Eldridge et al.	Seagrass Stress Response Model
Seagrass Stress Response Model: The importance of Light, Temperature, Sedimentation
and Geochemistry
Peter M. Eldridge and James E. Kaldy
Western Ecology Division, US EPA, Newport, Oregon
Adrian B. Burd
Department of Marine Sciences, University of Georgia, Athens
Abstract. The objective of our modeling was to better understand the relationship between
seagrass and water-column and sediment stressors (i.e., light, organic and particle sedimentation,
sediment nutrients and sulfides). The model was developed and optimized for sediments in
Thalassia testudinum seagrass beds of Lower Laguna Madre, Texas, USA and is composed of a
plant sub-model and a sediment diagenetic sub-model. Simulations were developed for a natural
stressor (harmful algal bloom) and an anthropogenic stressor (dredging event). The harmful
algal bloom (HAB) was of limited duration and the simulations showed no effect of the algal
bloom on biomass trends but did suggest that sediment sulfides could inhibit growth if the bloom
duration and intensity were greater. The dredging event resulted in sedimentation of a layer of
organically rich material and reduction of canopy light for a period of months. The simulations
suggested that the seagrass could have recovered from the effects of light but residual effects of
high sulfides in the sediments would make the region uninhabitable for seagrasses for up to 2.5
years. These modeling exercises point out the importance of using a geochemical model to
evaluate the impact of reduced light and enhanced organic loading from both natural and
anthropogenic stressors to seagrass.

-------
Eldridge et al.	Seagrass Stress Response Model
1. Introduction
Worldwide >90,000 ha of submerged aquatic vegetation (SAV) habitat has been lost from
estuarine habitat (Short and Wyllie-Echevarria, 1996) primarily as a result of natural and
anthropogenic aquatic stressors. The loss of these highly valued habitats is affecting estuarine
ecosystem health and function. The importance of seagrass habitat to ecosystem function and
biodiversity has convinced local and national governmental agencies to develop programs to
protect seagrass habitats from degradation (U.S. EPA, submitted). As a result academic and
national research laboratories have developed numerical models that can be used to evaluate the
response of seagrass to natural and anthropogenic stressors. These models are viewed as a
mechanism for relating present and future levels of nutrient enrichment, sediment input, and
other aquatic stressors to trends in seagrass production and bio mass. Models can be used to
evaluate regionally and locally important physical, biological and geochemical processes that
interact with seagrass and this information can be used to set priorities on how best to ameliorate
detrimental effects of aquatic stressors to the seagrass. Here we present a seagrass stress-
response model for the tropical seagrass Thalassia testudinum ("Thalassia model"). We use the
Thalassia model to examine how this seagrass may respond to both natural and anthropogenic
stressors in Laguna Madre, Texas, USA.
Seagrasses interact physically, biologically and geochemically with both the water column
and the sediments (Fig. 1). These interactions are often complex and location dependent. Many
of the interactions are described in the primary literature. It is not possible or even desirable to
include all the physical, biological, and geochemical processes that effect seagrass in a single
model, instead it is more important to define and incorporate the key processes that are expected

-------
Eldridge et al.
Seagrass Stress Response Model
to shape seagrass population dynamics. In the following we develop the conceptual
biogeochemical framework used for this model.
Light
Correal
Wave orfcital
motion
Epiphytes
/•-
TSS
Oxidized
%AK -W <**>•
Reduced ^
AVS .	'
	. HS-
Figure 1. Interactions of seagrass with its biological, physical and geochemical environment. Seagrass
illustration revised from Short 1989.
Seagrasscs arc "ecosystem engineers" (Koch 2001) and consequently alter the areas they
colonize through feedback mechanisms. Similar to top-down and bottom-up control of
phyto plank ton populations in estuaries, seagrass production responds to canopy-water column
and sediment-root/rhizome interactions (Hemminga and Duarte. 2000) (Fig. 2).
Seagrasses are most commonly limited by light (Zimmerman et al., 1995) and nutrients
(Orth, 1977; Alcoverro et al., 1997), although C02 limitation can occur in isolated quiescent
3

