Bates College SCARAB

Honors Theses Capstone Projects

Spring 5-2011 Nitrogen Isotopes in Zostera marina: a Potential Indicator of Anthropogenic Nutrient Loading in Casco Bay, Gulf of Maine Gregory Edward Flynn Bates College, [email protected]

Follow this and additional works at: http://scarab.bates.edu/honorstheses

Recommended Citation Flynn, Gregory Edward, "Nitrogen Isotopes in Zostera marina: a Potential Indicator of Anthropogenic Nutrient Loading in Casco Bay, Gulf of Maine" (2011). Honors Theses. 2. http://scarab.bates.edu/honorstheses/2

This Open Access is brought to you for free and open access by the Capstone Projects at SCARAB. It has been accepted for inclusion in Honors Theses by an authorized administrator of SCARAB. For more information, please contact [email protected]. Nitrogen Isotopes in Zostera marina: a Potential Indicator of Anthropogenic Nutrient Loading in Casco Bay, Gulf of Maine

An Honors Thesis

Presented to The Faculty of the Department of Geology Bates College

In partial fulfillment of the requirements for the Degree of Bachelor of Arts

by Gregory Flynn

Lewiston, Maine March 25, 2010 Table of Contents Acknowledgments...... v Abstract...... vi

1. Introduction...... 1 1.1 Summary of Objectives...... 2 1.2 The Nitrogen Cycle in Marine Environments...... 2 1.2.1 Nitrogen Isotopes...... 7 1.2.2 Fractionation...... 8 1.2.3 Anthropogenic Influence...... 9 1.3 Previous Studies...... 11 1.3.1 Previous Studies in Casco Bay...... 12 1.4 Purpose...... 15 1.4.1 Site Description...... 15 1.4.1.1 Geology...... 16 1.4.1.2 Land Use...... 21

2. Methods...... 24 2.1 Field...... 25 2.1.1 Mackworth Island and Maquoit Bay...... 25 2.1.2 Rivers...... 29 2.1.3 Open Ocean ...... 29 2.2 Laboratory...... 30 2.2.1 Grain Size Analysis...... 30 2.2.2 Chlorophyll...... 30 2.2.3 Nitrogen Isotope Analysis using the IRMS...... 31 2.2.3.1 Eelgrass...... 31 2.2.3.2 Sediment...... 31 i 2.2.3.3 Particulate Organic Matter ...... 31 2.2.3.4 Inorganic Nitrogen in Water...... 32 2.2.4 Statistic Analysis...... 38

3. Results...... 40 3.1 Standards ...... 41 3.2 Mackworth Island versus Maquoit Bay...... 45 3.2.1 Nitrate...... 45 3.2.2 Ammonium...... 45 3.2.3 Producers...... 48 3.2.3.1 Carbon and Nitrogen Isotopes...... 48 3.2.3.2 C/N Ratios...... 48 3.2.3.3 Chlorophyll ...... 48 3.2.3.4 Grain Size Analysis...... 48 3.3 Inputs: Rivers and Open Ocean...... 51 3.3.1 POM...... 51 3.3.2 Chlorophyll...... 51 3.3.3 Nitrate...... 51

4. Discussion...... 54 4.1 Source of DIN to Mackworth Island and Maquoit Bay...... 55 4. 2 Quantifying for Enriched Nitrogen in Eelgrass and Nitrate...... 55 4.2.1 Possible Causes of 15N Enrichment in Eelgrass and Nitrate...... 60 4.3 Temporal Variability of Eelgrass Isotope Data...... 66

5. Conclusions...... 70 References...... 73 ArcGIS Layers...... 77

ii Appendix 1...... 78 Appendix 2...... 79 Appendix 3...... 80 Appendix 4...... 81 Appendix 5...... 82 Appendix 6...... 83

iii Acknowledgments

I would like to start by thanking the Bates College Geology Department, some of the most enthusiastic, intellectual, and fun people I have spent time with. To Dysktra Eusden: thank you for supporting my love for the outdoors and showing me the incredible powers of education in the Field. To John Creasy: thank you for getting me through my first big Geology project and being supportive of my thesis work from the start. To Michael Retelle: thank you for being the teacher of my first college Geology class, and fostering my interest in the subject. It is my privilege to thank Beverly Johnson, who my thesis would not have been possible without. She spent endless hours with me in her office, in addition to the countless hours correcting my drafts. Her patience and knowledge helped me throughout the process, and she was always there for me, being more than willing to take time out of her night, weekend, or holiday. Her commitment to me and my project got me over even the worst times in the past year, only for the best. In addition, her enthusiasm and excitement promoted my interest in stable isotope geochemistry, introducing me to a field that I would have never found myself interest in without her. To Phil Dostie: thank you for taking care of me in the geochemistry lab, helping me with everyday tasks, bandaging my wounds from failed lab work, and cleaning up my messes. To William Locke: thank you for being great assistance to me in the field, and helping me understand instrumentation used in this thesis. To Matt Duvall and William Ash of the Bates College Imaging Center, thank you for helping me with GIS and InDesign tasks, and organizing my thesis. Thank you, Curtis Bohlen and the Casco Bay Estuarine Partnership for being a great support to this project, providing funding, data, and the opportunity to test samples from the outer Casco Bay. Thank you to the crew of the OSV Bold EPA ship for collecting ocean water sample for this project and giving me a great tour of the boat. Thank you, Mike Doan and Friends of Casco Bay for providing me with crucial data and insight to recent studies in Casco Bay.Thank you Hilary Neckles for offering insight to data and helping me understand some results. Thank you William Loopesko (Dots) for starting this project and working out the methods. I know you spent days in the geochemistry lab trying to get isotope data from ammonium and nitrate. Without your hard work, nitrate isotope data would not be available for this study. Thank you to my friends for spending nights in Coram and Carnegie basement with me, listening to my practice presentations, reading over my drafts, playing trash can basketball, going to late night commons, and making me confident about my rock-paper- scissors abilities. At last, thank you to my family. They are the biggest single support for whatever I do in life and have always pushed me to get the best education possible. They are the most inspirational people in my life.

iv Abstract

Estuaries that are in close proximity to densely populated areas and/or receive run- off from populated watersheds are particularly susceptible to nitrogen loading, which can lead to anthropogenic-caused eutrophication. In a past study by McClellend and Valiela (1998) stable nitrogen isotope ratios (δ15N) of dissolved inorganic nitrogen (DIN) and Zostera marina were proven to be enriched in 15N in densely populated estuaries relative to less poplulated estuaries in Cape Cod. Little is known about whether this technique for identifying the presence of anthropogenic nutrients can be used on the coast of Maine. In this paper, two areas in the Casco Bay were studied to see if a populated area (Mackworth Island, Portland) shows 15N enrichment relative to a less populated area (Maquoit Bay, Brunswick). The DIN and the Zostera marina from Mackworth Island were shown to have δ15N values ~2.5 ‰ enriched relative to Maquoit Bay, suggesting that nitrogen isotopes in Zostera marina can be used to detect the presence of anthropogenic nitrogen. This infor- mation has the potential to help indicate early signs of eutrophication and help prevent any further nutrient overloading in Casco Bay.

v Table of Figures Figure 1.1: The nitrogen cycle in marine environments...... 4 Figure 1.2: Nitrogen in an eelgrass meadow...... 6 Figure 1.3: Nitrogen isotopes in different sources (Kendall, 1998)...... 10 Figure 1.4: Nitrogen isotopes and trophic levels (Mack and Ostrom, 1994)...... 10 Figure 1.5: δ15N versus concentration of DIN in groundwater (McClellend and Valiela, 1998)...... 13 Figure 1.6: δ15N in groundwater versus δ15N of primary producers (McClellend and Valiela, 1998)...... 13 Figure 1.7: Nitrate/nitrite and ammonium concentrations in the Casco Bay (Libby and Anderson, 2010)...... 14 Figure 1.8: A simplified bedrock map of Casco Bay (Maine Geologic Survey, 2008) ...... 17 Figure 1.9: A surficial geology map of Mackworth Island and its surrounding area (Maine Geologic Survey, 2008) ...... 18 Figure 1.10: A surficial geology map of Maquoit Bay and its surrounding area (Maine Geologic Survey, 2008) ...... 20 Figure 1.11: Percent land use in the watersheds of Mackworth Island and Maquoit Bay (Loopesko, 2010)...... 23 Figure 2.1: A map of site locations used in this study ...... 26 Figure 2.2: Water sample collection and preparation ...... 28 Figure 2.3: The ammonium diffusion method adapted from Holmes et al. (1998)... 35 Figure 2.4: The nitrate diffusion method adapted from Sigman et al. (1997)...... 37

Figure 3.1: Results from running standards in the Holmes et al. (1998) method..... 43

Figure 3.2: Results from running standards in the Sigman et al. (1997) method ..... 44

Figure 3.3: Mean nitrate concentration versus δ15N of nitrate from Mackworth Island

vi and Maquoit Bay...... 47 Figure 3.4: Mean nitrogen versus carbon isotope ratios for POM, eelgrass, and sediment at Mackworth Island and Maquoit Bay...... 49 Figure 3.5: Grain size analysis from sediment samples taken at each site at Maquoit Bay and Mackworth Island...... 50

Figure 3.6: δ15N versus δ13C for POM in the Bunganuc River, Presumpscot River, and Fore River...... 52

Figure 3.7: Mean chlorophyll A levels (ug/L) in surface water in the Bunganuc River, Presumpscot River, and Fore River...... 52

Figure 3.8: Mean concentration of nitrate versus δ15N of nitrate for all the river and ocean water samples...... 53 Figure 4.1: δ15N of nitrate versus concentration of nitrate for Mackworth Island and its inputs and Maquoit Bay and its inputs...... 56 Figure 4.2: Average δ15N of eelgrass and nitrate at Mackworth Island and Maquoit Bay...... 57 Figure 4.3: Comparison of DIN sources for eelgrass at Mackworth Island and Maquoit Bay based on a model by Zimmerman et al. (1997)...... 59 Figure 4.4: Concentration of nitrate, ammonium, and total DIN colleceted by the Casco Bay Estuary Partnership in 2010...... 63 Figure 4.5: The locations of the sites in figure 4.4 ...... 64 Figure 4.6: A map of sewer overflow outfalls and wastwater treatment plant facility outfalls in the Mackworth Island area...... 67 Figure 4.7: A comparison of δ15N of eelgrass sampled at Mackworth Island and Maquoit Bay in differenc studies over the past 10 years ...... 69

vii Table of Tables

Table 2.1: A spreadsheet showing what samples were collected at each site, including geographical locations and dates collected...... 27

Table 2.2: Volumes required for analysis of ammonium (Holmes et al., 1998) and nitrate (Sigman et al., 1997) contents in order to recover adequate mirco moles of nitrogen to obtain accurate nitrogen isotope data...... 33

Table 2.3: Concentrations of ammonium and nitrate determined by North East Labs in Portland, Maine for all sample sites in this study...... 33

Table 3.1: Ammonium and nitrate diffusion method data for standards...... 42

Table 3.2: Results of t-test and ANOVA tests used to determine statistically significant differences in data...... 46

Table 3.3: Ammonia Recoveris: EPA 350.1 method versus Holmes et al. (1998) method...... 47

Table 4.1: Percent cover (density) and average height data for the eelgrass beds at Mackworth Island and Maquoit Bay collected in 2009 by the Casco Bay Estuarine Partnership...... 61

Table 4.2: Differences in temperature between Mackworth Island and Maquoit Bay on the days when water and eelgrass samples were collected...... 65

Table 4.3: Ammonium concentrations (uM) of pore water at Mackworth Island and Maquoit Bay. Concentrations were measured by North East Labs, using EPA method 350.1...... 65

viii 1. Introduction

1 1.1 Summary of Objectives Nitrogen is an essential element for all living organisms. It is essential for the formation of proteins, and nucleic acids (DNA and RNA), the two most important polymers of life (Canfield et al., 2010). Although nitrogen is necessary for all life, too much nitrogen can be harmful to ecosystems. Organic and inorganic fertilizers and human wastewater contain high concentrations of nitrogen, and are often transported into rivers, lakes, estuaries, and other water bodies. This increased nitrogen flux can result in eutrophic waters, particularly in coastal zones (Sharp, 2007). Estuaries that are in close proximity to densely populated areas and/or receive nitrogen rich inputs from large watersheds are particularly susceptible to anthropogenic- caused eutrophication. This can have devastating consequences for marine ecosystems. Efforts to establish nutrient criteria for coastal marine water are currently underway, but the challenge is in determining concentration levels that can be used simply and uniformly across the coast. Identifying sites that reflect the presence of excess nutrients prior to the onset of eutrophication is vital for determining nutrient criteria. In recent studies, stable nitrogen isotope analysis of eelgrass has been used to identify the presence of anthropogenic nutrients in Cape Cod (McClellend and Valiela, 1998). This approach has not yet been attempted in Casco Bay, Gulf of Maine. In this study, two sites in Casco Bay were examined for nitrogen isotope ratios to compare sources of nitrogen. One site is expected to have low anthropogenic nitrogen loading, while the other site is expected to have high anthropogenic loading due to its proximity to Portland and large river mouths. If nitrogen isotopes can be used to track sources of anthropogenic nitrogen in the Casco Bay, then they can potentially be used as an early indicator for excess delivery of anthropogenic nutrients and potential for eutrophication.

1.2 The Nitrogen Cycle in Marine Environments The large majority (97.98%) of nitrogen on earth is locked up in mantle and crustal rocks (Sharp, 2007), with an average residence time of 1 billion years (Canfield et

2 al., 2010), leaving only 2.02% of nitrogen readily available. Hence, the remaining nitrogen must be cycled through the atmosphere, hydrosphere, biosphere, and upper lithosphere in order for life to exists as it does (Sharp, 2007). The atmosphere is the largest reservoir of available nitrogen on earth, making up about 90% of available nitrogen (Schlesinger, 2007) (Figure 1.1). Almost all nitrogen

in the atmosphere is in its inert, diatomic form, N2, which is a very stable compound because of its triple bond. Due to the strength of this triple bond, most organisms are unable to assimilate atmospheric nitrogen directly (Sharp, 2007), so atmospheric nitrogen must undergo several steps in order to be in a state where it can be broken down by most organisms. In the first step certain bacteria and cyanobacteria containing the enzyme

nitrogenase convert N2 into ammonia (NH3) via fixation:

2 N2(g)+6H2O → 4NH3(g)+3O2 (Galloway, 2004)

The initial conversion, N2 → 2N, requires lots of energy to break the strong bond, about

226 kcal/mole of N2 (Schlesinger et al., 2004), while the second reaction, 2N+3H2 → 2NH3, releases a small amount of energy.

