Kafue Rivers
(Fig. 1) forming one of the major wetland systems
throughout the MOZ basin (Ringrose et al., 2005). The
link between the Zambezi and Makgadigadi seems to
have been breached and re-established on a number of
Figure 3. The Okavango Delta with major fault lines; major streams;
distribution of permanent swamps, seasonal and occasional grass-
lands and woodlands; and the buffalo fence. The total area is
28,000 km2.
Aquat. Sci. Vol. 68, 2006 Overview Article 313
by woodlands and savannas and without any very distinct
sub-divisions although taxonomic gradients occur in par-
ticular from south to north. White (1983) however, identi-
fi ed the “Zambezian Phytochoria” based on the number
and proportion of endemic species, whose area predomi-
nantly includes the Okavango and Zambezi River basins.
This “regional centre of endemism” (op. cit.) has a number
of huge wetlands (Fig. 1) in its central part including the
Okavango Delta that during the recent past have all been
hydrologically connected to each other (Timberlake and
Childes, 2004; Timberlake, 1998).
It should be noted that there are no permanent waters
south of the Delta and the extensive area of dry Kalahari
savanna has probably functioned as an effective barrier
against migrations of aquatic organisms.
As the Delta is on the axis of the East African Rift,
there is a strong possibility of early hominid migrations
into the region up to several million years ago (Butzer,
1984). Indication of human life comes in the form of
stone tools found in several places (Butzer, 1984; Tho-
mas and Shaw, 1991) and produced between 200,000 up
to 35,000 years ago (Butzer, 1984; Mendelson and el
Obeid, 2004). Farming could have started as early as
2,000 years ago and most sites showing such evidence
date from between 1,500 and 1,000 years ago. The Delta
itself has been fairly well protected against farming and
in particular against livestock by the tsetse fl y (Glossina
morsitans) that was abundant until recently. Early hu-
mans in the Delta were the San (“Bushmen”) living as
hunter-gatherers. They have now all left the Delta and are
living in villages in the periphery. In terms of human oc-
cupancy tourist lodges and adjacent satellite villages for
service personnel are now common features.
To k eep the livestock away from the wildlife carrying
diseases like the foot-and-mouth disease, a “buffalo
fence” was constructed in 1982 (Fig. 3) that is fairly ef-
fective in closing off all wildlife movements between the
Delta and areas to the west and south.
Habitat diversity
A mosaic-like vegetation pattern characterizes the Delta
from permanent swamps, over a gradient of seasonally
fl ooded types of swamps and grasslands to riverine wood-
lands and dry savannas that are never under water. The
complex pattern is mainly caused by ever changing river
courses and the growth of islands that in most cases seem
to have started as termite mounds (McCarthy et al., 1998).
There are about 150,000 islands in the Delta (Gumbricht
et al., 2004) each with a typical vegetation zonation.
There are large variations in vegetation patterns over
small distances, although the Delta is very fl at and is
made up of homogeneous sand (Fig. 4). In an intensely
studied area in a central part of the Delta two transects
occasions from about 0.7 million years BP possibly to as
late as 400,000 years BP (Ringrose et al., 2005). The
Kafue River was captured by the Middle Zambezi some-
time in the mid or late Pleistocene, while its upper stretch
with Lake Bangweulu became connected to the Congo
River sometime during this time period.
The most important feature that makes the Delta
unique is the extreme spatial and temporal variations in
the fl ooding pattern, which change over at least four time
scales. The longest is over geological periods and de-
scribed above. Since around 1850 the fl ooding pattern
has moved from a very westerly to a central and now a
very easterly distribution (McCarthy and Ellery, 1998),
with some recent evidence of a return to a westerly trend.
There are dry and wet periods as well with about 8 and 18
years interval (Tyson et al., 2002), which causes the sea-
sonally fl ooded areas to withdraw and expand. Finally
the fl ooding pattern can change from year to year due to
local factors such as vegetation blockages in the streams
causing damming and overfl ow of riverbanks. Controls
on recent changes to Delta ecosystems have been dis-
cussed by a number of authors and summarized in
Ringrose et al. (2003a; 2005).
As mentioned above a mean 96–98% of infl owing
water is lost as evapotranspiration in the Delta and the
rest leaves through outfl ow systems, most commonly
through the Thamalakene and Boteti Rivers (Figs. 1 and
3) which occasionally reach the fi nal sink, the huge Mak-
gadigadi saltpans in the Kalahari. Sometimes, however,
outfl ow is also deviated to the lake depressions, Ngami
and Mababe to southwest and northeast. This happens
only during wetter years while during dry periods there is
no outfl ow from the Delta. Almost all infl owing water
and sediment comes from heavily weathered Archean
and Proterozoic rocks, which are nutrient poor, hence the
Delta sediments and soils are poor in nutrients (Hunts-
man-Mapila et al., 2005). The conductivity in the inlet is
40–50 µS/cm, total phosphorus 0.02–0.03 mg P/L and
total nitrogen 0.36 mg N/L (Cronberg et al., 1996). These
are typical levels for the streams in the Delta, while the
seasonal fl oodplains have 5–10 times higher nutrient lev-
els (Högberg et al., 2002). Very high nutrient levels can
be found in isolated pools used by wallowing and drink-
ing wildlife.
The Okavango Delta occurs within a massive shallow
basin within the middle of southern Africa. The part of the
depression fi lled with sands is called the Kalahari Basin
(Fig. 1). In its southern sub-basin the Okavango River
fl ows into the Delta and eventually into the Makgadikgadi
Pans (Mendelsohn and el Obeid, 2004). Here there is a
vegetation gradient mainly caused by very low rainfall in
the south (200 mm in southern Botswana) to 1,200 mm in
the north in central Angola-Zambia (Ringrose et al., 1999,
2003b). The vegetation belongs to the vast Sudano-Zam-
bezian Region (Werger and Coetzee, 1978) characterized
314 L. Ramberg et al. Okavango biodiversity
sifi cation of habitats was based on a combination of life
form characters and dominant species. In total 46 habi-
tats were identifi ed. Landscape variety was then meas-
ured as suggested by the Corina Land use programme by
recording the number of different habitats in 3 × 3 km
squares (European Union, 2000).
The proportion of the Okavango Delta occupied by
the dominant class “channels and recently inundated
fl oodplains” is a fairly low 8.9 per cent (Table 1). As the
dominant class occupies a small area in total, the other
classes are likely to be abundant, but will occupy even
smaller areas.
In the Delta study area the average polygon (i.e. spe-
cifi c habitat fragment) is fairly small being 0.05 km2. The
number of polygons per unit area, known as Monmo-
Figure 4. A. Detailed vegetation map of the HOORC research area in the central Okavango Delta with two inserted transects. B. Distribution
of vegetation habitats along these two transects relative to elevation (Modifi ed from Meyer, 1999).
4,000 m and 2,700 m long, respectively, had a maximum
variation in elevation of two meters, but crossed vegeta-
tion types from permanent open water communities –
sedge lands – grasslands – riverine woodlands – dry
woodlands, and some of these habitats occurred several
times along the same transect (Fig. 4). These small differ-
ences in elevation are making large differences in the fre-
quency and duration of fl ooding, which causes the large
variations in vegetation in a seemingly fl at and homoge-
neous environment.
A vegetation/habitat map over the whole Ngamiland,
110,000 km2, including the Okavango Delta of 28,000 km2
was produced by HOORC 2001–02, based on Landsat
images, air photos and a large number of ground-truth
transects (502 transects of 90 or 100 m length). The clas-
Aquat. Sci. Vol. 68, 2006 Overview Article 315
nier’s (1982) fragmentation index, is on average 180 in a
9 km2 quadrate. The number of different habitat types in
these 9 km2 areas varies between 1 and a maximum of 31.
The distribution is skewed with clustering of cells in the
3–11 range and a heaping of frequencies occurs at values
5 and 6. Thus, the majority of 3 × 3 km cells contain 5 or
6 different vegetation classes. However, as there are as a
mean 180 polygons in each of these cells it means that
each habitat type is repeated 30–36 times in each cell. In
the EU study (2000) the number of habitats per cell var-
ied between a low of 2.4 in Austria to a high of 4.4 in
Luxembourg. The average for the EU is 3.6, which again
indicates that the habitat diversity as well as the habitat
fragmentation is high in the Okavango Delta.
Figure 5 shows the resulting distribution of cells with
varying degrees of habitat variety. The high variety zones
along the edges of the Okavango Panhandle and Delta
stand out and contrast with the lower degree of variety
within the interior Delta and the dry lands further from
the Delta. Of particular interest are the cells with excep-
tionally high vegetation variability (over 15 classes per
cell). These are mostly located along the perimeter of the
wet Delta, along the Panhandle, and along the major fl ow
channels to the east and west (Fig. 3). These are areas,
which tend to receive fl ooding at intermittent periods, i.e.
they are not permanently fl ooded nor permanently un-
fl ooded areas. These areas with the least stable and pre-
dictable environments have thus the highest habitat diver-
sity.
Biodiversity of different plant and animal
groups
Algae
As far as we know no taxonomic study of Delta algae has
been published. Cronberg et al. (1996) sampled rivers,
fl oodplains and isolated pools during four occasions
1991–92 and used quantitative methods for the determi-
nation of the dominating species’ biomass. In general riv-
ers had very low biomass, below 1 mg/L fresh weight.
Floodplains had a large variation in biomass with gener-
ally higher biomass than the rivers. In both habitats the
Chlorophyceae contributed most species (total 23) fol-
lowed by Cyanophyceae (9) and Bacillariophyceae (6).
The total number of species was about 50. The isolated
pools could have very high biomasses, particularly if they
had been fertilized by dung from hippopotamus or drink-
ing animals. Typical species here were the euglenophytes.
In lagoons and pools the plankton succession is often in-
terrupted by turbation from drinking and wallowing ani-
mals (Ramberg unpublished). Most species here are small
Chlorophyceae and diatoms or good swimmers like
Cryptomonas and Peridinium. In the very shallow waters
in the vegetation there is a typical fl ora of colonial swim-
ming forms. During the cold season Synura and Uroglena
species (Chrysophyceae) are predominant while Eudori-
na-Pandorina and Vol vox (Chlorophyceae) are very com-
mon during rest of the year. In these shallow habitats it is
diffi cult to separate planktic and attached algae. In the
latter group there is a high number of desmid species that
thrive in the soft, often humic waters, and there is a good
number of diatoms as well. The total number of algal spe-
cies cannot be estimated but must be in the order of sev-
eral hundred.
Higher Plants
The Zambesian Phytochoria is one of 16 areas in Africa
defi ned as having more than 50% endemic plant species
and more than 1,000 such species in total.
Figure 5. Number of habitats per 3 × 3 km squares in the Okavango
Delta.
Table 1. Ve getation/Habitat characteristics in the Okavango Delta
and surrounding Ngamiland.
Delta Rest of Ngamiland
Area in square km 26,662 81,245
Number of vegetation classes 45 46
Proportion of dominant class 8.9 % 9.0 %
Number of polygons 520,079 831,440
Average polygon size (m2)51,266 97,717
Dominant class in Ngamiland = Low open shrubbed grassland with
sage bush.
Dominant class in Okavango Delta = Channels and recently inun-
dated fl oodplains.
316 L. Ramberg et al. Okavango biodiversity
land settings on islands or sandveld tongues. However,
despite their terrestrial character many of these taxa are
absent in the surrounding savanna habitats as they require
a different air humidity or soil moisture regime or higher
ground water table. Thus they are intimately associated
with the wetland environments of the Okavango Delta
(Ellery and Tacheba, 2003).
A large number of species occur in the permanent
swamps (about 220 taxa), and many are connected to the
fl ooded grasslands (about 90 taxa) or to the combination
of fl ooded grasslands and dryland settings (80 taxa). A
small number of species are parasitic (18) or insectivo-
rous (12).
It is diffi cult to extract the number of aquatic and
semi-aquatic (palustric) species from the species pool
mentioned above. The reason lies in the ambiguous defi -
nitions of hydrophytes (aquatic macrophytes) and wet-
land plant species as discussed by Junk in this volume. If
one follows the defi nition of Reed (1988) for hydro-
phytes, which “ ... demonstrate the ability to achieve ma-
turity and reproduction in an environment where all or
portions of the soil within the root zone become, periodi-
cally or continuously, saturated or inundated during the
growing season”, the majority of species of the frequent-
ly inundated fl oodplains of the Okavango Delta belong to
this category. Smith (SMEC, 1989) with his extensive ex-
perience of the Delta fl ora classifi ed it as outlined below
(Table 2).
Of the 147 plants classifi ed as aquatic and semi-
aquatic only 10 are woody of which only three are trees.
The palm Phoenix reclinata and Syzygium guineense are
fully grown trees that occur in patches fringing islands or
termitaria, while the shrubby Ficus verruculosa lines the
lower reaches of river channels in the perennial swamp
(op. cit.).
The “conventional” aquatic plants such as submersed
species or species with fl oating leaves develop such large
morphological plasticity in the fl oodplains that the affi li-
ation to a certain class is dependent on the environmental
conditions at the time of the investigation. For example
Nymphoides indica, a hydrophyte with fl oating leaves in
permanent waters or during fl ood conditions, is able to
develop a terrestrial form with compact leave rosettes as
soon as the fl oodplain dries up and thus resemble terres-
trial geophytes. There are, therefore, unavoidable over-
laps in both life form classifi cation and in habitat group-
ing, which refl ects the fact that large variations in fl ooding
(see above) cause the selection of plants with large eco-
logical, physiological and morphological plasticity. From
Smith’s data (SMEC, 1989) it can be estimated that 35%
of all species occur in more than one habitat along the dry
– wet gradient (Table 3).
Species/area relationships. Ellery and Tacheba (2003)
list 1,259 species for the Okavango Delta for an area of
Species number. We owe the best fl oristic information on
the Okavango Delta to the late Peter A. Smith (Ö1999)
who already began the study of the Okavango fl ora in the
early sixties. Although his unexpected death impeded the
publishing of his enormous knowledge, some fragmental
work remained and it forms the foundation for the current
fl oristic research.
Smith compiled the fi rst plant species list (vascular
plants) for the Okavango Delta within the scope of the
Ecological Zoning of Okavango Delta (SMEC, 1989) us-
ing the analysis of records of authenticated plant speci-
mens from herbaria containing a major collection of Bot-
swanan material (Gaborone, Harare, Kew-London,
Pretoria) supplemented by references in the botanical lit-
erature and his own records. This list contains 1,061 dif-
ferent species (and with lower taxa 1,078). However,
Smith guessed that the fi nal total species number might
eventually approach about 1,200 at least, since many re-
moter parts of the Delta had still to be explored botani-
cally.
