Effects of macroalgal mats on sediment nutrient release and benthic-pelagic coupling

Jordan Stark, Skidmore College, Saratoga Springs, NY 12866

Abstract

Nutrient loading, one of the most severe human impacts on coastal bays, leads to a range of ecological effects including the decline of seagrass beds and proliferation of macroalgal mats as well as other changes in nutrient cycling, oxygen levels and habitat quality. Here, I examined the effects of the presence of macroalgal mats on fine-scale nutrient and oxygen dynamics and benthic-pelagic coupling in a controlled lab setting. Decomposition rates were high, leading to production of sulfides and high variability in many measures. However, oxygen was produced at the surface of the macroalgal mats, indicating that photosynthesis did occur. N:P ratios varied from 0.5 in the thick mat treatment to 36.7 in the no mat treatment, indicating strong N limitation in the presence of macroalgal mats and some P limitation in their absence. Over the course of five days, chlorophyll a concentrations were 21.8±2.0 μg/L in the no mat (control) treatment but only 0.2±0.08 μg/L in the thick mat treatment, even though inorganic nitrogen and phosphorus were both available. This study demonstrates clear ecological effects of macroalgal mats on nutrient stoichiometry and hints at non-nutrient mechanisms of benthic-pelagic decoupling in the presence of macroalgae.

Key Words

Macroalgal mats; nutrient cycling; decomposition; eutrophication; benthic-pelagic coupling

Introduction

Marine systems are generally limited by nitrogen and sensitive to human nitrogen inputs

(Vitousek and Howarth 1991). This is particularly evident in coastal systems, where productivity is high and human inputs of nutrients can affect ecosystem structure and function. On Cape Cod, nitrogen loading is particularly high due to nutrient inputs from wastewater and fertilizers

(Valiela et al. 1992). Nutrient loading has been identified at the major stressor to Waquoit Bay, a shallow estuary in Falmouth and Mashpee, MA, and other area estuaries, leading to a wide range of problems including anoxic events, loss of fish habitat and changes to the food web (Serveiss et al. 2004, Valiela et al. 1992, D'Avanzo and Kremer 1994).

One of the most noticeable effects of eutrophication is the replacement of high-quality habitat seagrass beds with thick mats of macroalgae which may reduce light levels and crowd out seagrasses (Hauxwell et al. 2003, McGlathery 2001, Valiela et al. 1997). In some systems, these mats can be more than 50 cm thick during blooms, making them the major primary producers and causing daily fluctuations in respiration and nutrient uptake (Krause-Jensen et al. 1999,

Valiela et al. 1992). It is not clear how such thick accumulations of macroalgae can persist, since light is extinguished in the top few cm of the mat and decomposition occurs rapidly below the light layer (Peckol and Rivers 1996). When mats of macroalgae accumulate, nitrogen limitation may become more severe due to the low C:N ratio of macroalgae (~6:1) when compared to seagrasses (~20:1) (Duarte 1990). This can reduce phytoplankton biomass, as phytoplankton are heavily reliant on recycled nitrogen from the sediments that may be intercepted by algal mats.

Here, I examine how macroalgal mats take up nutrients and alter nutrient cycling at the sediment-water interface using a novel system for sampling water at specific depths within macroalgal mats incubated in the lab.

In addition to taking up nutrients, the respiration and photosynthesis of thick mats of macroalgae create fluctuations in oxygen concentrations over the day-night cycle. Photosynthesis is high at the surface of the mat, so oxygen levels can become supersaturated with oxygen trapped in the physical structure of the mat (Herbert 1999, Krause-Jensen et al. 1996). However, respiration levels are also high, so anoxic events may occur when light levels are insufficient to sustain high photosynthesis rates. In Waquoit Bay, anoxic events have occurred during periods of low light and high temperatures, leading to fish kills, while hypoxia occurs on a nearly daily basis at night during the summer (D'Avanzo and Kremer 1994). Light cannot penetrate to the bottom of thick algal mats, so in undisturbed areas rapid decomposition also occurs, increasing oxygen demand (Peckol and Rivers 1996).

In addition to effects on animals, nitrifying and denitrifying bacteria may be impacted by nutrient demand and steep gradients in oxygen caused by macroalgal photosynthesis and respiration. Ammonia released by decomposition of organic matter in sediments may be converted to nitrate under aerobic conditions by chemoautotrophic nitrifying bacteria that fix carbon using energy from the oxidation of ammonium. Conditions that promote nitrification

(high oxygen and potentially high ammonium levels) occur primarily in the photosynthetically active layer of the algal mat. The nitrate produced by these bacteria can then diffuse into the anaerobic zones of the algal mat (below the light layer) where it can be used as an electron acceptor by denitrifying bacteria. These two processes may be coupled, removing ammonium from the system by first nitrifying and then denitrifying it. Photosynthetically active macroalgal mats can either increase coupled nitrification/denitrification if oxygen is limiting for the nitrifying bacteria or decrease coupled nitrification/denitrification rates if the macroalgae outcompete the bacteria for ammonium or nitrate (Dalsgaard 2003, Krause-Jensen et al. 1999).

