552
D. I. Walker, G. A. Kendrick, and A. J. McComb
Cambridge and McComb, 1984; Walker and McComb, 1992; Marba et al., 1996; Short and WyllieEcheverria, 1996; Kendrick et al., 2002). The processes that influence the presence and persistence
of seagrasses operate at scales somewhere between
these extremes. They consist of local destruction of
beds by exceptional storms that occur every 30–50
years, affecting spatial scales of 5–10 km; accumulations of sand sheets that influence 100 m to 2 km,
which can provide entirely new habitats over periods of 5–20 years; or the frequent formation of gaps
(1–100 m) in existing beds during winter storms.
Large-scale studies provide information on net rate
of change, but provide no insight into the controlling
mechanisms, making prediction difficult. In turn,
process studies at small scales (shoot) cannot be
safely extrapolated to the landscape scale (Vidondo
et al., 1997; Kendrick et al., 1999), as both models and field studies have shown that properties of
the landscape emerge at intermediate scales to create seagrass patches, which in turn render extrapolation from small- to large-scale processes uncertain.
This subject is discussed in many other places in this
book, especially, Duarte et al. Chapter 11 and Bell
et al., Chapter 26.
The seagrass literature also contains contradictory statements regarding recovery of seagrasses
from disturbance and the roles of recruitment from
seedlings and clonal growth. Recovery of seagrasses
from disturbance is reported to be slow, and take
many decades to centuries (Kirkman, 1985; Clarke
and Kirkman, 1989; Kirkman and Kuo, 1990; Meehan and West, 2000). Recovery is reported to be primarily from clonal growth (lateral growth of rhizomes) as recruiting seedlings do not survive in any
great number (Kirkman, 1998; but see Orth et al.,
Chapter 5). In contrast, however, Kendrick et al.
(1999, 2000) have documented landscape-scale increases in seagrass cover over 2–3 decades and these
increases are not solely due to clonal growth, suggesting recruitment from seedlings may also play a
role in colonization and recovery. Marb` a and Duarte
(1999) have demonstrated that growth rates of seagrass rhizomes measured at the scale of shoots (cm)
may not extrapolate to the growth dynamics of seagrass patches or meadows (see also Vidondo et al.,
1998). Patch growth of Cymodocea nodosa in the
Mediterranean accelerated with increased patch size
and age which was not due to an increase in rhizome
spreading rates (Marb` a and Duarte, 1995; Vidondo
et al., 1997). Relatively high rates of flowering have
been found in temperate Australian seagrasses in
general (Marba and Walker, 1999) and Posidonia
spp. specifically (Waycott et al., 1997; Walker et al.,
2001; Campey et al., 2002). Seedling survival rates
of Posidonia species are also relatively high (1–10%:
Kuo and Kirkman, 1996; Gobert et al., Chapter 17).
Seagrass recruitment from seed and clonal growth
can and does occur. The question remains: does the
rate of recruitment match the rates of seagrass decline? In this review, we assemble information to
allow comparisons between rates of recruitment and
rates of decline.
III. Loss of Seagrass Beds
Mechanisms of seagrass decline are covered in detail in Kenworthy et al., Chapter 25. The documented
loss of seagrass from direct and indirect human impacts amounts to 33,000 km
2 or 18% of the documented seagrass area, over the last two decades
(Green and Short, 2003), based on an extrapolation of known losses (Short and Wyllie-Echeverria,
1996). Such declines have been widespread in shallow coastal seas and estuaries (see reviews by Walker
and McComb, 1992; Short and Wyllie-Echeverria,
1996 and Chapter 25). Loss of seagrass is usually
correlated with decreases in light intensity and/or increases in sedimentation, or direct physical removal
by coastal engineering. Recovery of seagrass is by
recruitment and clonal expansion.
A. Seagrass Dynamics and Mapping Studies
Most large-scale studies of the distribution of seagrasses have reported losses, but generally those
studies were commissioned because of an environmental disaster, or because an issue had already been
reported (Table 1). In these studies the researcher
has documented the scale of loss, funded by and
for management agencies. These studies are well
summarized in recent research articles and reviews
(Walker and McComb, 1992; Marb` a et al., 1996;
Short and Wyllie-Echeverria, 1996; Kendrick et al.,
2000). Large-scale recovery of seagrasses on unvegetated intertidal and subtidal environments have
been increasingly reported from mapping studies
(Table 2). These increases in seagrass cover receive
less recognition than they should. They infer the
spatial scales and temporal rates of recovery over
a range of seagrass species, from fast growing tropical species (e.g. Halodule species—Robbins, 1997;
Robbins and Bell, 2000) to slow growing temperate
D. I. Walker, G. A. Kendrick, and A. J. McComb
Cambridge and McComb, 1984; Walker and McComb, 1992; Marba et al., 1996; Short and WyllieEcheverria, 1996; Kendrick et al., 2002). The processes that influence the presence and persistence
of seagrasses operate at scales somewhere between
these extremes. They consist of local destruction of
beds by exceptional storms that occur every 30–50
years, affecting spatial scales of 5–10 km; accumulations of sand sheets that influence 100 m to 2 km,
which can provide entirely new habitats over periods of 5–20 years; or the frequent formation of gaps
(1–100 m) in existing beds during winter storms.