Once diatomic nitrogen is converted into ammonia gas (NH3), it is almost

+ immediately converted into ammonium (NH4 ) via protonation (Fogel and Cifuentes, 1993):

+ + H3O + NH3 → H2O + NH4 (Fogel and Cifuentes, 1993) Ammonium can be oxidized by nitrifying bacteria and converted into nitrite and then further oxidized to form nitrate via nitrification: + - 2NH4 + 3O2 → 2NO2 + 4H + 2H2O

- - 2NO2 + O2 → 2NO3 (Galloway, 2004) The bacteria use the energy released by the oxidation of ammonia to nitrite and the oxidation of nitrite to nitrate (Sharp, 2007). Ammonium and nitrate are the two primary sources of inorganic nitrogen that can be assimilated by most organisms and often limit primary production in marine ecosystems (Canfield et al., 2010). When organic matter dies, it decomposes and the nitrogen released from the 3 Anaerobic Aerobic

N2

F 0

ix . 7 ‰

a

tion Atmosphere 90% of available nitrogen + NH3/NH4 Nitri cation -29 to -18 ‰ n NO o 2 i A Marine t N s c s it

a

i r f - m 0 i

i R

2 c r e ‰

7 Ammoni cationd t

‰ u a

i i

t 1 5 c l 8 t o

- (Mineralization) t n a

to io ion

2 n 0 o

e t t 9

‰

i ‰

D o 0

4 - n 0 ‰ - NO3 n tio ila im 0‰ ss o Organic N A 7 t 2 Bacterial N - Plant N Dead Organic N

Figure 1.1: The nitrogen cycle in marine environments. A range in fractionation values are presented for each step, with values in per mil. Based on figures in Sharp (2007) and Canfield et al. (2010).

4 decomposition is converted back into ammonium in a process called ammonification or mineralization (Galloway, 2004). In aerobic zones (shallower areas) this ammonium is often recycled, where the ammonium can either be assimilated by organic matter again or be oxidized into nitrate (Canfield et al., 2010). In in deep ocean anaerobic zones or in estuaries where oxidized and anoxic sediments occur in close proximity, the decomposition of organic matter creates a large pool of inorganic nitrogen within sediments (Schlesinger, 1997). In this case much of the nitrogen is often returned to the atmosphere in a process called denitrification:

+ - 6CH2O+4H +4NO3 → 2N2+5CO2+7H2O (Galloway, 2004) In this reaction diatomic nitrogen diffuses out of water and recharges the nitrogen available in the atmosphere. If denitrification did not occur, atmospheric 2N would be exhausted in 100 million years (Sharp, 2007). In surface marine systems, concentrations of dissolved inorganic nitrogen (DIN), usually in the form of ammonium and nitrate, are generally very small because primary producers assimilate this nitrogen very quickly (Schlesinger, 1997; Jaffe, 1992). Primary producers, such as phytoplankton and benthic algae assimilate nitrate and ammonium dissolved in the water column, while other primary producers, such as seagrasses, assimilate nitrate and ammonium dissolved in the water column, and ammonium dissolved in the pore water (Figure 1.2) (Zimmerman et al. 1987; Romero et al., 2006). The proportion of each of these sources assimilated depends on the concentration of each. In eelgrass (Zostera marina) specifically, roots usually assimilate the majority of the nitrogen from ammonium in pore water. However, roots usually do not take up more than 70% of the total nitrogen, even when pore water ammonium concentration is very high (Zimmerman et al., 1987) since eelgrass leaves also assimilated nitrogen from the water column. Eelgrass assimilation of DIN also depends on the proportion of other primary producers in the system, including ephiphytes that attach to the eelgrass leaves and inhibit the leaves from assimilating DIN (Romero et al. 2006). Ammonium concentration in pore

5 Figure 1.2: A diagrammatic representation of the main fluxes of nitrogen in an eelgrass an eelgrass in of nitrogen fluxes the main of representation 1.2: A diagrammatic Figure (2006). et al. Romero from Adapted meadow.

6 water usually ranges from 50 to 500 umol/L in eelgrass beds (Short, 1983), but increased grain size of the sediments allows there to be more diffusion of ammonium out of the sediments into the water column (Macko and Aller, 1984). Marine environments receive nitrogen from four main sources: rivers, groundwater, biological fixation from green and blue algae in the open ocean, and precipitation. Rivers alone contribute about 40% of the total nitrogen influx to the seas and play a very significant role in coastal areas and estuaries (Schlesinger, 1997). There is a positive relationship between nitrogen flux and primary production in marine aquatic environments (Rabalais, 2002). If the load of nitrogen exceeds the capacity for assimilation of nutrient-enhanced production, this can result in algal blooms, increased turbidity, loss of submerged aquatic vegetation, oxygen deficiency, and overall effects on habitat and biodiversity (Rabalais, 2002; York et al., 2007). Since almost all of the nitrogen found in estuaries comes from terrestrial run- off and groundwater discharge, studying the watersheds that feed these rivers and groundwater is very important. A study conducted by Boyer et al. (2002) showed that total nitrogen in a river is not reflective of the size of the watershed, but instead of land use within the watershed. Overall, developed/urbanized and agricultural dominated watersheds show an increase in riverine nitrogen concentration relative to forested ones (Boyer et al. 2002). The nitrogen released by urban and agricultural watersheds can also be reflected in the stable nitrogen isotope values when much of the nitrogen is derived from organic fertilizers and wastewater run-off (McClelland and Valiela, 1998), thereby suggesting that the nitrogen isotopes can be used as a tracer for anthropogenic sources. 1.2.1 Nitrogen Isotopes

There are two stable isotopes of nitrogen:14 N and 15N. In nature, 99.63% of nitrogen is 14N, while 0.37% of nitrogen is 15N. The ratio of14 N/15N in air is 272, which is used as a standard in delta notation (Sharp, 2007). Delta notation is used to determine nitrogen isotope ratios of samples compared to the standard in the equation:

7 15 15 14 15 14 δ N (‰ vs. AIR) = ((( N/ N)sample/( N/ N)AIR) - 1)*1000 (Sharp, 2007) Positive values indicate that that the sample contains more 15N relative to air, while negative values indicate that the sample contains more 14N relative to air (Fry, 2002). When a sample is referred to as isotopically enriched or heavy, it is enriched in the heavy isotope relative to the standard (δ15N of air). 1.2.2 Fractionation Variations in nitrogen isotope ratios (δ15N) are results of fractionation in the nitrogen cycle. Fractionation is the division of isotopes between two substances or two phases of the same substance with difference isotope ratios (Hoefs, 2004). There are two main categories of fractionation: equilibrium fractionation and kinetic fractionation. Equilibrium fractionation is the partial separation of isotopes between substances in chemical equilibrium. Kinetic fractionation is the separation of stable isotopes from each other based on mass. The amount of fractionation varies for each step in the nitrogen cycle (Figure 1.1) (Sharp, 2007). In fixation, nitrogen fractionation is relatively small. This fractionation is still unclear, but it has been suggested to be around 0.7‰ (Sharp, 2007). Ammonification and protonation also show little to no fractionation (Sharp, 2007). In the two step process of nitrification some fractionation occurs. In the first step, when ammonium is converted to nitrite, fractionation ranges from -29 to -18 ‰ (Sharp, 2007). This depends on whether or not all ammonium is converted into nitrate; if all the ammonium is converted into nitrate, then there is no fractionation (Sharp, 2007). In the second step, all nitrite is converted into nitrate due to the instability of nitrite in aerobic settings. This causes there to be no fractionation in this step (Sharp, 2007). Fractionation during assimilation in primary producers ranges from -27 to 0‰ and is dependent on a few variables: availability of nitrogen, enzymes responsible for nitrogen fixation, and diffusion of nitrogen through the cell walls (Sharp, 2007). Plants

and phytoplankton prefer to assimilate 14N, opposed to 15N (Fry, 2002). If a system

8 contains excess amounts of nitrogen, then there will be high rates of fractionation because

organisms will preferentially assimilate all 14N. If nitrogen concentrations are lower, then fractionation rates will be lower because primary producers cannot be as selective (Fry, 2002). This means that discrimination between isotopes is only possible when the concentration of ammonium and/or nitrate is high (Sharp, 2007). Different species of aquatic plants and microorganisms (phytoplankton and bacteria) contain different enzymes that control reaction rates, and isotopic fractionation (Fry, 2007). Fractionation occurring during denitrification ranges from -40 to -5‰ (Sharp, 2007). This large fractionation is due to the Rayleigh fractionation that occurs as nitrate

is converted to N2 , which diffuses out of the system and into the atmosphere. In this case, Rayleigh fractionation leads to progressive partitioning of light isotopes into the atmosphere, while the heavier isotopes stay in the given system (i.e. lake). In marine systems the δ15N values of the surface water nitrate and ammonium are usually positive due to fractionation during both denitrification and assimilation, which both take more 14N than 15N out of the water. This leaves the nitrogen in the water enriched in 15N (York et al. 2007; Schlesinger, 1997). Ammonium tends to be enriched in 15N relative to nitrate due to fractionation during nitrification (York et al. 2007; Sharp, 2007). 1.2.3 Anthropogenic Influence There are three main anthropogenic sources of nitrate and ammonium to surface and ground waters: fertilizer nitrogen, natural soil-derived nitrogen, and animal/human sewage (wastewater) nitrogen, which have varying δ15N values (Figure 1.3) (Macko and Ostrom, 1994). In addition, nitrogen released by the burning of fossil fuels can influence nitrogen concentrations in large watersheds any water bodies (Canfield et al., 2010). Natural soils have δ15N values ranging from 3-12‰ (Tiessen et al., 1984). Although this nitrogen is produced naturally, human cultivation can increase the rate of organic matter mineralization, which can greatly increase the amount of nitrogen moving

9 frequency (%)

Figure 1.3: A frequency distribution of δ15N between different wastes, soils, and fertilizers. Different patterns in column represent the various sources seen on the right side of the diagram (Kendall, 1998).

Figure 1.4: A graph illustrating the increase in δ15N moving up trophic levels. The x-axis is 15δ N, while the y-axis respresents changes in trophic levels. Based on empiracal evidence (Mack and Ostrom, 1994).

10 through the soils into the groundwater (Winteringham, 1984; Macko and Ostrom, 1994). Inorganic fertilizers are produced using the Haber process, which artificially fixes

atmospheric nitrogen into ammonium, which has δ15N values of about 0‰ (Macko and Ostram, 1994). Organic fertilizers produced from animal waste have δ15N values ranging from 10-22‰ (Gormly and Spaulding, 1979). A study by Hallberg (1989) showed that a quarter of all fertilizer applied may be lost to ground waters, showing that application of fertilizers has a large influence on nitrogen dissolved in surface and ground waters (McClellend and Valiela, 1997; Macko and Ostrom, 1994). Human wastewater also has high δ15N values due to the effect that rising trophic levels has on nitrogen isotopes (Figure 1.4) (Macko and Ostrom, 1994). Sewage that is directly deposited into a water body, either from septic tanks or sewage overflows, leaves

water enriched in δ15N. Treated sewage, however, limits the amount of DIN and other potentially harmful compounds being discharged from their outfalls, but the DIN released may still be enriched in δ15N (DeBruyn and Rasmussen, 2002). Nitrogen derived from fertilizers, soils, and wastewater that flows into estuaries can potentially be differentiated using stable nitrogen isotopes. Due to very similar 15δ N values in DIN derived from inorganic fertilizers and natural soil, isotopes cannot be used to distinguish inorganic fertilizers from non-anthropogenic sources. However, this is not the case for wastewater and organic fertilizers (McClellend and Valiela, 1997; Canfield et al., 2010). Ammonium and nitrate derived from both wastewater and organic fertilizers have enriched δ15N values that can be distinguished from natural soil, thus these forms of anthropogenic nitrogen can be identified using stable nitrogen isotopes. In addition, different δ15N of ammonium and nitrate in water can be reflected in the 15δ N of organic nitrogen in primary producers living in the estuary (McClellend and Valiela, 1998).

1.3 Previous Studies McClellend and Valiela (1998) studied anthropogenic nitrogen inputs in the Waquoit Bay estuary in Massachusetts. They examined three different parts of the estuary that

11 receive water from watersheds with low versus high degrees of urbanization. In this study, the Childs River has the highest population density, the Quashnet River has a medium population density, and Sage Lot Pond has the lowest population density. It was found that Sage Lot Pond concentrations of nitrate varied from ~0.8 to ~50

uM, while Childs River samples ranged from ~0.8 to ~1100 uM. δ15N values ranged from ~-1 to 1‰ for Sage Lot Pond and ~-2 to 20 ‰ for Childs River. Overall, the average concentration of nitrogen and δ15N of nitrogen was higher in Child’s River then Sage Lot Pond (Figure 1.5), showing that the δ15N and concentration of nitrate of groundwater entering the estuaries are positively correlated with population density. McClellend and Valiela (1998) explain that the source of the increased nitrogen load in the Child’s River is

septic waste. Although this figure shows that 15δ N values of nitrate is positively correlated with concentrates of nitrate, McClellend and Valiela (1998) emphasize that the amount of nitrate in groundwater does not increase δ15N values, but instead the proportion of water derived from wastewater increases the δ15N values (McClellend and Valiela, 1998). The δ15N of different primary producers in the estuaries were also examined. All of the primary producers studied showed increased δ15N values that correlate with increase δ15N values of DIN in the groundwater (Figure 1.6). The most dramatic effect was seen in Zostera marina. This finding shows that if eelgrass grows in estuaries it seems to be quite sensitive to the presence of enriched DIN in estuaries, and therefore may serve as a proxy for presence of some anthropogenic nitrogen (i.e. septic and human wastewater run-off). 1.3.1 Previous Studies in Casco Bay A survey of nitrate and ammonium concentrations of surface waters in the Casco Bay was done during May of 2006, 2007, and 2008 (Figure 1. 7) (Libby and Anderson, 2010). For all three years, concentrations of nitrate in the Presumpscot River and Fore River reached 5 uM, affecting the estuarine waters that these river mouths lead into, including the water surrounding Mackworth Island. Similar trends were seen for ammonium concentrations, where concentrations were as high as 9 umol/L around Mackworth

12 Figure 1.5: A graph illustrating δ15N versus concentration of DIN in groundwater from watersheds of varying population densities. Sage Lot Pond is a smaller watershed with a lower population density than Childs River, both located in the Waquoit Bay Estuary, MA. The closed circles represent Child’s River and the open circles represent Sage Lot Pond (McClellend and Valiela, 1998).

