AQUATIC CONSERVATION: MARINE AND FRESHWATER ECOSYSTEMS

                  Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                      Published online in Wiley InterScience
                   (www.interscience.wiley.com). DOI: 10.1002 /aqc.490




  Nitrogen loading alters seagrass ecosystem structure and
        support of higher trophic levels


  LINDA A. DEEGANa*, AMOS WRIGHTa, SUZANNE G. AYVAZIANb,c, JOHN T. FINNb,
     HEIDI GOLDENa, REBEKA RAND MERSONa and JOHN HARRISONa,d
         a
          The Ecosystems Center, Marine Biological Laboratory, Woods Hole, MA 02543, USA
    b
      Department of Natural Resource Conservation, University of Massachusetts, Amherst, MA 01002, USA
      c
       Western Australia Marine Research Laboratory, Estuarine and Coastal Finfish Section, PO Box 20,
                   North Beach, Western Australia 6020, Australia
         d
          Stanford University, Geological and Environmental Sciences, Stanford, CA 94305, USA

                             ABSTRACT
       1. Anthropogenic-derived nutrient inputs to coastal environments have increased dramatically
      worldwide in the latter half of the 20th century and are altering coastal ecosystems. We evaluated the
      effects of nitrogen loading on changes in macrophyte community structure and the associated fauna
      of a north temperate estuary. We found that a shift in primary producers from eelgrass to
      macroalgae in response to increased nutrient loading alters habitat physical and chemical structure
      and food webs. As nitrogen load increased we found increased macroalgal biomass, decreased
      eelgrass shoot density and biomass, decreased fish and decapod abundance and biomass, and
      decreased fish diversity.
       2. The central importance of macroalgae in altering eelgrass ecosystem support of higher trophic
      levels is evident in the response of the ecosystem when this component was manipulated. Removal of
      macroalgae increased eelgrass abundance and water column and benthic boundary layer O2
      concentrations. These changes in the physical and chemical structure of the ecosystem with lower
      macroalgal biomass resulted in higher fish and decapod abundance and biomass.
       3. Both a 15N tracer experiment and the growth of fishes indicated that little of the macroalgal
      production was immediately transferred to secondary consumers. d15N values indicated that the
      most abundant fishes were not using a grazing food web based on macroalgae. Fish tended to grow
      better and have a greater survivorship in eelgrass compared to macroalgal habitats.
       4. Watershed-derived nutrient loading has caused increased macroalgal biomass and degradation
      and loss of eelgrass habitat, thus reducing the capacity of estuaries to support nekton.
      Copyright # 2002 John Wiley & Sons, Ltd.

      KEY WORDS:  ecosystem alteration; nutrient loading; estuarine; seagrass; food webs; eutrophication; fish; decapods




*Correspondence to: Linda A. Deegan, The Escosystems Center, Marine Biological Laboratory, Woods Hole, MA 02543, USA.
E-mail: ldeegan@mbl.edu

Copyright # 2002 John Wiley & Sons, Ltd.                                Received 1 August 2000
                                                     Accepted 28 June 2001
194                      L.A. DEEGAN ET AL.


                       INTRODUCTION

Anthropogenic-derived nutrient inputs to coastal environments have increased dramatically worldwide in
the latter half of the 20th century and are altering coastal ecosystems (Valiela et al., 1992; Nixon, 1995;
Howarth et al., 1996). The increased delivery of nutrients to coastal waters has been driven primarily by
local changes in the watersheds (Howarth et al., 1996; National Research Council, 2000), although global
changes in atmospheric deposition also contribute (Vitousek et al., 1997). The concern over nitrogen
loading is particularly acute in estuaries because loading is increasing and because primary production in
coastal waters is limited by nitrogen (Howarth, 1988). This increase in nitrogen loading is having large but
only partially understood effects on estuarine ecosystem function. One of the prominent functions of
estuaries is their support of higher trophic levels such as fish, decapod crustaceans and shellfish. Over 50%
of all U.S. economically important and over 25% of all east-coast shelf fish species use estuaries at some
stage in their life history (Houde and Rutherford, 1993; Ray, 1997).
  Eutrophication can result in changes in ecosystem level dynamics such as productivity, dissolved oxygen
concentrations, nutrient cycling, trophic structure and energy flow which have the potential to affect fish
production (Nixon et al., 1986; Breitberg et al., 1997; Valiela et al., 1997a,b). Although there are links
between changes in estuarine ecosystem dynamics and fish production, in most cases these linkages have
not been quantified.
  In many temperate coastal systems, eelgrass (Zostera marina) is a dominant primary producer and is an
important habitat for fish and invertebrates (Adams, 1976; Orth et al., 1984; Virnstein, 1987; Deegan et al.,
1997). In many ecosystems, increased loading of the limiting nutrient alters the assemblage of primary
producers (Rapport and Whitford, 1999). In coastal systems, nitrogen enrichment enhances the
proliferation of faster growing phytoplankton, epiphytic algae and macroalgae that compete with seagrass
for light and space (Valiela et al., 1997b; Raffaelli et al., 1998; Hauxwell et al., 2001). This interaction is
thought to be a major factor contributing to the widely observed declines in seagrass abundance worldwide
(Orth and Moore, 1983; Cambridge et al., 1986; Duarte, 1995; Raffaelli et al., 1998).
  Although the shift in primary producers in response to nitrogen loading has been well documented, the
impacts of this change on estuarine food webs and the production of fishes and invertebrates are not well
known. Many studies have suggested that the high productivity of seagrass ecosystems provides abundant
food while their structural complexity benefits fish and invertebrates by providing protection from
predation (Heck and Thoman, 1981; Leber, 1985; Bell and Westoby, 1986; Heck and Crowder, 1991; Heck
et al., 1997). In some estuaries, such as the Chesapeake Bay, competition with phytoplankton and epiphytes
eliminates eelgrass and the resulting bare substratum does not support a diverse or abundant fish
assemblage (Orth and Moore, 1983; Wyda et al., in press). In shallow, low energy estuaries, macroalgae
replace eelgrass (Harlin and Thorne-Miller, 1981; Duarte, 1995; Valiela et al., 1997b; Hughes et al., in
press). The replacement of eelgrass by macroalgae potentially changes the structural complexity, food webs
and the chemical suitability of the habitat for nekton. The competition between macroalgae and eelgrass
leads to diminished eelgrass growth and stature, and declines in shoot density and total habitat area (Short
et al., 1995; Short and Burdick, 1996; Hauxwell et al., 2001). Low dissolved oxygen levels and anoxia often
occur in macroalgal mats due to the respiration of the algae and the decomposition of the accumulated
macroalgal detritus (Johnson and Welsh, 1985; Hull, 1987; D’Avanzo and Kremer, 1994; D’Avanzo et al.,
1996). The loss and degradation of seagrass habitats has undoubtedly affected the distribution and
productivity of the animals that use them, however, for the most part, these changes have not been
demonstrated (Dennison et al., 1993; Hoss and Thayer, 1993).
  The present study examines the effects of increased nutrient delivery from watersheds on eelgrass habitat
plant communities, ecosystem physical and chemical structure, and fish and invertebrate abundance and
diversity. We hypothesize that increased nutrient loading alters primary producers, replacing eelgrass with
macroalgae. This replacement decreases ecosystem physical structure by replacing the tall, dense eelgrass

Copyright # 2002 John Wiley & Sons, Ltd.          Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                   195

canopy with low lying, finely branched macroalgae mats, decreases chemical suitability by the development
of low dissolved oxygen at the benthic boundary and alters food webs. We hypothesize that as macroalgal
biomass increases the abundance of small fish and decapod abundance decreases because the changes in
ecosystem physical and chemical structure will make these animals more susceptible to predation. We also
hypothesize that these ecosystem structure changes will lead to little macroalgal primary production being
passed up the food web to secondary consumers. Thus, as macroalgal biomass increases in response to
nitrogen loading, the areal extent and suitability of eelgrass habitats as a nursery area for fish and decapods
declines.
  We used several approaches to test these hypotheses. We examined natural fish, invertebrate and
macrophyte abundance in three sub-estuaries of Waquoit Bay that are subject to different nitrogen loading
rates. This was a space-for-time substitution (Pickett, 1989), where different nitrogen loading rates in
similar estuaries simulated changes over time due to increased anthropogenic nutrient loading (Valiela et al.,
1992). We conducted a macroalgal biomass experiment creating decreased and increased macro-
algal biomass and examined the response of plants, invertebrates and fish. We conducted a macroalgal
15
  N-enrichment experiment to trace macroalgal organic matter through the food web and examined the
growth of fishes in eelgrass and macroalgal-dominated habitats.