Large-scale studies provide information on net rate
of change, but provide no insight into the controlling
mechanisms, making prediction difficult. In turn,
process studies at small scales (shoot) cannot be
safely extrapolated to the landscape scale (Vidondo
et al., 1997; Kendrick et al., 1999), as both models and field studies have shown that properties of
the landscape emerge at intermediate scales to create seagrass patches, which in turn render extrapolation from small- to large-scale processes uncertain.
This subject is discussed in many other places in this
book, especially, Duarte et al. Chapter 11 and Bell
et al., Chapter 26.
The seagrass literature also contains contradictory statements regarding recovery of seagrasses
from disturbance and the roles of recruitment from
seedlings and clonal growth. Recovery of seagrasses
from disturbance is reported to be slow, and take
many decades to centuries (Kirkman, 1985; Clarke
and Kirkman, 1989; Kirkman and Kuo, 1990; Meehan and West, 2000). Recovery is reported to be primarily from clonal growth (lateral growth of rhizomes) as recruiting seedlings do not survive in any
great number (Kirkman, 1998; but see Orth et al.,
Chapter 5). In contrast, however, Kendrick et al.
(1999, 2000) have documented landscape-scale increases in seagrass cover over 2–3 decades and these
increases are not solely due to clonal growth, suggesting recruitment from seedlings may also play a
role in colonization and recovery. Marb` a and Duarte
(1999) have demonstrated that growth rates of seagrass rhizomes measured at the scale of shoots (cm)
may not extrapolate to the growth dynamics of seagrass patches or meadows (see also Vidondo et al.,
1998). Patch growth of Cymodocea nodosa in the
Mediterranean accelerated with increased patch size
and age which was not due to an increase in rhizome
spreading rates (Marb` a and Duarte, 1995; Vidondo
et al., 1997). Relatively high rates of flowering have
been found in temperate Australian seagrasses in
general (Marba and Walker, 1999) and Posidonia
spp. specifically (Waycott et al., 1997; Walker et al.,
2001; Campey et al., 2002). Seedling survival rates
of Posidonia species are also relatively high (1–10%:
Kuo and Kirkman, 1996; Gobert et al., Chapter 17).
Seagrass recruitment from seed and clonal growth
can and does occur. The question remains: does the
rate of recruitment match the rates of seagrass decline? In this review, we assemble information to
allow comparisons between rates of recruitment and
rates of decline.
III. Loss of Seagrass Beds
Mechanisms of seagrass decline are covered in detail in Kenworthy et al., Chapter 25. The documented
loss of seagrass from direct and indirect human impacts amounts to 33,000 km
2 or 18% of the documented seagrass area, over the last two decades
(Green and Short, 2003), based on an extrapolation of known losses (Short and Wyllie-Echeverria,
1996). Such declines have been widespread in shallow coastal seas and estuaries (see reviews by Walker
and McComb, 1992; Short and Wyllie-Echeverria,
1996 and Chapter 25). Loss of seagrass is usually
correlated with decreases in light intensity and/or increases in sedimentation, or direct physical removal
by coastal engineering. Recovery of seagrass is by
recruitment and clonal expansion.
A. Seagrass Dynamics and Mapping Studies
Most large-scale studies of the distribution of seagrasses have reported losses, but generally those
studies were commissioned because of an environmental disaster, or because an issue had already been
reported (Table 1). In these studies the researcher
has documented the scale of loss, funded by and
for management agencies. These studies are well
summarized in recent research articles and reviews
(Walker and McComb, 1992; Marb` a et al., 1996;
Short and Wyllie-Echeverria, 1996; Kendrick et al.,
2000). Large-scale recovery of seagrasses on unvegetated intertidal and subtidal environments have
been increasingly reported from mapping studies
(Table 2). These increases in seagrass cover receive
less recognition than they should. They infer the
spatial scales and temporal rates of recovery over
a range of seagrass species, from fast growing tropical species (e.g. Halodule species—Robbins, 1997;
Robbins and Bell, 2000) to slow growing temperate