Between 2000 and 2002 several botanical studies
were conducted which lead to an extended list of 1,299
species and subspecies (Ellery and Tacheba, 2003). Based
on the mark-recapture method these authors estimated
that the total number of species and lower rank taxa in the
Okavango Delta is 1,405. However, most botanical stud-
ies have been done in the central 2/3 of the total area,
while the in-accessible eastern parts in particular have
hardly been studied. The total number of species is there-
fore likely to be considerably higher than that estimated
by Ellery and Tacheba (2003).
Species diversity. The best overview of the overall fl oris-
tic diversity is provided by Ellery and Tacheba, (2003),
and the following paragraphs refl ect their calculations.
As stated before, the currently known fl ora of the Oka-
vango Delta comprises about 1,300 taxa on the species
and lower levels, of which 1,260 taxa are on the species
level. They belong to 530 genera and 134 families. The
most diverse families are the grasses Poaceae, sedges Cy-
peraceae, followed by the Asteraceae and Fabaceae, each
of which have more than 20 genera and 50 taxa of species
and lower ranks. Most genera (73%) are represented by
one or two species only, whereas a small number of gen-
era (7%) are represented by 10 or more taxa of species
and lower ranks.
The life-form spectrum is clearly dominated by her-
baceous plant species (hemicryptophytes 55.5%, crypto-
phytes 4.4%, therophytes 7.6% and aquatic plants 8.1%).
Woody plants make up 18.1% of the fl ora, split approxi-
mately evenly between shrubs and trees, the chamae-
phytes contributing 6.3%. The proportion of the hydro-
phytes seems to be underestimated (see below).
Of the total number of taxa present in the Okavango
Delta, a signifi cant proportion of about 60% occur in dry-
Aquat. Sci. Vol. 68, 2006 Overview Article 317
25,000 km2. Since the ratio species (S)/area (A) is not
linear, but typically follows a function: c = S : A0.18
(Rosenzweig, 1995), comparisons of species richness be-
tween areas of different size must take this into account.
In the formula above c is the number of species expected
to be found in an area of one km2. This value calculated
for a number of biomes in Southern Africa (Table 4)
shows that the Okavango Delta has a species density of
210 species per km2 similar to the dry Karoo biomes in
South Africa. This is corroborated further by inserting the
Okavango data in a log/log diagram with 25 biomes from
South Africa outside of the extremely species rich Cape
Floral Kingdom (Rosenzweig, 1995). The Okavango data
falls almost exactly on the regression line refl ecting simi-
larity to the dryer and colder southern and western bi-
omes, while the species densities are more than twice as
high for the better watered and warmer grasslands and
savannas in the eastern and northern parts of the sub-con-
tinent (Table 4).
Plant communities, their dynamics and species richness.
Several vegetation studies have been done in the Delta on
the plant community level within the last two decades
(SMEC, 1989; Ellery et al., 1993; Ellery et al., 2000;
Bonyongo et al., 2000; Ellery and Tacheba, 2003; Her-
mann, 2003; Sliva et al., 2004), which have resulted in
various suggestions concerning vegetation classifi cation.
Most of these have been restricted to limited objectives,
but regardless if the methods employed were quantitative
and statistical or based on more qualitative assessments,
the over riding result is always the same. The major fac-
tor organizing plant communities in the Delta is the hy-
drology and more specifi cally the duration and depth of
fl ooding.
A classifi cation of plant communities was done based
on quantitative data covering a fair amount of Delta habi-
tats (Sliva et al., 2004), producing a dendrogram using
TWINSPAN (Fig. 6). This clearly shows that on each
level of division the hydrological conditions are decisive.
At the fi rst level of division the samples were split into
group 2, which represents the wet (inundated or frequent-
ly fl ooded) wing of the community spectrum, whereas
group 3 represents seasonal fl oodplain and island com-
munities. In the next step, water depth and fl ood duration
are responsible for the division: group 4 is characterised
Table 2. Analysis of Okavango Delta plant species by lifeform (After Smith in SMEC, 1989).
Total Dicots Monocots Ferns
AQUATIC & SEMI-AQUATIC PLANTS
Emergent grasses & sedges 61 61
Other emergent herbs 76 55 19 2
Trees & shrubs 10 9 1
Subtotal 147 64 81 2
Submerged sedges 2 2
Submerged other herbs 19 5 14
Subtotal 21 5 16
Emergent and submerged herbs and creepers 11 10 1
Free fl oating on surface or submerged 16 7 7 2
Floating leaved 13 9 4
Subtotal 40 26 11 3
TOTAL AQUATIC PLANTS 208 95 108 5
NON-AQUATIC PLANTS
Forbs 383 328 50 5
Grasses 168 168
Sedges 60 60
Creepers 78 77 1
Subtotal herbs 689 405 279 5
Woody shrubs & shrublets 85 84 1
Shrubs or trees 28 28
Climbers 8 8
Trees 60 59 1
Subtotal woody plants 181 179 2
TOTAL NON-AQUATIC PLANTS 870 584 281 5
GRAND TOTAL 1,078 679 389 10
Table 3. Number of plant species observed in each habitat (after
Smith in SMEC, 1989, simplifi ed). Note that some species occur in
more than one habitat.
Habitats Number of observed species
Perennial swamp 205
Seasonal swamp 240
Flooded grasslands 213
Drylands 686
Miscellaneous 84
Tota l 1,428
318 L. Ramberg et al. Okavango biodiversity
by the indicators Schoenoplectus corymbosus and Mis-
canthus junceus (frequently inundated fl oodplains),
whereas group 5 is indicated by the presence of Vossia
cuspidata, Ceratophyllum demersum and Cyperus papy-
rus (open water and permanent swamps). The broad spec-
trum of seasonal swamp and island communities is con-
tained in group 6, which is specifi ed further on the next
level of division, while group 7 contains a special pan
vegetation. After the fourth level of division nine mean-
ingful ecological vegetation groups were identifi ed: (1)
Vegetation of open water, (2) Cyperus-dominated channel
fringe and backswamp communities, (3) Phragmites-
dominated channel fringe and backswamp communities,
(4) Miscanthus-Ficus permanently fl ooded backswamp
communities, (5) Schoenoplectus corym bosus-Cyperus
articulatus communities of shallow backswamps and fre-
quently inundated fl oodplains, (6) Communities of sea-
sonal fl oodplains, (7) Island fringe communities, (8) Is-
land interior grassland communities, (9a) Pan communities
– upper level, (9b) Pan communities – bottom level.
In the DCA ordination graph (Fig. 7), these nine ma-
jor communities identifi ed by the TWINSPAN arrange
neatly along axis 1 which can be explained by the depth
and duration of fl ooding. Axis 2 is also predominantely
related to hydrology, but here the main gradient is up-
stream-downstream with low annual water level varia-
tions in the upper parts of the Delta (0.5 m) and much
higher variations in the lower parts (2 m). The third major
environmental factor is hydrological as well; the differ-
ence between lentic and lotic habitats.
Recognising the different seasons, scale and focus of
all the different vegetation studies and considering our
own latest data (unpubl.), about 26 meaningful ecologi-
cal plant communities can be preliminarily classifi ed in
the vegetation of permanent swamps, fl oodplains and is-
lands. The drylands remain unconsidered, which are the
never fl ooded vegetation of large sandveld tongues and
large islands representing the Acacia and Mopane wood-
lands and shrubland, as well as the non-inundated grass-
land types (Table 5).
Communities associated with permanent water (No.
1–14) are relatively species poor, harbouring about 50 to
70 species per community group. Cyperus papyrus as
well as Phragmites species tend to develop large and
dense monospecifi c stands, supported by the relatively
higher nutrient loadings in the upper reaches of the Delta,
making the establishment of other less competitive spe-
cies diffi cult. The species diversity increases on the open
boundaries, thus along the channel and lake fringes. In
open water areas (ledibas, oxbow lakes) the diversity rais-
Table 4. Number of plant species per 1 km2 area in the Okavango
Delta and other Southern African biomes. Data compiled from
Smith in SMEC (1989) and Ellery and Tacheba (2003).
Biome No. species Area (km2) Species
per 1 km2 (4)
Savanna 5,788 632,034 523
Desert 497 41,292 73
Grassland 3,788 111,888 467
Fynbos (1) 7,316 36,628 1,104
Nama-Karoo (2) 2,147 198,468 239
Succulent Karoo (3) 2,125 50,516 302
Okavango (Ellery) 1,259 25,000 210
(1) The maccia like vegetation which is dominant in the Floral King-
dom Capensis and covers the southern most part of Africa.
(2) Karoo: The semi-desert Floral Domain located north of Capensis
and south and west of the Zambezian Floral Domain and the
Kalahari. It stretches along the western part of southern Africa
up to Angola and is divided into a number of sub-domains of
which Nama Karoo Sub-domain is bordering the Kalahari.
(3) The Succulent Karoo sub-domain is located in the mountainous
area of western Cape Province and is characterized by many suc-
culent plant species (see Werger and Coetzee, 1978).
(4) The species per 1 km2 (c) is a measurement of species density
based on the formula: c = Species number : Area 0.18 (km2)
(Rosenzwieg 1995). This is the expected number of species in
one square kilometre.
1
2
3
4
Group 8Group 11
Group 22 Group 23
9a 9b8
community group
5671234
Group 24 Group 25 Group 26 Group 27
Group 12 Group 13Group 9Group 10
Group 5 Group 6 Group 7Group 4
Group 3
Group 1
Group 2
Figure 6. Dendrogram showing the TWINSPAN hierarchical division of samples into plant communities. See text for the description of the
community groups.
Aquat. Sci. Vol. 68, 2006 Overview Article 319
es with the shallowness of the water. The species compo-
sition and abundance in these communities remains simi-
lar through the whole year independent of the fl ood pulse
(Czekanowski index of similarity 0.70–0.95).
Compared with permanent aquatic communities, the
number of plant species rises up to twofold on seasonal
fl oodplains (No. 15–20). On regularly inundated fl ood-
plains the water fl uctuation causes periodical changes
between terrestrial and aquatic phases of the sites. The
aquatic-terrestrial-transition-zone, “ATTZ” (Junk, 2003)
is a dynamic system of steadily changing water and nutri-
ent status, of establishment and dying off of species. This
dynamic littoral zone provides good living conditions for
both terrestrial and semi-terrestrial short-lived plant spe-
cies during the low water period as well as for aquatic
species during the inundation, as long as these species are
able to survive the unfavourable period or to colonise and
occupy the new habitats rapidly enough. The availability
of various temporary habitats which are densely packed
within relatively small areas is responsible for the high
species diversity. If one compares the low and high water
season, the alteration of species and their abundances
within the fl oodplain communities is also expressed by
the signifi cantly lower similarity indices (Czekanowski)
of 0.25–0.50.
However, the highest species diversity is exhibited in
the riparian woodlands along the island margins (No.
21–22). During the fi eld campaign in February 2003 be-
tween 20 and 83 species were recorded per 70 m2 plot,
and altogether more than 150 species (e.g. more than one
eighth of the whole Okavango fl ora) were identifi ed with-
in only fi ve plots (Sliva et al., 2004). Island margins pro-
vide optimal habitat for a large number of woody species
(shrubs and trees), which increase the species diversity
considerably. After exclusion of woody species the
(Czekanowski) similarity index of vegetation recorded
during dry and rainy season is only 0.22–0.25, which re-
fl ects the distinctive seasonal variation.
The origin and the unique ecological functions of is-
lands and associated woodlands has been subject of sev-
eral in-depth studies (McCarthy et al., 1991; McCarthy et
al., 1993; Ellery et al., 1993; Ellery and Tacheba, 2003).
In this environment the classifi cation of the riparian
plants as dry land species is ambiguous as the riparian
zones are fed by shallow horizontal groundwater infi ltra-
tion from adjacent rivers and fl oodplains (Ramberg et al.,
2006). The majority of woody species (trees, shrubs and
lianas) which occur within these riparian bands in the is-
land fringes of the Okavango Delta are probably able to
tap this groundwater resource (Ringrose, 2003). Even
though these are not strictly wetland habitats, they are
central for the present structure and functioning of the
whole ecosystem (Ellery and Ellery, 1996; Ellery and
Tacheba, 2003), and the fact that we fi nd the highest spe-
cies diversity in these island fringe communities under-
pins their high ecological value.
Next to the hydrological factor complex, the salinity
of the island soils infl uences the diversity of species with-
in small areas. There is a gradient of increasing solute
concentration in the ground water from the edge of the
islands towards the centre (McCarthy et al., 1991; Ellery
et al.; 1993) which is refl ected by a typical zonation of
Figure 7. DCA ordination graph of the 116 vegetation samples (Sliva et al., 2004). For names of communities see Table 5.
320 L. Ramberg et al. Okavango biodiversity
vegetation and the establishment of characteristic com-
munities (No. 23–24). On sites with high solute concen-
trations species diversity declines considerably since
only few species are adapted to those harsh living condi-
tions. Although the communities of saline soils seem to
be the species poorest among all other communities in
the Okavango Delta (with only about 20 species), they
contribute to the overall species diversity because of the
occurrence of the specialised halophytes.
Small ephemeral water bodies (pans) occur during
the rainy season in drier habitats and carry the next spe-
cifi c plant communities (No. 25–26) with a distinct zona-
tion according to the water depth and duration.
It is obvious that the main reason for the high plant
species diversity of the Okavango Delta as well as for the
exceptionality of this ecosystem from a nature conserva-
tion point of view, lies in the interaction of periodical
natural phenomena – the annual fl ood in the dry season
and the distinct rainy season in time of low water level –
with the shifts in fl ooding pattern over short and long pe-
riods. Succession processes at different phases of devel-
opment are therefore ongoing in all plant communities in
the Delta. These processes are the main driving forces for
the species and habitat diversities in the Okavango Delta
and must be conserved in order to maintain the unique-
ness of this system.
Invertebrates
The data on invertebrate species in the Okavango Delta is
far from comprehensive and many taxonomic groups are
too diffi cult to collect, or nobody has tried to sample
them, while some are taxonomically not well known or
there are no taxonomists able to identify them. The pic-
ture will, therefore, be patchy and will probably remain
so for a long time.
Table 5. Overview of main vegetation groups in the Okavango Del-
ta (except dryland habitats). Twinspan: The numbers correspond
with the Twinspan classifi cation (Fig. 6).