                         METHODS

General site description
This study was conducted in 3 of the 9 interconnected sub-watersheds that make up the Waquoit Bay
estuarine system (Figure 1). Hamblin, Timms and Sage Lot Ponds, are shallow sub-embayments with
fringing salt marsh. The physical characteristics of these ponds are similar, with salinity ranging seasonally
between 20% and 32% and temperature between 98C and 308C. The maximum depth at mean low water is
between 1.1 and 1.3 m, with a tidal range of 0.2–0.3 m. Although eelgrass was historically abundant
throughout Waquoit Bay, it has been declining in extent and abundance since the middle 1970s (Valiela
et al., 1992; Short and Burdick, 1996).

Nitrogen loading gradient comparison
To compare eelgrass ecosystems under differing nitrogen loading rates, we sampled the nekton (fish and
decapod crustacea species), eelgrass and macroalgal communities of Hamblin, Sage Lot and Timms Ponds
during the summers of 1992 and 1993. The variation in watershed development determines the amount of
nitrogen a pond receives (Valiela et al., 1992, 1997a, 2000). Upland development and nitrogen loading
varies (Valiela et al., 2000) from very low for Timms Pond (0 houses haÀ1, 16 kg N yrÀ1) to light for Sage
Lot Pond (0.12 houses haÀ1, 534 kg N yrÀ1) to moderately high in Hamblin Pond (1.3 houses haÀ1,
1679 kg N yrÀ1). We sampled two eelgrass areas in Sage Lot and Timms Ponds, and four areas in Hamblin
Pond. Areas were sub-divided into quarters to facilitate distribution of samples around the site.
  Eelgrass and macroalgae were measured in June and August 1992 and August 1993. We measured
eelgrass shoot density in 0.063 m2 quadrats haphazardly placed in each quarter of a site (n=8 for a site at a
sampling period). Macroalgal and eelgrass biomass was sampled by lowering a 0.073 m2 cylinder, with a
mesh bag (500 mm mesh) attached, over the eelgrass down to the sediment and using a thin plate at the
sediment surface to cut and remove the enclosed eelgrass, algae and detritus. Samples (n=6 per site per
sampling period) were washed on a steel screen and algae were sorted to species. Plants were oven-dried at
608C for (at least) 24 h and weighed (Cowper, 1978).
  Nekton were sampled in June, July and August in 1992, and in June and August 1993. Nekton (fish and
larger individuals of decapod crustacea species) abundance and species composition were sampled using

Copyright # 2002 John Wiley & Sons, Ltd.          Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
196                          L.A. DEEGAN ET AL.




Figure 1. Waquoit Bay with sub-estuaries and sub-watersheds indicated. The comparison of ecosystem response to a nitrogen loading
gradient was conducted in Hamblin (1679 kg N yÀ1), Sage Lot (534 kg N yrÀ1), and Timms (16 kg N yrÀ1) Ponds. The macroalgal
                     removal study was conducted in Hamblin Pond.


1 m  1 m throw nets in each site (n=4 site per sampling period). Small mesh (3 mm mesh) throw nets were
used because they effectively sample nekton in high densities of aquatic vegetation (Heck and Wetstone,
1977; Gore et al., 1981; Kushlan, 1981). As the boat drifted towards a site, the nets were tossed away from
the boat into a site. Nekton were removed from the enclosed area by a bar seine made from PVC pipe and
netting that fit closely to the inside of the throw net. We swept the bar seine through the throw net eight
times, and all fish and decapods were removed after each sweep. The traps were then allowed to sit
undisturbed for 10 min (allowing animals to re-emerge from hiding), after which four more sweeps were
conducted. We continued the pattern until no more animals were caught by four consecutive sweeps. Initial
tests using marked fish indicated a 90–95% recapture rate with this procedure. Animals were frozen and
later identified to species and counted.

Macroalgal biomass experiment
To evaluate how the development of a thick macroalgae mat affected both eelgrass and the animal
community, we manipulated macroalgal biomass. The macroalgal biomass experiment was conducted

Copyright # 2002 John Wiley & Sons, Ltd.               Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                  197

during April–August in Hamblin Pond because this pond had high macroalgae biomass and still had
eelgrass. The high ambient biomass of macroalgae allowed us to create the greatest differences between
control plots (ambient macroalgal biomass) and experimental plots (lower or higher macroalgae biomass).
The macroalgae were primarily Cladophora vagabunda, and Gracilaria tikvahiae with some blooms of Ulva
lactuca, Enteromorpha plumosa in the spring and fall. To measure the effects in the higher trophic levels, and
to avoid edge effects, treatment areas were four 25 m  25 m blocks; each block was further divided into
four 10 m  10 m plots with an approximately 1 m buffer zone between both the edge and another plot.
Each treatment was randomly assigned to a block prior to any sampling or manipulation.
  Scuba divers created four macroalgal biomass treatments: (1) control } plots in this treatment were not
manipulated and had ambient macroalgal biomass; (2) low macroalgal biomass } macroalgae were
removed; (3) high macroalgal biomass } macroalgal biomass was doubled (2 Â ) over control levels; (4)
disturbance control } macroalgal biomass was not altered, however, macroalgae were removed and then
replaced in the same plots as a control for disturbance. To create the low macroalgal biomass treatment,
divers on snorkel or scuba removed noticeable macroalgae by hand from around the eelgrass and placed it
in bags. Large invertebrates (primarily hairy sea cucumbers, Sclerodactyla briareus) and fish, but not small
invertebrates such as amphipods, were removed from the macroalgae and returned to their original plots.
The removed macroalgae were added to the high macroalgal biomass treatments to increase macroalgal
biomass levels to approximately two times the control levels. In the disturbance control plots, divers
removed the algae and then returned the algae to the same plot without removing invertebrates or fish.
Macroalgae in the control plots were not disturbed. A crew of 8–10 divers completed the experimental setup
in about 3 weeks in late May and early June. Experimental treatments were maintained weekly until the end
of September.
  Once prior to macroalgal manipulation (pre-treatment: April–May 1990) and monthly after
manipulation (post-treatment: June–August), we assessed the eelgrass, macroalgae, nekton and macro-
epifauna according to the methods previously described in the ‘Nitrogen loading gradient comparison’
section with some minor modifications. To assess how the chemical environment might have changed we
measured temperature, salinity and oxygen periodically throughout the experiment. Vertical profiles of
dissolved oxygen (Yellow Springs Instrument, Dissolved Oxygen Meter) were taken at dawn on 26–27 July
and 2–3 August. Vertical dissolved oxygen profiles were taken from 6–7 AM in the four treatments, in the
order: low, control, disturbance, and high. Sampling treatments in this order gave us a conservative
estimate of differences in dissolved oxygen between treatments. One 1 m wide transect along the diagonal
(14 m long) of the plot was counted each month to produce an integrated estimate of eelgrass shoot density
for each plot. We measured macroalgal biomass and macro-epifauna abundance from 2 cylinder samples in
3 of the 4 plots of each treatment each month. All animals retained on a steel screen (0.5 mm mesh; Thrush,
1986) were preserved in 95% ethanol, and later identified to the lowest possible taxon (usually species) using
a dissecting microscope. Nekton were sampled using throw nets. Because throw net sampling disturbed
approximately a 1.5 m  1.5 m area, and we were concerned about the cumulative disturbance this would
cause over the course of the experiment, we took only one throw net sample in each of the 4 plots of a
treatment each month. Areas that had been sampled were marked and not re-sampled. Although this limits
our ability to detect differences because of small sample size, we felt that to sample more intensively would
cause excessive sampling disturbance.

15
   N macroalgal labelling experiment
To understand the contribution of the macroalgae to the food web that supported fishes, we labelled
macroalgae with 15N and traced the propagation of this label through the food web within a cage in
Hamblin Pond. On 13 July 1992, a 2 m  2 m cage constructed of PVC pipe and netting (10 mm mesh) was
placed in the pond, and the macroalgae (7605 g wet weight) were removed. Macroalgae were placed in a

Copyright # 2002 John Wiley & Sons, Ltd.          Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
198                     L.A. DEEGAN ET AL.


