ArticlePDF Available

Biomass and Production of Large African Herbivores in Relation to Rainfall and Primary Production

Authors:

Abstract and Figures

Standing crop biomass, energy expenditure and production by large mammalian herbivores in the African savannas show a high degree of correlation with mean annual precipitation and predicted above ground primary production. These relationships possess the potential for predicting carrying capacity and protein production from simple meteorological data.
Content may be subject to copyright.
Oecologia
(Berl.)
22,
341-3b4
(1976)
@
by
Springer-Verlag
1976
Biomass
and
Production
of
Large
African
Herbivores
in
Relation
to
Rainfall
and
primary
production
M.
J.
Coer,
D.
H.
Cumming2
and
J.
phillipsonl
lAnimal
Ecology_Research
Group,
Department
ol Zoology,
Oxford
2Department
of
National
parks
and
\,Vildlife,
Salisbury,"Rhodesia
Received
October
Z,
lgTB
,
-B.ummar.g.
standing
crop
biomass,
energy
expenditure
and production
by
rarge
mammalian
herbivores
in
the
A{rican_savannas
show
a
h-igh
degree
of
'correlation-iith
-""o
rrrrroul
precipitation
and predicted
above
ground
primary
production.
These
relationships
possess
the
potential
for predicting
carrying
capacity
and p"tt"tr
production
from
simple
m"t"o"otogicai
data.
Pastoral
and rvildlife
ecosystems
of the
African
sayannas
pose
challenging
management
problems
associated
with
expanding
human
populations
and long
or sliort
term
variations
in
climate.
Increased
human
populations
give
rise
to
i
greater
demand
for
land
and
protein
and
hence
to
enforced
increases
in
rryild
and
domestic
animal
densities
rvhich
can
exceed
the
carrying
capacity
of
the
ecosystem.
Climatic
variations,
particularly
in
precipitation,
affect
the
production
of plant
material
and
hence,
i'directiy,
the
carrying
capacity
of
the
icosystem.
If
effective
management
policies
are
to be
established
there
is
clearly
a
need
to
develop
and
refine
predictive
models
both
for
climatic
variations
and
for
animal
carrying
capacities
and protein
production.
Relationships
of
a predictive
nature
have
been
estabiished
betweeri
rainfall
and
primary
productiJn
(walter,
1954;
whittaker,
1970)
and
bet-'een
actuar
evapotranspiration
and
primary
production
(Rosenzrveig,
1968).
some
rvorkers
(watson,
1972;
LeuthJa,
tsza;
sinclair,
1974;
Western,
1975)
have
recently
noted
an
association
between
annual
rainfall
and
large
African
herbivore
biomass
and
one
author
(phillipson,
1975)
has
indicated
t'hat
elephant
populations
in
Tsavo
National
park
(E'ast)
may
be go-
verned
in
a predictable
manner
by
temporal
and
spatial
variations
in'prima=ry
production.
The
present
paper
examines
both pubiished
and
unpublished
information
on
standing
crop
biomass
of large
mammalian
herbivores
from
wildlife
and
rastoral
areas
in the
savannas
of
east
and
southern
Africa.
It
explores
the
relatioriships
of
these
standing
crop
biomasses
and
their
energy
expenditure
and
production
to
mean
annual
precipitation
and
predicted
above
groot
d
net primary
production.
The
hypothesis
is
advancecl
that,
for
the
majority
of
*uouoou,-*y*tems
examined,
and particula'ly
those
'where
mean
annuai
precipitation
is iess
ihan
700
mm,
it
is
possible
to
establish
statistically
significant
relationships
which
permit
the
pre_
diction
from
rainfall
data
of
large
irerbivore
biomasses,
their
iroduction
and
energy
expenditure.
Standing
crop
biomass
represents
the
sum
of
the live
weights
of
individ.uals
occupying
a given
area.
rn practice
this is
computed
by multiplying
the
density
II.
J.
Coe
el
al'
342
a20
lco
r8o
240
60
120
I80
240
100
3d
i00
Fig.1.Irap"'"3f
,ffiil'rH:":",if
lffi
*':i:',:i"",i'J?#;,:;'Jii3;llli'"il;t'i*censuses
of
each
species
by
its
average
live
I:igttt
and
then
summing
t'he
biomasses
of
In#".-i,,oi"*"i,:H
j:+fr"ilg;
jlXffi;ilili}"';IJT"Ti'fr
1lT*".'"?:3;
uoon
(i)
the
comPtelet
liin""nit
rveight
used
for
each
species'
Complete
and
adequat"
"""*ou
dat'a
Jor
tr'venty
four
predominantlv
wildlife
areas
lvere
obrained
;l;;;;it.r*J
una.onpolrri.h"d
*oit""s
(Table
1,
Fig'
1)'
The
results
u,r"
p'i*u'ffiJ*
tft"
ta*
Africal
Flains
and
li-*11tl^::"U"d
savannas
'lvhere
large
mammals
#"
*o'"
easily
countea
inan
in
the
more
extensive
savanna
woodlands
of
rvest'
t""t*f
and
souther*
-it"ita,'
Dat'a
from
Watson's
(1972)
censuses
of
domestic
iJ"-r".r.
and
wild
h""bi-ror".
in
six
pastoral
areas
of
Kenya
*";n;::tJi"*$t1t?
to
include
major
component's
or
the
large
mammal
community
i"
t'"t*"if-Itrt*"'
-p""*t
t'"i;t;l;;bivore
biomass
were
omitted
Biomass
and
Production
o{
African
Herbivores
843
from
our
anaiyses
(Bourlidre,-196b;
Tinley,
1969,
1g71;
Caughley,
19ZJ;
Stervart,
1963)
as
were
t'hose
for "'vhich
we
could
not
calculate
overa"ll
dean
densities
for
the
ecosystem
in question
(Lamprey,
1g62,
Ig68;
Dorvsett,
1966;
Montf
ort,
lg72;
Travassos
santas
Diaz,
!g7L).-a
wrae
variety
of
census
tecirniques
has
been
used
in
the
African
savannas
but
since
aclequate
tests
of
their
accuracy
are
gene-
rally
lacking
we
r\.'ere
not
able
to
reject
any
census
purely
on
the
basis
of
the
tech_
niques
emploved.
Neither
did
rve
reject
any
census
on
the
grounds
of
the
area
it
encompassed,
except
rvhere
the
resurts
clearly
referred
to
a
rimited
and
un_
represent'ative
part
of
a larger
ecosystem
as
in
the
dry
season
concentration
area
of
32 km2
studied
intensively
by
Lamprey
(1963).
A
major
source
of
.variation
r"
p"utlir"a
biomuss
figures
lies
in
the
great
range
of
estimated
animar
weight
used
by
different
uothor"
[.ab'e
2).
rn
an
attempt
to
reduce
this
largely
artificial
source
of
variation
rve
recalculated
bio-
masses
for
each
area
using
standart
unit
rveights
for
eacrr
*p".i"s
(Tabre
2).
we
appreciate
however
that
both
the
grorvth
rates
of individu-ars
and
the
age
and
sex
structure
of
species
popura.tions
may
differ
from
area
to
area
and
from
year
to
year.
Detailed
age
and
weight
data
and
grorvth
curves
are
available
for
onry
a few
species
and
even
fev'er
data
are
availa=ble
on
regional
l.u.iufiorr*
in
grorvth
and
population
structure.
The
information
availabll
for
buffalo
provides
an
illustration
of
the rlegree
of
error
which
may
be
involved
in
using
standard
r'r'eights
over
wide
areas.
The
average
individuai
weight
of
buffaro
in
the
serengeti
is
424kg
(Sinclair,
1972)
'while
in
lhe
southern
pariof
the
Kriiger
National
park
it
is
490kg
(Pienaar,
1g69);
rve
adopted
an
intermediate
weight
of
4b0kg.
The
effective
error
for
the
calculated
buffalo
biomass
of
Serengeti
is
an
overestimate
of
6% (at
3.9
buffalo
km-z)
(sinclair,
1973)
and
for
Iftiigerbark
u.rd
o.,a"".*timate
of
9%
(at
0.56
buffalo
km-z)
(pienaar,
t'ooo;.
such
errors
are
probabry
within
the
sampling
errors
of
census
procedures
and
annual
variations
in
iopuiation
structure
and
individual
grolvth.
Given
the
census
data
presently
available
we
consider
the
use
of
standard
unit
*'eights
for
each
species
as
being
justified
for
comparative
purposes.
Energy
expenditure
for
each
community
lvas
based
on
the
rerationship:
Basal
Metabolic
Rate
(BMR):70
\4ro*o,zs
koal
.d-1.