Open water communities: Twi nspan
(Fig. 6)
Communities of lakes and standing backwater
1 Nymphaea nouchalii communities 1
2Eleocharis dulcis communities 1
3Ceratophyllum/Lagarosiphon/Ottelia
communities 1
4Trapa natans communities 1,2
5Vossia/Echinochloa pyramidalis communities 1,2
Communities of fl oating waters (channel beds)
6Nesaea/Potamogeton communities 1,5
Channel/lake fringe and backswamp communities:
7Trapa natans communities 2,1
8Vossia/Echinochloa pyramidalis communities 2,1
9Scirpus cubensis/Pycreus mundii communities 2
10 Fimbrystilis dichotoma/Pycreus fl avescens
communities 2
11 Cyperus papyrus fringe and backswamp
communities 2
12 Phragmites australis/P. mauritianus fringe
communities 3
13 Miscanthus/Ficus verruculosa backswamp
communities 4
14 Ficus verruculosa/Syzigium cordatum fringe
communities –
Communities of frequently fl ooded seasonal
fl oodplains:
15 Schoenoplectus corymbosus/Cyperus
articulatus communities 5
Communities of the Aquatic-Terrestrial Transition
Zone ATTZ (sensu Junk)
16 Panicum repens grassland communities 6
17 Cynodon dactylon/Sida cordifolia
communities 6
18 Imperata cylindrica/Setaria sphacelata
grassland communities 6
Communities of rarely fl ooded fl oodplains
19 Urochloa mossambicense/Pechuel-loeschea
leubnitzae communities
–
20 Acacia/Colophospermum shrubland
communities –
Island communities
Riparian woodland communities
21 Phoenix reclinata/Ficus sycomorus woodland
communities
7
22 Hyphaene/Diospyros mespiliformis woodland
communities
7
Island interior communities
23 Eragrostis sp./Acacia sp. grassland
communities 8
24 Sporobolus spicatus grassland communities 8
25 Urochloa trichopus/Glinus bainesii/Litogyne
pan communities 9a
26 Eragrostis pilosa/Lemna pan communities 9b
Table 6. Overview of Odonata families, number of genera and spe-
cies found in the Okavango Delta.
Genera Species
ZYGOPTERA
Calopterygidae 1 1
Lestidae 1 5
Coenagrionidae 6 25
Platycnemididae 1 1
Protoneuridae 1 1
Subtotal 10 33
ANISOPTERA
Aeshnidae 1 4
Gomphidae 4 7
Cordulidae 2 5
Libellulidae 22 45
Subtotal 29 61
Grand total 39 94
Aquat. Sci. Vol. 68, 2006 Overview Article 321
Dragonfl ies (Odonata). Comparatively good information
exists for Odonata (Pinhey, 1967; 1976). Pinhey collected
samples in the Okavango area during a number of expe-
ditions until 1976 and also obtained access to other
smaller collections. During 2000–2002 a new study was
conducted (Kipping, 2003) which was designed to cover
the same areas and habitats as those of Pinhey. A total of
94 species were found, 33 Damselfl ies (Zygoptera) and
61 Dragonfl ies (Anisoptera) in the Okavango Delta (Ta-
ble 6) out of 114 species in all of Botswana.
Comparisons with species inventories from the coun-
tries surrounding Botswana (Zimbabwe, Namibia, Zam-
bia, RSA, Mocambique) resulted in a total number of 295
species for this area. Cluster analysis revealed that there
are strong similarities in Odonata composition in particu-
lar with Zambia and Namibia (the Caprivi strip); in other
words with the wetlands to the north, which at wetter
times have been directly connected to the Okavango.
Many of the Odonata in the Delta have a Central African
distribution and reach their southern most distribution
here (Kipping, 2003).
Pinhey (op. cit.) recorded a total of 92 species in the
Delta. Twenty-fi ve years later Kipping (op. cit.) found
one species new for the area and one new for science.
On the other hand he could only fi nd 70 out of the 92
species found by Pinhey, although his sampling inten-
sity compares well with that of Pinhey both in time and
space. Nine of the “missing” species are Zygopterans
and seven of them had been recorded from three or more
localities in the Delta. Out of the 13 Anisopterans that
could not be found again, fi ve had been found in three or
more localities. For the species found in only 1–2 lo-
calities the problems of sampling rare species arise.
However, when there are indications that species which
were fairly wide spread up to the mid-seventies are now
absent, there are reasons to look for other explanations.
There has been a gradual decline in fl ooding of the Del-
ta since the mid-seventies which could have resulted in
a loss of suitable aquatic habitats for the larvae or in a
loss of suitable fl ying prey for hunting adults. Another
factor is the aerial spraying against tsetse fl ies in the
Delta which took place during the eighties and then
again 2001–02. During the fi rst period fairly potent in-
secticides such as dieldrin were used but over smaller
areas in each year. In the recent spraying, however, the
entire Delta south of the Panhandle was sprayed; the
northern part in 2001 and the southern part in 2002; to-
taling to about 17,000 km2. Deltamethrin was used
which has some good properties such as its short half-
life in nature and its specifi city for invertebrates. Adult
Odonata experienced high mortality during the spraying
of deltamethrine and the same results were recorded for
larvae of the families living on the sediment surface or
on vegetation (Ramberg, 2004).
Butterfl ies (Lepidoptera). A preliminary checklist of but-
terfl ies of Botswana, including the Okavango Delta, was
published by Pinhey (1968; 1971; 1974; 1976). This
checklist is based on his own collecting expeditions, sup-
plemented by records from museums in Southern Africa.
Pinhey’s coverage of the Okavango Delta is limited
mainly to the southern and western Okavango from Maun
to Mohembo, due to the inaccessibility of the Delta at the
time. Nevertheless, 115 species were recorded from this
area. Pinhey’s work forms the baseline for butterfl ies in
the Okavango. Very little other information has been pub-
lished since 1976.
The Nymphalidae and Lycaenidae are the most di-
verse families in the Okavango Delta (Table 7), despite
the abundance and conspicuousness of the Pierids (Pin-
hey, op. cit.). The vast majority of butterfl ies encountered
in the Delta belong to this last family, and are restricted
mainly to two very abundant migratory species – Bele-
nois aurota (Brown-veined White) and Catopsilia fl orella
(African Migrant). Large numbers of these two butterfl y
species migrate in a north-easterly direction throughout
the region, including the Okavango, in mid-summer. The
Lycaenids by contrast are small, inconspicuous species,
which nevertheless contribute to over 30% of the ob-
served diversity.
Analysis of Pinhey’s Checklist of Butterfl ies of Bot-
swana shows that the Okavango Delta is a focus of but-
terfl y diversity in Botswana. This is not unexpected since
this is a wetland area surrounded by arid Kalahari semi-
desert – a wide variety of habitats exist with a wide range
of larval food plants, and angiosperms which provide
nectar for the adult butterfl ies. Some species such as
Danaus chrysippus and Vanessa cardui are cosmopoli-
tan, while others are characteristic of wetlands such as:
Hyalites rahira, Precis ceryne, Myrina silenus, Borbo
micans, Parnara monasi and Gegenes hottentota.
In the Okavango Delta, the butterfl ies most at risk are
the myrmecophilous (ant-associated) Lycaenids. As lar-
vae these species require both the host ant and the host
plant, as well as optimal climatic conditions to thrive.
Due to the paucity of information on the Delta’s butter-
fl ies, no Lycaenids have yet been identifi ed as threatened,
although it is possible that some species are at risk.
Table 7. Overview of butterfl y families and species in the Okavango
Delta. The taxonomy follows Pringle et al. (1994).
Family Nr. of species
Nymphalidae 36
Lycaenidae 39
Pieridae 23
Papilionidae 3
Hesperiidae 23
Tota l 124
322 L. Ramberg et al. Okavango biodiversity
Effects of aerial spraying against tsetse fl ies. The aerial
spraying of the entire Delta (except the Panhandle)
against tsetse fl ies 2001 and 2002 using deltamethrin
(0.26–0.30 g/ha) was repeated fi ve times during the cold
season May–August. A large number of samples were
collected for the environmental assessment of both aquat-
ic and terrestrial environments (Perkins and Ramberg,
2004a). This study has, however, several weaknesses
which makes it less useful in the present context. No
benchmark studies were conducted before the spraying
began and taxonomic identifi cation was only taken to the
family or genus level. For some taxonomic groups (usu-
ally the most common) the identifi cation went a step fur-
ther and species were identifi ed as such but not given a
species name only an identifi cation number, so called
“morphospecies”.
After spraying, aquatic invertebrate families showed
a 25–46% reduced total abundance (Palmer, 2004). Be-
fore spraying there were statistically signifi cant differ-
ences in species compositions between lagoons and
streams, but due to the disappearance of several families
after the spraying, these differences became less appar-
ent and a species poorer, less diverse composition re-
mained. Out of a total 65 taxa 23 were common, and of
these, six taxa with several species in each, declined
drastically during the spraying campaign and had disap-
peared by the fi fth spraying cycle. It is likely that at least
the same proportion of the less common taxa was elimi-
nated as well.
Terrestrial invertebrates were predominantly sampled
from under tree canopies (as knockdown) before, during
and after the fi ve spraying cycles (Dangerfi eld, 2004).
Abundances declined by up to 68%. The most affected
group was beetles. The composition of species changed
through the cycles. Around 30% of the species were only
collected before the spraying or in the fi rst spraying cy-
cle, whilst a lower proportion appeared in later cycles for
the fi rst time.
The recovery of the invertebrates was studied during
2003 at the same sites as in the previous year. Some of
the aquatic families affected by spraying remained at re-
duced levels, notably shrimps and small backswimmers.
Many of the affected families returned to pre-spraying
abundances and the composition of aquatic invertebrates
in the sampled habitats returned approximately to their
pre-spraying patterns (Palmer, 2004).
There are no documented ENDANGERED or VUL-
NERABLE butterfl y species in the Okavango (Henning
and Henning, 1989). The following species are RARE:
Anthene minima, Colotis doubledayi angolanus, Pseu-
donympha swanepoeli and Tuxentius malaena. Borbo
micans and Gegenes hottentota are placed in the INDE-
TERMINATE category.
Mollusca. The occurrence of aquatic snails has been fair-
ly well documented in several studies compiled by Mur-
ray (1997). These studies are combined in Table 8. Most
of the aquatic snails found in the Okavango Delta are
widespread in the Afrotropical region. The most south-
erly localities known in Africa for populations of Pila
and Gabbiella (op. cit.) are found here and only one spe-
cies out of 16 may have some affi nity with temperate cli-
mates. There are no endemic snail species and many of
them occupy seasonal waters scattered over huge areas of
the African savanna. Despite the great distance, (nearly
3,000 km), between the Okavango and the Sudd in the
Nile River, 9 species of their total fauna of 34 species live
in both areas (op. cit.).
Zooplankton. In total 37 microcrustaceans have been re-
corded in the Delta by Lindholm (2006), Hart (1997),
and Hart et al. (2003). There are 16 species of copepods
within six genera, with Microcyclops and Tropodiap-
tomus being dominant and 45 cladoceran species (Lind-
holm, 2006). Most are minute, refl ecting the strong pre-
dation pressure from visually feeding fi sh fry. Moina
micrura, Daphnia laevis and Simocephalus vetulus are
dominant and widespread species, especially on many
fl oodplains (Högberg et al., 2002).
Three different zooplankton habitats may be distin-
guished in the Delta: permanent lakes, seasonal fl oodplains
and isolated temporary rain pools. Among these, seasonal
fl oodplains offer the most diverse zooplankton fauna. Dur-
ing high water periods, the production of ostracods, copep-
ods and cladocerans can be extremely high temporarily.
Nearly 90 g DW L–1 zooplankton biomass has been record-
ed, making zooplankton on seasonal fl oodplains a crucial
link in the aquatic food web (Högberg et al., 2002).
Fish parasites. Jo Van As and his research team have
identifi ed no less than 200 fi sh parasite species in long
term studies, many new to science. Most of these species
have complex life cycles involving one, two or three spe-
cifi c hosts such as trematode – snail – fi sh – bird. Such
relationships can probably only evolve over very long
evolutionary time scales and may refl ect the old age of
the Delta ecosystems as a functioning whole, even if the
evolutionary processes have taken place elsewhere. Since
this fi eld is under-researched in tropical and subtropical
areas, no meaningful comparisons can be made (Van As,
pers. com.).
Table 8. Overview of number of aquatic Mollusca species found in
the Okavango Delta.
Genera Species
Gastropoda (snails) 13 16
Bivalvia (mussels) 4 6
Tota l 17 22
Aquat. Sci. Vol. 68, 2006 Overview Article 323
Abundance of terrestrial invertebrates returned to pre-
spraying levels within one year. However, about 30% of
the species found before or in the beginning of the spray-
ing in 2002 could not be found again in 2003. On the
other hand about the same number of new species oc-
curred for the fi rst time in this later year (Dangerfi eld,
2004). It is not known how much of this refl ects a natural
pattern and how much is a result of rare species taking
over niches that became empty due to extermination; in
particular as most of these changes occurred among the
rare species (Ramberg, 2004).
Fish
The Okavango system has been connected to the Upper
Zambezi drainage basin and its fi sh fauna can be consid-
ered as being part of the Zambezi system, which has
some 134 species of fi sh (Skelton, 2001). Of these 86 are
found in the Okavango basin and 71 within the Okavango
River and Delta below the Popa Rapids of the East Caprivi
Strip in Namibia (Table 9). More than 50% of the Zam-
bezi species also occur in the Congo Basin and Lake Ma-
lawi. There are also some similarities with the fi sh fauna
of the Limpopo and Phongolo River Systems to the south-
east. This is probably the result of river capture in the not
so distant past. The Zambezi fauna includes 23 (17%)
endemic species, most of which are restricted to the up-
per Zambezi (Skelton, 2001). There are, however, no en-
demic species restricted to the Okavango River and Delta
below the Popa Rapids. So far no alien introductions or
translocated fi sh have been found in the Okavango River
and Delta.
The most important factors infl uencing the distribu-
tion of fi shes in the Okavango system are the permanence
and the fl ow rate of the water. Specialist rheophilics and
species adapted to rocky habitats such as the slender
stonebasher Hippopotamyrus ansorgii (Mormyridae), the
river sardine Mesobola brevianalis (Cyprinidae), the
mountain catfi sh Amphilius uranoscopus (Amphiliidae),
the broadhead catfi sh Clariallabes platyprosopos (Clari-
idae) and the Okavango rock catlet Chiloglanis fasciatus
(Mochokidae) are confi ned to the Angolan headwaters of
the Okavango River and riverine fl oodplains and are not
found in the Delta below the Gomare fault.
Fish stock assessment. Compared with a large number of
African lakes and rivers the number of fi shermen per
square km is extremely low in the Okavango Delta (Mo-
sepele, 2000). Locally household fi shing may be impor-
tant, but large areas of the swamp are not accessible by
boat due to vegetation blockages and the risk of Hippo-
potamus attacks. The total yield of the fi shery is low, and
the catch per unit effort (CPUE) very low (0.4 kg/Lund-
gren gillnet set) compared with ten other aquatic systems
in Africa which had a range of 1.4–4.2 kg/set (op. cit.).
This does indicate an overall low fi sh biomass in the Del-
ta and refl ects its low nutrient status. Locally, however,
the seasonal fl oodplains have a much larger density of
fi sh than the permanent swamp and streams (Högberg et
al., 2002). Occasionally very high densities occur before
and during spawning, as well as in drying-up pools that
have been isolated from the streams. Here fi sheating birds
aggregate in large numbers.
