Temporal Trends and Biogeochemical Controls on Methane And

Home , Haro Strait, Saanich Inlet

TEMPORAL TRENDS AND BIOGEOCHEMICAL CONTROLS ON METHANE AND

NITROUS-OXIDE DISTRIBUTIONS IN COASTAL WATERS OF THE SUBARCTIC

PACIFIC OCEAN

David Capelle

B.Sc., The University of Winnipeg, 2010

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF

THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES

(Oceanography)

THE UNIVERSITY OF BRITISH COLUMBIA

(Vancouver)

August 2016

Abstract

This PhD thesis examines the marine cycling of the greenhouse gases methane (CH4) and nitrous-oxide (N2O) in coastal British Columbia waters. The primary objectives of the work were to increase spatial and temporal data availability in an under-sampled coastal region, and to examine the processes responsible for CH4 and N2O distributions, and their sensitivity to changing environmental conditions (e.g. O2-availability). Using a novel high-throughput analytical system, based on purge and trap gas chromatography-mass spectrometry (GC-MS), we measured a 6 year time-series of monthly water column CH4 and N2O profiles from Saanich

Inlet, British Columbia, as well as three years of water column profiles and surface measurements along the West coast of Vancouver Island (WCVI). The physical and biological processes responsible for the observed CH4 and N2O distributions were investigated using relationships with ancillary physical data and biological data, including recently available metagenomic information. The results presented in this thesis document a dominant role for O2 concentrations in driving spatial and temporal variability in CH4 and N2O concentrations over a range of scales. In Saanich Inlet, the seasonal cycle of anoxia and deep water renewal exerts a primary control on water column N2O and CH4 accumulation, with additional likely contributions from sedimentary processes and in situ cycling of various nitrogen species and methylated compounds in the upper water column. In both Saanich Inlet and the WCVI, inter- annual variability and longer-term trends are associated with changes in upwelling intensity and

El Nino events, and these changes are set against a background of declining O2 concentrations across the Subarctic Pacific. Results from our work suggest that coastal CH4 and N2O concentrations may be responding significantly to these long-term declines in O2 levels, with significant implications for regional sea-air fluxes of climate-active trace gases. ii

Preface

A version of Chapter 2 has been published: Capelle, D.W., Dacey, J.W., Tortell, P.D.,

2015. An automated, high through-put method for accurate and precise measurements of dissolved nitrous-oxide and methane concentrations in natural waters. Limnol. Oceanogr.

Methods 13(7), 345-355. doi:10.1002/lom3.10029. I received assistance with the design and construction of the instrument described in Chapter 2 from Philippe Tortell and John Dacey. I performed most of the software programming and testing for accuracy, precision, and sensitivity.

The calibration of our instrument against conventional detectors was part of our participation in the Scientific Committee on Ocean Research Working Group #143 on dissolved methane and nitrous oxide measurements. I was responsible for writing and producing the manuscript and figures, and received edits and suggestions from Philippe Tortell and John Dacey.

A version of Chapter 3 has been published: Capelle, D.W., Tortell, P.D., 2016. Factors controlling methane and nitrous-oxide variability in the southern British Columbia coastal upwelling system. Mar. Chem. 179, 56–67. doi:10.1016/j.marchem.2016.01.011. I collected and measured all of the dissolved gas samples. I also assisted in the collection of additional samples, which were measured by IOS scientists as part of the LaPerouse Sampling Program. I performed all of the data analysis, and produced the figures and text, with feedback from Philippe Tortell.

The data presented in Chapters 4, 5, and 6 were collected as part of the Saanich Inlet time-series program, which is collaborative effort between the Tortell, Hallam, and Crowe laboratories at the University of British Columbia. Most of the samples were collected and analyzed by grad students and technicians from the Hallam Lab and EOAS Department. The time-series data were compiled by Alyse Hawley and Mónica Torres Beltrán. I measured nearly iii

all of the dissolved N2O and CH4 samples from Saanich Inlet that were collected from 2009-

2015. The Saanich Inlet archive data presented in Chapter 5 was compiled by Frank Whitney. I was the primary author of these chapters and produced all figures except for Figure 4.8. I received manuscript revision suggestions for Chapter 4 from Philippe Tortell, Steven Hallam, and Alyse Hawley, for Chapter 5 from Philippe Tortell, Frank Whitney, and Steven Hallam, and for Chapter 6 from Philippe Tortell.

I received assistance in performing incubation experiments with Saanich Inlet water from

Mónica Torres Beltrán, Sean Crowe, Arne Sturm, and Céline Michiels.

Table of contents

Abstract ...... ii

Preface ...... iii

Table of contents ...... v

List of tables...... x

List of figures ...... xi

List of abbreviations ...... xix

Acknowledgements ...... xx

Chapter 1: Introduction ...... 1

1.1 Motivation ...... 1

1.1.1 Coastal regions and estuaries ...... 2

1.2 N2O cycling ...... 3

1.3 CH4 cycling ...... 6

1.4 Thesis objectives ...... 7

1.4.1 Chapter 2 - An automated, high through-put method for accurate and precise

measurements of dissolved nitrous-oxide and methane concentrations in natural waters ...... 7

1.4.2 Chapter 3 - Factors controlling methane and nitrous-oxide variability in the southern

British Columbia coastal upwelling system ...... 8

1.4.3 Chapter 4 - A multi-year time-series of N2O dynamics in Saanich Inlet, British

Columbia, a seasonally anoxic fjord ...... 9

1.4.4 Chapter 5 - Influence of expanding hypoxia on methane and nitrous oxide

concentrations in coastal waters of the Subarctic Pacific ...... 10

1.4.5 Chapter 6 - A multi-year time-series of CH4 dynamics in Saanich Inlet, British

Columbia, a seasonally anoxic fjord ...... 10

1.4.6 Conclusion ...... 11

Chapter 2: An automated, high through-put method for accurate and precise measurements of dissolved nitrous-oxide and methane concentrations in natural waters. ..12

2.1 Summary ...... 12

2.2 Introduction ...... 13

2.3 Materials and methods: ...... 16

2.4 Assessment ...... 24

2.4.1 Recovery efficiency ...... 24

2.4.2 Optimization of purge time ...... 26

2.4.3 Accuracy of gas mixer ...... 26

2.4.4 Accuracy of liquid samples ...... 27

2.4.5 Precision ...... 29

2.4.6 Detection limits ...... 29

2.4.7 System performance for the analysis of field collected samples ...... 30

2.5 Discussion ...... 32

2.6 Comments and recommendations ...... 33

Chapter 3: Factors controlling methane and nitrous-oxide variability in the southern

British Columbia coastal upwelling system ...... 34

3.1 Summary ...... 34

3.2 Introduction ...... 35

3.3 Methods...... 40 vi

3.3.1 Study site ...... 40

3.3.2 Field sampling and gas analysis ...... 42

3.3.3 Upwelling intensity ...... 43

3.3.4 River discharge ...... 44

3.3.5 Sea-air fluxes ...... 44

3.3.6 Calculation of in situ N2O production ...... 45

3.4 Results ...... 47

3.4.1 Hydrographic conditions - upwelling intensity, riverine inputs, and O2

concentrations ...... 47

3.4.2 Surface water concentrations and sea-air CH4 and N2O fluxes ...... 48

3.4.3 Depth-dependent N2O and CH4 concentrations ...... 50

3.4.3.1 Along-shelf variability ...... 50

3.4.3.2 Cross-shelf variability ...... 54

3.5 Discussion ...... 55

3.5.1 Sources of CH4 ...... 58

3.5.2 Sources of N2O ...... 58

3.5.3 Biological production of N2O ...... 61

3.5.4 Dominant transport mechanisms for N2O and CH4 ...... 64

3.5.5 Sea-air fluxes ...... 65

3.6 Conclusion ...... 68

Chapter 4: A multi-year time-series of N2O dynamics in a seasonally anoxic fjord: Saanich

Inlet, British Columbia...... 70

4.1 Summary ...... 70 vii

4.2 Introduction ...... 71

4.3 Methods...... 76

4.4 Results and discussion ...... 80

4.4.1 Overview ...... 80

4.4.2 Seasonal variability ...... 81

4.4.2.1 Surface waters ...... 81

4.4.2.2 Mean seasonal cycle in the oxycline ...... 85

4.4.2.3 Mean seasonal cycle in the deep basin ...... 87

4.4.3 Depth resolved variability ...... 89

4.4.4 Relationship between O2 and N2O-cycling ...... 92

4.4.5 Interannual variability ...... 95

4.5 Conclusion ...... 100

Chapter 5: Influence of expanding hypoxia on nitrous oxide concentrations in the coastal

Subarctic Pacific...... 102

5.1 Summary ...... 102

5.2 Introduction ...... 102

5.3 Methods...... 104

5.4 Results and discussion ...... 107

5.5 Conclusion ...... 113

Chapter 6: A multi-year time-series of CH4 dynamics in Saanich Inlet, British Columbia, a seasonally anoxic fjord ...... 115

6.1 Summary ...... 115

6.2 Introduction ...... 116 viii

6.3 Methods...... 120

6.3.1 Study area...... 120

6.3.2 Sampling and analysis...... 122

6.4 Results and discussion ...... 124

6.4.1 Overview ...... 124

6.4.2 Mean seasonal cycle ...... 124

6.4.3 Factors driving CH4 cycling in the upper water column...... 127

6.4.4 Factors affecting mid-depth CH4 minimum ...... 134

6.4.5 Factors affecting CH4 in the deep basin ...... 136

6.4.6 Interannual variability ...... 137

6.4.7 Decadal-scale trends ...... 140

6.5 Conclusion ...... 142

Chapter 7: Conclusion ...... 144

7.1 Research implications ...... 144

7.2 Future work ...... 148

References ...... 152

Appendix ...... 178

List of tables

Table 2.1 Results of air-equilibrated water test for accuracy. CH4 equilibrium calculated using

Wiesenburg & Guinasso (1979) assuming atmospheric CH4 concentration = 1.87ppm CH4

(measured in 2012; Cunnold et al., 2002). N2O equilibrium concentration calculated using Weiss and Price (1980) assuming atmospheric N2O concentration = 325 ppb (measured in 2012; Prinn et al., 1990). Concentrations of CH4 and N2O in laboratory air were measured using the PT-

GCMS, and were not significantly different from atmospheric values (1.84 +/- 0.18 ppm CH4 and 323 +/- 60 ppb N2O)...... 28

Table 2.2 Results of intercalibration test show no significant differences between measurements of replicate samples made using PT-GCMS and ECD/FID. Values in the table represent the mean of all replicates ± 1 standard deviation. The total number of participating laboratories was seven for CH4 and ten for N2O...... 29

Table 3.1 Mean N2O and CH4 fluxes, excess concentrations above atmospheric equilibrium

(ΔN2O and ΔCH4), mixed layer depths, and time-weighted piston velocities (kw) for each cruise.

The number of stations used to calculate fluxes for each cruise is denoted by n. See methods for details of weighted piston velocity calculations...... 50

List of figures

+ Figure 1.1. Summary of processes responsible for marine N2O cycling. NH4 -oxidation can be performed by bacteria and archaea, and is dominant N2O source at O2 concentrations above ~ 10

µM. Denitrification is restricted to hypoxic to anoxic water, and can be a net source of N2O in hypoxic water, but is a net sink under suboxia (<5µM O2) and anoxia (Bange, 2008)...... 4

Figure 2.1 Schematic of purge and trap GCMS system. See text for details of the various components and analysis steps. Capital letters denote the various system components referred to in the text. Dotted lines denote alternate flow paths...... 16

Figure 2.2 Photograph of auto-sampler showing the syringe pump (left), tubing, sample vials, and two 13-position selector valves. Samples vials are contained in a sealed plexiglass tray to prevent spillage of HgCl2-containing sample water...... 17

Figure 2.3 Sample chromatogram. For each sample and standard, the area under each peak is automatically determined by the GCMS Post-Run Analysis software package, and exported to a formatted data file using a custom LabVIEW program...... 21

Figure 2.4 Detector intensity vs. purge time five milliliters water sample replicates. Purging for three minutes yields the same amount of gas as purging for 10 min, regardless of sample concentration...... 25

Figure 2.5 Standard curves generated with and without gas mixer spanning a three-week period.

...... 27

Figure 2.6 Comparison of water column profiles from Saanich Inlet, BC measured by PT-GCMS and conventional methods. Lilley et al.’s (1982) June CH4 profile (FID) matches well with June

2011 profile measured by PTGCMS (left). N2O profiles taken in August 1977, measured by

ECD (Cohen, 1978) also show good agreement with a depth profile from August 2012, measured by PT-GCMS (right)...... 31

Figure 3.1 Map of the study area along the west coast of Vancouver Island (WCVI). Depth- resolved samples were collected from profile stations (black circles) during five cruises (Jun-

2012, Sep-2012, Jun-2013, Sep-2013, and Jun-2014) along the Coastal Transect and the cross- shelf LC Line Transect. During June 2014, 5 m samples were collected from a number of additional stations (black triangles). The locations of CH4-seeps (located using 12 kHz echo sounder data) are shown by black x’s (Vaughn J. Barrie, pers. comm.). The O2 concentrations at

800m from the World Ocean Atlas climatology (Garcia et al., 2009) are shown in the inset, with the 20µM O2 contour line shown, highlighting the subarctic North Pacific OMZ...... 41

Figure 3.2 Upwelling Index values from the WCVI study region (48 °N, 125 °W, panel a) and

Fraser River discharge values (panel b) between January 2012 and November 2014. Upwelling values are plotted with 1-day (grey line) and 14-day (black line) running means. White bars

(panel a) indicate mean upwelling index values during the 90-day period before each cruise.

Vertical dashed lines indicate approximate sample collection dates during the 5 cruises...... 47

Figure 3.3 Excess CH4 (panel a) and N2O (panel b) above equilibrium concentrations measured at 5 m depth during June, 2014...... 49

Figure 3.4 Distributions of salinity (panels a-e), N2O (panels f-j), CH4 (panels k-o), and O2

(panels p-t) from each cruise along coastal transect. The locations of discrete samples are shown by black dots...... 52

Figure 3.5 Relationship between mean salinity and mean CH4 concentrations in shallow waters

(less than 15 m depth) along the coastal transect (R = 0.61; p = 0.003; n= 22). Error bars denote

xii

±1 standard deviation from the mean of all available measurements between 0 – 15 m depth at each station...... 53

Figure 3.6 Mean pre-cruise upwelling index (90 day average; shown in panel a) and depth sections of N2O, CH4, and O2 along the cross-shelf LC Line Transect. N2O (panels b-f), CH4

(panels g-k), and O2 (panels l-p). Region with abundant CH4 seeps indicated by horizontal black lines in panels g-k (see Figure 3.1 for seep locations). The upper and lower isopycnals used to

- calculate along-isopycnal changes in N2O, NO3 , O2, and N2O-yields are indicated by black lines in panels b-p...... 56

Figure 3.7 Correlation between mean pre-cruise upwelling indices (90 day average) and average

CH4 (a) and N2O (b) concentrations in shelf waters shallower than 50 m. Average gas concentrations were derived from samples collected within the top 50 m at all on-shelf stations

(bottom depth less than 200 m), and error bars indicate ± 1 standard deviation. Both correlations are statistically significant (R> 0.95; p < 0.01; n=5)...... 57

- - Figure 3.8 Relationship between O2, N2O, and NO2 +NO3 across all samples for the 5 cruises.

The negative correlation indicates nitrification is the dominant source of N2O in our study

- - region. The absence of decreasing NO2 +NO3 or N2O under low O2 suggests that denitrification is not occurring at appreciable levels in the water column...... 59

Figure 3.9 Comparison of density-dependent profiles of CH4 and N2O at an on-shelf (LC04, black lines) and off-shelf (LC11, grey lines) station during September, 2012. The changes in O2,

- - NO2 +NO3 and N2O along isopycnals are ascribed to in situ nitrification during the transit of water masses onto the shelf. Figures from additional cruises are included in below...... 62

Figure 3.10 Relationship between mean O2 concentrations at LC04 (on shelf station) and N2O

- - yields from nitrification. N2O yields were derived from an analysis of N2O and NO2 +NO3 xiii

changes along isopycnals (see Figure 3.9 and methods for details). The negative relationship implies increased N2O yields under low O2 concentrations in our study area. Grey triangles represent mean values derived for each cruise (average of all points interpolated to 0.01 kg m-3 density intervals), with error bars representing ± 1 standard deviation. Small black diamonds represent the individual calculated points for each cruise...... 63

Figure 3.11 Schematic diagram showing CH4 and N2O sources, sinks, and physical transport processes along the WVCI under upwelling (panel a) and downwelling (panel b) conditions.

Thick blue arrows indicate water circulation, dashed blue lines represent isopycnals and wavy, black lines indicate diffusion gradients. Upwelling transports N2O-rich waters from the deep

N2O max off the shelf, and CH4 from seeps near the shelf break along isopycnals towards the

- - coast. During transport, water column nitrification contributes additional N2O and NO2 +NO3 , while CH4-oxidation mitigates on-shelf CH4 increases. Sedimentary fluxes also increase the water column inventory of CH4 and N2O in shelf waters. The higher on-shelf CH4 and N2O concentrations lead to enhanced sea-air flux of these gases. Under downwelling conditions

(panel b), surface waters near air-equilibrium concentrations in O2, CH4 and N2O are carried below the surface near the coast. Low surface primary productivity (due to limited nutrient supply) results in relatively low rates of water column N2O production from nitrification.

Subsurface concentrations gradually increase as water flows away from the coast due to supply from CH4 seeps and the N2O maximum near the shelf-break...... 67

Figure 4.1 Map of Saanich Inlet, BC, showing location of sampling (star, panel a), and location on southeast Vancouver Island (inset). Depth transect along thalweg of Saanich Inlet (panel b) shows elevated sill near mouth of inlet. Adapted from Anderson and Devol, (1973)...... 78

xiv

Figure 4.2 Mean annual cycle of dissolved oxygen (a), nitrate (b), nitrous-oxide (c), and hydrogen-sulfide (d) derived from monthly 2007-2014 water column measurements in Saanich

Inlet...... 84

Figure 4.3 Seasonal cycle of N2O sea-air flux estimates based on near-surface (10m) excess N2O concentrations and wind-speed data. Individual flux estimates are shown as small black dots, with derived mean monthly flux estimates shown as larger black circles...... 85

Figure 4.4 Average water column profiles of oxygen, nitrate and nitrous oxide during different seasonal periods. Mean values are shown by thick lines, with the shaded grey area indicating one standard deviation. Note the logarithmic scale of O2...... 91

Figure 4.5 Relationship between N2O and apparent oxygen utilization (AOU) derived from monthly measurements in Saanich Inlet (a). Grey circles show measurements where AOU<270

µM, hollow circles are for AOU>270 µM. Linear regression for AOU<270uM is shown by thick black line. Several N2O:AOU relationships from different ocean regions are plotted with the same y-intercept for reference. Panel b shows the mean N2O concentration (± one standard deviation) in 1 µM O2 bins between 0 and 20 µM O2, and in 5 µM bins between 20-300 µM O2.

Panel b inset shows a curve fit to the individual O2 vs N2O measurements for O2 concentrations

< 30 µM...... 93

Figure 4.6 Composite annual cycle of the slope of the N2O:O2 linear relationship for portions of the water column with O2 > 30 µM. Colored circles represent individual profiles, where the colour indicates the mean O2 concentration of the measurements from that profile and above 30

µM O2. Hollow circles, connected by the line, represent mean monthly values...... 96

Figure 4.7 Monthly time series measurements of oxygen (a), nitrate (b), nitrous-oxide (c), and hydrogen-sulfide (d) from 2007-2014. Note that CTD O2 collection began in 2009...... 97 xv

Figure 4.8 Bubble plot shows all available RPKM abundance values for nitrification and denitrification genes. Bubble colour indicates N2O concentration and bubble size indicates

RPKM value (gene abundance)...... 98

Figure 4.9 Mean concentrations (± one standard deviation) of oxygen (a), nitrate (b), nitrite (c), nitrous-oxide (d), and ammonium (d) below 180m in Saanich Inlet. Vertical dashed line indicate approximate times of deep basin renewal. Renewal times were derived based on density increases of >= 0.009 kg m-3 at 200m depth relative to the previous sampling date...... 100

Figure 5.1 Long term trends in mean water column (10 : 200 m) O2 (a) and temperature (b) in

Saanich Inlet from 1963 – 2015. Grey dots indicate archive data (1960-2007), and crosses indicate recent time-series data (2009-2015). Thick black lines show mean annual values.

Dashed lines indicate linear trends through the data. Both trends are statistically significant (p <

0.05)...... 107

Figure 5.2 Mean water column concentrations of O2 (a), NO3 (c), and N2O (d), and depth of hypoxic boundary (b) in Saanich Inlet from Jan 2009 – Dec 2015. Straight black line shows linear regression through the data. All trends are statistically significant (p < 0.05)...... 109

- Figure 5.3 Relationship between O2, NO3 , and N2O concentrations from the west coast of

Vancouver Island (WCVI) and Saanich Inlet. Small black dots represent discrete measurements from samples collected from 50-120m along the WCVI between Jun 2012 and Jun 2014. Large black diamonds represent mean ± 1 standard error from each of the five WCVI cruises. Black x’s show mean values from 50 – 120 m depth from each cruise in Saanich Inlet between 2009 and

2015...... 112

xvi

Figure 6.1 Map of Saanich Inlet, BC, showing location of sampling (star, panel a), and location on southeast Vancouver Island (inset). Depth transect along thalweg of Saanich Inlet (panel b) shows elevated sill near mouth of inlet. Adapted from Anderson and Devol, (1973)...... 121

Figure 6.2 Mean annual contour plot of O2 (panel a), NO3 (b), CH4 (c), and H2S (d). Each dot represents a single measurement, which have been condensed into a single calendar year to highlight seasonal variability. Panel e shows mean seasonal cycle of sea-air CH4 flux. Small black dots represent individual flux measurements, while mean monthly values are shown by the large black dots and thick black line...... 125

Figure 6.3 Mean monthly CH4 from upper 85m (panel a), fluorescence from upper 40m (panel b) from our time-series measurements. Sediment trap data from Timothy et al. (2003) showing annual cycle in total grams organic carbon (panel c) and g aluminum (d) recovered in sediment traps at 45m, 110m, and 150m depth...... 129

Figure 6.4 Relationship between mean monthly fluorescence in upper 40m and mean CH4 in upper 85m from 2008-2015...... 130

Figure 6.5 Depth profiles of transmissivity (panel a), O2 (b), and CH4 (c) from May 2013. The shaded patch highlights the transmissivity and CH4 minimum near the oxic-anoxic interface. . 135

Figure 6.6 Contour plots of O2 (panel a), NO3 (b), CH4 (c), and H2S (d) from 2008-2014. Black dots represent individual measurements...... 138

Figure 6.7 Time-series plot of mean CH4 in upper 85 m (a) and CH4 flux (b). Straight black line indicates line of best fit through the data...... 141

Figure A.1 N2O, NO2-+NO3- and O2 profiles from LC04 and LC11 for each cruise...... 178

Figure A.2 N2O, NO2-+NO3- and O2 profiles from LC04 and LC11 for each cruise...... 179

- - Figure A.3 N2O, NO2 +NO3 and O2 profiles from LC04 and LC11 for each cruise...... 180 xvii

- - Figure A.4 N2O, NO2 +NO3 and O2 profiles from LC04 and LC11 for each cruise...... 181

- - Figure A.5 N2O, NO2 +NO3 and O2 profiles from LC04 and LC11 for each cruise...... 182

xviii

List of abbreviations

AOU - Apparent oxygen utilization

CC - California current

CH4 - Methane

CUC - California undercurrent

H2S - Hydrogen Sulfide

N2 - Nitrogen

+ NH4 - Ammonium

NH3 - Ammonia

N2O - Nitrous-oxide

- NO2 - Nitrite

- NO3 - Nitrate

O2 - Oxygen

OMZ - Oxygen minimum zone

3- PO4 - Phosphate

PT-GCMS - Purge and trap gas-chromatography mass-spectrometer

SO4 - Sulfate

WCVI - West coast of Vancouver Island

VICC - Vancouver Island coastal current

xix

Acknowledgements

I wish to acknowledge my advisor, Philippe Tortell, for his encouragement and support throughout this project. I also want to thank my committee members, and grad students in the

TORNADO Lab for their constructive feedback. Thanks to John Dacey and Dave Jones for your help in building and programming the PT-GCMS. Thank you to the Hallam and Crowe Labs, especially to Alyse Hawley, Monica Torres Beltran, Celine Michiels and Steven Hallam for all your help in collecting samples, sharing data, and providing valuable feedback. Thanks to Frank

Whitney for your insightful suggestions and for sharing the Saanich Inlet archive data. Thank you to Chris Payne and Lora Pakhomova for helping to collect and analyze Saanich Inlet data, and to Doug Yelland, Marie Robert, Moira Galbraith, and all the scientists who I sailed with from the Institute of Ocean Sciences for your advice, and for helping with data collection. I wish to thank the Captains and crews of the R/V Strickland and CCGS J.P. Tully for accommodating me and assisting in sample collection. Thanks to the members of SCOR Working Group #143 for assisting in calibrating our PT-GCMS. Thanks to the National Science and Engineering Research

Council of Canada (NSERC) for financial support for this research. To my parents for encouraging me to pursue my passion. And finally, to my wife, Vicky, thank you for all your love and support, and for putting up with my prolonged absences from home.

Chapter 1: Introduction

1.1 Motivation

Methane (CH4) and N2O are the two most important greenhouse gases after water vapour and CO2. The atmospheric concentrations of these gases have fluctuated in accordance with CO2 over glacial-interglacial cycles of the last 700,000 years, and have increased significantly since the industrial revolution as a result of various anthropogenic activities (Loulergue et al., 2008;

Petit et al., 1999; Sowers et al., 2003; Wolff and Spahni, 2007). A number of natural and anthropogenic CH4 and N2O sources have been identified, but many unknowns remain (Altabet,

Francois, Murray, & Prell, 1995; L. A. Codispoti, Brandes, Christensen, & Devol, 2001; L. A.

Codispoti, 2010; Freing, Bange, & Wallace, 2012; Wolff & Spahni, 2007). The marine contribution to atmospheric CH4 and N2O is not well constrained, with most estimates failing to adequately represent emissions from coastal continental shelves, estuaries and near-shore upwelling environments, which are among the strongest marine sources of N2O and CH4 to the atmosphere (Bange, 2008; Freing et al., 2012; Naqvi et al., 2010; Reeburgh, 2007; Rehder et al.,

2002). The underlying processes driving production and consumption and of marine CH4 and

N2O are also poorly understood, as is the response of the CH4 and N2O cycles to environmental perturbations. These knowledge gaps limit our ability to predict climate-dependent feedbacks in marine emissions of CH4 and/or N2O to the atmosphere.

The purpose of this research is to better quantify CH4 and N2O distributions in coastal

British Columbia waters, investigate the underlying processes controlling production and consumption of these gases, and examine their potential sensitivity to various environmental factors. The motivation for this research is based on a body of research that will be summarized 1

in the following paragraphs. First, I outline the importance of coastal and estuarine environments in the marine N2O and CH4 budgets, and highlight the need for frequent and high-resolution sampling in these regions. Next, I summarize the biological sources and sinks of N2O and CH4 in the marine water column, emphasizing some of the key unknowns. I conclude with a statement of the specific objectives of my research. Additional introductory material about the oceanographic setting for each sampling region is presented in individual data chapters.

