Ecological Indicators 23 (2012) 211–219
Contents lists available at SciVerse ScienceDirect
Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind
Light thresholds derived from seagrass loss in the coastal zone of the northern
Great Barrier Reef, Australia
C.J. Collier a,∗ , M. Waycott a,1 , L.J. McKenzie b
a
b
School of Marine and Tropical Biology, James Cook University, Angus Smith Dve, Townsville Qld 4811, Australia
Fisheries Queensland, Department of Employment, Economic Development and Innovation, Northern Fisheries Centre, 38-40 Tingira St., Portsmith, Cairns, Qld 4870, Australia
a r t i c l e
i n f o
Article history:
Received 2 December 2011
Received in revised form 30 March 2012
Accepted 2 April 2012
Keywords:
Seagrass
Light requirements
Water quality
Thresholds
Guidelines
Halodule uninervis
a b s t r a c t
There is a world-wide trend for deteriorating water quality and light levels in the coastal zone, and this
has been linked to declines in seagrass abundance. Localized management of seagrass meadow health
requires that water quality guidelines for meeting seagrass growth requirements are available. Tropical
seagrass meadows are diverse and can be highly dynamic and we have used this dynamism to identify light thresholds in multi-specific meadows dominated by Halodule uninervis in the northern Great
Barrier Reef, Australia. Seagrass cover was measured at ∼3 month intervals from 2008 to 2011 at three
sites: Magnetic Island (MI) Dunk Island (DI) and Green Island (GI). Photosynthetically active radiation
was continuously measured within the seagrass canopy, and three light metrics were derived. Complete
seagrass loss occurred at MI and DI and at these sites changes in seagrass cover were correlated with
the three light metrics. Mean daily irradiance (Id ) above 5 and 8.4 mol m−2 d−1 was associated with gains
in seagrass at MI and DI, however a significant correlation (R = 0.649, p < 0.05) only occurred at MI. The
second metric, percent of days below 3 mol m−2 d−1 , correlated the most strongly (MI, R = −0.714, p < 0.01
and DI, R = −0.859, p = <0.001) with change in seagrass cover with 16–18% of days below 3 mol m−2 d−1
being associated with more than 50% seagrass loss. The third metric, the number of hours of light saturated irradiance (Hsat ) was calculated using literature-derived data on saturating irradiance (Ek ). Hsat
correlated well (R = 0.686, p < 0.01; and DI, R = 0.704, p < 0.05) with change in seagrass abundance, and
was very consistent between the two sites as 4 Hsat was associated with increases in seagrass abundance
at both sites, and less than 4 Hsat with more than 50% loss. At the third site (GI), small seasonal losses
of seagrass quickly recovered during the growth season and the light metrics did not correlate (p > 0.05)
with change in percent cover, except for Id which was always high, but correlated with change in seagrass
cover. Although distinct light thresholds were observed, the departure from threshold values was also
important. For example, light levels that are well below the thresholds resulted in more severe loss of seagrass than those just below the threshold. Environmental managers aiming to achieve optimal seagrass
growth conditions can use these threshold light metrics as guidelines; however, other environmental
conditions, including seasonally varying temperature and nutrient availability, will influence seagrass
responses above and below these thresholds.
© 2012 Published by Elsevier Ltd.
1. Introduction
Seagrass meadows are a critical element in tropical coastal
ecosystems for a number of reasons: they form habitat for diverse
fauna and flora and they are food for herbivores, including endangered turtles, dugongs and manatees (Sheppard et al., 2010); they
support the productivity of adjacent coral reefs and mangroves
∗ Corresponding author. Tel.: +61 7 4781 5745; fax: +61 7 4725 1570.
E-mail address: [email protected] (C.J. Collier).
1
Current address: School of Earth and Environmental Sciences, The University of
Adelaide, Adelaide, SA 5001, Australia.
1470-160X/$ – see front matter © 2012 Published by Elsevier Ltd.
http://dx.doi.org/10.1016/j.ecolind.2012.04.005
through energy and material subsidies (Heck et al., 2008); and,
they support fauna that are critical for the food security of many
tropical nations (Unsworth and Cullen, 2010). In addition, their primary productivity drives coastal nutrient cycling (Costanza et al.,
1997), they stabilize sediment and they are biochemical and hydrodynamic modifiers of their local environment (Marbà et al., 2006).
Their role in the coastal zone is therefore unquestionable and yet,
we do not know enough about their basic growth requirements
to provide environmental managers with the information needed
to protect them, particularly so for tropical seagrass meadows
(Waycott et al., 2005).
Seagrass loss has become a global concern with accelerating
rates of loss over the previous decades placing them amongst
212
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
the most threatened habitats in the world (Orth et al., 2006;
Waycott et al., 2009). The causes of loss are both anthropogenic
and natural but most commonly result from reductions in water
quality (Waycott et al., 2009). Water quality reductions are particularly devastating to seagrass meadows because losses can be
widespread, and chronic. In the Great Barrier Reef (GBR), water
quality is a strong driver of ecosystem health and biodiversity
(De’ath and Fabricius, 2010). For seagrass meadows, low light levels
associated with poor water quality are thought to be the primary
factor limiting the growth of coastal seagrasses (Waycott et al.,
2005). Because of its strongly influential role in modifying ecosystem processes, there are large watershed and land use management
programs in place to improve the quality of water entering the GBR
(DPC, 2011).
For the localized management of water quality, guidelines indicate levels at which, if exceeded, could result in unacceptable
ecosystem harm (GBRMPA, 2009). Exceeding these values should
trigger action to minimize harm. In order for guidelines to be
developed, basic ecosystem health requirements are needed. We
know that light availability is a critical determinant of seagrass
growth and abundance (Ralph et al., 2007), however we do not
really know how much light is enough for many habitats. As light
requirements are not known for many seagrass species, particularly locally-specific thresholds, they are frequently not included in
guidelines. Alternatively, guidelines are set against species in different ecosystems for which more data is available, such as coral reefs,
under the assumption that they will protect other adjacent habitat, such as seagrass (GBRMPA, 2009; De’ath and Fabricius, 2010).
Furthermore, surrogate indicators of light such as turbidity or secchi depth are frequently adopted, but these are difficult to apply in
shallow ecosystems, where changing sea level alters the impact of
turbidity (Sofonia and Unsworth, 2009).