-------
Eldridge et al.	Seagrass Stress Response Model
habitats. Most declines in seagrass are related to decreased light availability (Short and Wyllie-
Echeverria, 1996). Low light stress is manifested in a cascade of effects that are intimately
related to the geochemistry (Fig. 2).
Phytotoxic substances produced by anaerobic bacterial metabolism can limit seagrass
distributions and may be determinates of anthropogenic effects on seagrass distribution (Carlson
et al, 1994; Koch, 2001). A poorly understood yet key component controlling sediment
geochemistry is the input of organic particles that are available for remineralization. Complex
hydrodynamic interactions can cause seagrass beds to be either sources or sinks for particulate
matter (Nepf and Koch, 1999).
Sediment sulfides are the most important of the potentially toxic metabolites from sediment
diagenesis of organic matter (Koch, 2001). Experimental work has shown that anaerobic
conditions inhibit internal carbohydrate transport (Zimmerman and Alberte, 1996) and that
sulfides reduce seagrass photosynthesis (Goodman et al., 1995; Holmer and Bondgaard, 2001).
To counteract sulfide concentrations in the rhizosphere, seagrasses transport photosynthetically
produced O2 through lacunae to the roots (Kraemer and Alberte 1993). Thus transport
mechanism probably evolved to support aerobic root respiration, but excess 02 diffusing from
the roots has the additional benefit of oxidizing sulfide to non-toxic sulfate (Caffrey and Kemp,
1991). Model simulations indicate that O2 diffused from the seagrass root system effectively
reduce sulfide concentrations (Eldridge and Morse, 2000); additionally field experiments support
these model results (Lee and Dunton, 2000)
The US EPA as part of the national effort to protect sensitive habitats from degradation is
developing seagrass nutrient stress-response models. The Thalassia model is a prototype
developed using data from a regional program (funded by the US Corps of Engineers, Galveston,
4

-------
Eldridge et al.	Seagrass Stress Response Model
TX) and undertaken by local universities and agencies to evaluate the effect of dredging on
seagrasses in Laguna Madre, Texas, USA. While the plant model and the geochemical model of
seagrass processes have been previously published (Burd and Dunton, 2001; Eldridge and
Morse, 2000), this modeling study combines these models in a single coupled plant-sediment
model that describes the effect of the interactions of water-column and sediment stressors.
Unstressed seagrass
Stressed seagrass
High Light Penetration
Bioturbation &
Bio-irrigation
High Oxygen
Nutrient
t
iSulfidej
Nutrients
NH4*:F04
No
Toxic
effect
Low Light Penetration
• -O O •
Epiphytes
hytoplanktoii
wwifpaiiia

Low
Nutrient
uptake
v 7/ Sediment
Nutrients
Toxic sulfide
effect and
reduced
Roots/K litanies
Figure 2. Seagrass sediment geochemical interactions. Part of the photosynthetic Oj production is transported
through lacunae to roots and sediments. The 02 supports aerobic respiration in the roots and diffuses into sediments
to re-oxidize toxic sulfides to non-toxic sulfate. A) During periods of high photosynthesis, seagrass can reduce
sulfides to non-toxic levels. B) During periods of low photosynthetic activity, less 02 is transported in the root zone
and the plant is exposed to toxic sulfides.
5

-------
Eldridge et al.	Seagrass Stress Response Model
2. Model development
As part of a large multi-investigator project, we developed a series of seagrass models for
Laguna Madre, Texas (Fig. 3). Lower Laguna Madre (LLM) is a shallow, micro-tidal, wind-
mixed estuary that is generally not subjected to anoxia or hypoxic events. Because of restricted
access to the Gulf of Mexico and low fresh water inflows, the residence time of water and
nutrients is long. The system's nutrient status is oligo- to mesotrophic and as a result there is
little epiphyte interaction with seagrass. Thalassia testudinum (turtle grass) is the dominate
species in the southern part of LLM and is the subject of this study, T. testudinum is a broad leaf
seagrass with a canopy extending about 0.5 m from the bottom, and 80-90% of biomass in the
below ground root and rhizome structure. These sub-tidal beds are not subjected to desiccation
and wind-generated waves are rarely large enough to cause physical damage except during
hurricanes. We explicitly deal with sediment resuspension and deposition; however, sediment
transport issues were addressed with hydrodynamic and sediment transport models described
elsewhere (Teeter et al., 2001).
The Thalassia model explicitly focuses on how light and sediment diagenesis influence
seagrass biomass. To examine the interaction of stressors and seagrass we simulated two time
periods; 1) during an interval when there were two harmful algal blooms in LLM (natural
stressor), and 2) during a dredging event (anthropogenic stressor).
During both of these time periods, we collected time-series data on the physical properties of
the system most likely to affect seagrass physiology. These measurements included continuous
measurements of underwater irradianee at the seagrass canopy and temperature at a nearby site
(Kaldy and Dunton, 2000) and sediment geochemical data (Eldridge and Morse, 2000).
6