Figure 1.6: δ15N (per mil) of DIN in groundwater versus the δ15N of various primary pro- ducers in Waquoit Bay, MA. Averages for all three locations of each species are graphed. Sage Lot Pond samples are plotted on the far left and Childs River Samples are graphed on the far right; eelgrass is highlighted (McClellend and Valiela, 1998).

13 May 2008 May May 2007 May May 2006 May Figure 1.7: Nitrate/nitrite (uM) and ammonium concentrations (uM) in the Casco Bay in May of 2006,2007, of (uM) in the Casco in May Bay concentrations ammonium (uM) and 1.7: Nitrate/nitrite Figure 2010). Anderson, and 2008 (Libby and

14 Island. In Maquoit Bay nitrate and ammonium concentrations were low (~1 umol/L and ~4 umol/L, respectively) and similar to open ocean concentrations for all three years. In 2007, high concentrations of nitrate were found in north east Casco Bay. This most likely was coming from the Kennebec River, but could also be the result of potentially higher concentrations of nitrate in the outer bay/open ocean. William Loopesko’s (2010) thesis studied nitrogen enrichment of eelgrass, POM, and sediments at Mackworth Island and Macquoit Bay in the fall of 2009. He was unable to run isotope analysis on water due to the challenges associated with analyzing nitrate and ammonium nitrogen isotope values. Loopesko (2010) found that δ15N of POM was enriched at Maquoit Bay by 1.5‰ relative to Mackworth Island, the δ15N of sediment (benthic algae) was very similar for both locations, and the δ15N values of eelgrass were 1.5‰ enriched at Mackworth Island relative to Maquoit Bay.

1.4 Purpose Work has now been done that continues to examine nitrogen isotope fractionation in Maquoit Bay and Mackworth Island and determine if nitrogen isotopes can be used to track sources of anthropogenic nitrogen in Casco Bay. At Maquoit Bay and Mackworth Island, isotope data was analyzed for both DIN in the water column and pore water, in addition to eelgrass, sediment, and particulate organic matter (POM). Isotope and concentration data was also analyzed from river and open ocean sources to the two sites. The nitrogen isotope ratios and concentrations of samples from Mackworth Island and Maquoit Bay were studied to compare a densely populated area with a more pristine area. If nitrogen isotope data between the two sites is representative of the anthropogenic activity surrounding them, then nitrogen isotopes can be used as an early indicator of human-caused eutrophication in the Casco Bay. 1.4.1 Site Description Casco Bay is a 40 km long and 21 km wide bay located in southern Maine. It is enclosed by Cape Elizabeth to the south and Small Point to the north.

15 Mackworth Island is a small island located just north east of Portland close to the mouth of the Presumpscot and Fore Rivers, which both have heavily populated watersheds. Mackworth Island has relatively high levels of anthropogenic activity, which is expected to be reflected in nitrogen isotope data from the area. Maquoit Bay is located twenty miles northeast of Mackworth Island, and is a small inlet bound by Mereneck Point on the east and the mainland on the west side. This eelgrass bed is nearly twice the size of the bed at Mackworth Island and there has been little development on its shores and in the surrounding watershed. The pristine setting of Maquoit Bay should be reflected in the nitrogen isotope data taken from the area, and be differentiated from the data taken from Mackworth Island. 1.4.1.1 Geology The bedrock that composes the coast of Casco Bay is almost entirely Silurian- Ordovician metamorphic gneiss and schist, which is referred to as the Casco Bay Formation (Maine Geologic Survey, 2008) formed from volcanic and marine sedimentary rocks (Marvinney, 2002) (Figure 1.8). The upper section of both the Presumpscot and Fore River watersheds are primarely Ordovician-Devonian mixed volcanic rocks (Maine Geologic Survey, 2008), while the lower part of the watershed is made up of a variety of Silurian-Ordovician metamorphic schist and gneiss that are part of the Casco Bay Group (Maine Geologic Survey, 2008). Both Maquoit Bay and Mackworth Island’s bedrock is composed of the Casco Bay formation schist and gneiss (Maine Geologic Survey, 2008). The surface geology of Mackworth Island (Figure 1.9) is composed of till, overlying fine grained glaciomarine sediments (Presumpscot Formation) (Thomson and Born, 1985). The surficial geology of the upper watersheds of the Fore River and Presumpscot River are generally till, but the lower watersheds are mostly the Presumpscot Formation (Thomson and Born, 1985). The surficial geology of Maquoit Bay (Figure 1.10) is similar to Mackworth Island, in that the edge of the estuary is the Presumpscot Formation, while the Bunganuc River watershed is a mix of the Presumpscot Formation

16 N 1:500,000

Figure 1.8: A simplified bedrock map of Casco Bay (Maine Geologic Survey, 2008) with Maquoit Bay highlighted in blue and Mackworth Island highlighted in red.

17 Mackworth Island Portland 1:50,000 N 1:1,000,000 Figure 1.9: A surficial geology map of Mackworth Island and its surrounding area (Maine Geologic Survey, Survey, Geologic (Maine area surrounding its and Island Mackworth of 1.9: A surficialgeology Figure map 19. can be on page seen map the for legend 2008). The

18 The legend for the surficial the for The geologylegend on and 20. pages 18 maps

19 Maquoit Bay Maquoit 1:40,000 N 1:1,000,000 A surficial geology map of Maquoit Maquoit of 1.10 : A surficialgeology Figure map Geologic (Maine area surrounding its and Bay 19. can be on page seen legend 2008). The Survey,

20 and till (Maine Geologic Survey, 2008; Thomson and Born, 1985).

1.4.1.2 Land Use The watershed that feeds Casco Bay drains 1,585 square kilometers of mountain and coastal forests. About 17 percent of Maine’s population live within this watershed, even though it is only three percent of the total land mass (Libby and Anderson, 2010). There are many sub-watersheds in the Casco Bay watershed that drain into different areas of the estuary, affecting the water quality. The Maquoit Bay watershed (Figure 1.11) encompasses 30 square kilometers, containing three major streams, the largest being the Bunganuc River (Loopesko, 2010). Sixty seven percent of the watershed is forested, pastures make up another fifteen percent, and six percent is developed. Developed land is defined as land with buildings, impervious surfaces, or other artificial surfaces. Only 0.05 percent of this watershed is considered to be highly developed. The total population of the watershed is about 5,900 people (513 people per square mile), which includes a small portion of Brunswick. No wastewater plants (Brunswick’s wastewater treatment plant is on the Androscoggin River), storm drainage pipes, or sewage outfalls are found within the watershed (Loopesko, 2010). The Mackworth Island watershed (Figure 1.11) encompasses 1,837 square kilometers, which includes the Fore River and the Presumscot River watersheds, which is connected to the Crooked River watershed through Sebago Lake. There are forty towns in the watershed, including Portland metropolitan area. In 2000, the population of the watershed was 189,000 people (267.72 people per square mile), 86% of this being in the lower section of the watershed, closer to Mackworth Island. About fifty six percent of the watershed is forested, but the urban areas are densely developed relative to those in the Maquoit Bay watershed. It contains twenty eight wastewater treatment plants, fifteen of which are classified as major (Loopesko, 2010).

21 Overall, the Mackworth Island watershed contains much more urban areas and wastewater treatment plants than that of Maquoit Bay. For this reason, it is hypothesized that nitrogen isotopes in primary producers and DIN at Mackworth Island should be enriched relative to Maquoit Bay.

22 Figure 1.11: Percent land use in the watersheds of Mackworth Island and Maquoit Bay (Loopesko, 2010).

23 2. Methods

24 2.1 Field Three sites were selected within eelgrass beds at both Maquoit Bay and Mackworth Island (Figure 2.1), overlapping with those of Loopesko (2010). Fresh water inputs from the lower Presumpscot River and Fore River (for Mackworth Island), and the Bunganuc River (for Maquoit Bay) were sampled to evaluate terrestrial input of nitrogen. Open ocean inputs were assessed from two locations sampled by the OSV Bold along a transect in the outer Casco Bay. 2.1.1 Mackworth Island and Maquoit Bay An identical sampling scheme was employed for all the sites at Maquoit Bay and Mackworth Island. Surface water, eelgrass, sediment, pore water, and particulate matter were collected at all sites (Table 2.1). In addition to sample collection, temperature, salinity, dissolved oxygen, and pH were determined using a Hydrolab. Nine liters of surface water were collected from all sites, each in individual 1 liter Nalgene bottles, three of which were light-tight (Figure 2.2). Three liters were filtered for one POM sample, and the water was used to determine dissolved nitrate δ15N values and concentrations. Another three liters were collected for dissolved ammonium δ15N values and concentrations. These waters were acidified in the field using sulfuric acid to stabilize ammonium in the water samples (Holmes et al., 1998). The last three liters (light-tight containers) were filtered separately to obtain three filters for chlorophyll analysis. Three individual eelgrass plants were collected at each site. These were obtained by free diving and taking each sample individually, making sure to capture the whole plant, including the root and rhizomes. Eelgrass samples from each site were bagged together. One sediment sample was taken at each site. Sediment was collected by free diving in each eelgrass bed, and collecting the upper 5 cm (~100 grams) of the seafloor, making sure not to include large amounts of eelgrass leaves and shell debris. Sediment was bagged in Ziploc bags. At each site, one 13 cm wide by 40 cm deep core was taken to extract and analyze

25 0 2.5 5 10 15 20 Kilometers 1:400,000 Bunganuc River Site 2 Site 3 Site 1 Casco ±Bay Maquoit

Open Ocean Samples Bay

Bates R1-15 Presumpscot River

Mackworth Legend Eelgrass Cover (2005) Island Waypoints 0 0.375 0.75 1.5 2.25 3 Kilometers

Site 3 1:50,000 Site 2 Site 1 Figure 2.1: GIS Maps indicating the locations of Mackworth Island and Maquoit Bay. Sites at both locations, each river where samples were taken, and the open ocean sample sites are labelled. The large green areas represent eelgrass cover in 2005. Fore River

Fore River

26 3 (1 liter each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 N/A N/A N/A N/A Chlorophyll 1 Core (~65 cm) cubic (~65 1 Core cm) cubic (~65 1 Core cm) cubic (~65 1 Core cm) cubic (~65 1 Core cm) cubic (~65 1 Core cm) cubic (~65 1 Core N/A N/A N/A N/A N/A N/A N/A Grain Size Analysis Cores Analysis Size Grain 50 cmcubic 50 cmcubic 50 cmcubic 50 cmcubic 50 cmcubic 50 cmcubic N/A N/A N/A N/A N/A N/A N/A Sediment 3 plants 3 plants 3 plants 3 plants 3 plants 3 plants N/A N/A N/A N/A N/A N/A N/A Eelgrass 1 Core (~500 cubic cm) (~500 1 Core cubic cm) (~500 1 Core cubic cm) (~500 1 Core cubic cm) (~500 1 Core cubic cm) (~500 1 Core cubic cm) (~500 1 Core N/A N/A N/A N/A N/A N/A N/A Pore Water (Ammonium) Water Pore 3 (1 liter each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 2 (1 each) liter 2 (1 each) liter 2 (1 each) liter 2 (1 Surface Water (Ammonium) Water Surface 3 (1 liter each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 3 (1 each) liter 2 (1 each) liter 2 (1 each) liter 2 (1 each) liter 2 (1 POM collect to was filtered water this of 1-2 liters * Surface Water (Nitrate)* Water Surface 9' 69.1") ° (43° 39' 49.0" ,70° 9' 47.1") (43° 39' 49.0" ,70° 9' 47.1") (43° 34' 31.6",70 (43° 34' 31.6",70° 9' 69.1") (43° 51' 52.3584", 70° 1' 14.2") (43° 43' 1.9956", 70° 15' 52.7") (43° 38' 56.5296", 70° 17' 32.1") (43° 51' 27.5076", -70° 0' 35.5098") (43° 51' 43.3908", -70° 0' 10.5336") (43° 51' 32.5512", -69° 59' 55.1178") (43° 40' 55.3362", 70° 13' 57.6") (43° 40' 57.4284", 70° 13' 55.2") (43° 41' 2.5074", 70° 13' 56.2") Samples Collected Samples Site 2 Site 3 Site 29th 2010) July 13th and (July MackworthIsland 1 Site 2 Site 3 Site 20th, 2010) (July Rivers Pres. River River Fore Bung. River 2nd, 2010) (July Ocean R1-R15 Bottom Surface R1-R15 Bates Bottom Bates Surface Maquoit Bay (July 6th and July 21st,July and 6th 2010) Bay (July Maquoit 1 Site Table 2.1: A spreadsheet showing what samples were collected at collected at were samples what showing 2.1: A spreadsheet Table collected. dates and locations geographical including each site,

27 Collected in Field Lab Preparation

Filtered using a Whatman 0.45 um lter to collect water. Filters were disposed of after ltration, and the water was kept for three Acidi ed Acidi ed Acidi ed separate ammonium isotope analyses.

Ammonium Ammonium Ammonium

1 liter was ltered using a Whatman 0.45 um lter for chlorophyll analysis a total of 3 times for each site. The water was disposed of. Light-Tight Light-Tight Light-Tight Container Container Container

All water was ltered using a Whatman 0.45um lter and the water was kept for nitrate isotope analysis.

Nitrate Nitrate Nitrate

One lter from each site was kept for POM nitrogen isotope analysis.

Figure 2.2: A flow chart indicating what the water samples from each site were used for. Each square represents 1L of water.

28 pore water. In order to ensure that the pore water represent an area where eelgrass grows, the core was taken in an area of high eelgrass density. Upon retrieval the cores were sealed with duct tape. Another 5 cm wide by 13 cm deep core was taken for grain size analysis of the sediment in which the eelgrass was growing. In order to accurately reflect grain size in the eelgrass beds, cores were taken in areas of dense eelgrass cover. Once these cores were collected they were sealed with duct tape and transported upright to prevent any mixing of sediment layers. All samples were labeled and placed in a cooler immediately upon collection. At Maquoit Bay, all water samples, cores, and sediment samples were collected on 6 July 2010, and the eelgrass samples were collected on 21 July 2010. For Mackworth Island, all water and sediment samples were taken on 13 July 2010, and the cores and eelgrass samples were collected on 29 July 2010. 2.1.2 Rivers At each river, nine 1L water samples were collected on 20 July 2010, and treated identically to those at each site at Mackworth Island and Maquoit Bay (Figure 2.2). Hydrolab data were recorded at each river, as well. 2.1.3 Open Ocean Open ocean water samples were collected on 2 July 2010 by the EPA OSV Bold. Four 1L surface water and four 1L deep water samples were taken from two sites in Casco Bay, titled “R1-R15” and “Bates” (Figure 2.1). These sites were chosen to account for inorganic dissolved nitrogen flowing into the Casco Bay from the open ocean. The depth of the deep water samples were 30.5 meters at the Bates site and 26.5 meters at the R1-R15 site. All of these water samples were acidified and frozen upon collection.