Habitat selection, form and function. Morphological fea-
tures such as body shape, size and shape of fi ns, place-
ment of eyes and size and shape of mouth can be used to
construct an ecomorphological classifi cation of Okavan-
go fi shes. The structural similarities in unrelated species
of fi sh can be correlated with their habitat and niche.
These vary from open water, mid-water, surface and bot-
tom dwelling species, as well as species with special ad-
aptations to inhabit the dense vegetation in papyrus
beds.
Open and midwater fast swimmers. Fishes in this cat-
egory have a fusiform (tapering at both ends) body shape
with forked or lunate caudal fi ns and large, laterally
placed eyes. The dorsal and anal fi n is short and acts as a
rudder for maneuvering at high speed. These fi sh are all
predators mostly in the upper layer of open water of fl ow-
ing rivers usually in roving shoals. The typical example
of this category is the tigerfi sh Hydrocynus vittatus
(Characidae) which occurs in the Panhandle and upper
part of the Delta. Other fi shes in this category include
some of the barbs (Cyprinidae) and other characins. Un-
like the tigerfi sh, these are all small species.
The African pike Hepsetus odoe is the sole represent-
ative of its family, Hepsetidae and occurs widespread in
Africa in the Upper Zambezi, Congo and Niger drainage
Table 9. Fish families, number of genera and species in the Oka-
vango system and Okavango below the Popa Rapids. Compiled
from Skelton et al., 1985; Merron, 1993; Skelton, 2001; 2002 and
own data (Van As) from annual collections 1997–2004.
Family Genera (n) Species (n)
Okavango
System
Species (n)
Okavango below
Popa Rapids
Mormyridae 5 6 6
Kneriidae 2 2 0
Cyprinidae 525 17
Distichodontidae 2 3 3
Characidae 4 4 4
Hepsetidae 1 1 1
Claroteidae 1 1 1
Amphiliidae 2 3 2
Schilbeidae 1 1 1
Clariidae 2 7 6
Mochokidae 2 8 6
Poeciliidae 1 3 3
Cichlidae 718 18
Anabantidae 2 2 2
Mastacembelidae 1 2 1
Tota l 38 8 6 7 1
324 L. Ramberg et al. Okavango biodiversity
basins. Its body plan fi ts comfortably in this category, but
its hunting methods differ from that of the tigerfi sh. It
prefers the quiet waters of channels and lagoons where it
stalks its prey, going for the kill in a swift rush. They do
not co-exist in the same habitats as tigerfi sh, which pre-
fers the fast fl owing currents of the river.
Midwater slow swimmers. The body shape of fi shes
in this category is deep and laterally compressed. The
caudal fi n is truncated and deep, the dorsal fi n is long and
extends from directly behind the head to the base of the
tail stem, the anterior part with spines and the posterior
part with fl exible soft rays. All the fi ns are broad provid-
ing these fi sh with excellent agility and the ability to
hover in the open waters and in the vegetation of lagoons.
This category includes the tilapias of the genera Oreo-
chromis (2 species) and Tilapia (3 species) and some of
the bream of the genus Sargochromis (2 species), all from
the family Cichlidae. They all have relatively small
mouths and feed on plants, insects and detritus, while
some are omnivores. The Zambezi river bream Pharyn-
gochromis acuticeps also falls in this category, although
it prefers sheltered habitats. Other cichlids in this catego-
ry are the banded jewelfi sh Hemichromis elongatus and
the southern mouthbrooder Pseudocrenilabrus philan-
der. The latter species has a very wide distribution in
southern Africa ranging from the Orange-Vaal to the
Congo systems and extending to Lake Malawi. It is, to-
gether with Tilapia ruweti, the smallest of all the cichlids
in the Okavango system.
The agility of movement as a result of the laterally
compressed body plan and broad fi ns paved the way for a
group of species within this category to fi ll a niche as
predators. Their thin profi le facilitates stalking of small
fi sh in the vegetation of lagoons and along the riparian
vegetation. They have exceptionally large mouths, hence
their common name largemouth breams. All the predator
breams belong to the genus Serranochromis of which
there are six species in the Okavango system.
Surface feeders. This category comprises the topmin-
nows or lampeyes of the genus Aplocheilichthys (Poecilii-
dae), with three species and a possible new species present
in Botswana. They occur in small shoals throughout the
Okavango River and Delta and are abundant serving as a
food source for many birds and other fi sh. Lampeyes are
adapted to feed on insect larvae and plankton, very often
on the surface of the water. Their abdomen is rounded but
their dorsal surface is fl at allowing them to feed at the
surface of the water without exposing their bodies to
predators from above. They have exceptionally large eyes
placed near the top of their heads. This is no doubt an
adaptation for locating their prey as well as avoiding
predators. Another adaptation for its surface feeding life-
style is the fact that their mouths face upward.
Bottom dwellers. The bodies of fi shes in this category
are dorso-ventrally compressed and their mouths face
downward. Their eyes are placed on the dorsal side of the
head. These are the only features, which the bottom dwell-
ers have in common since the benthic habitat does not
present a uniform niche. Benthic dwelling fi sh occupy a
variety of trophic levels. Their food sources range from
detritus, benthic algae, and a wide range of adult and lar-
val insects, oligochetes, snails or even other fi shes.
The catfi shes (Clariidae) are represented by six spe-
cies in the Okavango system, of which the broadhead cat-
fi sh Clariallabes platyprosopos is restricted to rocky
habitats in Namibia and Angola. Catfi shes are all well
adapted benthic species and opportunistic omnivores that
will feed on almost any available food source. They have
a dominant ecological presence and at least the sharp-
tooth catfi sh, Clarias gariepinus, can be regarded as a
keystone species.
Other benthic species in the Okavango system include
representatives of the family Mochokidae. The Okavango
rock catlet or mouthsucker, Chiloglanis fasciatus, is
adapted to attach to rocks and plants by its mouth, which
forms a very effective sucker. Six species of squeakers of
the genus Synodontis co-exist in the Okavango system,
each adapted to fi ll a different niche, displaying a text-
book example of adaptive radiation. The spotted squeak-
er, Synodontis nigromaculatus, feeds on detritus and
hosts hundreds of individuals of fi ve different nematodes
and a whole range of ciliophorans in its rectum, probably
symbionts assisting in digestion (Moravec and Van As,
2004).
Although the body plan of the squeakers is ideally
suited for feeding in the benthal, it allows for a peculiar
type of upside down feeding as well. We have personally
observed large numbers of spotted squeakers swimming
upside down on the surface of the river feasting on ter-
mites which appeared in mass after a thunderstorm. This
feeding activity happened in the late afternoon and con-
tinued for a considerable time.
Another very specialized group of benthic dwellers is
the sand catlets of the genus Leptoglanis (Amphiliidae).
These small fi sh of no more than 40 mm bury themselves
in the sand with only their eyes protruding. They are nor-
mally found in submerged sandbanks behind little patch-
es of vegetation facing upstream in the main river and
large channels. Here they prey on small planktonic crea-
tures.
According to Skelton (2001) the taxonomy is in disar-
ray and the spotted sand catlet Leptoglanis rotundiceps
from the Okavango and Zambezi systems may represent
a complex of several different species.
Dense vegetation and rocky habitats. Papyrus beds
comprise a multitude of co-existing plant species repre-
senting a unique habitat for aquatic animals. Fish living
in this submerged jungle require specifi c adaptations in
order to fi nd food and avoid predators in a world of al-
most perpetual darkness. The evolutionary road to sur-
Aquat. Sci. Vol. 68, 2006 Overview Article 325
vival in this habitat may be varied resulting in morpho-
logically different models co-existing amongst the dense
foliage within the papyrus beds.
One very successful inhabitant of dense vegetation is
the spiny eel Aethiomastacembelus frenatus. This small
eel is slender, snake-like, has a very fl exible body, is very
tough and fast. They occur in abundance in papyrus beds,
but are so cryptic that fi shermen spending a lifetime fi sh-
ing the Panhandle have never encountered one. The only
way of fi nding live specimens of these ellusive spiny eels
is to physically lift a section of papyrus onto a boat and
search through the root mass. The same adaptation for
surviving in a papyrus root bed is advantageous when
surviving in crevices and rocky habitats as the second
species the ocellated spiny eel Aethiomastacembelus
vanderwaali does successfully in upper reaches of the
Okavango River.
The endemic African fi sh family, Mormyridae, is rep-
resented by six species belonging to fi ve genera, all su-
perbly adapted for life in dense vegetation in the Okavango
River and Delta: The Zambezi parrotfi sh Cyphomyrus dis-
corhynchus, the bulldog Marcusenius macrolepidotus, the
western bottlenose Mormyrus lacerda, the northern
churchill Petrocephalus catostoma, the southern churchill
P. wesselsi and the dwarf stonebasher Pollimyrus castel-
naui. One species, the slender stonebasher Hippopot-
amyrus ansorgii, is adapted for life in rocky habitats of the
upper Zambezi and not found in the Delta.
These unusual fi shes are laterally compressed with a
large anal fi n. The caudal fi n is forked and the tail stem is
pronounced. They have soft mouths adapted for taking in-
vertebrates from plants. Their adaptive advantage for sur-
vival in dense vegetation is that they discharge a weak elec-
trical current from an organ situated in the tail. Each species
has its own distinct signature discharge and this serves as a
means of communication, navigation and prey detection.
Strategies to cope with low oxygen levels. The oxygen
levels throughout the River and Delta below the Popa
Rapids are generally low. Even in the fast fl owing main
stream saturation levels are between 50–70%. In seasonal
fl oodplains and back waters oxygen levels rarely exceed
3 mg/L, and at night often drop below 1 mg/L which is
close to lethal for many fi sh species.
Despite the low oxygen level, the fl oodplains have
other advantages making it a suitable habitat for many
fi sh species. It has a high productivity and provides a
relatively sheltered environment against predators, there-
fore many fi sh species occur there and have adapted to
the low oxygenated environment.
All fi ve representatives of the catfi sh family, Clarii-
dae, found in the Okavango have accessory air-breathing
organs in their gill chambers, and the two climbing perch
species have accessory air-breathing organs in chambers
above the gills. Many of the cichlid species have physio-
logical adaptations to survive low oxygen levels, but for
a signifi cant number of species the strategies of overcom-
ing the fundamental problem of low oxygen levels are
still unknown.
Low oxygen levels could be lethal for eggs and lar-
vae. To deal with this problem, all cichlid species have
special strategies to provide a better-oxygenated environ-
ment for eggs and fry either by mouth brooding or by
fanning and guarding eggs laid in nests (Skelton, 2001).
The African pike and the climbing perches build nests
of bubble foam on the water surface under which spawn-
ing takes place. The eggs fl oat and are then trapped in the
foam where oxygen levels are high. These fl oating nests
are vulnerable to predation and are guarded by the males.
In some fi sh species, such as catfi sh, the eggs are sticky
and attach to vegetation, which will keep them away from
the frequently anoxic sediments. At least 19 species rep-
resenting 24% of the Okavango fi sh fauna have evolved
strategies to care for eggs and fry.
Endemic species. The fi sh fauna of the Okavango River
and Delta are now, for all practical purposes, isolated.
The link between the eastern part of the Delta and the
Zambezi, the Magwegqana River, also called the Selinda
spillway, no longer receives water. However, this river
was fl owing to the Zambezi even in the recent past during
wetter periods and there are no endemic fi sh species in
the Okavango. The physical conditions in the Delta espe-
cially in fl oodplains, can however be extreme in terms of
low oxygen levels at night and low temperatures during
the winter. These conditions could result in considerable
evolutionary pressure for those species surviving here,
but the time for speciation, since the last drying up of the
Delta, has probably been to short for new species to
evolve.
Reptiles and amphibians
In total 33 amphibians and 64 reptiles have been recorded
in the Okavango Delta (Murray, 1997).
All amphibians are dependent on water at least for
reproduction and/or deposition and hatching of the eggs.
Most of the 33 amphibian species occur close to water
and only 3–5 species are more terrestrial. The most pro-
nounced of these is Chiromantis xerampelina, which
spends the entire life in trees – even during reproduction.
The eggs are deposited in a bubble nest above temporary
water pools when they are still dry. When the rains come
the eggs are washed down into the water. On the other
hand only two amphibian species (Xenopus laevis and X.
muelleri) are fully aquatic. Most other species select hab-
itats close to water and some have adopted strategies of
hibernation or aestivation to survive temporal and sea-
sonal desiccation of habitats.
Out of the 33 species in Botswana (Table 10) twelve
(36%) have a distribution restricted to the Okavango and
326 L. Ramberg et al. Okavango biodiversity
the Chobe and eight (24%) are confi ned to the Okavango
Delta only. These species are tropical and the Okavango
Delta is commonly the southern end of their distribu-
tion.
Out of 64 reptile species (Table 11) recorded from the
Delta the four terrapins (Pelomedusidae), Varanus niloti-
cus, the Nile Crocodile and one snake (Crotaphopeltis
barotseensis) are confi ned to water, while the python and
four snake species in Colubridae mainly occur in swamp
habitats. Most reptiles, 52 in all, are thus terrestrial. Most
of these have a wide distribution in southern and central
Africa. There are on the other hand 10 species whose dis-
tribution in Botswana is restricted to the Okavango and
the Chobe. Seven of them are aquatic or swamp species
and have a northern-tropical distribution. Two species are
terrapins and the other fi ve are snakes.
Birds
The number and variety of birds in the Okavango Delta is
well documented, due largely to the efforts of amateur
birdwatchers who contributed substantial data to the Bird
Atlas of Botswana between 1980 and 1990. This data-
base, where birds have been recorded in a standardized
way for the whole country, has subsequently been kept
updated by the Records Sub-committee of Bird Life Bot-
swana. The analysis of bird diversity that follows is drawn
from these sources, supplemented by personal observa-
tions (Table 12).
There are 444 confi rmed bird species occurring in the
Okavango Delta. This makes the Delta together with the
Chobe River, the most species-rich area in Botswana.
Most are widely distributed species belonging to 74 fam-
ilies of which the most important in terms of number of
species are the following:
Accipitridae (eagles, hawks, buzzards, kites): 38 spe-
cies. This family includes the African Fish Eagle Haliaee-
tus vocifer and African Marsh Harrier Circus ranivorus
as two typical wetland species.
Sylviidae (warblers, apalises, cisticolas etc.): 31 spe-
cies. One of the warblers (Greater Swamp-Warbler A.
rufescens) and three cisticolas (Red-faced Cisticola C.
erythrops, Luapula Cisticola C. galactotes, Chirping Cis-
ticola C. pipiens) are wetland species with their ranges in
Botswana largely confi ned to the Okavango Delta.
