Food security, co-management and the

ecosystem approach to fisheries

African Great lakes Conference,

Entebbe 2-5 may

Jeppe Kolding (University of Bergen, Norway)

Paul van Zwieten (Wageningen University, The Netherlands

Michael Plank (University of Canterbury, New Zealand

Richard Law (University of York, UK)

Hans D. Gerritsen (Marine Institute, Ireland)

David Reid (Marine Institute, Ireland)

The big challenge

With a growing

population: How do we

reconcile food and

conservation?

Which fishing pattern gives the highest yield

and least structural impact on the community ?

Johannesburg 2002 Declaration § 31 (a):

«Stocks should be kept at biomass levels that can produce maximum sustainable yields (MSY).»

CBD Malawi principles for Ecosystem Approach:

«A key feature of the ecosystem approach includes conservation of ecosystem structure and functioning»

The aquatic food web is size structured…

..abundance is inversely correlated with size

Primary producers

Primary consumers

Secondary consumers

Tertiary consumers

Top predators

Log B

iom

ass

Size

Log B

iom

ass

Size

Community size spectrum

The distribution of biomass by body size follows regular patterns

phytoplankton

zooplankton

Small fish

Large fish

Under conventional selective fishing slope and

intercept will change

Changes in the North Sea

Log

Bio

mas

s

Size

Unfished

Garcia et al. 2012

1983–1987

1998–2002

Bio

mass

Size

Balanced harvesting… (Garcia et al 2012)

.. is fishing as many sizes and species as possible in

proportion to natural productivity

It will reconcile objectives

by maintaining community

structure while returning

highest yields

Lake Kariba

Zambezi River

Lake Kariba

Regulated and enforced

Yield = 1000 tonnes yr-1

Non regulated and non enforced

Yield = 6000 tonnes yr-1

Lake Kariba

Average mesh size (mm stretched), Zambia

R2 = 0.86

60

80

100

120

140

160

180

1960 1965 1970 1975 1980 1985 1990 1995

Zambia - Catch rates

0

1

2

3

4

5

6

7

8

9

10

1959 1964 1969 1974 1979 1984 1989 1994 1999

CP

UE

Artisanal CPUE (Kg/net) from CAS surveys

Experimental CPUE (Kg/45 m net set)

Artisanal CPUE (Kg/net/night) from Scholtz 1993

Zimbabwe - Catch rates

0

2

4

6

8

10

12

14

1962 1967 1972 1977 1982 1987 1992 1997

CP

UE

Artisanal CPUE (kg/45m net) from CAS surveys

Experimental CPUE (kg/45m)

Linear (Artisanal CPUE (kg/45m net) from CASsurveys)

Controlled effort Open access

Kolding et al. (2003)

0

5

10

15

20

25

30

35

40

45

2" 2.5" 3'' 3.5'' 4'' 4.5'' 5'' 5.5'' 6'' 6.5'' 7'' 7.5"

Rel

ativ

e fr

equ

ency

(%

)

Mesh size (inches)

Zambia

Zimbabwe

Mesh size distributions

Illegal

Mesh size distributions

and catch rates (Zambia)

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50 63 76 89 102 114 127 140 152 165 178

Rel

ativ

e no

of n

ets

and

CPU

E

Mesh size (mm)

CPUE

Mesh size

R² = 0.9036

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

4.00 6.00 8.00 10.00 12.00 14.00

Relation between CPUE and number of nets in each mesh size (76-178 mm)

Comparison between unfished and heavily fished areas in Lake Kariba

Comparison between unfished and heavily fished areas in Lake Kariba

Kolding et al. 2015

Comparison between unfished and heavily fished areas in Lake Kariba

Number of species has not

changed and slopes are parallel

meaning ecosystem structure is

maintained while yields are 6x

higher than under regulations

Kolding et al. 2015

Bangweulu swamps Northern Zambia

Carl Huchzermeyer

Bangweulu swamps

• Fish are getting smaller

• Rampant use of illegal gears

Bangweulu swamps Northern Zambia

Mesh size Mesh size

(mm) gillnets kusikila seines weir traps cum %

3 3,869 13

4 8,358 41

6 2,322 49

8 387 50

10

25 534 17 53 52

38 6,719 68 178 75

50 4,233 135 49 90

63 1,260 643 97

76 554 74 99

89 136 - 99

102 - -

114 -

127 255 100

140 -

Total: 13,691 937 280 14,936 29,844

% 46 3 1 50 100

% legal 22

Total number of gear by type

Ndang

a

Species and size composition by gear

Fishing pressure versus size

Large fish

in gillnets

SeinesCiclids

Small fish

illegal

legal

Only largest species in legal

gillnets are overexploited

Overfished

two multispecies fisheries

Celtic Sea (EU)

major commercial fishery

100,000 to 150,000 tonnes yr-1

>1000 vessels

quotas, minimum landing sizes

demersal, ~15 spp, trawls

small-scale artisanal fishery

~15,000 tonnes yr-1

~5000 fishers

largely unregulated

>30 spp, gillnets, seines, traps

Bangweulu Swamps (Zambia)

size distributions of yield

biomass yields as a function over body mass

(aggregated over species)