Lamprey
(1964)
and
Eltringham
(1974)
used
BMR
x
B
as
an
approximation
of
energy
expenditure
in
rvild
herbivores.
Results
reported.
ry
M"#irg73)
indicate
that
BMRx
1.b
is
a
more
rearistic
estimate
of
eiergy
e*i"nartu""
in
free
living
ungulates.
The
totar
metaboric
weight
of
each
.o-iorrity
n**
.ur.rrtuted
using
unit
metabolic
rveights
(wrgqru)
for
Jach
species.
Estimates
of
energy
expenditure
in
kJh-l
u-ere
calculated
foi
each
co-mri.ity,
Energy
Expenditure
(kJ
n-r;
-
70 (\'Vkg0'?5)_j.5'4.1868
(1
Cal:4.186S
J)
Estimates
of
secondary
production
were
derived
from
calculated
biomass
figures
using
production
to
biomass
ratios
forthreesizecategoriesoflargeherbiv_
ore'
Biomass
for
each
community
was
divided
into
large
laloo
tg;,
intermediate
(100-750
kg)
and
smail
(5-90
kg)
herbivores.
with
the
Jxceptiorri'ot
urrtiuto
urra
344
\1. J.
Coe
et al.
Tabie 1. Sqmmary
of calculated
standing
crop biomass,
energy
expenditure
and
production
'were
Locality
Authorit5'
for
census
Large
herbivore
biomass
kg km-r
A. Wi,ldl,i'te
areas
1.
Rwindi
plain,
Albert National
Park, Zaire
2.
Rrvenzori National
Park,
Uganda
3.
Bunyoro North, Ugancla
4. Ilanyara National
Park, Tanzania
5. Ngorongoro
Crater,
Tanzania
6.
Lake
Nakuru
Nat'ional
Park, Kenya
7.
Amboseli Game
Reserve, Kenya.
8.
Lochinvar ranch,
Zambia
9.
Lake
Rudolf
(East)
Kenya
10. Samburu-Isiolo,
Kenya
11.
Nairobi
National
Park,
Kenya
12.
Tsavo
National
Park
(East)
N. of
Voi River,
Kenya
13.
Tsavo Nat,ional
Park
(East)
S. of Voi
River,
Kenya
14.
Ilkomasi Game
Reserve,
Tanzania
15.
Loliondo
Controlled
Area, Tanzania
16.
Serengeti
National
Park,
Tanzania
17.
Ruaha Na,tional
Park,
Tanzania,
18.
Akagera
National
Parh,
Rwanda
19. Sengwa
\trrildlife
Research
Area, Rhodesia
20.
Hend.erson's ranch,
R,hodesia.
21.
Kruger National
Park,
Northern
section
22.
Kruger Na,tional
Park, Southern
section
23.
Willem
Pref,orius Nature
Reserve, S.
Africa
24. llfolosi
Game Reserve. South
Africa
t7l
[23]
[35,
36]
[71]
[66]
[28]
u2l
t51l
[42,
60]
t3l
124l
t38l
[38]
t26l
[70]
127,
43, s6, 48)
[43
[5e
[13
[15
[5e
[50
14
l3e
I7,448
19,928
t3,261
19,189
7,561
6,688
4,848
7,568
405
2,018
4,824
4,033
4,388
1,731
5,423
8,352
3,909
3,980
4,31s
2,869
984
3,?83
3,344
4,385
B. Pastoral,
Areas
25. trIandera district,
Kenya
26. \Irajir district,
Kenya
27.
Garissa
district,
Kenya
28. Turkana district,
Kenya
29. Samburu
district,
I{enya
30.
Kaput'ei
district, Kenya
[6e]
t6el
[6e]
[6e]
[6e]
t6el
1,901
1,151
3,818
2,446
6,514
7,884
Note:
Biomass clata
for
pastoral areas
are
from Watson
(1972) and were
not
recalculateC'
the authors.
l,ildebeest
production
to biomass
(P/B) ratios of 0.05,
0.20
and
0.35
were used
for large,
intermediate and small
herbivores
respeetively,
these ratios lr-ere based
on
the
limited
data available
on
production and
biomass
in the literature
(Petrides,
1965; Buechner
and
Golley,
1967; Du
Plessis, 1972;
X{ontfort, 1974). The P/E
ratios
for
buffalo
and wildebeest
weve
calculated
from population and
growth
data
provided by
Pienaar
(1969)
and Watson
(1930), respectively;
these ratios
approximated
tire value
of
0.2
used
for
intermediate
sized
herbivores.
Ferv
reports on
lalge
herbivore
densities
and
biomasses
included
adequate
climatological
data
o{ t'he
areas
to
rvhich
they
referred'
trYe
relied primarily
on
Thornthwaite
(1962)
and
reports
of
the
East
African
Meteorological
Deparbment
Biomass
and
Production
of
African Herbivores
of large
herbivores
together
s'ith
rainfall
data
for
areas
from
lvhich
complete
census
data
available
345
Energy
Production
Annual
expenciiture
kg
km-z
a-r
precipita-
kJ.km-r.h-l
(liveweight)
lion(mm)
lleteorological
station
Authority
for rainfall
62,175
76,253
43,979
65,905
56,105
35.498
22,970
40,591.
2,436
9,745
24,728
12,968
15,239
6,154
26,962
43,063
10 11t
t9,257
18,299
15,148
4,4I4
20,552
18,427
17,o17
1,936
2,5d4
1.,L45
2,105
1,503
1,409
934
1,983
6/
402
1,008
it.f I
a.ld
L4i
r,to+
JO+
871
722
684
t62
884
/uD
605
863
1l.|1n
1,150
915
893
878
350
813
165
J /O
844
553
553
784
803
625
/ 5i)
oul
446
312
650
520
650
t7l
L2l
t8l
[17]
12rl
[62]
t73l
[51]
162l
t3l
121l
L2r)
l17l
[26]
162l
l42l
t21l
15,
401
[13,
14]
l15l
[62]
t62l
t4l
tssl
Rwindi
Camp
Ilean
of
4 stations
District
Office
Nakuru
Lodrvar
Archers
post
Nairobi
^{irport
Isohyets
Integrated
llean
Several
stations
Loiiondo
Integrated
Mean
Several
stations
Gabiro
Senqrva
gorge
Pafuri
Komatipoort
228
218
398
330
500
710
trIandera
I{airbesrvein
f<u;rrao
12rl
t21l
calculated
from
isohyets
calculated
from
isohyets
caiculated
from
isohyets
[21]
Fuli
tables
giving
ilensities,
biomass
ancl
metaboiic
weights
in each
area
are
available
from
Logro
NAAP:Logro
AE (1.66
*0.27)-(1.66+0,07)
(L962,
1961)
for
data
on
mean
annual
precipitation
for
the regions
studied.
Rainfal
o'rer
large
areas,
encompassing
marked
rainfaii
grad.ieniis,
\.vas
carcurated
by
measuring
areas
falling
betr.veen
isohyets
and
so
cleriving
an
overali
mean
rainfall
figure
for
the
area.
trVe
have,
as
{ar
as possible,
attempiecl
to
cross
checl<
rainfall
data
against
&s
many
sources
as possibie.
_
Fen'if
any
data
are
avallabre
on primary
production
in
the
areas
covered
in
this
report,.
To gain
some
icea
of tire
possible
rJationship
bet.ween
rarge
herbivore
biomass
and
primary
procluction
ri,e
used
R,osenzrveig's
lloos;
preilictive
equation:
346
M' J'
Coe
et crl'
Table
2. Range
of unit
weights
used
by
thirteen
authors
in the
calculation
of large
herbivore
biomass
and
the
unit.nei"ghts
adopted
for
the
re-calculation
of
standing
crop
biomasses
Species
Av. 'rveights
Ratio
t]nit.
max./min.
rvt,.