302 litres tub under ambient conditions. Beginning 13 July 1992, the 15N label was added as 99% enriched
(15NH4)2SO4 in aqueous solution (8.03 Â 10À4 M) in daily increments for 6 d (total (15NH4)2SO4=0.4074 g).
On 23 July the 15N-labelled algae and representative species of animals were placed in the cage at
abundances approximating those found during our sampling in Hamblin Pond.
  Measurements of d15N in algae, invertebrates, and fish were performed on reference samples and at
weekly intervals for three weeks (30 July, 6 August, and 13–15 August). Macroalgae (aggregates of all
species), invertebrates and fish (1–5 whole individuals) were dried at 608C for 24 h and ground with a
mortar and pestle. d15N was analyzed at the Ecosystems Center’s Stable Isotope Facility using an
automated elemental analyzer with a cryogenic purification system coupled to a Finnigan Delta S iso-
tope ratio mass spectrometer. Stable isotope ratios are expressed using d notation defined thus:
d15N (%)=[(15N:14Nsample /15N:14Nstandard)À1] Â 1000. Air was used as the standard. Analytical precision
was Æ 0.1%.

Fish growth experiment
To determine if macroalgae affected the growth rate of fishes, fourspine stickleback (Apeltes quadracus) and
rainwater killifish (Lucania parva) growth rates were examined in eelgrass and macroalgal habitats in
Timms and Sage Lot Ponds. We placed cages (2 m high  1 m2, 10 mm mesh) in each habitat in each pond.
Individual fish were anaesthetized using phenoxyethanol and marked mid-dorsally with one of 6 colors of
acrylic paint using a 26-guage hypodermic needle. Five marked fish of each species were placed in each cage
on 21 August 1995, and collected approximately 21 d later. We measured individual fish wet weight
( Æ 0.01 g) and total length ( Æ 1 mm) at the beginning and end of the experiment, and calculated growth as
the difference between these two measurements. Fish used in this experiment were mid-sized juveniles
(initial length ranged between 26 and 30 mm for both species; initial weight 0.16 g for fourspine stickleback
and $0.25 g for rainwater killifish).


Statistical analysis
Data were analysed using either factorial analysis of variance or repeated measures analysis of variance
(ANOVA; SAS Institute Inc., 1998). Means are presented with one standard error. The Tukey–Kramer
post-hoc test was used to test differences in means if a significant main effect was found. Significance level
was a=0.05 for all analyses. Data were transformed as needed to meet the assumptions of ANOVA.
  In the macroalgal biomass experiment the treatment plots were considered the experimental unit (n=4).
Factorial ANOVA was used on pre-treatment samples when one sample was taken per plot in the pre-
treatment period (eelgrass shoot density and fish and decapod abundance). For macroalgal biomass and
macro-epifauna pre-treatment samples in which more than one sample was taken in a plot, the replicate
samples were considered repeated measures (n=3 because only three of the 4 plots were sampled).
Repeated measures ANOVA was used for all post-treatment measurements. Multiple samples taken in a
plot in the same month were replicate measures within that month. Monthly samples were treated as
repeated measures for the plot. For the nitrogen loading comparison, we used a nested factorial ANOVA
with plots nested with ponds. To meet the assumptions of ANOVA, macroalgal biomass, macro-epifauna,
and fish and decapods abundances were log-transformed (ln (x+1)) before analysis. Analysis of fish and
decapods was restricted to small-sized individuals (525 g individual total weight), because larger, more
mobile fish are not effectively sampled with throw nets. Eelgrass shoot density did not need to be
transformed. Equality of variance F-tests and K–S Normality tests indicated that the transformed variables
were normally distributed and more closely met the equality of variance assumption; therefore we used
ANOVA. Growth (mm or g wet weight per 21 d) of individual fish in the cage experiments was compared
between macroalgal and eelgrass habitats using ANOVA.

Copyright # 2002 John Wiley & Sons, Ltd.          Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                  NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                       199

                             RESULTS

Nitrogen loading gradient comparison
Macroalgal biomass increased, and eelgrass biomass and shoot density decreased with increased nitrogen
loading (Figure 2, Table 1). The response to nitrogen load was similar between years, although there
were some year-to-year differences in absolute abundances. The plant community was dominated by
eelgrass with little macroalgae at the lowest nitrogen loading (Timms Pond) and changed to thin stands of
eelgrass surrounded by a deep (4–14 cm) macroalgal mat at the highest nitrogen load (Hamblin Pond).
Eelgrass comprised 99% (Timms Pond), 35% (Sage Lot Pond), and 1% (Hamblin Pond) of the total plant
(algae+eelgrass) biomass. Eelgrass biomass (156 g mÀ2) and shoot density (300–600 shoots m2) were
highest in Timms Pond and declined with increasing nitrogen load. Three species of macroalgae, Gracilaria
tikvahiae, Cladophora vagabunda, and Chaetomorpha sp. accounted for greater than 95% of the total
macroalgal biomass. Macroalgal biomass was positively related to nitrogen loading and was highest in
Hamblin Pond (171 dry g mÀ2). Hamblin Pond also had a thick ($70 cm deep) layer of black, anoxic mud,
while Timms Pond had a thinner (5 cm) layer of organic matter overlying the sand layer.
  Fish abundance and average number of fish species were negatively related to increasing nitrogen loading
(Figure 2, Table 1). The pattern of fish community response to nitrogen loading was similar between years,
although year-to-year variation in community characteristics was also apparent. In both years, the mean




Figure 2. Mean biomass (g dry wt mÀ2) of eelgrass and macroalgae, eelgrass shoot density (number shoots mÀ2), and biomass (g dry
wt mÀ2), abundance (number mÀ2) and number of species (number mÀ2) of fishes and decapods during June–August, 1992 and 1993 in
Hamblin, Sage Lot and Timms Ponds. Data are mean Æ standard error for all plots within a pond. Some standard errors are smaller
                           than the symbol size.

Copyright # 2002 John Wiley & Sons, Ltd.               Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
200                          L.A. DEEGAN ET AL.


   Table 1. F-values for ANOVAs comparing eelgrass community characteristics of Timms, Sage Lot and Hamblin Pondsa

                       Abundance (number mÀ2)       Biomass (g mÀ2)     Species (number mÀ2)
PLANTS
 Eelgrass      Pond(plot)       30.7***               6.3***         }
           Year           2.9                22.3**          }
           Year*Pond(plot)     6.4**                2.3           }

  Macroalgae    Pond(plot)       }                  21.8***         }
           Year          }                  2.4           }
           Year*Pond(plot)     }                  0.52          }

ANIMALS
Fish         Pond(plot)       10.7***               5.6***          4.1***
           Year          19.7***               18.7***          0.9
           Year*Pond(plot)     6.2***               3.2**          0.7

  Decapod      Pond(plot)        1.8                 2.2**          2.1*
           Year          13.2**               18.3***         11.2**
           Year*Pond(plot)     1.2                 1.0           3.1**
a
Data collected June-August 1992 and 1993; ANOVAs were on ln(x+1) transformed data for macroalgae, fish and decapods. For each
variable the degrees of freedom were: Pond(plot), 7; Year, 1; Year*Pond(plot), 7. Residual degrees of freedom were: Eelgrass density
(number shoots mÀ2), 384; Eelgrass biomass (dry wt g mÀ2), 78; Macroalgal biomass (dry wt g mÀ2), 76; Fish and decapods (Density,
number mÀ2; Biomass, g wet wt. mÀ2; number of species), 128. Significance levels are: ***=0.001; **=0.01; *=0.05.




number of species per m2 declined with increased nitrogen loading (Figure 2). Total number of species
captured did not differ appreciably in the 2 years of our study (12 species in 1992 and 13 in 1993). Sage Lot
Pond had the highest number of total fish species (13), while Timms and Hamblin had the same number of
total species (8). Six species were common to all three ponds: fourspine stickleback, American eel (Anguilla
rostrata), northern pipefish (Sygnathus fuscus), mummichog (Fundulus heteroclitus), rainwater killifish
(Lucania parva) and Atlantic silverside (Menidia menidia). Fourspine stickleback was the most abundant
and most frequently caught fish species in all three ponds, and accounted for 82% of the total number of
fish caught in 1992 and 73% in 1993. In 1992 oyster toadfish (Opsanus tau) accounted for 56% of total
biomass, due to the capture of one 735 g individual in Hamblin Pond. American eel accounted for 42% of
total biomass, due to the capture of six large individuals (20–137 g). Large toadfish and eels were found
exclusively in Hamblin Pond. Because 1 m2 throw nets do not accurately sample large mobile fish such as
these, we excluded these larger fish from our analysis. With the exclusion of these fish, mean fish biomass
was highest in the pond with the lowest nitrogen load and was negatively related to increasing nitrogen
load.
  Although abundance of decapods was highly variable, we found a general decline with increasing
nitrogen loading (Figure 2, Table 1). Total number of species caught in 1992 (n=8) was greater than in
1993 (n=5). The daggerblade grass shrimp (Palaemonetes pugio) was the dominant species in both
abundance and biomass during both years. This species accounted for 67% of the total catch in 1992 and
95% in 1993.