1.1.1 Coastal regions and estuaries

Estuaries and coastal regions are considered the strongest marine sources of CH4 and

N2O to the atmosphere – accounting for an estimated 60% and 75% of the total marine N2O and

CH4 flux, respectively. It is important to note, however, that these estimates are based on relatively sparse data, and thus subject to significant uncertainty (Bange et al., 1996, 1994).

Furthermore, N2O and CH4 distributions in these regions are susceptible to small-scale variability due to riverine sources, upwelling, and coastal topography (Lueker et al., 2003;

Nevison et al., 2004; Rehder et al., 2002). The intermittent nature of these inputs makes them challenging to resolve with sporadic sampling efforts. This has prompted several researchers to point out the need for additional high-frequency measurements of coastal and estuarine CH4 and

N2O (Bange et al., 1994; Kohfeld et al., 2005; Middelburg et al., 2002), and the development of sustained and systematic time-series measurements. Such time-series programs have the potential to generate very large numbers of samples, creating a significant analytical workload.

The methods used in many laboratories for CH4 and N2O measurements require labour-intensive and time-consuming analysis of discrete bottle samples. This potentially limits the number of samples that can be analyzed over a reasonable time-frame. High through put, automated 2

methods are thus desirable in increasing the analytical capacity for CH4 / N2O measurements for long-term time series programs.

1.2 N2O cycling

The dominant microbial processes responsible for N2O cycling are nitrification and denitrification (Figure 1.1), and both these processes are sensitive to O2-availability. Water column N2O is produced mainly as a by-product during nitrification, the sequential oxidation of ammonia (NH3) to nitrite (NO2) and nitrate (NO3). Nitrifiers are chemo-autotrophic, using the energy derived from NH3 oxidation to support C-fixation. Light has been shown to inhibit nitrification (Hooper and Terry, 1973; Merbt et al., 2012), which suggests that most nitrification occurs below the euphotic zone. However, there is evidence for appreciable rates of nitrification in the euphotic zone as well (Grundle and Juniper, 2011). The N2O yield during bacterial nitrification increases under O2-limitation, and nitrification ceases in the absence of O2 (Goreau et al., 1980). Although nitrification was traditionally believed to be performed exclusively by bacteria, recent work has shown that the first step of nitrification (NH3-oxidation) can be performed by a number of archaeal groups (Francis et al., 2005), which are now believed to be the dominant nitrifying organisms (and potentially N2O producers) in the marine environment

(Martens-Habbena et al., 2009). At present, the effect of changing O2-availability on N2O-yields during archaeal nitrification is not known. N2O is also produced as an intermediate product

- during incomplete denitrification, the stepwise reduction of NO3 to N2 by heterotrophic bacteria

(Figure 1.1). This process is believed to contribute to the high N2O concentrations under low-O2 conditions (< 20 µM) in the marine environment (Naqvi et al., 2010). Under suboxic conditions

(O2 < 5uM), N2O is reduced to N2 during the last step of denitrification. This represents the only 3

+ Figure 1.1. Summary of processes responsible for marine N2O cycling. NH4 -oxidation can be performed by bacteria and archaea, and is dominant N2O source at O2 concentrations above ~ 10 µM. Denitrification is restricted to hypoxic to anoxic water, and can be a net source of N2O in hypoxic water, but is a net sink under suboxia (<5µM O2) and anoxia (Bange, 2008).

known biological sink of N2O in the oceans (Naqvi et al., 2010).

Several other N2O production pathways have been the subject of recent research focus.

- For example, nitrifier-denitrification, the process by which nitrifiers reduce NO2 to N2O at low oxygen concentrations, has been found to produce N2O at low oxygen concentrations in cultures of Nitrosopumilus maritimus and Nitrosomonas europaea (Frame and Casciotti, 2010; Poth and

Focht, 1985). Another potential N2O source is dissimilatory reduction of nitrate to ammonia

(DNRA), which was recently measured for the first time in a marine water column (Lam et al.,

2009). This process differs from denitrification primarily because the end product is a

+ bioavailable form of nitrogen (NH4 ), whereas denitrification produces the non-bioavailable N2

molecule (Koike and Hattori, 1978). Finally, N2O can be produced by heterotrophic nitrification, the oxidation of NH3 to NO3 by organisms that derive their carbon from organic sources

(Anderson et al., 1993). The relative contributions of these processes to the marine N2O budget is not yet well constrained, but they likely play a small role compared to nitrification and denitrification.

The general trend of high N2O production at low O2 concentrations has led to the suggestion that marine N2O concentrations will increase due to marine de-oxygenation resulting from climate change (Codispoti, 2010). Climate models predict that as the atmosphere warms, marine O2-concentrations will decline due to a combination of reduced supply and ventilation of subsurface waters (Falkowski et al., 2011; Keeling et al., 2010; Stramma et al., 2010). A 50-year time series of dissolved O2 from the subarctic North Pacific Ocean shows that de-oxygenation is

-1 -1 already occurring, with subsurface O2 declining at rates of 0.39-0.70 µmol kg y , and the depth of the 60µM O2 boundary shoaling from 400m to 300m between 1953 and 2007 (Whitney et al.,

2007). Similar O2 loss has also been documented across wide swaths of coastal waters in the eastern Subarctic Pacific, including the continental shelf region of British Columbia, Canada.

Continued O2 loss will increase the total volume of water available for rapid N2O-production, and bring high N2O water masses closer to the surface, leading to a potentially significant increase in marine N2O flux. However, our limited understanding of N2O cycling and its O2 sensitivity limit our ability to predict these increases (Bange, 2008; Codispoti, 2010; Codispoti et al., 2001; Naqvi et al., 2010).

1.3 CH4 cycling

Marine CH4 is primarily from CO2 and acetate reduction by methanogenic Archaea (Judd et al., 2002; Reeburgh, 2007). This process is generally confined to anoxic sediments where sulfate reduction has nearly proceeded to completion. Low sulfate concentrations are important for methanogenesis, because sulfate reducers out-compete methanogens for H2, a key substrate in

CO2 reduction (Reeburgh, 2007). Much of the anaerobically-produced CH4 is consumed within the upper portions of the sediment by either anaerobic or aerobic methanotrophs (i.e. CH4 oxidizing bacteria), but some CH4 may escape into the water column, particularly in regions with high organic carbon fluxes to the sediment surface (Ward et al., 1989; Zaikova et al., 2010). CH4 produced in freshwater sediments may also be transported into the marine water column near rivers, as evidenced by CH4-salinity relationships, and by high CH4 concentrations observed near rivers (Bange et al., 1998; Musenze and Grinham, 2013).

In addition to the well-known sedimentary sources of CH4, there is a shallow secondary

CH4 maximum in most ocean regions (Holmes et al., 2000; Reeburgh, 2007). Because CH4 production is not believed to occur in well-oxygenated waters, the appearance of a near-surface

CH4 maximum presents a so-called ‘methane paradox’. There is strong evidence that much of this near-surface CH4 is produced inside sinking particles and zooplankton intestinal tracts (De

Angelis and Lee, 1994; Marty, 1993; Oremland, 1979; Sansone et al., 2001; van der Maarel et al., 1999). Recent evidence suggests at least a portion of this CH4 may also come from the bacterially-mediated cleavage of naturally occurring methylated compounds, such as methylphosphonate (MPn) and dimethylsufide (DMS) (Damm et al., 2010; Florez-Leiva et al.,

2013; Karl et al., 2008).

1.4 Thesis objectives

This PhD research addresses key limitations in understanding marine CH4 and N2O cycling and budgets, namely the lack of adequate spatial and temporal data coverage from coastal regions, and limited information on the relationship between hydrographic variability and gas concentrations. We employ repeated sampling from multiple seasons in order to determine the natural variability in CH4 and N2O distributions in coastal BC waters. This is essential for identifying significant long-term changes in marine CH4 and N2O concentrations, which may already be occurring due to ongoing changes in the marine environment. Coupling these large data sets with ancillary measurements from our study region enables us to evaluate the physical, biological, and chemical factors driving this variability in a natural ecosystem. These findings can be used to identify oceanographic variables that control CH4 and N2O distributions, and to predict how gas concentrations may respond to specific ecosystem perturbations. We briefly outline the specific objectives and approaches used in each chapter of the thesis, below.

1.4.1 Chapter 2 - An automated, high through-put method for accurate and precise measurements of dissolved nitrous-oxide and methane concentrations in natural waters

This chapter describes an automated system to facilitate the measurement of large numbers of dissolved CH4 and N2O samples. Such high through put analysis enables the efficient processing of many hundreds of samples, and thus help to address the limited spatial and temporal data availability from coastal marine ecosystems. We constructed an auto-sampler based on purge-and-trap gas extraction for the analysis of dissolved CH4 and N2O in discrete water samples using commercial gas-chromatograph mass-spectrometer. The automated system is capable of producing calibration curves and blanks, and measuring a batch of 25 discrete 7

samples with minimal (~1 hr.) user input. This represents a significant improvement in sampling handling efficiency and precision over conventional head-space or purge-and-trap analysis based on manual sampling handling. Through careful calibration against NOAA certified gas standards, and participation in an international inter-calibration exercise, we demonstrated that our system achieves similar degrees of accuracy and precision to conventional methods. This chapter presents a preliminary set of field observations to demonstrate the utility of the new system to document spatial and temporal patterns in CH4 and N2O concentrations in coastal marine waters.

1.4.2 Chapter 3 - Factors controlling methane and nitrous-oxide variability in the southern British Columbia coastal upwelling system

We collected and analyzed ~1,500 discrete CH4 and N2O samples collected during five separate cruises over a three-year period from the coastal upwelling shelf waters of Vancouver

Island, British Columbia. We examined the influence of local currents, upwelling intensity, O2 availability, freshwater input, and sedimentary sources on N2O and CH4 distributions and sea-air fluxes. We determined that CH4 accumulation in the water column was primarily derived from sedimentary sources, particular CH4 seeps along the shelf break and coastal shelf waters. High concentrations of N2O were derived from the N2O maximum in the off-shelf oxygen minimum zone, as well as water column production by nitrification. We showed that upwelling enhanced the shoreward and vertical transport of both N2O and CH4, while also stimulating additional N2O production in the water column and/or sediments. We also estimated N2O-yields of nitrification, and demonstrated an increase in N2O yields under low O2 conditions.

1.4.3 Chapter 4 - A multi-year time-series of N2O dynamics in Saanich Inlet, British

Columbia, a seasonally anoxic fjord

We analyzed over 3,000 discrete N2O samples collected over a nine-year period from a seasonally anoxic fjord, Saanich Inlet, British Columbia. We were able to identify regular seasonal variability in N2O concentrations in different depth intervals and sea-air fluxes. We demonstrated that maximum N2O concentrations and sea-air fluxes occurred during periods when low O2 deep waters were transported into near surface waters. Deep basin renewal events supplied N2O to the deep basin waters, rather than stimulating in situ N2O production. We interpreted our measurements in the context of ancillary data and previously published research, including microbial metagenomic data, to estimate the relative importance of different potential

N2O sources and sinks. We showed that N2O sources and sinks were strongly influenced by O2 availability, with N2O produced by nitrification and partial denitrification in O2-depleted waters

(O2<20µM), whereas net N2O consumption by denitrification was initiated below ~10µM O2.

We observed seasonal differences in N2O:O2 ratios, suggesting seasonal differences in N2O yields from nitrification and/or net N2O production by denitrification. Metagenomics data indicated the gene for N2O reduction (nosZ) was widespread in Saanich Inlet, including well- oxygenated surface waters, potentially maintaining low N2O concentrations relative to open ocean OMZ waters. Metagenomics also suggest the denitrification pathway was mediated by a distributed pathway shared among various organisms, including the widely distributed gamma- proteobacteria SUP05. We also observed anomalous N2O concentrations during weak deep basin renewals, associated with El Niño events.

1.4.4 Chapter 5 - Influence of expanding hypoxia on methane and nitrous oxide concentrations in coastal waters of the Subarctic Pacific

There is concern that ongoing ocean de-oxygenation and shoaling of O2 minimum zones will lead to increased production and sea-air flux of N2O. We analyzed a ~60 year archive of O2, temperature, and salinity data from Saanich inlet, combined with our time-series data since 2009, to test for significant long-term trends in O2 and N2O in Saanich Inlet. We identified significant declines in O2 throughout the water column of Saanich Inlet dating back to the 1950s. Our time- series observations since 2009 demonstrate that this trend has accelerated in recent years, and is accompanied by a rapid shoaling of the hypoxic and suboxic zone. This is linked with significantly increasing N2O concentrations in the upper 120m and sea-air fluxes, which are likely related to enhanced N2O production in shallow waters due to higher N2O yields and partial denitrification. These results provide an important benchmark for future studies examining continued temporal trends in biogeochemical cycles in Saanich Inlet and other low oxygen coastal systems.

1.4.5 Chapter 6 - A multi-year time-series of CH4 dynamics in Saanich Inlet, British

Columbia, a seasonally anoxic fjord

We analyzed over 3,000 discrete CH4 samples collected over a nine-year period from a seasonally anoxic fjord, Saanich Inlet, British Columbia. We were able to identify regular seasonal variability in CH4 concentrations in different depth intervals and sea-air fluxes, and interpreted our measurements in the context of ancillary data and previously published research to estimate the relative importance of potential CH4 sources and sinks. Our observations suggest that CH4 is primarily derived from sedimentary sources in the deep basin, and likely from 10

multiple sources in the upper 85m, likely including sediment disturbance, transport from the deep basin, cleavage of methylated compounds, and production in anoxic microenvironments in particles. Methane consumption in the mid-depth CH4-minimum was enhanced during periods of high particle flux, potentially implicating methanotrophy associated with sediment resuspension.

We also observed consistent CH4 anomalies associated with El Nino events, both in the deep basin associated with weak/absent renewal events, and in the upper 85m associated with regionally elevated CH4 concentrations. We also observed significantly declining CH4 concentrations in the upper 85m and reduced sea-air fluxes. This may reflect reduced bioturbation of sediments by O2-depended fish and benthic organisms.

1.4.6 Conclusion

We conclude the dissertation in Chapter 7 with a discussion of the overall implications and significance of this research. This chapter also suggests some directions for future research, building on the results presented in this thesis.

Chapter 2: An automated, high through-put method for accurate and precise measurements of dissolved nitrous-oxide and methane concentrations in natural waters.

2.1 Summary

We describe a technique for measuring dissolved CH4 and N2O concentrations from discrete water samples using an automated purge and trap gas extraction system, coupled with a gas chromatograph-mass spectrometer (PT-GCMS). The automated system measures blanks, standards, and 25 samples in less than five hours with only ~30 minutes of operator involvement.

Rigorous testing of the PT-GCMS demonstrates sensitivity, accuracy and precision that is comparable or better than conventional methods for CH4 and N2O analysis. Measured concentrations of CH4 and N2O in air-equilibrated water samples showed good agreement with expected values derived from solubility calculations, and results of a multi-laboratory inter- calibration exercise showed that our measurements agree with those made using conventional methods. Precision of replicate water samples is 3.3% for CH4 and 3.0% for N2O. Detection limits are well below the expected concentrations in most natural waters with a 5 mL sample, and can be lowered substantially by analyzing a larger sample volume. To demonstrate the utility of the method, we present depth profiles of CH4 and N2O from Saanich Inlet, a coastal anoxic fjord in British Columbia. The Saanich Inlet water column exhibits rapid changes in CH4 and N2O across depth-dependent and seasonally variable redox conditions. Our high through- put, automated method facilitates the measurement of aqueous N2O and CH4 concentrations, and will thus help to improve our understanding of the natural cycling of these climate-active trace

gases.

2.2 Introduction

Methane and nitrous-oxide are potent greenhouse gases accounting for 17% and 6 %, respectively, of the total radiative forcing of well-mixed greenhouse gases in the atmosphere

(IPCC, 2013). The atmospheric concentrations of these gases have co-varied with CO2 over glacial-interglacial cycles, and, over the past two centuries, have risen above the highest values observed in the ice core record (Sowers et al., 2003; Wolff and Spahni, 2007). Most of the ocean surface is supersaturated in CH4 and N2O, making the oceans a net source of these gases to the atmosphere (Bange, 2008; Bange et al., 1994; Freing et al., 2012; Karl et al., 2008; Seitzinger et al., 2000). However, CH4 and N2O concentrations in marine surface waters vary significantly over a range of spatial and temporal scales, so that accurate estimates of the mean annual air-sea fluxes of these gases requires frequent and long-term sampling from numerous oceanic regions.

Furthermore, our understanding of the processes involved in CH4 and N2O cycling and their sensitivity to various environmental factors are not fully-resolved. This limits our understanding of how marine CH4 and N2O fluxes will respond to on-going oceanic changes, including progressive loss of oxygen from the ocean interior (Stramma et al., 2008; Whitney et al., 2007), nutrient loading, and warming (Codispoti, 2010; Freing et al., 2012; Naqvi et al., 2010, 2000;

Shakhova et al., 2010; Whitney et al., 2007). In order to address this uncertainty, there has been a recent increase in the number of time-series measurements of marine CH4 and N2O concentrations, including work at the following monitoring sites; Hawaiian Ocean Time Series

(HOTS) station ALOHA (Karl and Lukas, 1996), the Candolim time series (CaTS) off Goa,

India (Naqvi et al., 2009), Line P in the North Pacific (Tortell, unpublished), Boknis Eck in the 13

southwest Baltic Sea (Bange et al., 2010), Saanich Inlet in coastal British Columbia (Zaikova et al., 2010) and COPAS off the coast of Conception, Chile (Farias et al., 2009).

At present, most analytical measurements of CH4 and N2O in natural waters are conducted using time-consuming and laborious analysis of discrete samples. These measurements require two primary steps; gas extraction and gas detection. Gas extraction is typically achieved using either a headspace equilibration technique (Careri et al., 1999; Elkins,

1980) or a purge-and-trap (PT) approach, where gas is stripped out of a water sample, concentrated onto a trap and then rapidly desorbed for detection (Swinnerton and Linnenbom,

1967). While commercially available systems exist for automating the gas extraction step, these systems are costly, and many laboratories continue to use manual processing of individual samples and calibration standards (Karl et al., 2008; Schmitz et al., 2012; Walter, Bange,

Breitenbach, & Wallace, 2006; Yoshida, Inoue, Watanabe, Suzuki, & Noriki, 2011; Zhan &

Chen, 2009).

Once extracted from the liquid phase, CH4 and N2O are typically isolated using gas- chromatography, and measured using different detectors (often using separate sample vials).

Traditionally, most laboratories have used flame-ionization detectors (FID) for CH4 and electron capture detectors (ECD) or thermal-conductivity detectors (TCD) for N2O. Cavity ring-down systems, using infrared absorption, have also gained popularity in recent years, but these systems still require separate detectors for CH4 and N2O measurements(O’Keefe and Deacon, 1988;

Yvon-Lewis et al., 2011). By comparison, mass spectrometry has the potential for simultaneous analysis of many different gas analytes in a single sample, and this technique has become increasingly common as a tool for environmental gas analysis (Santos and Galceran, 2003). For example, Ekeberg, Ogner, Fongen, and Joner (2004) have used gas chromatography – mass 14

spectrometry (GCMS) to measure CH4 and N2O from air samples, while Punshon & Moore

(2004b) used GCMS for N2O analysis in water samples. To our knowledge, however, none of the major research laboratories investigating the marine distributions of CH4 and N2O use GCMS for simultaneous analysis of these gases in seawater samples.

Here we present a new method for the analysis of N2O and CH4 in aqueous samples using an automated purge and trap gas extraction system coupled with GCMS (PT-GCMS). Our goal was to develop a rapid and fully automated system for high precision measurements of these gases from a single vial, in order to improve efficiency and minimize labour in the analysis, while reducing the potential for human-error associated with manual sample handling. Our purge and trap auto-sampling system was built using commercially available and relatively inexpensive components to automatically extract gases from up to 25 aqueous samples in a single batch for analysis on a quadrupole mass spectrometer. The system includes automated routines for calibration and blank determinations, and requires almost no operator involvement, beyond a minimal set up for each batch of samples. We demonstrate that our automated method achieves analytical precision and accuracy comparable to conventional methods, and we provide an example of high quality CH4 and N2O data from Saanich Inlet, a seasonally anoxic coastal fjord. Our automated sample processing method could be easily adapted to other detectors and applied to a wide range of environmental systems, including lakes, rivers, wetlands and groundwater.

2.3 Materials and methods:

Discrete seawater samples are obtained from depth profiles using GoFLO or Niskin bottles. Sub-samples for gas analysis are immediately transferred to 60 or 160mL glass vials

(Wheaton, 20mm cap size, PN: 223746) via a thick-walled silicon tube, overfilling the sample bottles three full volumes, while being careful not to introduce bubbles. The sample water is then spiked with 100uL of saturated HgCl2, and crimp-sealed with butyl-rubber stoppers and aluminum caps (Wheaton PN: 224193-01 and 224100-174, respectively). Samples are then

Figure 2.2 Photograph of auto-sampler showing the syringe pump (left), tubing, sample vials, and two 13- position selector valves. Samples vials are contained in a sealed plexiglass tray to prevent spillage of HgCl2- containing sample water.

stored in the dark at 4°C until they can be analyzed. Samples can be stored for at least 45 days in these conditions without showing signs of degradation (Elkins, 1980), but should be run as soon as possible to minimize the potential for storage effects.

Figure 2.1 shows a schematic representation of our analytical system. The main components of the system are: i) the purge chamber, a sealed glass tube where sample water is bubbled with ultra-high purity (UHP) helium to strip the dissolved gases from solution (Figure

2.1, component C); ii) the trap, a U-shaped, 30 cm L x 1/8” O.D. stainless steel tube packed with

Porapak Q (Sigma-Aldrich, 80-100 mesh) that binds CH4 and N2O when cooled with liquid

nitrogen, and releases these gases when heated (Figure 2.1, component E), and iii) the GCMS

(Shimadzu, QP-2010), which uses gas chromatography to separate CH4 from N2O, and a mass- spectrometer (quadrupole, secondary electron multiplier with conversion dynode) to detect them.

All of the components of the PT system are controlled by a computer running LabVIEW 2010

(National Instruments). We designed a program that controls the components sequentially so that, once initiated, the batch of samples is analyzed without further human input. The program also controls the interface of the PT sampling system and the GC-MS detector, and automates the collection and analysis of detector data.

Up to 25 samples can be connected to the PT system by inserting two needles (Air-Tite,

22 gauge x 3” long) through the septum of each vial (Figure 2.2). The two needles are connected to 1/16” PEEK tubing by high pressure, zero-dead volume adapter fittings (VICI PN:

ZRU1.5TFPK). This tubing is plumbed into 13-position, 28 port electronic valves (Figure 2.1, component A –VICI PN: C35ZT-41813 EMT), with two ports for each sample, and a common inlet and common outlet. The common outlet on the valve is plumbed into a programmable electronic syringe pump (Norgren-Kloehn, PN: 55222, Figure 2.1, component B), which is used to transfer liquid from the sample vials to the purge chamber and waste container. The common inlet on the valve is connected to a UHP helium supply via a solenoid valve. The helium flow is used to maintain positive pressure in the sample vial during liquid extraction. We use two of these 13 position valves in series to obtain a total of 25 sample positions (one of the sample positions is used to connect the two valves). During automated sample processing, the electronic valve cycles through each position sequentially, effectively connecting one sample vial to the system at a time, while leaving the other sample vials isolated from the ambient air and from each other. 18

All liquid handling in our system is accomplished using the syringe pump (Figure 2.1, component B). This device consists of a gas-tight syringe connected to a stepper motor for precise control of liquid removal or dispensing. The syringe pump also contains a built-in electronic valve so that water can be picked up / dispensed from one of three positions at any given time (sample vial, purge chamber, or waste container). To reduce the effect of sample carryover, the syringe pump picks up 4 mL of sample water from each new sample and dispenses it to waste. This amount of liquid represents slightly more than three-times the internal volume of the tubing between sample bottles and the syringe-pump. After this rinse step, the syringe pump picks up sample water and dispenses the desired volume to the purge chamber. Occasionally, a small helium bubble (from sample tubing which was flushed with helium prior to analysis, see below) can become stuck to the plunger or the top of the syringe. To ensure this bubble is not transferred to the purge chamber, the syringe-pump withdraws an additional 1 mL from each sample, and the first 0.5 mL is dispensed to waste to remove any bubbles from the top of the syringe. The sample volume is then dispensed to the purge chamber, and the final 0.5 mL in the syringe is dispensed to waste, thus removing the bubble adhered to the plunger.

The purge chamber (Figure 2.1, component C) is a custom built elongated (40 cm high by

19 mm O.D.) glass tube with a round bottom and 13mm opening at the top. The opening is plugged with a thick, blue rubber stopper (Chemglass -PN: CLS-4209-14), through which three

1/16” O.D. stainless steel tubes are inserted. One tube is used to transfer sample water between the purge chamber and the syringe pump, the second tube sends UHP Helium (50 mL min-1) into the bottom of the chamber for gas stripping, and the third tube (which protrudes just below the bottom of the stopper) is the gas outlet from the purge chamber to the trap. Once the purge chamber is filled with sample water, the helium flow to the chamber is turned on with a solenoid 19

valve in order to strip dissolved gas from the sample. Gases exit the purge chamber via the outlet tube and pass through a Nafion drying tube (Figure 2.1, component D, Permapure - PN:

MD-070-24S-2) to remove any water vapour, which can clog the trap with ice and interfere with sample analysis. We operate the Nafion tube with a counter-flow of dry N2 flowing at 100 mL min-1. The dry gas exiting the Nafion tube is then passed through a 2-position, 6-port Valco injection valve (Figure 2.1, component F, VICI -PN: DC6WE). In this ‘load’ position, the sample gases are passed through the trap and the He purge flow is vented to waste, while the GC receives a 10 mL min-1 flow of UHP He.

Prior to the transfer of a liquid sample into the purge chamber, the trap is automatically lowered into a liquid-nitrogen bath using an electronic linear actuator (Thomson – PN: S-24-

27A8-06). The trap is submerged ~1cm below the surface of the liquid nitrogen, using a slightly lower trap position for each sample in the sequence to compensate for evaporation of liquid nitrogen in the Dewar. By the time the sample water is transferred to the purge chamber, the trap has been submerged in liquid nitrogen for a full-minute, ensuring it has reached the desired temperature prior to gas stripping.

After the sample has been completely purged of analytes (4 minutes for a 5mL sample - see below), the syringe pump extracts the liquid in the purge chamber and dispenses it to waste.