29 2.2 Laboratory 2.2.1 Grain Size Analysis For each core (5cm x 13cm), the top 8 centimeters were used for grain size analysis. These sediment samples were placed in a clean plastic cup and rinsed with tap water and de-ionized water to solubilize the salts. After decanting the water, the sediments were put in a drying oven overnight. Once dry, about 30 grams of each sample was weighed, then soaked in 40% hydrogen peroxide for 12 hours. The sediment was

then placed in a 40Ø sieve to separate the sand from the silt and clay. The clay and silt was drained out of the sieve using de-ionized water. The sand from each sample was then collected, dried, and weighed. Percent sand and percent silt/clay data was determined from the weights. 2.2.2 Chlorophyll Chlorophyll samples were prepared by filtering 1 liter of surface water through a combusted Whatman 0.45 um glass filter. Prior to use, all Whatman filters were combusted at 400OC for 4 hours to volatilize any organic matter that could cause contamination. The filters were then placed in a 50 mL tube and 10 milliliters of 90% acetone were added, submerging the whole filter. Samples were then capped, placed in a freezer for about 12 hours, removed, and centrifuged for approximately 5 minutes at 2.5 rpm to separate the supernatant from the filter. One mL of supernatant was removed, placed in a cuvette with 9 mL of 90% acetone, and mixed. Samples were placed in a Turner Designs 10-AU-005-CE fluorometer to determine a Fo value, which is the measured fluorescence when a flash a light is applied to the sample. Once the first value is recorded, 3 drops of 6 M HCl were added to the cuvette, mixed, and placed in the fluorometer again to get a Fa value, which is a measure of the fluorescence in the absence of photosynthetic light. The Fo and Fa values were then both used to calculate the Chlorophyll A levels in a Microsoft Excel worksheet.

30 2.2.3 Nitrogen Isotope Analysis using the IRMS Eelgrass, sediment, POM, and water samples were all analyzed for their nitrogen isotope composition. All of these sample types require different preparation before being analyzed in the Costech Instruments elemental-analyzed-combustion system attached to a ThermoFinnigan Delta V Advantage stable isotope ratio mass spectrometer (EA-C- IRMS). In order to obtain accurate isotope data, 0.8 umols of nitrogen and carbon have to be recovered from samples during analysis by the EA-C-IRMS. 2.2.3.1 Eelgrass On the same day as collection, each eelgrass sample was rinsed in de-ionized water, the youngest leaves were separated from the rest of the plant, then the separated leaves were scraped with a glass slide to remove epiphytes and other surface material. All of the separated and cleaned leaves were frozen and freeze dried using a Lacono Freezezone 6. Once freeze dried, all three leaves per site were homogenized using a Spex Industries mixer/shaker. One to two milligrams of powdered eelgrass from each site were sub-sampled for analysis in the EA-C-IRMS. 2.2.3.2 Sediment Sediment samples were immediately placed in a freezer when returned to the lab and eventually freeze dried. Eight to ten milligrams of sediment from each sample were weighed out for the EA-C-IRMS. When selecting sediment it was essential to avoid organic matter and carbonate in each sample. In order to make sure carbonate was not in the sediments, a droplet of hydrochloric acid was added to the sediments. If the hydrochloric acid bubbled, that indicated that the sediments contain carbonate. None of the sediment samples bubbled appreciably. 2.2.3.3 Particulate Organic Matter To collect POM, one to two liters of water collected at the three rivers and six eelgrass sites were filtered through Watman 0.45 um pre-combusted glass filters upon arrival to the lab. In order to accumulate an adequate amount of POM, water was filtered

31 until it could no longer move through the filter and the volume filtered was recorded. Filters were freeze dried then stored in air tight containers. One eighth of a filter was analyzed in the EA-C-IRMS. 2.2.3.4 Inorganic Nitrogen in Water Prior to isotope ratio analysis of dissolved inorganic nitrogen (ammonium and nitrate) it is necessary to know the concentration of dissolved inorganic nitrogen species (Table 2.2). Small samples of about 250 ml were collected in the field along with the other samples and sent to North East Labs (Table 2.3) in Portland, ME. Smaller concentrations

require large volumes of water because in order to obtain accurate δ15N values at least 0.8 micro moles of nitrogen have to be analyzed. Ammonium Both surface water and pore water were analyzed for ammonium isotope analysis. Upon returning to the lab, surface water samples were filtered with combusted Whatman 0.45 uM filters in order to clear the water of suspended matter. For pore water analysis, the top 20-25 cm of sediment from the 13x40cm core was centrifuged in 50 mL conical test tubes. The pore water was decanted off and pooled until about 200 mL of pore was collected. This water was filtered through Watman 0.45 uM filters.

The δ15N of ammonium dissolved in the water column and pore water was analyzed using methods described by Holmes et al. (1998). This method (Figure 2.3) reduces ammonium into ammonia in water samples, then isolates the ammonia released into the air space in the container onto a diffusion filter; Teflon encased Whatman 2.7 uM GF/D filter disk (pre-combusted to remove any organics), acidified with 25 uL of 2 M sulfuric acid (Keeney and Nelson, 1982). These filter disks within the Teflon were analyzed for δ15N values of ammonium in the IRMS once this method is complete. To create the diffusion filters Holmes et al. (1998) suggest using Teflon disks, but this proved unsuccessful in multiple trials. These disks would break open, soaking the filter, thus ruining the trial. The filter disks cannot get wet because they can only absorb

32 Sample NH4+ or NO3- Concentration (uM) Volume Needed (mL) 20 100 10 200 5 400 2.5 800 Table 2.2: Volumes required for analysis of ammonium (Holmes et al., 1998) and nitrate (Sigman et al., 1997) contents in order to recover adequate mirco moles of nitrogen to obtain accurate nitrogen isotope data.

Ammonia Concentration Nitrate Concentration Site EPA 350.1(uM) EPA 9056 (uM) Fore River 25.56 N/A Presumpscot River limit of detection N/A Bunganuc River limit of detection N/A

Mackworth 1 23.33 N/A Mackworth 2 limit of detection N/A Mackworth 3 limit of detection N/A Mackworth Pore 1 77.78 N/A Mackworth Pore 2 43.89 N/A Mackworth Pore 3 66.67 N/A

Maquoit 1 29.44 N/A Maquoit 2 14.44 N/A Maquoit 3 limit of detection N/A Maquoit Pore 1 722.22 N/A Maquoit Pore 2 472.22 N/A Maquoit Pore 3 388.89 N/A limit of detection= 11.11 uM limit of detection=8.06 uM Table 2.3: Concentrations of ammonium and nitrate determined by North East Labs in Portland, Maine for all sample sites in this study. Not all locations have an indicated concentration due to the labs limits of dectection.

33 nitrogen from the diffused ammonia to get an accurate 15δ N value for ammonium. Instead of the Teflon disks, 1.5 cm wide Teflon tape was used to encase the filter (Colorado Plateau Stable Isotope Laboratory, 2010). This proved more successful to construct the Teflon encased filter disks. It was first necessary to cut Teflon tape into 1.5 cm by 3 cm rectangles, which were folded in half, creating a dual layered 1.5 cm by 1.5 cm Teflon square. Then a Whatman filter disk was place on top of one layer of Teflon tape and

acidified with 25 ul of 2M 2H SO4. Another 1.5 cm by 1.5 cm dual layered Teflon tape square was laid on top of the filter disk. The two pieces of tape were sealed together by pressing the tape edges with a smooth edged metal cylinder, 1.2 centimeter diameter. In order to keep pressure consistent, filters were placed on tin foil with layers of papers towels below. Plastic Nalgene bottles were used as containers to run the ammonium diffusion method. All surface water samples were run in 1L bottle, and the pore water samples were run in 250mL bottles. For all river samples (salinities less than 10 ppt) about 3%-5% by

weight of NaCl (pre-combusted at 450oC for four hours) was added, to prevent osmosis from breaking apart the packets during incubation. One diffusion packet, and about 300 milligrams of MgO (combusted at 400oC for four hours) per 100 milliliters were added to each sample. Adding the MgO brings the pH up to about 9.2, so that all ammonium in the sample will be converted to ammonia. Ammonia then diffuses from the water into the head space in the container, and then onto the acidified filter disk (Figure 2.3). Once the filter packet and MgO were added to all samples, they were agitated in a New Brunswick Scientific C25 incubator shaker at 40oC and 60 rpm for 14 days. After 14 days the filter packets were taken out of the water samples and placed in a desiccator with a beaker of concentrated acid for 24 hours to dry and remove any trace amount of ammonia. Once dry, the filters were taken out of the desiccator, and the Whatman filter was removed from the Teflon tape and placed in a tin cup to be analyzed in the EA-C-IRMS.

34 NH3 Filter Pack

NH3 MgO (pH=9.2)

+ NH4

NaCl (if necessary)

Shake (60 rpm) and incubate at 65oC for 14 days + - NH4 + OH NH3 + H2O Figure 2.3: An illustration of the ammonium diffusion method adapted from Holmes et al., 1998.

35 Nitrate Upon return to the lab, surface water samples were filtered with combusted Whatman 0.45 uM filters. Nitrate isotope analysis was not done with pore water because reducing environments contain almost no nitrate (Schlesinger, 1997 and Jaffe, 1992). The method for extracting nitrate from water was adapted from Sigman et al. (1997) paper, and is very similar to that of Holmes et al. (1998), but includes two extra steps (Figure 2.4). One liter samples were used for this method. Filter packets were assembled as described above. One liter Wheaton glass bottles were used for each water sample in this method. Again, about 3%-5% by weight of ashed NaCl was added to each sample with low salinities to prevent osmosis from occurring later in the procedure. In the first step (Figure 2.4), three grams of ashed MgO were added to each sample, bringing the pH to about 9.2 and convert ammonium to ammonia. Samples were then pre-incubated at 65oC for about 5 days to decompose labile dissolved organic nitrogen into ammonia. Samples were then placed into an oven at 95oC to allow for the evaporation of each sample down to 15%-20% of their original volumes (Step 2 in figure 2.4), driving off any dissolved ammonium. This step can take up to three days, but once each sample is down to 15%- 20% of their original volume the bottles can be removed from the oven. Each sample was monitored closely to prevent the full evaporation of water. The pre-incubation and evaporation steps removed all ammonium and ammonia from the water samples, leaving dissolved nitrate as the only form of nitrogen in the water samples (Figure 2.4). Once each sample was boiled down and the salts were re-dissolved (if necessary) one diffusion packet and 0.75 grams of Devarda’s alloy was added to each sample. The cap was immediately tightened and sealed. In this third and final step (Figure 2.4), Devarda’s alloy reduces the nitrate to ammonia, which is diffused into the headspace and captured on the filter disks. Each sample was placed in an incubator shaker at 60oC and 60 rpm for 14 days. After 14 days the filter packets were taken out of the water samples and placed

36 - ] 4 + 8[Al(OH) 3 C for 14 days C for Filter Pack Filter o - 3 NO O 2NH 3 2

Devarda’s Alloy NH Step 3 Step 3 + 18H - NH at 65 C for 14 days 65 C for at Shake (60 rpm) and incubate Shake (60 rpm) and incubate + 8Al + 5OH - 3 Shake (60 rpm) and incubate at 65 at incubate and (60 rpm) Shake 3NO + 4 O 2 NH + H 3 3 Step 2 Step C until volume is is volume C until NH o NH

MgO (pH=9.2) - volume. volume. - 3 3 + OH + NH NO 4 NH Evaporate at 95 at Evaporate resuced to 15%-20% of its original its 15%-20% of to resuced + 4 3 NH NH 3 STEP 1 C for 5 Days C for NH o 65

MgO (pH=9.2) 3 - 3 65 C for 5 days 65 C for Figure 2.4: An illustration of the three steps in the nitrate diffusion method adapted from Sigman et al., 1997. et al., Sigman from method adapted diffusion in the nitrate thesteps three of illustration 2.4: An Figure + NaCl (if necessary) NH NO

37 in a desiccator with a beaker of concentrated sulfuric acid for 24 hours to remove trace amounts of ammonia. Finally, the filter packets were taken out of the dessicator, and the Whatman filters were taken out of the Teflon tape and placed in a tin for immediate analysis by the EA-C-IRMS. Standards Analysis of standards are a way to test the ammonium and nitrate diffusion methods and to determine if the method causes fractionation. If it is determined that the method itself causes consistent fractionation, a correction can be applied to the samples to

determine the true δ15N values. Standards were created and run for both the Holmes et al. (1998) and Sigman et al. (1997) methods. Solid ammonium sulfate and sodium nitrate salts were used to make standards

and were both analyzed in the EA-C-IRMS directly to determine the true δ15N values. These salts were used to make standard solutions over a range of concentrations. For ammonium, a stock solution containing the dissolved ammonium sulfate of known isotopic composition was diluted into three different concentrations of 5 uM, 15uM, and 25uM. These standards were run with the samples through the ammonium diffusion method. For nitrate a stock solution containing the dissolved sodium nitrate of a known isotopic composition was diluted into three concentrations of 5uM, 20uM, and 40uM. These standards were run with the samples through the nitrate diffusion method. Once these standards were analyzed in the EA-C-IRMS, concentrations and δ15N values were compared to those of the solid ammonium sulfate and sodium nitrate. Corrections were applied if the δ15N value was affected by fractionation caused by the method. 2.2.4 Statistic Analysis In order to determine statistically significant differences in data t-test and ANOVA tests were run using Minitab software. Statistically significant differences in data are indicated when the p-value is <0.05. All data collected from Mackworth Island and Maquoit Bay were tested for significant differences using a 2-sample t-test. All data

38 collected from the three rivers were compared using an ANOVA test. Average data from all the rivers were compared to open ocean samples using a 2-sample t-test. T-tests are used when two sites are being compared and ANOVA tests are used when more than two sites were being compared.