Seagrass meadows are highly dynamic marine habitats and
tropical seagrass meadows exemplify their dynamic nature
(Waycott et al., 2005). Cycles of meadow development and loss are
thought to be a natural part of the system, but have been rarely documented (Birch and Birch, 1984; Campbell and McKenzie, 2004).
The frequency and intensity of these cycles in seagrass loss and
meadow development are exacerbated by anthropogenic activities,
which have increased the stressors on these systems (Grech et al.,
Fig. 1. Study sites of Green Island, Dunk Island and Magnetic Island in the northern Great Barrier Reef, Australia and the rivers and their water sheds (grey fill) that primarily
influence water quality at the study sites.
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
2011). Tropical seagrass meadows are also complex and spatially
variable (Carruthers et al., 2002; Waycott et al., 2011). This makes it
difficult to apply standard techniques for the establishment of basic
growth requirements. For example, minimum light requirements
(MLR) can be determined by measuring light at the long-term depth
limit of seagrass meadows (Dennison et al., 1993), but only if the
meadow edge is clearly defined and relatively static which is rarely
the case in the dynamic seagrass habitats of the GBR. Alternatively,
using experimental approaches, light treatments are applied and
responses detected, but this approach is limited by treatment resolution (Ruiz and Romero, 2001; Collier et al., 2009; Lavery et al.,
2009; Collier et al., in press).
The objectives of this work were to measure light levels that
indicate integrated and ecologically harmful conditions for tropical seagrass meadows. In this study we utilize the dynamic nature
of seagrass meadows to measure light conditions associated with
seagrass loss and gain. We measured changes in seagrass cover over
four years from 2008 to 2011 and correlated it with three light metrics. This monitoring period was associated with the incremental
decline in seagrass cover at two of the three locations, which ended
in the complete disappearance of seagrass at the monitoring sites
in 2011.
2. Methods
2.1. Study sites
Data collection began in 2008 at three island locations in the
northern Great Barrier Reef, Australia (Fig. 1). The study sites were
located at Magnetic Island (MI, −1.5 m below lowest astronomical tide; −19◦ 10.9, 146◦ 50.6), Dunk Island (DI, −0.9 m below LAT,
−17◦ 55.9 146◦ 08.4), and Green Island (GI, −1.2 m below LAT, −16◦
45.5, 145◦ 58.4) (Fig. 1). The study sites were nearshore reef habitats, with varying degrees of influence from terrigenous run-off.
The species composition at the study sites was variable and the
only species to occur at all sites was Halodule uninervis, which is
common throughout the Indo-west Pacific (Waycott et al., 2004).
2.2. River discharge
River discharge data from 1975 to 2011 were obtained for the
rivers that have the greatest influence over the study sites (Devlin
et al., in press). Data were sourced from the State Government of
Queensland, Department of Environment and Resource Management. These are the Mulgrave River, affecting Green Island, the Tully
River, affecting Dunk Island, and the Burdekin River, affecting Magnetic Island (Fig. 1). River discharge was totaled for each month and
a monthly average from 1975 to 2011, and average from 2009 to
2011 was derived to compare recent discharge (previous 3 years)
with the long-term average.
213
Total seagrass cover, as well as the percentage of each species
present, were recorded in each quadrat. Mean total percent cover
and percent of each species, was generated for each site and each
time. Change in percent cover was calculated from one sampling
event to another according to the following equation
[%CovT 2 − %CovT 1 ]
× 100
%CovT 2
(1)
where %CovT1 and %CovT2 are the total percent cover at the first and
second measurement time in a row.
2.4. Light measurements
Photosynthetically active radiation (PAR) was measured continuously at all sites from the start of the measurement period.
Light loggers (2␲ light loggers; Submersible Odyssey Photosynthetic Irradiance Recording System, Dataflow Systems Pty Ltd.,
New Zealand) were installed at seagrass canopy height. The loggers were attached to a star picket, which held the sensor at
seagrass canopy height (∼15 cm from the sediment surface). PAR
was recorded every 30 min. Light loggers were calibrated against
a certified reference PAR sensor ((LI-CORTM LI-192SB Underwater
Quantum Sensor) in controlled laboratory conditions. A custombuilt calibration device held the light loggers a consistent distance
from the source light and excluded all external light. The loggers
were calibrated before each deployment. Because the light loggers
were calibrated in air, a multiplication factor of 1.33 was applied
to data to allow for the differences in light absorption properties
between air (calibration medium) and water (deployment media)
(Kirk, 1994).
The light loggers were checked and/or replaced every 6–12
weeks, i.e. on each biological sampling event and at one time in
between sampling events. Each light logger had a custom-made
wiper unit to keep the sensor clean while deployed. Some fouling
of the sensors still occurred, therefore, duplicate light loggers (i.e.
two at each site) were deployed resulting in one redundant data
set. If one logger or wiper unit failed, data from the other logger
was used. Despite this added to effort to ensure continuous light
data some data gaps exist where both systems failed for short periods. If minor epiphyte fouling of the light sensor occurred (0–25% of
the sensor covered) because the wiper was not working effectively
(as opposed to not working at all), the data were corrected. A linear back-calculation was applied to the data to compensate for the
light absorbing properties of the epiphytes. A linear calculation was
applied, as opposed to an exponential (growth) calculation, as the
wiper retards the rate of epiphyte growth even when not working
effectively. If epiphyte cover was greater than 25%, the data were
excluded in the analysis as there is a reduced confidence in its quality and our ability to make appropriate adjustments. If the sensor
was fouled from sediment being deposited on the sensor, the data
was also excluded as it was not possible to estimate the rates and
dynamics of sediment deposition and removal.
2.3. Seagrass cover
2.5. Describing the light environment
Seagrass cover (percent of sediment surface covered in seagrass)
was measured approximately every 3 months, or as weather conditions permitted. Seagrass cover was measured on self-contained
underwater breathing apparatus (SCUBA). Three parallel transects
were established at each site. The start of each transect was permanently marked with a star picket. Percent cover was estimated
within a 50 × 50 cm quadrat every 5 m starting from 0 to 50 m (i.e.
11 quadrats per transect; 33 in total for each site) (McKenzie et al.,
2001; http://www.seagrasswatch.org/monitoring.html). Transects
were 2–3 m apart from each other for logistical purposes of SCUBA
diving in often poor visibility and were aligned along the depth
contour.