Depending on the overall effect on denitrification, macroalgal mats may worsen or alleviate some human nutrient loading to coastal systems where they occur. For this study, I artificially isolated macroalgal mats in mesocosms with sediment to test the effects of macroalgal mat presence and thickness on nutrient uptake, oxygen fluctuations, denitrification rates and benthic-pelagic coupling. I also examined the decomposition rate of the mats using three different models. The physical structure of macroalgal mats is an important factor in their effects on ecosystems, so I created sampling devices with permanently installed tubes at various depths within the macroalgal mats to allow sampling without disturbing the mats. I then sampled these over daily and weekly time periods to test three hypotheses:

1. Greater densities of macroalgae increase fluctuation in oxygen levels temporally and

spatially.

a. High oxygen should be present near the surface of the mat and during the day,

while anoxia should occur at night and at the base of the mat where light does not

penetrate.

2. Macroalgae take up a significant fraction of ammonium and phosphate released from

sediments, decreasing nutrient levels in the overlying water.

a. Temporal and spatial patterns in nutrient concentrations should occur, including

nutrient uptake by macroalgae and trapping of nutrients within the physical

structure of macroalgal mats.

b. Macroalgal mats should reduce phytoplankton biomass in the overlying water.

3. Coupled nitrification and denitrification will change based on the interaction between

nutrient levels and oxygen levels in two possible ways.

a. Nitrification should produce high nitrate levels where oxygen is present at the

surface of the mat, while denitrification should reduce nitrate levels and produce

N2 gas near the base of the mat. b. If macroalgae outcompete bacteria for nitrogen, nitrification and denitrification

will decrease with increased macroalgal presence.

Methods

Mesocosm preparation and maintenance

On 11 November, I collected macroalgae and sediment from the mouth of the Child’s

River where it enters Waquoit Bay in Falmouth, MA. I collected the sediment from near the edge of the water at low tide and macroalgae from slightly further up the Child’s. The macroalgae was in mats about 10 cm thick with a mixture of Gracilaria and Ulva, which I collected in the proportion that they occurred in the field (about 80% Gracilaria, 20% Ulva). I brought the algae and sediment back to the lab, where I sorted through it to remove animals, including snails, sea cucumbers and grass shrimp. I then stored this macroalgae in a slow flow- through tank with running seawater until I established my mesocosms. On 12 November, I sieved the sediment with a 4 mm coarse sieve to remove animals and large pieces of plants and homogenized all collected sediment to ensure even distribution of sediment among the tanks.

On 13 November, I began establishing the mesocosms. First, I placed sampling devices made from PVC and plastic tubing into each of six clear fiberglass mesocosms (30 cm diameter).

Each sampling device had tubes running from specified depths to the surface of the mesocosm to allow me to take samples with minimal disturbance to the macroalgal mat (Figure 1, Table 1).

These tubes each had a volume of 15 mL. I then added sieved sediment to about 10 cm above the base of the mesocosms and gently filled them with seawater to 35 cm above the sediment. On 14

November, I added 10 cm of the mixed macroalgae to the thin mat treatment tanks (n=2) and 40 cm of macroalgae to the thick mat treatment tanks (n=2). I did not add any algae to the sediment only treatment tanks (n=2). I then filled all tanks with seawater to 70 cm above the sediment

(total volume=51 L water) and filled a larger tank surrounding all six experimental tanks with water to 40 cm to buffer temperature changes. At this time, I established a 12 hour day/night cycle with a growth light suspended over the tanks and placed an air stone in the overlying water of each tank (about 20 cm into the water) to keep it oxygenated. On 19 November, I wrapped black plastic around the mesocosms with macroalgae to the height of the macroalgal mat to reduce light entering the mat from the sides.

After establishment, I replaced about 15% of the water in each tank on 16 and 19

November with fresh seawater. On 20 November, after I completed sampling for nutrients, I replaced about 40% of the water in each tank to clear the water, which had become sulfurous and cloudy due to decomposition. I measured temperature daily over the course of the experiment; in all cases it stayed between 17 and 22 °C, with the temperature generally being warmest at the end of the day due to heating from the lights and coolest in the morning before the lights turned on.

Dissolved oxygen measurements

I inserted an oxygen microelectrode into a cut off one milliliter syringe and measured DO in small subsamples of water from each depth directed into the syringe through a three-way valve (to avoid atmospheric contamination). I measured dissolved oxygen levels at each sampling point (Table 1) on 19 November at 07:00 (before the lights turned on) and 18:40

(before the lights turned off) and on 21 November at 10:30.

Nutrient sampling and analysis

On 14 November (Day 0), four hours after adding the macroalgae, I collected 16 initial samples for nutrient analysis. To determine how nutrient gradients became establish, I collected water samples from all sampling points at 7:00 on 17 November (Day 3) and from 20 sampling points at 07:00 on 19 November. To determine the variation in nutrient concentrations over the daily cycle, on 20 November I collected nutrient samples from all sampling points at 06:30,

12:20 and 18:15.