-------
Eldridge et al.
Seagrass Stress Response Model
Lower Laguna Madre
Hypersaline negative estuary
(evaporation is greater than freshwater
input),
670 km2 in which 70% of area is
covered with seagrass,
Micro-tidal with only two openings to
the Gulf of Mexico,
Average depth is ~1 m,
Oligotrophic to mesotrophic levels of
primary production,
Few epiphytes on seagrass, drift
macroalgal mats causing bare areas.
Site used for model calibration
Dredging event used for validation
Figure 3. LLM was the site for the development work on the stress response model for Thalassia testudinum.
The location map shows the position of the data collection site used for model calibration, and for the model
validation during a dredging event. See Kaldy and Dunton (2000) for additional site description. Source of map is
www.glo.state.tx.us/gisdata/gisdata.html
The factors that can inhibit seagrass production explicitly incorporated in the model are
reduced light, nutrient limitation and increased phytotoxin concentrations. The effect of light
was determined with a plant model (Burd and Dunton, 2001), while the production of nutrients
and the sulfide phytotoxin were estimated using a sediment diagenetic model (Eldridge and
Morse, 2000). We incorporated nutrient uptake kinetics for both above- and below-ground
tissues (Lee and Dunton, 1999) and inhibition kinetics for toxic sulfide and ammonium
concentrations (Carlson et al., 1994; Pulich, 1989)
__ Land Cut
to Upper Laguna
Location
Mansfield Pass
30 Kilometers
i

-------
Eldridge et al.
Seagrass Stress Response Model
2.1 Plant model
The governing equations used in the plant model are shown below. The Smith-Tailing
function (Tailing 1957) was used to estimate gross primary production (P(I)) (EQ. 1) while the
derivative function (EQ. 2) (Burd and Dunton, 2001) was used to estimate changes in above
ground biomass (Ca). The parameter definitions for these equations are shown in Table 1. T.
testudinum photosynthetic parameters were from Herzka and Dunton (1997).
The Thalassia model contained the below ground compartment implicitly. The two years
of below ground biomass data available for this analysis showed no seasonal trends (Kaldy and
Dunton, 2000). Below ground losses were interpolated from an inverse analysis of the root zone
for a summer and winter analysis (Kaldy and Eldridge, unpubl. data). The inverse analysis is a
constrained optimization in which gross primary production, shoot, rhizome net production,
above and below ground respiration and other data are used to estimate unknown flows such as
DIC, DOC, and detritus in the below ground compartment.
2,2 Diagenetic model
There are many pathways by which organic matter may be oxidized and the oxidation may
form various organic intermediates (Lovley and Phillips, 1989; Kristensen and Blackburn, 1987;
Postma and Jakobsen, 1996). To keep the analysis manageable, only one generalized biogenic
reaction for each oxidant is included and we assume the reaction goes to completion (CO2 and
water). The relationships between growth and substrate concentration are incorporated in the
(1)
(1- r)P(/)^rCa 1
(2)
8

-------
Eldridgeetal.	Seagrass Stress Response Model
model using hyperbolic Monod relationships, while inhibition is modeled with a hyperbolic
feedback (see equations 58 through 62 in Boudreau, 1996).
The diagenetic model has 13 geochemical compartments encompassing solid and porewater
organic and inorganic species that are important in Laguna Madre sediment diagenesis (Table 2).
The model was calibrated with vertical sediment geochemical profiles (2 cm increments) from T.
testudinum beds and adjacent bare sediment sites (Eldridge and Morse, 2000). The model includes
organic matter loading, diffusion processes (molecular, bioturbation, and irrigation), and advective
processes (burial and porewater flow) as well as important geochemical interactions (see Eldridge
and Morse (2000) for details).
Table 1. Symbols and parameters used in Thalassia model.
Parameter
Symbol
value
Reference
time
t
d

Irradiance at canopy
depth
1(2)
Continuous measurements by
LiCor spherical (Ait) light sensor
Kaldy and Dunton (2000)
Carrying capacity
K
gdw m 2

DOC release
8
As a fraction of production
Kaldy and Eldridge (in prep)
Temperature
difference
AT
Current temperature and reference
temperature (31) at which
measurements were made (°C)
Burd and Dunton (2001)
Max. Photosynthesis
P max
190 (|imol O2 gdw"1 hr"1)
Herzka and Dunton (1997)
Initial slope of P vs I
curve
a
0.4 (|imol O2 mg chl"1 hr-1)/(p.mol
photon m"2 s"1)
Herzka and Dunton (1997)
Above ground plant
respiration
Ra
35 (p.mol O2 gdw"1 h"1)
Herzka and Dunton (1997)
Above ground plant
mortality
Ma
0.0052 (d1)
Kaldy and Eldridge (in prep)
Our goal in the study was to determine how seagrass modified geochemical sediment profiles.
The model simulates root-zone fluxes and leaf-detrital inputs. Comparing the model results with
the actual sediment chemical profiles provides an assessment of seagrass-geochemical interactions.
9