The GCMS acquisition is then initiated, and the 2-way trap valve (Figure 2.1, component F,

VICI –DC6WE) is switched to the inject position. In this position, the GC carrier (UHP He) passes through the trap into the GC, while the He purge flow passes directly to waste, by-passing the trap. The trap is then raised out of liquid nitrogen and heated to room temperature by passing high-current electricity (~70 amps, at 3 volts) through the trap for 8 seconds. This heating releases the adsorbed CH4 and N2O from the trap, allowing them to flow with the carrier-gas to 20

the GCMS. After the trap has been flushed onto the GC for one minute, the trap valve is switched back to the load position, and the analysis of the next sample is initiated. Flushing the trap for one-minute enables heat from the stainless steel trap to be distributed evenly throughout the Porapak lining so that all analytes are thoroughly flushed from the trap. The time needed to raise and heat the trap is 17 seconds, which leaves 43 seconds for the stainless steel trap to heat the Porapak and flush the analytes into the He carrier stream. Given the total flow of the GCMS carrier (10 mL min-1), the trap will be flushed with ~7.2 mL He, or roughly nine times the internal volume of the trap (~0.8 mL) during this interval.

The detector is a Shimadzu QP2010 quadrupole mass spectrometer equipped with an electron-impact ion source. For GC gas separation, we use a 30 m CarbonPLOT column (J&W

Scientific – PN: 113-3132), with a column flow rate of 1.5 mL min-1 (total flow = 10 mL min-1, purge flow = 1 mL min-1), held at a constant temperature of 40°C, and an injection split ratio of

5:1. Figure 2.3 shows an example chromatogram obtained for a seawater sample. Methane has a shorter retention time than N2O, and is measured at its secondary mass-to-charge ratio (m/z) of

15.00 atomic mass units (amu) to eliminate peak interferences from oxygen at m/z = 16.00 amu

(Figure 2.3). Nitrous-oxide elutes after roughly three minutes, and is also measured at its secondary m/z (30.00 amu) to avoid interference with CO2 at m/z = 44.00 amu, as per the method described by (Punshon and Moore, 2004b). The GCMS analysis software automatically integrates the area under each peak in the chromatogram, and writes them to an ASCII file on the computer. These values are subsequently written to a data file generated by the LabVIEW program. The automation of the analysis sequence ensures consistent peak retention times, allowing for completely automated peak processing by the GCMS Analysis software. We routinely inspect the automated peak-integration to ensure consistency, and can adjust peak integration settings or manually re-integrate peaks, if needed.

Two blanks are automatically measured at the start of each batch, by performing all of the above steps, without transferring any liquid into the purge vessel (i.e. cooling the trap, turning on the purge He gas flow for 4 minutes, equivalent to the purge time, then heating and injecting the trap to the GCMS). These blanks provide a measure of any residual analyte present in the purge vessel, sampling lines and valve flow paths. Following the blank determination, a five-point standard curve is automatically generated by measuring the detector response to a series of precisely measured quantities of CH4 and N2O. Additional single standard measurements are automatically performed after 8, 16, and 24 samples (roughly once per hour) to monitor instrumental drift during the batch. We prepare the calibration samples by diluting a standard gas mixture (Praxair Canada; 10.3 ppm CH4 and 10.0 ppm N2O, balance N2) with UHP Helium using two electronic mass-flow controllers (Figure 2.1, component G, Advanced Specialty Gas 22

Equipment – 202-4117-5FRC/9FHC). The accuracy of our Praxair standard tank was calibrated against a certified primary standard gas mixture prepared by the National Oceanic and

Atmospheric Administration (Boulder, CO). To make low (high) concentration standards, we increase (decrease) the He flow-rate and decrease (increase) the standard gas flow-rate. The diluted gas mixtures are flushed for one minute through a stainless steel standard loop of known internal volume (VICI – CSL-series) mounted on a 2-way, 6 port injector valve (Figure 2.1, component H, VICI –DC6WE). A gas solenoid valve then stops the gas flow for five seconds to equilibrate the pressure in the standard loop with air, and the valve is switched to inject the standard into the purge chamber. For each standard, number of moles (n) of CH4 and N2O injected is calculated using the ideal gas law (PV = nRT). The partial pressure (P) of CH4 and

N2O in the sampling loop is derived from the total gas pressure in the sampling loop (measured by a Keller Digital Quartz Pressure sensor -Model: Preciseline 0-30 PSIA), and the mixing ratio of standard gas and He set by the mass flow controller. The temperature of the gas inside the sampling loop is assumed to be equal to the laboratory temperature, which is measured by thermocouple wired into an Omega CN7500 control module. Once injected into the purge chamber, the standard mixtures are analyzed in the same way as liquid samples. Our multi-point calibration method is fully automated, unlike conventional calibrations, which require manual syringe injections or manual changing of a sample loops.

A data file is generated by the LabVIEW program during the batch, and includes information such as the sample name, depth, date collected, date/time analyzed, detector response (i.e. peak area) for CH4 and N2O, and, in the case of calibrations, the number of moles of CH4 and N2O injected. We employ a Matlab script to read this data file and derive the sample concentrations. The script first calculates the linear slope and intercept of the calibration curve 23

for the batch (moles of gas injected vs. peak area), and uses this to calculate the number of moles measured in each sample based on the measured peak areas. Because the purge and trap technique removes all dissolved CH4 and N2O from sample water, there is no need to employ gas-solubility equations to derive concentrations from the number of moles of gas detected (as required by the headspace technique). Rather, the number of moles in each sample is divided by the total volume of water that is purged to calculate concentration.

2.4 Assessment

2.4.1 Recovery efficiency

To determine the recovery efficiency of the trap, we compared the peak areas obtained for standard samples directly injected onto the GCMS (DirectPA) against the peak areas obtained when the same quantity of gas was dispensed into the purge vessel and subjected to our full analytical procedure (i.e. cryotrapping and injecting; P&TPA).

% Recovery = 100 * P&TPA / DirectPA (1.1)

Percent recovery of the purge and trap was determined to be 107±3.1% for CH4 and

99.6±1.7% for N2O. The 107% recovery for methane results from the different peak characteristics observed during direct injection of standards vs. cryotrapping and injection.

Relative to the directly-injected standards, standards that were cryotrapped exhibit sharper and better separated peaks. The cryo-focusing allowed the CH4 peak to fully-separate from the tail of

15 the preceding peak at m/z 15.00 ( N). In contrast, CH4 injected directly from the standard loop exhibited a broader peak that overlapped slightly with the tail of the 15N peak. This caused the

CH4 peak from directly-injected standards to have an artificially high baseline and a smaller 24

effective peak-area than the cryotrapped standards, resulting in an apparent trapping efficiency slightly higher than 100%. This analytical artefact is not a concern for our analysis since all of our samples, blanks and standards pass through the purge and trap system and thus show similar peak shapes. Moreover, the problem is not applicable to N2O, as there are no adjacent peaks on m/z = 30.00. We thus conclude that our purge and trap system is able to fully recover N2O and

CH4 gas from aqueous samples.

2.4.2 Optimization of purge time

The minimum time needed to strip all dissolved gases from sample water was determined by examining the relationship between detector response and purge time for replicate water samples. We prepared a set of air-equilibrated water sample replicates by filling clean sample vials with MilliQ water and placing them, uncapped, in a shaker bath for 5 days in the laboratory at 22°C. The sample vials were then crimp-sealed and analyzed by PT-GCMS (5 mL sample volume), with purge times ranging from 30seconds to 11 minutes. A set of high-concentration liquid samples (1000 nM CH4 and 217 nM N2O) was also prepared by bubbling MilliQ water with CH4 and N2O standard mixtures for ~1 hour, and carefully overfilling sample vials with this water without introducing air bubbles. These vials were immediately crimp-sealed and measured in the same way as the air-equilibrated samples. The results of this experiment show that all of the dissolved CH4 and N2O is stripped from sample water after approximately three minutes of purging, regardless of the concentration, and that analytes are not lost from the trap during extended purging (Figure 2.4). We thus chose 4 minutes as an optimal purge-time for 5mL samples.

2.4.3 Accuracy of gas mixer

To test the accuracy of our gas mixing system, we compared standard curves generated using the mass flow controller against standard curves generated using various sizes of standard loops filled with undiluted standard gas. Over a three week period, we performed one standard curve using each method once per day, and examined the relationship between detector response and moles injected from each method. As shown in Figure 2.5, the two calibration methods produced very similar results, with no consistent offsets observed. 26

2.4.4 Accuracy of liquid samples

Accuracy was measured by preparing a set of five air-equilibrated MilliQ water samples in the manner described above. The expected concentration of CH4 and N2O in the samples was calculated based on the temperature of the water bath, the mean atmospheric pressure in the lab during equilibration, atmospheric concentrations of CH4 and N2O (Cunnold et al., 2002; Prinn et al., 1990), and the temperature-dependent solubility of these gases in fresh water (Weiss and

Price, 1980; Wiesenburg and Guinasso, 1979). We confirmed that lab air has atmospheric concentrations of CH4 and N2O by measurements of lab air samples on the PT-GCMS (Table

2.1), calibrated against the NOAA standard gas mixture. Accuracy was further evaluated by analyzing replicate seawater samples collected at the Hawaii Ocean Time-series (HOT) program long-term monitoring site (Station ALOHA) during November 2013. The analysis of dissolved

CH4 and N2O concentrations was conducted by members of the Scientific Committee on Oceanic

Research (SCOR) Working Group #143 (http://www.scor-int.org/) whereby seawater samples were collected from 25 m and 700 m in replicate and distributed to twelve laboratories around

Figure 2.5 Standard curves generated with and without gas mixer spanning a three-week period. 27

the world who regularly measure CH4 and N2O in the marine environment. This is part of an ongoing project by SCOR Working Group 143 which is currently focused on the production and distribution of CH4 and N2O gas standards for distribution among analytical laboratories.

As shown in Table 2.1, our PT-GCMS measured 2.8 ± 0.1nM CH4 and 8.5 ± 0.3nM N2O in air-equilibrated MilliQ water standards. These values are in excellent agreement with the expected equilibrium concentrations of these gases in fresh water (Table 2.1). Furthermore, the laboratory inter-calibration exercise revealed good agreement between our CH4 and N2O measurements and those of other laboratory groups (Table 2.2). These results indicate that our system is able to produce highly accurate measurements of CH4 and N2O concentrations in marine water samples.

Table 2.1 Results of air-equilibrated water test for accuracy. CH4 equilibrium calculated using Wiesenburg

& Guinasso (1979) assuming atmospheric CH4 concentration = 1.87ppm CH4 (measured in 2012; Cunnold et al., 2002). N2O equilibrium concentration calculated using Weiss and Price (1980) assuming atmospheric N2O concentration = 325 ppb (measured in 2012; Prinn et al., 1990). Concentrations of CH4 and N2O in laboratory air were measured using the PT-GCMS, and were not significantly different from atmospheric values (1.84

+/- 0.18 ppm CH4 and 323 +/- 60 ppb N2O).

Gas nM observed nM expected

CH4 2.8 ± 0.1 (n=5) 2.7

N2O 8.5 ± 0.3 (n=5) 8.5

Gas Depth Capelle et al. SCOR WG#143

CH4 25m 2.3 ± 0.3 (n=3) 2.3 ± 0.3 (n=30)

700m 1.2 ± 0.5 (n=3) 1.0 ± 0.4 (n=31)

N2O 25m 6.4 ± 0.1 (n=3) 6.6 ± 0.5 (n=40)

700m 39.9 ± 0.3 (n=3) 38.8 ± 5.0 (n=38)

2.4.5 Precision

Precision was tested by measuring sixteen replicate, air-equilibrated water samples, prepared as described above. The coefficient of variation between the 16 replicates was then calculated, using:

σ C. V. = 100 ∗ (1.2) mean

Where, σ represents the standard deviation of the measurements. The coefficient of variation between 16 replicate air-equilibrated water samples was 3.3% for CH4 and 3.0% for N2O. These values are similar to conventional methods (Careri et al., 1999; Elkins, 1980; Lilley et al., 1982;

Yoshida et al., 2011).

2.4.6 Detection limits

The detection limits were calculated according to Currie (1968), using:

L.D. = 3.29 * σBlank (1.3)

Where L.D. is the limit of detection, and σBlank is the standard deviation of repeated blank

measurements. To derive σBlank, we calculated the standard deviation of the observed CH4 and

N2O peak areas for 29 blank measurements, and converted these peak areas to molar concentrations using appropriate calibration curves.

The limit of detection is dependent on the volume of sample water purged. When purging a 5 mL sample, the limits of detection for CH4 and N2O were 0.40 nM and 0.43 nM, respectively. These values are well below the atmospheric equilibrium values of most oceanic waters, and below the vast majority of values observed in the open ocean (Bange, 2008;

Valentine, 2011). If necessary, however, detection limits can be further reduced. For example, concentrations as low as 0.10 nM CH4 and 0.11 nM N2O can be detected by purging 20mL samples.

2.4.7 System performance for the analysis of field collected samples

For nearly a decade, we have been collecting monthly time-series samples of dissolved gas depth profiles in Saanich Inlet, British Columbia (Zaikova et al., 2010). This system is a seasonally anoxic fjord, where annual changes in redox chemistry drive large variability in N2O and CH4 concentrations (Cohen, 1978; Lilley et al., 1982; Ward et al., 1989). Initially, we processed the Saanich Inlet samples using a time-consuming manual gas extraction method using head-space equilibration. Using this method, we could process up to 16 samples per day, with extensive operator involvement required for each sample. We now use our automated system to processes the samples using PT-GCMS, and are able to measure more than one complete depth profile in a single batch run, with more than 50 samples (2 batches) analyzed per day.

Additional samples can also be run overnight, with the system left unattended. The analysis of 25 samples requires ~30 minutes of operator involvement for the initial set up of sample vials. 30

Our N2O and CH4 measurements from Saanich Inlet show the expected behaviour of these gases in the water column. In particular, we observed the disappearance of N2O in the anoxic water column, which results from N2O reduction during the last step in denitrification. In contrast, CH4 accumulates in sub-surface waters due to diffusion from anoxic sediments, where it is produced during anaerobic decomposition of organic matter. Our results show good agreement with previously published N2O and CH4 measurements from Saanich Inlet, covering a relatively wide range of concentrations. N2O profiles from Saanich Inlet from August 1977

(Cohen, 1978) closely resemble the profiles we measured during August 2012 (Figure 2.6).

Similarly, CH4 profiles from June, 1982 (Lilley, Baross, & Gordon, 1982) show the same

Figure 2.6 Comparison of water column profiles from Saanich Inlet, BC measured by PT-GCMS and conventional methods. Lilley et al.’s (1982) June CH4 profile (FID) matches well with June 2011 profile measured by PTGCMS (left). N2O profiles taken in August 1977, measured by ECD (Cohen, 1978) also show good agreement with a depth profile from August 2012, measured by PT-GCMS (right).

dominant features we observed in the June, 2011 profile (Figure 2.6). Any small differences between our measurements and those Lilley et al. (1982) and Cohen (1978) are likely due to natural inter-annual variability.

2.5 Discussion

The results of our assessment indicate the PT-GCMS achieves comparable accuracy and precision to conventional methods for dissolved CH4 and N2O analysis in discrete water samples.

Purging a 5 mL aliquot achieves adequate sensitivity for typical oceanic measurements. The automation of gas extraction and standards significantly reduces the labour required both during sample analysis and data processing. Moreover, the use of a single detector allows N2O and CH4 to be measured from a single sample, minimizing the number of discrete water samples required, and eliminating potential sources of uncertainty and sampling errors. The automated PT-GCMS can thus facilitate the analysis of large numbers of samples, thereby improving the spatial and temporal coverage of CH4 and N2O observations in time-series programs or high-resolution surveys.

Our automated purge and trap system offers advantages over commercially available auto-samplers. It consists of individual components that can be modified, replaced, or repaired with minimal technical expertise. Whereas many commercial auto-samplers can only accommodate specific vials, our system can be used with virtually any sample container (from

20 mL to >500 mL). Most importantly, our system includes built-in routines for sample calibration and blank determinations.

2.6 Comments and recommendations

We would recommend that any research group analyzing large numbers of dissolved CH4 and N2O samples consider implementing an automated sample processing scheme as described here. The custom-built purge and trap auto-sampler we have developed could be added on to the front-end of conventional detectors for dissolved CH4 and N2O (e.g. GC-FID and GC-ECD, respectively). The MS detector offers the potential for future applications covering a wide range of analysis options, including other volatile gases and isotopic species. For example, with some modifications to the front-end auto-sampler, it would be possible to add a headspace equilibration analysis to measure dissolved O2 and N2 from the same sample vials used for CH4 and N2O measurements (Upstill-Goddard et al., 1996). The purge and trap system is not suitable for O2 or N2 analysis because these gases are not readily adsorbed by our trap. Moreover, the high concentrations of these gases in most aqueous samples would likely over-saturate the detector when used with a purge and trap system. The MS could also be used to measure isotopes of CH4 and N2O in tracer incubation experiments to determine the production rates of various gases from labelled substrates (Punshon and Moore, 2004a, 2004b; Sich and Russow,

1999). The method, however, is not suitable for the analysis of natural isotope variations due to the relatively low precision (± ~ 3 %) of the quadrupole GCMS as compared to dedicated isotope ratio mass spectrometers. In such tracer experiments, our auto-sampler could be programmed to automatically take measurements from live incubation experiments at specific time points, thereby eliminating the inconvenience and potential errors associated with time-sensitive sub- sampling. The high cost of the MS impede the widespread adoption of this technique for dissolved CH4 and N2O measurement, but these additional analyses could not be accomplished using conventional detectors, or commercially available purge and trap systems. 33

Chapter 3: Factors controlling methane and nitrous-oxide variability in the southern British Columbia coastal upwelling system

3.1 Summary

Coastal upwelling systems are important marine sources of methane (CH4) and nitrous- oxide (N2O). Current understanding of the controls on CH4 and N2O distributions in these coastal waters is restricted by limited data availability. We present the first multi-year measurements of

CH4 and N2O distributions from the seasonally upwelling shelf waters of British Columbia,

Canada, a coastal end-member of the north Pacific oxygen minimum zone (OMZ). Our data show significant seasonal differences in CH4 and N2O distributions and fluxes driven predominantly by upwelling. Methane is supplied to the water column primarily from sediments

(especially near methane seeps), and is transported to the surface mixed layer by upwelling. A positive correlation between CH4 concentrations and salinity indicates limited inputs from Fraser

River estuary waters to the study site. Shelf waters receive N2O from a deep, off-shelf N2O maximum in the OMZ core, and from nitrification in the water column and possibly sediments.

Both the physical transport of N2O and its apparent in situ production are enhanced under upwelling conditions. N2O yields from nitrification, estimated from changes in N2O and nitrate +

- - nitrite (NO3 +NO2 ) along isopycnals, ranged from 0.04 – 0.49%, with the highest values

-2 observed under low ambient O2 concentrations. Sea-air fluxes ranged from -4.5 – 21.9 µmol m

-1 -2 -1 day for N2O and 2.5 – 34.1 µmol m day for CH4, with the highest surface fluxes observed following summer upwelling over the broad continental shelf of southern Vancouver Island. Our results provide new insight into the factors driving spatial and inter-annual variability in marine

CH4 and N2O in high productivity coastal upwelling regions. Continued time-series measurements will be invaluable in understanding the longer-term impacts of climate-driven variability on marine biogeochemical cycles in these dynamic near-shore waters.

3.2 Introduction

Methane (CH4) and nitrous-oxide (N2O) are the most important greenhouse gases after carbon-dioxide and water-vapour, accounting for ~ 17% and 6% of the global radiative forcing of all greenhouse gases, respectively (IPCC, 2013). These gases are actively cycled in low oxygen sub-surface ocean waters and sediments, where intensive microbial activity drives a diverse suite of metabolic pathways.

The major processes driving marine N2O cycling are nitrification and denitrification. N2O

+ is produced as a by-product of nitrification (step-wise oxidation of ammonium (NH4 ) to nitrite

- - (NO2 ) and nitrate (NO3 )), which is carried out by a variety of chemo-autotrophic bacteria and archaea under oxic to nearly anoxic conditions (Casciotti and Buchwald, 2012; Freing et al.,

- - 2012). N2O yields from marine nitrification (i.e. mol N2O produced per mol NO2 +NO3 produced) are highly variable, ranging from 0.004 – 0.4 % (De Wilde and De Bie, 2000; Frame and Casciotti, 2010; Goreau et al., 1980; Punshon and Moore, 2004a; Santoro et al., 2011;

Stieglmeier et al., 2014), and have been shown to increase under low oxygen conditions (Frame and Casciotti, 2010; Goreau et al., 1980; Stieglmeier et al., 2014). The change in N2O yield may

- be due to the tendency of nitrifiers to preferentially reduce NO2 to N2O (nitrifier-denitrification) under O2-limitation (Frame and Casciotti, 2010). Denitrification (the step-wise reduction of NO3

- to N2 via NO2 , nitric oxide (NO) and N2O) is typically confined to waters with <5 µM O2

(Codispoti et al., 2001), and is ultimately a sink of N2O under anoxic conditions. However, the 35

enzyme N2O-reductase is more O2-sensitive than the other N-reductase enzymes in denitrification, resulting in N2O accumulation by partial/incomplete denitrification under very low (sub-micromolar) O2 concentrations (Betlach and Tiedje, 1981; Dalsgaard et al., 2014).

Indeed, denitrification appears to be a dominant source of N2O in suboxic marine waters such as the Arabian Sea and West-Indian continental shelf (Bange et al., 2001; Codispoti et al., 2001;

Jayakumar et al., 2009; Naqvi et al., 2000), and at the peripheries of OMZs (Bange, 2008;

Castro-González and Farías, 2004). The highest N2O production rates thus occur in low oxygen waters, where high N2O-yields from nitrification co-occur with net- N2O production from denitrification. Recent evidence suggests that N2O may be produced during dissimilatory reduction of nitrate to ammonium (Welsh et al., 2001), but this process appears to be confined to anoxic or very low O2 (< 10µM) waters (Lam et al., 2009).

Oxygen levels also exert a significant control on the marine CH4 cycle. Until relatively recently, this gas was thought to be produced exclusively under anaerobic conditions during the biological or thermogenic breakdown of organic matter. Anaerobic CH4-producing environments are generally confined to organic-matter rich sediments or within the earth’s crust, although they can also be present inside sinking particles or digestive tracts of marine organisms (De Angelis and Lee, 1994; Holmes et al., 2000; Oremland, 1979; Sansone et al., 2001). Sediment-derived

CH4 is often mostly consumed by methanotrophic organisms, thus limiting CH4 fluxes to the atmosphere, although the abundance of methanotrophs and their ability to consume CH4 from sediments can be highly variable (Reeburgh, 2007; Steinle et al., 2015). Methane from subsurface organic deposits may migrate upward through coarse grained sediments or tectonic faults and escape the water column in seep-derived bubbles, which can enhance CH4 transport into the mixed layer (Reeburgh, 2007; Rehder et al., 2009, 2002; Solomon et al., 2009). The 36

release of CH4 from these seeps shows strong spatial and temporal variability over a range of time-scales (from hours to years; Boles, Clark, Leifer, & Washburn, 2001; Leifer & Boles, 2005;

Tryon et al., 1999), and represents a potentially underestimated source of atmospheric CH4.

Moreover, the potential destabilization of CH4-rich clathrate deposits under various ocean warming scenarios has prompted significant research effort in recent years (Archer, 2007;

Solomon et al., 2009; Sowers, 2006). There has also been increased interest in other in situ sources of CH4 in oxygenated marine surface waters, including the cleavage of methyl-groups from larger molecules, such as methylated sulfides (Damm et al., 2010; Florez-Leiva et al.,

2013) and methylphosphonate (Cooke et al., 2012; Karl et al., 2008). Due to their proximity to the ocean-atmospheric interface, these surface water pathways of CH4 production may be important controls on sea-air CH4 fluxes.

Coastal upwelling regions are sites of active CH4 and N2O cycling, and disproportional contributors to the global marine emissions of these gases to the atmosphere (Bange, 2008;

Nevison et al., 2004; Rehder et al., 2002; Sansone et al., 2001). High CH4 and N2O fluxes have previously been demonstrated in a number of upwelling regions (Bange et al., 1996, 1994;

Pierotti and Rasmussen, 1980; Sansone et al., 2001), and recent research efforts have explicitly examined the effects of upwelling on coastal CH4 and N2O distributions in the waters of coastal

Peru (Kock et al., 2015), Chile (Cornejo and Farías, 2012; Farías et al., 2015), Mauritania in NW

Africa (Kock et al., 2008; Wittke et al., 2010), California (Cynar and Yayanos, 1992; Lueker et al., 2003; Nevison et al., 2004), Oregon (Rehder et al., 2002), and the Arabian Sea (Bange et al.,

2001). In these upwelling systems, high surface productivity results in significant fluxes of organic carbon to sub-surface waters, fueling microbial O2-demand and driving redox gradients that favor N2O and CH4 production at relatively shallow depths (Naqvi et al., 2010; Sansone and 37

Popp, 2001). Upwelling can also act to transport CH4 and N2O-rich sub-surface water into the mixed layer (Bange et al., 2001; Cornejo and Farías, 2012; Lueker et al., 2003; Naqvi et al.,

2010; Nevison et al., 2004; Rehder et al., 2002). Short-term variability in upwelling over periods of hours to months has been shown to influence CH4 and N2O fluxes from coastal upwelling systems (Bange et al., 2001; Cornejo and Farías, 2012; Lueker et al., 2003; Rehder et al., 2002;

Wittke et al., 2010). Upwelling of O2-depleted water has been linked to high N2O production rates and sea-air fluxes in a number of coastal systems, including the equatorial Pacific upwelling zone of Chile and Peru (Cornejo and Farías, 2012; Farías et al., 2009), and the Arabian

Sea (Bange et al., 2001). The ongoing expansion and intensification of OMZs (Falkowski et al.,

2011; Keeling et al., 2010; Stramma et al., 2010; Whitney et al., 2007) and the intensification of coastal upwelling due to stronger land-sea atmospheric pressure gradients (Bakun, 1990;

Bylhouwer et al., 2013; Wang et al., 2015) – both of which are predicted effects of climate change – may thus lead to increased CH4 and N2O fluxes from coastal upwelling systems

(Codispoti, 2010; Codispoti et al., 2001; Naqvi et al., 2010; Rehder et al., 2002). In addition to upwelling, other factors including eutrophication, sedimentary diffusion, freshwater inputs, and local bathymetry also appear to influence coastal N2O and CH4 distributions, resulting in high spatial and temporal variability in surface water concentrations and sea-air fluxes. This variability, combined with a scarcity of data, limits our ability to quantify CH4 and N2O emissions in coastal upwelling systems, and our understanding of longer-term (e.g. inter-annual) responses to environmental forcing.

In this article, we present new field data documenting the seasonal and inter-annual variability in CH4 and N2O concentrations and sea-air fluxes along the west coast of Vancouver

Island (WCVI), British Columbia (BC). This coastal region is characterized by high seasonal 38

productivity, resulting from wind-driven summer time upwelling. Our study site lies in close proximity to the large oxygen-minimum zone (OMZ) of the subarctic North Pacific, which supplies O2-depleted water to the shelf during upwelling (Crawford and Peña, 2013). The persistently low O2 levels in these waters have been declining in recent decades (Crawford and

Peña, 2013; Whitney et al., 2007), and this has increased the likelihood of periodic upwelling of hypoxic water onto the shelf (Roegner et al., 2011), potentially enhancing the sea-air flux of CH4 and N2O. Our study site also contains a number of sedimentary bubble plumes (seeps), which have been identified as important sources of CH4 to the water column along the Oregon Coast

(Grant and Whiticar, 2002; Heeschen et al., 2005; Suess et al., 1999). The combination of upwelling, intensifying shelf hypoxia, and presence of seeps make the WCVI a potentially significant site for CH4 and N2O production and high sea-air fluxes. To date, CH4 and N2O distributions in this region have not been examined systematically.