Macroalgal biomass experiment
The distribution of plants and most animals did not differ among treatments prior to the macroalgal
biomass manipulation (Figure 3, Table 2). Eelgrass shoot density was low (5 shoots m2) and macroalgae

Copyright # 2002 John Wiley & Sons, Ltd.                Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                      201




Figure 3. Mean ( Æ standard error) macroalgal biomass (g dry wt mÀ2) and eelgrass shoot density (number of shoots mÀ2) and
macroepifauna, fish and decapod abundance (number mÀ2) over time in the macroalgal manipulation experiment. Some standard
                      errors are smaller than the symbol size.




biomass high ($100 g dry wt mÀ2) with no significant differences across the plots. Gracilaria tikvahiae,
Cladophora vagabunda and Chaetomorpha sp. were the dominant species of algae, representing
approximately 50%, 20% and 5% of total macroalgal dry mass, respectively. Initial fish abundance did
not differ among the treatments (4 Æ 0.4 individuals mÀ2), although the low treatment had the highest
(5 Æ 0.4 mÀ2) and the high treatment the lowest (3 Æ 0.3 individuals mÀ2) pre-treatment mean abundances.
Abundance of decapods did not differ among the treatments (20 Æ 2.9 individuals mÀ2). Only the initial
abundance of macro-epifauna differed among treatments (Figure 3), with the control and low treatments
having similar and higher abundance (2.5 Â 104 individuals mÀ2) compared to the high and disturbance
treatments (1.0 Â 104 individuals mÀ2) (Tukey–Kramer test, p50.05).
  Altering the biomass of macroalgae changed many aspects of the eelgrass ecosystem, including relative
plant and animal abundance and biomass and chemical suitability. Macroalgal biomass was highest in the
spring and declined over the summer except in the High macroalgal biomass treatment (Figure 3, Table 2).

Copyright # 2002 John Wiley & Sons, Ltd.              Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
202                          L.A. DEEGAN ET AL.


  Table 2. Pre- and post-macroalgal manipulation experiment ANOVA analyses of algal biomass, eelgrass density, and fish and
                           invertebrate abundancea

Pre-treatment                                               DF      F-value
Plants
 Algal biomass (in dry wt mÀ2)                Treatment                 3         2.47
                                Replicate                 1         0.02
                                Replicate  treatment           3         1.78
  Eelgrass density (number of shoots mÀ2)          Treatment                 3         1.98

Animals
Macro-epifauna abundance (in number mÀ2)           Treatment                 3         4.86*
                                Replicate                 1         3.26
                                Replicate  treatment           3         0.17
  Fish abundance (in number mÀ2)               Treatment                 3         1.01

  Decapod abundance (in number mÀ2)             Treatment                 3         1.08

Post-treatment
Plants
 Algal biomass (in dry wt mÀ2)                Treatment                 3        21.75***
                                Month                   2         4.70*
                                Month  treatment             6         2.53
                                Replicate                 1         1.23
                                Replicate  treatment           3         0.03
                                Month  replicate             2         2.16
                                Month  replicate  treatment       6         0.62

  Eelgrass density (number shoots mÀ2)            Treatment                 3       487.0***
                                Month                   1       1018.0***
                                Month  treatment             3       225.0***
Animals
Macro-epifauna abundance (in number mÀ2)           Treatment                 3         4.75*
                                Month                   2         8.82**
                                Month  treatment             6         2.94*
                                Replicate                 1         0.43
                                Replicate  treatment           3         0.42
                                Month  replicate             2         3.17
                                Month  replicate  treatment       6         0.38

  Fish abundance (in number mÀ2)               Treatment                 3         4.75**
                                Month                   2         0.73
                                Month  treatment             6         0.15

  Decapod abundance (in number mÀ2)             Treatment                 3         2.77
                                Month                   2        37.2***
                                Month  treatment             6         1.08

a
Treatments are: reference, removal, disturbance, addition. Pre-treatment samples were collected in April–May 1990, before
experimental manipulation of macroalgae. Pre-manipulation algal biomass and macro-epifauna were analyzed with repeated measures
ANOVA, with replicate samples treated as the repeated measure; Eelgrass, fish and decapod abundance were factorial ANOVA as no
replicate within plots samples were taken. Pre-treatment residual degrees of freedom were: Algal biomass and macro-epifauna, 8; Fish
and decapods, 12; Eelgrass density, 8. Post-treatment samples were collected in June-August, 1990, after macroalgal manipulation, and
analyzed with repeated measures ANOVA. Significance levels are: ***=0.001; **=0.01; *=0.05.




Copyright # 2002 John Wiley & Sons, Ltd.                 Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                   203

Macroalgal biomass was lowest in the low and highest in the high treatments compared to the other
treatments. Throughout the course of the experiment, macroalgal samples were collected the day before the
weekly maintenance of the macroalgal biomass treatments. The doubling time of macroalgae in these
systems is less than a week (Peckol et al., 1994; Peckol and Rivers, 1996) indicating macroalgae could grow
substantially between treatments and sampling. Thus, our estimates of macroalgal biomass represent the
peak biomass of the macroalgae in the treatments for a week and are a conservative estimate of the
effectiveness of experimental manipulations on macroalgal biomass. We found no differences in macroalgal
species composition over the course of the experiment (data not shown).
  The removal of macroalgal biomass resulted in a denser cover of eelgrass that persisted into the following
spring. Eelgrass shoot density was highest in the low macroalgal biomass treatment compared to all other
treatments by August of the same year and was over 4 Â higher the following spring (Figure 3). We
observed small plant sprouts as well as more leaves per shoot at the end of the season in the low macroalgal
biomass treatments, but not in the other treatments. The greatest difference among treatments was seen the
following spring when eelgrass shoot density in the low macroalgal biomass treatment was 4 Â higher than
in the control, 2 Â higher than in the disturbance control, and 10 Â higher than the high macroalgal
biomass treatment.
  High macroalgal biomass reduced O2 concentrations and removal of macroalgae increased O2 near the
sediment–water interface (Figure 4). Vertical profiles indicated higher oxygen content in the water and near
the bottom in the low macroalgal biomass treatment compared to the other treatments. On 26 July, after
several warm cloudy days and with a high tide at 3–4 AM the low macroalgal biomass treatment had twice
as much oxygen at the surface as did the other treatments. On 3 August, the low macroalgal biomass
treatment had consistently higher dissolved oxygen levels throughout the profile than did the other
treatments. The higher O2 content in the low macroalgal biomass treatment is probably due to the
combination of less respiration because of low macroalgal biomass and higher oxygen production by the
more abundant eelgrass blades. The dissolved oxygen in the high macroalgal biomass treatment was
consistently lower than in the other treatments. We found no vertical stratification of either temperature or
salinity during the experiment (data not shown).
  Macro-epifauna showed no consistent pattern of response to the macroalgal biomass manipulation.
Macro-epifauna abundance declined over the summer in the control and low macroalgal biomass
treatments, while they increased in the disturbance control and high macroalgal biomass treatments
(Figure 3, Table 2). The higher levels of macro-epifauna in the high macroalgal treatment could be
attributed to transferring the animals with the macroalgae. A total of 53 species were caught; of these 10
genera/species comprised 90% of the small invertebrates caught. These were amphipods (Microdeutopus
sp., Corophium sp., Ampithoe sp, Lysianopsis alba), tunicates (sea grape Molgula sp.), polychaetes
(Haploscoloplos robustus, Podarke obscura), isopods (Erichsonella attenuata), burrowing sea cucumbers
(Leptosynapta sp.) and a small snail (Hydrobia minuta). The total number of species present in the low (43)
and high (42) macroalgal biomass treatments was higher than in the control (39) or disturbance (36)
treatments.
  Removal of macroalgae created a more suitable habitat for fishes as evidenced by the higher
fish abundance and diversity in the low macroalgal biomass treatment compared to all other treatments
(Figure 3, Table 2). A total of 10 fish species were collected: threespine stickleback (Gasterosteus aculeatus),
Atlantic silverside, winter flounder, northern pipefish, mummichog, oyster toadfish, American eel, inland
silverside (Menidia beryllina), rainwater killifish and northern searobin (Prionotus carolinus). More fish
species were found in the low macroalgal biomass treatment (8) than in any of the other treatments
(5 species in the control; 6 in the disturbance; and 4 in the high macroalgal biomass treatments).
Some benthic-associated fish, such as winter flounder, were found only in the low macroalgal biomass
treatment. In June, July and August the fish were primarily young-of-the-year fish from adults that
spawned earlier.

