Lessons for ecology, conservation and society from the SerengetiAnthony R.E. Sinclair

Centre for Biodiversity ResearchUniversity of British Columbia, Canada

Dr. Simon MdumaHead of TAWIRI & the SerengetiBiodiversity Progam

The Macaulay InstituteFrankfurt Zoological SocietyNSERC, CanadaNational Geographic SocietyWildlife Conservation Society

THANK YOU

THE TEAM

Simon MdumaStephen MakachaAlly NkwabiJohn MchettoAnne SinclairJohn FryxellRay HilbornRoy TurkingtonKris Metzgerand indirectly

Markus BornerSimon Thirgoodand many others

Species are being lost at unprecedented rates.

How much can we lose before the ecosystems in which we live become unsustainable?

Species exist within a matrix of other speciesand are subject to the effects of their environment

We must understand how the whole system behaves-lack of understanding could lead to surprise, inappropriate management, or even system collapse

THE WORLD PROBLEM

- therefore, one must understand important properties of ecosystems to apply

- effective management

- effective conservation

THE NEED

Changes in Ecosystems can lead to unexpected outcomes

Flathead Lake, Montana

Opossum shrimp added to provide extra food forKokanee SalmonBald eagles

0

20

40

60

80

100

120

140

160

1978 1980 1982 1984 1986 1988 1990

Abu

ndan

ce In

dex

Bald Eagles x 7Kokanee x 1000Opossum shrimp

Spencer et al. 1991 BioScience

Flathead lake, Montana

Unexpected events occur because ofcomplex interactions.

Illustrated by events in Serengeti, Tanzania

Rift

val

ley

Rift

val

ley

Isuria

Scarp

Ol DoinyoLengai

Meru4565

Olduvai

gorge

Ketumbaine2942

Gelai2962

The Serengeti Ecosystem

Salaiplains

Serengetiplains

Loitaplains KENYA

TANZANIA

From Grant Hopcraft

SERENGETI IS COMPOSED OF TREELESS PLAINS

..AND SAVANNA WOODLANDS

Wildebeest is the dominant species

Wildebeest migration patterns

TANZANIA KENYA

40 0 40 80 Kilometers

Mean Annual Precipitation

Mean Annual Precipitation (cm)36 - 4646 - 5555 - 6464 - 7373 - 8383 - 9292 - 101101 - 111111 - 120No Data

SNP Boundary

Kris Metzger

The reasonfor dry seasonmigration

1200

400

Serengeti Wildebeest Population outbreak –the event that changed everything – to understand it we must go back a century

0

300

600

900

1200

1500

1800

1950 1960 1970 1980 1990 2000

Popu

latio

n Si

ze (x

100

0)

The Great RinderpestThe Great Rinderpest

• The Epizootic of 1889 • Ethiopia to Cape by 1896• Die-off of cattle and other ruminants 95%• These include African buffalo• Wildebeest yearling disease present up to

1963• THIS IS THE PERTURBATION THAT UNDERLIES

OUR UNDERSTANDING OF THIS SYSTEM

Rinderpest

0

20

40

60

80

100

1955 1960 1965 1970

% w

ith A

ntib

odie

s

Years

Wildebeest

Buffalo

Incidence of Rinderpest in Serengeti

Complex events that followed the wildebeest increase

- Grass fires: between 1920s and 1960s 80-90% of savanna was burnt each dry season

Grass fires prevent tree regeneration below 2 m height

Repeated burning prevents regeneration and producesa distorted age structure of old trees

% Acacia tree canopy cover drops rapidly in the 1960s

05

1015202530

1940 1960 1980 2000Year

% A

caci

a C

anop

y

H.Dublin 1986

Northern Serengeti, Mara triangle 1944

Photo Syd Downey

1983

Wildebeest grazing reduces grass fuel and area burnt

0102030405060708090

100

1960 1970 1980 1990 2000

% A

rea

Bur

ntSERENGETI AREA BURNT IN DRY SEASON

0

20

40

60

80

100

0 500 1000 1500

Wildebeest number (thousands)

% a

rea

burn

t

Increase in wildebeest causes decreasein burning

Complex interactions of wildebeest and the environment

Complex interactions of wildebeest and the environment

The extent of grass fires is determined by thedegree of grazing imposed by wildebeest

….this had consequences on savanna trees

1980 1986

20031991

Savanna

Martin and Osa Johnson 1928

Osa Johnson filming from plane 1933

NYARABORO FROM EMAKAT

1928 1982

2003

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

1920 1930 1940 1950 1960 1970 1980 1990 2000 2010

rate

of c

hang

e in

tre

es

Instantaneous rate of change in tree density negative 1920s-1960s, then increases rapidlyin 1980s, 1990s following wildebeest

SERENGETI TREE DENSITY

Lions use thickets to capture prey- more thicket improves success

Lions use thickets to capture prey- more thicket improves success

0

50

100

150

200

250

300

350

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Lion

num

ber

0

200

400

600

800

1000

1200

1400

1600lionwildebeest

Serengeti woodland lion numbers Serengeti woodland lion numbers increase in the 1990sincrease in the 1990s

Wildebeest x 1000

Wildebeest x 1000

Lion data from C. PackerLion data from C. Packer

0

10,000

20,000

30,000

40,000

50,000

1970 1975 1980 1985 1990 1995 2000 2005

Impa

la n

umbe

r

0

2,000

4,000

6,000

8,000

10,000

12,000Topi num

ber

impalatopi

Decrease in most resident prey

Disturbance and multiple statesDisturbance and multiple states

Ecosystems are continually being disturbed

weather eventshuman harvestinginvasions of speciespredation

Disturbance can cause a change of state

Disturbance changes ecosystems – which do notalways return to the original state afterwards

The role of elephants in Serengeti

Elephants knock down mature trees andblamed for tree decline

Trees and elephant predationTrees and elephant predation

1970s – fire rather than elephant shown to be the cause of decline (Norton-Griffiths work in 1970s)

Elephant play another role by feeding on seedlings

Dublin observed elephants removing almost all seedlings- They hold the system in a grassland state

Removal of elephants in Serengeti by ivorypoachers in 1980s but not in Mara…..