Authority
max,
Elephant
1,700d
Black
rhino
660
r
\Thite
rhino
1,360k
Zebta
160r
Hippopotamus
1,000d
Warthog
30i
Bushpig
Forest
hog
6801
310
f
210
f
136i
70e
25r
140
e
IALC
159r
130
f
30f
40e'
i
90f
82m
4,990c
1,000
m
2,000b
290b
1,400b
70b
140b
800
m
664
m
544e
220a
104e
50b'
e
159
m
t7\il
250b
181
c
205r
54r
70b
141
c
130b
68a
226s
60b
r,725*
t5ru
1,500
200
1,000
45+
90
750
450*
340
136*
70
lo
30r
150
150
220
185
160
72*
40
5d-
125
100
91
53*
t23*
120
40
2.9
1.5
1.5
1.8
L.4
OQ
t33l
[3e]
Compromise
between
[48]
and
[39]
t26l
[23]
[13]
[39,
18]
6496
of 140
kg
and
in
similar
proportion
as
for
rvarthog
(45:70)
[26]
Based
on
[56]
and
[50]
[26]
and see
notes
[39,77]
[26,38]
t3el
[78]
138l
See
notes
859i,
adult
t
rveight
[76]
85 91,
adult
t
v'eight
[76]
123l
[51]
[26]
tel
[38]
[23]
[4e]
[1e]
I70l
t4l
126,491
Giraffe
Buffalo
Eland
Grea'ter
kudu
Lesser
kudu
Nyala
Bushbuck
Beisa
oryx
Gemsbok
R,oan
Sable
lYater
buck
Lechrve
R,eed.buck
Kob
Hartebeeste
Topi
Tsessebe
Blesbok
Wildebeest
\Vildebeest
Black
Lnpala
1.2
2.1
2.6
1.6
1.5
2.0
1.1
1.8
1.1
1.6
l.It
1.8
1.6
1.6
108
I
901
1.9
*
Unit
weights
based
on
age rveight
data
and
population structure'
+
Includes
duiker,
klipspringer,
Grysbok,
Steinbok'
'1+1;
o1S1;
cilSl;
dl2il;
et2?l;
fl2il;
e[50];-h[!8];
i[3e];
:[a0];-k[49];
1[55]; m[61]'
,qaaiti"""i
data"on
iody
weights
w.ere
consulted
in
[37]
and
[54]'
Biomass and
Production of
African
Herbivores
Table 2
(continued)
Spet'ies
Ar'.
welghts
Authority
mln. max.
Grant's Gazelle
Thomson's
Gazelle
Gerenuk
Springbok
Others-
Cattle
Shoat'
atl
12t
50m
15b
[26]
173,
521
t38l
[4]
and calculated from
[45]
Followed most authors
[6e]
l6el
26*
10
180
18
where
NAAP:
net &bove
ground
primary
production
in
g m-25,-r
(dry
rveight)
and
AE:annual
actual evapotranspiration
in mm a*1.
Act'ual
evapotranspiration
may be
regarded as
a simultaneous
measure
of \yater
availability
and solar
radiation-both
of
rvhich
limit the
rate of
photosynthesis
(Rosenzu'eig, 1968).
In
semi-arid
and
alid envilonments
$'hefe potential
evapotranspiration
is greater
than annual
precipitation
(AP),
annual
AE:AP3;
we have
taken AE to be
equal
to AP up
to
700
mm a-1. Vaiidation
of
Rosenz'weig's
formula
for the African
savannas
has not been attempt'ed
but
Phillipson
(in press)
found good agreement
bet'ween
predicted NAAP and
net primary
production
measured by
Cassady
(1973) at Buchuma, adjacent
to the
Tsavo
National
Park
(East)
in Kenya.
The result's
for 24 u-ildlife areas are summarised
in
Tabie 1. The biomasses
for
tlre six
pastoral areas in Kenya are
those
given by
Watson
(L972).
He un-
fortunately did
not give
densities of
the'rvild herbivore
species in these areas
and
'we
could not recalculate biomass estimates or
calculate comparable
energy
expenditure
and secondary production.
The relationship betlveen large herbivore biomass
and
mean
annual rainfall is,
over the
range
of data
available, curvilinearlvitir
an
inflection in t'he region
of
800
to
900
mm a-l of
rainfall. A
least
squares
regression of
logro large herbivore
biomass on
log10 rainfall shox-ed a
highly significant'
(P:0.001)
relationship
betlreen
the tu'o variables
(X'ig.2a).
Simiiarly
significant
relationships
obtained
bet'ween estimates
of energy expendit'ure
and
rainfall and bet'ween estimates of
secondary
production and rainfall
(n'ig.2b
and c).
Note that biomass, energy expenditure and
production in solely rvildlife
communities
and in those comprising
rvildlife and domestic
livestock
were
very
similar.
In
semi-arid
rvildlife areas receiving less than 700
mm a-l
(belo'w
which AE:
AP) there was a significant
relat'ionship
betu'een
large herbivore
biomass and
predicted
above
ground
primary production
(Fig.
3).
Adequate data
on AE for
area,S
leceiving more
than 700
mm a-l of
rainfall
t'ere
not, available so
u'e could
not test
this relationship
for
such
areas.
Biomass
in pastoral areas
\Yas higher
R,atio Unit.
max./min.
rvt.
1.6
t.J
40
15
t--
348
M. J. Coe et al.
Fig.
?a c. Relationship
betwcen
Iarge
herbivore standing crop biomass
(a),
energy
expenditure
(b),
and secondary
production (c),
and mean annual
precipitation
for wildlife
(
o)
and
pastoral
systems
(f
)
in east
and
southern
Africa. Least squarcs regression
equations
and regression lines are indicated
on each
graph
togetherw-ith the correlation
coefficient
(r)
anci
its significance
4
).
E
;
=
-
d-
=
i
-
o
ts
;-
2.0
3,C
r.6
2
3.0
2.5
LoC.^
RAINFALL
{
mm a
-r)
1U
r=0.832
P<0.001
N=24
2.5
LoGICRAlNFALL rnn
a-r
)
than
in wildlife
areas
but'
also shorved
a significant, reiationship
to predicted.
NAAP
(r:O,gSt,,
P:0.001).
An equally
significant relationship
was found
bet'ween
tire values
of large
herbivore
biomass
and rainfall in
the semi-arid
wildlife
ecosystems (r:0.94,
D.n'.:10,
P:0.001).
The calculated regression
equation for
the tv'elve
wildlife ecosystems
receiving
less than
700
mm
a-l of rainfall lvas:
Large
irerbivore
biomass:8,684
(+2.28)
AP-1205.9
(+156.6).
There is
a clear empirical
relationship
bet'lveen iarge herbivore
biomass and
mean
annuai
rainfall l'hich
provides
a
basis for first order predictions
of iarge
herbivore
biomass
from
meteorological
data in
the AJrican
sav&nnas. In this
cont'ext
it is rvorth
noting
that
Walter
(1954)
found
a
linear
relationship
betv.een
primary
production
and
rainfall
in
semi-arid regions
cf
south
\\rest
Africa.
Similarly,
\Yhittal<er
(1970)
has shorvn
an almost linear
relationship
betu'een
primary production
and rainfall
in a range
of vegetation t;pes grorving
in rainfali
regimes
of betu'een 100
and
800 mm
a-1.
Since
iarge herbivore
communities are
ultimately
limited
b;z their food resources (Sinclair,
1974; Lack, 1954; Hairston
eta,1.,
1960;
Wynne
Edrvards,
1962) the ca,usal
link
betneen
rainfall and large
herbivore
biomass
is rnost
iikelv
to ooerate throush
the
effects
of rainfall on
Locr0RAlNFAt[
nn a-ll
Biomass
and Production
of African
Herbivores
2.5
3.0
2.5
3.0
LOq0RAINFALL
{ mm a-l
)
LOG-^NAAP
(qmm-d')
tu
-
Tig.
3
Fig. 4
Fig'
3. Relationship
betrveen
large
herbir.ore
biomass
and
predictecl
NAAF
for
wildlife
eco-
systems (
o)
rvhere
annual precipitation
is less
than
700 mm a-1.
Note
the
higher
biomasses
in
pastoral
areas (-f-)
r,r.irich
u'ere
omitted from
the calculated
regression
Iig.4.
Carrying
capacity
in African
rvildlife
areas
in terms o{
large herbivore
standing
crop
biomass
and
mean annual precipitation.
Except'ional
area,s
(o)
discussed
in
the text,
and
omitted
from
the calculated
regression
l-ere 1. Rrvindi piain,
2.
Iiwenzori
National Park,
4. Manyara
National Park,
7. Amboseli and pastoral
systems (f
).
The
mean
and individual
prediction
limits are
shon-n by
full and
broken iines,
respectively
primary
production.
The
close
relationship
betlreen large
irerbivore
biomass
and
predicted
NAAP (Fig.
3)
in
semi-aricl regions
provides
some
support,
for
this
hypothesis.
We should
emphasise
tliat
our model
refers
to large
herbivore
communities
as
a rvhole
and not
to
particular
species.
Sinclair (1972)
reported
a significant
relation-
ship
betu'een
buffalo
density
and rainfali
in
a number
of
east
African
liabitats.