Copyright # 2002 John Wiley & Sons, Ltd.           Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
204                        L.A. DEEGAN ET AL.




Figure 4. Water column dissolved oxygen profiles from the macroalgal manipulation experiment. Measurements were taken in the
                   early morning (6–7 AM) after a 3–4 AM high tide.



 We observed a seasonal pattern in the response of decapods to manipulation of macroalgal biomass
(Figure 3, Table 2). Decapod abundance was highest in April and declined to near zero in all treatments
during June and July. In August the low macroalgal biomass treatment had a higher decapod abundance
compared to all other treatments (Tukey–Kramer test, p50.05), due primarily to the recruitment of newly
hatched daggerblade grass shrimp. Decapods were primarily adults and young-of-the-year of shrimp and
crabs: daggerblade grass shrimp (the dominant species), sand shrimp (Crangon septemspinosa), spider crab
(Libinia dubia), blue crab (Callinectes sapidus), and mud crabs (Rhithropanopeus harrisii, Neopanope sayi
and Panopeus herbsteii).

15
   N Macroalgal labelling experiment
Changes in the d15N of primary producers and consumers in the 15N enrichment experiment indicated
that very little of the macroalgal production was transferred up the food web to the common fish species

Copyright # 2002 John Wiley & Sons, Ltd.              Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                  NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                       205

Table 3. The d15N of primary producers and consumers in the 15N enrichment of macroalgae experiment in Hamblin Pond, July 1992

                                Date
Trophic level          Species             Initial     1 week     2 weeks      3 weeks
                                23 July     30 July     6 August      13 August
Primary producers
Macroalgae           Gracilaria sp.         979               456
                Cladophora sp.
Eelgrass            Zostera marina           7.3               10.1

Primary consumers
Herbivores
Amphipod            Gammarus mucronatus        5.5      1370      411
Isopod             Ericsonella filiformis       4.3       30.2               10.1
Isopod             Idotea baltica           6.3                       32.9
Detritivores
Hairy cucumber         Sclerodactyla briareus       7.1       13.5               9.7
Polychaete           Nereis sp.             7.3       89.1

Secondary consumers
Oyster toadfish         Opsanus tau            10.5       24               31.7
Atlantic Silverside       Menidia menidia          10.0               10.5       11.1
Threespine stickleback     Gasterosteous aculatus      10.6       11.2      9.7       10.5
Sand Shrimp           Crangon septimspinosa       9.5               9.4       9.4




(Table 3). The d15N of macroalgae was enriched 160 Â (970%) over baseline values (6%) when it was
placed in the enclosure. After 2 weeks, the algae had lost about 50% of the initial 15N enrichment, but
remained 80 Â as enriched as baseline (450%). These enriched values clearly distinguish the macroalgae
from all other primary producers (d15N values of around 6–10%) in the estuary.
  The d15N values of some primary consumers (invertebrate herbivores and detritivores) were enriched
2–250 Â that of natural values indicating they were consuming enriched macroalgae. Background variation
in natural d15N value for a consumer species may be 1–2%, therefore we considered any increase above 5%
to indicate enrichment over natural abundance levels (Hughes et al., 2000). Within 1 week, the herbivorous
amphipod Gammarus mucronatus was enriched with d15N values 250 Â its natural abundance d15N value.
The d15N value for Gammarus (1370%) after 1 week was higher than that of the algae mix (979%), possibly
due to preferential consumption of epiphytic diatoms on the macroalgae. Although it was not possible to
determine a separate d15N value of the epiphytic diatoms and macroalgae, it is likely that because of its fast
turnover this diatom film would be enriched relative to the bulk macroalgae. The rate of the decline in the
d15N value of Gammarus during the experiment was proportional to the decline in the macroalgae
suggesting macroalgae was its primary food source. Two other herbivores, the isopods Ericsonella and
Idotea, were slightly enriched in d15N indicating minimal consumption of the enriched algae. The degree of
enrichment in the detritivores was highly variable. Nereis was rapidly enriched (15 Â baseline; 89% after
2 weeks) indicating direct consumption of the macroalgae. In contrast, the hairy cucumber showed little to
no enrichment after 3 weeks.
  The most abundant secondary consumers, Atlantic silverside and three-spine stickleback, showed little to
no enrichment indicating they were not depending on a macroalgal based grazing food web. Of the three
fish species sampled only a benthic predator, a juvenile oyster toadfish (10 g wet weight), was found to have

Copyright # 2002 John Wiley & Sons, Ltd.               Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
206                         L.A. DEEGAN ET AL.




Figure 5. Mean growth ( Æ standard error) of fourspine stickleback (Apeltes quadracus) and rainwater killifish (Lucania parva) in
                     eelgrass and macroalgae-dominated habitats.




an enriched d15N value (31.7% versus a baseline of 10.5%). The predatory sand shrimp did not show an
increase over natural abundance d15N values.

Fish growth experiment
There was a tendency for growth of fish to be higher in eelgrass compared to in macroalgal habitats
(Figure 5), however, the difference was significant for rainwater killifish (p=0.03 for weight; p=0.3 for
length; n=17) but not fourspine stickleback (p=0.6 for weight; p=0.4 for length; n=21). Individuals of
both species lost weight in macroalgal but not in eelgrass habitats. Fewer fourspine sticklebacks were
recovered from macroalgal habitats (n=4) compared to eelgrass habitats (n=17), indicating poorer
survival in macroalgal areas. The low survival of fourspine stickleback in macroalgal areas made it difficult
to determine if the growth difference between the habitats was statistically important. In both habitats, fish
growth rates were relatively low (range: 0.06–0.08 g 21 dÀ1; 0.0–2.2 mm 21 dÀ1). This was perhaps because
the experiment was run late in the growing season (late August to September).