Based on data obtained from five spring and summer cruises, we present detailed observations of the spatial and temporal variability of N2O and CH4 concentrations and sea-air fluxes, and use these observations to examine the processes affecting the distributions of these gases in the water column. In particular, we examine the influence of upwelling and fresh water

(Fraser River) fluxes on CH4 and N2O distributions along the BC continental shelf, the potential contribution of sedimentary bubble-plumes to water column CH4 budgets, and the oxygen- dependent changes in the N2O yields from nitrification. This work represents a starting point for future time-series observations of CH4 and N2O dynamics in coastal BC waters.

3.3 Methods

3.3.1 Study site

The WCVI region is located at the northern end of the eastern, North Pacific upwelling region (Figure 3.1). The upwelling season typically runs from June to September each year, while downwelling occurs between October and May (Bylhouwer et al., 2013). The onset, duration and intensity of upwelling is variable on an inter-annual basis, and this variability has been associated with the Pacific Decadal Oscillation (PDO) and El-Niño Southern Oscillation

(ENSO) (Bylhouwer et al., 2013). There is a permanent oxygen minimum zone (OMZ; defined as O2 concentrations less than 20 µM) located between 800-1200 m depth in the water directly adjacent to the continental shelf (Figure 3.1, inset). The study area is influenced by several local water masses. The Vancouver Island Coastal Current (VICC) is a buoyancy-driven freshwater current that runs in a northerly direction along the coast of Vancouver Island (shore-ward of the

150m depth contour), fed by the Fraser river freshet (i.e. snow-melt runoff) during the spring and summer, and by coastal mountain rain runoff from Vancouver Island during the fall and winter

(Foreman, 2000; Masson and Cummins, 1999). The VICC can extend to the bottom of the water column (Masson and Cummins, 1999). Over the outer continental shelf region, surface currents are driven by seasonal winds, flowing predominantly northward during winter and south during summer. Sub-surface currents in this outer shelf region are dominated by the northward-flowing warm, saline and nutrient-rich California Undercurrent (CUC), with a core depth ranging from

125 – 250 m (Foreman, 2000; Thomson and Krassovski, 2010). A persistent cyclonic eddy (Juan de Fuca Eddy) is found in the southeastern region of our study area (near the coastal LB and LC

Figure 3.1 Map of the study area along the west coast of Vancouver Island (WCVI). Depth-resolved samples were collected from profile stations (black circles) during five cruises (Jun-2012, Sep-2012, Jun-2013, Sep-

2013, and Jun-2014) along the Coastal Transect and the cross-shelf LC Line Transect. During June 2014, 5 m samples were collected from a number of additional stations (black triangles). The locations of CH4-seeps

(located using 12 kHz echo sounder data) are shown by black x’s (Vaughn J. Barrie, pers. comm.). The O2 concentrations at 800m from the World Ocean Atlas climatology (Garcia et al., 2009) are shown in the inset, with the 20µM O2 contour line shown, highlighting the subarctic North Pacific OMZ.

Line stations, Figure 3.1) during summer and fall, associated with the Juan de Fuca and Tully canyons. Here, local currents and bottom topography enhance the upward transport of low O2, nutrient-rich deep-water on to the shelf (Crawford and Peña, 2013).

Continental shelf sediments along the WCVI are primarily sands and gravels with low organic matter content (<1% by weight), though fine-grained silts and clays with ~3% organic matter content are present in some near-shore locations (Carter, 1973). Large CH4-hydrate deposits have been mapped at depth (>1200 m) off the Vancouver Island coast (Riedel et al.,

2002), and CH4 has been observed leaking into seawater along the Cascadia tectonic margin off the west coast of Oregon (Suess et al., 1999).

3.3.2 Field sampling and gas analysis

Sample collection took place during 5 coastal research cruises on the CCGS John P.

Tully in June and September between 2012 and 2014 (cruises 2012-25, 2012-59, 2013-38, 2013-

58, and 2014-21). During each cruise, we collected depth profile samples from 11 stations along two perpendicular transects; one along-shore coastal transect, and one cross-shelf transect

(Figure 3.1). Additional surface (5m) samples were collected during the June 2014 cruise (Figure

3.1) in order to derive more spatially resolved sea-air flux estimates. Discrete water column samples for dissolved CH4 and N2O analysis were collected using a rosette equipped with 12 L

Niskin bottles. Water from the Niskin bottle was transferred to 60 mL glass vials (two replicates for each sample depth) using a flexible silicon tube in a manner that eliminated bubbles, with vials overfilled three volumes to prevent air contamination. Each vial was immediately poisoned with 100 µL saturated HgCl2 solution, crimp-sealed with rubber butyl stoppers, and stored at 4°C until analysis by automated purge and trap gas-chromatography-mass-spectrometry (PT-GCMS). 42

Our method, described in detail by Capelle et al. (2015), provides an average precision of 3%, and detection limits of 0.4 nM for both CH4 and N2O when purging 5 mL of sample water. The median difference between all duplicate concentration measurements in this study was less than

5%. The rosette was equipped with a CTD (SBE-911plus) and oxygen sensor (SBE 43) to

- - measure salinity, temperature, and oxygen. Discrete measurements of oxygen and NO2 +NO3 were also made at each depth following protocols employed the Institute of Ocean Sciences

(Barwell-Clarke and Whitney, 1996).

3.3.3 Upwelling intensity

Upwelling indices for 48.0°N, 125.0°W were obtained with 6-hourly resolution from the

Pacific Fisheries Environmental Laboratory (http://www.pfeg.noaa.gov/products/). The upwelling index provides a measure of the strength of upwelling or downwelling-favourable winds. It is derived from the atmospheric pressure gradients along the coastal ocean margin and

Ekman transport calculations. Upwelling (downwelling) occurs when winds are blowing parallel to the coastline in a southerly (northerly) direction. To determine a characteristic time-scale over which upwelling influenced shelf water properties, we computed the mean values of salinity,

- - temperature, oxygen, and NO2 +NO3 in sub-surface (below 50 m) waters on the continental shelf

(maximum depth 200m), and correlated these properties with the mean upwelling intensity derived over a range of time intervals, from 1 to 150 days prior to each cruise (in 10 day intervals). From this analysis, we observed the highest correlation coefficients when using an averaging period between 80 and 110 days. This time-scale agrees well with the estimates of shelf residence time of ~ three months determined by Ianson et al. (2009). We thus used an averaging window of 90 days to determine the mean upwelling state prior to each cruise. The 43

onset of the upwelling season was calculated according to the method of Bylhouwer et al.

(2013), based on the date when the cumulative annual upwelling value equals 10% of the total annual upwelling.

3.3.4 River discharge

The primary source of freshwater to the WCVI between spring and fall is the Fraser

River (Masson and Cummins, 1999). Mean daily discharge values were obtained from the hydrometric gauge station at Hope, BC (Station 08MF005, Environment Canada, 2015).

3.3.5 Sea-air fluxes

-2 -1 Sea-air fluxes of CH4 and N2O (µmolm day ) were calculated as the product of the air- sea disequilibrium (ΔC) and a gas exchange coefficient (i.e. piston velocity, kw)

Flux = kw * ΔC = kw * (Cobs - Ceq) (3.1)

The excess (or deficit) of CH4 or N2O in surface waters relative to saturation values, was calculated from the difference between the mean observed (Cobs ) and air-equilibrium (Ceq) concentrations in the mixed layer (or upper 15 m of the water column where the mixed layer was not sampled). Air equilibrium CH4 and N2O concentrations were calculated using the equations of Weiss & Price (1980) and Wiesenburg & Guinasso (1979) for N2O and CH4 solubility, respectively, using mean monthly atmospheric CH4 and N2O concentrations from 2012, 2013 and 2014 measured at Barrow, Alaska (Data provided by NOAA ESRL Global Monitoring

Division, Boulder, Colorado, USA; http://esrl.noaa.gov/gmd/). Piston velocities were computed as a function of the Schmidt number for each gas and wind speeds during the period two weeks prior to sampling. Daily wind speeds were derived from the mean of measurements from two 44

moored buoys in our study area (46206 and 46132, data provided by Fisheries and Oceans

Canada http://www.meds-sdmm.dfo-mpo.gc.ca/isdm-gdsi/waves-vagues/index-eng.htm) and the

NCEP/NCAR daily wind speeds from four locations in our study area (47.5N, 125.0W; 47.5N,

127.5W;50.0N, 127.5W; and 50.0N, 130.0W) (http://www.esrl.noaa.gov/psd/data/, Kalnay et al.,

1996). We found reasonably good agreement between these different sources of wind speed data, with an overall standard deviation between the daily buoy and NCEP/NCAR wind speeds of 1.8m/s. We followed the approach of Reuer et al. (2007) to derive a weighted piston velocity over the two weeks prior to our measurements. This approach takes into account recent wind speed history to obtain a weighting function based on the fraction of the mixed layer depth ventilated on any given day (see Reuer et al., 2007 for full details). The weighted piston velocity provides a less biased estimate of gas exchange coefficients over the residence time of gases in the surface mixed layer.

3.3.6 Calculation of in situ N2O production

To estimate the amount of N2O in shelf waters derived from in situ production, we calculated differences in N2O concentrations across isopycnals between an off-shelf (LC11) and on-shelf (LC04) station along the LC transect (see Figure 3.1 for station locations, and Figure 3.6 for positions of upper and lower isopycnals used for this calculation from each cruise). For this

-3 analysis, discrete N2O measurements from each of these stations were interpolated to 0.01 kg m density intervals, and the difference between on-shelf and off-shelf N2O concentration (dN2O) across each of these density surfaces was computed. The density surfaces, ranging from 25.3 to

26.8 σθ, represented depths between 50 m depth and the bottom of the shelf station (see Figure

3.6 for specific isopycnal ranges used for each cruise). Using this same approach, we computed 45

- - - - the change in NO2 +NO3 concentrations (dNO2 +NO3 ) along these density surfaces over the cross-shelf transect. N2O yields of nitrification on each isopycnal were then derived by dividing

- - the dN2O by dNO2 +NO3 , and computing the average N2O-yield for each cruise. This method also enabled us to assess the influence of O2 availability on N2O-yields from nitrification, by examining the relationship between derived N2O-yields and the corresponding O2-concentrations at the on-shelf sampling station.

The calculations described above rely on some key assumptions. First, we assume that

N2O concentrations along isopycnals should be constant in the absence of biological N2O cycling, so that changes in N2O along an isopycnal can be ascribed to in situ production.

Moreover, we assume that no appreciable denitrification occurs in the water column of our study region. As discussed in the results section, this assumption is supported by an examination of

- - - N2O, NO2 +NO3 and O2 data. Also, since primary producers (which consume NO3 ) are restricted to the euphotic zone, nitrification is assumed to be the primary biological factor

- - affecting both N2O and NO3 in sub-surface waters. Sedimentary nitrification could supply NO3 and N2O to the water column near the sediment water interface, but our analysis does not

- - distinguish between N2O and NO2 +NO3 derived from sediments vs. the water column.

Similarly, our analysis does not account for alongshore transport or diapycnal mixing, which also

- likely contribute to the dissolved N2O, O2, and NO3 gradients in our study area.

3.4 Results

3.4.1 Hydrographic conditions - upwelling intensity, riverine inputs, and O2 concentrations

Upwelling along the continental shelf of our study area was highly variable over short timescales (Figure 3.2a), but the 14-day running mean showed distinct seasonal patterns over the

2.5 years of our time-series. In general, upwelling occurred over much of the summer between

May and September. Mean upwelling values were consistently positive during the 90-days before the two September cruises (between ~ 15 and 20 m3 s-1 100 m coastline-1), and negative

(net downwelling) prior to the June 2012 and 2014 cruises (~ -20 to -30 m3 s-1 100 m coastline-1).

Figure 3.2 Upwelling Index values from the WCVI study region (48 °N, 125 °W, panel a) and Fraser River discharge values (panel b) between January 2012 and November 2014. Upwelling values are plotted with 1- day (grey line) and 14-day (black line) running means. White bars (panel a) indicate mean upwelling index values during the 90-day period before each cruise. Vertical dashed lines indicate approximate sample collection dates during the 5 cruises. 47

In contrast to June cruises in 2012 and 2014, weak positive upwelling was observed in June

2013. Moreover, the onset of the upwelling season was roughly one month earlier in 2013 than in 2012 or 2014 (data not shown).

Daily Fraser River discharge values between 2012 and 2014 are shown in Figure 3.2b.

Peak discharge values typically occurred during May or June, fed by melting snowpack in the coastal mountains of SW British Columbia. Discharge values were lower during September cruises, and of similar magnitude in 2012 and 2013. The earliest peak discharge occurred during

2013, nearly one month earlier than the 2014 peak discharge, and almost two months earlier than the 2012 peak discharge. However, there were small inter-annual differences between the date at which the cumulative annual discharge reached 50% of the total annual discharge, and the cumulative annual discharge on June 1 of each year.

Oxygen concentrations in our study area ranged from 5 µM to 450 µM. The lowest oxygen concentrations were observed in deep waters near the shelf sediments, and in the off- shelf OMZ waters between 800-1000 m, while the highest O2 concentrations were found near the surface

(Figure 3.4 and Figure 3.6). The minimum O2 in shelf waters (depth < 200 m) was 60 µM.

3.4.2 Surface water concentrations and sea-air CH4 and N2O fluxes

Measurements conducted on the June 2014 cruise provide a broad spatial overview of surface CH4 and N2O concentrations and sea-air fluxes in the waters adjacent to Vancouver

Island. We observed CH4 supersaturation in the mixed layer of all stations, and N2O supersaturation for most stations during that cruise (Figure 3.3). Super-saturations were strongest in the southeastern portion of the study area, overlying the broad continental shelf in the vicinity of the Juan de Fuca Eddy (Figure 3.3). We also observed a significant decrease in 48

Figure 3.3 Excess CH4 (panel a) and N2O (panel b) above equilibrium concentrations measured at 5 m depth during June, 2014.

CH4 and N2O supersaturation along the cross-shelf gradient, with generally higher values observed in near-shore waters, decreasing beyond the shelf break (~1000 m isobath).

To examine seasonal and inter-annual variability in regional CH4 and N2O sea-air fluxes and surface supersaturation, we used surface gas measurements collected from a more limited set of stations (Figure 3.1, black circles). Mean fluxes, surface supersaturation and other parameters used to calculate fluxes are listed in Table 3.1. Methane fluxes and saturation values were always positive (net flux to atmosphere), and September fluxes (3.4 to 34.1 µmol m-2 d-1) were significantly higher than June fluxes (2.5 – 20.7 µmol m-2 d-1; Wilcoxon rank sum, p<0.05).

Nitrous-oxide fluxes were also significantly higher during September (0.5 – 21.9 µmol m-2 d-1)

-2 -1 than June (-4.5 – 9.9 µmol m d ), and N2O fluxes were significantly higher following periods of upwelling than downwelling.

3.4.3 Depth-dependent N2O and CH4 concentrations

3.4.3.1 Along-shelf variability

In the along-shore transect (Figure 3.4), CH4 concentrations ranged from 5.1 – 35.9 nM, and N2O concentrations ranged from 9.0 – 33.0 nM. Depth profile measurements along the coastal transect (see Figure 3.1) enabled us to examine the influence of freshwater inputs on gas concentrations. As shown in Figure 3.4, there was a clear signature of low salinity water along the southern portion of the transect, derived from the Fraser River. This salinity-gradient sets up the buoyancy-driven Vancouver Island Coastal Current (VICC), which flows northward along the coast, extending down to the bottom of the water column within a few km from the coast. In

Table 3.1 Mean N2O and CH4 fluxes, excess concentrations above atmospheric equilibrium (ΔN2O and

ΔCH4), mixed layer depths, and time-weighted piston velocities (kw) for each cruise. The number of stations used to calculate fluxes for each cruise is denoted by n. See methods for details of weighted piston velocity calculations.

Cruise Jun-2012 Sep-2012 Jun-2013 Sep-2013 Jun-2014

N2O

Mean ± Std. Dev.

(µmol m-2 day-1) -0.3 ± 3.3 7.6 ± 4.7 4.7 ± 2.9 6.0 ± 6.3 3.0 ± 2.3

Range

(µmol m-2 day-1) -4.5 - 3.9 1.5 - 14.2 0.9 - 9.9 0.5 - 21.9 0.5 - 7.9

N2O (nM) 0.0 ± 2.4 3.7 ± 2.5 3.8 ± 2.6 4.3 ± 4.2 1.4 ± 1.1

CH4

Mean ± Std. Dev.

(µmol m-2 day-1) 10.8 ± 6.1 19.7 ± 10.0 8.5 ± 3.7 11.2 ± 5.8 10.3 ± 5.6

Flux

(µmol m-2 day-1) 4.4 - 20.7 4.5 - 34.1 2.5 - 16.6 3.4 - 19.9 4.3 - 19.3

CH4 (nM) 6.7 ± 2.8 9.5 ± 5.1 6.8 ± 3.0 8.0 ± 4.9 5.0 ± 2.5

Other Parameters

Mixed Layer Depth (m) 7.5 ± 0.6 7.0 ± 0.6 7.0 ± 0.8 9.6 ± 0.8 8.6 ± 1.4

-1 kw (m d ) 1.6 ± 0.4 2.1 ± 0.2 1.3 ± 0.2 1.5 ± 0.4 2.1 ± 0.3 n (stations with data) 8 10 11 10 10

Figure 3.4 Distributions of salinity (panels a-e), N2O (panels f-j), CH4 (panels k-o), and O2 (panels p-t) from each cruise along coastal transect. The locations of discrete samples are shown by black dots.

Figure 3.5 Relationship between mean salinity and mean CH4 concentrations in shallow waters (less than 15 m depth) along the coastal transect (R = 0.61; p = 0.003; n= 22). Error bars denote ±1 standard deviation from the mean of all available measurements between 0 – 15 m depth at each station.

general, the fresh water signature was most apparent during the June cruises (closest to the peak

Fraser River discharge), but there was significant variability in the intensity and spatial extent of this signal. For example, in June 2013 (the year with the earliest peak river discharge), low salinity extended beyond our northernmost sampling station (~ 50 °N). In contrast, the freshwater signatures in the northern region of the transect were more limited during June 2012 and, particularly, in 2014.

In general, maximum N2O concentrations (~ 30 nM) were observed in the deep saline and O2-depleted waters below ~100 m depth, with increasing concentrations towards the north 53

where the influence of Fraser River waters was diminished (Figure 3.4a-e). For most cruises, the low salinity surface waters were associated with relatively low N2O concentrations (minimum values ~8 nM). During June 2012, subsurface N2O concentrations were much lower than any other cruise. A shallow (~ 20 – 40 m depth), subsurface N2O maximum (~22 nM) was also observed in the southernmost stations between September 2012 and September 2013.

In contrast to the distribution of N2O, the lowest CH4 concentrations (~ 6 nM) were observed in the deep saline, O2-depleted waters in the northern section of Vancouver Island. The highest CH4 concentrations (~30 nM) were observed near the sediments in the southern portion of the transect during the two Sept. cruises, indicating the importance of sedimentary CH4 sources in the wide, southern portion of the shelf. A shallow (~ 10 – 80 m depth) subsurface CH4 maximum (~15 nM) was apparent in the northern portion of the transect for all cruises. This feature was strongest during Sept. 2012 and June 2013, and weakest following a period of downwelling in June 2012. The location of this feature was not associated with any appreciable subsurface turbidity maxima. In surface waters (<15 m depth), we observed a positive relationship between CH4 concentrations and salinity along the coastal transect (Figure 3.5, R =

0.78, p= 0.003). No such correlation with salinity was apparent for N2O.

3.4.3.2 Cross-shelf variability

Cross-shelf gradients in hydrography and gas concentrations reflected changes in the transport of deep water masses onto the continental shelf via upwelling. The influence of upwelling on N2O concentrations can be clearly seen in our across-shelf transect (Figure 3.6).

For all cruises, maximum N2O concentrations (> 40 nM) were found in the deep, off-shore waters of the OMZ core (Figure 3.6, b-f). During periods of upwelling, these deep, N2O-rich, 54

and O2-depleted waters appear to be transported onto the continental shelf, resulting in elevated

N2O levels in the mid-shelf waters, particularly near the location of the Tully canyon and the

Juan de Fuca Eddy (Figure 3.6). Indeed, mid-shelf N2O concentrations were highest (>30nM) during the three cruises that followed periods of net upwelling (i.e. Sep-2012, Jun-2013, and

Sep-2013), and we observed a strong positive correlation between mean upwelling intensity and

N2O concentrations in near surface waters (shallower than 50 m) over the shelf (Figure 3.7a). As discussed below, in situ production may also account for the elevated N2O concentrations over the continental shelf.

Methane concentrations along the cross-shelf transect ranged from 1.5 – 104 nM. Unlike

N2O, minimum CH4 concentrations were observed in off-shelf deep waters, while the highest concentrations were observed on the outer shelf region, in the vicinity of known bubble seeps

(indicated by the horizontal black bars in Figure 3.6, g-k; see also Figure 3.1). Even though offshore waters were low in CH4, we did observe a positive relationship between mean upwelling intensity and near-surface water (<50 m) CH4 concentrations over the continental shelf (Figure

3.7b).

3.5 Discussion

Our field data provide new measurements of water column N2O and CH4 distributions and sea-air fluxes along the WCVI shelf, an important, yet under-sampled region. Our work is the first multi-year study from the WCVI upwelling region that includes both surface and depth- resolved water-column measurements of CH4 and N2O. Such depth-resolved data are needed to link the distributions and fluxes of N2O and CH4 to local sources and transportation processes

(e.g. sedimentary diffusion, water column production, upwelling, and freshwater inputs). Our 55

Figure 3.6 Mean pre-cruise upwelling index (90 day average; shown in panel a) and depth sections of N2O, CH4, and O2 along the cross-shelf LC Line

Transect. N2O (panels b-f), CH4 (panels g-k), and O2 (panels l-p). Region with abundant CH4 seeps indicated by horizontal black lines in panels g-k (see

- Figure 3.1 for seep locations). The upper and lower isopycnals used to calculate along-isopycnal changes in N2O, NO3 , O2, and N2O-yields are indicated by black lines in panels b-p.

Figure 3.7 Correlation between mean pre-cruise upwelling indices (90 day average) and average CH4 (a) and

N2O (b) concentrations in shelf waters shallower than 50 m. Average gas concentrations were derived from samples collected within the top 50 m at all on-shelf stations (bottom depth less than 200 m), and error bars indicate ± 1 standard deviation. Both correlations are statistically significant (R> 0.95; p < 0.01; n=5).

results can thus contribute new insight into how variability in oxygen-availability, upwelling intensity, sedimentary processes and fresh water inputs influence N2O and CH4 cycles in the coastal waters of southern British Columbia. In the discussion below, we examine the processes driving the distributions and fluxes of CH4 and N2O along the WCVI.

3.5.1 Sources of CH4

The absence of a negative correlation between salinity and CH4 in our study area indicates that CH4 was not supplied from freshwater or estuarine sources (Figure 3.5).

Considering the nearly 200km distance between the Fraser River (the primary spring / summer freshwater source) and the WCVI, it is likely that much, if not all, of the CH4 in the river water would have been ventilated to the atmosphere before reaching our study area (e.g. Sansone et al.,

1999). Similarly, the lack of correlation between turbidity (beam transmissivity) and CH4 indicates that high CH4 concentrations are not associated with high particle loads or resuspended sediments. Although we cannot exclude or confirm water column production of CH4, the high concentrations of CH4 near sediments, particularly in regions with abundant seeps, suggests that CH4 is supplied primarily from seeps and other sedimentary sources to the water column. To date there have been few direct measurements of dissolved CH4 concentrations in the immediate vicinity of known seeps. We aim to obtain such measurements in future work.

3.5.2 Sources of N2O

Across all of the station depth profiles we examined, N2O exhibited strong correlations

- - with O2 and NO2 +NO3 (Figure 3.8), indicative of nitrification. In contrast, the absence of any

- - concomitant loss of N2O and NO2 +NO3 at low O2 concentrations indicates that denitrification 58

- - Figure 3.8 Relationship between O2, N2O, and NO2 +NO3 across all samples for the 5 cruises. The negative correlation indicates nitrification is the dominant source of N2O in our study region. The absence of

- - decreasing NO2 +NO3 or N2O under low O2 suggests that denitrification is not occurring at appreciable levels in the water column.

and dissimilatory nitrate reduction to ammonia are not likely significant processes in our study region. This is not surprising given that the lowest O2 concentrations in our study area are above the nominal O2 threshold for denitrification (Codispoti et al., 2001). However, shelf O2 concentrations along the WCVI have been shown to fall as low as 30 µM at some times

(Crawford and Peña, 2013), and the continued decline of shelf O2 could allow denitrification to become an important process in the future. 59

Our results suggest that much of the N2O in shelf waters is supplied directly from off- shelf N2O maximum in the OMZ. The observed changes in N2O concentrations along isopycnal surfaces between off-shelf and on-shelf waters of the LC transect support this idea. We can determine how much N2O was supplied by advection vs. on-shelf production by assuming that the total N2O on the shelf is the sum of N2O produced over the shelf and N2O supplied by advection, i.e.;

N2Oon-shelf = N2Oproduced + N2Oadvected (3.2)

Since we assume that N2O is transported along isopycnal surfaces, we would expect on-shelf

N2O concentrations at a given isopycnal to be equal to the off-shelf N2O concentrations at the same isopycnal in the absence of on-shelf sources, such that:

N2Oadvected = N2Ooff-shelf (3.3)

We can rearrange equation (2) to solve for the amount of N2O produced over the shelf, and substitute in N2Ooff-shelf (which we measured), yielding:

N2Oproduced = N2Oon-shelf – N2Oadvected = N2Oon-shelf – N2Ooff-shelf (3.4)

We applied this approach to the concentration vs. density data from our on-shelf (LC04) and off- shelf (LC11) stations (Figure 3.9, below) to estimate that 70-75% of N2O on the shelf was supplied by advection, and the remaining 25-30% was produced in the water column and/or

- - supplied by shelf sediments or mixing. We observed a corresponding increase in NO2 +NO3 and decrease in O2 across the same isopycnals for all cruises except June 2012 (Figure 3.9, b and c, and below), strongly suggesting that nitrification was the process responsible for the excess N2O in shelf waters. The excess N2O in the shelf water was greater during post-upwelling cruises

(dN2O ~10 nM), with the greatest N2O production observed during September 2013. These results suggest an enhancement of nitrification during periods of upwelling. 60

3.5.3 Biological production of N2O

- - Based on the differences in N2O and NO2 +NO3 along isopycnals, we determined the

- - N2O-yield during nitrification (moles of N2O per mole NO2 +NO3 ). Our computed N2O yields ranged from 0.04 – 0.45%, and were highest under low ambient O2 conditions (Figure 3.10). This range of values does not include one negative N2O yield derived from the June 2012 cruise

(discussed below).