This
single
species
correlation
does
not liold
over
the
full range
of
ecosystems
we
have
examined
and biomasses
of particular
dominant
species (elephant,
buffalo,
t'ildebeest,
zebra
and impala
for
example)
did
not
shov, cleal
correlations
1-ith
rainfall.
This
suggests
that
the
total large
mammal
biomass
in
African
savanna
ecosystems
is
a
reflection
not,
only
of levels
of primary
production
but
also
of
competitive
interactions
betrveen
species
comprising
the communities.
Species
composition
also
varies greatly
in
the different
ecosystems
and
is
presumabft'
a
reflect'ion
of
the varying
habitais
they
contain.
carrying
capacity,
or the
abilitv
of a given
area
to
support
a certain
population
of
animals
on
a continuing
basis (De
vos, 1g69)
may
be altered
by
both long
and
short
term
variations
in
climate
and particularly
in precipitation
(phillipson,
1e75).
The
majority
of
censuses
referred
to in
our
ana,iyses
cor,,ered
oni;,
short
time
periods
and
some
areas may
have
been
overstockecl,
a,ncl
others
understocked,
at,
the
time
of the
census.
In
some ecosystems
the influence
of rainfall
on primary
product'ion
and hence
on
carrying
capacity
may
be modified
by factors
such as
soil
and drainage
patterns
and particularly
ground
u'ater leveis,
e.g. Lake
Manyara
349
44.
+
=
;
o
=
-dn
=
;
.
a
a
22
2
y
=
J.665
rt0.Z)8 x-1.@5
i
o.atbc2,z
r 0.96
Y'.'/.
l[. J.
Coe
ef
al.
National
Park
(Douglas-Harnilton,
1972).
For
these
reasons
the
regressron
equations
of
Figs.
2-4
mav
not reflect
the
carrying
capacity
of
all systems.
The areas
g,hich
lie most
noticeably
above
the
regression
line of
Tig'
2a are
the
pastoral
systems
censusecl
by watson
(1972),
the
Rt'indi
plain
and
Rlvenzori
National
Park
rvhich
lie in
the east,ern
rift valley
and
those
of Nlanyara
and
Amboseli
(see Table
1).
The biomasses
in the
six
pastola'l
areas
may be
relativell'
higher
than
in other
areas
because Watson
used
different
unit
rveights
for
l'ild
he.-rbivor"s.
Biomass
figures
for the
pastoral
Northern
Frontier
districts
of
Kenya
(Mandera,
\Yajir
and
Garissa)
includecl
a high
proport'ion
of elephant'
for
rvhich
watson
(1972)
used
a
unit
$'eight
of
2,251kg.
Biomasses
using
a unit
rveight
of
1,725
kg
for elephant
(Table 2)
.rvould.
thus
be
lorver
than
those
given in
Table
1.
These
plstoral
areas
are,
horrever',
also
those
inlvhich
high
mortality
occurred
duringlhe
drought
of
1973-1974
(Casebeer
and
l\Ibai,
L974)
suggesting
that
the
biomasses
of
these
aleas
\Yere
also above
the
long
term
carrying
capacities
of
the
ecosystems.
The
stand.ing
crop
biomasses
of
Rtindi
and
Rl'r'enzori
are
exceptio-
nally
nlgn;
both
areas
are associated
rvith
eutrophic
volcanic
ash
soils
and
alluvial
soils
1Anon.,
1962)
.w.hich,
in conjunction
$'ith
high
rainfall,
could
rvell
result
in a
higher
primary
production
than
in
other
areas
of
comparable
rainfa,ll
but on
less
mirkediy
tertite
soils.
Lake
Manyara
National
Park
supports
a
particularly
higlr
populatitn
of
elephant
and
buffalo
but
because
it
is
supplied
by
ground
v'ater
iro*
th"
rift valLy
rvall
it, also
supports
a richer
vegetation
than
surrounding
areas
of
comparabie
rainfall
(Dougla,s-Hamilton,
1972).
The
Amboseli
system
is
similarly
supported
b;r
an
internal
clrainage
system
resulting
in
swamps
and
a
far
richer
food
*"""
than
$,ould
normally
obtain
under
the
prevailing
rainfall
regime
(\Yestern,
1973).
If
these
areas
are
excluded
from
general
considerations
of
carrying
capacity
in the
African
savannas
then
the
relationship
between
large
he"birror"
bi,omass
ancl
rainfall
in the
remaining
tu'ent'y
systems
results
in a
higher
correlation
coefficient,
(r:0.96,
P:0.001),
a
mote
conservat'ive
prediction
of
biomass
and one
probably
more closely
related
to the
long t'erm
carrying
capacity
of rvildlife
areas
(n'ig. 4).
The broad
relationship
betr,veen
biomass
and
rainfall
does
not take
into
account
the
local
temporal
and
.spatial
vadations
rvithin
individual
ecosystems.
Field
and
Laws
(1970), for example,
reported
a
range
of
betlr'een
5,136
and
29,493
kgkm-2
in
four
major
habitat
types
in
the
Rrvenzori
National
Park,
uganda;
similar
but
less extreme
variations
are
to be
found
in ot'her
game areas
and
t'he
phenomenon
of
seasonal
game
concentrations
is
*'ell
knor-n.
such
spatial
and
i"-por*.y
local
plienomena
rvithin
ecosystems
do
not
invalidate
the
findings
of
this
study
.rvhich
focus
on
1,he
'lvhole
area
used
by a
muitispecies
large
herbivore
communit'y
on an
annual
and
long
term
basis.
Broad
comparisons
of
herbivore
biomass
in
relat'ion
to
habitat
(vegetation)
types
have
failed
to
yield generalisations
of
a
predictive
nature
on
questions
of
standing
crop
biomass,
productivity
and
caff;-ing
capacity
(e.9.
Pienaar,
1966;
De Vos,
1969;
Bourlidre
and
Hadley,
1970)
;management
decisions
have
thus
perforce
been
based
largely
on
evidence
of
vegetation
change
(e.9. Laws,
1970;
inderson
and
\Yalker,
Ig71)
or
on
the
responses
of
large
herbivore
species
t'o
varying
conditions
of
their
habitat
(sinclair,
1974;
La$'s,
1970).
A
major
de-
Biomass
and
Production
of African
Herbivores
351
parture
from
this
approach
rvas
phillipson's
(197b)
attempt
to
predict
and
explain
variations
in
carrying
capacitv
in
Tsavo
National
ea*t
lnust;.
clearly
the
empirical
rerationship
bet*.een
large
herbivore
biomass,
energy
expendit'ure
and
product'ion
ancl
mean
annual
precipitation
demonstrated
in
this
studv
is
an
advance
on
earlier
attempts
to
vL*.
carrying
capacity
in
terms
of
vegetation
type
or herbivore
response
and
'lvill
allorv
first
order
predictions
of
carrying
capacity
of
specific
ecosystems
to
be
made.
These
results
also
underline
the
need
for
more
comprehensive
studies
of
entire
ecosystems
and
the
full
com_
plement
of Iarge
mammalian
herbivores
thev
support.
More
importanily,
perhaps,
they
provide
a
basis
for
crude
estimated
of
trr" potential
prJduction
of
animal
protein
from
the
African
savannas.
Acknowled'gements.
We thank
Dr.
II.
Norton-Griffiths
for
providing
us
rvith
unpublished
data
and
Mr.
K.
Tinley
for
information
on
.r.vildlife
areas
in rroci,mbiquelone
of
us
(D.H.M,O.)
u'as
supported
by
the
Department
of
Nationai
parks
and
wild
Life'Mu"ug"-"rrt,
Rhodesia,
whilst
on
sabbatieal
leave
in
Oxforcl.
Referenees
1. Anderson,
G.D.,
\valker,
B.H.:
vegetation
composition
and
erephant
damage
in the
Sengwa
\Vildtife
Research
Area.
J.
Sth. Atr.
Wildt.
Mgmt.
Ass.
4(1j,
Fta
$W+)
2'
Anonymous:
Atlas
of
uganda.
Department
of Lands
and.
surveys,
uganda,
g3
pp.
(1g62)
3'
Barkham,
J.P.,
Ridding,
c.J.:_The
large
animal
populations
anJ
vegetation
of
the
sambum-Isiolo
Game-Reserve,
Kenya.
A preriminary
i.,r"y,
school
of
'En.r-ironmental
Sciences,
University
o{ East
Anglia,
Norrvi.t
lfOZ+y
4' Borquin,
O.: Utilisation
and
aspects
of
management
of fhe
William
pretorius
Game
Re-
serve.
J.
Sth.
Afr.
Mgmt.
Ass.
B(2),
65_78
(t971\
5.