                           DISCUSSION

These results indicate that anthropogenic nutrient enrichment is causing a shift in primary producers and
altering fish and invertebrate communities and food webs in estuarine ecosystems. We found both

Copyright # 2002 John Wiley & Sons, Ltd.               Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                  207

watershed nutrient loading induced and experimentally manipulated alteration in macroalgal biomass
resulted in changes in fish and invertebrate abundance and diversity, as well as fish survival and growth.
Several aspects of our study indicate that the finely branched, filamentous macroalgal mats that replaced
eelgrass do not provide support to fish equivalent to that of eelgrass ecosystems.
  We observed an increase in macroalgae and a decline in eelgrass biomass in response to increased
nutrient loading as has been found worldwide (Duarte, 1995; Valiela et al., 1997b; Raffaelli et al., 1998).
Macroalgae are a natural component of eelgrass ecosystems, but in pristine watersheds, such as Timms
Pond, with low nutrient conditions macroalgal biomass is low relative to eelgrass (Short and Burdick, 1996;
Duarte, 1995). Nutrient loading enhances the growth of macroalgae leading to increased biomass as we
observed in Hamblin Pond. Removal of macroalgae increased abundance of eelgrass suggesting that
excessive macroalgae interferes with eelgrass growth through light or space competition (Short et al., 1995;
Short and Burdick, 1996; Hauxwell et al., 2001).
  Fish abundance and diversity and decapod abundance decreased with increased nitrogen load and
macroalgal dominance. Other studies have found a greater abundance and diversity of fish and
invertebrates in seagrass habitats compared to bare areas (Orth et al., 1984; Virnstein, 1987) or macroalgal
habitats (Phil et al., 1994). Increased biomass of macroalgae as a result of eutrophication has been
suggested to reduce juvenile plaice (Pleuronectes platessa) recruitment in Sweden (Phil and van der Veer,
1992) and juvenile cod (Gadus morhua) recruitment in Norway (Tveite, 1984). Sogard and Able (1991)
found higher densities of fish and invertebrates in macroalgae (Ulva lactuca) compared to unvegetated
habitats, but macroalgae did not provide fishes with an equivalent substitute for eelgrass.
  The central importance of macroalgae in altering eelgrass ecosystem support of fish and decapod
crustacea is evident when macroalgal biomass was manipulated. When macroalgal biomass was lowered,
eelgrass shoot density increased 2–10-fold and fish and decapod abundances were higher compared to other
areas with high macroalgal biomass. Interestingly, despite the higher abundance of invertebrate prey in the
high macroalgal biomass treatment, fish and decapods preferentially occupied the low macroalgal biomass
treatment.
  Excessive macroalgae may alter food webs by changing the physical or chemical structure of the
habitat. Both the 15N enrichment experiment and the growth of fishes in eelgrass and macroalgal
habitats indicated that little of the macroalgal production was transferred to the dominant fishes. We
found that 15N enriched macroalgae could be traced to herbivorous and detritivorous invertebrates, but
that little was found in the most abundant fishes. The high 15N value for Gammarus mucronatus is
consistent with feeding trials that have shown that Gammarus is an important herbivore on macroalgae in
Waquoit Bay (Hauxwell et al., 1998). The lack of 15N enrichment in the dominant fishes and the lower
growth of fishes suggest that macroalgal habitats are not providing the same food web support as eelgrass
habitats.
  Alteration of the physical structure of the habitat resulting from the increase in compact, filamentous and
finely branched algal species may prevent fish from feeding on benthic invertebrates. Macroalgae hinders
predator efficiency and provides a refuge for invertebrates (Heck and Thoman, 1981; Kulczycki et al., 1981;
Wilson et al., 1990; Isaksson et al., 1994; Dorf and Powell, 1997). Algal biomass and morphology have been
shown to be important in preventing fish from feeding on amphipods, crabs and shrimp (Holmhund et al.,
1990). Both laboratory and field studies have confirmed that fish have more difficulty capturing prey in fine,
filamentous algae compared to broad bladed algae, eelgrass or bare substrate (Dean and Connell, 1987;
Isaksson et al., 1994; Preisser and Deegan, 1995).
  Low dissolved oxygen near the bottom associated with high levels of macroalgal biomass may be a
chemical barrier that prevents fish from feeding on the benthos. Many fish will not cross a dissolved oxygen
cline to feed on the bottom if dissolved oxygen concentrations are less than 2–3 ppm (Nestlerode and Diaz,
1998). O2 concentrations of 2 ppm or lower are typical in high macroalgal biomass areas (this study,
D’Avanzo et al., 1996). Oyster toadfish are very tolerant of low dissolved oxygen levels and burrow in even

Copyright # 2002 John Wiley & Sons, Ltd.          Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
208                      L.A. DEEGAN ET AL.


anoxic substrate for shelter (Bigelow and Schroeder, 1953). These characteristics of oyster toadfish may
explain why this fish can exist and feed in macroalgal mats while other fish species do not. The alteration of
the physical and chemical structure of the habitat by macroalgae results in fewer and weaker connections
between in the benthic food web and the dominant fish species.
  The change in the physical and chemical structure of the habitat due to macroalgae may also affect
survival of small fishes. Fish and invertebrate survival is suggested to be higher in seagrass because of the
structural complexity of seagrass ecosystems compared to adjacent open areas (Heck et al., 1997). The
species of macroalgae that proliferate in response to nutrient enrichment tend to form unattached low-lying
mats that do not provide the same degree of structural complexity as a tall eelgrass canopy. Some fish have
impaired escape behaviors in hypoxic environments resulting in increased predation (Breitburg et al., 1997).
In severe cases short-term anoxic events (524 h) in eutrophied areas with high macroalgal biomass can kill
an entire year-class of fish, especially those with high site fidelity such as winter flounder (Deegan and
Buchsbaum, 2001).
  There are at least three implications of our work relevant to the conservation of eelgrass habitats and
managing ecosystems to prevent biotic impoverishment. Biologists agree that the major proximate causes of
biotic impoverishment today are habitat loss, degradation, and fragmentation (Soul! , 1991). First, our
                                             e
work suggests that all plants do not provide equal ecosystem function and that conversion of eelgrass
habitats to macroalgal dominated areas is the equivalent of habitat loss. Breitburg (1998) suggested that the
negative effects of nutrient loading on fish communities might be weaker if nutrient additions caused
changes in which species provides structure rather than a complete decline in macrophyte biomass. In
addition, compensatory replacement (some species increase while others decrease) has been suggested as a
mechanism that buffers the effects of stress at the ecosystem level resulting in similar levels of production
and biomass despite a change in species composition (Fogarty and Murawski, 1998; Rapport and
Whitford, 1999). Macroalgae, despite their high primary productivity and biomass, did not provide a
suitable alternative habitat for most eelgrass-dependent fishes and compensatory replacement by other
species did not occur.
  Second, our work suggests that we need to consider both the quality of the remaining habitat and
the surrounding ecosystem when applying habitat fragmentation theory to the conservation of eelgrass
habitats (Robbins and Bell, 1994). The study of declining or fragmenting habitats has been dominated
by two classical paradigms, island biogeography and metapopulation dynamics. Both of these approaches
invoke a habitat patch–matrix model where number of species is considered a function of habitat area
(species–area relationship) and isolation is measured as the distance across the matrix between habitat
patches (Rosenzweig, 1995; Ricketts, 2001). The habitat patch–matrix model has been successfully used
in conservation biology, however recent work has indicated that fuller development of this theory is
needed (Ney-Nifle and Mangel, 2000; Vos et al., 2001). An underlying assumption of the species-area
relationship is that the habitat remaining is equivalent in function to the original habitat (i.e., that patch
carrying capacity is constant; Vos et al., 2001). In this paper and others (Hughes et al., 2001; Wyda et al.,
2001), we have shown that that the loss of an ecosystem can be qualitative and involve a change
or degradation in the structure, function, or composition of an ecosystem (Noss, 1990). Increasing biomass
of macroalgae in an area that would still be considered eelgrass habitat decreases the capacity of
the ecosystem to support species even if the area of the habitat does not change. An additional
consideration is that seagrass patches maybe more isolated than simple distance would indicate depending
on the intervening matrix (Ricketts, 2001). Our work suggests that the physical and chemical structure
of the surrounding macroalgal matrix causes patches to be more effectively isolated than if they are
surrounded by sand or bare bottom (Robbins and Bell, 1994; Frost et al., 1999). For example, fish will
tend to remain in an eelgrass patch because the oxygen content of the eelgrass patch is more favourable
to their survival than the oxygen content of the surrounding macroalgal matrix. Thus, nutrient loading
leading to the conversion of eelgrass habitat to macroalgal dominated areas, degradation by increasing

Copyright # 2002 John Wiley & Sons, Ltd.          Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                    209

biomass of macroalgae within existing eelgrass habitat and fragmentation of eelgrass habitat into small
patches surrounded by high macroalgal biomass, has reduced the overall capacity of estuaries to support
animal populations.
 Third, our work suggests that macroalgal removal offers a potential short-term management option for
conservation of eelgrass habitat. The macroalgal biomass experiment shows that some of the effects of
macroalgae on plant, animal and chemical components of the ecosystem can be reversed by removal of the
macroalgae and that the results persist beyond a single growing season. This sustained response, at least at
the plant level, has also been seen in another study that examined the effect of macroalgae canopy height on
eelgrass (Hauxwell et al., 2001). While it is clear that survival of seagrass ecosystems ultimately depends on
reductions in nutrient loading, removal of macroalgae provides a short-term approach to maintaining
seagrass habitats while long-term controls on nutrient loading are developed.