The N2O yield of nitrification we observed was largely within the range (0.004 – 0.4%) observed in other marine environments (De Wilde and De Bie, 2000; Punshon and Moore,

2004a), and the 0.25 – 0.31% yields observed in pure cultures of ammonia-oxidizing bacteria by

(Goreau et al., 1980) under similar O2 concentrations (~50 µM). However, the N2O yields we observed were higher than those observed in pure culture of ammonia-oxidizing archaea (0.004 –

0.11%; Santoro et al., 2011; Stieglmeier et al., 2014). This may suggest that N2O was produced mostly by bacterial ammonia oxidizers rather than archaeal ammonia oxidizers in our study area, although we lack data to directly support this idea. Our inferred N2O yields were also higher than the range of N2O yields (0.028-0.04%) previously measured at station LC11 (station P4 on the Line P transect), at the western limit of our study area (Grundle et al., 2012). This discrepancy could reflect a signature of sedimentary nitrification over the continental shelf,

- - which would also supply a high ratio of N2O: NO2 +NO3 to the water column due to the very low

O2 concentrations in sediments.

During the June 2012, cruise, we observed negative N2O yields, resulting from a small

- negative dN2O (~ -2 nM) and positive dNO2 +NO3 (~ +5 µM) between the on-shelf and off-shelf stations (below). N2O loss associated with sea-air exchange could provide one possible

- explanation for this apparent N2O consumption. For example, if upwelled waters (rich in NO2 61

Figure 3.9 Comparison of density-dependent profiles of CH4 and N2O at an on-shelf (LC04, black lines) and

- - off-shelf (LC11, grey lines) station during September, 2012. The changes in O2, NO2 +NO3 and N2O along isopycnals are ascribed to in situ nitrification during the transit of water masses onto the shelf. Figures from additional cruises are included in below.

- +NO3 and N2O) are returned to the sub-surface after a brief surface residence time, gas exchange could ventilate N2O to the atmosphere more rapidly than phytoplankton could consume the dissolved NO3, resulting in an apparent N2O deficit. Alternatively, the rapid downwelling during

2012 could have reduced the residence-time of shelf water, effectively flushing the N2O from the shelf more quickly than it was produced. The strong downwelling observed prior to the

June 2012 cruise is consistent with both of these mechanisms, but we lack definitive data to

firmly establish the cause for the apparent negative N2O yields during this cruise.

Figure 3.10 Relationship between mean O2 concentrations at LC04 (on shelf station) and N2O yields from

- - nitrification. N2O yields were derived from an analysis of N2O and NO2 +NO3 changes along isopycnals (see

Figure 3.9 and methods for details). The negative relationship implies increased N2O yields under low O2 concentrations in our study area. Grey triangles represent mean values derived for each cruise (average of all points interpolated to 0.01 kg m-3 density intervals), with error bars representing ± 1 standard deviation.

Small black diamonds represent the individual calculated points for each cruise.

3.5.4 Dominant transport mechanisms for N2O and CH4

Figure 3.11 presents a simplified schematic diagram illustrating the primary mechanisms likely influencing cross-shelf variability in CH4 and N2O distributions. Upwelling and downwelling are the dominant transport mechanism for CH4 and N2O across the continental shelf. During upwelling favourable conditions, N2O and CH4 are transported along isopycnals towards the coast, though in reality, this transport is not uniform or unidirectional due to short- term variability in tides and upwelling. Moreover, enhanced upwelling in the vicinity of the Juan de Fuca and Tully Canyons appears to increase local supply of N2O and CH4 from deep waters.

As water is transported towards the coast, nitrification acts to increase N2O concentrations, while

CH4 concentrations decrease due to aerobic CH4-oxidation and mixing. Excess N2O and CH4 in near surface waters can be ventilated to the atmosphere via sea-air flux, resulting in elevated fluxes in near shore regions. In contrast, under downwelling conditions, air-equilibrated surface waters low in dissolved CH4 and N2O are transported into the sub-surface and advected off the shelf. Water column and sedimentary nitrification still supply N2O to the sub-surface water column, but likely at lower rates than under upwelling conditions due to reduced supply of organic matter during downwelling. Low CH4 concentrations remain throughout the water column, except near the sediments and seeps where CH4 diffuses into the bottom water before being consumed by water-column methanotrophs.

In the along-shelf transect, the dominant transport mechanism is the VICC, which carries freshwater northward, gradually mixing it with the more saline waters beneath. Here, sediments appear to be a source of CH4 (seeps are less conspicuous along this transect), while the Fraser

River likely supplies only a small amount of CH4 (if any) based on the positive correlation between CH4 and salinity (Figure 3.5). N2O is supplied to the water column by nitrification and 64

the dominant advective source is from deep, N2O-rich marine water. Upwelling appears to bring

CH4 and N2O rich waters from sediments closer to the surface, as well as N2O from hypoxic deep waters. This can be seen in June 2013, where the early onset of upwelling coincided with high subsurface CH4 and N2O concentrations near the southern end of the transect, relative to

June 2012 and June 2014 (Figure 3.4), and by the positive correlations between upwelling and mean CH4 and N2O concentrations in the upper 50 m of shelf water (Figure 3.7). In contrast, strong downwelling appears to transport air-equilibrated CH4 and N2O depleted water into the sub-surface, while simultaneously preventing the shore-ward transport of CH4 and N2O-rich waters. This was observed during June 2012, when anomalously low sub-surface CH4 and N2O

(and high O2) concentrations were observed in the central and northern parts of the coastal transect (Figure 3.4). Taken together, our results thus provide evidence for a critical role of upwelling on sub-surface distributions of N2O and CH4 along the WCVI continental shelf. As discussed below, variable upwelling intensity also influences sea-air fluxes of these gases.

3.5.5 Sea-air fluxes

Sea-air fluxes in our study region (Table 3.1) were much higher than open ocean values

-2 -1 (< 1 µmol m day for both CH4 and N2O; Naqvi et al., 2010). Our maximum N2O flux values

(21.9 µmol m-2 day-1) were significantly lower than the fluxes observed in shelf waters off the coast of Peru (up to 1,800 µmol m-2 day-1; Arévalo-Martínez et al., 2015) and in the Arabian Sea

-2 -1 (up to 3200 µmol m day ; Naqvi et al., 2010), which experience rapid N2O production from both nitrification and denitrification in suboxic (O2< 5µM) near-surface waters (Arévalo-

Martínez et al., 2015; Naqvi et al., 2010). In contrast, surface N2O measurements along the

WCVI (concentrations and sea-air fluxes) were similar to values observed in the surface waters 65

off southern California (~2.3 – 7.4 nM N2O excess; Pierotti & Rasmussen, 1980), in the

-2 -1 Benguela upwelling system (-1.8 – 43.4 µmol N2O m d ; Frame et al., 2014), and higher than

-2 -1 the fluxes (0 -2.2 µmol N2O m d ) observed near Mauritania in NW Africa (Wittke et al.,

2010), where O2 minima are relatively less intense and further from the surface.

Surface CH4 surface supersaturation along the WCVI was higher than previous measurements off the Oregon coast (2 – 7 nM excess CH4; Rehder et al., 2002), but lower than

-2 -1 the surface concentrations and fluxes (~800 nM CH4 and 8.6 – 1300 µmol CH4 m d ) reported from coastal California waters (Coal Oil Point; Mau et al., 2007). These differences (i.e. Coal

Oil Point > WCVI > Oregon Coast) may be due to the different depths of CH4 seeps in these different region. Along the WCVI, seeps are located mainly between 50 - 200 m (Figure 3.1), as compared to 600-800m along the Oregon coast and <70 m in the Coal Oil Point region. Regions with the shallowest seeps (Coil Oil Point) thus appear to have the highest CH4 sea-air fluxes, possibly due to the limited time available for water column CH4 oxidation. Of course, this is likely an oversimplification since many additional factors can affect bubble-mediated CH4 transport to the surface, such as microbial activity, films, water column stratification, and the partial pressure of CH4 in gas bubbles (Mau et al., 2007; Schmale et al., 2011).

Previous field studies have demonstrated that upwelling can exert a strong control on sea- air N2O and CH4 fluxes (Lueker et al., 2003; Nevison et al., 2004; Rehder et al., 2002). For example, a continuous record of atmospheric N2O at Trinidad Head, CA shows strong negative correlations between N2O fluxes and sea surface temperatures (SSTs), with high atmospheric

N2O mixing ratios and low SST occurring during periods of strong upwelling (Lueker et al.,

2003). Similarly, Rehder et al. (2002) found strong correlations between SST, upwelling favorable winds, and surface ocean CH4 concentrations off the coast of Oregon, and argued that 66

Figure 3.11 Schematic diagram showing CH4 and N2O sources, sinks, and physical transport processes along the WVCI under upwelling (panel a) and downwelling (panel b) conditions. Thick blue arrows indicate water circulation, dashed blue lines represent isopycnals and wavy, black lines indicate diffusion gradients.

Upwelling transports N2O-rich waters from the deep N2O max off the shelf, and CH4 from seeps near the shelf break along isopycnals towards the coast. During transport, water column nitrification contributes

- - additional N2O and NO2 +NO3 , while CH4-oxidation mitigates on-shelf CH4 increases. Sedimentary fluxes also increase the water column inventory of CH4 and N2O in shelf waters. The higher on-shelf CH4 and N2O concentrations lead to enhanced sea-air flux of these gases. Under downwelling conditions (panel b), surface waters near air-equilibrium concentrations in O2, CH4 and N2O are carried below the surface near the coast.

Low surface primary productivity (due to limited nutrient supply) results in relatively low rates of water column N2O production from nitrification. Subsurface concentrations gradually increase as water flows away from the coast due to supply from CH4 seeps and the N2O maximum near the shelf-break.

this supports the link between upwelling and increased CH4 flux. We have shown that upwelling is positively correlated with N2O and CH4 concentrations in the upper 50m of the water column

(Figure 3.7), suggesting that upwelling increases the potential for high sea-air fluxes of these gases in our study area. This was corroborated by the significantly higher N2O fluxes following upwelling periods (mean 6.0 ± 4.9 µmol m-2 day-1) relative to post-downwelling sampling (mean

-2 -1 1.5 ± 3.1 µmol m day ; Wilcoxon rank sum, p<0.05). Although CH4 also showed a tendency towards higher fluxes following upwelling (12.6 ± 8.1 vs. 10.5 ± 5.6 µmol m-2 day-1, respectively), the difference was not statistically significant. However, sea-air fluxes were

-2 -1 significantly higher in September than June for both CH4 (15.2 ± 8.9 µmol m day and 9.5 ±

-2 -1 -2 -1 -2 - 4.8 µmol m day , respectively) and N2O (8.6 ± 5.2 µmol m day and 5.9 ± 4.0 µmol m day

1, respectively), highlighting the importance of capturing seasonal variability in WCVI regional flux estimates.

3.6 Conclusion

Our results show that seasonally variable upwelling exerts an important control on CH4 and N2O distributions and sea-air fluxes in the WCVI. Sea-air fluxes and the surface supersaturation of these gases were within the ranges reported for other coastal upwelling systems, with highest values near the coast in the vicinity of the Juan de Fuca Eddy, and strong spatial variability across the continental shelf. Seeps are a potentially significant (albeit localized) source of water column CH4, which are unevenly distributed throughout the region. We determined that upwelling leads to increased subsurface N2O concentrations due to advection of

N2O rich water, and through the enhancement of nitrification in the water column and potentially sediments. The N2O-yield from nitrification along the coast of Vancouver Island appears to be 68

consistent with other studies, and increases under O2-limitation, as expected. The continued decline of O2 (Crawford and Peña, 2013) and intensification of summertime upwelling in the

WCVI and other coastal upwelling systems (Bakun, 1990; Bylhouwer et al., 2013; Wang et al.,

2015) may thus lead to higher CH4 and N2O fluxes, which would act as a positive feedback on climate change.

Our study represents the start of a coastal upwelling time series of N2O and CH4 measurements that may provide valuable insights into longer term changes in the concentrations of these gases and their response to ecosystem changes. The continuation of these time-series observations will provide an important source of information on inter-annual variability in coastal CH4 and N2O cycling. This information is needed to describe the effects of hypoxic upwelling events and long-term changes in upwelling intensity on the distribution and cycling of these gases in the region.

Chapter 4: A multi-year time-series of N2O dynamics in a seasonally anoxic fjord: Saanich Inlet, British Columbia.

4.1 Summary

We present an 8-year time-series of monthly water column N2O measurements from

Saanich inlet, a seasonally anoxic fjord on the eastern coast of Vancouver Island, British

Columbia. Our observations document regular seasonal cycles in N2O concentrations driven by physical and biological forcing, with significant inter-annual variability. Near-surface N2O concentrations were typically supersaturated, and increased with depth to a maximum near the oxic-anoxic transition where multiple N2O sources such as nitrification, nitrifier-denitrification, and partial denitrification could co-occur. Genes associated with both the nitrification and denitrification pathways were widely distributed throughout the water column suggesting the potential for simultaneous N2O production from nitrification and incomplete denitrification. We observed consistently higher N2O concentrations and sea-air fluxes during the late summer/early fall, coincident with the period of deep basin renewal and enhanced nitrification and denitrification gene abundance. In most years, deep basin renewal during the late summer led to an increase in bottom water N2O, which was largely attributable to lateral water mass advection, rather than enhanced in situ production. Subsequent denitrification in the deep basin completely consumed N2O by late spring. There was a strong negative relationship between N2O and O2 concentrations in the water column, with evidence of N2O consumption at O2 levels below

~10µm, and a seasonally variable N2O:O2 slope likely reflecting apparent O2-dependent changes in N2O yields. Our results characterize the seasonal and interannual variability in N2O

concentrations that can be used as a baseline for identifying and studying the effect of longer- term changes, such as declining O2-availability, on N2O cycling. Identifying and studying these changes in Saanich Inlet could help improve our understanding of these changes on local and ocean basin scales.

4.2 Introduction

Nitrous-oxide (N2O) is an important greenhouse gas, and a catalyst of stratospheric ozone destruction (IPCC, 2013; Portmann et al., 2012). Atmospheric N2O concentrations have risen

~20% above pre-industrial levels due to anthropogenic activities such as fertilizer application and fossil fuel use (IPCC, 2013), and have also fluctuated over natural glacial-interglacial cycles

(Sowers et al., 2003; Wolff and Spahni, 2007). Roughly one-third of all naturally-derived N2O in the atmosphere is emitted from the oceans (Seitzinger et al., 2000), with disproportionately high contributions from estuaries and eastern-boundary upwelling systems, where high primary productivity fuels rapid O2-demand and N2O production in near-surface waters (Bange et al.,

1996; Nevison et al., 2004).

In both marine and terrestrial systems, N2O is produced primarily by nitrification and

+ - denitrification. Nitrification, the aerobic oxidation of ammonium (NH4 ) to nitrate (NO3 ) via

+ - hydroxylamine (NH3OH ) and nitrite (NO2 ), yields N2O as an intermediate by-product and primarily drives chemoautotrophic microbial growth. This process, which is believed to be the dominant source of N2O throughout the world's oceans, is restricted to oxygenated waters since it requires O2 as an electron acceptor (Ward, 2011). Nitrification has long been considered a distributed process mediated by distinct microbial groups including ammonia oxidizing bacteria

+ - (AOB) and archaea (AOA) mediating NH4 conversion to NO2 , and nitrite oxidizing bacteria 71

- - (NOB) mediating conversion of NO2 to NO3 (Costa et al., 2006). More recently, Nitrospira

- bacteria have been identified with the capacity to drive ammonia oxidation (comammox) to NO3 for energy conservation, although the environmental distribution of this process remains to be determined (Lam and Kuypers, 2011; van Kessel et al., 2015).

The strong linear relationship between apparent oxygen utilization (AOU) and N2O concentrations observed in many sub-surface ocean waters is taken as an indicator of N2O production by nitrification (Cohen and Gordon, 1979). However, the slope of the N2O:AOU relationship is regionally variable, owing to differences in the nitrogen content of sinking organic matter, mixing of different water masses, or O2-dependent changes in the relative N2O yields from various sources (Nevison et al., 2003). The N2O yield of nitrification is known to increase under O2-limitation (Frame and Casciotti, 2010; Goreau et al., 1980), while N2O production by

- nitrifier denitrification (NO2 reduction to N2O by nitrifiers) is more permissive under O2- limitation (Frame and Casciotti, 2010; Poth and Focht, 1985). Both of these factors could potentially alter the N2O:AOU relationship, with higher N2O production expected under suboxic conditions.

- - Denitrification – the step-wise reduction of NO3 to NO2 , to NO, to N2O, and N2 – can be

- a net-source of N2O under low O2 conditions, but prolonged anoxia in the absence of NO3 or

- NO2 ultimately leads to net N2O-consumption (Betlach and Tiedje, 1981; Dalsgaard et al., 2014).

The last step of denitrification, which is believed to be sole biological N2O sink in the oceans, is restricted to anoxic and near anoxic waters (O2≤ ~5µM), though the O2 threshold for denitrification is not clearly defined, and appears to be higher in natural waters than in laboratory culture experiments (Betlach and Tiedje, 1981; Castro-González and Farías, 2004). This may result from the presence of naturally occurring micro-anoxic environments inside sinking 72

particles and organisms. Compared to nitrification, the metabolic capacity for denitrification is extremely widespread and distributed among many taxonomic groups. In marine oxygen- deficient waters, bacteria affiliated with the SUP05 group have been implicated in sulfur oxidation coupled to incomplete denitrification resulting in the production of N2O (Lavik et al.,

2009; Walsh et al., 2009).

Significant research has focused on determining the relative contributions of nitrification and denitrification to regional N2O budgets (Castro-González and Farías, 2004; Sutka et al.,

2006; Yamagishi et al., 2005), and the O2 sensitivity of these processes (Betlach and Tiedje,

1981; Bianchi et al., 2012). Biogeochemical observations, along with incubation-based rate measurements have been used to identify the dominant sources of N2O in various ocean regions.

Isotopic evidence from the western Pacific indicates that N2O is mostly derived from denitrification, rather than nitrification (Yamagishi et al., 2005; Yoshida et al., 1989). In contrast, isotopic data point to a dominant nitrification source of N2O in the North Pacific shallow N2O maximum (Dore and Karl, 1996), while coupled nitrification-denitrification has been identified as a key N2O production term in the Arabian Sea (Naqvi et al., 1998). Incubation experiments conducted in coastal Chilean waters with selective inhibitors showed ~20-80% reduction in net

N2O production when nitrification was inhibited, indicating that both nitrification and denitrification were important sources of N2O (Farias et al., 2009). Additional studies in this region demonstrated that the N2O yields of denitrification increased under suboxic conditions

(O2<20µM), highlighting the O2 sensitivity of net N2O production (Castro-González and Farías,

2004).

Quantifying the O2-dependence of N2O cycling is particularly important in light of the recently documented expansion of marine oxygen-deficient waters (Stramma et al., 2008; 73

Whitney et al., 2007), which could promote higher marine N2O concentrations and drive biogeochemical climate feedbacks (Codispoti, 2010). Understanding the effects of O2 availability on marine biogeochemical processes requires field study sites with regular and predictable changes in oxygen concentration over a prolonged period of time. To date, such time- series work examining N2O cycling has been limited to a few low O2 regions, such as the Indian continental shelf (Naqvi et al., 2009), coastal Chile (Farías et al., 2015), The subtropical North

Pacific (Karl and Lukas, 1996), and the shallow, seasonally-anoxic waters of Eckenförde Bay,

Germany (Lennartz, 2013). Bange et al. (2001) compiled surface N2O measurements from the

Arabian Sea between 1977 and 1997, and found strong seasonal forcing driven by the monsoon, which enhances subsurface hypoxia in the coastal waters, resulting in rapid N2O production and increased sea-air fluxes. A ten-year time-series of monthly N2O measurements from coastal

Chile (Farías et al., 2015) demonstrated high N2O concentrations associated with O2-depleted waters during the productive upwelling season, and the intermittent appearance of small-scale short-lived N2O-rich water masses in the vicinity of spatially-constrained O2-gradients. These small-scale, ephemeral features highlight the need for high frequency sampling in time-series programs.

For close to a decade (since 2007), our group has been conducting time-series monitoring in Saanich Inlet, a seasonally anoxic fjord on the eastern coast of Vancouver Island, British

Columbia, Canada (Figure 4.1). During the summer months, strong tidal mixing in this system drives the upwelling of nutrient rich sub-surface water (down to 60 m), fueling high surface primary production on a fortnightly cycle (Gargett et al., 2003). The high surface primary production supplies organic matter to the deep basin waters, which are isolated during much of the year by a 70m deep glacial sill (Figure 4.1b) that restricts O2 supply through lateral 74

advection. As a result, decomposition of organic matter leads to O2 deficiency in the deep basin, creating a highly reducing environment with significant accumulation of N2, CH4, and H2S

(Anderson and Devol, 1973; Lilley et al., 1982; Manning et al., 2010). The deep basin is ventilated during the fall (and to a lesser extent during spring), when tidal currents allow dense, oxygenated, nutrient-rich water to flow over the sill and displace the resident deep water

(Anderson and Devol, 1973). Given its strong and regular cycles of deep basin anoxia and renewal, Saanich Inlet provides a model ecosystem in which to study the effect of seasonal O2- availability on N2O cycling.

The microbial community in oxygen-deficient waters of Saanich Inlet is broadly representative of other open ocean oxygen-minimum zones (OMZs) and anoxic regions, highlighting the utility of the inlet as a model for microbiological and biogeochemical studies of marine oxygen-deficient waters (Chow et al., 2015; Hawley et al., 2014; Labonté et al., 2015;

Roux et al., 2014; Wright et al., 2012; Zaikova et al., 2010). Previous work in Saanich Inlet has

+ examined various aspects of water column N-transformations, including NH4 -oxidation

— (Grundle and Juniper, 2011; Ward and Kilpatrick, 1990), NO2 oxidation (Grundle and Juniper,

2011) and N2 production by denitrification (Manning et al., 2010). These studies have highlighted the rapid cycling of redox-active nitrogen species in Saanich Inlet, but have not provided measurements of N2O. To our knowledge, the only published N2O measurements in this system are two profiles measured during July 1977 (Cohen, 1978). These observations provided the first evidence of complete N2O consumption in an anoxic water column through denitrification, and also demonstrated strong sea-air fluxes of N2O from Saanich Inlet surface waters.

Here, we document the seasonal, inter-annual and depth-dependent variability of N2O 75

concentrations in Saanich Inlet, to better understand the biogeochemical controls on N2O cycling and their response to changing O2 availability over time. Our observations, which span nearly a decade, enable us to examine variability in N2O concentrations linked to regional forcing including El-Nino and altered hydrography and physical transport processes. In addition to chemical measurements, we use metagenomic data to identify the relative abundance and variability of genes involved in N2O production in relation to water column N2O measurements.

Our results demonstrate a tight coupling between N2O concentrations, O2 dynamics and regional hydrography, with variability observed over multiple time-scales. Results from this study underscore the utility of time-series measurements in capturing temporal changes in coastal biogeochemical processes.

4.3 Methods

Beginning in February 2007, monthly sampling was conducted from the MSV Strickland at a single station near the deepest part of Saanich inlet (SI03, 48.59N, 123.51W; Figure 4.1).

Measurements of in situ temperature, salinity, photosynthetically active radiation (PAR), and transmissivity were obtained from a conductivity-temperature-depth (CTD) sensor (SBE 25) equipped with an oxygen sensor (SBE 43) and PAR sensor (Biospherical Instruments QSP-

200PD). CTD O2 measurements were calibrated against discrete Winkler O2 measurements from each month.

- - 3- + Discrete samples for measurements of macro-nutrients (NO3 , NO2 , PO4 , Si and NH4 ), hydrogen-sulfide (H2S), and dissolved gases (including O2 and N2O) were collected using 8 or

12 L GO-FLO bottles. Samples for macro-nutrients were collected through 0.2 um syringe filters and stored frozen in the laboratory prior to measurement by air segmented continuous-flow 76

colorimetric analysis using a Bran Luebbe Auto Analyzer 3 (Armstrong et al., 1967).

+ Ammonium (NH4 ) was measured in freshly collected samples using the fluorometric method of

Holmes et al. (1999). Reagents were added to samples, standards and blanks in the field, and the resulting fluorescence signal was measured in the laboratory within 8 hours. Ship-board analysis for hydrogen-sulfide (H2S) was conducted with freshly collected samples according to Cline

(2015).

Duplicate samples for dissolved N2O analysis were collected by transferring water from

GO-FLO bottles to 60 mL serum vials using silicon tubing. Sampling bottles were over-filled three times without introducing air bubbles, and 100µL saturated HgCl2 solution was added to stop microbial activity. The vials were subsequently crimp-sealed and stored at 4°C until analysis by purge-and-trap gas-chromatography mass-spectrometry (Capelle et al., 2015). This method provides analytical precision of 3% or better, and we observed an average standard deviation of 16% between replicate samples. The accuracy of our method is ensured by calibration against a NOAA reference standard, and by inter-comparison against measurements from other laboratories (see Capelle et al. 2015 for details).

The sea-air flux of N2O was calculated from the excess mixed layer gas concentration, relative to atmospheric equilibrium, (ΔC) and a weighted piston-velocity (kw):

Flux = (Cobs - Ceq) * kw (4.1)

Mixed layer depths, defined as the depth with a density of >0.125 kg m3 greater than the density at 1m, were typically very shallow, with a mean of 3.7± 3.6m. Since our shallowest standard sampling depth was 10m for much of the time series, we applied a correction factor to the 10m N2O samples to estimate mixed layer concentrations. The correction factor was

Figure 4.1 Map of Saanich Inlet, BC, showing location of sampling (star, panel a), and location on southeast

Vancouver Island (inset). Depth transect along thalweg of Saanich Inlet (panel b) shows elevated sill near mouth of inlet. Adapted from Anderson and Devol, (1973). 78

determined by the mean ratio of N2Osurface:N2O10m from those cruises where we had both a mixed layer and 10m N2O sample (0.84 ± 0.07; n=12). For the few samples where N2O was under- saturated at 10m depth (n=11 out of 113 total), we assumed that the mixed layer concentration was equal to the concentration at 10m. Excess concentrations were calculated from the difference between mixed layer concentrations (Cobs) and the equilibrium concentrations (Ceq) calculated from the temperature and salinity of the mixed layer (Weiss and Price, 1980). Air-equilibrium

N2O concentrations were derived from mean monthly atmospheric N2O mixing ratios measured at Barrow, Alaska (Data provided by NOAA/ESRL halocarbons in situ program; http://www.esrl.noaa.gov/gmd/dv/data/index.php?site=brw). The weighted piston-velocity (kw) was determined from the wind-speed/kw relationship derived by (Nightingale et al., 2000), and mean daily wind speeds from the two-week period leading up to sample collection, according to

Reuer et al. (2007). Briefly, the weighted scheme is derived from the daily derived piston velocity and the proportion of the mixed layer ventilated for each of the 14-days leading up to sample collection (see Reuer et al., 2007 for additional details). Wind-speeds were obtained from the Environment Canada meteorological buoy in Saanich Inlet (buoy c46134).