Bourlidre'
F.: Densities
and
biomass
of
some
ungulate
populations
in
Eastern
congo
and
Rwanda'with
notes
on popurat'ion
structure
and
iion/ungurate
ratios.
zoor.
art.r,
1gg-207
(1e65)
6. Bourlidre,
F.,
Hadley,
II.
J.:
The
ecology
of tropical
savannas.
Ann.
Rev.
Ecol.
and
Systematics
1,
125-b2
(1970)
Bourlidre,
F.,
versehuren,
J.:
rntroduction
d
l'6cologie
des
ongures
du
parc
National
Albert'
Brussels,
rnstitut
des
parcs
Nationaux
du
corigo
lag",
iia pp.
(r9d01
Buechner,
H.
K.,
Buss,
L*O.-,_
L_gn*hurst,
W.
II.,
Brooks,
A.
C.:
Numbers
and
migration
of
elephants
in
Murchison
Fars
Natonal
park,
uganda.
i.
witu.
Mg;t.
22,-eo_58
(1968)
Buechner,
H'
K.,
Golley,
,F.
B^.:^se-conclary
prJuctivity
of terreJriar
ecosystems
(K.
Petruservicz,
ed.), p.
248-2b4.
Oxford:
Stacl;tt
f
C6Z
casebeer,
R,. L.,
Mbai,
H.
T.
M.:
Animal
mortality
rg7}f4
Kajiado
district.
uNDp/nAo
Wildli{e
Ma,nagement
Project,
N?iro!j,
Kenya,
project'!\rorkirrg
Docom"rri
No.
i
(1974)
cassady,
J. T.:
The
effect
o{
rainfall,
soil
toisture
.and
harvesting
intensity
on grass
product'ion
on two
ransela"d,uilgl
in
Kenya.
E.
Afr.
Agric.
for.
J.
86,
zo-s6
(rgz:)
Caughley,
G.: FAO/DF/Zambl68lEr0
Workid
J;;ument
No.
t,
Rome (1923)
Cumming,
D.
H.
Il.:
Unpublished
field
data
uumming,
D.
H.nr.:
A
contribution
to
the
biology
of
warthog
(phacochoerus
afri,canus
Gmelin)
in
the
sengwa
region
of
Rhodesia.
Tro.t"-"s
Nt. Irus.
.i
ilrr.a"ri,
trn
press)
Dasmann,
R'.tr'.,
Mossman,
A.
s.:
commercial
utilisation
of game,"";;;i;
on
a
-Rho-
desian
ranch.
Ilimeogr.
Report,
l1
pp.
(1961)
De
Vos,
A':
rcological
"ottditio.rs
uit".ii"g
tire production
of wild
herbivorous
mammals
on grasslands.
Advances
in^ecological
""J"u."li
(J.
B.
Cragg,
"a.1,".f.
O,
pp.
1BZ_188.
London:
Academic
Press
1g6g
Douglas-Hamilton,
r.:
on the
ecology
and
behaviour
of the
African
elephant.
E.
phil.
Thesis,
University
of
Oxford
(1972)
I
o
10.
11.
12.
13.
14.
15.
16.
17.
352
\l'
J'
Coe
etal'
18.
Do'wset't,
R'. J':
Wet
season
game
populations
and
biomass
in
the
Ngoma
area
of
the
Kafue
Nation"i
purt.
-irr"-f"r.".
b.".
rup""..-o"pt.
Game
and
Fisheries,
zamhia,
4,
,r.
f:;ru.:Jiltl'*.,
"""a*y
of
the
Blesbok
rv*h
special
reference
ro
producrivitv.
\yildl.
ltonograPhs
30,
1-70
(1972)
20.
East
African
U"t"otoiogi"i
Department:
Summary
of
R'ainfall
qata
11962)
21.
East
A{rican
lf"tuotoiolitul
Deiartment':
Sumnary
of
Rainfall
data
(1964)
22.Eltringham,S'K':Changesin-thelarge.mammalcommunityolthelllveyaPeninsula'
riwenzori
National
park,
uganda,
following
removal
of
hippopotamus'
J'
appl'
Ecol'
11'
,r.
i5,:rd:T.
g::ii"",
R..
M.:
The
disrrib'tion-of
rhe
larger
herbivores
in
rhe
Queen
Elizaberh
-"'
Nlii"".f
Park,
Uganda'
J'
appl'
EcoL
?'
273-294
(1970)
--
24.
Foster,
J.
8.,
coe,
li.
i.,
rrr"
Lior]1u*,
or
gu-"
uri-.ls
in
Nairobi
National
Park.
J'
Zool'
Lond.
155,
413-425
(1968)
25.
IIaiston,
N.
G.,
s;l"d;
F'E''
Slobo$lil
t'B':
Community
structure'
population
""
;;;t
"1;"d
competition'
A^""'
Nat'
9t['
42L425
(1960)-
26.
Harris,
L.D.:
An
"t"i"gi*f
description
of
a.semi
arid
East
African
ecosystem'
R'ange
Science
D"p..tm"ni
il:;
sb;i;;
N9'
1-L.
Colorado
State
rJniYersitv
(1972)
27.
Hendrichs,
rr.,
s.rritlo"g""
J"r-H"rti"r-Bio-u**u
in
der
Dornbusch-savanne
ntirdlich
und
westlich
d"t
S";;;;fi:S1"*"
'1t-ott'1tika
nach
einem
neuen
Verfahren
und
Bemer-
kungen
zur
Bio-u..""a;."un;".iir
pftunr"of""*.Lr.d"n
Ti"tarten'
Siiugetierk'
Mitt'
18'
237-
255
(1970)
2E.
Kutilek,
M'J':
The
<lensitV
1nd^l]oT:::'of
large
mammals
in
Lake
Nakuru
National
-"'
i,;.
E.
Afr.
\\
ildl'
J.
12,20r-2r2
(1e74)
.
29.
Lack,
D.:
The
"'t;;i;"1'#;l';;imal
numbe's'
343pp'
oxford:
Clarendon
1954
30.
Lamprey,
H.
1..,
A**fi#?i'lrr"
""rl"gy
of
the
mammal
population
of
a
game
reserve
ln
the
Acacia
*.r"ooul'J*t""g;;yik",
;tlfth
particula*ef^".".tc"
to
animal
n*mbers
and
il;;;;.
b.
Phil.
Thesis,
universitv.of
ox{ord{1962)
-
1
-^^i^-
31.
Lamprey,
H.F''
;;;i;;""i--"pt#igl..oi
th"
large
mamm-al
species
in
the
Tarangire
Game
Reserve,
,r.rffilitr.
n.
at..
\\rildl.
J.
1,
63-92
(1963)
,,-i
.
32.
Lamprey,
II.
F'
t
Eu[itut'ion
of
large
almmal
densities'
biomass
andenergy
cxchange
in
the
Tarangire
cr*;;l;*;.*-rna
ifr"
U**i-Sffi"
in
Tanganyika'
E'
Afr'
Wildl'
J'
2'
,r.
it'f":tg:*Jr.:
Age
criteria
for
the
African
Erephanr
Lorod,onta
'.
afri,cana.
E.
Afr.
wildl'
,-.
i,;*;1;1il1,n$llnn""r-
",
agents
o{
habitat
anct
landscape
cha'ge
in
East'
A{rica'
oikos
21'
,u.
1,.11"1til.0it.,
Parker'
r'c'
S''
Jo!n$o1e,
R'C'D':
Eiephants
and
habitats
in
North
*
il;;;,
Uganda,
n'
'q't'
tt'itat'
J'
8'
163-180
(1970)
36.
Lalvs,
R.
1\I.,
Park;
i'
C'
S''
itft"sione'
R'
C'
D':
Elephants
and
their
habitats'
376
pp'
,r.
?"i;'"1;fl5::*?;Jff:.ri"Ti
E:*,A1iican
mammars'
E'
Arr'
w'dl'
r'
1'
123-124
(1e63)
38.
Leurhold,
llr.,
r,""ih"ia,^s.
li"rn"a"gi""1
;;;;-;i
ungulates
in-Tsavo
(East)
National
Park,
Kenya'
T'"t;;"B;;n
ptj""i'
Kenya
Nationai
Par"ks'
67
pp'
(1973)
3g.
Mentis,
M.
T.,
Esti;;tes
of
natural
biomu*s'"s
o{
Iarge
herbivores
in
the
unfolosi
Game
""
n"*"t""
Area'
Ilammalia
34'
363-393
(1970)
40.