30 min PAR data were converted to total daily irradiance. Daily
irradiance was then averaged for the full duration between seagrass
percent cover measures (∼3 months). This gave an average daily
irradiance (Id ), which was correlated against change in seagrass
cover over the same monitoring period.
Frequency distributions for Id were calculated with daily irradiance ranging from 0 to 14+ mol m−2 d−1 at increasing increments of
1 mol m−2 d−1 (i.e. number of days <1 mol m−2 d−1 , <2 mol m−2 d−1 ,
up to 14+ mol m−2 d−1 ). The frequency distribution of Id was calculated for each sampling period of approximately 3 months, i.e.
the same duration over which percent cover was measured. The
214
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
Table 1
Values of the half-saturation constant (Ek ) and the ambient light intensity when Ek
was measured from published and unpublished sources, and the median diurnal
peak in irradiance at each site used to derive site adapted Ek value.
Location
Ek (mol m−2 s−1 )
Light intensity
(mol m−2 s−1 )
Reference
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
Mesocosm
In situ
In situ
57.1
159.7
191.4
199.5
190.7
251.5
331.9
345.4
163.9
426.0
425.2
329.1
601.0
548.3
40.0
140.7
153.6
172.1
270.9
306.0
347.3
361.3
400.0
590.9
638.3
680.8
865.0
998.0
Collier et al. (2011)
Collier (unpub.)
Collier (unpub.)
Collier (unpub.)
Collier (unpub.)
Collier (unpub.)
Collier (unpub.)
Collier (unpub.)
Collier et al. (2011)
Collier (unpub.)
Collier (unpub.)
Collier (unpub.)
Campbell et al. (2007)
Campbell et al. (2007)
Location
Median of maximum
light (mol m−2 s−1 )
Site-adapted Ek
Green Island
Dunk Island
Magnetic Island
531.9
409.2
275.1
231.7
229.2
159.7
Data sources: (Campbell et al., 2007; Collier et al., 2011) and C. Collier, unpublished
data.
proportion of days within each incremental range of Id (% of days)
was then correlated against change in seagrass cover over the same
sampling period. The range in Id that had the highest correlation
with change in percent cover was used to describe light thresholds
in terms of frequency of days at or below that irradiance.
Hours of saturating irradiance (Hsat ) were calculated for
the measurement period (i.e. ∼3 months) using published and
unpublished values for the half-saturation constant of seagrass
photosynthesis (Ek ). The full set of available Ek data for H. uninervis were collated (Table 1). There is a wide range in the reported
Ek values as the photosystems adapt to their light environment
(Ralph and Gademann, 2005) and Ek data from high light sites (e.g.
Green Island and Lizard Island (Campbell et al., 2007)) were substantially higher than those held under constant low light (Collier
et al., 2011). To estimate an appropriate Ek value to apply for each
site, the median maximum daily light intensity was calculated for
each site. Ek values for the corresponding light intensity were averaged up to the median value. This resulted in a literature-derived Ek
value that was tailored to the light conditions of the site. The number of hours above Ek in each day was then calculated (HSat ). This
is a species-specific measure that was generated for H. uninervis
only, as it was a dominant species and it occurred at all sites. Hsat
was therefore correlated against change in H. uninervis abundance
only, whereas the other two light metrics were correlated against
change in total seagrass abundance.
The relationship between change in seagrass cover and the light
indices were calculated using linear regression in SPSS 19.0.
3. Results
3.1. River discharge
The wet seasons (January–March) of 2009–2011 had elevated
discharge from rivers adjacent to the study sites compared to the
long-term average (Fig. 2). This was particularly evident in the discharge from the Tully and Burdekin Rivers, which have an influence
over the water quality at DI and MI, respectively. Furthermore,
Tropical Cyclone Yasi crossed the coast near Tully in the northern GBR on 3rd February 2011. This brought devastating winds and
Table 2
Light at the study sites including maximum daily instantaneous irradiance (I InstMax ,
␮mol m−2 s−1 ), mean daily instantaneous irradiance (␮mol m−2 s−1 ) and total daily
irradiance (Id , mol m−2 d−1 ) at each site for the duration of the study. Values are
means for the duration of the study with standard deviation in parenthesis.
GI
DI
MI
I InstMax
I InstAv
Id
526.8 (187.7)
340.0 (207.2)
280.8 (164.5)
122.8 (53.4)
86.1 (48.9)
61.2 (38.7)
10.6 (4.5)
7.4 (4.2)
5.3 (3.3)
heavy rainfall to regions south of Tully, including DI and to a lesser
extent, MI.
3.2. Seagrass cover
All sites began as healthy, multispecies seagrass meadows. At
MI there was substantial seagrass loss with a starting cover of 47%,
comprised largely of Cymodocea serrulata and Halodule uninervis,
which peaked at 54% in November 2008 and then declined thereafter to 0% in May 2011 (Fig. 3). The DI site had characteristics of a
recovering meadow when monitoring began as it was dominated
by Halodule uninervis and Halophila ovalis with only few Cymodocea
serrulata shoots. At DI, seagrass cover started at 3%, reached 7% in
December 2009 and declined to 0% after Tropical Cyclone Yasi (3
February 2011). However, the meadows at DI were in a strongly
declining trajectory prior to TC Yasi. The seagrass losses observed
at DI and MI, were typical of trends throughout the southern and
central GBR over this period (McKenzie et al., in press). The largest
declines occurred in the wet season (January–March) or just after
the wet season (Fig. 3). Seagrass cover at GI was variable but largely
stable over 2008–2011 (Fig. 3). Seagrass cover at GI underwent seasonal declines whereby cover was lowest in winter (around July)
and peaked just prior to or early in the wet season. At GI, seagrass
cover did not decline by more than 50% at any one time, therefore
losses of greater than 50% at the other two sites were taken to be
‘atypical’ (McKenzie et al., 1998) and were used to derive thresholds
associated with losses of concern.
3.3. Light
Daily irradiance (Id ) at the three sites differed from each other
during the measurement period (2008–2011). At GI mean Id was
10.6 mol m−2 d−1 and maximum daily instantaneous irradiance
(I InstMax ) was on average 526.8 ␮mol m−2 s−1 (Table 2). Id at
DI was lower at 7.4 mol m−2 d−1 on average and I InstMax was
340.0 ␮mol m−2 s−1 . MI had the poorest light conditions with Id
being half of GI at 5.3 mol m−2 d−1 on average and I InstMax was
280.8 ␮mol m−2 s−1 .