                      ACKNOWLEDGEMENTS

The work presented in this paper could not have been accomplished without the contributions of the following people:
David Basler, Dixie Berthel, Robert Billard, Chris Capone, Brad Colvin, Susan Conant, Tania Lewandowski, Laura
Lynch, Kristin O’Brien, Melissa Weaver. We thank Jason Wyda, Jeffrey Hughes, Matt Cieri and an anonymous
reviewer for comments on the manuscript. This work was funded by grants from: NOAA National Estuarine Research
Reserve Program, The Mellon Foundation, The Jessie B. Cox Foundation, and the Environmental Protection Agency
(R823606-01-0, R825757-01-0) and the NOAA Office of Protected Resources (40AANF803410).


                         REFERENCES
Adams SM. 1976. The ecology of eelgrass, Zostera marina (L), fish communities. I. Structural analysis. Journal of
 Experimental Biology and Ecology 22: 269–291.
Bell DB, Westoby M. 1986. Importance of local changes in leaf height and density to fish and decapods associated with
 seagrasses. Journal of Experimental Biology and Ecology 104: 249–274.
Bigelow HB, Schroeder WC. 1953. Fishes of the Gulf of Maine. Fishery Bulletin 74, Fishery Bulletin of the Fish and
 Wildlife Service vol. 53, Washington, DC; 577pp.
Breitburg DL, Loher T, Pacey CA, Gerstein A. 1997. Varying effects of low dissolved oxygen on trophic interactions in
 an estuarine food web. Ecological Monographs 67: 489–507.
Breitburg DL. 1998. Scaling eutrophication effects between species and ecosystems: the importance of variation and
 similarity among species with similar functional roles. Australian Journal of Ecology 23: 280–286.
Cambridge ML, Chiffings AW, Brittan C, Moore L, McComb AJ. 1986. The loss of seagrass in Cockburn Sound,
 Western Australia. II. Possible causes of seagrass decline. Aquatic Botany 24: 269–285.
Cowper SW. 1978. The drift algae community of seagrass beds in Redfish Bay, Texas. Contributions in Marine Science
 21: 125–131.
D’Avanzo C, Kremer JN. 1994. Diel oxygen dynamics and anoxic events in a eutrophic estuary of Waquoit Bay,
 Massachusetts. Estuaries 17: 131–139.
D’Avanzo C, Kremer J, Wainwright SC. 1996. Ecosystem production and respiration in response to eutrophication in
 shallow temperate estuaries. Marine Ecology Progress Series 141: 263–274.
Dean RL, Connell JH. 1987. Marine invertebrates in algal succesion. III. Mechanisms linking habitat complexity with
 diversity. Journal of Experimental Marine Biology and Ecology 109: 249–273.
Deegan LA, Finn JT, Ayvazian SG, Ryder-Kieffer CA, Buonaccorsi J. 1997. Development and validation of an
 estuarine biotic index. Estuaries 20: 601–617.
Deegan LA, Buchsbaum RN. 2001. The Effect of Habitat Loss and Degradation on Fisheries. In The Decline of
 Fisheries Resources in New England: Evaluating the Impact of Overfishing, Contamination, and Habitat Degradation.
 Buchsbaum RN, Robinson WE, Pederson J (eds). Massachusetts Bays Program, University of Massachusetts Press:
 MA, in review.
Dennison, WC, Orth RJ, Moore KA, Stevenson JC, Carter V, Kollar S, Bergstrom PW, Batiuk RA. 1993. Assessing
 water quality with submerged aquatic vegetation. Bioscience 43: 86–94.
Dorf BA, Powell JC. 1997. Distribution, abundance, and habitat characteristics of juvenile tautog (Tautoga onitis,
 Family Labridae) in Narragansett Bay, Rhode Island, 1988–1992. Estuaries 20(3): 589–600.

Copyright # 2002 John Wiley & Sons, Ltd.            Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
210                       L.A. DEEGAN ET AL.


Duarte C. 1995. Submerged aquatic vegetation in relation to different nutrient regimes. Ophelia 41: 87–112.
Fogarty MJ, Murawski SA. 1998. Large scale disturbance and the structure of marine systems: fishery impacts on
 Georges Bank. Ecological Applications 8(1) Supplement: S6–S22.
Frost MT, Rowden AA, Atrill MJ. 1999. Effect of habitat fragmentation on the macroinvertebrate infaunal
 communities associated with the seagrass Zostera marina L. Aquatic Conservation: Marine and Freshwater Ecosystems
 9: 255–263.
Gore RH, Gallaher EE, Scotto LE, Wilson K. 1981. Studies on decapod crustacea from the Indian River region of
 Florida. XI Community composition, structure, biomass and species–areal relationships and algae-associated
 macrocrustaceans. Estuarine, Coastal and Shelf Science 12: 485–508.
Harlin MM, Thorne-Miller B. 1981. Nutrient enrichment of seagrass beds in a Rhode Island coastal lagoon. Marine
 Biology 65: 221–229.
Hauxwell J, McClelland J, Behr P, Valiela I. 1998. Relative importance of grazing and nutrient controls of macroalgal
 biomass in three temperate shallow estuaries. Estuaries 21: 347–360.
Hauxwell J, Cebrian J, Furlong C, Valelia, I. 2001. Macroalgal canopies contribute to eelgrass (Zostera marina) decline
 in temperate estuaries. Ecology 82: 1007–1022.
Heck KL Jr, Crowder LB. 1991. Habitat structure and predator-prey interactions in vegetated aquatic ecosystems. In
 Habitat Structure: The Physical Arrangements of Objects in Space. Bell SS, McCoy ED, Mushinsky ER (eds).
 Chapman & Hill: London, pp. 281–299.
Heck KL Jr, Nadeau DA, Thomas R. 1997. The nursery role of seagrass beds. Gulf of Mexico Science 15: 50–54.
Heck KL Jr, Wetstone GS. 1977. Habitat complexity and invertebrate species richness and abundance in tropical
 seagrass meadows. Journal of Biogeography 4: 135–142.
Heck KL Jr, Thoman TA. 1981. Experiments on predator–prey interactions in vegetated aquatic habitats. Journal of
 Experimental Marine Biology and Ecology 53: 125–134.
Holmhund MB, Peterson CH, Hay ME. 1990. Does algal morphology affect amphipod susceptibility to fish predation?
 Journal of Experimental Marine Biology and Ecology 139: 65–83.
Hoss DE, Thayer GW. 1993. The importance of habitat to the early life history of estuarine dependent fishes. American
 Fisheries Society Symposium 14: 147–158.
Houde DE, Rutherford ES. 1993. Recent trends in estuarine fisheries: predictions of fish production and yield. Estuaries
 16: 161–176.
Howarth R. 1988. Nutrient limitation of net primary production in marine ecosystems. Annual Review of Ecology and
 Systematics 19: 89–110.
Howarth RW, Billeti G, Swaney D, Townsend A, Jaworski N, Lajtha K, Downing JA, Elmgren R, Caraco N, Jordan
 T, Berendse F, Freney J, Kudeyarov V, Murdoch P, Zhao-Liang Z. 1996. Regional nitrogen budgets and riverine N
 and P fluxes for the drainages to the North Atlantic Ocean. Biogeochemistry 35: 75–139.
Hughes JE, Deegan LA, Peterson BJ, Holmes RM, Fry B. 2000. Nitrogen flow through the food web in the oligohaline
 zone of a New England estuary. Ecology 81(2): 433–452.
Hughes JE, Deegan LA, Wyda JC, Weaver MJ, Wright A. In press. The effects of eelgrass habitat loss on fish
 communities in estuaries of Southern New England, USA. Estuaries.
Hull SC. 1987. Macroalgal mats and species abundance: a field experiment. Estuarine Coastal and Shelf Science 25:
 519–532.
Isaksson I, Phil L, van Montfrans J. 1994. Eutrophication-related changes in macrovegetation and foraging of
 young cod (Gadus morhua L.): a mesocosm experiment. Journal of Experimental Marine Biology and Ecology 177:
 203–217.
Johnson DA, Welsh BL. 1985. Detrimental effects of Ulva lactuca (L.) exudates and low oxygen on estuarine crab
 larvae. Journal Experimental Marine Biology and Ecology 86: 73–83.
Kulczycki GR, Nelson WG, Virnstein RW. 1981. The relationship between fish abundance and algal biomass in a
 seagrass drift algae community. Estuarine Coastal and Shelf Science 12: 341–347.
Kushlan JA. 1981. Sampling characteristics of enclosure fish traps. Transactions of the American Fisheries Society 110:
 557–562.
Leber KM. 1985. The influence of predatory decapods, refuge, and microhabitat selection on seagrass communities.
 Ecology 66: 1951–1964.
National Research Council. 2000. Committee report on the causes and management of coastal eutrophication, R. W.
 Howarth (chair). National Academy Press, Washington, DC.
Nestlerode JA, Diaz RJ. 1998. Effects of periodic environmental hypoxia on predation of a tethered polychaete, Glycera
 americana: implications for trophic dynamics. Marine Ecology Progress Series 172: 185–195.
Ney-Nifle M, Mangel M. 2000. Habitat loss and changes in the species–area relationship. Conservation Biology 14: 893–
 898.