Water for metagenomic analysis was collected using GO-FLO or Niskin bottles from selected depths (10m and between 100-200m) from 14 time points using sterile silicon tubing and filtered through 0.22 um Sterivex polycarbonate filters. Genomic DNA was extracted from the filters using the methods of Zaikova et al. (2010). The extracted DNA was used to generate metagenomic datasets as described in Hawley et al. (submitted). Metagenomic sequence information provides a genetic inventory that can be used to reconstruct the metabolic potential of microbial communities. This potential can be rendered in the form of diagnostic functional genes, enzyme complexes or entire metabolic pathways driving specific biogeochemical 79

transformations such as nitrification or denitrification. The functional genes we examined play key roles in ammonia oxidation (amoA), archaeal nitrifier-denitrification (nirK-arc), nitrate reduction (napA/narG), nitrite reduction (nirK/nirS), nitric oxide reduction (norB) and nitrous oxide reduction (nosZ). Genes were identified in the metagenomic datasets using the

Metapathways pipeline (Konwar et al., 2013). Resulting KEGG and NCBI RefSeq (2014) annotations were cross-validated by clustering protein sequences at 80% identity. Additionally, the RefSeq taxonomy for genes annotated as nitrate reductase (narG) was used to further differentiate between nitrate reductase (present in multiple taxonomic lineages) and nitrite oxidoreducase (present only in Planctomycetes and nitrite oxidizing groups). We used the unit of read fragments per kilobase mapped (RPKM) (Konwar et al., 2015) as a proxy for gene abundance. For complete method details, see Hawley et al. (submitted). Although our 83 metagenomes (~20GB of assembled metagenomic sequence data) represent a large dataset, the spatial and temporal coverage of these molecular observations is limited relative to the nearly

1,500 sampling points represented in the current study.

4.4 Results and discussion

4.4.1 Overview

Nitrous-oxide concentrations in Saanich Inlet showed significant temporal and depth- dependent variability over the course of our time series, with values ranging from <0.5 nM to

37.4 nM and a mean value of 14.7nM. The maximum concentration of N2O in Saanich Inlet is comparable to other land-locked anoxic basins, which range from 20-60 nM N2O (e.g. Baltic

Sea, Black Sea and Cariaco Basin; see Naqvi et al., 2010), but somewhat lower than that reported 80

for a number of hypoxic to anoxic waters in the open ocean (~80 nM in ETNP, Gulf of

California, ETSP, Arabian Sea and Bay of Bengal; see Naqvi et al., 2010).

As expected, seasonal processes and O2-dependent gradients accounted for much of the observed variability in N2O concentrations. We also identified significant inter-annual variability in N2O concentrations. Genes involved in N2O cycling processes, including nitrification (amoA, nxr), archaeal nitrifier-denitrification (nirK-arc) and various steps in the denitrification pathway

(napA/narG, nirK/nirS, norB, nosZ), were widespread throughout the Saanich Inlet water column, and typically increased in abundance during the period of deep-basin renewal in late summer into fall. However, we found that the abundance of these functional genes was not correlated with any of the measured chemical variables (e.g. oxygen, nutrients, temperature, salinity or N2O concentrations). In the following sections, we consider three different regions of the water column including the euphotic zone, oxycline, and deep basin and discuss patterns in

N2O concentrations, hydrography and functional gene abundance. Each of these regions is subject to a unique set of physical, chemical and biological influences, resulting in distinct N2O dynamics.

4.4.2 Seasonal variability

4.4.2.1 Surface waters

The euphotic zone represents the upper ~30m of the water column (based on PAR data), where primary production dominates biogeochemical cycles during the summer, and nitrification is believed to be light-inhibited (Grundle and Juniper, 2011). Estuarine circulation periodically transports water into the euphotic zone from the upper oxycline (Gargett et al., 2003). Figure 4.2 81

- shows the mean annual cycle of O2, NO3 , N2O, and H2S derived from all of our individual monthly measurements in the euphotic zone. Variability in this upper layer was driven by near- surface primary production and physical circulation. The summertime estuarine circulation supplies nutrient-rich sub-surface water to the surface, promoting summertime phytoplankton blooms. This same mechanism also likely supplies N2O to the surface, as positive correlations

- between subsurface NO3 and N2O concentrations have been previously documented in the region (Capelle and Tortell, 2016). We consistently observed high Chla fluorescence values in the upper ~30m of the water column between April and September, indicative of increased phytoplankton abundance (data not shown). These spring/summer phytoplankton blooms

- account for the seasonal reduction of NO3 and O2 accumulation in surface waters (Figure

4.2a,b). Euphotic zone N2O concentrations were typically lowest during the winter, and increased over the course of the summer, reaching a maximum in late summer/early fall (Sept. -

Oct.). This seasonal pattern could result from increased in situ N2O production, reduced ventilation of N2O during the summer, or transport of N2O rich waters into the euphotic zone.

Although nitrification is often assumed to be light-inhibited, several authors have demonstrated nitrification as a source of N2O in the euphotic zone of Saanich Inlet (Grundle and Juniper, 2011;

Ward and Kilpatrick, 1990). We detected nitrification genes (amoA and nxr) in euphotic zone waters (10m, Figure 4.8), reinforcing the potential for euphotic zone N2O production from nitrification. Seasonal variability in nitrification rates could therefore contribute to euphotic zone

+ N2O variability, with maximum rates during the summer time when NH4 oxidation rates are presumably highest. However, Grundle and Juniper (2011) observed little variability in nitrification rates in the upper 120m of Saanich inlet between April and October (note they did not measure rates during winter). This observation suggests that N2O variability over the spring 82

and summer was not primarily driven by seasonal changes in in situ production. As discussed below, our results suggest that the upward mixing of oxycline waters during late summer likely accounted for much of the observed increase in euphotic zone N2O concentrations, while increased sea-air N2O fluxes likely reduced euphotic zone N2O concentrations during the fall and winter.

As shown in Figure 4.3, the accumulation of N2O in the euphotic zone and oxycline during summer was associated with increased N2O sea-air fluxes between May and Sept.

(maximum monthly mean of ~ 4 µmol m-2 d-1in September). The seasonal average sea-air fluxes exhibited significant inter-annual variability, with values ranging from -1.5 – 9.9 µmol m-2 d-1, and an overall mean of 1.8 µmol m-2 d-1 (positive numbers indicate transfer from the surface ocean to the atmosphere). On average, sea-air flux removed 5% of the total N2O inventory from the Saanich Inlet mixed layer per day, but this increased to 10% per day during Sep-Nov. This suggests that air-sea exchange was an important sink of N2O from near surface waters during fall.

Somewhat surprisingly, we found evidence for potential biological consumption of N2O in surface waters. In particular, we detected the presence of the N2O reductase gene (nosZ) in

10m waters (Figure 4.8) suggesting that microbes in the mixed layer possessed the metabolic capability for N2O consumption. It has typically been assumed that biological N2O reduction

(which occurs as part of the denitrification pathway) only occurs under suboxic or anoxic conditions. However, it is possible that low O2 conditions in gut micro-biota or within particles create micro-environments favourable for this process. In Saanich Inlet, denitrification or partial denitrification inside micro-anoxic environments in oxygenated water appears possible given the high productivity rates and sediment flux in the Inlet (Timothy and Soon, 2001; Timothy et al., 83

Figure 4.2 Mean annual cycle of dissolved oxygen (a), nitrate (b), nitrous-oxide (c), and hydrogen-sulfide (d) derived from monthly 2007-2014 water column measurements in Saanich Inlet.

2003), and this process has been measured in waters with high suspended sediment loads at

~90% of saturation O2 values (Liu et al., 2013), and in marine sediments at up to 90 µM O2 (Gao et al., 2009). We note that the identification of functional genes in the metagenome does not necessarily translate into metabolic pathway activity. Moreover, in the absence of direct rate measurements, we are currently unable to estimate the potential contribution of biological N2O reduction in Saanich Inlet surface waters. Further measurements of biological activity of N2O reduction in surface waters including process rates and gene expression in the water column and suspended particles would help to elucidate the activity of N2O reducing microorganisms and their role in controlling surface water N2O dynamics.

4.4.2.2 Mean seasonal cycle in the oxycline

The oxycline is typically situated between 30 and 120m, and represents a region of strong depth-dependent O2 gradients associated with high rates of nitrification and organic matter

remineralization (Grundle and Juniper, 2011). In this depth interval, estuarine circulation supplies water from Haro Strait during the spring and summer. The seasonal cycle of N2O in oxycline waters (between ~30 and 120 m depth) differed significantly from that observed in the

- euphotic zone. In oxycline waters, the distribution of O2, NO3 , and N2O was likely influenced primarily by the seasonal cycle of nitrification. By late summer/early fall, O2 and N2O in the upper 120m reached their respective minima and maxima, before gradually increasing and decreasing, respectively, over the less productive winter months (Figure 4.2a,c). The summertime N2O maximum could be driven by rapid nitrification and reduced O2 availability, which could have enhanced N2O yields from nitrification and partial denitrification. A previous study demonstrated very high rates of nitrification between April and October, which should supply large amounts of N2O to the oxycline (Grundle and Juniper, 2011).

Conversely, N2O declines in the oxycline during light-limited winter months could be driven by lower primary production, which should reduce the supply of NH3 from sinking organic matter to the oxycline, thus reducing nitrification and N2O production in these depths.

+ Indeed, we observed significantly lower NH4 concentrations at the base of the euphotic zone

+ + (40m) during winter (0.1 ± 0.1µM NH4 ) relative to summer (2.9 ± 2.8µM NH4 ), suggesting a restricted substrate supply for nitrification during winter. A reduction in stratification and enhanced mixing during winter could also help to bring N2O concentrations in the oxycline closer to air equilibrium values, by exposure to the overlying atmosphere. Ward and Kilpatrick

(1990) found a significant reduction of CH4 concentrations in the oxycline during the fall, which they attributed to increased ventilation of subsurface waters through surface cooling and increased storm intensity. Similarly, we consistently observed surface cooling and isopycnal shoaling during the fall period (not shown), and this decreased stratification appeared to be the 86

primary driver of N2O ventilation during fall and winter, with the maximum in sea-air N2O fluxes (Figure 4.3) observed during the fall.

We observed a widespread abundance of nitrification genes across strong O2 gradients between 100-200m, with the highest abundances of amoA (NH3-oxidation genes) near the base

- of the oxycline (between 100-120m, Figure 4.8), whereas nxr (NO2 -oxidation genes) were evenly distributed between 100-200m (Figure 4.8). Denitrification genes were also relatively abundant near the base of the oxycline (100-120m) throughout the time series (though not as abundant as in the deep basin), suggesting potential N2O production from both nitrification and partial denitrification. Both nitrification and denitrification gene abundances were typically highest during the deep water renewal period in the lower oxycline (100-120m, Figure 4.8), consistent with maximum N2O concentrations in the oxycline. Nitrite oxidation genes (nxr) also increased during Jan-Feb, when moderate intensity renewals can occur down to ~150m

(Manning et al., 2010).

4.4.2.3 Mean seasonal cycle in the deep basin

In the deep basin, below ~120m, lateral advection is restricted for most of the year due to the sill at the mouth of Saanich Inlet, and renewal during late summer and fall dominates the physical circulation. The lack of circulation for much of the year allows anoxia to develop and denitrification to occur, while H2S from sedimentary sulfate reduction accumulates in the water column. Prolonged microbial degradation of sinking organic matter gradually consumed the

- available O2, NO3 and N2O over the winter, spring, and summer (Figure 4.2), and these changes in redox chemistry exert a strong influence on N2O cycling in the deep basin, where N2O concentrations exhibited significantly higher seasonal variability than either the euphotic zone or 87

oxycline.

- After consumption of all available NO3 through denitrification, sulfate-reduction (likely within sediments) is used to fuel organic matter remineralization, resulting in the accumulation of water column H2S during the mid to late summer period (Figure 4.2d). Deep water renewal during late summer (Sep-Oct) led to the rapid oxidation of H2S, and the accumulation of O2,

- - NO3 and N2O. The accumulation of O2 was typically more short-lived than NO3 , due to O2 consumption during oxidation of H2S and other reduced species. During renewal events, higher

- density water appeared to subduct beneath the resident anoxic water, displacing low O2, NO3 and N2O waters into the mid-depth water column, and leading to a shoaling of the oxycline

(Figure 4.2a). The inflowing renewal water was a net source of N2O to deep basin waters below

150m, and this lateral transport, rather than stimulated in situ production, was likely the dominant process leading to increased N2O concentrations in deep basin waters following renewal events. Support for this assertion comes from the similarity in N2O concentrations in

Haro Strait (the source of renewal waters), and deep basin waters of Saanich post-renewal (data not shown).

Denitrification genes (napA/narG, nirK/nirS, norB, and nosZ) were typically most abundant during renewal periods in the deep basin, and declined over the winter and spring along

- - with NO3 , NO2 and N2O concentrations (Figure 4.7 and Figure 4.8). The abundance of denitrification genes in the deep basin waters suggests that denitrification could be rapidly initiated following a return to suboxic or anoxic water column conditions (Figure 4.7 and Figure

4.8). A delay in the onset of N2O reduction has been suggested as a potential explanation for the short-lived N2O accumulation (up to ~1500 nM; Naqvi et al., 2000; Rönner, 1983) observed in some marine systems with transient oxic-suboxic transitions (Codispoti et al., 2001; Jayakumar 88

et al., 2009). This short-lived N2O accumulation may result from changes in the microbial interactions driving a distributed denitrification pathway. Consistent with this idea, a recent study of size fractionated microbial communities in the Chilean OMZ identified an overabundance of nosZ transcripts on particles (Ganesh et al., 2015). In Saanich Inlet, SUP05 can reach up to 50% of total bacteria (Walsh and Hallam, 2011) expressing ~50% of total denitrification proteins

(with the exception of nosZ) (Hawley et al., 2014). During our sampling, we observed that nosZ was nearly ubiquitous in the Saanich Inlet oxycline and deep basin waters, and correlated (R =

0.56 to 0.75, n = 76) with other denitrification genes (nirK/nirS and norB). This correlation suggests direct metabolic coupling between SUP05 and other (as of yet unidentified) microbial community members involved in the distributed denitrification pathway. This metabolic coupling helps to explain the lack of massive N2O accumulation in Saanich Inlet.

4.4.3 Depth resolved variability

To better illustrate the depth-dependent seasonal changes in Saanich Inlet, Figure 4.4

- shows the mean depth profiles of N2O, O2 and NO3 (± one standard deviation) during three seasonal periods: post-renewal/Fall (Sep-Dec), winter (Jan-Apr), and summer (May-Aug; Figure

- 4.4). The effect of summertime estuarine circulation and primary production on O2, NO3 , and

N2O depth profiles can be seen in the euphotic zone and oxycline. Primary production accounts

- for a ~15 µM drawdown of NO3 _in the euphotic zone relative to winter, and a 20% increase in

O2% saturation. N2O concentrations in the upper 100m are ~20-40% higher during fall than winter or summer, and the N2O maximum shoals from 100m to 75m between summer and fall.

Figure 4.4 also demonstrates that the N2O maximum was typically found within the steepest part of the oxycline, where O2 concentrations decreased by as much as 90 µM over a 40m depth 89

range. Steep O2-gradients are known to be sites of intensive microbial activity (Edgcomb and

Pachiadaki, 2014). The high N2O concentrations in this portion of the oxycline (up to 38nM) may thus result from multiple sources, such as nitrification, nitrifier-denitrification, and

- incomplete denitrification. The mean annual cycle of O2, NO3 , and N2O in the deep basin

(below 120m), is also clearly evident in Figure 4.4. Variability is driven by gradual consumption of these chemical species during winter and summer, and resupply by renewal in late-

- summer/fall. Denitrification caused both NO3 and N2O to decrease with depth in the deep basin,

- but it is important to note that NO3 concentrations began to decrease at depths 10-20m shallower than N2O. This observation suggests that the rate of N2O consumption by denitrification was slower than N2O production by nitrification and/or partial denitrification.

We observed genes associated with both nitrification and denitrification throughout the water column, but no significant correlation (either positive or negative) between the genes of these respective pathways. This result indicates that the potential for nitrification and denitrification is widely distributed with potential process overlap under different water column redox conditions. An overlap of nitrification and anammox has been measured in the Peruvian upwelling system (Lam et al., 2009) and proteins from the archaeal amo gene cluster have been observed under the same conditions as anammox and denitrification proteins in Saanich Inlet

(Hawley et al., 2014) with implications for biological control over nitrogen loss processes. The widespread distribution of nitrification and denitrification genes in the Saanich Inlet water column underscores the importance of integrating process rates and multi-omic sequence information (DNA, RNA, protein) to differentiate between metabolic potential and water column activity.

Note the logarithmic scale of O2.

4.4.4 Relationship between O2 and N2O-cycling

- The discussion above highlights the relationship between O2, NO3 and N2O concentrations in Saanich Inlet. Figure 4.5 presents the relationship between O2 and N2O concentrations across all depths and sampling dates over the course of our time-series. As shown in this figure, we observed a linear relationship between apparent oxygen utilization (AOU) and

N2O for AOU values less than 270µM. This relationship shows a large degree of scatter, which results partly from seasonal variability in N2O:O2 stoichiometry, as discussed below. Based on molecular evidence indicating that nitrification dominates in Saanich Inlet above ~ 1µM O2,

(Hawley et al., 2014), we ascribe the increase in N2O concentrations between 0 and ~270 µM

AOU (Figure 4.5a) as a primary result of nitrification. Above this AOU threshold, denitrification consumes N2O without affecting O2, and this results in the rapid decline in N2O concentrations with increasing AOU (open symbols in Figure 4.5a).

The N2O - AOU relationship observed in a number of marine systems has been used to estimate N2O concentrations from O2 measurements, and the N2O-yield of nitrification (Cohen and Gordon, 1979; Nevison et al., 2003). To compare our observations with other datasets, we plotted the N2O:AOU slopes from other marine systems on Figure 4.5a, with a correction for differences in y-intercepts among the various data sets. This compilation shows that N2O:AOU relationships are regionally variable, yielding best-fit lines with slopes ranging from 0.01 to 0.3 nmol N2O : µmol AOU across different oceanic regions (Figure 4.5a). The line of best fit through our data (based on the points between 0-270µM AOU) is most similar to that observed in the Bedford Basin, a semi-enclosed, periodically anoxic basin (Punshon and Moore, 2004a).

By comparison, the slope of our N2O:AOU relationships was significantly lower than the that observed in the Arabian Sea and subtropical N. Pacific, suggesting that N2O yields are higher in 92

Figure 4.5 Relationship between N2O and apparent oxygen utilization (AOU) derived from monthly measurements in Saanich Inlet (a). Grey circles show measurements where AOU<270 µM, hollow circles are for AOU>270 µM. Linear regression for AOU<270uM is shown by thick black line.

Several N2O:AOU relationships from different ocean regions are plotted with the same y-intercept for reference. Panel b shows the mean N2O concentration (± one standard deviation) in 1 µM O2 bins between 0 and 20 µM O2, and in 5 µM bins between 20-300 µM O2. Panel b inset shows a curve fit to the individual O2 vs N2O measurements for O2 concentrations < 30 µM.

these systems than in Saanich Inlet. Regional differences in N2O:AOU and N2O:O2 may be related to changes in N2O contributions from nitrification, denitrification and nitrifier denitrification, which are sensitive to O2 availability. In addition, the effects of changing water temperature on O2 solubility following ventilation, or external N2O inputs/losses associated with mixing of different water masses (see Nevison et al., 2003) could drive regional variability in the

N2O:AOU relationship. N2O:AOU ratios are generally low in the Atlantic where O2 concentrations are high. In contrast, O2-depleted Pacific waters tend to have higher N2O:AOU ratios due to higher N2O yields and additional sources of N2O, such as partial denitrification

(Nevison et al., 2003).

To determine the O2 concentration thresholds separating net N2O production from consumption, we calculated the mean N2O concentration in 1µM O2 bins from 0-20µM O2, and 5

µMO2 bins from 20-300µM O2 across all depths and sampling periods (Figure 4.5b). As shown in Figure 4.5b, there is a strong linear decrease in N2O concentrations between ~ 20 and 300

µMO2. As shown in the inset figure, mean N2O concentrations decline sharply with decreasing

O2 below ~ 10µM O2, indicating that this is the threshold below which N2O consumption outstrips production from nitrification and partial denitrification. The wide scatter in our

N2O:AOU and N2O:O2 data (Figure 4.5a,b) suggest variable stoichiometry between N2O production and O2 consumption in Saanich Inlet. To investigate potential sources of this variability (Figure 4.5), we examined seasonal changes in the slope of the depth-dependent

N2O:O2 relationship. For each monthly depth profile, we determined the slope of the N2O:O2 relationship for all discrete samples with O2>30µM. As shown in Figure 4.6, the N2O:O2 slope became progressively steeper (i.e. increasing net yield of N2O per mole O2 consumed) over the

-1 summer months, with a maximum observed during October (-6 nmol N2O µmol O2 ). 94

Thereafter, the slope relaxed through the fall and winter, eventually reaching a value near zero

(i.e. little/no net N2O production). The seasonal difference in the N2O:O2 slope may be attributable to changes in water column O2 concentrations, with lower O2 concentrations during summer driving higher N2O-yields from nitrification and potentially net N2O production from denitrification. The seasonal variability in the N2O:O2 relationship in Saanich Inlet may also be important in other marine systems, suggesting the need for time-resolved information on N2O:O2 stoichiometry.

4.4.5 Interannual variability

Our observations demonstrate strong inter-annual variability superimposed on the seasonal cycle

- in Saanich Inlet. Figure 4.7 shows the full data record for O2, NO3 , N2O and H2S concentrations over the course of our time-series (note that CTD O2 data collection began only in 2009), as well as all available RPKM abundance estimates for genes involved in nitrification and denitrification. The repeated seasonal cycle is clearly evident over the time-series, as seen by the

- annual shoaling/deepening of the 90µM O2 contour and near-surface consumption of NO3 during summer. In general, N2O concentrations in the euphotic zone and oxycline fluctuated within a relatively narrow range from 10-20 nM throughout the time-series. However, there were some periods with notably high (20-30 nM) N2O concentrations between 80-120 m during late 2009/early 2010, and in late 2013 and 2014. These periods of increased N2O concentrations were coupled with changes in the timing and intensity of deep water renewal events, associated

- with lower O2 and NO3 accumulation below 150m during the fall of 2009, 2010, and 2013

(Figure 4.7 a,b).

O2 concentration of the measurements from that profile and above 30 µM O2. Hollow circles, connected by the line, represent mean monthly values.

To provide a quantitative assessment of inter-annual chemical variability in the deep basin, we examined the time series of mean substrate concentrations between 185 and 200m

- (Figure 4.9). Typical renewals led to an increase of between ~ 25-30 µM O2, 20-25 µM NO3 , and 18-28 nM N2O in deep basin waters. We observed strong correlations between N2O and

- - NO3 concentrations at 200m following renewal (R= 0.91, n= 11). Nitrite (NO2 ) concentrations

- often showed transient increases at the time of renewal, and again a few months later when NO3

- concentrations approached zero. As shown in Figure 4.9, there was no increase in O2 or NO3

Figure 4.7 Monthly time series measurements of oxygen (a), nitrate (b), nitrous-oxide (c), and hydrogen- sulfide (d) from 2007-2014. Note that CTD O2 collection began in 2009.

+ during fall 2009. The decline in NH4 (Figure 4.9), and H2S (not shown) during this period,

- - suggest a small amount of oxygen, NO3 and/or NO2 was supplied to the deep basin, but these species were quickly consumed in oxidation reactions. The moderate increase in N2O during fall

2009 also suggest a weak renewal took place, supplying some N2O to the deep basin. This period coincided with a moderate El Nino event, which has been shown to reduce deep basin renewal in the Strait of Georgia (Masson, 2002). Saanich Inlet renewal water has the same origin as the

Strait of Georgia, thus the El Nino could explain the apparently weak deep basin renewal in

Saanich during 2009.

Figure 4.8 Bubble plot shows all available RPKM abundance values for nitrification and denitrification genes. Bubble colour indicates N2O concentration and bubble size indicates RPKM value (gene abundance).

Our time-series of deep basin N2O concentrations (Figure 4.9) allowed us to investigate the direct vs. indirect N2O inputs associated with deep water renewal. The transient supply of O2

- and NO3 during renewal creates conditions conducive to in situ N2O production by both partial denitrification and nitrification. This is reflected in alternative patterns of gene abundance in the water column during periods of deep water renewal (e.g. Sep 2012; Figure 4.8). Stimulation of in situ N2O production should lead to increased N2O concentrations after the observed density increases at 200m. However, we found no evidence of increased N2O concentrations between the termination of renewal and subsequent sampling dates, suggesting that renewal did not induce in situ N2O production (Figure 4.9). Moreover, as discussed above, we found that post-renewal

N2O concentrations in the deep basin (18-28 nM; Figure 4.9) were similar to those measured in source renewal waters of Haro Strait during September 2014 (24-27nM), providing further evidence that post-renewal N2O increases resulted from direct supply within renewal waters.

However, our results do not preclude the possibility that N2O production and consumption are also stimulated during renewal periods, potentially due to simultaneous nitrification and

- denitrification where anoxic NH3-rich basin water mixes with oxygenated NO3 -rich renewal water.

Figure 4.9 Mean concentrations (± one standard deviation) of oxygen (a), nitrate (b), nitrite (c), nitrous-oxide

(d), and ammonium (d) below 180m in Saanich Inlet. Vertical dashed line indicate approximate times of deep basin renewal. Renewal times were derived based on density increases of >= 0.009 kg m-3 at 200m depth relative to the previous sampling date.

4.5 Conclusion

The time-series data presented here provide comprehensive N2O measurements from

Saanich Inlet, a model oxygen minimum zone. Relatively few marine N2O time-series are currently available (Farías et al., 2015; Lennartz, 2013; Naqvi et al., 2009), making our observations significant to understanding the temporal dynamics of N2O in productive coastal regions. Our monthly time-series observations enabled us to identify pronounced seasonal and

100

inter-annual variability in N2O distributions and sea-air fluxes, resulting from changes in physical circulation and redox chemistry within the inlet. The combination of metagenomic and chemical measurements provides robust evidence for the importance of O2-availability on coupled metabolic processes driving N2O cycling. However, additional work is needed to assess the activity of genes involved in nitrification and denitrification, and to measure the associated

N2O production/consumption rates associated with these processes. Indeed, integrated process and multi-omic datasets when combined with environmental parameters and thermodynamic principles have the potential to inform gene-centric models of microbial community metabolism at ecosystem scales. The identification of clear seasonal differences in N2O:O2 ratios may be important to consider when estimating N2O concentrations from O2 measurements in other regions with less resolved seasonal data coverage. The nearly decades-long time series of observations we have collected provide an important baseline against which to assess long-term climate-driven shifts in redox chemistry in coastal marine waters.

101

Chapter 5: Influence of expanding hypoxia on nitrous oxide concentrations in the coastal Subarctic Pacific

5.1 Summary

The expansion of marine oxygen minimum zones (OMZs) has the potential to affect oceanic concentrations and sea-air flux of the greenhouse gas nitrous-oxide (N2O), yet few datasets exist to quantitatively examine this possibility. Based on monthly observations between

2009-2015 in Saanich Inlet, a seasonally anoxic fjord on the coast of Vancouver Island, British

Columbia, we demonstrate significant increases in water column N2O concentrations and sea-air flux, associated with O2 loss and hypoxic boundary shoaling. These changes likely result from the regional-scale expansion of oxygen deficient waters, where N2O accumulates due to production by nitrification and incomplete denitrification. Our results are consistent with recent observations of oxygen-dependent changes in N2O concentrations in British Columbia coastal waters, which could act as a biogeochemical climate feedback.