39 3. Results

40 All samples run in the EA-C-IRMS were analyzed with internal standards (acetanilide, caffeine, and cod meat) of known isotopic composition. All showed adequate recoveries and isotope ratios within 0.2‰ of the true values, indicating that the machine was running properly throughout the course of this experiment. All isotope data for Mackworth Island, Maquoit Bay, the Bunganuc, Fore, and Presumpscot Rivers, and the open ocean samples can be seen in Appendix 1-4, respectively. Chlorophyll data for all sample sites can be seen in Appendix 5. All grain size analysis data for Maquoit Bay and Mackworth Island can be seen in Appendix 6.

3.1 Standards The amount of ammonium recovered from the ammonium diffusion method (Holmes et al., 1998) was nearly equal to the amount in the standard solutions (R2=1), suggesting that all ammonium in the standard solutions was converted to ammonia and captured on the filter disk (Table 3.1; Figure 3.1). The amount of fractionation that occurred during the ammonium diffusion method was greatest at lower ammonium concentrations (~5±2‰ at 5uM) (Figure 3.1). For concentrations between 5uM and 15 uM, a linear relationship between the ammonium captured on the filter disk and the amount in the standard solution where y= 0.1514x – 5.6137 was used to correct for fractionation caused by the method. For concentrations between 15 uM and 25 uM the equation y=0.0681x – 4.3472 can be used to correct for method caused fractionation in samples (y=the amount of fractionation, x=concentration). The amount of nitrate recovered in the nitrate diffusion method (Sigman et al., 1997) was nearly equal to the amount in the standard solutions (R2=0.9996), suggesting that all nitrate in the standard solutions were converted to ammonia and captured on the filter disks (Figure 3.2). The amount of fractionation that occurred during the nitrate diffusion method was greatest at lower concentrations; ~2.5±0.5‰ for the 3 uM sample. At concentrations of 20 uM to 40 uM minimal fractionation seemed to occur from the

41 A)

Sample ID Sample Type N conc. (uM) Average st. dev. d15N (3pts) 3uM 1L std NO3 3.63 5.5 3uM 1L std. NO3 5.27 4.45 1.16 4.6 20 uM 500 ml std NO3 22.19 8.0 20 uM 500 Ml std NO3 22.74 22.47 0.39 7.7 40 uM 250 ml std. NO3 47.17 7.4 40 uM 250 ml std. NO3 43.45 45.31 2.63 7.8

5.0 uM (NH4)2SO4 NH4 5.18 -6.9 5.0 uM (NH4)2SO4 NH4 5.41 5.30 0.16 -2.7 15.0 uM (NH4)2SO4 NH4 15.48 -1.6 15.0 uM (NH4)2SO4 NH4 15.17 -4.2 15.0 uM (NH4)2SO4 NH4 14.94 15.20 0.27 -4.2 25.0 uM (NH4)2SO4 NH4 24.61 -4.8 25.0 uM (NH4)2SO4 NH4 25.74 -1.4 25.0 uM (NH4)2SO4 NH4 25.66 25.34 0.63 -1.6

B)

Sample ID SampleType Mass (mg) % N N (umol) d15N 7730 NaNO3 0.37 14.92 4.26 7.2 7731 NaNO3 0.51 14.94 5.82 7.2 7732 NaNO3 1.27 13.76 13.35 7.4 7733 Ammonium Sulfate 0.62 19.73 9.35 -0.5 7734 Ammonium Sulfate 0.95 19.94 14.47 -0.3 7735 Ammonium Sulfate 0.75 19.60 11.23 -0.3 Table 3.1: Ammonium and nitrate diffusion method data for standards. Table A shows results from the standard solutions made to test the ammonium and nitrate diffusion methods, and table B is the results from analyzing solid sodium nitrate and ammonium sulfate used to make the standard solutions.

42 A)

30.0

25.0 y = 1.0021x + 0.2461 R² = 1 20.0

15.0

10.0

Captured on Captured Filter Disk 5.0 Ammonium Concentration (uM) (uM) Ammonium Concentration 0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Concentration of Ammonium Standards (uM)

B) 0.0

15 δ N for Solid (NH4)2SO4 -1.0

-2.0 y = 0.0681x - 4.3472 R² = 1 -3.0 n=3 y = 0.1514x - 5.6137 R² = 1 n=3 -4.0

-5.0 n=2 -6.0

-7.0 N of Ammonium Standards (‰) N of Ammonium Standards 15 δ -8.0

-9.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Concentration of Ammonium Standards(uM)

Figure 3.1: The results of running standards for the ammonium diffusion method (Holmes et al., 1998). Graph A shows ammonium concentrations captured on filter disks versus actual ammonium concentration of standards. An R2 value of 1.0021 indicates adequate recoveries for all concentrations. Graph B shows mean δ15N of ammonium versus con- centration of ammonium data for standards of known isotopic composition. The known isotopic composition is indicated by the red line, which is the δ15N of solid ammonium sulfate, the salt used to make the standard solutions. The bars indicate standard deviation. The equations listed are used to correct for fractionation that occurs in the method in samples with concentrations between 5uM and 15 uM and between 15 uM and 25uM.

43 A)

50.0 y = 1.1054x + 0.8602 45.0 R² = 0.9996 40.0

35.0

30.0

25.0

20.0

15.0

10.0 Captured on Filter Disk on Filter Captured

Nitrate Concentration (uM) (uM) Concentration Nitrate 5.0

0.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 Concentration of Nitrate Standards (uM)

B)

9.0

8.0 n=2 n=2 15 δ N for solid NaNO3 7.0 y = 0.1576x + 4.3415 R² = 1 6.0

5.0

4.0 n=2

3.0

N of Nitrate Standards (‰) Standards of Nitrate N 2.0 15 δ 1.0

0.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 Concentration of Nitrate Standards (uM)

Figure 3.2: The results from running standards in the nitrate diffusion method (Sigman et al., 1997). Graph A shows nitrate concentratrations captured on the filter disks versus ni- trate concentration of the standards. An R2 value of 0.9996 indicates adequate recoveries of nitrogen at all concentrations. Graph B shows mean δ15N of nitrate versus concentra- tion of nitrate for standards of known isotopic composition. The known isotopic compo- sition is indicated by the red line, which is the δ15N of solid sodium nitrate, the salt used to make the standard solutions. The bars indicate standard deviation. The equation listed is used to correct for fractionation that occurs in the method for samples with concentra- tions below 20uM.

44 method. For concentrations below 20 uM the equation y= 0.1576x + 4.3415 can used to correct for method-caused fractionation in samples.

3.2 Mackworth Island versus Maquoit Bay 3.2.1 Nitrate There were no statistically significant differences in concentration of nitrate between Mackworth Island and Maquoit Bay (Table 3.2; Figure 3.3), however nitrate concentration in Maquoit Bay is more variable than at Mackworth Island (Table 3.2). There is a statistically significant difference in15 δ N of nitrate between Mackworth Island and Maquoit Bay (p=0.004) (Figure 3.3). The average offset between the15 δ N at the two sites is 3.1‰, with more enriched values at Mackworth Island. 3.2.2 Ammonium Ammonium concentrations were determined using two different methods: from water samples sent directly to North East Labs for analysis using EPA method 9056, and from the amount of nitrogen recovered in the EA-C-IRMS from the filters that are products of the ammonium diffusion method of Holmes et al. (1998). These data are not in agreement. When data are available from both types of analysis, the amount of ammonium recovered via the ammonium diffusion method (Holmes et al., 1998) is always lower than the ammonium measured at North East labs. Because the ammonium values are greater than the limits of detection at North East Labs (Table 3.3), this suggests that the ammonium diffusion method is not capturing all of the ammonium in the samples. Though this appears to contradict the standard experiments, standards were run in a different manner and some of the concentrations indicated by North East Labs are much higher than any of the standard’s concentrations. Ammonium isotope data derived from the ammonium diffusion method were extremely depleted in 15N, ranging from about -2‰ to -37‰, way more depleted than published data of York et al. (2007) and McClellend and Valiela (1998). Because the amount of nitrogen recovered was much smaller than expected, the depleted isotope

45 - 0.4 1.0 N/A N/A 0.08 Standard Deviation 1.1 4.3 -2.3 N/A N/A Fore Average 1.9 1.6 6.3 1.1 4.4 0.1 0.1 2.5 2.5 1.1 0.6 3.8 0.6 5.7 0.1 N/A N/A N/A N/A N/A N/A 0.63 0.12 1.00 10.54 Standard Deviation Standard Deviation Standard Deviation 3.9 0.4 9.2 8.6 3.9 7.1 1.5 8.1 -1.6 N/A N/A N/A N/A N/A N/A 2.28 18.0 0.77 1.98 25.4 -12.9 -16.1 -10.1 -10.5 73.66 Mackworth Island Presumpscot Ocean Average Average Average 9.3 0.5 0.8 1.3 3.0 0.1 1.2 0.4 3.5 5.5 3.0 2.9 1.8 3.0 0.9 N/A N/A N/A N/A N/A N/A 0.22 1.72 0.95 1.10 Standard Deviation Standard Deviation Standard Deviation 2.8 5.3 5.3 9.3 6.9 5.7 -2.7 -5.4 -3.1 N/A N/A N/A N/A N/A N/A 3.93 10.4 11.5 2.85 1.60 19.0 -10.0 -19.7 -12.4 99.76 Maquoit Bay Maquoit Average Bunganuc Average Rivers Average yes insufficient data insufficient insufficient data insufficient no no yes no no no yes yes insufficient data insufficient yes insufficient data insufficient no data Insufficient insufficient data insufficient no no no yes no no yes Statistically Significant? yes Statistically Significant? Statistically Significant? N/A N/A N/A N/A N/A N/A 0.26 0.12 0.025 0.258 0.153 0.004 0.208 0.019 0.001 0.033 0.412 0.647 0.119 0.183 0.001 0.292 0.154 0.011 0.046 P-Value P-Value P-Value ‰) Percent clay/silt dN15 (‰) concentration (uM) (ug/L) concentration concentration (uM) dN15 (‰) C/N (molar) dC13 (‰) dN15 (‰) Concentration* (ug/L) Concentration* C/N (molar) Concentration (uM) dC13 (‰) dN15 (‰) dN15 (‰) dN15 (‰) Concentration (uM) (ug/L) concentration** Concentration* (uM) Concentration* C/N (molar) Concentration (uM) Data Being Tested Data Being Tested (‰) dN15* Data Being Tested dN15 ( dC13 (‰) dN15 (‰) Mackworth Island vs. Maquoit Bay vs. Island Maquoit Mackworth Rivers vs. Rivers Rivers vs. Ocean Grain Size Ammonium Ammonium Chlorophyll Nitrate Nitrate Sediment Sediment Sediment Chlorophyll POM Ammonium POM ANOVA test *indicates ship EPA OSV the Bold by collected **data Ammonium POM Ammonium Ammonium Chlorophyll Nitrate Eelgrass Nitrate Sample Type Sample Type Nitrate Sample Type Eelgrass Eelgrass Nitrate Table 3.2: Results of t-test and ANOVA tests used to determine statistically significant differences in data. Green columns indicate sta - indicate columns Green data. in differences significant usedstatistically determine to tests ANOVA and t-test of 3.2: Results Table compar when done were tests ANOVA and areas, two comparing when done were T-tests data. inthe differences tistically significant areas. three ing

46 3.00

2.00

1.00

0.00

-1.00 N of Nitrate (‰) of Nitrate N 15 δ -2.00

-3.00

-4.00 0.00 5.00 10.00 15.00 20.00 25.00 Nitrate Concentration in Surface Water (uM) Mackworth Island Mackworth Island Average Maquoit Bay Maquoit Bay Average Figure 3.3: Mean nitrate concentration versus δ15N of nitrate from Mackworth Island and Maquoit Bay. Bars indicate standard deviation; n=6 for both Mackworth Island and Maquoit Bay, where two samples were analyzed from each of the three sites at both Mack- worth Island and Maquoit Bay.

Ammonia Concentration Nitrogen Recovered in the EA-C-IRMS Site using EPA 350.1(uM) from the Holmes et al. (1998) method (uM) Fore River 25.56 2.0 Presumpscot River limit of detection N/A Bunganuc River limit of detection 1.2

Mackworth 1 23.33 1.4 Mackworth 2 limit of detection 1.0 Mackworth 3 limit of detection 1.0 Mackworth Pore 1 77.78 N/A Mackworth Pore 2 43.89 N/A Mackworth Pore 3 66.67 N/A

Maquoit 1 29.44 1.5 Maquoit 2 14.44 0.3 Maquoit 3 limit of detection 0.6 Maquoit Pore 1 722.22 N/A Maquoit Pore 2 472.22 59.4 Maquoit Pore 3 388.89 7.5 limit of detection= 11.11 uM N/A= filter pack breakage Table 3.3: Ammonium concentration data for samples collected at the same sites on the same day. Ammonia EPA 350.1 method is used by North East Labs and Holmes et al. (1998) is the ammonia diffusion method used in this study.

47 values are attributed to the incomplete diffusion of ammonia onto the filter disk in the Holmes et al. (1998) method, likely due to the inadequate mixing on magnesium oxide. For these reasons, ammonium isotope data are not considered in this study. 3.2.3 Producers 3.2.3.1 Carbon and Nitrogen Isotopes

No significant differences were determined for13 δ C values of eelgrass and sediment between Mackworth Island and Maquoit Bay, however differences in 13δ C values for POM are significant (p=0.033) (Table 3.2; Figure 3.4). The offset between the two sites is an average of 3.6‰, with more enriched values at Mackworth Island. The δ15N values for POM and sediment (Figure 3.4) did not show a significant difference between the two sites, yet Eelgrass showed a significant difference in15 δ N between Maquoit Bay and Mackworth Island (p= 0.046). The offset of eelgrass δ15N averages between the two sites is 2.4‰, with more enriched values at Mackworth Island. 3.2.3.2 C/N Ratios The C/N ratios of eelgrass and sediment showed no significant difference between Maquoit Bay and Mackworth Island, however there is a significant difference in the C/N ratio of POM (p=.0001). The offset between the average C/N values at the two sites is 3.3, with a higher ratio at Mackworth Island (Table 3.2). 3.2.3.3 Chlorophyll There is no statistically significant difference between chlorophyll A concentrations in surface water at Maquoit Bay and chlorophyll A concentrations at Mackworth Island (Talbe 3.2). 3.2.3.4 Grain Size Analysis A statistically significant difference was determined for the grain size (Table 3.2; Figure 3.5) in the top 5 cm of sediment between Maquoit Bay and Mackworth Island (p=0.025). Maquoit Bay was shown to have an average of 26.1% more clay/silt relative than Mackworth Island.