Copyright # 2002 John Wiley & Sons, Ltd.             Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
                 NITROGEN LOADING ALTERS SEAGRASS ECOSYSTEM                     211

Nixon SW, Oviatt CA, Frithsen J, Sullivan B. 1986. Nutrients and the productivity of estuarine and coastal and marine
 ecosystems. Journal of the Limnological Society of South Africa 12: 43–71.
Nixon SW. 1995. Coastal marine eutrophication: a definition, social causes and future concerns. Ophelia 41: 199–219.
Noss RF. 1990. Indicators for monitoring biodiversity: a hierarchical approach. Conservation Biology 4:
 355–364.
Orth RJ, Heck KL, van Montfrans J. 1984. Faunal communities in seagrass beds: a review of the influence of plant
 structure and prey characteristics on predator-prey relationships. Estuaries 7:339–350.
Orth RJ, Moore KA. 1983. Chesapeake Bay: an unprecedented decline in submerged aquatic vegetation. Science 222:
 51–53.
Peckol P, DeMeo-Anderson B, Rivers J, Valiela I, Maldonado M, Yates J. 1994. Growth, nutrient uptake capacities
 and tissue constituents of the macroalgae Cladophora vagabunda and Gracilaria tikvahiae related to site-specific
 nitrogen loading rates. Marine Biology 121: 175–185.
Peckol P, Rivers J. 1996. Contribution by macroalgal mats to primary production of a shallow embayment under high
 and low nitrogen loading rates. Estuarine Coastal and Shelf Science 43: 311–325.
Phil L, Wennhage H, Nilsson S. 1994. Fish assemblage structure in relation to macrophytes and filamentous epiphytes
 in shallow non-tidal rocky- and soft-bottom habitats. Environmental Biology of Fishes 39: 271–288.
Phil L, Van Der Veer HW. 1992. Importance of exposure and habitat structure for the population density of 0-group
 plaice, Pleuronectes platessa L. in coastal nursery areas. Netherlands Journal of Sea Research 29: 145–152.
Pickett STA. 1989. Space-for-time substitution as an alternative to long-term studies. In Long-Term Studies in Ecology:
 Approaches and Alternatives. Likens GE (ed.) Springer-Verlag: New York, p. 110–135.
Preisser MC, Deegan LA. 1995. Effect of changing plant morphology on invertebrate susceptibility to predation in
 eelgrass beds. Biological Bulletin 189: 242–243.
Raffaelli DG, Raven JA, Poole RJ. 1998. Ecological impacts of green macroalgal blooms. Oceanography and Marine
 Biology: Annual Review 36: 97–125.
Rapport DJ, Whitford WG. 1999. How ecosystems respond to stress. Bioscience 49(3): 193–203.
Ray GC. 1997. Do the metapopulation dynamics of estuarine fishes influence the stability of shelf ecosystems? Bulletin
 of Marine Science 60: 1040–1049.
Ricketts TH. 2001. The matrix matters: effective isolation in fragmented landscapes. The Amercian Naturalist 158(1):
 87–99.
Robbins BD, Bell SS. 1994. Seagrass landscapes: a terrestrial approach to a marine subtidal environment. Trends in
 Ecology and Evolution 9: 301–304.
Rosenzweig ML. 1995. Species Diversity in Space and Time. Cambridge University Press: New York; 436pp.
SAS Institute Inc. 1998. Statview II. SAS Institute, SAS Campus Drive, Cary, North Carolina.
Short TS, Burdick DM. 1996. Quantifying eelgrass habitat loss in relation to housing development and nitrogen loading
 in Waquoit Bay, Massachusetts. Estuaries 19: 730–739.
Short TS, Burdick DM, Klady JE III. 1995. Mesocosm experiments quantify the effects of eutrophication on eelgrass,
 Zostera marina. Limnology and Oceanography 40: 740–749.
Sogard SM, Able KW. 1991. A comparison of eelgrass, sea lettuce macroalgae, and marsh creeks as habitats for
 epibenthic fishes and decapods. Estuarine Coastal and Shelf Science 33: 501–519.
Soul! ME. 1991. Conservation: tactics for a constant crisis. Science 253: 744–750.
   e
Thrush SF. 1986. The sublittoral macrobenthic community structure of an Irish sea lough: effects of decomposing
 accumulations of seaweed. Journal of Experimental Marine Biology and Ecology 96: 231–242.
Tveite S. 1984. 0-group cod investigations on the Norwegian Skagerrak coast. The propagation of cod Gadus morhua L.
 Flodevigen Rapportser. Series. 1: 582–590.
Valiela I, Foreman K, LaMontagne M, Hersh D, Costa J, Peckol P, DeMeo-Anderson B, D’Avanzo C, Babione M,
 Sham C, Brawley J, Lajtha K. 1992. Couplings of watersheds and coastal waters: sources and consequences of
 nutrient enrichment in Waquoit Bay, Massachusetts. Estuaries 15: 443–457.
Valiela I, Collins G, Kremer J, Lajtha K, Geist M, Seely B, Brawley J, Sham CH. 1997a. Nitrogen loading from coastal
 watersheds to receiving estuaries: new method and application. Ecological Applications 7: 358–380.
Valiela I, McClelland J, Hauxwell J, Behr PJ, Hersh D, Foreman K. 1997b. Macroalgal blooms in shallow estuaries:
 controls and ecophysiological and ecosystem consequences. Limnology and Oceanography 42: 1105–1118.
Valiela I, Geist M, McClelland J, Tomasky G. 2000. Nitrogen loading from watersheds to estuaries: verification of the
 Waquoit Bay nitrogen loading model. Biogeochemistry 49: 277–293.
Virnstein RW. 1987. Seagrass-associated Invertebrate Communities of the Southern U.S.A.: A Review. Florida Marine
 Research Institute Publications No. 42, St. Petersburg, FL; 209pp.
Vitousek PM, Aber JD, Howarth RW, Likens GE, Matson PA, Schindler DW, Slesinger WH, Tilman DG. 1997.
 Human alteration of the global nitrogen cycle: sources and consequences. Ecological Applications 7: 737–750.

Copyright # 2002 John Wiley & Sons, Ltd.             Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)
212                       L.A. DEEGAN ET AL.


Vos CC, Verboom J, Opdam PFM, Ter Braak CJF. 2001. Towards ecologically scaled landscape indices. The American
Naturalist 183(1): 24–41.
Wilson KA, Able KW, Heck KL Jr. 1990. Predation rates on juvenile blue crabs in estuarine nursery habitats: evidence
for the importance of macroalgae (Ulva lactuca). Marine Ecology Progress Series 58: 243–251.
Wyda JC, Deegan LA, Hughes JE, Weaver MJ. In press. The response of fishes to submerged aquatic
vegetation complexity in two ecoregions of the Mid-Atlantic bight: Buzzards Bay and Chesapeake Bay. Estuaries.




Copyright # 2002 John Wiley & Sons, Ltd.            Aquatic Conserv: Mar. Freshw. Ecosyst. 12: 193–212 (2002)