-------
Eldridge et al.
Seagrass Stress Response Model
Removal of the seagrass fluxes in some model simulations then provides a means to quantify
modifications of sediment chemistry caused by the seagrass. The model simulations suggest that
between 25 to 50% of the photosynthetic 02 had to be transported below ground to account for the
observed chemical profiles (Fig 4., not all simulations are shown).
Table 2, Solid and dissolved species used in the sediment diagenesis model. The model assumes an oxidation
state of zero for organic material. C:N:P of surface flux is that of seagrass above ground tissues. Root zone C:N:P flux
is that of the below ground tissues.
OM1
OM2
DOM
02
NOj1"
NHj
SO,2"
TS
Fe(OH)3
Fe2+
FeS,
DIC
ALK
HS"
H2S
HC03"
CO,2"
ALKC
PH
Explicit species
labile organic matter
refractory organic matter
dissolved organic matter
oxygen
nitrate
ammonia
sulfate
total sulfides
amorphous
ferrous
pyrite
dissolved inorganic carbon
total alkalinity (treated as a species)
Implicate (calculated) species
sulfide species
sulfide species
sum of hydrated and
carbonic acid
bicarbonate
carbonate
carbonate alkalinity
unhydrate<
Solid
Solid
Porewater
Porewater
Porewater
Porewater
Porewater
Porewater
Solid
Porewater
Solid
Porewater
Porewater
Porewater
Porewater
Porewater
Porewater
Porewater
Porewater
Porewater
3. Model results and discussion
We combined the plant and sediment diagenetic model in a coupled model that solved the
seagrass and sediment geochemical equations simultaneously. This allowed us to use the outputs
10

-------
Eldridgc et al.	Seagrass Stress Response Model
of the seagrass model - detritus, dissolved organic matter, etc - as inputs to the sediment model.
The simultaneous solutions to both models allowed us to formulate feedbacks between sulfides,
ammonium, and seagrass production. We then used the combined model to simulate the response
of seagrass to a natural and anthropogenic aquatic stressor.
10
15l*->
-*rr-
-u*.
GT


0 1 2 0 50 100 0 5 10 0 200 400 0 10 20 0 500 1000 0 200 400
POC (%) [DOM] [DIG] [O,] [NOj-1 [NH/] [R.S]
Figure 4. Thalassia testudinum site sediment geochemistry. Estimated flux of DOM, POC, and NH»+, flux to the
rootzone (shaded area) is estimated using fractions of primary production from the T. Testudium optimization. Red
line simulation has no root zone fluxes (case 1), green line is 50% of the estimated root zone flux (case 2), and the
black line is the model results with the estimated root zone fluxes (case 3). Asterisks are data from the calibration site
in September 1996. [DOC] and [DIC] as mM, POC in percent solid, [02],[N0,1, [NH.1, and fH2S] (J.M.
3.1. "Brown tide" bloom - a natural aquatic stressor.
LLM is subjected to harmful algal blooms (HAB) of Aureoumbra lagunensis (Brown tide)
which is periodically advected through the land cut from Upper Laguna Madre (ULM) (Kaldy
11

-------
Eldridge et al.	Seagrass Stress Response Model
and Dunton, 2000). The duration and intensity of these HABs are considerably greater in ULM
where similar modeling studies demonstrated a reduction in the seagrass Halodule wrightii
production and bio mass (Burd and Dunton, 2001). Field studies (Onuf, 1996) also document the
long-term impact of brown tide on seagrasses in ULM.
The impact of short term, episodic blooms on seagrass production is not known. Two short-
term events occurred in the spring and winter of 1995 and 1996 (HAB 1 and HAB 2) during which
Kaldy and Dunton (2000) sampled water-column chlorophyll a (Fig. 5). We used this time-series
to compute organic loading to the sediment caused by the HAB. There is only one data point for
the high chlorophyll levels during March 1995 that can be corroborated by increased light
attenuation for about a week. During the HAB 2 bloom, similar chlorophyll a concentrations were
measured at a site 2 km away from the calibration site (Kaldy and Dunton 2000, Kaldy, pers. obs).
Chlorophyll a concentrations remained high for about a month longer at this reference site than
was measured at the Thalassia collection site, suggesting that our estimate of the bloom's duration
was conservative.
Both surface irradiance and the seagrass canopy light field were measured throughout 1995
and 1996 using LI-COR quantum sensors of PAR (photosynthetically active radiation 400-700
nra). An advantage of collecting canopy level light data is that it includes natural temporal
stochasticity encountered by the plant. The disadvantage is that the sensors can be fouled by
drift macroalgae and attached microalgae. This was the case on two occasions during the
collection of the canopy light record (Herzka and Dunton, 1998).
The HAB affects both the underwater light environment and the flux of organic matter to the
sediments. Consequently, the bloom influences seagrass directly through reduced photosynthesis
and indirectly through geochemical cycling of settling organic material (settled brown tide cells).
12