100 102 104

101

weight (g)

103

105

Celtic Sea

yield (tonnes)

101

103

Bangweulu

yield (kg)

Celtic-Sea landings

Celtic-Sea landings+discards

Bangweulu

landed catch

size distributions of yield

Plank et al. 2016

size distributions of catch

smallest fish:

~250 g

~10 g

size limits not operating in Bangweulu

100 102 104

weight (g)

101101

103

103

105

Celtic Sea

(tonnes)

Bangweulu

(kg)

range of body sizes

~20-fold

~1000-fold

largest fish

~5 kg (fish caught when smaller)

~10 kg (upper limits to growth)

both fisheries are sustainable

how is this possible?

model of an unregulated fishery:

fishers share an aquatic ecosystem (commons)

each fisher decides what size of fish to catch

what happens to the fish stock and catch?

how does fishing mortality F(x), aggregated over fishers, get distributed over over fish body size x?

bac Kwww awbw

bookkeeping of biomass

Dynamic size-spectrum model a different ecological model:

+

stochastic event

Model overview (Law et al. 2015)

dwwUwGwE1

MUGUw

Et

U

1. The fish population

2. The fishers

• know nothing about ecology or size-spectrum dynamics

• do know their own catch, and the catch obtained by their neighbours – and their gears

From time to time he will evaluate his own catch (and gears) compared to those of his neighbours. If he catches less, he will tend to shift gear

rule for each individual fisher

fishing effort is area under curve

each fisher i chooses a net mesh centred on an arbitrary fish size

body weight x (log scaled)

Fi(x

) b

y fi

sher

update net mesh from time to time

more likely to move if his yield (cpue) is small compared to others: 1 – Yi/max(Yj).

new net mesh size chosen uniformly at random on range of body sizes

body weight x (log scaled)

Fi(x

)b

y fi

sher

random shift

Fi(x

)b

y fi

sher

aggregate fishing mortality F(x)

sums over all fishers (simple SEES model)

body weight x (log scaled)

few fishers (100)

110-2 102

weight (g)

1

10-2

10-4

bio

mas

s (g

m-3

)

unfished

fished

110-2 102

weight (g)

pro

du

ctiv

ity

(g m

-3 y

r-1)

0

0.2

0.4

0.4

productivityfishing mortality

0.6

0.2

fish

ing

mo

rtal

ity

rate

(yr

-1)

emergent

F(x) aggregated close to biomass peak

many fishers (8000)

110-2 102

weight (g)

110-2 102

weight (g)

4

8

0

0.2

0.4

fish

ing

mo

rtal

ity

rate

(yr

-1)

fishers remove biomass peak: flat ‘Sheldon’ biomass spectrum

(constant biomass on log-scaled intervals of body mass)

F(x) spreads out over body size x on RHS of productivity peak

F(x) ~proportional to productivity

110-2 102

weight (g)

bio

mas

s (g

m-3

)

10-4

1

10-2

biomass spectrum (unfished)

F(x)aggregated

Result = emergence of balanced harvesting

F(x) proportional to productivity balanced harvesting

a fisher cannot increase yield by a change in gear

constant biomass all fishers get equal biomass yield

Nash equilibrium: No fisher has an incentive to change fishing strategy, given the strategies of all other fishers

aggregate biomass yield

total fishing effort must still be controlled

caveats:

biomass yield and profit are very different things

balanced harvesting

without min. landing

weight regulation‘Bangweulu-like’

= collapse of stock and fishery

BH gives

much greater

biomass yield100 g min. landing

weight regulation‘Celtic-Sea-like’

Conclusions

Mesh size regulations in small-scale fisheries are

impeding maximum yield and healthy resilient

ecosystems

When food security (biomass) is more important

than commercial value, then catch-rates (CPUE)

will regulate the fishing pattern towards a

Balanced harvest regime.

For co-management to work the State will have to

abandon size and gear regulations.

Conclusions:

A trait-based size spectrum model is able to quite

accurately simulate the Kariba fishery

The model confirms that starting fishing on

smaller individuals and fishing ‘balanced’

according to size will:

Increase total yields

Minimize impact on fish community

By fishing ‘illegally’ the Zambian fishers are able

to catch 6x more without causing deep structural

changes to the fish community

Thank you for your attention