48 Maquoit Bay

Maquoit Bay 16.00

14 14.00

12 12.00

n=3 10 10.00

n=3 8 ) 8.00 ‰ N (

6 15 6.00 n=3 n=3 δ N of Producers of N

4 4.00 15 n=3 δ n=3 2 2.00

0.00 0 -25.00 -20.00 -15.00 -10.00 -5.00 0.00 -2.00 -2 -22 -20 -18 -16 -14 -12 -10 -8 -6 13 δ13C (‰) δ C of Producers

Eelgrass POM Sediment Maquoit Bay Mackworth Island Figure 3.4: Mean nitrogen versus carbon isotope ratios for POM, eelgrass, and sediment at Mackworth Island and Maquoit Bay. Bars and circles indicate standard deviation . Statistically significant differences were determined for the15 δ N of eelgrass and the δ13C for POM.

49 Mackworth Island site weight of sample (grams) weight of sand weight of silt/clay % Clay/Silt (<0.06mm) % Sand 1.00 39.13 6.05 33.08 84.54 15.46 2.00 30.63 9.60 21.03 68.66 31.34 3.00 30.11 10.66 19.45 64.60 35.40 Mackworth Island 100% Maquoit Bay site weight of sample (grams) weight of sand weight of silt/clay % Clay/Silt (<0.06mm) % Sand 90% 1.00 31.41 0.04 31.37 99.87 0.13 2.00 32.74 0.16 32.58 99.51 0.49 80% 3.00 33.13 0.03 33.10 99.91 0.09 70%

60%

50% Silt/Clay 40% Sand

Percent of total Percent 30%

20%

10%

0% 1 2 3 Site Maquoit Bay

100%

90%

80%

70%

60%

50% Silt/Clay

40% Sand Percent of total Percent 30%

20%

10%

0% 1 2 3 Site

Figure 3.5: Grain size analysis from sediment samples taken at each site at Ma- quoit Bay and Mackworth Island. Percent sand (blue) versus percent clay/silt (red) were graphed.

50 3.3 Inputs: Rivers and Open Ocean 3.3.1 POM POM in the Fore River had δ13C values 8.5‰ enriched relative to the Bunganuc River, and 3.8‰ enriched relative to the Presumpscot River (Figure 3.6). The Presumpscot River had the most enriched δ15N values; 1.43‰ enriched relative to the Bunganuc River and 4.28‰ enriched relative to the Fore River. The C/N for the POM in the rivers showed that the Bunganuc River has the highest ratio, and the Presumpscot River and Fore River have relatively similar ratio values (Table 3.2). One POM sample was run per river, so no ANOVA could be used to determine differences between the 15δ N, δ13C, and C/N ratios of POM between the three rivers. No POM data were analyzed from the open ocean samples. 3.3.2 Chlorophyll A significant difference (p=0.019) was determined for chlorophyll A concentrations between the Presumpscot River and Bunganuc River (Table 3.2; Figure 3.7), where the Bungnauc River contained an average of 2.08 ug/L more chlorophyll A than the Presumpscot River. No significant differences were found for chlorophyll A concentration between the open ocean samples and the overall average for the rivers. 3.3.3 Nitrate No statistically significant differences in nitrate concentration or15 δ N values were determined between the Fore River, Presumpscot River, and Bunganuc River. However, when comparing average nitrate data in the rivers to average nitrate data for the open ocean samples (Table 3.2; Figure 3.8) a significant difference was determined in both nitrate concentration (p=0.001) and nitrate δ15N (p=0.011). There was an average offset of 5.38 umol/L, showing higher concentrations in the rivers. The offset between average15 δ N of the river and ocean is 7.4‰, with more enriched values in the rivers.

51 7.00

6.00

5.00

4.00 7.00

6.00 3.00

5.00

4.00 2.00 Rivers in POM for N 15 δ

3.00 1.00

2.00 Rivers in POM for N 15 δ

1.00 0.00 -26.00 -24.00 -22.00 -20.00 -18.00 -16.00 -14.00 -12.00 -10.00 4 0.00 -26.00 -24.00 -22.00δ13C -20.00 for POM-18.00 in Rivers-16.00 -14.00 -12.00 -10.00 δ13C for POM in Rivers 3.5 4 PresumopscotPresumopscot River River ForeFore River River BunganucBunganuc River River Figure 3.6: δ15N versus δ13C for POM in the Bunganuc River, Presumpscot River, and Fore 3 3.5 River.

3 4 2.5 3.5 2.5

2 3 2 a 2.5 1.5

1.5 2 ab 1 1 1.5 Chlorophyll A (ug/L) Chlorophyll Rivers for Chlorophyll A (ug/L) Chlorophyll Rivers for 1 N=3 0.5 0.5 ab Chlorophyll A (ug/L) Chlorophyll Rivers for b N=4 0.5 N=2 N=3 0 0 0 Bunganuc River Presumpscot River Fore River Open Ocean Bunganuc River Bunganuc RiverPresumpscotPresumpscot River River ForeFore River River Open OceanOpen Ocean

Figure 3.7: Mean chlorophyll A levels (ug/L) in surface water in the Bunganuc River, Pre- sumpscot River, and Fore River. Bars indicate standard deviation; “a” “ab” and “b” indicate a statistically significant difference between the rivers. Any rivers that share the same letter show no significatn difference.

52 5.00

5.00

0.00 5.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00

0.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 0.00 -5.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00

-5.00 -5.00 -10.00

-10.00 -10.00 N of Nirate (‰) for Inputs for (‰) Nirate of N

15 -15.00 δ N of Nirate (‰) for Inputs for (‰) Nirate of N N of Nirate (‰) for Inputs for (‰) Nirate of N 15 -15.00 δ 15 -15.00 δ

-20.00 -20.00 Concentration of Nitrate (uM) for Inputs Concentration of Nitrate (uM) for Inputs -20.00 Concentration of NitratePresumpscot (uM) River for Average Inputs Fore River Average Bunganuc River Average Ocean Surface Average Ocean Bottom Average Presumpscot River Average Fore River Average Bunganuc River Average Ocean Surface Average Ocean Bottom Average

Presumpscot River Average Fore River Average Bunganuc River Average Ocean Surface Average Ocean Bottom Average

Figure 3.8: Mean concentration of nitrate versus δ15N of nitrate for all the river and ocean water samples. Standard deviation is indicated by the bars; n=3 for all the rivers and the ocean bottom, and n=4 for the ocean surface.

53 4. Discussion

54 4.1 Source of DIN to Mackworth Island and Maquoit Bay The concentration and 15δ N value of nitrate in surface water at Mackworth Island show no significant difference between that of the Fore River and Presumpscot River, however there is a significant difference in the δ15N and concentration of nitrate between Mackworth Island and the open ocean samples (Figure 4.1). This indicates that the majority of nitrate in the water column at Mackworth Island comes from the rivers, not an open ocean source. In addition to the Presumpscot and Fore Rivers, it is possible that inland run-off through groundwater and direct shore run-off also influence the nitrate concentration and δ15N values at Mackworth Island. The same trends are apparent for Maquoit Bay (Figure 4.1); almost all nitrate in the Bay is derived from the Bunganuc River, and other sources of inland run-off. These findings suggest that nearshore estuarine regions are significantly influenced by river influxes, in agreement with Schlesinger (1997).

4. 2 Quantifying for Enriched Nitrogen in Eelgrass and Nitrate Both eelgrass and nitrate are enriched in 15N at Mackworth Island relative to Maquoit Bay. The fractionation between eelgrass and nitrate in the surface water is consistent between Mackworth Island and Maquoit Bay, at approximately 8‰, where the nitrate is depleted in 15N relative to eelgrass (Figure 4.2). Published fractionation values for primary producers that rely solely on nitrate as a source of nitrogen show that nitrate is enriched in 15N by 5-10‰ relative to the primary producers (summarized in York et al., 2007). Thus, other sources of nitrogen (ammonium, nitrite, and dissolved organic nitrogen (DON)) are also, not surprisingly, being assimilated by eelgrass. Although water column nitrite and DON concentration data do not exist, nitrate and ammonium are the most preferred sources of nitrogen to be assimilated by primary producers (York et al., 2007). In addition nitrite is very unstable in the presence of oxygen, thus is almost immediately converted into nitrate during nitrification (Sharp, 2007). Numerous studies have shown that ammonium is preferentially assimilated over

55 A) 5.0 Maquoit Bay

0.0 n=6

(‰) -5.0 n=3

n=7

N of Nirate Nirate of N -10.0 15 δ

-15.0

-20.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Nitrate Concentration (uM) Bunganuc River Average Open Ocean Average Maquoit Bay Average

B) 5.0 Mackworth Island n=6 0.0 n=6

-5.0

n=7 -10.0 N of Nirate (‰) (‰) of Nirate N 15

δ -15.0

-20.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Nitrate Concentration (uM) River Average (Presumpscot and Fore) Open Ocean Average Mackworth Island Average

Figure 4.1: δ15N of nitrate versus concentration of nitrate for A) Maquoit Bay and its inputs, and B) Mackworth Island and its inputs.

56 10.00

Eelgrass 8.00 n=5

Eelgrass 6.00 n=3 7.68 ± 1.72 4.00 N (‰) 15

δ 2.00 8.44 ± 1.37 Nitrate

0.00 n=6

-2.00 Nitrate n=6

-4.00 Maquoit Bay Mackworth Island

Figure 4.2: Average δ15N of eelgrass and nitrate at Mackworth Island and Maquoit Bay. Bars indicate standard deviation for each point.

57 nitrate (Dortch, 1990; L’Helguen et al., 1996). Eelgrass specifically, can assimilate nitrogen from the water column through its leaves or from the pore water through its roots. Often ammonium concentrations are much higher in the pore water than the surface water (Romero et al. 2006), so eelgrass assimilates the large majority of its nitrogen from pore water. Using a model developed by Zimmerman et al. (1987) and the average pore water ammonium concentrations measured at both sites, the proportion of water column and pore water ammonium assimilated by eelgrass can be estimated at Maquoit Bay and Mackworth Island (Figure 4.3). According to the model, eelgrass from Mackworth Island assimilates ~50% nitrogen from pore water ammonium, ~35% from water column ammonium and ~15% from water column nitrate. Eelgrass at Maquoit Bay assimilates ~70% of its nitrogen from ammonium in the pore water, ~20% from ammonium in the water column, and ~10% from nitrate in the water column. Assuming eelgrass and nitrate isotope data in the study are correct and the Zimmerman et al. (1987) model is legitimate, a two end member mixing model can be used to determine the average δ15N of ammonium being assimilated by eelgrass at Maquoit Bay and Mackworth Island. δ15N Eelgrass= (δ15N Nitrate-En)(fn)+(δ15N Ammonium-Ea)(fa) fn=fraction of nitrogen derived from nitrate fa=fraction of nitrogen derived from ammonium En= fractionation between eelgrass and nitrate= -7‰ (Horrigan et al., 1990) Ea= fractionation between eelgrass and ammonium= -5‰ (Montoya et al. 1991) Maquoit Bay: 5.7‰= (-2.7‰-7‰)(0.10)+(δ15N Ammonium-5‰)(0.90) δ15N Ammonium= 12.4‰

Mackworth Island:

8.1‰= (0.4‰-7‰)(0.15)+(δ15N Ammonium-5‰)(0.85)

δ15N Ammonium= 15.7‰ 58 Pore Water Ammonium

Water Column Ammonium

Water Column Nitrate

Figure 4.3: The effect of pore water ammonium concentrations on patterns of nitrogen uptake by eelgass from pore water ammonium, water column ammonium, and water column nitrate. Each component is graphed as a percentage of total nitrogen taken up each day in a model developed by Zimmerman et al. (1997). The red line represents the average concentration of ammonium in pore water at Mackworth Island. This shows that eelgrass at Mackworth Island accumulates about 15% of its nitrogen from water column nitrate, 35% from water column ammonium, and 50% from pore water ammonium. The blue line represents Maquoit Bay, where the eelgrass accumulates 10% of its nitrogen from water column nitrate, 20% from water column ammonium, and 70% from pore water ammonium.

59 The two-end-member mixing model calculates that the 15δ N of ammonium at Mackworth Island is about 15.7‰, and the δ15N for ammonium at Maquoit Bay is about 12.4‰, in agreement with Middelburg and Nieuwenhuize (2001). Both of these values were corrected for fractionation during assimilation based on published data from studies done at the Chesapeake Bay, which analyzed the fractionation between phytoplankton and ammonium/nitrate (Horrigan et al., 1990b; Montoya et al., 1991). It should be noted that in this model, ammonium in pore water is not differentiated from ammonium in surface water. It is unlikely that both sources of ammonium will have identical δ15N values or be fractionated consistently. Thus these calculated ammonium 15δ N values provide only rough estimates of what is truly expected. Inputs from rivers and run-off drive the nitrate and ammonium in the water column, while ammonium in pore water is supplied by regeneration within the estuary (Romero et al., 2006). Future work should focus on isotopic analysis of ammonium in the water column and pore water.

4.2.1 Possible Causes of 15N Enrichment in Eelgrass and Nitrate Nitrogen-15 enrichment in eelgrass and nitrate at one site over another may be due to one or a combination of three factors: 1) rates of primary production, where high rates of primary production can result in more enriched 15N values if concentration of DIN is similar between the sites (Nixon et al. 1986; Fogel and Cifuentes, 1993), 2) the degree to which internal cycling of nitrogen occurs, and 3) the amount and sources of nitrogen inputs, where high amounts of anthropogenic waste could result in more 15N enriched nutrients and primary producers (Schlesinger, 1997). The possibility of each of these influences on15 N values at Mackworth Island and Maquoit Bay will be considered below. In 2009, the Casco Bay Estuary Partnership collected data that showed the eelgrass bed at Maquoit Bay was significantly more dense than the eelgrass bed at Mackworth Island (p=0.022) (Table 4.1), suggesting higher rates of primary productivity in the eelgrass beds at Maquoit Bay. Primary production of eelgrass is controlled by three

60 Location % Cover Average Height (cm) Maquoit Bay 96.25 79.75 100.00 105.00 100.00 102.50 55.00 65.25 100.00 103.25 81.25 69.50 100.00 110.00 100.00 80.00 96.25 85.75 100.00 107.00 100.00 89.00 100.00 61.25 86.25 73.50 80.00 57.00 Average 92.50 84.91

Mackworth Island 71.25 43.00 13.75 71.00 61.25 88.50 65.00 77.50 78.75 72.75 87.50 86.25 97.50 75.75 45.00 77.00 Average 65.00 73.46 p-value 0.022 0.132 Table 4.1: Percent cover (density) and average height data for the eelgrass beds at Mackworth Island and Maquoit Bay collected in 2009 by the Casco Bay Estuarine Partnership.