-------
Eldridge et al.	Seagrass Stress Response Model
Enhanced organic matter flux is explicitly modeled; hence, the coupled plant/sediment
geochemical model may be uniquely capable of examining the effect of the HABs on Thalassia.
35
30
25
20
00
A
<3
V»
1995	1996
Figure 5. Water column chlorophyll a values from the calibration site measured monthly between February 1995
and January 1997.
The actual rate of bloom sedimentation to the bottom is not known. The small cell size and
wind-mixing in the shallow lagoon limit the sedimentation of this alga (DeYoe and Suttle, 1994).
However, the limited exchange with the ocean and minimal grazing (Buskey and Hyatt, 1995;
Buskey et al., 1997) suggest that much of the A. lagunensis production eventually does sink to
the bottom. We ran a series of simulations using available data in concert with A. lagunensis
morphological and physiological data from DeYoe and Suttle (1994). During a dense bloom of
A. lagunensis C:N and C:Chl were 28 and 176, respectively, and the growth rate was 0.45 d"1.
We used these data in the coupled seagrass diagenetic model to estimate the effect of the two
HABs on seagrass in Laguna Madre.
13

-------
Eldridge et al.
Seagrass Stress Response Model
Surface Light
Canopy Light
Probe was fouled
with raacroalgae
HABS2
HABS 1
Figure 6. Surface photon flux density (PFD) from a fixed monitoring station 6 km from the calibration site and
PFD at canopy level at the calibration site.
During the development of the HAB simulation we noted that different formulations for
temperature variations made a large difference in the simulation results. This was unexpected
and suggests it would be instructive to provide a comparison of results using modeled
temperature (a sine wave function) and actual temperature measurements (Kaldy and Dunton,
2000). Both simulations were consistent with the biomass data (given the standard deviation of
samples) and showed the expected seasonal variations, but were still substantially different from
each other. The mean biosmass data in the time-series varied by about 100 gdw m~2, the
simulated biomass with measured and a sine wave function for temperature varied by 70 and 55
gdw m~2 respectively. While none of the models exactly simulate the biomass variations in the
data, the actual temperature model included significantly more of this variation (Fig. 7). Overall,
14

-------
Eldridge et al.	Seagrass Stress Response Model
the natural seasonal variation was greater than any HABs effect. HAB-1 occurred during the
winter when growth rates were reduced (Kaldy and Dunton, 2000) and had a shorter duration and
predictably had no obvious effect on seagrass biomass. HAB-2 may have caused an early onset
of the seasonal decline in biomass but this is not distinguishable from normal variation between
years.
220
Biomass data
model (temp data)
model (sin wave)
200
180
160
'£
I 140
a
120
i
£ 100
1
80
60
1 X
40
HAB 1	HAB 2
- 	-^0		""V"*
1995	1996
Figure 7. Comparison of biomass data and modeling results using time-series temperature data and using a sine
wave approximation of temperature.
We also examined the influence of the brown tide on modeled seagrass habitat sediment
biogeoehemistry. Assuming that sufficient HABs biomass settled and was metabolized, seagrass
production could be effected by the additional sulfide production. Seagrass is not sensitive to
sulfide concentrations below 200 [imol 1"' in the root zone (Pulich, 1989) and production is not
strongly inhibited until mmol concentrations (Carlson et al, 1994). The model predicted that the
amount of material settled from the HABs in concert with material entering the sediments as leaf,
root, rhizome and other detritus based on our earlier model calibration were unlikely to cause
15

