20. Biotic Manipulations of Aquatic Ecosystems
Mechanical methods and herbicides are probably
the most effective, but are usually not taxonspecific. Consideration of the spatial pattern of
macrophyte removal should be included in the experimental design because the geometry of remaining macrophyte beds may have important implications for how other parts of the system respond to
manipulation. For example, the amount of edge
habitat produced by mechanical cutting of macrophyte beds affects the nature of predator-prey interactions among littoral zone fishes and their prey
(Trebitz and Nibbelink 1996). Water level regulation is viable in reservoirs where exposure of root
systems to drying and/or freezing can have speciesspecific effects (Cooke et al. 1993).
Macrophyte introduction is usually accomplished by transplanting healthy cuttings, tubers, or
rosettes to the system to be restored. This process
is labor intensive and will only succeed if water and
sediment chemistry are appropriate for the species
of interest, and if light availability is high enough
to support plant growth. In systems where large
herbivores such as grass carp or waterfowl will impede macrophyte restoration, these must be removed or excluded from areas where macrophyte
growth is desired (Lauridsen et al. 1993; Cooke et
al. 1993).
Restoration of Other Structural Features
Manipulations of structural features of aquatic ecosystems have been designed to evaluate how habitat
architecture affects ecosystem processes. Large
woody debris and rock formations are important
structural features of aquatic ecosystems (Johnson
and Stein 1979; Harmon et al. 1986) that are a focus
for ecosystem restoration (NRC 1992). Large
woody debris has been depleted in many rivers and
lakes due to loss of riparian trees and active removal from inshore habitats (Maser and Sedell
1994; Christensen et al. 1996). Loss of rock formations can result from channel straightening and
simplification of fluvial systems, and from waterlevel change and structure removal in lakes. These
losses of structural features have important implications for hydrology, nutrient cycling, and
predator-prey interactions. Most experiments have
been performed in streams with the goal of evaluating the effects on fish production. These manipulations have involved construction of crosschan311
nel log sills (Gowan and Fausch 1996), boulder
groupings, cobble-filled wire gabions, and other
current deflectors (House and Boehne 1985; Moore
and Gregory 1988), and replacement of large
woody debris (Hilderbrand et al. 1998). Large-scale
manipUlation of structural features such as large
woody debris to determine their effects on ecosystem processes is an important avenue for future research in all aquatic habitats and is critical to restoration of many aquatic ecosystems (Hilderbrand
et al. 1998).
Wetland Restoration
Extensive loss of wetlands has focused the attention
of scientists and managers on restoration of disturbed sites throughout North America and Europe
(Zedler 1996). A call for experimental approaches
through use of reference systems and extensive
postmanipulation monitoring (Brinson and Rheinhardt 1996) likens mitigation efforts to ecosystem
experiments. Although the methods of wetland restoration are extensive and often site-specific, some
general principles are useful. Most wetland restoration projects involve restoration of the natural hydrology, elimination of contaminants and pollutants, and reestablishment of lost species (NRC
1992). Revegetation may require active seeding because of depleted seed banks in many restored wetlands (Galatowitsch and van der Valk 1996). The
appropriate time frame to evaluate the result of a
wetland restoration effort is debatable, but usually
underestimated (Mitsch and Wilson 1996). By putting wetland restoration into an adaptive management context (Gunderson et al. 1995), the appropriate time frame becomes a dependent variable in
these ecosystem manipulations.
Simulation Modeling, Manipulation
Strength, and Statistical Power
Simulation modeling is a powerful tool for planning ecosystem experiments, exploring possible responses to manipulations, and interpreting results
(Kitchell 1992; Carpenter and Kitchell 1993b;
Scheffer 1997; Trebitz et al. 1997). Modeling is
especially valuable in the implementation of management experiments because it helps crystallize
Mechanical methods and herbicides are probably
the most effective, but are usually not taxonspecific. Consideration of the spatial pattern of
macrophyte removal should be included in the experimental design because the geometry of remaining macrophyte beds may have important implications for how other parts of the system respond to
manipulation. For example, the amount of edge
habitat produced by mechanical cutting of macrophyte beds affects the nature of predator-prey interactions among littoral zone fishes and their prey
(Trebitz and Nibbelink 1996). Water level regulation is viable in reservoirs where exposure of root
systems to drying and/or freezing can have speciesspecific effects (Cooke et al. 1993).
Macrophyte introduction is usually accomplished by transplanting healthy cuttings, tubers, or
rosettes to the system to be restored. This process
is labor intensive and will only succeed if water and
sediment chemistry are appropriate for the species
of interest, and if light availability is high enough
to support plant growth. In systems where large
herbivores such as grass carp or waterfowl will impede macrophyte restoration, these must be removed or excluded from areas where macrophyte
growth is desired (Lauridsen et al. 1993; Cooke et
al. 1993).
Restoration of Other Structural Features
Manipulations of structural features of aquatic ecosystems have been designed to evaluate how habitat
architecture affects ecosystem processes. Large
woody debris and rock formations are important
structural features of aquatic ecosystems (Johnson
and Stein 1979; Harmon et al. 1986) that are a focus
for ecosystem restoration (NRC 1992). Large
woody debris has been depleted in many rivers and
lakes due to loss of riparian trees and active removal from inshore habitats (Maser and Sedell
1994; Christensen et al. 1996). Loss of rock formations can result from channel straightening and
simplification of fluvial systems, and from waterlevel change and structure removal in lakes. These
losses of structural features have important implications for hydrology, nutrient cycling, and
predator-prey interactions. Most experiments have
been performed in streams with the goal of evaluating the effects on fish production. These manipulations have involved construction of crosschan311
nel log sills (Gowan and Fausch 1996), boulder
groupings, cobble-filled wire gabions, and other
current deflectors (House and Boehne 1985; Moore
and Gregory 1988), and replacement of large
woody debris (Hilderbrand et al. 1998). Large-scale
manipUlation of structural features such as large
woody debris to determine their effects on ecosystem processes is an important avenue for future research in all aquatic habitats and is critical to restoration of many aquatic ecosystems (Hilderbrand
et al. 1998).
Wetland Restoration
Extensive loss of wetlands has focused the attention
of scientists and managers on restoration of disturbed sites throughout North America and Europe
(Zedler 1996). A call for experimental approaches
through use of reference systems and extensive
postmanipulation monitoring (Brinson and Rheinhardt 1996) likens mitigation efforts to ecosystem
experiments. Although the methods of wetland restoration are extensive and often site-specific, some
general principles are useful. Most wetland restoration projects involve restoration of the natural hydrology, elimination of contaminants and pollutants, and reestablishment of lost species (NRC
1992). Revegetation may require active seeding because of depleted seed banks in many restored wetlands (Galatowitsch and van der Valk 1996). The
appropriate time frame to evaluate the result of a
wetland restoration effort is debatable, but usually
underestimated (Mitsch and Wilson 1996). By putting wetland restoration into an adaptive management context (Gunderson et al. 1995), the appropriate time frame becomes a dependent variable in
these ecosystem manipulations.
Simulation Modeling, Manipulation
Strength, and Statistical Power
Simulation modeling is a powerful tool for planning ecosystem experiments, exploring possible responses to manipulations, and interpreting results
(Kitchell 1992; Carpenter and Kitchell 1993b;
Scheffer 1997; Trebitz et al. 1997). Modeling is
especially valuable in the implementation of management experiments because it helps crystallize