Ilontfort,
A.,
U"*il"*,
biomasses
et
struciurcs
des
populationsd'Ongules
sallvages
au
Parc
de
1'Ak"g";'iR;;;"Jo1'
f'ur"-"t"
et
la
Vie
20'
210'256
(Is72)
41.
Montfort,
A.
,
a""i;il;;;-
deIa
biologie
des
phacocher
e,s_(Pha,cochoerus
aethiopictts)
au
Parc
Nationat'dlTffiC*t
n*u"au'
i["*-uliu
38(2)'
1?7-200
(1974)
aZ.
No*ot'Grif{iths,
M':
PersJnal
communication
(1975)^
'
'---^^^l
43.
Norton-Griffiths,
ii.,
fit"
"""rfrers
and
distribui'ion
of large
mammals
in
R'uaha
National
Park'
'l'anzama'
E'
A'
Wildl'
J'
lS
(2)'
I2L-I40
(1975)
44.
Iroen,
A.
N.
,
Wiiliif;
;;;i;;y,
;"
o"rryd"ui
upp'o"rh.
sur'
{}a,nci1co
: Freeman
1e73
45.
penzhorn,
B.
I.;;;;
urJ
ul"
"o*porition*uniii*.tt.iott.
of
the
sprinebok
(Antidorcas
marsupi,at'i's)p'n*i":i;'t;;i"
ii"""""i"
Z";;N;;;iPark'
J'
st^t''
e1"'
wildl'
Mgmt'
Ass.
4(1),
63-65
(1974)
Biomass
and Production
oi African
lferbivores
358
46.
Fet'rides,
G.
A.,
swank,
\trr.
G.:
population
densities
ancl the
range
carrying
capacity
for
large
mammals
in
Queen
Elizabeth
National
park,
uganda. zool."aft.ltil,
ios_2zs
dsasl
47.
Phiilipson,
J.: Rainfall,
primary
production
and.,carrying
capacity',
of
rsavo
National
Park,
Kenya.
E.
A{r.
\\'ildt.
J,
1g
(B
and,4),
l7t_201 (1'975J
48. Pienaar,
u. de
v.:
observations
on the
developmental
biology,
gtrorvth
and
some
aspects
of
the population
biology
of
the
African
Buffalo
(Syncenrs
iifi"i.
c.17"r
Sparrman)
ii
the
Kruger
National
Park.
Koedoe
lZ, 29_82
(1g69a)
49'
Pienaar,
U.
de
V.:
Predator-prey
relat'ionships
amongsf
the larger
mammals
oJ the
Kruger
National
Park.
Koedoe
Lg,
10g_176
(1969bi
50' Pienaar,
u.dev.,
van\k,
p.,
n'aisail,
N.: An
aerial
censug
of
elephant
and
buffalo
in_
the
Kruger
Nat'ionar
^par\
and
the
implications
thereof
on
inteided
-urrug"*.rrt
schemes.
Koedoe
9,40-102
(1966)
51. Robinette,
w.L.:
Notes
on t'he
biology
of
the
Lechrve (Kodu,s
teeke).
The
puku
occ.
-
lapers
Dept.
Game
and
Fisheries,
N. ilhodesia
g,
g4_112
(196.1)
52' R'obinette,
w. L.,
A19he_r,
A,
L.:
Notes
on ageing
criteria
and reproduction
of
.r.homson,s
Gazelle.
E.
A{r.
Wildl.
J.
g,
SB_98
(1971)
53'
R'osenzrveig,
lI'
L.:
Net primary
productivity
of
terrestrial
communities:
prediction
from
climatological
data.
Amer.
Nat.
102,
6T_24
(1968)
51.
9:.:T' *R;
Live*eights
and
body
measurements
of
serengeti
same
animals.
E.
Afr.
\\rildl.
J.
5,24_36 (1567)
55.
schaller,
G.B.: The_
serengeti
T,ion,
4g0pp.
chicago:
u'iv.
of
chicago
press
1gT2
56.
sinclair,
A.R.E.:
Long
term
monitoring-of
mamtai populations
iir
the
serengeti:
-
99"rls
of non-migratory
ungulates.
E.
Afr.
\Vildl.
J. 4,287:Zg7
(1572)
57.
Sinclair,
A.R,.E.:
Population
increases
of
buffalo
ancl
wildebeeste
in the
Serengeti.
E.
Afr.
IYildl.
J.
11,
93-107
(1923)
58.
Sinciair,
A' R.
E.:
Natural
regulation
of a
buffaio population
in East
A{rica.
E,
Afr.
Wildl,
J.12, 135-154
(r974\
59.
spinage,
c. A.,
Guiness,
tr'.,
Eltringham,
s.
tr{.:
Estimation
of iarge
mammal
numbers
in
the
Akagera
National
Park
ancl }lirtara
hunfing
reser\re,
Rr-anclal.
La
Terre
et la
Vie 4,
567-570
ft972\
60.
stewart,
D.
R,.
M.:
wildlife
census-Lake
rtudolf.
E.
Afr.
\\iildl.
J. l,
121 (1968)
61.
Ste$'art,
D. R.
nI.,
Zaphfuo,
D.
R'.
P.: Biomass
and
density
of
wild
her.bivores
in clifferent
East
African
habitats.
Mammalia
ZZ(4),
489_496
(1968)
62' Thornthv-aite,
C.
W': Average
climatic
s.ater
balance
data
of thc
continents,
Africa.
Lab.
Climatol.
Pub.
Climatol.
15,
11b-128
(1962)
63'
Tinley,
K'
L.:
First
air
count
of
the
buf{alo
of }Iarromou.
Veterin.
lfocamb,
Lourenco
Marques
1,
15b-1?0
(1969)
64.
Tinley,
K.
L.:
Nature
conservation
as a
{orm
of rand
use.
proc.
s-{Rccus
symposium
^_
I"ld
at
Gorongosa
Naiionaj
park,
Ifocambique,
1TB
pp.
(1971)
65. Travassos
santas
Diaz,
J. A.
:
propost
a
pata
a criacao
da
,,Reserva
cle
caca
d.a zinave,,
.
Rev.
Ci6n.
Veterin.
4,
(Ser.
A)
51*-6b
(19t1)
66. Tur'er,M.,\4'atson,R.]I.:Acensusof
gameinNgorongoroerater.E.Afr.wildl,J.2,
165-168 (1964)
67'
\I'alter,
H.:
Le
facteur
eau dans
les rdgions
arides
et
sa signification
pour
l,organisatio'
de
la v6g6tation
dans
les contr6es
sub-tiopicales.
In:
Les
di,isions
ecoliogiques
Ju
monde,
Centre
Nat.
de
la
p"es.
Scient.,
paris,
p.
2Z*ZS
(1,ge+l
68' \Tatson,
Ii'.M.:
The population
ecology
of the
wilclebeeste (Connoclnetes
taurinus
albo-
--
jubatus
Thomas)
in
the
serengeti.
ph.
.b.
Thesis,
carnbridge
university
1rsozl
69'
\Yatson,
R. trf.
: Resu]ts
of
aerial
iivestock
surveyB
of Kapitei
divisionl
S'umbur.r
district
and
North-eastern
Province.
Statistics
division,
Min.
of iinuno"
and
planning,
Il,epublic
_-
of
Kenya,
111 pp.,
85 X,igs.
(1972)
70'
\\ratson,
R''
M.,
Graham,
A.
D.,
parker,
r.
s.
c.: A census
of the
large
mammalsof
Loriondo
controlled
area.
Northern
Tanzania.
E.
Afr.
Wilcll.
J.
1, 4B_b9 (1569)
71.
\Iratson,
R.
M.,
Turner,
{.
r.
t.,
A
co-unt,
of
rarge
mammars
of
the
rake
}ranyara
National
Parh;
results
and
discussion.
E.
Afr.
I4rildl.
J..-g,
9b_98 (196b)
72. western,
D.:
The
structure,
dynamics
and
changes
of the Amboseli
ecosystem.
ph.
D.
Thesis,
University
of
Nairobi
(1973)
4
Oecologia
(Berl.)
M.
J.
Coe
ei
ol.
73.
Western,
D.:
Personal
communication
(1975)
74.
Western,
D.:
The
r.,t"."-"i-tt"Jitional"'primitive"
societies'
I\Ianuscript
in
preparat'ion
75.
Whittaker.
R,.
lL:
Communities
and
ecosystems'
163
pp' London:
Collier-lfacmillan
1970
76.
Wilson,
\'.
J.:
\Yeighis;i;;;;
-'^-ti*
f'or)a
"utti'i"
Zambia'
Arnoldia
(R'hod')
3(32)'
,r.