-------
Eldridge et al.	Seagrass Stress Response Model
sulfide concentrations to reach even 200 jimol l"1. Thus, model results suggest that neither of
the HABs blooms in LLM during 1995 had an impact on Thalassia habitats. However, the
higher sulfide concentration due to the bloom continued for several months suggesting that
concentrated blooms of greater duration could reduce Thalassia biomass.
a
a
-------
Eldridge et al.	Seagrass Stress Response Model
placement could change seagrass distribution, biomass and production patterns prompted local,
state, and federal agencies including the US Army Corps of Engineers to assess alternative
dredging practices. As part of this, we developed a series of simulations using the Thalassia
model to examine loss and recovery of seagrass associated with a dredging event.
3.2.1.	Validation of the combined plant-sediment diagenetic model.
To provide a verification of the combined plant/sediment model, several Thalassia habitats
were monitored during a dredging event in late 1998. The validation process required that the
model predict the outcome of the dredging event on seagrass biomass and rhizosphere
geochemistry from light data and the thickness and composition of the depositional layer from
dredged material disposal. Because of logistical problems (exact placement areas, sensor burial,
etc), there was no single data set of continuous light or biomass, and sediment geochemistry.
Consequently, for this simulation we combined light and initial biomass data from other sites
with the sediment geochemistry of a site buried with 7 to 10 cm of dredge materials. Much of
the leaf material at this site remained above the sediment layer while the roots and rhizomes were
buried to 9-12 cm.
Data for the model validation was collected for about 1 year. Trends in measured biomass
and model results were similar (Fig. 9) although the model was not able to simulate the rapid
reduction in biomass in the first 2 months, possibly as a result of sediments settling on leaves - a
process not considered in the model. The model was able to simulate the 1800 to 2600
concentration of sulfides and the 600 to 1000 p.M ammonium (shown as the first year of the
futures analysis, Fig 10).
3.2.2.	Simulation of loss and recovery of seagrass after a dredging event—a future analysis.
17
-------
Eldridge et ai.	Seagrass Stress Response Model
The objectives of these analyses were to predict trends in bio mass after a dredging event and
to identify water-column or sediment stressors that inhibit seagrass. The assumption of any
multiple year futures analysis is that annual cycles of physical, biotic and geochemical processes
are similar. Unusual storm events, and other unexpected events, such as prolonged cloud cover,
can make the simulations meaningless. Likewise, because models are simplifications of real
systems, model results must be interpreted based on local knowledge. Our simulations
encompassed 3 rates of organic matter reactivity and 3 root-zone depths in order to provide an
envelope of possible responses.
Based on our assumption of similar seasonal trends our light time-series started with the
record from the dredging event and was followed with annual light time-series from a nearby site
unaffected by the dredging event. For additional years of the simulation we simply added the
same annual time-series multiple times. As is typical in any underwater time-series of
measurements, there were time periods for which we had no data. Gaps were filled by
connecting measurements from the last date collected to the first measurement for the next set
(Fig. 10A).
Model simulations show a continued reduction in bio mass (Fig. 10B) even though light
levels had returned to a normal annual cycle of irradiance (Fig 10A). Microbial metabolism of
organic material in the dredged sediments, combined with the reduction in plant productivity
(i.e., small root zone O2 flux), produced lethal concentrations of sulfide in the root zone (Fig.
IOC).
18
-------
Eldridge et al.
Seagrass Stress Response Model
250
200
150
e
s
tJ
00
£ 100
S3
E
o
a
50
-50
Figure 9. Comparison of observed (symbols) and simulated seagrass biomass. Biomass data were collected before,
during and after a dredging event in LLM. Initial biomass was taken from a nearby site, because of the difficulty of
making an apriori prediction of dredge deposition.
Simulations suggest that sediment conditions would not be suitable for seagrass re-
colonization until 2.5 years after the dredge event. However, the model is not formulated to
account for seagrass re-colonization (e.g. seed germination); consequently, we interpret the
"recovery" as a return to conditions conducive to seagrass colonization. We expect to
incorporate seagrass recruitment mechanisms in future models. Ammonium concentrations
increase rapidly to levels found in the validation sediment cores and followed a similar duration
concentration profile found in the sulfides. The literature is not clear on what the toxic levels of
ammonium are (Van Katwijk et al., 1997), but these levels may also have been toxic to seagrass.
19
-------
Eldridge et al.
Seagrass Stress Response Model
cu
S S
o s
PQ 00
tn
X
60
40
20
0
200
100
0
X
2:
™1			1	1-
J	LLl	L
r 	i	t	 i i 	t	 i J \ r	 » —
Red 5 cm root zone
''-t—	T	1	X	1	
B
Blue 9 cm root zooe