In all cases, I flushed each tube by removing 25 mL water and air, then removed additional sample (15-50 mL depending on what analyses I was conducting). I immediately filtered each sample with a GFF. Some samples were analyzed within 24 hours; all other samples for ammonium and phosphate analysis were acidified with 1 μL 5 N HCl per mL sample.

Samples for nitrate analysis were frozen until analysis.

I analyzed ammonium in all samples (including initial samples and samples on Nov. 17,

19, and 20) using the modified Strickland and Parsons (1972) method based on the Solorzano

(1969) phenol-hypochloric method. I analyzed phosphate in samples from 20 November at all timepoints using a modification of the method from Murphy and Riley (1962). I analyzed nitrate in selected samples from 20 Nov. and from the 15N addition using a modification of the cadmium reduction method from Wood et al (1967) with Lamotte ® kits. To increase precision and accuracy, color intensity was measured on a Shimadzu UV-1800 Spectrophotometer and standards ranging from zero to 100 μM were used to convert from absorbance to concentration.

Sulfide sampling

Over the course of my experiment several tanks of macroalgae began to decompose and give off a sulfurous smell. Given this, I took samples on 21 November at 10:30 to determine sulfide concentrations. I used the method from Gilboa-Garber (1971) with 25 μL sample from each depth and analyzed the results spectrophotometrically on a Shimadzu UV-1800

Spectrophotometer on 25 November. Chlorophyll sampling

Between 22 November and 25 November, I took a 1 L surface water sample from each tank every day to determine chlorophyll levels. I filtered each using a 47 mm GF/F filter immediately following collection. On 21 November, I also took samples, but filtered them with a

GF/D filter. I froze all samples after filtering and extracted the chlorophyll in acetone on 2

December. On 3 December, I determined the chlorophyll concentrations using a modification of the method from Lorenzen (1967).

Denitrification rate

At 15:40 on 25 November, I withdrew 40 mL of water from the 5 cm depth within the algal mat and mixed in 5 mL of 100 mM 15N-labeled nitrate and 5 mL 100 mM bromine tracer.

In a separate syringe, I withdrew 30 mL water from the 5 cm depth. I then reinjected the 15N labelled mixture (50 mL) followed by the 30 mL of unlabeled water to ensure that all labeled nitrate and tracer was purged from the sample tubes. After 35 minutes, 1 hour 40 minutes and 3 hours I removed samples from the 5 cm depth tube in each tank. I created a siphon, flushed an

Exetainer tube and collected a sample. The water was preserved with 20 µL of saturated ZnCl, then capped and stored underwater in a cooler until analysis on 6 December. At the same time, I collected about 10 mL water from the same tube for bromine and nitrate analysis. I capped these

30 samples tightly and refrigerated until analysis on 9 December. I analyzed these samples for N2

29 and N2, using the isotope pairing technique to determine denitrification rates (Steingruber et al.

2001).

Results

Dissolved oxygen concentrations

In all tanks with algal mats, average oxygen levels within the algal mat were low (~ 1.0 mg/L) before the lights turned on and increased during the day to ~3.0 mg/L, while in tanks without mats oxygen levels remained constant between 6 and 7 mg/L (Figure 2). Depth profiles indicate that in the no mat treatment oxygen levels are consistent over both depth and time, while in the thick mat treatment oxygen increased over the course of a day from 1 mg/L to 6 mg/L at the surface of the mat (Figure 3). Increases deeper in the mat were smaller. The same trend occurred in the thin mat treatment, but variability was high.

Average nutrient concentrations

Concentrations of phosphate and nitrate varied significantly among treatments

(F2,164=223.5, P<0.0001; F2,27=18.71, P<0.0001 respectively), while concentrations of ammonium did not vary among treatments (p=0.5854; Figure 4). Phosphate concentrations

(weighted average ± SE) were very high in the thick mat treatment (18.4±4.1 μM), and much lower in the thin mat (1.7±0.5 μM) and no mat (0.3±0.1 μM) treatments. Nitrate was highest in the thin mat treatment (18.2±1.8 μM), and lower in the thick mat (7.2±0.6 μM) and no mat

(9.7±1.7 μM) treatments.

Phosphate concentrations in were not significantly different in the water column than within the macroalgal mat for the thick mat treatment (P=0.3711) or the thin mat treatment

(P=0.4169; Figure 5). For the thin mat treatment, ammonium concentrations were not significantly different in the mat and water column (P=0.0883). However, in the thick mat treatment ammonium concentrations were higher within the mat (P=0.0023; Figure 6). Nitrate concentrations were measured only at the surface of mats and in the no mat treatment, so comparisons between the mat and water column are not possible.

The DIN:P ratio (molar) was 0.5 in the thick mat treatment, 11.6 in the thin mat treatment and 36.7 in the no mat treatment (Table 2).