5.2 Introduction

Climate-dependent changes in surface ocean hydrography and circulation have led to significant declines in sub-surface O2 concentrations over the past several decades (Crawford and

Peña, 2013; Whitney et al., 2007). These changes are expected to continue, and perhaps intensify, over the coming century, with important implications for marine ecosystem structure and biogeochemical cycling (Keeling et al., 2010). In oxygen minimum zones, O2 deficiency shifts energy away from multi-cellular life forms into microbial community metabolism based on

102

alternative terminal electron acceptors. As a result, OMZs are hotspots for fixed nitrogen loss processes including anammox and denitrification, with rapid cycling of the climate-active trace gas nitrous oxide (N2O; Dalsgaard et al., 2014). The world's most expansive marine OMZ is located in the poorly ventilated intermediate waters (between ~ 400 and 1500 m depth) of northeast Subarctic Pacific Ocean (NESAP). Long-term monitoring in this region at Ocean

Station Papa provided early evidence of on-going de-oxygenation in the ocean interior (Whitney et al., 2007), with decreases of 0.39 - 0.70 μmol kg−1 y−1 observed between 1950 and 2005.

A recent compilation of 30 years of O2 data from Canadian Pacific waters (Crawford and

Peña, 2013) has demonstrated that the rate of de-oxygenation in the coastal Subarctic Pacific (~

0.8 μmol kg−1 y−1) may be higher than that observed in open ocean waters. Decreasing coastal

O2 concentrations have been linked to fish mortality associated with the upwelling of oxygen deficient sub-surface waters along the North American continental shelf (Roegner et al., 2011).

These continental shelf waters contribute disproportionately to global oceanic sources of N2O to the atmosphere (Bange et al., 1996; Naqvi et al., 2010; Nevison et al., 2004), and it has been suggested that on-going de-oxygenation in these systems could intensify oceanic N2O emissions, leading to a potentially significant positive feedback on global climate warming (Bange, 2006;

Codispoti, 2010; Naqvi et al., 2010).

At present, the link between ocean de-oxygenation and N2O cycling remains largely speculative, due to a scarcity of dedicated time-series observations. Long-term monitoring at the

−1 −1 Boknis Eck station in the Baltic Sea demonstrated a loss of 0.9 μmol O2 kg y in the near bottom (25 m) waters since 1957, yet no significant trends in N2O concentrations have been detected since measurements began in ~2006 (Lennartz, 2013). Similarly, no statistically significant temporal trends in O2, or N2O have been detected in the productive coastal upwelling 103

waters at the COPAS time-series off the coast of Chile since its inception in 2002 (Farias et al.,

2009; Farías et al., 2015).

Here we report new time-series observations of N2O from Saanich Inlet, a coastal fjord in

SW British Columbia that provides a model system to study the effects of de-oxygenation on biogeochemical cycling in marine OMZs. Our results demonstrate significant increases in N2O concentrations and sea-air fluxes, associated with progressive de-oxygenation in this system. We discuss our observations in the context of a long-term (> 50 year) archive of dissolved O2 in

Saanich Inlet, examining how decadal scale changes in O2 concentrations may influence the cycling of redox-active nitrogen species in coastal waters of the Subarctic Pacific.

5.3 Methods

Saanich Inlet is coastal fjord in SW British Columbia with a maximum depth of ~230 m.

This system experiences seasonal cycles of oxygen-deficiency, which result from high rates of organic matter remineralization during spring and summer in poorly ventilated bottom waters, and subsequent inflow of nutrient rich-oxygenated waters during deep water renewal in the early fall (Anderson and Devol, 1973). Much of the water in Saanich Inlet can be traced back to the

N2O-rich continental shelf waters off the coast of British Columbia and Washington, which are derived from a mixture of NESAP water and California Undercurrent water (Anderson and

Devol, 1973; Thomson and Krassovski, 2010; Waldichuk, 1957). The inlet is characterized by a pronounced vertical O2 gradient that supports a wide range of microbially-driven processes, including nitrification and denitrification, which directly influence water column N2O concentrations (Bourbonnais et al., 2013; Cohen, 1978; Grundle and Juniper, 2011). The inlet has been used as a model system to examine microbiological and biogeochemical processes in 104

low O2 waters, and has been the site of a number of seminal studies of marine N2O cycling

(Cohen, 1978; Devol et al., 1984). Yet, there have been no systematic time-series observations of N2O in this system.

Archived temperature, salinity and O2 data are available from station SI03, near the deepest part of Saanich Inlet starting in 1930, with varying sampling intervals and a number of significant data gaps over this time period (http://www.pac.dfo-mpo.gc.ca/science/oceans/data- donnees/search-recherche/profiles-eng.asp). Sampling has been consistent enough for trend analyses since 1960. Winkler titrations (Carpenter, 1965) were used to for the analysis of all O2

- data in the archive. We calculated mean monthly O2, NO3 , salinity, and temperature in the 10 -

200 m depth horizon from all depth profiles for each month containing adequate data in the archive. To reduce the likelihood of biasing our average values towards a particular depth interval, we excluded any depth profiles that did not contain at least one measurement from each of the four 50 m depth intervals in the upper 200m of the water column. This removed ~20% of all O2 profiles in the original archive from further analysis, and 15% of all temperature and salinity profiles. The remaining data were relatively well distributed across all seasons throughout the time-series. We obtained roughly 29% of all usable O2 profiles from summer

(Jun-Aug), compared with 19% from fall (Sep-Nov), 24% from winter (Dec-Feb), and 28% from spring (Mar-May). We derived the equation of the line of best fit through the mean values vs. time, and tested for a significant correlation using either Pearson’s linear regression for normally distributed data, or Spearman’s for non-normally distributed data, using p values less than 0.05 as a threshold of statistical significance.

For the past decade, we have been conducting monthly time-series observations of water column microbiology and biogeochemical dynamics in Saanich Inlet, with the aim of 105

documenting the drivers of and responses to seasonal and inter-annual O2 variability. Beginning in late 2006, monthly samples were collected at station SI03 (48.59°N, 123.51°W), near the deepest part of Saanich Inlet (~ 220 m depth), for a suite of measurements including nutrients and dissolved gases (Hawley et al., submitted). Our initial trace gas measurements were made using membrane inlet mass spectrometry, with relatively low sensitivity and precision. Starting in 2009, we began using an automated purge and trap GC-MS system for N2O analysis (Capelle et al., 2015), resulting in significantly improved data quality. For the purposes of this study, we only present N2O measurements derived from this latter method. Samples were collected at 16 depths between the surface and 200 m, with O2 concentrations measured using a Seabird SBE43 coupled to a CTD (Seabird SBE25) and calibrated against discrete bottle samples analyzed by

Winkler titration. Nutrient concentrations were measured by colorimetry using a Bran Luebbe

Auto Analyzer (Armstrong et al., 1967). Sea-air fluxes were derived from wind speeds at 10m height (Environment Canada meteorological buoy c46134) and excess surface N2O concentrations, using a weighted piston velocity (kw) as described by Reuer et al. (2007; modified from Nightingale et al. 2000). Excess surface concentrations were derived from the difference between observed and equilibrium values of N2O, which were determined from temperature, salinity, and atmospheric concentrations according to Weiss and Price (1980), using atmospheric measurements from Barrow, Alaska (http://www.esrl.noaa. gov/psd/data/; Kalnay et al., 1996).

106

Figure 5.1 Long term trends in mean water column (10 : 200 m) O2 (a) and temperature (b) in Saanich Inlet from 1963 – 2015. Grey dots indicate archive data (1960-2007), and crosses indicate recent time-series data

(2009-2015). Thick black lines show mean annual values. Dashed lines indicate linear trends through the data.

Both trends are statistically significant (p < 0.05).

5.4 Results and discussion

Analysis of the long-term Saanich Inlet data archive (Figure 5.1) shows notable changes in water column temperature and O2 concentrations that are consistent with long-term trends observed in the NE Pacific source waters of Saanich Inlet. The rate of warming since 1960

(0.01°C yr-1) is similar to that observed over the same period in the offshore waters of the

-1 NESAP (0.005-0.012 °C yr ; Whitney et al., 2007). The long-term change in O2 concentrations observed in Saanich Inlet was -0.35 µmol kg-1 yr-1 averaged over 10 - 200 m. This rate fall below the ranges reported from Ocean Station Papa between 1956 - 2006 (0.4 – 0.7 µmol kg-1 yr-1), and the BC continental shelf between 1979 - 2011(0.83 - 1.22 µmol kg-1 yr-1; Crawford and Peña,

2013; Whitney et al., 2007). These apparently lower rates of O2 loss in Saanich Inlet may be

107

explained by time and depth-dependent differences in de-oxygenation rates. When trends are examined using only data more recent than 1979 (as done in Crawford and Pena, 2013), the rate

-1 -1 of O2 decline averaged over the water column (10 : 200 m) in Saanich Inlet (0.74 µmol kg yr ) agrees well with the BC continental shelf. Moreover, there was significant depth-dependent variability within this overall water column average, with little or no O2 loss in the upper 50 m, which likely skewed our reported rates of de-oxygenation relative to the NESAP. The most

-1 -1 rapid rate of de-oxygenation (mean -0.69 µmol O2 kg yr from 1960 - 2015) occurred near the

-1 hypoxic transition (< 60 µmol O2 kg ) between 80 - 100 m depth. This rate is near the upper range observed in the NESAP between 1950-2005 (Whitney et al., 2007). The strong coherence between changes in Saanich Inlet and its source waters suggest that long-term trends in the

NESAP and BC continental shelf are reflected in Saanich Inlet.

Declining O2 in the source waters of Saanich Inlet has been attributed to reduced rates of intermediate water formation, weakened ventilation due to increased stratification, and changes in the relative supply from the NESAP and California Undercurrent (Whitney et al., 2007,

Crawford and Peña, 2013). Changes in the composition of these source waters could also impact

Saanich Inlet. Irrespective of the driving mechanism, the historical data provide evidence that

Saanich Inlet is undergoing a significant long-term shift in hydrographic properties, associated with a progressive regional de-oxygenation. Shorter-term forcing including the El Niño events in

2009-10 and 2015 and the recent appearance of a strong warm temperature anomaly in summer

2015 (Bond et al., 2015; Whitney, 2015) have also influenced regional hydrography and O2 concentrations. These episodic events have occurred against a background of decadal scale trends.

Our recent intensive time-series work in Saanich Inlet suggests that the rate of de- 108

Figure 5.2 Mean water column concentrations of O2 (a), NO3 (c), and N2O (d), and depth of hypoxic boundary

(b) in Saanich Inlet from Jan 2009 – Dec 2015. Straight black line shows linear regression through the data.

All trends are statistically significant (p < 0.05).

oxygenation in this system since 2009 (Figure 5.2a) has been significantly accelerated relative to the long-term average. Over the ~ 7 years of our monthly sampling program, mean water

-1 -1 column O2 concentrations have decreased by an average of ~ 2.6 µmol kg yr (p <0.01), faster

109

than the rate of de-oxygenation observed for the BC continental shelf since ~1980 (Crawford and

Peña, 2013; Whitney et al., 2007). Consistent with the patterns observed in the long-term data

-1 archive, de-oxygenation has been most-rapid between 80 - 100 m (- 6.54 µmol O2 kg ) since

2009. Decreasing O2 concentrations have been associated with a shoaling of the hypoxic zone (<

-1 -1 -1 -1 60 µmol kg O2) and suboxic (< 5 µmol kg O2) zone at rates of ~ 2 m y and 3.4 m y , respectively (Figure 5.2b; p <0.05). This result implies a shoaling of the oxycline by ~ 15 - 25 m in less than a decade, more than four times greater than that observed in the subarctic NE Pacific between 1956 - 2006 (Whitney et al., 2007). It is important to note, that the rapid expansion of

O2-deficient waters may have been exaggerated by the recent warm anomaly in the NESAP

(Bond et al., 2015; Whitney, 2015). However, even if data from 2014-2015 are excluded, the rate of O2 decline observed in our time-series remains higher than that observed between 1960 and 2015.

The extent to which declining O2 has influenced other biogeochemical processes in

Saanich Inlet remains largely unknown. Our results show that recent de-oxygenation in Saanich

- Inlet has been associated with a significant positive trend in water column NO3 and N2O

- concentrations (Figure 5.2c,d). Mean water column NO3 and N2O concentrations increased by

0.59 mol kg-1 y-1 and 0.56 nmol kg-1 y-1, respectively (Figure 5.2c,d). Notably, the fastest rate of

-1 -1 N2O increase was associated with the 80 - 100m depth horizon (1.3 nmol N2O kg y ), where de-oxygenation was also most rapid. The increase in N2O concentrations was associated with a statistically significant increase in sea-air fluxes of N2O (p < 0.01), indicating an enhanced local

N2O source to the atmosphere. In addition to the sea-air flux of N2O from Saanich Inlet surface waters, additional N2O may be ventilated to the atmosphere during the rapid lateral exchange of surface water with Satellite Channel, a region of active mixing. 110

- The changes in N2O and NO3 concentrations in Saanich Inlet are consistent with an increase in the relative contribution of older, poorly ventilated waters to the region, and with the

- observations of (Whitney et al., 2007) who reported NO3 increases between 100-600 m in the

- subarctic NE Pacific. The coupling between NO3 and N2O concentrations reflects the fact that

+ - they are both produced primarily through nitrification (NH4 oxidation to NO3 ). Nitrification

+ produces between 0.04 and 4 nmol N2O per µmol NH4 oxidized (De Wilde and De Bie, 2000;

- Goreau et al., 1980; Yoshida et al., 1989). Nitrous oxide thus accumulates along with NO3 in poorly ventilated waters, provided that O2 levels are sufficiently high to inhibit denitrification,

- which consumes both NO3 and N2O.

In addition to changes in source water properties, increasing N2O concentrations in

Saanich Inlet could also be attributable to in situ, O2-dependent changes in N2O cycling.

Decreasing O2 concentrations have been shown to increase N2O yields from nitrification in laboratory experiments (Goreau et al., 1980) and marine systems (Nevison et al., 2003; Punshon and Moore, 2004a), including coastal waters along the BC continental shelf (Capelle and Tortell,

2016). In addition, incomplete denitrification has been suggested as a mechanism of N2O production in suboxic waters (Freing et al., 2012; Yamagishi et al., 2005). This process can be facilitated within anoxic micro-zones inside particles (Jia et al., 2016), whose formation would be enhanced by water column de-oxygenation. Increasing N2O could also be related to increasing

+ zooplankton grazing and the subsequent release of NH4 , which could fuel more rapid N2O production from nitrification. Our observations do not allow us to explicitly examine the relative importance of these different processes.

- The coupled changes in O2, NO3 , and N2O observed in Saanich Inlet reflect larger-scale patterns across the BC continental shelf and other NESAP coastal waters. Considering only 111

- Figure 5.3 Relationship between O2, NO3 , and N2O concentrations from the west coast of Vancouver Island

(WCVI) and Saanich Inlet. Small black dots represent discrete measurements from samples collected from

50-120m along the WCVI between Jun 2012 and Jun 2014. Large black diamonds represent mean ± 1 standard error from each of the five WCVI cruises. Black x’s show mean values from 50 – 120 m depth from each cruise in Saanich Inlet between 2009 and 2015.

- waters above the anoxic zone in Saanich inlet (where denitrification consumes NO3 and N2O

- locally), the NO3 :N2O relationship observed between 50-120 m in Saanich Inlet is consistent with that observed over the same depth interval in the continental shelf and open ocean waters along the west coast of Vancouver Island WCVI (Figure 5.3a). This suggests that N2O production is associated primarily with nitrification in both systems, with similar N2O yields.

Examination of mean values derived from five WCVI cruises over a two year period (large

- symbols in Figure 5.3), shows an O2-dependent increase in both N2O and NO3 over seasonal and

- inter-annual time scales. These data thus suggest a tight coupling between NO3 and N2O

112

concentrations in coastal NESAP waters (Figure 5.3), that is mirrored by the patterns observed in

Saanich Inlet. However, data derived from Saanich Inlet show a significant negative offset in

N2O concentrations relative to O2 as compared to WCVI samples, suggesting a low N2O-yield per mole O2 consumed in this system. The relatively low N2O concentrations at a given oxygen concentration in Saanich Inlet as compared to the coastal NESAP is consistent with observations from other landlocked anoxic basins (Naqvi et al., 2010). Such behavior could be attributable to the higher abundance of fish and zooplankton in the highly productive waters of Saanich Inlet

(Herlinveaux, 1962), which could reduce the proportion of O2 consumption associated with nitrification, thus reducing N2O yields per mole O2 consumed. This would act to reduce the O2- dependent accumulation of N2O in Saanich Inlet relative to other regions of the NESAP.

5.5 Conclusion

The lack of long-term data sets, and a limited understanding of the sensitivity of N2O cycling to environmental conditions, makes predictions of future marine production rates and sea-air fluxes of this climate-active gas highly uncertain (Zamora and Oschlies, 2014; Zamora et al., 2012). The results presented here, despite their limited temporal extent, provide support for the idea of O2-dependent feedback on oceanic N2O cycling driven by OMZ expansion in marine waters. Marine emissions of N2O are globally significant, accounting for roughly one third of the total N2O released to the atmosphere (Freing et al., 2012). Roughly 50% of this flux is from coastal upwelling regions and estuaries (Bange, 2008), so any increase in fluxes from these systems may have important climate implications. Expanded time-series monitoring of coastal marine systems is needed to better constrain the changes in the concentration of N2O and other climate-active gases in the face of rapid environmental change. Our observations are likely 113

influenced by the extreme temperature anomaly during 2014 and 2015 across much of the north

Pacific Ocean (Bond et al., 2015; Whitney, 2015). Therefore, while O2 loss and associated N2O increases have likely accelerated in recent years, these accelerated rates may not be indicative of longer-term trends. However, the recent temperature anomalies in Saanich Inlet may become increasingly common episodic events as the North Pacific warms, further enhancing the tendency towards elevated N2O concentrations and sea-fluxes in the coming years.

114

Chapter 6: A multi-year time-series of CH4 dynamics in Saanich Inlet, British

Columbia, a seasonally anoxic fjord

6.1 Summary

We present seven years of water column methane (CH4) measurements from Saanich

Inlet, a seasonally anoxic fjord in British Columbia, Canada. Our time-series data show that CH4 concentrations and sea-air fluxes vary significantly over seasonal cycles, driven primarily by changes in O2-availability throughout the water column. Methane concentrations were always supersaturated near the surface (mean 1160 ± 590% saturation at 10m), and remained high between 10-85m. Over this depth range, CH4 may have been supplied to the water column by production inside sinking particles, sediment resuspension, cleavage of methylated compounds, and/or transport from the deep basin. A persistent mid-depth CH4-minimum occurred between

85-110m, associated with a turbidity maximum near the base of the oxycline. Sediment resuspension in this depth range may have supplied particle-associated methanotrophs, which facilitated rapid oxidation of the high CH4 concentrations in oxygenated waters. Below the oxic- anoxic transition zone, CH4 concentrations increased significantly with depth towards the bottom of the water column, and concentrations in the deep basin waters (> 150 m) increased steadily over the seasonal cycle from the onset of anoxia until the subsequent renewal. The seasonal and depth-dependent distribution of CH4 in deep basin waters suggests a diffusive sedimentary CH4 source during periods of summer anoxia, with deep water renewal supplying oxygenated water with near-equilibrium CH4 concentration. Beyond the strong seasonal cycle in CH4 concentrations, we also observed inter-annual variability. Most notably, the moderate 2009 -

115

2010 El Niño resulted in a weak deep basin renewal, and anomalously high CH4 concentrations in both the deep basin and upper 85m. We also observed a significant decreasing trend in upper water column (10 – 85 m) CH4 concentrations and sea-air fluxes over our time-series. Although the cause of this trend is unclear, it may be related to declining O2 concentrations and shoaling of the suboxic boundary. Our time-series observations provide a better understanding of the seasonal and inter-annual dynamics of CH4 in Saanich Inlet, and baseline data against which to better identify longer term, climate-driven changes in the biogeochemistry of a model coastal oxygen minimum zone.

6.2 Introduction

Methane (CH4) is an important greenhouse gas, accounting for ~30% of the total atmospheric radiative forcing (IPCC, 2013). Oceanic sources account for an estimated 2-4% of the total CH4 released to the atmosphere, with 75% of this coming from continental shelf areas and estuaries (Bange et al., 1994). Much of the oceanic CH4 originates in organic-rich anoxic coastal sediments, which can have extremely high CH4 concentrations due to the activity of heterotrophic microbial methanogens that use CO2 as a terminal electron acceptor. Non- biological methanogenesis can also occur thermogenically deep within the earth’s crust, with the produced CH4 migrating upward through tectonic faults (Valentine, 2011). It has been suggested that the sedimentary CH4 deposits in some coastal regions are large enough to cause significant and rapid atmospheric warming if ventilated to the atmosphere over short time scales (Dickens et al., 1995). Yet, the contribution of the coastal ocean to global CH4 cycling is subject to significant uncertainty due to limited seasonal and spatial coverage (Bange et al., 1994), and potential underestimation of CH4 emissions in estuaries, and regions with active seabed CH4 116

seeps or significant sediment resuspension (Bange, 2006; Schneider von Deimling et al., 2011,

Bussmann, 2005).

Microbial CH4 oxidation in or near marine sediments prevents the majority of sedimentary CH4 from being released to the water column and ventilated to the atmosphere

(Reeburgh, 2007; Valentine, 2011). Methane can be oxidized under aerobic conditions using oxygen as a terminal electron acceptor, or anaerobically, using sulfate, nitrate, nitrite or various transition metals (Beal et al., 2009; Ettwig et al., 2010; Nauhaus et al., 2002; Raghoebarsing et al., 2006; Reeburgh, 2007). Under certain conditions, microbial CH4 consumption can be overwhelmed by rapid CH4 fluxes, for example when sedimentary CH4 escapes in bubbles

(Valentine, 2011), and when sediments are disturbed by wave action or bioturbation (Bussmann,

2005; Shakhova et al., 2013). Regions where this sedimentary CH4 reaches the surface mixed layer tend to be disproportionately large sources of atmospheric CH4 relative to other ocean regions (Middelburg et al., 2002; Solomon et al., 2009).

An additional marine CH4 source to the atmosphere is derived from a secondary, shallow

CH4 maximum below the euphotic zone in fully aerobic conditions. The observations presents the so-called ‘methane paradox’, since CH4 production has traditionally been thought to be restricted to anaerobic environments. One possible source of CH4 in the upper water column is anaerobic methanogenesis inside anoxic microenvironments, such as sinking particles and the digestive tracts of fish (Oremland, 1979; van der Maarel et al., 1999) and zooplankton (De

Angelis and Lee, 1994). More recently, aerobic CH4 production has also been documented during the cleavage of methyl-groups from methylated compounds such as methylphosphonate

(Karl et al., 2008) and dimethylsulfide (DMS) (Damm et al., 2010; Florez-Leiva et al., 2013) in surface waters. These near-surface CH4 sources, though small on a volumetric basis, may be 117

widespread over vast areas of the aerobic surface ocean, thus representing an important marine source of atmospheric CH4.

Marine CH4 cycling is influenced by a number of factors, including oxygen availability, primary productivity, freshwater inputs, upwelling, sedimentary processes and microbial and invertebrate community dynamics (Bange, 2006; Naqvi et al., 2010; Reeburgh, 2007; Valentine,

2011). Of all these factors, oxygen availability may be the single most important control on water column CH4 concentrations. Methane concentrations are very high (5-10 µM) in the water column of anoxic basins such as the Cariaco Basin and Black Sea, and in suboxic basins (tens to hundreds of nanomolar; see Valentine, 2011). The ongoing expansion of marine oxygen-depleted waters (Crawford and Peña, 2013; Whitney et al., 2007) and anthropogenic nutrient loading may therefore increase marine CH4 concentrations and sea-air fluxes (Bange, 2006). Distinguishing variability in CH4 cycling over various time-scales requires prolonged, repeated sampling efforts spanning different seasons and multiple years. Time-series sampling projects are useful in meeting these requirements, but are presently restricted to a few oceanic regions, including the productive upwelling waters of Chile (Farias et al., 2009), the shallow, seasonally anoxic waters of Eckenförde Bay in the Baltic Sea (Bange et al., 2010), and the Hawaii Ocean Time-Series station ALOHA in the oligotrophic, subtropical North Pacific (Karl and Lukas, 1996).

In this article, we document a 7-year time-series of monthly water column CH4 concentrations from Saanich Inlet, a highly productive coastal fjord, which experiences seasonal anoxia in the deep basin waters. Methane concentrations in Saanich Inlet have been measured previously by Lilley et al. (1982) and Ward et al. (1989), with depth profiles typically showing significant surface super-saturation (~800%), and a shallow concentration maximum (40-60nM between 30-70m depth). Below this sub-surface maximum, CH4 concentrations decrease to a 118

mid-depth minimum (~10nM) at the oxic-anoxic interface, then increase with depth to values greater than 1500 nM near the sediments (Lilley et al., 1982; Ward et al., 1989). Methane oxidation rates previously measured in Saanich Inlet were high enough to consume the entire upward flux of CH4 from the deep basin (Ward et al., 1989), suggesting the upper CH4 maximum

(above 85m) was not derived from same source as the deep basin CH4 maximum.

While existing data provide a general picture of CH4 distributions in Saanich Inlet, the lack of systematic, and seasonally resolved measurements limits our understanding of the temporal dynamics of CH4 in relation to dominant environmental controls (e.g. O2 availability) in this system. Recent metagenomic work in Saanich Inlet has established this site as a model system for coastal low oxygen marine waters, with bacterial and archaeal communities that appear to be representative of other similar marine waters (Orsi et al., 2012; Walsh and Hallam,

2011; Walsh et al., 2009; Zaikova et al., 2010). Our objective is to contribute detailed measurements of CH4 to the growing understanding of molecular signatures in Saanich Inlet.

Our results document regular, predictable seasonal variability in CH4 distributions, and longer- term trends associated with regional hydrographic shifts, providing new insight into the temporal dynamics of CH4 in Saanich Inlet, and a baseline dataset against which future changes can be assessed. The results are significant in the context of declining O2 concentrations in coastal

British Columbia waters.

119

6.3 Methods

6.3.1 Study area

Saanich Inlet is a 26km glacial fjord with a maximum depth of 230m, and an elevated sill at a depth of ~70m (Figure 6.1). The primary sources of river-derived freshwater in Saanich Inlet are located outside of the fjord (the Cowichan and Fraser rivers), which drive an inverse- estuarine circulation through much of the year (Gargett et al., 2003). During summer, strong tidal mixing near the mouth of the inlet periodically drives normal estuarine circulation above the sill, delivering nutrients to the euphotic zone and stimulating high primary productivity (Gargett et al., 2003; Timothy and Soon, 2001). Deep basin circulation is restricted during most of the year by the shallow sill (Figure 6.1b), except during late summer / early fall, when dense, oxygenated and nutrient-rich water flows over the sill and sinks to the bottom of the inlet (below 150 m), displacing the resident deep water (Anderson and Devol, 1973). Heterotrophic decomposition consumes all available oxygen in the deep basin between October and June, after which H2S accumulates in the water column. Anoxia typically persists until the subsequent deep basin renewal. The intensity of renewal, measured as a percentage of the total volume of water in the deep basin that is replaced during renewal, varies on an interannual basis, typically ranging from

30-70% (Anderson and Devol, 1973; Manning et al., 2010).