Green 7 cm roo* zone
iiiti
t	!	r~
2000
1000
AJOJAJOJAJOJAJOJAJ
Year 1
Year2
Year 3	Year 4
Figure 10, Four year time series of (A) canopy light, and simulated (B) Thalassia above ground biomass, (C)
sulfide (H2S signifies all porewater sulfides), and porewater (D) ammonium. The initial year of the simulations is
the validation while years 2 through 4 are the futures analysis. Red, green, and blue lines are simulations run with
root zones of 5, 7 and 9 cm depth. For each root zone depth we varied the amount of labile sediments from 40 to
50% of total sediments (KlabU,,= 25 y"! and K«fac,ory= 0.12 y"1).
4. Conclusions
Seagrass through their architecture arid anatomy are successful bioengineers that alter their
physical and geochemical environment to meet its physiological requirements. Our modeling
simulations of Thalassia testudinum suggest that this seagrass can withstand exposure to an
episodic natural stressor-a HAB event, but was extinguished by a chronic anthropogenic stressor
-a dredging event. Short-term exposure to a HAB caused no obvious response in biomass
trends, although longer exposure and higher bloom concentrations might eventually cause a
20
-------
Eldridge et al.	Seagrass Stress Response Model
decline. Exposure to dredged material eliminated the seagrass due to development of high
sediment sulfide concentrations caused by metabolism of organically-rich dredged material.
This example was extreme and would probably only occur close to the actual dredging event;
however, it illustrates the importance of the interaction of seagrass productivity with sediment
geochemical processes. An important feature the geochemical model was the time lag between
stressor exposure (i.e. settling phytoplankton or sediment), and the increase in sulfide
concentrations. In each simulation (HAB or dredge deposition) sulfide concentrations remain
elevated for a period considerably longer than the actual exposure to the aquatic stressor. This
suggests that the sediment will contain a record of aquatic stressor events that may last for
months to several years.
Acknowledgement.
The information in this document has been funded in part by the U.S. Environmental Protection
Agency. It has been subjected to the Agency's peer and administrative review, and it had been
approved for publication as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use. We thank Cheryl Brown, an EPA
Post-doctoral fellow for her comments and editing of this manuscript. We also thank John Morse
for the development work he did on the original diagenetic model.
5. References
Alcoverro, T.D., J. Romero, C.M. Duarte, and N. Lopez, Spatial and temporal variations in
nutrient limitation of seagrass Posidonia oceanic a growth in the NW Mediterranean.
Mar. Ecol. Prog. Ser. 146, 155-161, 1997
-------
Eldridge et al.	Seagrass Stress Response Model
Burd, A.B. and K.H. Dunton, Field verification of a light-driven model of biomass changes in
the seagrass Halodule wrightii. Mar. Ecol. Prog. Ser., 209, 85-98, 2001
Boudreau. B. P., A method-of-lines code for carbon and nutrient diagenesis in aquatic sediments,
Computers & Geosciences, 22, 479-496, 1996
Buskey, E.J. and C.J. Hyatt, Effects of the Texas "brown tide" alga on planktonic grazers, Mar.
Ecol. Prog. Ser., 126, 285-292, 1995.
Buskey, E.J., P.A. Montagna, A.F. Amos and T.E. Whitledge, The initiation of the Texas brown
tide algal bloom: disruption of grazer populations as a contributing factor, Limnol. and
Oceanogr., 42,1215-1222, 1997.
Caffrey, J. M., and W.M. Kemp, Seasonal and spatial patterns of oxygen production, respiration
and root-rhizome release in Potamogeton perfoliatus L. and Zostera marina L., Aquat.
Bot. 40, 109-128, 1991.
Carlson. P. R., L. A. Tarbro and T. R. Barber, Relationship of sediment sulfide to mortality of
Thalassia testudinum in Florida Bay, Bulletin of Marine Science, 54, 733-746, 1994.
DeYoe, H.R., and C.A. Suttle, The inability of the Texas "Brown tide" alga to use nitrate and the
role of nitrogen in the initiation of a persistent bloom or this organism, J. Phycol,. 30,
800-806, 1994.
Eldridge, P.M. and J.W. Morse, A diagenetic model for sediment-seagrass interactions, Marine
Chemistry, 70, 89-103, 2000.
Goodman, J.L., K.A. Moore, and W.C. Dennison, Photosynthetic responses of eelgrass (Zostera
marina L.) to light and sediment sulfide in a shallow barrier island lagoon, Aquatic
Botany, 50, 37-47, 1995.
-------
Eldridge et al.	Seagrass Stress Response Model
Hemminga, M.A. and C.M. Duarte, Seagrass Ecology, Cambridge University Press, Cambridge,
UK, pp.298.
Herzka, S.Z. and K.H. Dunton, Seasonal photosynthetic patterns of the seagrass Thalcissia
testudinum in the western Gulf of Mexico, Mar. Ecol. Prog. Ser. 152, 103-117, 1997.
Herzka. S.Z. and K.H. Dunton, Light and carbon balance in the seagrass Thalcissia testudinum:
evaluation of current production models. Marine Biology 132:711-721, 1998.
Holmer, M.H. and e.J. Bondgaard, Photosynthetic and growth response of eelgrass to low
oxygen and high sulfide concentrations during hypoxic events. Aquatic Botany 70, 29-38,
2001.
Kaldy, J.E. and K.H. Dunton, Above- and below-ground production, biomass and reproduction
ecology of Thalassia testudinum (turtle grass) in a subtropical coastal lagoon, Mar. Ecol.
Prog. Ser., 193, 271-283, 2000.
Koch, E.W., Beyond light: physical, geological, and geochemical parameters as possible