Ploceidae (sparrows, weavers, bishops, widows, que-
leas): 25 species. Eight members of this family are wet-
land species, with substantial populations in the Okavan-
go Delta, as follows: Thick-billed Weaver Amblyospiza
albifrons, Spectacled Weaver Ploceus ocularis, Village
Weaver P. cucullatus, Golden Weaver P. xanthops, South-
ern Brown-throated Weaver P. xanthopterus, Southern
Red Bishop Euplectes orix, Yellow-crowned Bishop E.
afer, and Fan-tailed Widowbird E. axillaris.
Ardeidae (herons, egrets, bitterns): 18 species. All of
the ardeids are primarily wetland species, with the excep-
tion of the Black-headed Heron Ardea melanocephala
and Cattle Egret Bubulcus ibis which are widely distrib-
uted throughout Botswana (although both breed exten-
sively in the Okavango Delta).
The Slaty Egret Egretta vinaceigula is the Okavan-
go’s only near-endemic bird species. It has the Okavango
Delta as the centre of its distribution, and it is estimated
that 85% of the global population of this species occurs
here.
All members of the Anatidae are strictly wetland de-
pendent in Botswana, and 12 species are found in the
Okavango Delta. Only four of the nine kingfi shers found
in the Okavango Delta are strictly aquatic and piscivorous
– the remainder are woodland, insectivorous species.
Table 10. Amphibian families in the Okavango Delta, number of
genera and species.
Family Genera Species
Pipidae 1 2
Bufonidae 1 5
Microhylidae 2 3
Ranidae 6 15
Hemisotidae 1 2
Rhacophoridae 1 1
Hyperoliidae 3 5
Tota l 15 33
Table 11. Reptile families in the Okavango Delta, number of genera,
species and wetland/aquatic species.
Family Genera Species Swamp/Aquatic
Testudinidae 3 3
Pelomedusidae 2 4 4
Gekkonidae 5 8
Agamidae 1 3
Chamaeleonidae 1 1
Scincidae 5 10
Lacertidae 3 5
Gerrhosauridae 1 2
Varan i dae 1 2 1
Typ hlo pidae 1 2
Leptotyphlopidae 1 1
Boidae 1 1 1
Colubridae 22 31 5
Elapidae 4 6
Viperidae 2 2
Amphisbaenidae 3 5
Crocodylidae 1 1 1
Tota l 57 64 1 2
Table 12. Ta xonomic composition of confi rmed bird species in the
Okavango Delta.
Major groups Families Species
Non-passerines 33 186
Near passerines 18 79
Passerines 23 179
TOTAL 74 444
Aquat. Sci. Vol. 68, 2006 Overview Article 327
Geographic and habitat distribution. The Okavango Del-
ta falls within the Afrotropical region. However, all
around the Delta to the west, east and south there is an
abrupt change from tropical vegetation to Kalahari wood-
land or dry savannah. Consequently, the distributions of
many bird species, particularly waterbirds, closely mirror
the extent of the Okavango. Nevertheless, it is not easy to
classify the avifauna of the Okavango into wetland-re-
stricted species and those that are not restricted to the
wetland. For the purpose of comparison, the following
three categories may be defi ned:
1. Aquatic species (112 spp.) – those that feed by diving,
swimming or wading, or feed on shores or mudfl ats in
the vicinity of water. These are all non-passerines with
the exception of the coucals and some of the kingfi sh-
ers, which are near-passerines.
2. Non-aquatic species (57 spp.) inhabiting wetland hab-
itats such as fl oodplain forests, palm swamps, marshes
and reed beds. The majority of these are passerines.
3. Terrestrial species not restricted to wetlands (275
spp.). These are mostly near-passerines and passer-
ines, and members of the family Accipitridae (non-
passerines).
Conservation status. Tyler and Bishop (1998) list six glo-
bally threatened and near-threatened bird species which
occur in the Okavango Delta. They are shown in Table 13,
updated to include two additional species listed in ‘Threat-
ened Birds of the World’ (BirdLife International, 2000).
An estimated 85% of the global Slaty Egret popula-
tion is restricted to the Okavango Delta. The Delta is also
very important for the Wattled Crane – it currently sup-
ports the largest, single population of this species and
over 15% of the global population (Beilfuss et al., 2002).
The other globally threatened species are occasional visi-
tors to the Delta or palaearctic migrants.
Seventeen range-restricted or biome-restricted spe-
cies occur in the Okavango (Tyler and Bishop, 1998).
One of these, the Chirping Cisticola is aquatic, and in
Botswana is confi ned to the Okavango. The others are
more widespread, and most are common in their respec-
tive habitats.
Following the criteria laid down by BirdLife Interna-
tional, the Okavango Delta is also of conservation impor-
tance for a substantial number of congregatory water-
birds; it supports over 1% of the global populations of 20
species, and 0.5% of the global populations of another 12
species.
The vast majority of the birds found in the Okavango
Delta are breeding residents (339 or 76%) as shown in
Table 14. There is, however, a signifi cant number of pal-
aearctic migrants all of which are waders (29.3%), that
visit the Okavango specifi cally because of its wetland
habitats.
Mammals
The Okavango Delta has a wide variety of large mam-
mals occurring locally in high numbers, and which are
the main attractions in the growing tourism industry
(Mbaiwa, 2003). However, most mammals in the Delta
are fairly small and often overlooked. The overall mam-
mal biodiversity of this entire community is determined
by such factors as habitat diversity, connectivity to spe-
cies pools in the Southern African region and the envi-
ronmental history of the Delta.
Number of species. Some 122 mammal species of 12 or-
ders and 34 families live in the Okavango Delta (Table 15).
All the larger species are wide spread across the African
Savanna region. The distributional ranges of some of the
larger mammals are marginally within the Delta. One of
these, the Sable Antelope (Hippotragus niger), is common
in the broad-leaved woodlands and the grasslands close to
Table 13. Globally threatened or near-threatened bird species occur-
ring in the Okavango Delta.
Common name Scientifi c name Status
Vul nerable
Slaty Egret
Lesser Kestrel
Cape Vulture
Wattled Crane
Corn Crake
Lappet-faced Vulture
Near-threatened
African Skimmer
Data defi cient
Black-winged Pratincole
Egretta vinaceigula
Falco naumanni
Gyps coprotheres
Grus carunculatus
Crex crex
Tor gos tracheliotus
Rhynchops fl avirostrisi
Glareola nordmanni
Resident
Palaearctic
migrant
Vagrant
Resident
Palaearctic
migrant
Resident
Resident
Palaearctic
migrant
Table 14. Numbers of resident and migratory bird species in the Okavango.
Residents Intra-African migrants Palaearctic migrants
Status uncertain
Breeding Breeding Non-breeding Non-breeding
No. of species 339 40 2 58 5
% 76.4 9.0 0.4 13.1 1.1
328 L. Ramberg et al. Okavango biodiversity
the Delta (Skinner and Smithers, 1990). Similarly, the
Eland (Taurotragus oryx) and the Gemsbock (Oryx gazel-
la) prefer drier landscapes and rarely spend time in the
Delta (Skinner and Smithers 1990), while the White Rhino
(Ceratotherium simum) was recently introduced after their
local extinction (Mosojane personal communication, Bot-
swana Department of Wildlife and National Parks).
Typical forest species do not occur in the Delta al-
though the riverine woodlands in the Delta often have
closed canopies. Their patchiness might be preventive for
the establishment of such species. Similarly, there are no
rocky outcrops in the Delta, so those mammals typical of
this habitat such as the Klipspringer (Oreotragus oreotra-
gus), the Rock Dassies (Procaviidae) and the Dassierat
(Petromus typicus) are not found either. Some species
that occur adjacent to the Delta in the dry Kalahari envi-
ronment such as the Springbok (Antidorcas marsupialis),
the Black-footed Cat (Felis nigripes), and the South Afri-
can Hedgehog (Erinaceus frontalis) seem not to have
been recorded in the Delta either. For these the wetter
habitats here might be preventive. It is, however, diffi cult
to understand why the typical South African mammal
family with many species; the burrowing golden moles
(Chrysochloridae), have not established themselves in
the Okavango region. Similarly the Oribi (Ourebia oure-
bi) is missing, although it is common in the non-distant
Chobe National Park and for which the Delta habitats
seem to be suitable (Bonyongo, 2004).
Species composition and size distribution. Almost half of
the mammal species are bats or rodents (n = 57). Most of
these are small and weigh less than 100 g (Table 16). A
third of the mammals (n = 40 species) are heavier than
10 kg and 11 of these are carnivores. At least 18 species
weigh more than 100 kg. These include the large African
antelopes, the Burchell’s Zebra (Equus burchelli) and the
African Lion (Panthera leo). The four species that are
heavier than 1,000 kg include the White Rhinoceros, the
Hippopotamus, the Giraffe (Giraffa camelopardalis), and
the African Elephant.
The two most common of these megaherbivores
(Owen-Smith, 1988) signifi cantly affect the physical en-
vironment. Elephants that feed on a large variety of
plants, including trees and shrubs, modify the terrestrial
habitats of other species in the Delta (e.g. Gilson and
Lindsay, 2003), for instance by changing woodlands to
grasslands. Hippopotami on the other hand, change both
the aquatic and fl oodplain habitats for species living in
the Delta (McCarthy et al., 1998) by opening up channels
and facilitating fl ooding.
Common species and total biomass. The Impala is the
most common large mammal in the Delta (Table 17), fol-
lowed by the Buffalo and the Red Lechwe (Bonyongo,
2004). Elephants are also very abundant. This species has
increased in numbers from 2,300 (1975/76) to 5,700
(1984/85) (SMEC, 1989), 15,000 in 1988, and 35,000 in
2002 (Bonyongo, 2004). Similarly, the numbers of two
other large herbivores, Hippopotamus and Buffalo, have
increased remarkably during the last fi fteen years, while
most small and medium sized herbivores have declined
(op. cit.). The abundance of elephants is now so high that
they constitute a signifi cant threat to woodlands, especial-
ly when considering that about 1/4 of the Delta is perma-
Table 15. Mammal orders, families and the number of species found
in the Okavango Delta.
ORDER & FAMILY COMMON NAME SPECIES
Order Insectivora 5
Family Soricidae Shrews 5
Order Macroscelidea 1
Family Macroscelididae Elephant shrews 1
Order Chiroptera 26
Family Pteropodidae Fruit bats 2
Family Emballonuridae Tomb bats 2
Family Molossidae Free-tailed bats 6
Family Vespertilionidae Vesper bats 13
Family Nycteridae Slit-faced bats 1
Family Rhinolophidae Horseshoe bats 1
Family Hipposideridae Leaf-nosed bats 1
Order Primates 3
Family Lorisidae Bush babies 1
Family Cercopithecidae Baboons & monkeys 2
Order Pholidota 1
Family Manidae Pangolin 1
Order Lagomorpha 1
Family Leporidae Hares 1
Order Rodentia 31
Family Bathyergidae Molerats 1
Family Hystricidae Porcupine 1
Family Pedetidae Springhare 1
Family Gliridae Dormouse 1
Family Sciuridae Squirrels 1
Family Thryonomyidae Canerats 1
Family Muridae Rats & mice 25
Order Carnivora 28
Family Protelidae Aardwolf 1
Family Hyaenidae Hyaenas 2
Family Felidae Cats 6
Family Canidae
Foxes & dogs
& jackals 4
Family Mustelidae Otters & polecat 4
Family Viverridae
Civets & genets
& mongooses 11
Order Tubulidentata 1
Family Orycteropodidae Aardvark 1
Order Proboscidea 1
Family Elephantidae African elephant 1
Order Perissodactyla 2
Family Rhinocerotidae Rhinoceroses 1
Family Equidae Zebras 1
Order Artiodactyla 22
Family Suidae Pigs 2
Family Hippopotamidae Hippopotamus 1
Family Giraffi dae Giraffe 1
Family Bovidae Antelopes & Buffalo 18
Aquat. Sci. Vol. 68, 2006 Overview Article 329
nently wet (P. Mundy, pers.comm.). The Hippopotamus
numbers given at about 2,500 are likely to be a minimum,
due to the problems in aerial counting of these often sub-
merged animals. For similar reasons the numbers based on
aerial counts given in Table 17 for Sitatunga and Kudu, are
likely to be very under-estimated. The numbers derived by
aerial counts for Impala are very under-estimated as well,
but here it has been possible to correct for this error by
integrating the relationship between aerial and ground-tru-
thed density estimates (Bonyongo, 2004).
The total mammal biomass for the Moremi Game Re-
serve (7,000 km2) in the Okavango Delta has been esti-
mated as being 12,000 kg/km2 (Bonyongo, 2004), which
is much higher than for most wildlife areas in southern
Africa and comparable with the rich savannas in the East
African Rift valley. Compared with regression models
between rainfall (Coe, Cumming and Phillipson, 1976)
and rainfall + nutrient level (East, 1984) the Okavango
wildlife biomass is 4–8 times higher than expected. The
extended productive period caused by the annual fl ood is
certainly one of the causes for this. On the other hand the
generally low nutrient levels in the Delta should limit
biological production (op. cit), but the dynamic vegeta-
tion successions caused by fl ooding with periodically and
locally high mobilization of nutrients may cause high nu-
trient levels for forbs and hence favorable production
conditions for grazing mammals.
Habitat assemblages. All the common species mentioned
above, except the Buffalo, depend on more than one hab-
itat. For instance, the Impala inhabits fl oodplains and
grasslands adjacent to riparian woodlands, while the Red
Lechwe prefers the seasonal fl oodplains close to deeper
waters of the Delta (Skinner and Smithers, 1990). The
Hippopotamus grazes at night often several kilometres
from the rivers and lakes which it uses during the day
(Skinner and Smithers, 1990). Elephants are also water
dependent. They are mixed feeders and use most of the
habitats in the Delta (op. cit.).
There are some clear differences in species composi-
tion along the wet-dry habitat gradient in the Delta. The
Hippopotamus, the Sitatunga, the Cape Clawless Otter
(Aonyx capensis) and the Spotted-necked Otter (Lutra
maculicollis) live in the deeper, usually permanent waters
of the rivers, lagoons and lakes (Skinner and Smithers,
1990). The Reedbuck (Redunca arundinum) occurs in the
seasonally fl ooded areas with lower Cyperaceae species,
while the Red Lechwe frequents the fl oodplain grasslands
in large numbers (Skinner and Smithers, 1990). These wet
habitats support 3 and 21 species each (Table 18) each,
and are different in species composition from each other
and from the drier habitats. They have a high proportion
of grazers while insectivores (mainly bats) are absent.
The drier habitats across the Delta (riverine forests,
riverine woodlands, savanna woodlands, dry woodlands
and dry scrub), on the other hand, support similar groups
of species (Table 18). These habitats are more species-
rich with a total of 110 species (Table 18) and with a
Table 16. Number of species in six body mass classes (kg) recorded for 12 mammalian orders in the Okavango Delta (Data from various
authors in Skinner and Smithers, 1990).