61 main factors: total influx of DIN (from outside sources and recycling within the system), temperature, and light transparency of the water column. Data determined in this study show no significant difference in total nitrate concentrations in surface water between Mackworth Island and Maquoit Bay. In addition, data collected in 2010 by the Casco Bay Estuary Partnership shows no significant difference in total DIN concentrations at sites very close to the sites used in this study in June and July of 2010 (Figure 4.4; Figure 4.5). At the time of sampling for this study temperature (Table 4.2) was significantly higher at Maquoit Bay relative to Mackworth Island (p=0.000). Chlorophyll A concentrations in the water column were not statistically different between the two sites, suggesting that temperature had minimal effect on primary production. No data on water transparency was collected, although this would be very important for a future since light availability is a major control for eelgrass production. One reason for increased amounts of eelgrass at Maquoit Bay may be due to differences in pore water ammonium concentrations, which were significantly higher at Maquoit Bay (p=0.044; Table 4.3). The increased concentration of ammonium in the pore water is a function of breakdown of organic matter in the soil (ammonification) and exchange of ammonium between the pore water and surface water. Since the eelgrass bed at Maquoit Bay is currently denser than at Mackworth Island, more organic matter is most likely being broken down at Maquoit Bay, which can foster increased rates of ammonification in the sediments. Maquoit Bay also has finer sediments in the top 8 cm of the sea floor compared to Mackworth Island, which may allow less ammonium to diffuse out of the pore water into the water column, where it is more easily diluted by circulating water. It appears that one reason for higher rates of eelgrass production at Maquoit Bay is higher concentrations of ammonium pore water. This may be due to increased cycling of nitrogen within the system, lower diffusion rates of ammonium out of the pore water

62 40 Fore River

35

30

25

20

15

10 Concentration (uM) Concentration 5

0 January February March April May June July August September

East Portland 40

35

30

25

20

15

10

Concentration (uM) Concentration 5

0 January February March April May June July August September

Nitrate Ammonium DIN OutsideEast Maquoit Portland Bay 40 40

35 35

30 30

25 25

20 20

15 15

10

10 (uM) Concentration

Concentration (uM) Concentration 5 5

0 0 January JanuaryFebruaryFebruary MarchMarch April MayMay June June July JulyAugust AugustSeptember September

Nitrate Ammonium DIN

Figure 4.4: Concentration of nitrate, ammonium, and total DIN colleceted by the Casco Bay Estuary Partnership in 2010. Figure 4.5 shows that location of each of these sites.

63 Outside Maquoit Bay

East Portland Fore River

Figure 4.5: The locations of the three sites where DIN concentrations were sampled by the Casco Bay Estuary Partnership in 2010.

64 Site Date Temp C Average temp Maquoit 6-Jul 23.42 Maquoit 6-Jul 28.9 Maquoit 6-Jul 23.09 25.14 Mackworth 13-Jul 19.16 Mackworth 13-Jul 19.08 Mackworth 13-Jul 19.58 19.27 p=0.00 Table 4.2: Differences in temperature between Mackworth Island and Maquoit Bay on the days when water and eelgrass samples were collected.

Ammonia Concentration Average Ammonium Site using EPA 350.1(uM) Concentration (uM) Mackworth Pore 1 77.78 Mackworth Pore 2 43.89 Mackworth Pore 3 66.67 62.78

Maquoit Pore 1 722.22 Maquoit Pore 2 472.22 Maquoit Pore 3 388.89 527.78 p=0.044

Table 4.3: Ammonium concentrations (uM) of pore water at Mackworth Island and Maquoit Bay. Concentrations were measured by North East Labs, using EPA method 350.1.

65 (due to the fact that the sediments are finer at Maquoit Bay), or a combination of both.

Increased primary production could yield enriched δ15N values, if DIN concentrations are constant, but since Maquoit Bay shows higher total concentrations of DIN (surface and pore water combined) than Mackworth Island, the differences in primary productivity may not impact the δ15N of eelgrass. However, eelgrass at Maquoit Bay is shown to assimilate more ammonium from pore water (increased internal cycling of nutrients) than eelgrass at Mackworth Island, which can effect the δ15N value of eelgrass. Little to no fractionation occurs during mineralization/ammonification, thus the ammonium produced in pore water likely has δ15N values that are relfective of the primary producers decomposed in the sediment. These producers, assuming to have assimilated DIN from

the same system, have δ15N values about 3‰ depleted relative to its nitogen source (DIN) due to the -3‰ fractionation that occurs during assimilation (Macko and Ostrom, 1994). So, it is probable that ammonium in the pore water is depleted in 15N relative to ammonium and nitrate in the water column. If this is true, this can cause nitrogen depletion of eelgrass at Maquoit Bay relative to Mackworth Island since Maquoit Bay eelgrass seems to assimilate more nitrogen from the pore water than Mackworth Island eelgrass. Although river data show no significant difference in15 δ N of nitrate between the Presumpscot and Fore Rivers and the Bunganuc River, enrichment of the water could be caused by wastewater treatment plants very close to Mackworth Island (Figure 4.6). This group of sewer overflows and wastewater outfalls includes the overflows and outfalls of Falmouth Wastewater Treatment Facility and Portland Water District, two treatment plants that are considered major outflows by the Maine Department of Environmental Protection (2010). These two facilities could possibly be the primary cause of DIN enrichment in the surface water at Mackworth Island (Heaton, 1986). In order to determine if this is true, the average discharge, concentration, and δ15N of DIN would have to be collected from each facility’s outfalls.

66 2 Kilometers

N

Figure 4.6: A map of sewer overflow outfalls (red) and wastwater treatment plant facility outfalls (blue) in the Mackworth Island area.

67 With these findings, it is probable that increased 15δ N values in eelgrass found at Mackworth Island are caused by both increased δ15N of source DIN and a decreased proportion of ammonium assimilated from pore water relative to Maquoit Bay.

4.3 Temporal Variability of Eelgrass Isotope Data Data from Loopesko (2010) show δ15N values in eelgrass about 3‰ lower at both Maquoit Bay and Mackworth Island compared to samples presented in this study (Figure 4.7). These results suggest that some systematic shift in nitrogen isotope cycling has occurred in Casco Bay over the last year. However, a comparison of data collected nearly 10 years ago by Neckles (2001) in Maquoit Bay show δ15N values of eelgrass that are very similar to those at Maquoit Bay in this study. More continuous sampling (annual and monthly) is needed to determine the significance and implications of the shifts in15 N data through the last decade.

68 Maquoit Bay 10.0 9.0 8.0 7.0 6.0 5.0 4.0 3.0 δ15N of δ15N Eelgrass 2.0 of Eelgrass δ15N 1.0 0.0 2000 2002 2004 2006 2008 2010

Mackworth Island 10.0 9.0 8.0 7.0 6.0 5.0 4.0 3.0

δ15N of δ15N Eelgrass 2.0 1.0 0.0 2008 2009 2010

Figure 4.7: A comparison of δ15N of eelgrass sampled at Mackworth Island and Maquoit Bay in difference studies over the past 10 years. In 2001 Hilary Neckles (blue diamond) sampled eelgrass is Maquoit Bay, in 2009 William Loopesko (red square) sampled eelgrass from both Maquoit Bay and Mackworth Island, and in 2010 eelgrass was sampled for this study (green triangle).

69 5. Conclusions

70 Enriched 15N values were found in eelgrass in the more heavily populated system at Mackworth Island. While δ15N in eelgrass would appear useful for identifying sites that are influenced by anthropogenic nutrients, it is not definitive. The15 δ N of eelgrass reflects a combination of nitrogen sources and isotopic fractionations that occur during plant synthesis, all of which vary from site to site. To identify sites in Casco Bay that may be affected by the input of anthropogenic nutrients, it is recommended to focus on the nitrate and ammonium δ15N and concentration of source water and pore water, and secondarily eelgrass δ15N values. Source waters (rivers, groundwater, etc.) must be sampled for DIN instead of estuarine water because the concentration and δ15N is unaffected by primary production, which largely occurs in the estuary. Both source water and wastewater treatment plant inputs must be accounted for when sampling for ammonium and nitrate coming into the areas. It is also essential to analyze pore water ammonium in order to account for nitrogen that is recycled with the system. Once all waters are analyzed, eelgrass and other primary producers should be tested for nitrogen isotopes to determine if enrichment of inputs is affecting the nitrogen isotope ratios of primary producers. If so, then it can be proven that an area is being affected by anthropogenic deposition of nitrogen via wastewater and organic fertilizer input. Perhaps a better focus of research, in terms of identifying nutrient hot spots in Casco Bay, would be to expand on the water/nutrient sampling program currently underway by Friends of Casco Bay and Mike Doan. The influence of both inorganic fertilizer runoff and atmospheric nitrogen from fossil fuels (especially in larger watersheds) on near shore systems is expected to grow with time, and is best detected by measuring the concentrations in key locations. Ammonium created from the Haber process is very common today, and the large majority of large farms use inorganic fertilizers (Canfield et al., 2010). Both synthetic fertilizers and nitrogen precipitated out of the atmoshere (from fossil fuels) are not isotopically distinguishable from natural soils,

71 thus would be difficult to document using δ15N approaches presented in this study. Coastal areas with high influxes of terrestrial water sources containing high concentrations of DIN are often not noticed until overgrowth and potential eutrophication has already begun to occur. Although concentrations of DIN in the Casco Bay are currently not thought to be a big threat to near shore systems, if the population of Maine continues to grow and more land becomes farmland, the influx of nitrogen from inorganic fertilizers, in addition to organic fertilizers and wastewater must be accounted for in order to prevent nitrogen pollution that can lead to eutrophication.

72 References

Boyer, E.W., Goodale, C.L., Jaworski, N.A., Howarth, R.W. 2002. Anthropogenic nitrogen sources and relationships to riverine nitrogen export in the northeastern USA. Biogeochemistry, v. 57-58, pp. 137-169.

Canfield, D.E., Glazer, A.N., Falkowski, P.G. 2010. The Evolution and Future of Earth’s Nitrogen Cycle. Science. v. 330, pp. 192-196.

Casco Bay Esuary Partnership. 2010. State of the Bay Report. Portland, Maine: CBEP

Colorado Plateau Isotope Laboratory. 2010. Northern Arizona University: Flagstaff, AZ.

DeBruyn, M.H. and Rasmussen, J.B. 2002. Quantifying Assimilation of Sewage-Derived Organic Matter by Riverine Benthos. Ecological Applications, v. 12, pp. 511-520.

Deegan, L. and Wright, A. 2002. Nitrogen loading alters seagrass ecosystem structure and support of higher trophic levels. Aquatic Convervation: Marine and Freshwater Ecosystems, pp. 193-212.

Doan, M. 2010. Personal Communication. Friends of Casco Bay.

Dortch, Q. 1990. The interaction between nitrate and ammonium uptake in phytoplankton. Marine Ecology. Ser. 61. 183-201.

Engel, Michael and Macko, Stephan. 1993. Organic Geochemistry: Plenum Press. New York.

Fogel, M.L., Cifuentes, L.A., 1993, Isotope fractionation during primary production, in, Engel, M.H., and Macko, S.A., Organic Geochemistry Principles and Applications: Plenum Press, New York, pp. 73-100.

Fry, B., 2002, Stable Isotope Ecology: Springer science and business media, New York.

Galloway, J.N. 2004. The Global Nitrogen Cycle in Schlesinger, William H., Holland, H.D., and Turekian, K.K., Treatise on Geochemistry. v. 8, Biogeochemistry: Elsevier Pergamon, Amsterdam, pp. 557-584.

Gormly, J.R., and Spaulding, R.F. 1979. Sources and Concentrations of nitrate-nitrogen in ground water of the Central Platte Region, Nebraska. Groundwater, v. 89, pp. 1264- 1269.

73 Heaton, T. H. E. (1986). Isotopic studies of nitrogen pollution in the hydrosphere and atmosphere: a review. Chemical Geology. 59: 87-102.

Hoefs, J. 2004. Stable Isotope Geochemistry. Springer: New York, New York.

Holmes, R.M., McClelland, J.W., Sigman, D.M., Fry, B., Peterson, B.J. 1998. Measuring 15 + N-NH4 in marine, estuarian and fresh waters: An Adaption of the ammonia diffusion method for samples with low ammonia concentrations. Marine Chemitry, v. 60, pp. 235-243.

Horrigan, G. and Springer, A.L. 1990. Oceanic and Estuarine Ammonium Oxidation: Effects of Light. American Society of Limnology and Oceanography. v. 45, pp. 479- 482.

Jaffe, D.A. 1992. The Nitrogen Cycle, in, Butcher, Samuel S., Charlson, Robert J., Orians, Gordon H., Wolfe, Gordon V. Global Biogeochemical Cycles: Academic Press, London, pp. 263-282.

Kendall, C., and McDonnell, J.J. 1998. Nitrogen Isotopes in Catchment Hydrology. Elsevier: Amsterdam.

Lajtha, K. and Michener, R.H. 1994. Stable Isotopes in Ecology and Environmental Science. Blackwell Scientific Publications. Oxford, England.

Lee, K.S. and Short, F.T. 2004. Development of nutrient pollution indicator using the seagrass, Zostera marina, along nutrient gradients in three New England estuaries. Aquatic Botany, pp. 197-216.

L’Helguen, S., Madec, C., and Le Corre, P. 1996. Nitrogen uptake in permanetaly well- mixed temperate coastal waters. Estuarine Coastal Shelf Science. v. 42, pp. 803-818.

Libby, S. and Anderson, D. 2010. Red Tides in Inshore and Offshore Casco Bay and Their Relationship to Local and Gulf of Maine Physical and Biological Conditions. Casco Bay Estuarian Partnership.

Loopesko, W. 2010. Stable Isotopes in Zostera marina as a Proxy for Anthropogenic Nitrogen Loading in Casco Bay, Maine. Bates College Department of Geology. Lewiston, Maine.

Macko, S.A., Ostrom, N.E. 1994. Pollution studies using stable isotopes, in, Lajtha, K., Michener, R.H., Stable Isotopes in Ecology and Environmental Sciences. Blackwell

74 Scientific Publications. Oxford, England.

Maine Geologic Survey. 2008. Bedrock Geology of the Bath 1:200,000 Map Sheet, Coastal Maine.

Marvinney, R.G. 2002. Simplified Bedrock Geologic Map of Maine. Maine Geologic Survey: Department of Conservation.