submersed aquatic vegetation habitat requirements, Estuaries, 24, 1-17. 2001
Kraemer, G.P., and R.S. Alberte. Age-related patterns of metabolism and biomass in
subterranean tissues of Zostera marina (eelgrass). Mar Ecol Prog Ser, 95, 193-203, 1993.
Kristensen, E. and T. H. Blackburn, The fate of organic carbon and nitrogen in experimental
marine sediments: influences of bioturbation and anoxia. J. Mar. Res., 45, 23-257, 1987.
Lee, K. and K.H. Dunton, Diurnal changes in pore water sulfide concentrations in the seagrass
Thalassia testudinum beds: the effects of seagrasses on sulfide dynamics, J. Exp. Mar.
Biol, and Ecol., 255, 201-214, 2000
Lee, K. and K.H. Dunton, Inorganic nitrogen acquisition in the seagrass Thalassia testudinum:
development of whole-plant nitrogen budget, Limnol Oceanogr., 44,1204-1215, 1999.
23
-------
Eldridge et ai.	Seagrass Stress Response Model
Lovley, D.R., E.J.P. and E. J. P. Phillips, Requirements for a microbial consortium to completely
oxidize glucose in Re(III)-reducing sediments, Appl. Environ. Microbiol., 54, 3234-3236,
1989.
Nepf, H.M. and E.M. Koch, Vertical secondary flows in submersed plant-like arrays, Limnol.
Oceanogr., 44, 1072-1080, 1999.
Onuf, C.P., Seagrass responses to long-term light reduction by brown tide in upper Laguna
Madre, TX: distribution and biomass patterns, Mar. Ecol. Prog. Ser., 138, 219-231, 1996.
Orth, J.J., Effect of nutrient enrichment on growth of eelgrass Zostera marina in Chesapeake
Bay, Virginia, USA., Mar Biol, 44, 187-194, 1977.
Postma, D. and R, Jakobsen. Redox zonation: Eqiulibrium constraints on the Fe(III)/S04-reduction
interface, Geochimica et Cosmochimica Acta. 60, 3169-3175, 1996.
Pulich, W. M., Effects of rhizosphere macronutrients and sulfide levels on the growth physiology
of halodule wrightii Aschers. and Ruppia maritima L. s.L, J. Exp. Mar. Biol. Ecol., 127,
69-80, 1989.
Short, F.T. Eelgrass and the wasting disease. Univ. of New Hampshire Cooperative Extension,
Durham NH. 1989.
Short, F.T. and S. Wyllie-Echevarria, Nature of Human-induced distributions of seagrass,
Environmental Conservation, 23,17-27, 1996.
Tailing, J.F., A new model for leaf photosynthesis incorporating gradients of light environment
and the photosynthetic properties of chloroplasts within a leaf, Annls Bot, 56, 489-499,
1957.
Teeter. A.M. B.H. Johnson. C. Berger. G. Stelling. N.W. Scheffner, M.H. Garcia and T.M.
Parchure, Hydrodynamic and sediment transport modeling with emphasis on shallow-
24
-------
Eldridge et al.	Seagrass Stress Response Model
water, vegetated areas (Lakes, reservoirs, estuaries and lagoons, Hydrobiologia, 444, 1-
23,2001.
Van Katwijk, M.M., L.H.T Vergeer, G.H.W. Schmitz, J.G.M. Roelofs. Ammonium toxicity in
eelgrass Zostera marina. Mar. Ecol. Prog. Sers., 157, 159-173, 1997
Zimmerman. R.C., JL. Reguzzoni, R.S. Alherte. Eelgrass (Zostera marina L.) transplants in San
Francisco Bay: role of light availability on metabolism, growth and survival. Aquat. Bot.
51, 67-86, 1995.
Zimmerman, R.C. and R.S. Alberte, Effect of light/dark transition on carbon translocation in
eelgrass Zostera marina seedlings. Mar Ecol Prog Ser, 136, 305-309, 1996.
25
-------
WET)~0Z~/2f
TECHNICAL REPORT DATA
WED-02-121 (Please read Instructions on the reverse before completing)
1. REPORT NO- 2.
EPA/600/A-02/066
3, RECIPIENTS ACCESSION NO
4. TITLE AND SUBTITLE n AA , / ^
Stress /tesponst mode/the Zntpor&n
of J tight, T^mper £> , _/
P.M. E/dndgt; X£. Kdldy, A- &. Buret
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Western £cohc,y 3>/ USEPA
ZOO S\N 3& St
(U>r/At(fe, 0& 47333
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
OA ml ds dfose
13. TYPE OF REPORT AND PERIOD COVERED
COn-Pureyice proceedt'ms
14. SPONSORING AGENCY CODE J
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The objective of our modeling was to better understand the relationship between seagrass and water-
column and sediment stressors (i.e., light, organic and particle sedimentation, sediment nutrients and
sulfides). The model was developed and optimized for sediments in Thalassia testudinum seagrass beds
of Lower Laguna Madre, Texas, USA and is composed of a plant sub-model and a sediment diagenetic
sub-model. Simulations were developed for a natural stressor (harmful algal bloom) and an anthropogenic
stressor (dredging event). The harmful algal bloom (HAB) was of limited duration and the simulations
showed no effect of the algal bloom on biomass trends but did suggest that sediment sulfides could inhibit
growth if the bloom duration and intensity were greater. The dredging event resulted in sedimentation of
a layer of organically rich material and reduction of canopy light for a period of months. The simulations
suggested that the seagrass could have recovered from the effects of light but residual effects of high
sulfides in the sediments would make the region uninhabitable for seagrasses for up to 2.5 years. These
modeling exercises point out the importance of using a geochemical model to evaluate the impact of
reduced light and enhanced organic loading from both natural and anthropogenic stressors to seagrass.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
SeadftLSs/c&rbofl sequestration/
segment/'geochemistry


18 DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21. NO. OF PAGES
20. SECURITY CLASS (This Page}
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE	forms/aamin/teehrpt.Srm 7/8/99
-------