<0.1 0.1–1.0 1.1–10.0 10.1–100 101–1000 >1000
Insectivora 3 2
Macroscelidea 1
Chiroptera 23 3
Primates 1 1 1
Pholidota 1
Lagomorpha 1
Rodentia 24 5 1 1
Carnivora 41310 1
Tubulidentata 1
Proboscidea 1
Perissodactyla 11
Artiodactyla 8122
Tota l 5 1 1 51622144
Table 17. Number of large mammals in the Okavango Delta in
2002, calculated for an area of 20,000 km2, based on 10 aerial counts
done 1988–2002 by the Department of Wildlife and National Parks,
Government of Botswana (from Bonyongo, 2004). For Impala the
numbers have been corrected based on ground counts.
Species Total number
Elephant, Loxodonta africana 35,000
Zebra, Equus burchelli 14,000
Warthog, Phacochoerus aethiopicus 2,000
Hippopotamus, Hippopotamus amphibius 2,500
Giraffe, Giraffa camelopardalis 5,000
Wildebeest, Connochaetes taurinus 8,000
Tsessebe, Damaliscus lunatus 3,000
Impala, Aepyceros melampus 140,000
Buffalo, Syncerus caffer 60,000
Kudu, Tr agelaphus strepsiceros 300
Sitatunga, Tragelaphus spekei 500
Red Lechwe, Kobus leche 60,000
330 L. Ramberg et al. Okavango biodiversity
typical mammal composition of the East-South African
savannas. These include the big cats (Lion, Leopard
(Panthera pardus) and Cheetah (Acinonyx jubatus) which
are common as well as the Spotted Hyaena (Crocuta cro-
cuta); and the endangered African Wild Dog (Lycaon pic-
tus) has a stronghold here. The large number of species is
due to the species-rich groups of bats and rodents (Table
15) which predominantly occur here. The most common
herbivores are listed in Table 17. It is obvious that these
woodlands provide habitats for a larger number of frugi-
vores, browsers, granivores and insectivores than the
open sedge- and grasslands (Table 18).
The diversity of the habitats over relatively small ar-
eas in the Delta should enhance the number of species
(Rosenzweig, 1995). Bonyongo (2004) has corroborated
this further. He found a highly signifi cant positive regres-
sion between habitat heterogeneity for fi ve large protect-
ed areas in southern Africa – including the Okavango
Delta – and their herbivore species richness.
Reproduction. In the Southern African region at least
75% of the species living in the Okavango Delta give
birth during the summer months (Smithers 1971; 1983),
while six species breed during winter only. These same
species breed during winter in the Delta as well (Table
19). In the Delta, however, another 12 species are winter
breeders which are summer breeders in the region and an
additional 27 species are winter breeders which breed
year round in the region. This might indicate that the pro-
longed breeding is a response to the relative predictabil-
ity in abundance of resources during the fl ooding.
Discussion
Species and habitat richness
As the number of species increases with size of the study
area following a log/log relationship (see Rosenzweig
(1995) for an overview), a proper comparison of biodi-
Table 18. Habitat-specifi c number of species recorded in functional trophic groups for nine broadly defi ned habitat types in the Okavango
Delta.
Swamp Reedbed Aquatic
grass
Grassl. Riverine
forest
Riverine
woodl.
Savanna
woodl.
Dry
woodl.
Dry
scrub
Insectivores 2
14%
2
17%
3
19%
1
3%
8
32%
13
31%
29
33%
10
32%
8
24%
Herbivores 8
57%
6
50%
8
50%
17
53%
8
32%
13
31%
31
35%
15
48%
12
35%
Frugivores –––– 2 2 1––
Browsers 1 1 1 1 5 6 5 5 2
Granivores 2– 1 4– 5 7 5 4
Grazers 6 4 414– 115 4 5
Tuberivores –––– 1 1 1 1 1
Carnivores 2
14%
2
17%
3
19%
7
22%
4
16%
7
17%
14
16%
10
32%
9
26%
Omnivores 2
14 %
2
17%
2
13%
7
22%
5
20%
9
21%
14
16%
6
19%
5
15%
Tota l 14 12 16 32 25 42 88 31 34
Table 19. Comparison of seasonal breeding patterns for mammals in the Okavango Delta with that of the Southern African region.
Winter breeders
in SA region
and Delta
Summer breeders
in SA region and
winter breeders
in Delta
Whole year breeders
in SA region and
winter breeders
in Delta
Same breeding in
SA region and Delta
or unknown
Insectivora 2 1 2
Macroscelidea 1
Chiroptera 26
Primates 2 1
Pholidota 1
Lagomorpha 1
Rodentia 6 7 18
Carnivora 4 1 5 18
Tubulidentata 1
Proboscidea 1
Perissodactyla 1 1
Artiodactyla 1 12 9
Tota l 6 12 27 77
Aquat. Sci. Vol. 68, 2006 Overview Article 331
versities between the Okavango Delta and other areas in
the region requires that this relationship is known for
each taxonomic group. This is, of course, not the case. A
crude comparison is nevertheless attempted in Table 20
between the Okavango Delta and the countries in the
Southern African region. The calculated total number of
species per one square kilometer for the Okavango Delta,
329 for the six biological groups, is slightly higher than
for Botswana, probably refl ecting the larger contribution
of aquatic species. The species density, however, de-
creases within the dry part of the gradient and is consid-
erably lower than for the wetter countries to the north
with values of 500–700 species per km2. South Africa
stands out with a much higher species density which is
caused by the extremely species-rich Cape Floral king-
dom. With the exception of this unique area, the species
richness in the Okavango Delta is in the range of the oth-
er biomes in the southern part of the Southern African
sub-continent.
The species rich Odonata fauna is probably a true fea-
ture of the Okavango as this group is well studied in the
region. The large variety of aquatic habitats may be the
reason.
The number of habitats identifi ed in the Delta is not
signifi cantly higher than in the surrounding Kalahari. The
density of habitats, with as a mean 5–6 habitats repeated
30–36 times in each area of 9 km2, may however be high.
Comparisons with other data are diffi cult since it is hin-
dered by the particularities of the habitat classifi cation
used. A high number of habitats per area will give a high
“edge effect” (sensu Leopold) which in turn will favor
species that are using more than one habitat. The high
density of Impala which is the most numerous antelope
in the Delta and uses the woodland-grassland inter-phase,
may be an example of this edge effect.
The highest habitat density occurs at the Delta fringes
(Fig. 5), where the hydrological gradients are likely to be
steepest and have the widest total range. These are, there-
fore, also the areas which probably have the highest total
biodiversity and those species which are dependent on
more than one habitat for their development or their daily
activities will be particularly favored here. This is a new
insight that challenges the conservation efforts of the
Delta (see below).
Flooding, productivity and habitat succession
There are two major processes organizing productivity
and habitat succession in the Okavango Delta: Firstly the
fl ood pulse without which the seasonal fl oodplains with
its entire fl ora and fauna would disappear, and biological
productivity would be dramatically reduced. The second
process is the shift in fl ood distribution over different
time scales. This creates a dynamic patch system of dif-
ferent nutrient levels and at different stages of biological
succession (Fig. 8). During a fl ooding phase – irrespec-
tive of long or short – there is an accumulation of dead
organic matter. If permanently wet conditions continue
for years and decades a layer of peat will accumulate and
keep nutrients such as phosphorus inaccessible in the or-
ganic matrix. When dry conditions reoccur due to drought
or river avulsion which moves the fl ooding elsewhere, the
peat will be oxidized by fi re or microbes and the nutrients
released. A highly productive grassland is formed which
will attracts grazing wildlife and livestock as well as alert
agriculturalists. The high wildlife biomass in the Delta of
12,000 kg/km2, which is about six times higher than ex-
pected (Bonyongo, 2004), indicates the importance of the
mobilization of nutrients for biological productivity. This
is, of course, also caused by the direct effect the annual
fl ood has on the production of grazing for herbivores.
Similarly, this fl ood is signifi cant for the annual cycle of
aquatic productivity by causing the high nutrient levels,
with resulting high zooplankton- and fi sh production on
seasonal fl oodplains (Högberg et al., 2002). As the soils
are composed of fi ne sand, the mobilized nutrients will
not be retained easily on site, but disperse gradually and
Table 20. Number of species in different groups in the Okavango Delta (from this study and for Odonata from Kipping, 2003), as compared
with other countries in the region (from Cumming, 1999).
Okavango
Delta
Botswana Angola Namibia South
Africa
Zambia Zimbabwe
Area 103 k m 2 25 600 1,247 824 1,221 752 390
Mammals 122 154 276 154 247 229 196
Birds 444 569 872 640 774 732 634
Reptiles 64 143 150 140 301 160 156
Amphibians 33 36 80 32 95 83 120
Odonata 94 114 250 117 147 222 157
Fish 71 81 268 97 220 156 132
Flowering Plants 1,300 2,000 5,000 3,159 20,300 4,600 6,000
Plant density per 1 km2 210 182 400 272 1,629 403 591
Total (excl. Odonata) 2,034 2,983 6,646 4,222 21,937 5,960 7,238
Total species density
(excl. Odo.) per 1 km2
329 285 531 364 1,761 522 713
Note: The species density has been calculated using the same formula as in Table 4.
332 L. Ramberg et al. Okavango biodiversity
hence the fertile grassland will become more and more
nutrient-poor and eventually become encroached by
woody species and end up as a dry woodland-bushland.
Sooner or later the fl ood will return and drown the trees
and again form the typical wet-dry gradient from perma-
nent streams, lagoons and sedge-lands over seasonal
grasslands to riparian woodlands. This ever changing
system of biological successions is caused by three fac-
tors: The fact that the Delta is a slightly conical alluvial
fan causes a lack of spatial stability of stream channels;
the substantial sediment transport in the stream channels
in the Delta, causes blockages and forces the water to
fl ow into other areas (McCarthy and Ellery, 1998); and
the predictable seasonal fl ood pulse that is the direct fac-
tor leading to fl ooding of large areas, but in turn depends
on the rainfall pattern in the Angola highlands and the
hydrographic characteristics of its drainage basin and
river channels. In combination these geophysical, hydro-
logical and meteorological factors create the unique Oka-
vango Delta biological landscape.
Speciation and biogeography
There are at least four factors which should enhance the
evolution of new species in the Okavango Delta. In its
present phase it is almost entirely isolated from other
wetlands and is an oasis in the huge Kalahari dry savan-
na. It has a unique hydrology with fl ooding during the
cold season, while usually in the tropics and sub-tropics
the cold season is also the dry one. In these areas the
warm season is also the wet one, often with torrential
rains and fl ooding. This is the case in the Delta as well,
and therefore it has two wet periods with high biological
production. Due to a pronounced continental climate the
shallow waters on fl ood plains at an elevation of about
1,000 m asl often freeze during the cold season. There is
no other large wetland in the Southern African region
with this combination of features. Finally, most aquatic
and wetland species in the Delta: fi shes, snails, odonates,
amphibians, reptiles (the aquatic and wetland species),
originate predominantly from the north and more or less
tropical environments and are likely to be less well adapt-
ed to the specifi c Delta environment, and thus are proba-
bly under evolutionary pressure to adapt. An indication
of ongoing evolutionary changes may be that a consider-
ably larger proportion of mammal species as compared to
the Southern African region are winter breeders here,
which is probably an adjustment to exploit the high bio-
logical production during the cold fl ooding season. As
mating for most of these species takes place a year to a
month before the fl ooding arrives – which thus cannot
trigger reproduction – it is likely that genetic mechanisms
have already evolved.
There are, however, no confi rmed endemic species in
the Okavango Delta. This is probably due to a combina-
tion of geographic features and climate variability. The
interior central part of southern Africa with the upper
Zambezi and Okavango Rivers is an ancient highland
plateau at about 1,000 m asl with extremely low gradi-
ents. Over a distance of more than 500 km from the
Liuwa fl oodplains downstream (Fig. 1) the gradient is
1:10,000. The fl ow in the hydrological connection (the
Chobe tributary) between the Okavango Delta and the
Zambezi, can go in either direction depending on which
of the rivers has the highest water level (Davies, 1986).
Although the longterm climatic variations are not known
in detail, it is certain that there have been very dry periods
as evidenced by fossil sand-dunes around the Okavango
Delta as well as wet periods indicated by widespread al-
luvia extending from the Delta and into the Zambezi (see
Mendelsohn and el Obeid, 2004). Several such climate
Figure 8. The principal habitat successions in the Okavango Delta
following an initial period of increased fl ooding and after a time
span of decades to centuries a shift to a drying phase. The bottom
part of the fi gure illustrates how organic matter and nutrients accu-
mulate, mobilize and disperse during such a sequence. The upper
part illustrates how communities along a typical wetland gradient
will change during this sequence.
Aquat. Sci. Vol. 68, 2006 Overview Article 333
swings have probably taken place during the last 100,000
years. The Okavango Delta has thus been fairly isolated
from the Zambezi system in dry periods and perhaps even
completely dried out, while in wet periods it was proba-
bly part of a huge wetland complex of several 100,000 km2
which occupied the central and southern part of western
Zambia, southwestern Angola, the Caprivi strip in Na-
mibia and northern Botswana with the Okavango Delta
and the Makgadikgadi Pan (Fig. 1). The high connectiv-
ity between the Okavango and the Zambezi is indicated
by the large similarities between them in species compo-
sition of many aquatic biological groups such as amphib-
ians, fi sh, dragonfl ies and mollusks.
During transitional periods changing from dry to in-
creased wetness the Delta expanded and established an
irregular link with the Zambezi system. Few individuals
of wetland species may have dispersed through this nar-
row pathway and found an environment with unused or
under-used niches which are not identical to the ones of
their origins. This is a situation of competitive speciation
(Rosenzweig, 1995), which is a comparatively fast proc-
ess. However, in the next phase when the Delta became a
part of a large wetland system with an easy migration of
species and fl ow of genes, it is likely that the newly estab-
lished genetic diversity – probably mainly below species
level – merged with the larger gene pool of the species
and thus cannot be detected. Although geographical spe-
ciation is expected to occur in this large wetland complex
with an open fl ow of genes, the rate of speciation in this
period was probably low (Rosenzweig, 1995). During
periods of receding wetlands on the other hand, many
niches decreased in their extent, the competition became
more intense and the risk of extinction increased. These
processes may explain the low degree of endemism in the
Okavango Delta. However, for a better understanding
both the frequency of climate changes and the rates of
speciation must be known.
The larger Okavango-Zambezi wetland complex, on
the other hand, has a fair number of endemic species.
White (1983) defi ned the Zambezi phytochorion based
on more than 50% endemic plant species and also identi-
fi ed the “Barotse Centre of Endemism” (White, 1965)
basically identical in extent with the huge wetland com-
plex described above. The whole antelope sub-family
Reduncini, which is almost entirely confi ned to African
wetlands, has a high biodiversity in this area with four
biological species and more than ten mostly endemic
subspecies (Cotterill, 1998). Similarly, 23% of all fi sh
species in this wetland complex are endemic, and more
than half of them originate from the Congo River system
to the north, where the Zambezi has tapped into several
river systems and thus expanded its catchment (Skelton,
1993). This tropical origin of species is likely to also be
the case for many other groups of aquatic biota, espe-
cially because the Kalahari to the south has formed a for-
midable migration barrier for a long time.