Nutrient profiles over time

Nutrient profiles varied over time after establishment of the mesocosms (Figure 7) but on average did not show any trends (Figure 8). Neither ammonium nor phosphate decreased significantly over the course of 20 Nov. in any treatment. Variability in nutrient concentrations was high, and in general there were no significant differences among treatments (Table 3). In the thick mat treatment, there were significant differences in ammonium concentration over time, but the average concentrations decreased over the course of the day.

While the average concentrations remained constant, there were differences in the shape of the profiles between treatments (Figures 9, 10). Variability in ammonium concentration

(coefficient of variation) increased toward the surface of the mat for the thick mat treatment

(p=0.0249; Figure 11). The thin mat exhibited the same trend, but this trend was not significant

(p=0.1688; Figure 11).

Sulfide concentrations

Sulfide concentrations were low (<0.5 mM) in all samples except those below 20 cm in tank 3 (thick mat treatment) where concentrations reached 1.5 mM (Figure 12). In tank 1 (thin mat treatment) some sulfides (0.1-0.2 mM) were present within the macroalgal mat, but did not occur in the overlying water. Similarly, in tank 3 (thick mat) concentrations were high within the mat and declined rapidly at the surface of the mat. However, in tank 2 (thin mat) sulfides were not present at any depth, while in tank 5 (thick mat) sulfides were only present below 10 cm. Chlorophyll concentrations in overlying water

Phytoplankton abundance measured by chlorophyll concentrations in the water column above the mat or sediment varied slightly over time, but were consistently more than 20 times higher in the no macroalgae treatment (21.8±1.99 μg/L (mean ± SE)) than in the thick mat treatment (0.9±0.3 μg/L) or the thin mat treatment (0.2±0.1 μg/L; Figure 13).

Denitrification rates

29 30 15 Denitrification was measured by the production of N2 and N2 after the NO3 addition.

30 2 N2 increased significantly in both tanks of the thin mat treatment tanks over time (R =0.9523,

2 30 P<0.0001; R =0.9916, P<0.0001; Figure 14). In one thick mat treatment, N2 also increased significantly over time (R2=0.9351, P<0.0001), but in the other thick mat treatment there was no

29 change over time (P=0.0868). N2 did not increase in any tanks after the addition, indicating there was no coupled nitrification-denitrificaton. The total denitrification rate was 0.215±0.073

μM/day in the thick mat treatment and 0.017±0.017 μM/day in the thin mat treatment.

Discussion

Oxygen and Gross Production

Dissolved oxygen concentrations remained constant in the no mat treatment but increased during the day in both treatments with macroalgae, indicating that production occurred even in the tanks where decomposition was rapid. Oxygen concentrations increased most at the surface of the algal mats (top 5-10 cm), particularly in the thick mat treatment (Figure 3). This corresponds with PAR measurements in Gracilaria mats by Peckol and Rivers (1996), who found PAR reduction of 90% 4 cm into the mat and nearly complete light extinction 16 cm into the mat. This surface photosynthesis produced enough oxygen to increase average DO over the course of a day (Figure 2), even though decomposition was high.

Nutrient concentrations and stoichiometry

Despite the diel pattern in oxygen, clear daily patterns in nutrient concentrations were not evident in these systems. However, the presence of macroalgal mats, both thick and thin, did have a significant effect on nutrient dynamics. One striking result is the change in N:P ratios between the treatments from 0.5 in the thick mat treatment to 36.7 in the no mat treatment (Table

2). This suggests strong N limitation of autotrophs in the thick mat treatment that is not matched in the other treatments. While N limitation is common in coastal and marine systems, other studies have found that phosphorus may actually be limiting in Waquoit bay during the late fall, when the macroalgae for this study was collected (Peckol et al. 1994). However, phosphate concentrations were not significantly different within the macroalgal mats and the water column, indicating that phosphorus was not taken up in large quantities.

Ammonium concentrations were significantly lower in the water column than in the macroalgal mat in both the thin and thick treatments (Figure 4). This is indicative of nitrogen uptake by macroalgae or denitrification by bacteria within the tanks. Phytoplankton, while present in all treatments, were at very low concentrations when macroalgae were present, so their nitrogen uptake (based on an assumed C:Chl ratio of 50 and the Redfield ratio) accounts for less than 25% of the nitrogen decrease in both treatments. Therefore, even though decomposition was the dominant process occurring in these mesocosms, benthic-pelagic decoupling did occur when macroalgae were present.

Based on the difference between measured phosphorus concentrations between the algal mat and no mat treatments and the Redfield ratio, N uptake in the thick mat treatment was very high (about 280 μmol/L). However, this result assumes that all phosphorus release was caused by decomposition of macroalgae and therefore accompanied by nitrogen in the Redfield ratio.

However, phosphorus release from sediments can also occur under anoxic conditions, particularly when sulfides are present (Jensen et al. 1995). While this may explain some of the phosphate release, the N:P ratio is so skewed that it is likely that nitrogen was taken up or released in a form that we did not measure in this study.

The most obvious explanation for the accumulation of phosphorus, but not nitrogen would be a high denitrification rate. While the addition of 15N labeled nitrate did not result in a

30 large production of N2 gas, denitrification may have occurred at higher levels at other times.

