Restoration of freshwater lakes - Restoration Ecology - page 169

the resultant changes in shallow lakes the develop-
ment of macrophytes is most striking; the macrophytes
compete for nutrients with phytoplankton and limit
the latter's growth, so that the clear-water phase
often observed in many lakes during spring or early
summer is a culmination of the trophic cascades.
Lake biomanipulation as a restoration technique,
started in the 1980s, is now well documented in the
literature on lake restoration (Gulati et al. 1990,
Lammens et al. 1990, Kufel et al. 1997, Harper et al.
1999, Walz & Nixdorf 1999, Kasprzak et al. 2002,
Gulati & van Donk 2002). In the Netherlands, which
is leading lake biomanipulation research and its
application, more than 20 lakes and ponds ranging
in area from 1.5 to 2650 ha and depth from 0.8 to
2.5 m have been biomanipulated. In virtually all
these cases the standing stocks of planktivorous fish
were reduced drastically. More than a 75% reduction
of the fish stock in winter has been found to be
critical for generating pronounced effects on water
clarity by early spring. The spring peak densities and
grazing maxima of Daphnia spp. in lakes that gener-
ally precede the clear-water state were invariably not
prolonged (Gulati 1990a, 1990b) if reductions of
nutrients or planktivorous fish, or both, were inade-
quate. In case of failure, it was difficult to ascertain
whether Daphnia populations declined due to poor food
quality (high cyanobacterial densities), or due to pre-
dation by planktivores, or both. The northern pike ( Esox
lucius ), which was introduced to control the plank-
tivorous fish, did not establish well over the years in
many lakes. Size-selective predation of larger-bodied
zooplankton ( Daphnia ) by the planktivores led to
a decrease in Daphnia grazing on phytoplankton,
which thus increased and retarded the improvement
in quality.
The biomanipulation theory and its applications
have developed concurrently. The hypothesis of
alternative stable states - a turbid state dominated
by phytoplankton, and a clear state dominated by
macrophytes (Moss 1990, 1998, Scheffer et al. 1993)
- is interesting. The literature evidence to support the
existence of these alternative stable states is, how-
ever, not overwhelming (Gulati & van Donk 2002).
Extreme disturbance may be needed for a lake to shift
from a turbid state to clear-water state; repeated and
sustained reductions of the planktivore fish stocks may
be required to ensure the establishment of macrophytes.

Thus, improvement in the underwater light climate
has been used as the main indicator of success of
the top-down, cascading effects (Meijer et al. 1994a,
1994b, Hosper 1997, Meijer et al. 1999, Van den Berg
et al. 1999, Meijer 2000, Van Nes 2002).
The literature on lake biomanipulation reveals more
long-term failures than successes, mainly because
bottom-up (i.e. nutrient input) effects on the structure
of pelagic foodweb tend to persist even after strong
top-down manipulation (McQueen et al. 1986). It
needs stressing that reduction of nutrients from the
catchment is an important prerequisite for success of
biomanipulation measures (Benndorf 1987): the P
input rate must fall below a certain threshold for gra-
zers to be able to contain phytoplankton biomass.
De Melo et al. (1992), who reviewed the results of 18
enclosure and 26 whole-lake experiments, cast
doubts on the trophic cascade theory of Carpenter and
Kitchell (1992, 1993), mainly due to a weakening of
the cascading effect or absence of the top-down
response at the zooplankton/phytoplankton level in 80%
of the cases analysed. On the other hand, Benndorf
et al. (2002) have attributed the failures of most
biomanipulation work in deeper lakes to extremely high
P loading, implying some bias in the analysis of De
Melo et al. (1992), probably because of more deep lakes
than shallow ones in the analysis. Drenner and
Hambright (2002), however, found no support for the
analysis of Benndorf et al. (2002). We expect the nutri-
ent dynamics and the efficacy of restoration measures
to markedly differ with lake depth (Moss 1998), due
to differences in the sediment-water interactions. In
addition, shallow lakes are more likely to be colonized
by macrophytes and shift to a clear-water state
earlier than the deeper lakes.
Here we highlight a number of conditions that
should be met before applying biomanipulation as a
technique for lake restoration.

Importance of fish in lake restoration

In shallow lakes, the fish are relatively easy to manip-
ulate (Lammens 1999) to produce virtually instantan-
eous effects (Jeppesen 1998). Fish removal varies from
25 to 100%, but biomanipulation measures seem to
be more effective, at about a 75% reduction of the
fish community (Hansson et al. 1998, Moss 1998, Meijer
2000). However, this percentage appears to be rather