The uniqueness of the biology of the Okavango Delta
landscape is intuitively felt. It is, however, not caused by
the occurrence of endemic species nor a high diversity of
species. These features are normal for the Southern African
region. Two other factors, however, may be more unique:
Habitat density – not the number of habitats – is probably
high resulting in a high “edge effect” (sensu Leopold)
which favors species using more than one habitat. Second-
ly, and probably more importantly, the biological produc-
tivity, best refl ected in the very high biomass of large mam-
mals, is much higher than expected. This is in all likelihood
caused by the large scale shifts in fl ooding patterns over
time in combination with the annual fl ood-pulse; mecha-
nisms that both accumulate and mobilize nutrients.
Threats to biodiversity
Development and planning in most human societies strive
towards stability and predictability, which in the case of
the Okavango Delta is in direct confl ict with its inher-
ently unstable nature. There are several examples where
the lack of understanding the Delta’s nature has caused
concerns. Inside the Delta the channelling of water, clear-
ing and dredging of waterways have been done in the past
and are still being proposed even now by the Department
Table 21. Number of species in taxonomic groups of originally terrestrial origin observed in each major habitat in the Okavango Delta. For
plants and mammals the data is for actual number of species observed in each habitat i.e. overlapping species are included, whereas for
reptiles and birds the data is based on the authors’ classifi cation of species into either aquatic-, wetland- or dry land – whereby habitat over-
lapping is excluded.
Taxonomic group Number
of species
Sum observed
in each habitat
Percent habitat
overlap
Aquatic/
Perennial swamp
Wetland/
Seasonal swamp
Dryland/
Terrestrial
Plants (1) 1,061 1,428 35 % 205 519 704
Reptiles (2) 64 7 5 52
Birds (3) 444 112 57 275
Mammals (4) 122 134 10 % 3 21 110
(1) Data from Smith (SMEC, 1989).
(2) Data from this study but observations per habitat are missing.
(3) Data from this study but observations per habitat are missing.
(4) Data from this study.
334 L. Ramberg et al. Okavango biodiversity
of Water Affairs, Government of Botswana (Ramberg
2002; 2004b). The largest scheme of this kind “the South-
ern Okavango Integrated Water Development Project”
(1985–92) was fi nally canceled after a seven year plan-
ning period and a cost of many million dollars due to
strong local opposition and a critical international review
(IUCN, 1993).
Upstream of the Okavango Delta, Namibia has fairly
large plans to pump water from the river for irrigation,
and in Angola there are a number of (old) plans for the
construction of as many as 16 hydro-electrical power
plants (Mendelsohn and el Obeid, 2004). If implemented,
these schemes are likely to have severe negative impacts
on the bio-diversity of the Delta since they will reduce
both the total infl ow and the peak fl ows and thus the ex-
tent of the fl ooded areas. In addition, dam operation will
level out discharge variability, cut off fl ood fl ows and
again reduce the fl ooded areas (Ramberg, 1998). Due to
the trapping of sediments in the reservoirs the frequency
of shifts in fl ooding locations will also be reduced. The
river avulsions driving these changes during fl oods are
caused by transported silt and sand, which eventually set-
tle as sediment in the Delta channels building up channel
blockages (McCarthy, 1992), resulting in shifts of the
channel network.
The highest habitat diversity is found in the fringe ar-
eas of the Delta. As discussed above, it is of course also
highly likely that total species diversity is highest here.
The highest losses of species are, therefore, likely to be
caused by the fi rst water development schemes. The im-
plications for the management of biodiversity in the Del-
ta are immense since it cannot be concentrated on pre-
serving some kind of core area, and is complicated even
more by the fact that these fringe areas with highest bio-
diversity are under strongest local human exploitation
pressure as well.
There are other threats to the biodiversity of the Delta,
in particular from the livestock industry which – through
the Ministry of Agriculture – were responsible for the
aerial spraying against tsetse fl ies (Perkins and Ramberg,
2004a, b). This might have caused a loss of invertebrate
species as indicated by the reduced number of dragonfl y
species over the past 30 years and the disappearance of
many invertebrates after the recent spraying. As usual in
such cases, the lack of background and benchmark data
makes the results inconclusive. The Ministry of Agricul-
ture has also constructed “veterinary fences” to prevent
transmission of diseases from wildlife to livestock, which
have effectively blocked the migratory routes for mam-
mals between the Delta and the Kalahari (Conservation
International, 2003). This has probably been decisive for
the observed reduction of wildlife numbers in the whole
north-western Botswana.
The forces counteracting all these eager developers
are the thriving and developing tourism industry, for
which a pristine Okavango Delta is vital, and the interna-
tional conservation forces. The tourism industry in the
Okavango Delta is hampered by the same weaknesses as
in Africa in general (Ramberg, 1993); the benefi ts to lo-
cal populations are comparatively small and consequent-
ly it has weak local political support, while the livestock
sector which is competing for land is well embedded in
the local and central governance structures. The interna-
tional interests in the conservation of the Delta were
boosted when the Government of Botswana ratifi ed the
Ramsar Convention in 1997 and designated the Okavan-
go Delta “a wetland of international importance”. This
was as a direct response to the threat from Namibia to
draw water from the river through a pipeline to its capital
Windhoek (Ramberg, 1997). In particular the Ramsar bu-
reau and IUCN have worked for a management plan for
the Delta together with the Government of Botswana. For
management of the entire river basin the three countries,
Angola, Namibia and Botswana have established a joint
commission, OKACOM (Permanent Okavango River
Basin Commission) in 1994 which has been fairly pas-
sive up to now, probably due to the civil war in Angola
which ended in 2002. Recent international support has
revitalized the organization which will now embark on
the development of a joint management plan for the
whole river basin.
Acknowledgments
We acknowledge Tom Gonser (Eawag: Swiss Federal In-
stitute of Aquatic Science and Technology) for the lan-
guage editing of our manuscript and Mr. Sampie Ferreira,
University of Pretoria, who provided useful compilations
of mammal data.
References
Andersson, A., T. Gumbricht, D. Hughes, D. Kniveton, S. Ringrose,
H. Savenije, M. Todd, J. Wilk and P. Wolski, 2003. Water fl ow
dynamics in the Okavango river basin and Delta-a prerequisite
for the ecosystems of the Delta. Physics and Chemistry of the
Earth 28: 1165–1172.
Arnold, T. H. and B. C. de Wet, 1993. Plants of Southern Africa:
Names and distribution. Memoires of the Botanical Survey of
South Africa, 62, National Botanical Institute, Pretoria.
Barnes, J. E., L. M. Turton and E. Kalake, 1994. A List of the Flow-
ering Plants of Botswana in the herbaria at the National Muse-
um, Sebele and University of Botswana. Revised Edition by
The Botswana Society and national Museum, Monuments and
Art Gallery, Gaborone.
Beilfuss, R., C. Bento, P. Hancock, B. Kamweneshe, K. McCann, K.
Morrison and L. Rodwell, 2002. Water, wetlands and Wattled
Cranes: A regional monitoring and conservation program for
Southern Africa.
Bird Life International, 2000. Threatened Birds of the World. Barce-
lona and Cambridge, UK: Lynx Editions and Bird Life Interna-
tional.
Aquat. Sci. Vol. 68, 2006 Overview Article 335
Bonyongo, M. C., 2004. The ecology of large herbivores in the Oka-
vango Delta, Botswana. PhD thesis, School of Biological Sci-
ences, University of Bristol, UK, 125 pp.
Bonyongo, M. C., G. J. Bredenkamp and E. Veenendaal, 2000.
Floodplain vegetation in the Nxaranga Lagoon Area, Okavango
Delta, Botswana. S. Afr. J. Bot., 66: 15–21.
Butzer, K. W., 1984. Archeogeology and Quaternary environment in
the interior of southern Africa. In: Klein, R.G. and A. A. Balke-
ma (eds.). Southern African Prehistory and Paleoenvironments,
Rotterdam, pp. 1–64.
Coe, M. J., D. H. Cumming and J. Phillipson, 1976. Biomass and
production of large African herbivores in relation to rainfall and
primary production. Oecologia 22: 341–354.
Conservation International, 2003. A line in the sand. The unsustain-
able expansion of Botswana’s beef industry at the expense of
local communities and the Okavango Delta, CD and briefi ng
document, EIA US, P.O. Box 53343, Washington DC 200009,
USA.
Cooke, H. J., 1980. Landform evolution in the context of climatic
changes and neotectonics in the middle Kalahari of north-cen-
tral Botswana. Transactions Institute of British Geographers,
NS 5: 80–99.
Cronberg, G., A. Gieske, E. Martins, J. Prince Nengu and I.-M. Sten-
strom, 1996. Major ion chemistry, plankton, and bacterial assem-
blages of the Jao/Boro River, Okavango Delta, Botswana: The
swamps and fl oodplains. Arch.Hydrobiol./Suppl. 107: 335–407.
Cotterill, F., 1998. Reducine antelope of the Zambezi Basin. In: Tim-
berlake, J. (ed.), Biodiversity of the Zambezi Wetlands. Biodiver-
sity Foundation for Africa, P.O. Box FM 730, Famona, Bulawayo,
Zimbabwe, pp. 145–199.
Cumming, D., 1999. Study on the development of transboundary nat-
ural resource management areas in Southern Africa – Environ-
mental context: Natural resources, land use, and conservation.
Biodiversity support program, Washington DC, USA, 67 pp.
Dangerfi eld, M. J., 2004. Monitoring of terrestrial invertebrates. In:
Perkins, J. and L. Ramberg (eds.), Environmental monitoring
of tsetse fl y aerial spraying impacts in the Okavango Delta –
2002. Okavango Report Series 2: 52–150. (ISBN 99912-948-8-
8).
Davies, B. R., 1986. The Zambezi river system. In: Davies, B. R.
and Walker, K. F. (eds.), The Ecology of River Systems, W.
Junk, Dordrecht, Netherlands, pp. 225–267.
Dincer, T., H. H. Heemstra and B. B. J. Khupe, 1987. Study, using
stable isotopes of fl ow distribution, surface-groundwater rela-
tions and evapotranspiration in the Okavango swamp, Botswa-
na, Isotope Hydrology, Proceedings Neuherberg Symposium,
IAEA, Vienna.
East, R., 1984. Rainfall, soil nutrient status and biomass of large
African savanna mammals. African Journal of Ecology 22:
245–270.
Ellery, W. N., K. Ellery, K. H. Rodgers and T. S. McCarthy, 1995.
The role of Cyperus papyrus in channel blockage and abandon-
ment in the northeastern Okavango Delta, Botswana. African
Journal of Ecology 33: 25–49.
Ellery, W. N. and B. Tacheba, 2003. Floristic Diversity of the Oka-
vango Delta, Botswana. In: Alonso, L.E. and L.-A. Nordin
(eds.) A Rapid Biological Assessment of the Aquatic Ecosys-
tems of the Okavango Delta, Botswana: High Water Survey.
RAP-Bulletin of Biological Assessment 25: 69–96.
Ellery, W. N., K. Ellery and T. S. McCarthy, 1993. Plant distribution
on islands of the Okavango Delta: determinants and feedback
interactions. African Journal of Ecology 31: 118–134.
Ellery, W. N., T. S. McCarthy and J. M. Dangerfi eld, 2000. Floristic
diversity in the Okavango Delta, Botswana as an endogenous
product of biological activity. In: Gopal, B., W. J. Junk and J. A.
Davis (eds.), Biodiversity of wetlands: assessment, function and
conservation, Backhuys Publishers, Leiden, The Netherlands,
pp. 195–226.
Ellery, W. N. and K. Ellery, 1997. Plants of the Okavango Delta, A
Field Guide, Tsaro Publications, Durban, South Africa, 224 pp.
European Union (EU), 2000. From Landcover to Landscape Diver-
sity in the European Union. EU report on website: http://europa.
eu.int/comm/agriculture/publi/landscape.
Gibbs Russell, G. E., 1987. Preliminary fl oristic analysis of the ma-
jor biomes in southern Africa. Bothalia 17: 213–227.
Gieske, A., 1995. Vegetation driven ground water recharge below
the Okavango Delta (Botswana) as a solute sink mechanism –
an indicative model. Conference on groundwater recharge and
rural water supply. Johannesburg, South Africa, September
1995. pp: 119–124.
Gilson, L. and K. Lindsay, 2003. Ivory and ecology: changing per-
spectives on elephant management and the international trade in
ivory. Environmental Science and Policy 6: 411–419.
Gumbricht, T. J. and T. S. McCarthy, 2004. Channels, wetlands and
islands in the Okavango Delta, Botswana, and their relation to
hydrological and sedimentological processes. Earth Surf. Proc-
ess. Landforms 29: 15–29.
Hancock, P., 2005. Butterfl ies. In Mosepele, K. and B. Mosepele
(eds.): AquaRap II: Rapid Assessment of the Aquatic Ecosys-
tems of the Okavango Delta, Botswana: Low Water Survey.
Draft Final Report, Conservation International, Washington
DC., USA, pp. 149–151.
Hart, R. C., 1997. A limnological profi le of the upper Okavango Delta
at low water level. Sth. Afr. Aquat. Sci. 23: 21–33.
Hart, R. C., N. A. Rayner and H. Mosepele, 2003. A brief commentary
on Okavango Delta Micro-Crustacea. In: Alonso, L. E. and L.-A.
Nordin (eds.), A rapid biological assessment of the aquatic eco-
systems of the Okavango Delta, Botswana: High water survey.
RAP Bullentin of Biological Assessments, Conservation Interna-
tional, Washington DC, USA.
Henning, G. A. and S. F. Henning, 1989. Butterfl ies. South African
Red Data Book, African National Scientifi c Programmes Re-
port No. 158.
Hermann, J.-M., 2003. Vegetation study in a seasonal fl oodplain of
the Okavango Delta. MSc Thesis, Technische Universitaet
Muenchen, Chair of Vegetation Ecology, Freising-Weihen-
stephan, Munich, Germany.
Högberg, P., M. Lindholm, L. Ramberg and D. O. Hessen, 2002.
Aquatic food web dynamics on a fl oodplain in the Okavango
Delta, Botswana. Hydrobiologica 470: 23–30.
Huntsman-Mapila, P., A. B. Kampunzu, B. Vink and S. Ringrose,
2005. Cryptic indicators of provenance in the geochemistry of
the Okavango Delta sediments, NW Botswana. Sedimentary
Geology 174: 123–148.
IUCN, 1993. The IUCN review of the Southern Okavango Integrat-
ed Water Development Project, IUCN, Gland, Switzerland, 543
pp.