For example, denitrification may be coupled on a temporal, rather than spatial scale. This has been demonstrated with the presence of benthic microphytes encouraging nitrification during the day and denitrification at night (Risgaard-Petersen et al. 1994). The large fluctuations in dissolved oxygen levels over time in these tanks (Figure 3) may have promoted such coupling, and denitrification rates would need to be measured at night to examine this possibility.

However, the presence of sulfides can inhibit both nitrification and denitrification, leading to the production of nitrous oxide rather than N2 gas (Seitzinger 1988). This mechanism may have prevented denitrification in these experimental systems due to high sulfide concentrations (up to

1.5 mM; Figure 12), but is probably less important in most natural systems.

Another explanation for low measured nitrogen levels is bacterial uptake. Particularly in the thick mat treatment tanks, sulfide oxidizing bacterial activity was evident from films of white sulfurous material. Around the time of nutrient sampling, elemental sulfur or other white sulfur compounds were present in both thick mat tanks at such high concentrations that it was impossible to see the surface of the algal mat through the overlying water. Sulfur oxidizing bacteria (Thioploca) have been shown to store nitrate in vacuoles for later use as an electron acceptor (Jørgensen and Gallardo 1999), which would be consistent with the observed sulfur availability and low nitrate concentration in the thick mat treatment relative to the thin mat treatment.

The macroalgal mats may have also directly taken up some nitrogen; even though decomposition levels were high, photosynthesis did significantly increase oxygen concentrations over the course of a day (Figure 2). This occurred even in the thick mat tanks, which had very cloudy, sulfurous water. While average oxygen levels showed an increase over the day, oxygen profiles were highly variable. This may be partly an artifact of the experimental setup due to lack of mixing, but variability within macroalgal mats also occurs in nature. In fact, variability in both oxygen and nutrient levels may promote coupled nitrification and denitrification, as well as other processes which have different nutrient and oxygen requirements (Seitzinger 1988).

Decomposition rate

Portions of all algal mats decomposed over the course of the study, leading to anoxic zones within the tanks and sulfate reduction to sulfides (Figure 12) and elemental sulfur. Both tanks with thick mats of algae had water that became milky-white and smelled strongly of sulfur as the experiment proceeded. Clearly, decomposition was an important process affecting the nutrient cycling and oxygen dynamics within these mats, even over the relatively short time frame of this study. However, I did not have a direct measure of decomposition rate. Therefore, I estimated decomposition using three different methods and compared the results.

First, I calculated a range of expected decomposition values for my algal mats based on the results of Peckol and Rivers (1996), who measured photon flux within algal mats. Using their equations for the biomass of macroalgal mats of given height (biomass (g dry wt/m2) = thickness(cm)*36.2-12.7) and for the decomposition rate of Gracilaria below the light layer in algal mats (percent of tissue remaining= -2*days+105.2), I calculated expected decomposition rates for both treatments. Second, I calculated respiration rates based on the change in oxygen from the end of the day to the beginning of the day, and assumed that this respiration was all due to decomposition. Based on this respiration rate and an RQ of 1, I calculated the carbon loss.

Third, I used measured phosphate concentrations and the Redfield ratio to calculate the expected carbon loss needed to produce this concentration of phosphorus over a week.

Decomposition model 1 (Peckol and Rivers 1996) produced expected decomposition values of 5.5-7.9% of the thick mat and 0-5.3% of the thin mat over the course of a week (Table

3). Lower values result from assuming that decomposition occurs below 15 cm (light extinction point), while higher values result from assuming that decomposition occurs below 4 cm (10% of mat surface irradiance). Model 2 (based on respiration rates) produced 10% decomposition for the thick mat and 7.3% decomposition for the thin mat. Model 3 (based on phosphorus release and the Redfield ratio) produced 1.9% decomposition for the thick mat and 0.6% decomposition for the thick mat (Table 3).

All three methods resulted in high decomposition levels, which correspond with the appearance of the mesocosms and with previously published findings of 55% per month decomposition in unlit portions of Gracilaria mats (Peckol and Rivers 1996). Of the three methods, estimating decomposition based on phosphate concentration produced the lowest rate.

This relatively low rate may be the result of release in ratios slightly off from the Redfield ratio or phosphate uptake by the surface of the macroalgal mat, bacteria or phytoplankton. The fact that this method produces a low rate supports the assumption that all released phosphorus came from decomposition; if phosphorus was released from the anoxic sediment, this method should produce a very high decomposition rate. The highest decomposition rate came from estimation based on respiration. This high rate could be the result of respiration that was not decompositional, such as respiration of living macroalgae.

Further comparisons of both decomposition rates and processes within macroalgal mats are needed to understand the survival of thick macroalgal mats in the field. While the rate measured here was similar to that measured in previous studies, some aspects of the process may be dissimilar (such as high production and accumulation of sulfides). Changing these variables may have additional impacts on nutrient cycling and benthic-pelagic decoupling.