McClellend, J. and Valiela I., 1998. Linking Nitrogen In Estuarian Producers to land- derived sources. Liminology and Oceanography, v. 43, pp. 577-585.

Middleburg, J.J., Nieuwenhuize, J. 2001. Nitrogen isotope tracing of dissolved inorganic nitrogen behavior in tidal estuaries. Estuarine, Costal and Shelf Sciences, v. 53, pp. 385-391.

Montoya, J.P. 1994. Natural abundance of 15N in marine planktonic ecosystems, in Michener, Robert H., and Lajtha Kate. Blackwell, Oxford. pp. 174-194.

Neckles, H. 2010. Personal communication.

Nixon, S.W., Oviatt, C.A., Frithsen, J., Sullivan, B. 1986. Nutrients and the productivity of estuarine and costal marine ecosystems. Journal of the Limnological Society of Southern Africa,v.12, pp. 43-71.

Nixon, S. W.; Oviatt, C. A.; Frithsen, J.; Sullivan, B.1986. Nutrients and the productivity of estuarine coastal marine ecosystems. Journal of the Limnological Society of Southern Africa. v. 12.

Rabalais, N.N. 2002. Nitrogen in Aquatic Ecosystems. Ambio v. 31, pp. 102-112.

Romero, J., Lee K.S., Perez, M., Mateo, M.A., Alcoverro, T. 2006. Nutrient Dynamics in Seagrass Ecosystems, in, Larkum, A.W.D., Orth, R.J., Duarte, C.M., Seagrasses: Biology, Ecology, and Conservation. Springer, The Netherlands.

Schlesinger, W.H. 1997. Biogeochemistry: An analysis of Global Change: Acedemic Press, San Diego.

Sharp, Z. 2007. Principles of Sable Isotope Geochemistry. Pearson Education, Upper Saddle River, NJ.

75 Sigman, D.M., Altabet, M.A., Michener, R., MCorkle, D.C., Fry, B., Holmes, R.M. 1997. Natural abundance-level measurement of the nitrogen isotopic composition of oceanic nitrate: an adaptation of ammonia diffusion method. Marine Chemistry, v. 57, pp. 227-242.

Thomson, W.B., Born, H.W. 1985. Surficial Geologic Map of Maine: Maine Geologic Survery, Department of Conservation.

Tiessen H., Karamanos, R.E., Stewart, J.W.B., and Stelles F. 1984. Natural nitrogen-15 abundance as an indicator of soil organic matter transformations in native and cultivated soils. Soil Science Society of America Journal, v. 48, pp. 312-315.

Winteringham, F.P.W. 1984. Soil and Dertilizer Nitrogen. IAEA. Vienna.

York, J. and Tomasky, G. 2007. Stable isotopic detection of ammonium and nitrate assimilation by phytoplankton in the Waquoit Bay estuarine system. Liminology and Oceanography, pp. 144-155.

Zimmerman, R.C., Smith, R.D., Alberte, R.S. 1987. Is growth of eelgrass nitrogen limited? A numerical simulation of the effects of light and nitrogen on the growth dynamics of Zostera marina. Marine Ecology, v. 41, pp. 167-176.

76 ArcGIS Layers

Coastl_04202006. Downloaded February 7th, 2007. Maine Office of GIS. http://www. maine.gov/megis/

Eelgrass05. Downloaded November 11th, 2009. Maine Department of Marine Resources. http://megis.maine.gov/data/metadata/eelgrass05/htm

National Agriculture Imeragy Program (NAIP). Downloaded February 20th, 2011.USDA- FSA-APFO Aerial Photography Field Office, U.S. Geologic Survey. http://seamless. usgs.gov/website/seamless/viewer.htm.

77 Appendix 1 Mackworth Island Isotope Data Mackworth Island: Eelgrass, POM, and Sediment Sample ID Type (mg) % N N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) Site 7736 Eelgrass 2.17 1.62 2.72 8.1 37.68 69.76 -10.6 25.66 1 7737 Eelgrass 2.32 1.58 2.81 8.0 37.18 73.34 -10.6 26.07 1 7738 Eelgrass 2.40 1.58 2.92 8.0 37.95 77.54 -10.6 26.59 1 7739 Eelgrass 2.84 1.62 3.54 8.3 37.78 91.30 -9.8 25.81 2 7740 Eelgrass 2.93 1.93 4.36 8.2 40.14 100.34 -9.2 22.99 3

7741 POM 0.00 1.26 6.1 0.00 10.67 -16.0 8.50 1 7742 POM 0.00 1.44 4.2 0.00 12.40 -16.0 8.62 2 7743 POM 0.00 1.30 1.2 0.00 11.16 -16.1 8.61 3

7744 Sediment 8.41 0.08 0.47 14.4 0.96 6.84 -12.9 14.42 1 7745 Sediment 8.55 0.08 0.53 7.0 1.64 11.88 -14.0 22.41 2 7746 Sediment 7.68 0.08 0.49 6.5 1.01 6.56 -11.8 13.51 3

Mackworth Island: Water DIN Sample ID Type N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) Site 7747 NO3 7.62 -0.7 0.00 0.87 -28.6 0.11 1 7748 NO3 3.88 2.7 0.00 0.76 -27.2 0.19 1 7749 NO3 3.19 -1.9 0.00 1.04 -28.2 0.33 2 7750 NO3 2.74 1.4 0.00 1.24 -27.9 0.45 2 7751 NO3 3.20 0.9 0.00 0.90 -28.6 0.28 3 7752 NO3 2.83 0.2 0.00 1.06 -28.3 0.38 3

7753 NH4 1.41 -5.1 0.00 0.82 -27.3 0.58 1 7754 NH4 0.93 -5.9 0.00 1.07 -25.7 1.15 2 7755 NH4 1.12 -1.5 0.00 0.74 -27.1 0.66 2 7756 NH4 0.55 -7.9 0.00 0.91 -26.3 1.67 3 7757 NH4 1.04 -7.5 0.00 0.97 -26.1 0.93 3

Yellow columns indicate that nitrogen recovered is less than limit of detection for the EA- C-IRMS (0.8uM). For this reason isotope data is inaccurate.

78 Appendix 2 Maquoit Bay Isotope Data Maquoit Bay: Eelgrass, POM, and Sediment Sample ID Type (mg) % N N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) Site Eelgrass 7758 Eelgrass 2.59 2.85 5.69 5.1 42.15 93.02 -12.2 16.34 1 7759 Eelgrass 2.56 3.36 6.64 5.2 46.66 101.77 -14.3 15.34 2 7760 Eelgrass 3.03 1.76 4.12 6.8 40.23 103.96 -10.7 25.26 3

7761 POM 0.00 2.95 5.5 0.00 15.52 -20.3 5.26 1 7762 POM 0.00 2.69 5.6 0.00 14.05 -20.4 5.22 2 7763 POM 0.00 1.63 4.8 0.00 8.96 -18.3 5.48 3

7764 Sediment 8.03 0.17 1.01 4.2 1.82 12.37 -8.5 12.26 1 7765 Sediment 8.35 0.15 0.91 4.8 1.51 10.67 -10.4 11.67 2 7766 Sediment 7.54 0.03 0.19 -0.7 0.32 2.05 -11.0 10.63 3

Maquoit Bay: Water DIN Sample ID Type N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) Site 7767 NO3 7.14 -3.6 0.00 1.06 -28.1 0.15 1 7768 NO3 4.05 -2.4 0.00 0.95 -26.7 0.23 1 7769 NO3 2.91 -2.2 0.00 1.67 -23.4 0.57 2 7770 NO3 23.03 -3.0 0.00 1.11 -27.8 0.05 2 7771 NO3 21.63 -2.8 0.00 0.86 -27.6 0.04 3 7772 NO3 3.91 -2.3 0.00 1.15 -27.7 0.30 3

7773 NH4 1.47 -8.1 0.00 1.03 -27.3 0.70 1 7774 NH4 1.49 -3.7 0.00 0.93 -26.0 0.62 1 7775 NH4 0.27 -22.1 0.00 0.89 -26.9 3.32 2 7776 NH4 0.25 -15.5 0.00 0.93 -26.9 3.70 2 7777 NH4 0.64 -7.3 0.00 0.89 -26.8 1.40 3 7778 NH4 0.64 -1.0 0.00 0.97 -27.0 1.52 3

7779 NH4 Pore 59.36 -37.7 0.00 0.73 -26.4 0.05 2 7780 NH4 Pore 7.45 -35.3 0.00 1.03 -26.4 0.55 3 Yellow columns indicate that nitrogen recovered is less than limit of detection for the EA- C-IRMS (0.8uM). For this reason isotope data is inaccurate.

79 Appendix 3 River Isotope Data

Bunganuc River Sample ID Type (mg) % N N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) 7781 POM 0.00 1.63 4.9 0.00 12.62 -24.4 7.74

7782 NO3 0.00 7.47 -2.1 0.00 1.09 -27.9 0.15 7783 NO3 0.00 7.12 -6.4 0.00 0.94 -29.1 0.13 7784 NO3 0.00 13.28 -7.6 0.00 0.75 -28.5 0.06

7785 NH4 0.00 0.98 -27.2 7786 NH4 0.00 0.97 -26.9 7787 NH4 0.00 1.16 1.4 0.00 0.97 -25.3 0.83

Presumpscot River Sample ID Type (mg) % N N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) 7788 POM 0.00 1.47 6.4 0.00 6.76 -19.6 4.59

7789 NO3 0.00 4.42 1.4 0.00 1.00 -27.1 0.23 7790 NO3 0.00 7.64 -5.9 0.00 0.73 -28.5 0.10 7791 NO3 0.00 9.25 -0.2 0.00 0.95 -26.8 0.10

7792 NH4 0.00 0.91 -26.7 7793 NH4 0.00 0.96 -26.9 7794 NH4 0.00 0.96 -27.8

Fore River Sample ID Type (mg) % N N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) 7795 POM 0.00 2.08 2.2 0.00 8.44 -15.8 4.05

7796 NO3 0.00 4.29 -1.8 0.00 1.16 -28.8 0.27 7797 NO3 0.00 3.94 -1.7 0.00 0.87 -28.2 0.22 7798 NO3 0.00 4.67 -3.5 0.00 0.88 -28.1 0.19 Average 4.30 -2.3 0.23 std. dev. 0.37 1.0 0.04

7799 NH4 0.00 0.70 -17.9 0.00 0.94 -26.5 1.34 7800 NH4 0.00 4.53 -19.3 0.00 1.06 -26.1 0.23 7801 NH4 0.00 0.70 -16.4 0.00 0.89 -27.0 1.27

Yellow columns indicate that nitrogen recovered is less than limit of detection for the EA- C-IRMS (0.8uM). For this reason isotope data is inaccurate.

80 Appendix 4 Open Ocean Isotope Data

Ocean Inputs Sample ID Location Type (mg) % N N (uM) d15N (3pts) %C C (uM) d13C C/N (Molar) 7802 Bates Surface A NO3 0.00 1.29 -18.3 0.00 0.61 -24.8 0.47 7803 Bates Surface B NO3 0.00 2.65 -7.3 0.00 0.56 -25.9 0.21 7804 r1-r15 Surface A NO3 0.00 1.32 -5.8 0.00 0.47 -27.8 0.35 7805 r1-r15 surface B NO3 0.00 1.33 -11.6 0.00 0.57 -26.2 0.43

7806 r1-r15 Bottom A NO3 0.00 1.50 -5.5 0.00 0.41 -26.3 0.27 7807 r1-r15 Bottom B NO3 0.00 1.35 -17.1 0.00 7808 Bates Bottom A NO3 0.00 0.75 -14.9 0.00 0.42 -24.5 0.56 7809 Bates Bottom B NO3 0.00 2.12 -7.9 0.00 0.55 -25.0 0.26

7810 r1-r15 Bottom B NH4 0.00 0.18 5.6 0.00 0.43 -28.3 2.38 7811 Bates Surface B NH4 0.00 0.21 1.7 0.00 0.46 -23.9 2.25 7812 Bates Bottom B NH4 0.00 0.41 9.0 0.00 7813 r1-r15 Surface B NH4 0.00 0.22 5.7 0.00 0.41 -26.0 1.86 7814 Bates Bottom A NH4 0.00 0.18 9.8 0.00 7815 r1-r15 Bottom A NH4 0.00 0.39 9.4 0.00 7816 r1-r15 Surface A NH4 0.00 0.15 3.2 0.00 0.40 -25.8 2.68

Yellow columns indicate that nitrogen recovered is less than limit of detection for the EA- C-IRMS (0.8uM). For this reason isotope data is inaccurate.

81 Appendix 5 Chlorophyll Data Site Sample Date Replicate Chlorophyll A (ug/L) Average Chlorophyll A (ug/L) Maquoit Bay 1 6-Jul A 2.82 B 1.78 C 2.21 2.27 Maquoit Bay 2 6-Jul A 4.56 B 7.00 C 5.54 5.70 Maquoit Bay 3 6-Jul A 3.49 B 3.39 C 4.60 3.83 Site Average 3.93 Site std. dev. 1.72

Mackworth Island 1 6-Jul A 2.82 B 1.68 C 1.73 2.07 Mackworth Island 2 6-Jul A 2.60 B 3.01 C 3.33 2.98 Mackworth Island 3 6-Jul A 1.60 B 1.94 C 1.78 1.78 Site Average 2.28 Site std. dev. 0.63 Site Sample Date Replicate Chlorophyll A (ug/L) Average Chlorophyll A (ug/L) Bunganuc River 19-Jul A 3.83 B 2.80 C 1.92 2.85 std. dev. 0.95

Presumpscot River 19-Jul A 0.90 B 0.68 C 0.73 0.77 std. dev. 0.12

Fore River 19-Jul A 1.16 B 1.05 1.10 std. dev. 0.08

R1-R15 (Ocean) 2-Jul A 2.45 B 2.67 C 0.83 1.98 std. dev. 1.00

Bates (Ocean) 2-Jul N/A

82 Appendix 6 Grain Size Data Mackworth Island site weight of sample (grams) weight of sand weight of silt/clay % Clay/Silt (<0.06mm) % Sand 1.00 39.13 6.05 33.08 84.54 15.46 2.00 30.63 9.60 21.03 68.66 31.34 3.00 30.11 10.66 19.45 64.60 35.40

Maquoit Bay site weight of sample (grams) weight of sand weight of silt/clay % Clay/Silt (<0.06mm) % Sand 1.00 31.41 0.04 31.37 99.87 0.13 2.00 32.74 0.16 32.58 99.51 0.49 3.00 33.13 0.03 33.10 99.91 0.09

83