$":#l$i,r.,
ou,u
from
the
c.Iling
of
krdr.Tragelaphus
strepsiceros
pailas
i.
the
K1'le
"'
N;i;#
Park,
Rhodesia.
ArnoidialRhod')
4(36)'
1-25
(1970)
28.
\\rilson,
v.
J.,
child,
C.
r.
i.,-N.t".'on
bo.h
bo.il
lTragilaphus
scripfas)
from
a
tsetse
fly
control
area
in
x*til""r,
nrroa.uia.
The
puku
occ.
Papers
Dept.
Game
and
Fisheries'
N.
Rhodesia
2,
118-128
(1964)
?9.
\Vvnne
Ed$'ura*,
.l
'
'
iltiltui-d'i'peo*ion
itt
telation
to
social
behaviour'
653
pp' Edinburgh-
London.
Olir-er
and
Bo-vd
1962
Dr.
M.
J.
Coe
Animal
EcologY
Research
GrouP
Department
of
ZoologY
South
Parks
Roacl
Oxford
OX1
3PS,
England
... While all these variables may influence population growth, I thought several would contribute to the occurrence and length of prolonged lags. Considering the environmental conditions that typically reduce reproductive output and mortality in ungulates, I predicted that species with long gestation periods living in areas with more competitors and longer dry seasons would be more likely to exhibit prolonged lags (Coe et al. 1976;Fryxell et al. 1988;Garel et al. 2004). In contrast, I expected species with shorter gestation periods, introduced to areas with no native competitors, and consistent rainfall throughout the year would exhibit earlier, faster population growth. ...
... The availability of water greatly influences habitat selection in many mammals (Coe et al. 1976;Smit et al. 2007;Longshore et al. 2008), especially those that live in drier environments (Noy Meir 1973;Bleich et al. 2010;Letnic et al. 2015). Bleich et al. (2010) found that increasing the number of artificial water points could increase the area of suitable habitat for mountain sheep (Ovis canadensis) by up to 92%. ...
... As canopy cover can influence most abiotic and biotic factors that affect how ungulates use habitat (Elton 1939;Mysterud & Ostbye 1999), it is likely to contribute to habitat selection. Tracks and roads are frequently avoided by deer, as these anthropogenic features often represent areas of high disturbance (Rost & Bailey 1979;Sibbald et al. 2011;Scholten et al. 2018 (Caughley 1970;Coe et al. 1976;Fryxell 1987;Mduma et al. 1999). ...
Thesis
Full-text available
Invasive species are a leading cause of biodiversity loss worldwide. Deer have been introduced to environments around the world, and many species have gone on to become invasive. Feral deer potentially compete with native species and livestock, pose risks to vehicles or by acting as vectors of disease, as well as contributing to economic and social losses. Presently there are six free-living deer species in Australia: chital (Axis axis), fallow (Dama dama), hog (Axis porcinus), red (Cervus elaphus), rusa (Rusa timorensis), and sambar (Rusa unicolor). Four chital deer were liberated on Maryvale Station in Northern Queensland in 1886 and since then, the number and range of chital deer has slowly increased. The factors that contributed to both the delayed growth as well as their sudden range and population increase are not known, such as how they select and use habitat as well as the cues that drive their reproduction. Considering the broad economic and environmental impact chital could have on this region, understanding their ecology is critical to developing more effective management and control strategies, as well as predicting where these feral species are likely to spread next.
... Climatic conditions impact large herbivores via multiple pathways. Because precipitation and temperature govern primary production, and large herbivores have substantial forage and water intake requirements, climate indirectly affects herbivore fitness, production (Coe et al., 1976;Raynor et al., 2020) and diversity (Olff et al., 2002;Veldhuis et al., 2019). Climate also acts as a direct control on individuals and populations by influencing physiological functioning and behavior. ...
Article
Link for free access to paper: https://authors.elsevier.com/c/1ewHq-4PS2FI8 As the dominant large herbivore in midcontinent North America since the terminal Pleistocene, bison (Bison spp.) have been a fundamental component of ecosystems and economies. Despite the importance of bison in late Quaternary North America, large-scale (regional to continental) patterns of bison biogeography are not well understood. Here we integrate archaeological and paleontological bison occurrence data with simulated climate data to better understand long-term drivers of bison distribution and abundance in North America. We used these records to model bison distribution and abundance over the past 20 thousand years at 1-thousand-year intervals. Our results show that late Quaternary changes in the distribution and abundance of bison were influenced by large-scale trends in temperature and precipitation. The distribution of bison since the Bølling–Allerød Interstadial (ca. 14 ka) is primarily explained by seasonal temperature patterns (mean temperature of the coldest quarter is the most important variable for 12 of the 14 1-thousand-year intervals). The modeled climate of bison distributions progressively narrowed since the Last Glacial Maximum (ca. 20 ka) as bison populations retracted from disjunct Pleistocene refugia and congregated in midcontinent rangelands. Through the Pleistocene-Holocene transition, bison experienced rapidly warming summer temperatures that increased faster in midcontinent North America than other regions and the continent as a whole. Model results suggest that Holocene bison abundance was influenced by hydroclimatic shifts that affected the quality and availability of forage. Bison abundances decreased through the dry early and mid-Holocene and increased when moisture availability improved in the late Holocene. We infer that bison have thrived under a broad range of environmental conditions since the Last Glacial Maximum and that the climatic and biogeographic space occupied by bison narrowed in recent millennia.
... Globally, as evidenced by the unprecedented fire events of recent years, including those in Cyprus (July 2021), western Canada/USA (June/July 2021), California (August 2020) and Australia (2019-2020), Greece (2018), Spain and Portugal (2017), and France (2016), climate change is not only bringing heatwaves and drought, but also catastrophic wildfires, the latter driven by the hot and drying environment, as well as the direct correlation between rainfall and primary biomass (Coe et al. 1976). In Africa, human modification of ecosystems appears to have put forest and savanna extents out of synch with its wetting and drying cycles (Gee 1998;Archibald et al. 2012). ...
Article
Full-text available
Across much of southern Africa’s savanna wildernesses, wildfires burn unchecked. This is particularly true in the woodland savannas of northern Botswana, where wildfires originate outside of management activities, and are left to burn uncontrolled, because of limited resources and remoteness. There is concern that severe annual wildfires are resulting in the ‘savannisation’ of large tracts of wooded land across forest reserves, protected areas and surrounding wilderness areas. Because the current fire regime is unknown, management interventions are hard to introduce. We examine the recent 20-year (2001–2020) fire history in northern Botswana using MODIS satellitederived fire products to reveal fire frequency and seasonality. Six wildfire hotspots are identified for exploration of fire frequency and possible origins. Annual fire frequencies are far higher than would be expected without anthropogenic ignition. Extensive areas in some hotspots are shown to have burned between 14 and 16 out of the 20-year period. Fires peak in September, several weeks before the onset of the rainy season and associated lightning strikes, and when the fuel load is at its maximum and conditions at their driest. Adaptive fire management practices, such as those being followed in neighbouring South Africa and elsewhere should inform Botswana’s fire management policies.
... Areas covered by each of the classified land cover classes were estimated based on pixel representation and are presented in Figure 5. Rainfall plays a significant role in the ecosystem and has been positively related to the primary productivity of biomass in the savanna system (Phillipson, 1975;Coe et al., 1976;East, 1984). Therefore, a Linear Pearson Correlation (LPC) was computed (Lenhard and Lenhard, 2014) to assess the relationship between the area covered by V. stuhlmannii and rainfall between 2013 and 2017. ...
... The mean productivity of Nubian Ibex in our study (0.81 neonate:female) was well below the species potential (Nowak, 1991;Mungall & Sheffield, 1994;Loison et al., 2002). Lower productivity than seen in Alpine Ibex (Capra ibex) was likely due to limited forage characteristic of arid environments as found in other species (Coe et al., 1976;Sowell, 2001). Productivity ratios are frequently used as a management metric because they allow determination of the maximum total adult mortality that a population can sustain without declining, and high production of juveniles is often an indicator of good individual and population health (Gaillard et al., 2000;Eberhardt, 2002;Bender, 2006Bender, , 2019. ...