Phytoplankton suppression

Chlorophyll a concentrations were significantly higher in the no mat treatment than in either of the other two treatments at all measured times (Figure 13). This is somewhat enigmatic, because although nutrient concentrations were not as high in the treatments with macroalgae, both phosphate and DIN were still available. However, phytoplankton bloomed only in the no mat treatment. This indicates that while benthic-pelagic coupling of nutrient cycles may be interrupted by the presence of macroalgal mats, other mechanisms may also be at play. These could involve changing ecosystem stoichiometry, enabling bacterial competition (such as by release of carbon from decomposing macroalgae), producing sulfide or many other changes.

Additional studies are needed to determine which of these mechanisms is important and whether any are relevant in the field.

Conclusions

In this study, macroalgal mats did not exhibit the predicted trends in nutrient gradients, but did demonstrate several other meaningful ecological effects. First, the high rate of decomposition compared well with decomposition measured in other studies, although further studies will be needed to determine if the process of decomposition was similar in the lab and in nature. This indicates that persistence of mats in natural systems probably requires some level of disturbance to cause turnover of the algae and periodic exposure to light and oxygen. The presence of macroalgal mats alters ecosystem stoichiometry (leading to very low N:P ratio in the thick mat treatment) and benthic-pelagic nutrient coupling. While this study did not show high rates of nitrification or denitrification, sulfide inhibition or other factors (such as diel coupling) may have led to an underestimation of these rates as they occur in nature. Finally, phytoplankton suppression was strong in all tanks with macroalgae. Since nutrients were still available, this hints at other mechanisms of benthic-pelagic decoupling and the reduction of phytoplankton populations in the presence of macroalgal mats.

Acknowledgements

Many thanks to: Ken Foreman for mentoring, help with design and carrying out project;

Anne Giblin for help with several analyses; Scott Lindell for providing tanks and experimental

15 layout; Jane Tucker for analysis of NO3 addition results; Sam Kelley for help with bromine analysis; Rich McHorney, Alice Carter, Sarah Nalven and Fiona Jevon for help advice and help with all steps of the project; Tyler Messerschmidt and Hannah Kuhns for help with setting up and sampling mesocosms.

Literature Cited

Dalsgaard, T. 2003. Benthic primary production and nutrient cycling in sediments with benthic

microalgae and transient accumulation of macroalgae. Limnology and Oceanography

48:2138-2150.

D'Avanzo, C., J. N. Kremer. 1994. Diel Oxygen Dynamics and Anoxic Events in an Eutrophic

Estuary of Waquoit Bay, Massachusetts. Estuaries 17:131-139.

Duarte, C. M. 1990. Seagrass nutrient content. Marine ecology progress series. Oldendorf 6:201-

207.

Gilboa-Garber, N. 1971. Direct spectrophotometric determination of inorganic sulfide in

biological materials and in other complex mixtures. Analytical Biochemistry 43:129-133.

Hauxwell, J., J. Cebrian, and I. Valiela. 2003. Eelgrass Zostera marina loss in temperate

estuaries: relationship to land-derived nitrogen loads and effect of light limitation

imposed by algae. Marine Ecology Progress Series 247:59-73.

Herbert, R. A. 1999. Nitrogen cycling in coastal marine ecosystems. FEMS microbiology

reviews 23:563-590.

Jensen, H. S., P. B. Mortensen, F. O. Andersen, E. Rasmussen, and A. Jensen. 1995. Phosphorus

Cycling in a Coastal Marine Sediment, Aarhus Bay, Denmark. Limnology and

Oceanography 40:908-917.

Jørgensen, B. B., V. A. Gallardo. 1999. Thioploca spp.: filamentous sulfur bacteria with nitrate

vacuoles. FEMS microbiology ecology 28:301-313.

Krause-Jensen, D., K. McGlathery, S. Rysgaard, and P. B. Christensen. 1996. Production within

dense mats of the filamentous macroalga Chaetomorpha linum in relation to light and

nutrient availability. Marine ecology progress series 134:207-216. Krause-Jensen, D., P. B. Christensen, and S. Rysgaard. 1999. Oxygen and nutrient dynamics

within mats of the filamentous macroalga Chaetomorpha linum. Estuaries 22:31-38.

Lorenzen, C. J. 1967. Determination of chlorophyll and phaeo-pigments: spectrophotometric

equations. Limnology and Oceanography 12(2):342-346.

McGlathery, K. J. 2001. Macroalgal blooms contribute to the decline of seagrass in nutrient-

enriched coastal waters. Journal of Phycology 37:453-456.

Murphy, J., J. P. Riley. 1962. A modified single solution method for the determination of

phosphate in natural waters. Analytica Chimica Acta 27:31-36.

Peckol, P., J. S. Rivers. 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.

Peckol, P., B. DeMeo-Anderson, J. Rivers, I. Valiela, M. Maldonado, and J. Yates. 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.

Redfield, A. C. 1958. The biological control of chemical factors in the environment. American

Scientist 46:230A-221.