3.4. Seagrass cover and light correlations
Change in total seagrass cover correlated significantly with total
daily irradiance (Id ) at MI, but not at the other sites (Fig. 4). Increases
in seagrass cover occurred at Id ranging from 5 to 8.4 mol m−2 d−1
(Fig. 4, Table 3). At DI and MI – the two sites where substantial
seagrass loss occurred – losses of more than 50% were associated
with average Id below 8.2 and 5.2 mol m−2 d−1 . At DI, although
there was not a significant trend between change in seagrass
cover and Id , there were clear distinctions in daily irradiance
leading to loss (less than 8.2 mol m−2 d−1 ) and gain (greater than
8.4 mol m−2 d−1 ). There is a small gap in between these thresholds
(8.2–8.4 mol m−2 d−1 for which we recorded no change in seagrass
cover. At GI, there was considerable overlap in the irradiance leading to loss or gain, and average daily irradiance was always above
7.1 mol m−2 d−1 .
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
215
Fig. 2. River discharge from the rivers that primarily influence water quality at the study sites. The Mulgrave River, near Green Island, the Tully River near Dunk Island, and
the Burdekin River near Magnetic Island showing monthly totals for (2009–2011) in solid lines and long-term monthly average (1975–2011) in dashed lines. Data courtesy
of the State Government of Queensland, Department of Environment and Resource Management.
Change in seagrass cover was also correlated against frequency
of days when daily irradiance was <1 mol m−2 d−1 through to
<14 mol m−2 d−1 in 1 mol m−2 d−1 increments (Table 4). At MI,
change in seagrass cover correlated highly with frequency of
days when daily irradiance was <1 through to <7 mol m−2 d−1 .
However, the strongest correlation was for frequency of days (%)
below 3–5 mol m−2 d−1 . At DI, change in seagrass cover correlated
with frequency of days where daily irradiance was <1 through to
<6 mol m−2 d−1 but the highest correlation was for % days below
2–4 mol m−2 d−1 . We based our threshold values on % days below
3 mol m−2 d−1 for both sites (Table 3, Fig. 5), but also describe the
threshold values for the other highly correlated light levels. At MI,
Id below 3, 4 and 5 mol m−2 d−1 for 18%, 51% and 78% of days or
more, was associated with more than 50% loss of seagrass cover. At
DI Id below 2, 3 and 4 mol m−2 d− 1 for 10%, 16% and 22% of days
were associated with losses greater than 50%.
There were also good correlations between change in seagrass
cover and hours of light saturated irradiance (HSat , Fig. 6). Change
in the cover of H. uninervis correlated with Hsat at MI and DI. Furthermore, there were distinct differences in HSat associated with
H. uninervis gain and loss and they were consistent between sites.
Gains in H. uninervis at MI and DI occurred above 3.7 and 4.2 h and
losses greater than 50% occurred below 3.9 and 4.2 h (Table 2). The
GI subtidal site has a similar cut-off associated with gains in H. uninervis (4.5 h), however, there was considerable overlap between light
levels associated with gain and loss at GI.
GI
60
40
20
Percent cover
0
DI
60
40
20
0
MI
60
40
20
0
Jan
Apr
Jul
2008
Oct
Jan
Apr
Jul
2009
Oct
Jan
Apr
Jul
2010
Oct
Jan
Apr
Jul
Oct
Jan
2011
Fig. 3. Total seagrass cover (bold circles) and cover of Halodule uninervis (white triangles) in permanent transects at Green Island (GI), Dunk Island (DI) and Magnetic Island
(MI) from 2008 to 2011.
216
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
Change in total seagrass cover (%)
200
GI
DI
MI
150
R = 0.639
p < 0.05
100
R = 0.649
p < 0.05
R = 0.537
p =n.s.
50
0
-50
-100
2
0
4
6
8
10
12
14
0
2
4
Id (mol m-2 d-1)
6
8
10
12
14
0
2
4
Id (mol m-2 d-1)
6
8
10
12
14
Id (mol m-2 d-1)
Fig. 4. Change in daily irradiance (Id , mol m−2 d−1 ) reaching seagrass canopy height and change in total seagrass cover between monitoring times (∼3 months) at Green
Island (GI), Dunk Island (DI) and Magnetic Island (MI) for the period 2008–2011.
200
GI
DI
MI
Change in seagrass cover (%)
150
100
R=-0.459
p=n.s.
R=-0.859
p<0.001
R=-0.714
p<0.01
50
0
-50
-100
0
20
40
60
100 0
80
20
40
60
100 0
80
20
40
60
80
100
Days below 3mol m-2 d-1 (%)
Fig. 5. Percent of days when daily irradiance was lower than 3 mol m−2 d−1 and change in seagrass cover between monitoring times (∼3 months) at Green Island (GI), Dunk
Island (DI) and Magnetic Island (MI) for the period 2008–2011.
4. Discussion
occurred. We identified three different light metrics that correlated
with changes in seagrass cover and we identified critical thresholds
associated with seagrass loss and gain. Exceedance of these thresholds occurred at two out of the three monitoring sites (MI and DI)
during successive wet season events that led to seagrass loss.
This study presents a unique approach to identifying light
thresholds from real-time changes in subtidal seagrass cover,
among a variety of locations where complete seagrass loss had
Change in H. univeris cover (%)
200
GI
DI
MI
150
100
50
R = 0.349
p > n.s.
R = 0.704
p < 0.05
R = 0.686
p < 0.01
0
-50
-100
0
1
2
3
4
5
HSat (Hrs)
6
7
8 0
1
2
3
4
5
HSat (Hrs)
6
7
8 0
1
2
3
4
5
6
7
8
HSat (Hrs)
Fig. 6. Mean HSat (∼3months) and change in the cover of H. uninervis over the same duration at Green Island, Dunk Island and Magnetic Island during the period 2008–2011.
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
Table 3
Light thresholds associated with seagrass loss (more than 50% loss) or gain at the
three monitoring sites from January 2008 to May 2011 for total seagrass cover compared to daily irradiance (Id ) and days below 3 mol m−2 d−1 , and change in Halodule
uninervis cover compared to mean hours of light saturated photosynthesis (Hsat ).