0

500

1000

1500

2000

2500

3000

3500

1950 1960 1970 1980 1990 2000 2010

Num

ber

POACHING

SERENGETI ELEPHANT

IVORY BAN

..resulting in much food for elephants in the 1990s and 2000s inSerengeti but not the Mara (wildebeest in both)

KENYA

TANZANIA

SERENGETI - MARA 2005

-hence ecosystems can have more than one state under the same conditions

Tibetan yak

Tibetan grasslands near Naqu

0

2

4

6

8

10

12

0 2000 4000 6000 8000 10000Livestock equivalents

Num

ber o

f out

brea

ksMONGOLIA – Brandt’s vole outbreaks

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 20050

10000

20000

30000

40000

50000

60000

70000

80000

Wahine Storm, 1968

Num

ber o

f sw

ans

CHANGE IN STATE: swans on Lake Ellesmere, New Zealand

Disturbance and Ecosystem Processes

What goes wrong if we ignore ecosystems? Their processes for ecosystem functioningFor example

• Hydrology, flux and storage

• Biodiversity and Stability (resilience)

Saline upwelling in Western Australia

Biodiversity and stabilityBiodiversity and stability

• species diversity and community resilience

e.g. Australian Eucalypt woodlands

Original closedeucalyptwoodland

Degraded eucalypt woodland

Drawings Frank Knight

Noisy miners reduce or exclude…..

…white-plumed honeyeaters and otherspecies in degraded Eucalypt woodland

Loss of bird diversity in Australian woodland

Psyllid outbreaksin isolated treesof farmland

Eucalypt dieback – eastern Australia

Psyllid insect outbreaks cause dieback in exposed trees-disruption of intact forest causes biodiversity loss andecosystem disfunction

Long time scales - The effects of historyLong time scales - The effects of history

Disturbance events take a long timeTo understand ecosystems we must understandthe long-term events. History is important

Rinderpest – the ecological event in 1889 that changed the course of human history in Africa.It decimated human populations and allowed the colonization of Africa

The ecological effects are still seen in Serengeti today

0

10000

20000

30000

40000

50000

60000

70000

80000

1890 1910 1930 1950 1970 1990 2010

Num

ber

ObservedReported

The GreatRinderpestepizootic

removed

Buffalo long-term change 1890-2003

Poaching

Data from Sinclair & Mduma

0

1

2

3

4

5

6

1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980

Log

Freq

uenc

y The great rinderpest

Acacia tree density 1880-1980

Disturbance and history

THE IVORYTRADE

1840s – 1890s

TIPPU TIPThe great slavetrader of easternAfrica

- 1860s to 1890s

Ivory from Zanzibar

0

50

100

150

200

250

300

350

1840 1850 1860 1870 1880 1890 1900kg

x 1

000

Ivory from Khartoum

0

50

100

150

200

250

1840 1860 1880 1900 1920

kg x

100

0EAST AFRICAN IVORY EXPORTS

Data from C. Spinage 1973

0

500

1000

1500

2000

2500

3000

3500

1850 1900 1950 2000

Num

ber

countsreports

The Ivory trade

Poaching

SERENGETI ELEPHANT

CITESIvory ban

The effects of historyThe effects of history

Ivory hunting – affected the vegetation in Protected areas such as Tsavo (Kenya) andChobe (Botswana), and altered the long-termcourse of conservation in Africa.

It also decimated human populations and prevented human advancement

Disturbance and historyDisturbance and history

Therefore, - ecosystems are always changing,- they do not return to the same state- long-term consequences

- e.g. Scotland- e.g. New Zealand

LESSONS FOR SOCIETY: THE FUNCTION OF MIGRATION

Migration allows use of ephemeral resources on the plains and so larger populations than if they were resident. This rule applies to all migrations

SAHEL

Traditional Humanmigrations in the Sahel follow rain

Boreholes have been sunk in the Sahel since the 1960s topresent. Sedentary life style has resulted in overgrazing

Photo A. de Vos 1975

.. and repeated famines. MALI FAMINE 1973

MALI –

TheGreen Polygon1973demonstratesovergrazingnot drought

The ranchboundary

The role of Protected Areas

- The Green Polygon illustrates the need for baselines

- Protected Areas act as such ecological baselines to provide insight for human ecosystems

- BUT Ecosystems are continually evolving and do notreturn to where they began

- Protected areas will not stay as they are currently

Take home messageConservation has to focus on ecosystems

- they are complex, have long time scales and multiple states

- they are subject to disturbance which can change state

- human disturbance can be monitored by reference toProtected Areas

- ecosystems are continually changing

Current protection strategy is not addressing this issue- we need to find a new way to accommodate change