Risgaard-Petersen, N., S. Rysgaard, L. P. Nielsen, and N. P. Revsbech. 1994. Diurnal Variation

of Denitrification and Nitrification in Sediments Colonized by Benthic Microphytes.

Limnology and Oceanography 39:573-579.

Seitzinger, S. P. 1988. Denitrification in Freshwater and Coastal Marine Ecosystems: Ecological

and Geochemical Significance. Limnology and Oceanography 33:702-724. Serveiss, V. B., J. Bowen, D. Dow, and I. Valiela. 2004. Using Ecological Risk Assessment to

Identify the Major Anthropogenic Stressor in the Waquoit Bay Watershed, Cape Cod,

Massachusetts. Environmental management 33:730-740.

Solorzano, L. 1969. Determination of ammonia in natural waters by the phenolhypochlorite

method. Limnol.Oceanogr 14:799-801.

Steingruber, S. M., J. Friedrich, R. Gächter, and B. Wehrli. 2001. Measurement of denitrification

in sediments with the 15N isotope pairing technique. Applied and Environmental

Microbiology 67:3771-3778.

Strickland, J. D. H., T. R. Parsons. 1972. A practical handbook of seawater analysis. Fisheries

Research Board of Canada 167:.

Valiela, I., J. McClelland, J. Hauxwell, P. J. Behr, D. Hersh, and K. Foreman. 1997. Macroalgal

blooms in shallow estuaries: controls and ecophysiological and ecosystem consequences.

Limnology and Oceanography 42:1105-1118.

Valiela, I., K. Foreman, M. LaMontagne, D. Hersh, J. Costa, P. Peckol, B. DeMeo-Andreson, C.

D'Avanzo, M. Babione, C. Sham, J. Brawley, and K. Lajtha. 1992. Couplings of

Watersheds and Coastal Waters: Sources and Consequences of Nutrient Enrichment in

Waquoit Bay, Massachusetts. Estuaries 15:443-457.

Vitousek, P. M., R. W. Howarth. 1991. Nitrogen limitation on land and in the sea: how can it

occur? Biogeochemistry 13:87-115.

Wood, E., F. Armstrong, and F. Richards. 1967. Determination of nitrate in sea water by

cadmium-copper reduction to nitrite. Journal of the marine biological association.UK

47:23-31.

Figures and Tables

Table 1. Depths (cm above sediment) for nutrient sampling in each treatment. Thick mat Thin mat No mat 60 60 60 45 45 40 41 15 20 40 11 10 39 10 5 37 9 35 8 30 7 20 5 10 3 5 1 2

Table 2. Average nutrient concentrations (±SE) in each treatment weighted by height of sampling interval to determine concentration throughout tank.

PO4 (μM) NH4 (μM) NO3 (μM) N:P ratio Thick mat 18.4±4.1 2.0±0.4 7.2±0.7 0.5 Thin mat 1.7±0.5 1.8±0.5 18.3±1.8 11.6 No mat 0.31±0.1 1.7±0.5 9.7±1.7 36.7

Table 3. Results of ANOVA tests on change in nutrient concentrations over the daily cycle. Treatment PO4 NH4 No mat F2,26=0.93 P=0.4072 F2,26=0.21 P=0.8146 Thick mat F2,63=0.22 P=0.8022 F2,63=0.32 P=0.7255 Thin mat F2,69=0.18 P=0.8352 F2,68=3.63 P=0.0317

Table 4. Modeled decomposition in each treatment over one week using three different models. Percent decomposition is based on algal mat biomass estimates from Peckol and Rivers (1996) in all cases. Ranges for model 1 are based on assumptions of decomposition occurring below 4 cm or below 15 cm in the algal mat. Decomp. model 1 Decomp. model 2 Decomp. model 3 (Peckol and Rivers) (Oxygen uptake) (PO4 release) dry wt. (g C) percent (%) dry wt. (g C) percent (%) dry wt. (g C) percent (%) Thick mat 1.6-2.2 5.5-7.9 1.2 10.0 6.0 1.9 Thin mat 0.0-0.4 0-5.3 0.1 7.3 1.1 0.6

Figure 1. Sampler used to collect water samples from different depths within thick algal mats. Samplers were constructed from PVC and thin plastic tubing and depths were chosen to provide detail at the surface of the algal mat (Table 1). Thin tubes extended out of tanks and were capped with stopcocks to allow water to be retained or released from tubing as needed and tight connection to syringes.

7

6

5

4

3

Dissolved (mg/L) Dissolved Oxygen 2

1

0 6:00 9:00 12:00 15:00 18:00 21:00 Time

Figure 2. Change in dissolved oxygen concentrations over time within the thick mat (blue), thin mat (red) and water from the no mat treatment (green). Thick mat and thin mat points show average dissolved oxygen at all points within the algal mat, while no mat points show the average of all sampled points. Error bars show standard error.