Bold type indicates a significant correlation between change in seagrass cover and
the light metric. Note that at Green Island there were no losses greater than 50%.
Id (mol m−2 d−1 )
Location
Max light levels leading to >50% loss
–
Green Island
Dunk Island
8.2
Magnetic Island
4.0
Id (mol m−2 d−1 )
Location
% of days below
3 mol m−2 d−1
H. uninervis
(HSat )
–
15.9
18.1
–
4.2
3.9
Days above
3 mol m−2 d−1
H. uninervis
(HSat )
Minimum light levels associated with seagrass gain
9.7
5.3
Green Island
8.4
10.8
Dunk Island
5.0
16.7
Magnetic Island
4.5
4.2
3.7
Bold type indicates that the relationship between seagrass cover and the light metric
was significantly correlated (Figs. 4–6).
The ‘thresholds’ identified in this study, are light levels required
for the long-term survival of these tropical seagrass meadows, identified during ‘real-time’ loss of seagrass. Complete loss of seagrass
after the wet season in 2011 occurred at DI and MI. These two sites
showed clearly defined light levels that were associated with either
seagrass loss or gain, i.e. losses occurred at very low light intensities that caused dramatic seagrass loss (i.e. greater than 50%) over
the sampling periods. We suggest that for these sites, low light was
probably the main cause of seagrass loss, given the good correlation between change in seagrass cover and our measured light
indices. However, at DI and MI, low light periods tended to occur
during the wet season when other flood-related impacts could also
contribute to the stress (e.g. reduced salinity (Touchette, 2007). In
contrast, at Green Island, there were periods of seagrass loss and
gain that followed seasonal cycles (i.e., minimum cover in winter).
Thus the fluctuations in seagrass cover did not result in any more
than 50% loss at any time and were not associated with specific
light levels. Despite this, Id at GI was significantly correlated with
seasonal change in seagrass cover, suggesting that even in habitats with relatively clear water, light is an important environmental
driver.
There are a number of ways to derive light thresholds, including experimental manipulation of the light environment (in situ
or in aquaria) (Grice et al., 1996; Thom et al., 2008; Collier et al.,
in press), derivation of compensation irradiance using O2 fluxes
Table 4
Pearson correlation co-efficient (two-tailed) between percent of days within an
increasing range (in mol m−2 d−1 increments) of daily irradiance (Id ) and change
in total seagrass cover 3.
Id range
0–1
0–2
0–3
0–4
0–5
0–6
0–7
0–8
0–9
0–10
0–11
0–12
0–13
0–14
Green Island
Dunk Island
Magnetic Island
R2
p
R2
p
−0.436
−0.245
−0.459
−0.522
−0.609
−0.602
−0.532
−0.559
−0.539
−0.585
−0.600
−0.511
−0.502
−0.512
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
−0.661
−0.837
−0.859
−0.827
−0.744
−0.761
−0.63
−0.496
−0.172
−0.088
0.085
0.036
−0.111
0.079
<0.05
<0.01
<0.01
<0.01
<0.05
<0.05
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
R2
−0.622
−0.621
−0.718
−0.834
−0.772
−0.664
−0.616
−0.534
−0.597
−0.376
0
p
<0.05
<0.05
<0.01
<0.001
<0.01
<0.01
<0.05
n.s.
n.s
n.s
n.s
217
(photosynthesis–irradiance curves) (Masini and Manning, 1997;
Thom et al., 2008) and maximum depth penetration (Dennison,
1987; Collier et al., 2007; Thom et al., 2008). In this paper, we
outline an alternative approach, where the dynamic changes in
seagrass abundance (i.e. loss and gains) were correlated with light
metrics over the same duration. The benefits of this approach are
that thresholds were recorded during real reductions in light. This
allows natural features of the light environment to be captured
such as changes in light colour, diurnal variation of incoming irradiance resulting from incoming solar insolation and tidal changes in
sea level. This allows comparison with associated impacts such as
increased run-off, elevated nutrients, and lower salinity (e.g. Devlin
and Brodie, 2005). Furthermore, changes measured throughout
the year will incorporate seasonal variability (e.g. changes in
water temperature), accounting for these effects. This approach
also measured meadow-scale responses and therefore removed
“scaling-up” issues of small-scale experimental approaches. Disadvantages of our approach relate to the potential for other
environmental changes associated with riverine discharge to cause
changes in seagrass abundance thus limiting a more direct interpretation of change being caused by light reduction alone. However,
we can infer from the good correlations at DI and MI that light
was a very dominant factor. In contrast, experimental approaches
would enable manipulation of individual factors such as light,
albeit using artificial light reducing conditions. Our approach was
not limited by treatment resolution to the same extent as during
experimental manipulations; however, the method is restricted to
light levels that occurred between sampling events. As a result
there was not always a close overlap between light levels associated with gain and those with >50% loss. For example, at MI,
>50% loss occurred at Id <4.0 mol m−2 d−1 while gains occurred at
>5.0 mol m−2 d−1 . We would not necessarily expect these values
to overlap exactly, given the loss is for >50% while gains are for
any increase in seagrass abundance; however, we are left with a
gap and a potential margin of error. Furthermore, significant loss
of seagrass in these communities takes around 1.5 months (Collier
et al., 2011), while our sampling frequency was every 3 months.
Therefore, significant changes may have occurred within a portion
of the measured intervals, although light metrics were derived from
the entire 3 month duration. Despite these potential drawbacks,
we have derived thresholds using three light metrics that are relatively consistent between MI and DI – the two sites where loss
occurred – providing a robust starting point for ongoing refinement
of thresholds.
In addition to the detection of thresholds for the decline of seagrass cover, we were able to detect light thresholds associated with
gains in seagrass abundance. These were: 5 mol m−2 d−1 at MI; 15.9
and 18.1 days below 3 mol m−2 d−1 at DI and MI and 3.7 and 4.2
Hsat at DI and MI. This is an important and applied outcome of
this study as these are useful targets against which to set water
quality guidelines to enable seagrass growth. Achieving these light
levels will not necessarily assure gains in seagrass abundance as
other environmental factors, such as water temperature, nutrient
availability and disturbance regime are critical in affecting seagrass
growth rates (Udy et al., 1999; Rasheed, 2004; Collier et al., 2011).
