Community Ecology Processes, Models, and Applications

referred to as hysteresis (Fig. 8.8a). Scheffer and
Carpenter (2003) described a series of criteria for
identifying regime shifts, none of which alone diag-
noses a regime shift but which, together, constitute
strong evidence. In field data these include abrupt
shifts in time series, a bi- or multimodal frequency
distribution of states in a time series, and dual (or
multiple) relationships between ecosystem state
and a forcing variable. Experimental evidence in-
cludes dependence of final state on initial state (e.g.

order of colonization during succession), shift to-

wards a distinctly different stable state after a pulse

perturbation and hysteresis, i.e. change of the eco-

system along different pathways when the forcing

variable is increased compared with when it is de-

creased. Two important questions for applied ecol-

ogy are whether rapid shifts between relatively

long-lasting states of an ecosystem can be forced

by gradual changes in conditions, and whether

these shifts are reversible.

1940

400

Abundance^300

internal Structure

Thousand tonnes

200

100

0

1960

0.0 0.2

80

Smooth

External Forcing

Abrupt

Discontinuous^60

40

20

0

0

–0.5

–1

–1.5

–1 0 1 2

96

Gelatinous plankton

Planktivorous fish

Zooplankton

Zooplankton (^9893) Phytoplankton
89
2000
83
80 2
1.5
1
0.5
0
–0.5
–1
–1.5 –1 –0.5 0
83
81
2000
79
89
91
99
97
94
0.5
78
75
85
85
82 88
91
(^9490)
77
(^200097)
2
1.5
1
0.5
0
–0.5
–1
–1012
80
60
40
20
0
0.4
Fishing mortality
Catch (kt) Catch (kt)
Fishing mortality
Fishing mortality
f g h
d e
a b c
0.6
1931–1965
1966–2000
0.0 0.2 0.4 0.6
1980 2000 0 100
15
10
5
0
200 300 400
Biomass
Biomass
Catch
Frequency
Year
87
Figure 8.8(a) A conceptual model of the three forms of regime shift, and empirical evidence consistent with
discontinuous regime shifts in (b–e) Georges Bank haddock and (f–h) the Black Sea pelagic ecosystem. In Georges Bank
haddock, the regime shift is illustrated by (b) a discrete shift from higher to lower abundance after ~1965, (c) a bimodal
distribution of biomass across the time series, (d) a different functional relationship between fishing mortality and catch
in the two time periods and (e) hysteresis in this functional relationship in simulations of a model fit to empirical data.
Reproduced with permission from Collieet al. (2004). In the Black Sea, the regime shift is similarly supported by (f–h)
changing functional relationships through time between predators and prey. Reproduced with permission from
Daskalovet al. (2007).
STRUCTURE AND FUNCTIONING OF EMERGING MARINE COMMUNITIES 111