Article
Full-text available
The population status of the IUCN vulnerable Nubian Ibex Capra nubiana is largely unknown in the Middle East because of difficulties conducting intensive surveys. We tested the utility of camera-trapping at permanent water sites to estimate demographics and population trend of the species in the Dana Biosphere Reserve, Jordan. We also studied the relationship between precipitation and the Normalized Difference Vegetation Index NDVI on water site use by the Nubian Ibex, demographics, and population trend. The Nubian Ibex showed a mean annual productivity of 0.81 neonate/female and an adult female survival of 0.87 which is a lower survival rate than found in other ibex species. The mean potential annual rate of increase of this population (λ=1.14) was similar to those reported from other ibex species and indicated a viable population. The capture rate of camera traps (number of photographed animals/camera day) was found not to be a reliable index to monitor population trends because water site use was related to environmental conditions that could not be standardized or predicted a priori. Quantifying population demographic ratios using camera trapping allowed for monitoring the status of key population parameters and growth potential of Nubian Ibex populations in the Dana Reserve, and likely can be used to monitor population status and facilitate informed management of similar populations in the Middle East.
... Land managers have used prescribed burning in African savannas to prevent catastrophic fires that negatively impact vegetation (e.g., riparian forests) and infrastructure, and to create benefits for wildlife and livestock, including improving forage quality, controlling bush encroachment, reducing tick-borne diseases in livestock, and attracting higher densities of grazers. 77 More productive African savannas (based on rainfall and soil nutrients) support greater herbivore biomass, which in turn supports greater carnivore biomass, [78][79][80] and lion density in PAs is closely related to prey biomass. 64 Furthermore, EDS fires could improve lion fitness by leaving taller grass in the LDS, as lions hunt more frequently and with greater success with increased cover. ...
Article
Full-text available
Carbon emissions from savanna burning contribute to global climate change. Improved fire management in Africa could dramatically reduce carbon emissions and build ecosystem resilience, reduce threats to biodiversity and provide much needed financial support to local economies. Potential carbon revenues could substantially reduce protected area funding gaps that are in crisis due to COVID-19 and diversify income to augment tourism. More funding to pilot projects is needed to realize this potential and accelerate the UN’s Decade of Ecological Restoration.
... 1.5-2 (Coe et al., 1976;East, 1984;Fritz & Duncan, 1994). ...
Article
Full-text available
Large herbivores provide key ecosystem processes, but have experienced massive historical losses and are under intense pressure, leaving current ecosystems with dramatically simplified faunas relative to the long‐term evolutionary norm. Hampered by a shifting baseline, natural levels of large‐herbivore biomass are poorly understood and seldom targeted. This ‘Decade of ecosystem restoration’ calls for evidence‐based targets for restoring the natural diversity and biomass of large herbivores. We apply the scaling of the consumer–producer relationship to a global dataset of large‐herbivore density in natural areas. The analyses reveal that African ecosystems generally have much higher large‐herbivore biomass and also the strongest consumer–producer relationship. For Europe, Asia and South America, there are no significant relationships with primary productivity indicative of impoverished faunas. Compared to expectations from the African scaling relation, large‐herbivore biomass in ecosystems outside Africa is considerably lower than expected. Synthesis and applications. Ecological restoration and rewilding entail restoration of a natural grazing process. Our findings indicate that many nature reserves are depleted in large‐herbivore biomass, judged from their primary productivity. Meanwhile, overexploitation by seasonal livestock grazing takes place in other areas. It is thus difficult, but urgent, to reach scientific consensus regarding a natural baseline for large‐herbivore biomass. Until such agreement has been reached, we recommend to manage, or rewild, large herbivores in year‐round near‐natural grazing and without predefined density targets, but following natural and fluctuating resource availability with minimal management intervention. The establishment of experimental rewilding sites with reactive herbivore management is needed to further advance our understanding of natural grazing density. Ecological restoration and rewilding entail restoration of a natural grazing process. Our findings indicate that many nature reserves are depleted in large‐herbivore biomass, judged from their primary productivity. Meanwhile, overexploitation by seasonal livestock grazing takes place in other areas. It is thus difficult, but urgent, to reach scientific consensus regarding a natural baseline for large‐herbivore biomass. Until such agreement has been reached, we recommend to manage, or rewild, large herbivores in year‐round near‐natural grazing and without predefined density targets, but following natural and fluctuating resource availability with minimal management intervention. The establishment of experimental rewilding sites with reactive herbivore management is needed to further advance our understanding of natural grazing density.
... The lack of moisture limits plant growth, resulting in little biomass of relatively high nutrient concentrations (Fornara and du Toit, 2008a). On the other hand, where mean annual rainfall > 650 mm nutrients are depleted before plant growth ceases, leading to comparatively large plant biomass of low nutrient concentrations (Coe et al., 1976;Huntley, 1982;Teague and Smit, 1992;Skarpe, 1995). Moist-dystrophic soils are highly leached while arid-eutrophic soils are subjected to little or no leaching (Huntley, 1982;Verweij et al., 2006;Sankaran and Anderson, 2009). ...
Article
Full-text available
Impala are medium-sized antelopes, widely distributed throughout East and Southern Africa. They consume grasses in the wet season, switching to browsing in the dry season. African savannas are classified into arid-eutrophic and moist-dystrophic. The arid-eutrophic savannas have fertile soils which support plant biomass of high nutritive value, utilised by small and medium-sized herbivores. The moist-dystrophic savannas have nutrient-poor soils and high plant biomass of relatively low nutritive values, utilised by large-bodied herbivores. Impala occur in both savannas, though more abundant in the arid-eutrophic savannas. I studied browsing behaviour of impala in Mikumi National Park, a moist-dystrophic savanna, and Serengeti National Park, an arid-eutrophic savanna. A car was driven in all access roads in the Parks during the wet and the dry season. The car was stopped for recording whenever impala were sighted browsing within 50 m from the road. I recorded tree species browsed, browsing height, number of bites, and habitat type. The study showed differences between the seasons and the savannas. The animals took more bites during the dry season compared to the wet. Biting rate was higher during the dry season and when spineless trees were utilised. Small and large leaves reduced biting rate while medium size leaves promoted biting rate. The impala frequently browsed at head level in both savannas, however variations between the savannas were observed during the wet season. The differences in forage quality and quantity between the savannas and the seasons, and in herbivore community between the savannas are likely to attribute the results.
... These features, in addition to the hindgut fermentation, allow elephants to consume large quantities of low-quality forage (Bell 1971;Jarman 1974). Thus, elephants dominate savannas with abundant but lowquality plant biomass such as the moist-dystrophic savannas (Coe et al. 1976;Bell 1982;Huntley 1982;East 1984). The nutrient-rich or arid-eutrophic savannas comprise relatively little plant biomass as a result of intense herbivory but are generally high in nutrient concentrations (Huntley 1982). ...
Article
Full-text available
Sexual differences in foraging behaviour of African elephant are comparatively well studied. However, the behaviour is known to vary between savannas due to differences in plant and herbivory communities which determine the level of forage selection and competitions. The sexual variation in foraging behaviour is attributed by sexual dimorphism, where male elephants are larger than females. The different body sizes result in divergent foraging strategies due to the differential scaling coefficients of metabolism and gut size, where the larger bodied individuals can tolerate greater quantities of low-quality forage. We studied the foraging behaviour of male and female elephants in Serengeti National Park, a nutrient-rich savanna in Tanzania, based on the forage selection and the scramble competition hypotheses. Data were collected for 2 months in 2014 by driving in road transects and make visual observations of browsing elephants within 50 m from the road. We recorded the number of bites taken per tree, parts of the tree browsed, browsing height, and tree species. Females targeted a wider range of wood species, browsing at lower heights compared to males. The sexes had a comparable length of browsing bouts, spending an average of 4 min per tree. The results support the scramble competition hypothesis through the feeding height stratification and the forage selection hypothesis through the higher diversity of tree species utilised by females. We suggest that sexual differences in forage requirements interact with forage quality and competing herbivore communities in shaping the foraging behaviour of elephants.
Article
Full-text available
The time period between 560 and 360 ka (MIS14 to MIS11) was critical for the evolution of the Neanderthal lineage and the appearance of Levallois technology in Europe. The shifts in the distribution of the human populations, driven by cyclical climate changes, are generally accepted to have played major roles in both processes. We used a dataset of palaeoclimate maps and a species distribution model to reconstruct the changes in the area of Western Europe with suitable environmental conditions for humans during 11 time intervals of the MIS14 to MIS 11 period. Eventually, the maximum sustainable human population within the suitable area during each time interval was estimated by extrapolating the relationship observed between recent hunter-gatherer population density and net primary productivity and applying it to the past. Contrary to common assumptions, our results showed the three Mediterranean Peninsulas were not the only region suitable for humans during the glacial periods. The estimated total sustainable population of Western Europe from MIS14 to MIS11 oscillated between 13,000 and 25,000 individuals. These results offer a new theoretical scenario to develop models and hypotheses to explain cultural and biological evolution during the Middle Pleistocene in Western Europe.