However, as light is typically the primary growth limitation, if light
levels can be achieved through management actions the first step in
providing conditions for potential gains in seagrass abundance will
be met. Because of this we would recommend that environmental
managers use light levels associated with gains as a guideline value,
rather than those associated with loss.
4.1. Daily irradiance
Light levels associated with seagrass gain ranged from
5 mol m−2 d−1 at MI, to 8.4 mol m−2 d−1 at DI (though not
218
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
significantly correlated to seagrass change at DI). This is slightly
lower than the minimum light requirements identified for H. uninervis (the dominant species that also occurred at all sites) in
the Gulf of Carpentaria in northern Queensland (Longstaff and
Dennison, 1999) which was 9–12 mol photons m−2 d−1 (equating to 14–19% surface light) based on Id at the deepest edge
of the meadow. However, in the field study of Longstaff and
Dennison (1999), the meadows were intertidal and very shallow
(0.1 m below lowest astronomical tide) and in these environments
very large peaks in instantaneous irradiance can occur during
low tide. These peaks increase mean Id but do not necessarily contribute equally to net carbon gain due to photosynthetic
down-regulation, photosynthetic damage in high light or carbon
limitation (Ralph and Gademann, 2005; Beer et al., 2006). Low
light-adapted plants are particularly inefficient at high light intensities and so extremely high light levels, though contributing to Id
do not translate into photosynthetic gain (Ralph and Gademann,
2005). Therefore, it is not surprising that the apparent MLR for
very shallow turbid environments would be higher than those from
subtidal habitats. The contribution of high peaks in instantaneous
irradiance to averaged light conditions is probably one of the reasons that Id was not the most highly correlative metric derived
here.
4.2. Frequency of low light days
The strongest correlation with change in seagrass cover was
found for the percent of days below 3 mol photons m−2 d−1 . At
very low light levels, photosynthetic carbon gains are insufficient
to meet respiratory requirements and so net carbon gain falls
below the compensation point (Lee et al., 2007; Ralph et al., 2007).
Seagrass plants make some adjustments to the efficiency with
which they capture light and convert it to chemical energy (Ralph
et al., 2007), however the effectiveness of this in improving photosynthetic carbon gain is limited and so plants will also draw on
carbohydrate reserves to support respiratory and growth requirements (Collier et al., 2009). As low light conditions are prolonged,
growth rates slow and plants drop leaves and shoots, thus reducing their abundance (Ralph et al., 2007; Collier et al., in press).
This is relatively easy to demonstrate under constant low light
conditions using experimental mesocosms. In natural conditions,
light is highly variable and so very low light days may be interspersed with days when light levels are increased. This study has
shown that, the percentage of days within specific ranges of Id correlates well with changes in seagrass abundance and can make
a useful environmental indicator of impacts to seagrass meadows. Percentage of days below 3 mol photons m−2 d−1 correlated
strongly with change in seagrass cover at MI and DI. If there are
more than 15.9–18.1% of days below 3 mol photons m−2 d−1 then
some seagrass loss can be expected, while 10.8–16.7% of days
below 3 mol photons m−2 d−1 could lead to more than 50% loss
of seagrass. There were reasonable correlations with days below
5, 4, 2 and 1 mol m−2 d−1 also. At the lower end of this range, i.e.
frequency <2 mol photons m−2 d−1 , these extreme low light conditions are less frequent, and they may not be as useful in applied
situations.
The cause of site variability in mean Id and percent of days below
Id between MI and DI remains unclear, however there are a number
of possibilities. There were slight differences in the species composition at these sites. Early during the monitoring phase there was
around 50% Cymodocea serrulata at MI, whereas DI was dominated
by H. uninervis. This might account for some of the differences when
looking at total change in seagrass abundance. However, most of the
significant loss occurred when H. uninervis was dominant, so there
may be population-level differences in the sensitivity of H. uninervis. It is not uncommon for species specific levels of adaptability
to limit the capacity of plants to respond to changing environmental
conditions. This adaptability is sometimes referred to as phenotypic
plasticity and has been rarely measured in seagrasses. The sites
could also vary in a number of other environmental features not
captured here, including nutrient availability, exposure to salinity, sediment types, etc. Furthermore, this data set was based on
seagrass monitoring at intervals over ∼3 months between 2008
and 2011. Therefore, it captured net change in abundance over the
measured 3 month time intervals. There may have been more temporal variability that was not captured by the sampling frequency
used here. Higher frequency of monitoring and an extended duration of monitoring should reveal if similar thresholds occur among
sites.
4.3. HSat
There was good correlation between change in seagrass abundance and hours of saturating irradiance (HSat ). This metric also
showed very good consistency between sites. HSat differentiates between light limited and light saturated photosynthesis
(Ralph and Gademann, 2005). In contrast, total daily light can be
‘swamped’ by high light intensities that can occur in the middle of
the day but do not necessarily translate into increased photosynthesis as described above (Ralph et al., 2002; Ralph and Gademann,
2005; Beer et al., 2006). Adding to their more precise nature, HSat are
developed from Ek values and are species-specific. This is important because different seagrass species have different minimum
light requirements (Lee et al., 2007). Unfortunately due to complete
seagrass loss at these sites we were unable to develop site-specific
Ek , so they were generated from published and unpublished data
from other studies. Greater than 4 Hsat was required for gains in
H. uninervis. This is similar to findings for Halodule wrightii, in
which 3–8 Hsat was required for seagrass growth (Dunton, 1994).
Higher but more variable Hsat values have been reported for Zostera
species, including 5.2 h (Thom et al., 2008), 7 h (Zimmerman et al.,
1997) and 10 h (Dennison and Alberte, 1985; Dennison, 1987;
Ruiz and Romero, 2001, 2003). Cumulative Hsat has also correlated well with seagrass responses (Lavery et al., 2009), but this
metric could not be explored here as the time interval between
sampling events was not always exactly the same duration and
therefore cumulative indices could be skewed by the variable time
intervals. Although Hsat did result in precise and consistent thresholds, the downside of using Hsat in guidelines, in an applied sense
(e.g. for the environmental regulation), is that ideally you do need
to have locally-generated Ek . If resources are available we would
recommend that future monitoring include the measurement of a
site-specific Ek values in order to generate locally-specific Hsat for
inclusion as light thresholds.