DO (mg/L) 0 1 2 3 4 5 6 7 8 70 60 50 40 30 20 10 0 -10

70

60 50 40 30 20 10 0

Height above sediment Height (cm)above sediment -10

70 60 50 40 30 20 10 0

-10

Figure 3. Change in dissolved oxygen levels over time from 07:00 (blue line) to 10:30 (red line) and 19:00 (green line) in the no mat treatment (top), thin mat treatment (middle) and thick mat treatment (bottom). Brown boxes represent the sediment; green boxes represent the algal mat. All points are the average of the two treatments for that depth and time ± SE.

25 thick mat thin mat no mat

20

M) μ 15

10 Concentration ( Concentration 5

0 Phosphate Ammonium Nitrate Figure 4. Concentrations (weighted average ± 1 SE) of nutrients in the thick mat treatment (blue), thin mat treatment (red) and no mat treatment (green). All points are averages over all depths and over three sampling points on Nov. 20.

20 Within Mat 18 Water Column

16 M)

μ 14 12 10 8 6

Concentration ( Concentration 4 2 0 thick mat thin mat

Figure 5. Phosphate concentrations within algal mats and in the water column above the mats. Each bar represents the average concentration over all sampled points at three sampling times on 20 Nov. Error bars are ± 1 SE.

5 Within Mat Water Column

4

M) μ 3

2

Concentration Concentration ( 1

0 thick mat thin mat

Figure 6. Ammonium concentrations within algal mats and in the water column above the mats. Each bar represents the average concentration over all sampled points at three sampling times on 20 Nov. Error bars are ± 1 SE.

Ammonium concentration (uM) 0 2 4 6 8 10 70 60 50 Day 0 40 Day 3 30 Day 5 20 Day 6 10 0 -10

70

60 50 40 30 20 10 0

Height above sediment Height (cm)above sediment -10

70 60 50 40 30 20 10 0 -10

Figure 7. Change in ammonium levels during establishment of mesocosms on day 0 (blue lines), day 3 (red lines), day 5 (green lines) and day 6 (purple lines) in the no mat treatment (top), thin mat treatment (middle) and thick mat treatment (bottom). Brown boxes represent the sediment; green boxes represent the algal mat. All points are the average of the two treatments for that depth and time ± SE.

16

M) 14 μ 12 10 8 6 4

2 Ammonium concentration ( Ammonium 0 0 2 4 6 8 Days after establishment Figure 8. Change in average ammonium concentration over time in the thick mat (blue), thin mat (red) and no mat (green) treatments. Each point is the average of all sampled points on that day (n=2-12 depending on day and treatment) ± 1 SE.

Ammonium (uM) 0 3 6 9 12 15 18 70 60 50 40 30 20 10 0 -10

70

60 50 40 30 20 10 0 Height above sediment Height (cm)above sediment -10 70 60 50 40 30 20 10 0 -10

Figure 9. Ammonium concentrations profiles at 07:00 (blue lines), 12:30 (red lines) and 19:00 (green lines) in the no mat treatment (top), thin mat treatment (middle) and thick mat treatment (bottom). All points are the average of the concentrations at a given depth in both tanks ± 1 SE.

Phosphate (uM) 0 5 10 15 20 25 70 60 50 40 30 20 10 0 -10 70 60 50 40 30 20 10 0

-10 70 60 50 40 30 20 10 0 -10 Figure 10. Phosphate concentrations profiles at 07:00 (blue lines), 12:30 (red lines) and 19:00 (green lines) in the no mat treatment (top), thin mat treatment (middle) and thick mat treatment (bottom). All points are the average of the concentrations at a given depth in both tanks ± 1 SE. 2.5

2 y = 0.0175x + 0.6199 y = 0.0547x + 1.2143 R² = 0.4911 R² = 0.2897 1.5

1

Coefficient variation of Coefficient 0.5

0 0 10 20 30 40 Height above sediment (cm) Figure 11. Variability (CV) of ammonium concentrations within the macroalgal mats of the thick and thin mat treatments.

Sulfide concentration (mM) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 70

60

50

40

30

20 Height above sediment (cm) sediment Height above 10

0

Figure 12. Sulfide concentrations in mesocosms with thick mats (blue), thin mats (red) and no mats (green) on 21 November. Concentrations were likely higher during the rest of the experiment because I changed out ~50% of the water on 20 November.

40

35

30

25

20

15

10

Chlorophyll A Concentration (ug/L) Concentration A Chlorophyll 5

0 21-Nov 22-Nov 23-Nov 24-Nov 25-Nov Date

Figure 13. Chlorophyll concentrations in mesocosms with thick mats (blue), thin mats (red) and no mats (green) over time.

125 y = 12.004x + 86.057 120 R² = 0.9523

115

110 y = 5.9254x + 83.404 105 R² = 0.9915 100

95 30N2 concentration (nM) 30N2 concentration 90 y = 0.5885x + 83.583 R² = 0.9351 85 y = -0.1048x + 84.507 R² = 0.3615 80 0 1 2 3 4 Time (hours after 15N addition)

30 Figure 14. N2 concentration in thick mat (blue) and thin mat (red) tanks over time after 15 addition of N labeled NO3.