Junk, W. J., 2003. Tropical/Subtropical Wetland Biodiversity: Status
of Knowledge, Threats and Sustainable management. In: Ber-
nard, T., K. Mosepele and L. Ramberg (eds.), Environmental
monitoring of tropical and subtropical wetlands, Okavango Re-
port Series 1: 45–69. (ISBN 99912-949-0-2)
Kipping, J., 2003. Die Libellenfauna (Odonata) Botswanas – eine
aktuelle kommentierte Checkliste mit Angaben zu Phaenologie,
Habitatpreferenz und Schutzbedarf der Arten. Diplomarbeit an
der Hochschule Anhalt, Bernburg, Germany, 230 pp.
Lindholm, M., 2006. Zooplankton diversity and successions in the
Okavango Delta, Botswana – diversity and ecosystem interac-
tions in land/water ecotones. PhD dissertation, Department of
Biology, University of Oslo, Norway.
Mbaiwa, J. E., 2003. The socio-economic and environmental im-
pacts of tourism development on the Okavango Delta, north-
western Botswana. Journal of Arid Environments 54: 447–467.
McCarthy, T. S., 1992. Physical and biological processes controlling
the Okavango Delta – a review of recent research. Botswana
Notes and Records 24: 57–86.
McCarthy, T. S. and W. N. Ellery, 1998. The Okavango Delta. Trans-
actions of the Royal Society of South Africa 53: 157–182.
McCarthy, T. S., M. Barry, A. Bloem, W. N. Ellery, H. Heister, C. L.
Merry, H. Ruther and H. Sternberg, 1997. The gradient of the
336 L. Ramberg et al. Okavango biodiversity
Okavango fan, Botswana, and its sedimentological and tectonic
implications. Journal of African Earth Sciences, 24: 65–78.
McCarthy, T. S., W. N. Ellery and J. M. Dangerfi eld, 1998. The role
of biota in the infi ltration and growth of islands on the fl ood-
plain of the Okavango alluvial fan, Botswana. Earth Surface
Processes and Landforms 23: 291–316.
McCarthy, T. S., W. N. Ellery, J. R. McIver and B. T. Verhagen,
1991. Groundwater evolution, chemical sedimentation and car-
bonate brine formation on an island in the Okavango Delta
swamp, Botswana. Applied Geochemistry 6: 577–596.
McCarthy, T. S., W. N. Ellery and K. Ellery, 1993. Vegetation-in-
duced subsurface precipitation of carbonate as an aggradational
process in the permanent swamps of the Okavango (delta) fan,
Botswana. Chemical Geology 107: 111–131.
McCarthy,T. S., W. N. Ellery and A. Bloem, 1998a. Some observa-
tions on the geomorphological impact of hippopotamus (Hip-
popotamus amphibius L.) in the Okavango Delta, Botswana.
African Journal of Ecology 36: 44–56.
Mendelsohn, J. and S. el Obeid, 2004. Okavango River. The fl ow of
a lifeline, Struik Publishers, Cape Town, 176 pp.
Merron, G. S., 1993. The diversity, distribution and abundance of the
fi shes in the Moremi Wildlife Reserve, Okavango Delta, Bot-
swana. South African Journal of Wildlife Research 23: 115–122.
Meyer, T., 1999. Ecological mappings in the research area of the
Harry Oppenheimer Okavango Research Centre, Okavango
Delta, Botswana. MSc thesis, Hochschule Anhalt, Bernburg,
Germany, 181 pp.
Modisi, M. P., E. A. Atekwana, A. B. Kampunzu and T. H. Ng-
wisanyi, 2000. Rift kinematics during the incipient stages of
continental expansion: Evidence from the nascent Okavango
rift basin, northwest Botswana. Geology 28: 939–942.
Monmonier, M.,1974. Measurements of complexity for choropleth-
ic maps. The American Cartographer 1: 2.
Moravec, F. and J. G. Van As, 2004. Some nematodes from the
squeaker fi sh Synodontis nigromaculatus and S. vanderwaali
from the Okavango River, Botswana, including three new spe-
cies. Systematic Parasitology, 59: 169–187.
Mosepele, K., 2000. Preliminary length based stock assessment of
the main exploited stocks of the Okavango Delta fi shery. MSc.
thesis, Department of Fisheries and Marine Biology, University
of Bergen, Norway, 139 pp.
Murray, M., 1997. Fauna of the Okavango River basin (Botswana
sector). Permanent Okavango River Basin Commission, Pre-
paratory Assessment Study, 45 pp. and Annexes A–F.
Nicholson, S. E., 2000. The nature of rainfall variability over Africa
on time scales of decades to millennia. Global Planetary Change
26:137–158.
Owen-Smith, N., 1988. Megaherbivores. The infl uence of very large
body size on ecology, Cambridge Studies in Ecology, Univer-
sity Press, Cambridge, 369 pp.
Palmer, C. G., 2004. Monitoring of aquatic invertebrates. In: Per-
kins, J. and L. Ramberg (eds.), Environmental monitoring of
tsetse fl y aerial spraying impacts in the Okavango Delta – 2002.
Okavango Report Series 2: 31–51 (HOORC, Maun).
Partridge, T. C., P. Demenocal, S. A. Lorentz, M. J. Paiker and J. C.
Vo gel, 1997. Orbital forcing of climate over South Africa: a
200,000-year rainfall record from Pretoria Salt Pan. Quaternary
Science Reviews 16: 1125–1133.
Partridge, T. C., L. Scott and J. E. Hamilton, 1999. Synthetic recon-
structions of southern African environments during the Last
Glacial Maximum (21-28 kyr) and the Holocene Altithermal (8-
6 kyr). Quaternary International 57/58: 207–214.
Penry, H., 1994. Bird Atlas of Botswana. University of Natal Press,
Natal, 319 pp.
Perkins, J. and L. Ramberg (eds.), 2004a. Environmental monitoring
of tsetse fl y aerial spraying impacts in the Okavango Delta –
2002, Okavango Report Series 2: 1–150 (HOORC, Maun).
Perkins, J. and L. Ramberg (eds.), 2004b. Environmental recovery
monitoring of tsetse fl y spraying impacts in the Okavango Delta
– 2003, Okavango Report Series 3: 1–157 (HOORC, Maun).
Petit, J. R., 1999. Climate and atmospheric history of the past
420,000 years from the Vostock ice core, Antarctica. Nature
399: 429–436.
Pinhey, E. C. G., 1967. Odonata of Ngamiland. Arnoldia Rhod. 3:
1–17.
Pinhey, E. C. G., 1976. Dragonfl ies (Odonata) of Botswana, with
ecological notes. Occas. Pap. Natl. Mus. Rhod. B., Nat. sci.
5/10: 524–601.
Pinhey, E. C. G., 1968. Checklist of the Butterfl ies (Lepidoptera,
Rhopalocera) of Botswana. Part 1. Botswana Notes and Records
1: 85–92.
Pinhey, E. C. G., 1971. Checklist of the Butterfl ies (Lepidoptera,
Rhopalocera) of Botswana. Part 2. Botswana Notes and Records
3: 148–152.
Pinhey, E. C. G., 1974. Checklist of the Butterfl ies (Lepidoptera,
Rhopalocera) of Botswana. Part 1. Botswana Notes and Records
6: 197–200.
Pinhey, E. C. G., 1976. Checklist of the Butterfl ies (Lepidoptera,
Rhopalocera) of Botswana (fi nal part). Botswana Notes and
Records 8: 269–288.
Pringle, E. L. L., G. A. Henning and J. B. Ball (eds.), 1994. Pen-
nington’s Butterfl ies of Southern Africa, Second Edition, Struik
Publishers, Cape Town, 800 pp.
Ramberg, L., 1993. African communities in conservation: A human-
istic perspective. Journal of African Zoology 107: 5–18.
Ramberg, L., 1997. A pipeline from the Okavango river? Ambio 26:
129.
Ramberg, L., 1998. Problems facing the management of the Oka-
vango Delta. Reports of the Proceeding of the 2nd Wetlands
Conference, 24th–26th November 1997. NCSA, Ministry of Lo-
cal Government, Lands and Housing, pp. 41–50.
Ramberg, L., 2002. Statement on the proposed removal of aquatic
vegetation and vegetation blockages in the Okavango Delta.
Open letter to Principal Hydrological Engineer, Department
of Water Affairs, Government of Botswana. HOORC web-
page.
Ramberg, L., 2004a. An overview of environmental effects caused
by deltamethrin spraying of the Okavango Delta 2001, 2002 and
recovery monitoring in 2003. In: Perkins, J. and L. Ramberg
(eds.), Environmental recovery monitoring of tsetse fl y spraying
impacts in the Okavango Delta – 2003. Okavango Report Series
3: 1–30. (ISBN 99912-949-8-8).
Ramberg, L., 2004b. Preliminary assessment of EIA for the Maun
Groundwater Development Project. Report to Chief Technical
Advisor, Okavango Delta Management Plan and subsequently
presented to Department of Water Affairs and stakeholders.
HOORC website.
Ramberg, L., P. Wolski and M. Krah, 2006. Water balance and infi l-
tration in a seasonal fl oodplain in the Okavango Delta, Botswa-
na. Wetlands 26(3), in press.
Reed, P. B., 1988. National List of Plant Species that Occur in Wet-
lands: National Summary. U.S. Fish and Wildlife Service,
Washington D.C. Biological report 88(24).
Ringrose, S., 2003. Techniques for assessing the use of near surface
groundwater by riparian trees in the distal Okavango Delta, Bot-
swana. Applied Geography 32: 281–302.
Ringrose, S, D. Lesolle, T. Botshoma, B. Gopolang, C. VanderPost
and W. Matheson, 1999. An analysis of vegetation cover com-
ponents in relation to climatic trends along the Botswana Kala-
hari Transect. Botswana Notes and Records 31: 33–52.
Ringrose, S., C. VanderPost and W. Matheson, 2003a. Mapping eco-
logical conditions in the Okavango delta Botswana using fi ne
and coarse resolution systems including simulated SPOT VEG-
ETATION imagery. International Journal of Remote Sensing
24: 1029–1052.
Ringrose, S. and W. Matheson, 2003b. Assessment of vegetation
cover trends and local ecological factors along the Botswana
Kalahari Transect. Journal of Arid Environments 54: 297–317.
Ringrose, S., P. Huntsman-Mapila, A. B. Kampunzu, W. Downey, S.
Coetzee, B. Vink, W. Matheson and C. VanderPost, 2005. Sedi-
Aquat. Sci. Vol. 68, 2006 Overview Article 337
mentological and geochemical evidence for palaeo-environ-
mental change in the Makgadikgadi subbasin, in relation to the
MOZ rift depression, Botswana. Palaeogeography, Palaeocli-
matology and Palaeoecology 217: 265–287.
Rosenzweig, M. L., 1995. Species diversity in space and time. Cam-
bridge University Press, Cambridge.
Skelton, P. H., 2001. A complete guide to the freshwater fi shes of
southern Africa, Southern Book Publishers (PTY) LTD, Half-
way House, 395 pp.
Skelton, P. H., 2002. Changes to the scientifi c and common names
of southern African freshwater fi shes. African Journal of Aquat-
ic Science 27: 171–174.
Skelton, P. H., M. N. Bruton, G. S. Merron and B. C. W. Van der
Waal, 1985. The fi shes of the Okavango drainage system in An-
gola, South west Africa and Botswana: taxonomy and distribu-
tion. Ichthyological Bulletin of the J. L. B. Smith Institute of
Ichthyology 50: 1–21.
Skinner, J. D. and R. H. N. Smithers, 1990. The mammals of the
southern African subregion, University of Pretoria, Pretoria,
771 pp.
Sliva, J., M. Murray-Hudson and M. Heinl, 2004. Vegetation species
diversity inventory in the Okavango Delta, Botswana. In: Mo-
sepele, K. and B. Mosepele (eds.) AquaRap II: A Rapid Assess-
ment of the Aquatic Ecosystems of the Okavango Delta, Bot-
swana: Low Water Survey. Draft Final Report, Conservation
International, Washington DC, USA, pp. 69–83.
SMEC, 1989. Snowy Mountains Engineering Corporation Cooma
NSW Australia: Ecological Zoning Okavango Delta. Final re-
port, Volume I, Main report to the Ministry of Local Govern-
ment and Lands, Government of Botswana.
Smithers, R. H. N., 1971. The mammals of Botswana. Museum
memoirs of the National Museums and Monuments, Rhodesia
4: 1–340.
Smithers, R. H. N., 1983. The mammals of the southern African
subrgeion, 1st Edition, University of Pretoria, Pretoria.
Stokes, S., D. S. G. Thomas and R. Washington, 1997. Multiple epi-
sodes of aridity in southern Africa since the last interglacial pe-
riod. Nature 388: 154–158.
Thomas D. S. G. and P. A. Shaw, 1991. The Kalahari Environment,
Cambridge University Press Publishers, Cambridge, 248 pp.
Tiercelin, J. J. and K. F. Lezzar, 2002. A 300 million year history of
rift lakes in Central and East Africa: an updated broad review.
In: Odada, E. O. and D. O. Olago (eds.), The East African Great
Lakes, Limnology, Palaeolimnology, and Biodiversity, Kluwer
Academic Publishers, Dordrecht, pp. 3–60.
Timberlake, J. R., 1998. Biodiversity of the Zambezi basin wet-
lands: Review and preliminary assessment of available infor-
mation. Volume 1 (Summary and synthesis), Volume 2 (Techni-
cal
reviews). Biodiversity Foundation for Africa, Bulawayo/Zam-
bezi Society, Harare, Zimbabwe. 241 pp. and 652 pp. resp.
Timberlake, J. R. and S. L. Childes, 2004. Biodiversity in the Four
Corners Area: Technical Reviews, Volume 1 and 2. Occasional
Publications in Biodiversity No. 15, Biodiversity Foundation
for Africa, Bulawayo/Zambezi Society, Harare, Zimbabwe. 496
pp.
Tyler, S. J. and D. R. Bishop, 1998. Important Bird Areas of Bot-
swana. In: Barnes, K. (ed.), The Important Bird Areas of south-
ern Africa. BirdLife South Africa, Johannesburg, pp. 333–354.
Tyson, P. D. R. Fuchs, C. Fu, L. Lebel, A. P. Mitra, E. Odada, J.
Perry, W. Steffen and H. Virji, (eds.), 2002. Global-Regional
Linkages in the Earth System, START, Springer-Verlag, New
Yo rk , 198 pp.
Werger, M. J. A. and B. J. Coetzee, 1978. The Sudano-Zambezian
Region. In: M. J. A. Werger (ed.), Biogeography and Ecology of
Southern Africa, Dr. W. Junk Publisher, The Hague, pp. 301–
462.
White, F., 1983. The vegetation of Africa. A descriptive memoir to
accompany the Unesco/AETFAT/UNSO vegetation map of Af-
rica. Unesco, Paris, 356 pp.