5. Conclusions
This study has identified three light metrics that can be used
to describe changes in seagrass abundance. These thresholds can
be applied as guideline trigger levels by environmental managers
aiming to achieve optimal seagrass growth conditions. However,
achieving these light levels would not necessarily lead to gains in
seagrass, because, for example, there are natural cycles of loss associated with seasonal variability as seen at GI. Furthermore, there are
other site-specific factors that also limit seagrass growth including
nutrient availability. A priority area for future research will be to
investigate the drivers of spatial differences in the thresholds, e.g.
environmental or population differences. A critical step forward
for the application of these indicators for ecological management
will be the development of seasonally-specific thresholds and the
generation of site-specific Ek values.
C.J. Collier et al. / Ecological Indicators 23 (2012) 211–219
Acknowledgements
We thank the Marine and Tropical Sciences Research Facility
(MTSRF), the Reef Rescue Marine Monitoring Program and the Great
Barrier Reef Marine Park Authority for funding to support this work.
We thank A. Giraldo Ospina, JK van Dijk and the many volunteers
who assisted with field work.
References
Beer, S., Mtolera, M., Lyimo, T., Bjork, M., 2006. The photosynthetic performance
of the tropical seagrass Halophila ovalis in the upper intertidal. Aquat. Bot. 84,
367–371.
Birch, W.R., Birch, M., 1984. Succession and pattern of tropical intertidal seagrases
in Cockle Bay, Queensland, Australia: a decade of observations. Aquat. Bot. 19,
343–367.
Campbell, S.J., McKenzie, L.J., 2004. Flood related loss and recovery of intertidal seagrass meadows in southern Queensland, Australia. Estuar. Coast. Shelf Sci. 60,
477–490.
Campbell, S.J., McKenzie, L.J., Kerville, S.P., Bité, J.S., 2007. Patterns in tropical seagrass photosynthesis in relation to light, depth and habitat. Estuar. Coast Shelf
Sci. 73.
Carruthers, T.J.B., Dennison, W.C., Longstaff, B.J., Waycott, M., Abal, E.G., McKenzie,
L.J., Lee Long, W.J., 2002. Seagrass habitats of northeast Australia: models of key
processes and controls. Bull. Mar. Sci. 71, 1153–1169.
Collier, C.J., Lavery, P.S., Masini, R.J., Ralph, P.J., 2007. Morphological, growth and
meadow characteristics of the seagrass Posidonia sinuosa along a depth-related
gradient of light availability. Mar. Ecol. Prog. Ser. 337, 103–115.
Collier, C.J., Lavery, P.S., Masini, R.J., Ralph, P.J., 2009. Shade-induced response and
recovery of the seagrass Posidonia sinuosa. J. Exp. Mar. Biol. Ecol. 370, 89–103.
Collier, C.J., Uthicke, S., Waycott, M., 2011. Thermal tolerance of two seagrass species
at contrasting light levels: implications for future distribution in the Great Barrier Reef. Limnol. Oceanogr. 56, 2200–2210.
Collier, C.J., Waycott, M., Giraldo-Ospina, A. Responses of four Indo-West
Pacific seagrass species to shading. Mar. Pollut. Bull., http://dx.doi.org/
10.1016/j.marpolbul.2011.06.017, in press.
Costanza, R., D’arge, R., De Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg,
K., Naeem, S., O’neill, R.V., Paruelo, J., Raskin, R.G., Sutton, P., Van Den Belt, M.,
1997. The value of the worlds ecosystem services and natural capital. Nature
387, 253–260.
De’ath, G., Fabricius, K., 2010. Water quality as a regional driver of coral biodiversity
and macroalgae on the Great Barrier Reef. Ecol. Appl. 20, 840–850.
Dennison, W.C., 1987. Effects of light on seagrass photosynthesis, growth and depth
distribution. Aquat. Bot. 27, 15–26.
Dennison, W.C., Alberte, R.S., 1985. Role of daily light period in the depth distribution
of Zostera marina (eelgrass). Mar. Ecol. Prog. Ser. 25, 51–61.
Dennison, W.C., Orth, R.J., Moore, K.A., Stevenson, C., Carter, V., Kollar, S., Bergstrom,
P.W., Batiuk, R.A., 1993. Assessing water quality with submersed aquatic vegetation: habitat requirements as barometers of Chesapeake Bay health. Bioscience
43, 86–94.
Devlin, M.J., Brodie, J., 2005. Terrestrial discharge into the Great Barrier Reef Lagoon:
nutrient behavior in coastal waters. Mar. Pollut. Bull. 51, 9–22.
Devlin, M.J., Mckinna, L.I.W., Harkness, P.L. Characterisation and spatial extent
of plume water types in the Great Barrier Reef. Mar. Pollut. Bull.,
http://dx.doi.org/10.1016/j.marpolbul.2012.03.001, in press.
DPC, 2011. Reef Water Quality Protection Plan. For the Great Barrier Reef World
Heritage Area and Adjacent Catchments. Department of Premier and Cabinet,
The State of Queensland.
Dunton, K.H., 1994. Seasonal growth and biomass of the subtropical seagrass Halodule wrightii in relation to continuous measurements of underwater irradiance.
Mar. Biol. 120.
GBRMPA, 2009. Water Quality Guidelines for the Great Barrier Reef Marine Park
Authority.
Grech, A., Coles, R., Marsh, H., 2011. A broad-scale assessment of the risk to coastal
seagrasses from cumulative threats. Mar. Policy 35, 560–567.
Grice, A.M., Loneragan, N.R., Dennison, W.C., 1996. Light intensity and the interactions between physiology, morphology and stable isotope ratios in five species
of seagrass. J. Exp. Mar. Biol. Ecol. 195, 91–110.
Heck, K.L., Carruthers, T.J.B., Duarte, C.M., Hughes, A.R., Kendrick, G., Orth, R.J.,
Williams, S.W., 2008. Trophic transfers from seagrass meadows subsidize
diverse marine and terrestrial consumers. Ecosystems 11, 1198–1210.
Kirk, J.T.O., 1994. Light and Photosynthesis in Aquatic Ecosystems, 2nd ed. Cambridge
University Press.
Lavery, P.S., Mcmahon, K., Mulligan, M., Tennyson, A., 2009. Interactive effects of
timing, intensity and duration of experimental shading on Amphibolis griffithii.
Mar. Ecol. Prog. Ser. 394, 21–33.
219
Lee, K.-S., Park, S.R., Kim, Y.K., 2007. Effects of irradiance, temperature, and nutrients on growth dynamics of seagrasses: A review. J. Exp. Mar. Biol. Ecol. 350,
144–175.
Longstaff, B.J., Dennison, W.C., 1999. Seagrass survival during pulsed turbidity
events: the effects of light deprivation on the seagrasses Halodule pinifolia and
Halophila ovalis. Aquat. Bot. 65, 101–121.
Marbà, N., Holmer, M., Gacia, E., Barròn, C., 2006. Seagrass beds and coastal biogeochemistry. In: Larkum, A.W.D., Orth, R.J., Duarte, C.M. (Eds.), Seagrasses: Biology,
Ecology and Conservation. Springer.
Masini, R.J., Manning, C.R., 1997. The photosynthetic responses to irradiance and temperature of four meadow-forming seagrasses. Aquat. Bot. 58,
21–36.
McKenzie, L., Collier, C.J., Waycott, M. Reef Rescue Marine Monitoring Program:
Intertidal seagrass, annual report for the sampling period 1st September
2010–31st May 2011. Fisheries Queensland, Cairns, in press.
McKenzie, L.J., Campbell, S.J., Roder, C.A., 2001. Seagrass-Watch: Manual for Mapping
and Monitoring Seagrass Resources by Community (Citizen) Volunteers. QFS,
Northern Fisheries Centre, p. 100.
McKenzie, L.J., Lee Long, W.J., Roelofs, A., 1998. Port of Mourilyan seagrass
monitoring—first 4 years. EcoPorts Monograph Series No. 15 Ports Corporation
of Queensland, p. 30.
Orth, R.J., Carruthers, T.J.B., Dennison, W.C., Duarte, C.M., Fourqurean, J.W., Heck, K.L.,
Hughes, A.R., Kendrick, G.A., Kenworthy, W.J., Olyarnik, S., Short, F.T., Waycott,
M., Williams, S.L., 2006. A global crisis for seagrass ecosystems. Bioscience 56,
987–996.
Ralph, P.J., Durako, M.J., Enriquez, S., Collier, C.J., Doblin, M.A., 2007. Impact of light
limitation on seagrasses. J. Exp. Mar. Biol. Ecol. 350, 176–193.
Ralph, P.J., Gademann, R., 2005. Rapid light curves: a powerful tool to assess photosynthetic activity. Aquat. Bot. 82, 222–237.
Ralph, P.J., Polk, S.M., Moore, K.A., Orth, R.J., Smith, W.O.J., 2002. Operation of the xanthophyll cycle in the seagrass Zostera marina in response to variable irradiance.
J. Exp. Mar. Biol. Ecol. 271, 189–207.
Rasheed, M.A., 2004. Recovery and succession in a multi-species tropical seagrass
meadow following experimental disturbance: the role of sexual and asexual
reproduction. J. Exp. Mar. Biol. Ecol. 310, 13–45.
Ruiz, J.M., Romero, J., 2001. Effects of in situ experimental shading on
the Mediterranean seagrass Posidonia oceanica. Mar. Ecol. Prog. Ser. 215,
107–120.
Ruiz, J.M., Romero, J., 2003. Effects of disturbances caused by coastal constructions on
spatial structure, growth dynamics and photosynthesis of the seagrass Posidonia
oceanica. Mar. Pollut. Bull. 46, 1523–1533.
Sheppard, J.K., Marsh, H., Jones, R.E., Lawler, I.R., 2010. Dugong habitat use in
relation to seagrass nutrients, tides, and diel cycles. Mar. Mamm. Sci. 26,
855–879.
Sofonia, J.J., Unsworth, R.K.F., 2009. Development of water quality thresholds during
dredging for the protection of benthic primary producer habitats. J. Environ.
Monitor. 12, 159–163.
Thom, R., Southard, S., Borde, A., Stoltz, P., 2008. Light requirements for growth
and survival of Eelgrass (Zostera marina L.) in Pacific Northwest (USA) estuaries.
Estuar. Coast. 31, 969–980.
Touchette, B.W., 2007. Seagrass–salinity interactions: physiological mechanisms
used by submersed marine angiosperms for a life at sea. J. Exp. Mar. Biol. Ecol.
350, 194–215.
Udy, J.W., Dennison, W.C., Long, W.J.L., McKenzie, L.J., 1999. Responses of seagrass to nutrients in the Great Barrier Reef, Australia. Mar. Ecol. Prog. Ser. 185,
257–271.
Unsworth, R.K.F., Cullen, L.C., 2010. Recognising the necessity for Indo-Pacific seagrass conservation. Conserv. Lett. 3, 63–73.
Waycott, M., Duarte, C.M., Carruthers, T.J.B., Orth, R.J., Dennison, W.C., Olyarnik, S.,
Calladine, A., Fourqurean, J.W., Heck, K.L., Hughes, A.R., Kendrick, G.A., Kenworthy, W.J., Short, F.T., Williams, S.L., 2009. Accelerating loss of seagrasses
across the globe threatens coastal ecosystems. Proc. Natl. Acad. Sci U.S.A. 106,
12377–12381.
Waycott, M., Longstaff, B.J., Mellors, J., 2005. Seagrass population dynamics and
water quality in the Great Barrier Reef region: a review and future research
directions. Mar. Pollut. Bull. 51, 343–350.
Waycott, M., McKenzie, L.J., Mellors, J.E., Ellison, J.C., Sheaves, M.T., Collier, C.,
Schwarz, A.-M., Webb, A., Johnson, J., Payri, C.E., 2011. Vulnerability of mangroves, seagrasses and intertidal flats in the tropical Pacific to climate change.
In: Bell, J.D., Johnson, J.E., Hobday, A.J. (Eds.), Vulnerability of Tropical Pacific
Fisheries and Aquaculture to Climate Change. Secretariat of the Pacific Community.
Waycott, M., Mcmahon, K., Mellors, J., Calladine, A., Kleine, D., 2004. A
Guide to Tropical Seagrasses of the Indo-West Pacific. James Cook
University.
Zimmerman, R.C., Kohrs, D.G., Steller, D.L., Alberte, R.S., 1997. Impacts of CO2 enrichment on productivity and light requirements of eelgrass. Plant Physiol. 115,
599–607.