Elevated primary production in Saanich Inlet contributes to high particle fluxes relative to other fjords in the region (Timothy et al., 2003), with fecal pellets making up a significant proportion (between 15-99%) of sinking particles in shallow depths (above 60m) (Sancetta,

1989). An increase in lithogenic material with depth indicates that sediment resuspension also contributes a significant amount of material to the total particle fluxes in Saanich Inlet 120

Figure 6.1 Map of Saanich Inlet, BC, showing location of sampling (star, panel a), and location on southeast

Vancouver Island (inset). Depth transect along thalweg of Saanich Inlet (panel b) shows elevated sill near mouth of inlet. Adapted from Anderson and Devol, (1973). 121

(Katz et al., 2012; Sancetta and Calvert, 1988; Timothy et al., 2003; Yahel et al., 2008). Much of this resuspended material originates from the elevated sill near the mouth of Saanich Inlet

(Sancetta and Calvert, 1988; Timothy et al., 2003). Tidal activity is believed to be responsible for some of the sediment disturbance in Saanich, as turbidity has been shown to vary on tidal cycles

(Matabos et al., 2015). However, sediment resuspension by fish and other large heterotrophs is also important (Katz et al., 2012; Yahel et al., 2008). The activity of these organisms is influenced by O2-availability (Matabos et al., 2015), which fluctuates on fortnightly and seasonal cycles, and is expected to change over longer time scales as part of larger regional trends of declining O2 availability (Crawford and Peña, 2013; Whitney et al., 2007).

6.3.2 Sampling and analysis

Samples were collected at monthly intervals at single station near the deepest part of the inlet (SI03, 48.59N, 123.51W; Figure 6.1) using the MSV Strickland. Water was collected from

- various depth using 8 or 12L GO-FLO bottles and subsampled for dissolved O2, nutrients (NO3 ,

- 3- NO2 , PO4 ), H2S and CH4. Additional parameters (temperature, salinity, O2, Chla fluorescence, transmissivity, and photosynthetically active radiation, PAR) were collected by a conductivity- temperature-depth (CTD) sensor (Sea-Bird SBE 25) equipped with a Sea-Bird SBE 43 O2- sensor, a Biospherical Instruments QSP-200PD PAR Sensor, a WET Labs C-Star transmissivity sensor, and WET Labs WETStar fluorometer. Additional discrete samples were collected for oxygen measurements by Winkler titration, and these data were used to calibrate CTD O2

- - 3- measurements. Macronutrient samples ( NO3 , NO2 , PO4 , and Si) were filtered through a 0.2

µm syringe filter and stored frozen until measurement by segmented continuous-flow colorimetric analysis by a Bran Luebbe Auto Analyzer (Armstrong et al., 1967). Ammonium 122

+ (NH4 ) was measured using the fluorometric method of Holmes et al., (1999) within 8 hours of collection. Hydrogen-sulfide (H2S) was measured on board the ship using the methods of Cline

(2015). Dissolved CH4 was measured using purge and trap gas-chromatograph mass- spectrometry (PT-GCMS), using the method of Capelle et al. (2015). Briefly, water was transferred from GO-FLO bottles via silicon tubing to 60mL glass serum vials (in duplicate), which were overfilled three times without the introduction of air bubbles. Samples were then preserved with 100µL of saturated HgCl2 solution, and crimp-sealed without a headspace using butyl rubber stoppers and aluminum caps. Samples were stored in a laboratory cold room before analysis by PT-GCMS. This method provides better than 3% analytical precision, and we obtained a mean standard deviation of 15% between replicate samples.

Sea-air flux of CH4 was calculated from the mixed layer CH4 concentration in excess of atmospheric equilibrium (ΔC) and a weighted piston velocity (kw), where:

Flux = (Cobs - Ceq) * kw = ΔC * kw (6.1)

The mixed layer was typically restricted to the upper 5m of Saanich Inlet, and thus not captured by our shallowest sampling depth (10m). To estimate the CH4 concentration of the mixed layer, we applied a correction factor to the 10m samples. This correction factor was determined by the mean ratio between CH4_surface : CH4_10m from the available profiles with CH4 measurements from both the surface (~ 1 m) and 10m (0.68 +/- 0.13; n=12). Excess CH4 concentrations in the mixed layer were derived from the difference between measured CH4 concentrations (Cobs) and atmospheric equilibrium values (Ceq), which were computed using the formulation of

Wiesenburg and Guinasso (1979), based on the temperature and salinity of the mixed layer, and the atmospheric CH4 concentration during that month measured in Barrow, Alaska

(NOAA/ESRL halocarbons in situ program; 123

http://www.esrl.noaa.gov/gmd/dv/data/index.php?site=brw). We computed a weighted piston- velocity (kw) for each cruise based on the relationship between kw and wind speed (Nightingale et al., 2000), using wind speed data from the 14-days prior to and including sample collection according to Reuer et al. (2007). Wind speed data were obtained from Environment Canada meteorological buoy c46134 in Saanich Inlet.

6.4 Results and discussion

6.4.1 Overview

Across all depths and sampling dates, CH4 concentrations ranged from below our detection limit (<0.5 nM) to 1850 nM, with mean and median values of 168 nM and 28.7 nM, respectively. Methane concentrations were always supersaturated at our shallowest sampling depth (10m), ranging from 8.0 nM (250% saturation) to 145 nM (2900% saturation). Depth profiles of CH4 typically displayed an upper maximum (mean 32 ± 25 nM CH4) between 10-

85m, a mid-depth minimum (18 ± 25 nM CH4) between 85-110m, and a deep maximum below

110m (mean 697 ± 294 nM CH4 at 200m). The CH4 time-series displayed regular seasonal cycles, which appeared to be driven by changes in primary production, particle fluxes, O2- availability, and deep basin renewal. We discuss these trends in the following sections.

6.4.2 Mean seasonal cycle

Figure 6.2 shows the mean annual cycle of CH4 concentrations and sea-air fluxes in

Saanich Inlet, derived from all of our individual measurements. The figure also shows the mean

- annual cycles of O2, NO3 , and H2S to provide a chemical context for the CH4 data. In this 124

125

section, we briefly describe the dominant seasonal patterns observed over different depth ranges in the water column. Processes driving these distributions, and their inter-annual variability are discussed in the following sections.

In the upper 85m, CH4 concentrations ranged from 1 nM (30% saturation) to 225 nM

(nearly 4000% saturation) (mean 33 ± 26 nM CH4). The CH4 maximum in this depth range

- coincided with the spring phytoplankton bloom, as indicated by the reduction of NO3 concentrations and accumulation of O2 near the surface (0-50m) (Figure 6.2b). This feature intensified during spring and early summer. Following the mid-summer maximum, CH4 concentrations between 0-85m decreased between July and November, corresponding with declining primary productivity, as indicated by reduced consumption of near-surface (0-50m)

- NO3 and the return of O2 levels to atmospheric equilibrium levels (Figure 6.2).

The seasonal cycle of surface CH4 concentrations in the upper water column was mirrored by sea-air fluxes, which showed maximum values during mid-summer. The highest fluxes (> 60 µmol m-2 d-1) were observed in late June / early July (Figure 6.2e), with an overall

-2 -1 -2 -1 range from 3.8 to 64.3 µmol m d , and a mean of 19.7 ± 13.2 µmol m d . The CH4 sea-air flux estimates are similar to those reported in other land-locked anoxic basins such as the Black

Sea (26.6-50 µmol m-2 d-1; Amouroux, 2002) and Eckenförde Bay in the Baltic Sea (6-15 µmol m-2 d-1; Bange et al., 2010; Naqvi et al., 2010).

At depths between 85 and 110 m, we observed a significant and persistent minimum in

CH4 concentrations (mean 23.8 +/- 39 nM CH4). This feature was most pronounced during late summer / early fall (Aug-Oct), corresponding with the end of the productive season prior to deep basin renewal (Figure 6.2). Deep basin waters (below 120 m), were characterized by extremely high CH4 concentrations (mean 341 ± 409 nM CH4; maximum concentration: 1800 nM CH4), 126

with increasing values towards the bottom of the water column. This distribution is consistent with a sedimentary source of CH4 resulting from methanogenesis. Methane concentrations in the deep basin increased from winter to late summer, following the onset of anoxia (as indicated by

- the absence of O2, accumulation of H2S, and absence of NO3 due to denitrification; Figure 6.2).

- Deep basin renewal in the fall (indicated by the increase in O2 and NO3 below 120m) typically led to a rapid loss of CH4 below 120m, with concentrations decreasing by as much as 1800nM between monthly sampling periods.

6.4.3 Factors driving CH4 cycling in the upper water column

The strong super-saturation of CH4 in the upper 85m of Saanich Inlet suggests either in situ CH4 production in these waters, or active supply from an adjacent CH4 reservoir. Previous work has demonstrated a number of potential CH4 sources in the upper water column of coastal marine environments. Physical supply mechanisms include river water input (Bange et al.,

1998), lateral advection from sediments (Bussmann, 2005) and upward transport of deep waters

(Rehder et al., 2002). In situ CH4 sources may include production within the anoxic interior of particles and digestive tracts of macro-organisms (De Angelis and Lee, 1994; Holmes et al.,

2000; Oremland, 1979), and cleavage of methylated compounds such as DMS (Florez-Leiva et al., 2013) and methylphosphonate (Karl et al., 2008). Ward et al. (1989) used CH4 oxidation rates and diffusive flux estimates to determine that CH4 was produced at a rate of ~0.03 nmol

-2 -1 -1 cm d between 20-30m, which is equivalent to 0.03 nM CH4 d integrated over that depth range. We use our time-series data to investigate the potential contributions of each of different

CH4 sources to the upper water column of Saanich Inlet. Although our data are useful in identifying potential CH4 sources, additional tracer and rate measurement studies are needed to 127

quantitatively evaluate the strength of these source terms.

The regular seasonal cycle of CH4 concentrations in the upper 85m (Figure 6.2) suggests that CH4 sources are strongest during spring and early summer in Saanich Inlet. This corresponds to a period of maximum freshwater flux from the Fraser River. However, we did not observe a correlation (either positive or negative) between salinity and CH4 at any depth range between 10 and 85m, suggesting that riverine CH4 was not a significant source term over this depth interval.

River input may be a more important CH4 source above 10 m, in the shallow fresh water layer.

Unfortunately, we did not make regular CH4 measurements in the very top layer of the water column. However, the few near-surface (~1 m) CH4 samples we did measure (n=12) never exceeded those from 10 m depth and were correlated with CH4 concentrations between 20-40 m

(R = 0.75), suggesting that CH4 in the surface was derived from subsurface sources rather than freshwater inputs.

The seasonal pattern of CH4 concentrations in the upper 85 m of the water was consistent with seasonal variability in phytoplankton biomass (inferred from Chla fluorescence data). On average, phytoplankton biomass began to increase significantly in the upper water column by late March, and this timing coincided with increasing CH4 concentrations, with peak values observed during May-July (Figure 6.3a). Methane concentrations in the upper 85m were lowest between August and March, during periods of relatively low phytoplankton abundance through late fall and winter. Over the full annual cycle, we found a strong relationship between CH4 concentrations and Chla fluorescence in the upper water column (R = 0.93, p < 0.01; n= 12;

Figure 6.4), suggesting a link between primary production and CH4 cycling.

Despite the potential link between primary production and CH4 in the upper 85m, the upper CH4 maximum was typically found below ~30m, indicating little CH4 production occurred 128

in the euphotic zone. One possible explanation for the relationship between CH4 concentrations and phytoplankton abundance is the induction of methanogenesis in anoxic micro-zones in particles derived from phytoplankton-derived detritus, or zooplankton. Previous observations of

Hydrogen supersaturation in oxic waters of Saanich Inlet (Lilley et al., 1982) have been taken as evidence of anoxic micro-environments in surface waters. Whereas particles could provide a potential environment favourable to methanogenesis, we observed that the seasonal CH4 129

Figure 6.4 Relationship between mean monthly fluorescence in upper 40m and mean CH4 in upper 85m from

2008-2015.

maximum preceded the maximum organic particle fluxes (including fecal pellets) by 1-2 months

(Timothy et al., 2003). This temporal offset could suggest that the spring-time CH4 maximum was not necessarily associated with methanogenesis inside sinking particles or fecal pellets.

However, since the two data sets (particle fluxes and CH4 concentrations) were not measured during the same sampling interval, part of the offset might be attributable to inter-annual variability in the seasonal production cycle. Alternatively, it is possible that smaller suspended particles (which are not efficiently collected in traps) are important as sites of anoxic micro- anaerobic environments, though the smaller sizes of these particles should act to limit O2 depletion in their interior spaces.

130

The relationship between phytoplankton abundance and CH4 concentrations could also be attributable to bacterial activity stimulated by the products of algal metabolism. Methane production from cleavage of methylated compounds has been demonstrated in multiple marine systems (Florez-Leiva et al., 2013; Karl et al., 2008). The production of CH4 from DMS cleavage is a potential source in Saanich, as the highly productive waters of Saanich Inlet are likely to be high in DMSP, which is derived from algal metabolism. Although we lack systematic DMS/P measurements in Saanich Inlet, we have previously documented high (> 15 nM) DMS/P concentrations in coastal BC waters (Asher et al., 2011; Nemcek et al., 2008; Tortell et al.,

2012). Recent isotope tracer studies in the productive upwelling waters of Chile have

-1 -1 demonstrated CH4 production rates of up to 5 nmol L d from DMS (Florez-Leiva et al., 2013), significantly higher than previous estimates of CH4 production in Saanich Inlet (Ward et al.,

1989). Methane production from methylphosphonate has also been suggested as a mechanism for aerobic CH4 production (Karl et al., 2008). However, this pathway appears to be stimulated by phosphorus limitation, which is not a regular feature of Saanich Inlet, making this an unlikely source of CH4.

Sediment resuspension is another potential source of CH4 in the upper 85m of Saanich

Inlet. Rapid sediment resuspension has been previously documented in Saanich Inlet, based on the persistence of Al-rich particle fluxes (Figure 6.3d), which have been attributed to tidal disturbance of sediments on the elevated sill (Matabos et al., 2015; Timothy et al., 2003).

Bioturbation is another important cause of sediment resuspension in Saanich Inlet, with groundfish disturbing the upper 1-2cm of sediments at a rate of >100 events m-2 d-1, and smaller resuspension events (disturbing the upper few mm) occurring >500 times m-2 d-1 (Katz et al.,

2012; Yahel et al., 2008). Oxygen availability is known to control both the distribution and 131

activity of fish and other benthic organisms responsible for sediment resuspension. Fish are restricted to oxygenated waters (>65 uMO2, upper ~60-100m) (Katz et al., 2012), and the activity of these fish and benthic organisms is also influenced by tide-driven changes in O2- availability (Matabos et al., 2015). Therefore, both O2-availability and the distribution/population of sediment-disturbing organisms likely control the rate of CH4-release by bioturbation, potentially exerting a significant influence on CH4 concentrations in the upper

85m. In our dataset, we found evidence for O2-dependent changes in CH4 concentrations, with higher CH4 values during spring, when relatively high water column O2 concentrations would have enabled widespread and active biological sediment disturbance in the upper 85m (Figure

6.2).

The amount of CH4 released by sediment resuspension is unknown. However, a recent microcosm study using sediment cores from an oxygenated lake showed that sediment resuspension events that disturbed the upper 5mm of sediment released between 50-400 µmol

-2 -1 CH4 m h (Bussmann, 2005). Assuming that sediment re-suspension in the upper 85m of

-2 -1 Saanich Inlet releases CH4 at the lower end of this observed range (50 µmol CH4 m h ), and multiplying this by the ~13 km2 sediment surface in the upper 85m (based on surface area of

Saanich Inlet sediments with a depth less than 85m), we estimate a supply of roughly 240 µmol

-2 -1 CH4 m d to the upper 85m of Saanich Inlet. This source term is significantly larger than our

-2 -1 sea-air flux estimates (mean ~20 ± 13 µmol CH4 m d ). This difference is expected since microbial oxidation and lateral advection would remove much of the released CH4 from the inlet before it reached the atmosphere. This simple calculation suggests that sediment resuspension could be an important source of CH4 in the upper 85m of Saanich Inlet. Future work should aim to directly quantify the impact of sedimentary release on the dynamics of CH4 in the upper water 132

column.

Beyond CH4 producing processes in the upper water column, transport from the CH4-rich waters in the deep basin could also contribute CH4 to the upper 85m of Saanich Inlet. Renewal events act to transport stagnant deep basin waters into shallower layers of the water column, potentially bringing CH4-rich water closer to the surface (Anderson and Devol, 1973; Manning et al., 2010). While the strongest renewals are restricted to the late summer/fall period, shallower renewals (above ~120m) can occur throughout the year (Manning et al., 2010), and could therefore supply CH4 to the upper 85m (Ward and Kilpatrick, 1990). This transport process could be further enhanced by the estuarine/reverse-estuarine circulation that brings nutrients to the euphotic zone during spring and summer (Gargett et al., 2003). As discussed below (section

6.4.6), our data show temporally coherent CH4 concentration anomalies in the deep basin and upper 85m on inter-annual time-scales, lending support to the idea that CH4 is supplied to the upper 85m from the deep basin.

As discussed above, seasonal changes in the strength of surface water CH4 sources could account for elevated spring and summer CH4 concentrations in the upper 85m. However, it is equally important to consider seasonal variability in CH4 loss terms, including biological oxidation and sea-air flux. The late summer / early fall reduction in near surface CH4 concentrations is likely to be at least partially attributable to weakening thermal stratification, which would enable deeper mixing and enhanced exchange with the atmosphere, lowering CH4 concentrations in the upper 85 m. Ward and Kilpatrick (1990) proposed that deeper mixing due to reduced thermal stratification and strong winds could account for a decreasing CH4 maximum in the upper 50m of Saanich Inlet. In support of this, we observed surface cooling and isopycnal shoaling beginning in late summer and persisting until spring (data not shown), which would 133

have acted to enhance the ventilation of CH4 in the upper 85m to the atmosphere. Methane oxidation rates have only been measured during summer in Saanich Inlet (Ward et al., 1989), so it is unclear whether seasonal differences in CH4 oxidation rates are important in Saanich Inlet.

Additional CH4 oxidation rate measurements during fall and winter would be useful in evaluating the significance of this potential source of seasonal variability in the CH4 budget.

6.4.4 Factors affecting mid-depth CH4 minimum

The persistent CH4 minimum between 85-110m implies active methanotrophy in this depth horizon. Previous work has shown that CH4 oxidation rates are highest near the oxic- anoxic interface (~2nmol L-1 d-1), and much more rapid than in the upper 85m (max. 0.15nmol L-

1 d-1) (Ward et al., 1989). This suggests that methanotrophs are more abundant and/or more metabolically active in the oxycline than in the upper 85m of the water column. Bacterial abundance is high in the oxycline (Lilley et al., 1982), and recent work from our group has shown the presence of a highly active, diverse metabolic community in this depth zone, including several groups of C1-metabolizing bacteria potentially capable of CH4 -oxidation

(Zaikova et al., 2010).

As shown in Figure 6.5, we observed that the CH4 minimum in the upper 85 m was often associated with a transmissivity minimum, just above the oxic-anoxic transition. This observation suggests that particle loads were elevated in the vicinity of the CH4-minimum, with potential sources derived from resuspended sediments (due to tidal activity and bioturbation, as discussed in section 6.4.3). Sediment resuspension releases both sediments and sediment- associated methanotrophs (Bussmann, 2005). The sediment particles (and any associated sedimentary methanotrophs) sink as they drift towards the center of the inlet (Katz et al., 2012), 134

and could thus contribute to the rapid CH4-oxidation rates observed in the CH4 minimum.

Indeed, the period of maximum particle flux (Jul-Oct, Figure 6.3c,d) roughly corresponded to the period when CH4 concentrations in the mid-depth CH4 minimum were lowest (Jul-Oct, Figure

6.2c), suggesting CH4-oxidation was enhanced during periods when particle loads were highest.

However, we did not detect any significant seasonal differences in transmissivity levels to support this hypothesis.

Figure 6.5 Depth profiles of transmissivity (panel a), O2 (b), and CH4 (c) from May 2013. The shaded patch highlights the transmissivity and CH4 minimum near the oxic-anoxic interface.

135

6.4.5 Factors affecting CH4 in the deep basin

CH4 concentrations increased with depth below the CH4 minimum, and displayed regular seasonal variability, reaching the highest values below 180m during August (mean 1,260 ± 250 nM CH4), and lowest values only ~1month later following renewal (1.2 - 17.7 nM CH4; Figure

6.6d). Oxygen concentrations appear to be the main driver of CH4 variability in the deep basin waters, and the seasonal cycle in O2 levels in these bottom waters is driven by renewal events and subsequent heterotrophic activity.

Renewal waters typically supplied ~20µM O2 to the deep basin, which was consumed within 2-4 months (Figure 6.6a). The absence of oxygen allowed CH4 to accumulate in the deep basin water column over the subsequent months. We estimated rates of CH4 accumulation based on the change in CH4 concentrations at a given depth in between successive sampling periods.

Based on this approach, we found that the accumulation rate of CH4 in the deep basin (below

180m) in the interval between successive renewal events (Oct-Aug) did not exhibit significant

-1 -1 temporal trends (mean 7.5 +/- 6.6 nmol L d ). This suggests that CH4 accumulation rates in the deep basin were not affected by seasonal differences in organic matter flux associated with the spring and summer phytoplankton blooms, implying that CH4 was steadily supplied to the water column from sediments. This is supported by the increasing CH4 concentration with depth below the CH4-minimum. Although previous studies have reported very high CH4 concentrations a few cm below the sediment surface ( >100 µM CH4 at 5cm depth; Devol, 1983), anaerobic CH4 oxidation prevents the vast majority of sedimentary CH4 from entering the water column.

Moreover, any CH4 released from sediments is likely mostly due to molecular diffusion, as the lack of oxygen in the deep basin and low levels of turbulence make sediment resuspension negligible (Manning et al., 2010). Although it may also be possible for CH4 production to occur 136

in the anoxic interior of sinking particles in the water column, we cannot confirm this with our dataset.

The rapid decrease in CH4 concentrations following renewal events (up to -1,800nM

-1 month ) is attributable to the low CH4-content of renewal water, and/or to in situ CH4 oxidation resulting from the influx of O2. We measured the CH4 concentration between 60-250m during

September 2014 in Haro Strait, where renewal water originates, and found an average value of

21.1 +/- 0.8 nM. This concentration is on the upper end of values observed in the post-renewal bottom waters of Saanich Inlet. The general similarity of the bottom water and renewal source water CH4 concentrations, provides reasonably good evidence that much of the post-renewal decrease in CH4 results from flushing of the deep water mass. However, our data do not allow us to rule out a contribution of enhanced CH4 oxidation in the post-renewal period.

6.4.6 Interannual variability

Our discussion above highlights the average seasonal patterns derived from our 7 year time-series. As shown in Figure 6.6, however, we observed significant inter-annual variability in the distribution of CH4 and other biogeochemical properties across our sampling program.

For example, CH4 concentrations remained unusually high in the upper 85m throughout the year from 2009-2010 (Figure 6.6c). This anomaly was not associated with unusual trends in O2 or

- NO3 concentrations or Chla fluorescence near the surface during that period, suggesting it was not related to changes in primary production (Figure 6.6a,b). However, the 2009 – 2010 period

- did exhibit an anomalously weak renewal, shown by the lack of O2 and NO3 accumulation, and incomplete drawdown of H2S and CH4 below 150m during fall of 2009 and 2010 (Figure 6.6).

The simultaneous and similar nature of CH4 anomalies in the deep basin and upper 85m during 137

Figure 6.6 Contour plots of O2 (panel a), NO3 (b), CH4 (c), and H2S (d) from 2008-2014. Black dots represent individual measurements.

138

2009-2010 suggests they are linked, possibly by upward advective transport of CH4-rich water from the deep basin.

The cause of weak renewals during 2009-2010 may be related to El Niño, since a moderate El Niño occurred during 2009-2010, and we also observed a weak/absent deep basin renewal during the weak El-Niño in 2014 (Figure 6.6). El Niño events are associated with reduced summertime upwelling in coastal British Columbia waters (Bylhouwer et al., 2013), and the reduced supply of dense, upwelled water can prevent deep-water renewal in Haro Strait

(Masson, 2002), where Saanich Inlet renewal water originates. The lack of sufficiently dense water in Haro Strait during summer would have reduced the density gradient needed for deep basin renewal in Saanich Inlet, thus limiting the influx of oxygenated, low CH4 water into the deep basin of Saanich Inlet. As a result, deep basin O2 concentrations were relatively low following the 2009, 2010, and 2014 renewals, allowing O2 to be consumed more quickly than during typical years. The rapid consumption of O2 appeared to allow CH4 concentrations to rebound relatively quickly during the post-renewal phase and reach higher concentrations during the subsequent summer (1460 +/- 265 nM in 2010 and 2011) relative to years with larger renewals (1187 +/- 233 nM CH4 in 2012, 2013, and 2014) (Figure 6.6c). Interestingly, the post- renewal CH4 concentrations in Saanich Inlet during El Niño years (12.1 and 17.7 nM during

2009 and 2010, respectively) were higher than during non El Niño years (6.7, 1.2, and 2.8 during

2011, 2012, and 2013, respectively), and similar to the CH4 concentration in Haro Strait during the weak El Niño in September 2014 (21.1 +/- 0.8 nM between 60-250m). This indicates that high CH4 concentrations could be a regional feature in coastal BC waters during El Niño years, and could therefore explain the high CH4 concentrations observed in the upper 85m of Saanich

Inlet during 2009-2010. 139

6.4.7 Decadal-scale trends

Over the seven year period of our time-series (2009-2015), CH4 concentrations in the upper water column (average of 0 - 85 m depth) declined significantly (Spearman’s correlation, p

-1 < 0.05), with a mean trend of ~ -2 nM y (Figure 6.7a). Decreasing CH4 concentrations were associated with a significant decrease in sea-air CH4 flux (Figure 6.7b). In contrast, we observed no statistically significant changes in CH4 concentrations in the vicinity of the water column CH4 minimum (85 - 120 m), or in deep basin waters (below 120 m). Declining CH4 concentrations in the upper portion of Saanich Inlet was accompanied by a significant reduction in O2 throughout the water column (10 - 200 m) over the same period. This apparent de-oxygenation appears to be part of a long-term trend, which can be seen in the long-term data archive from Saanich Inlet,

-1 -1 with a mean rate of O2 loss of 0.37 µmol kg yr (Capelle et al., in prep.).

The apparent decrease in CH4 in the upper 85 m is difficult to explain in the context of declining O2. In theory, the expansion of hypoxia should increase the total volume of anoxic waters conducive to CH4 production, thus increasing upper water CH4 concentrations and sea-air flux. However, de-oxygenation could theoretically affect the depth of the CH4 minimum by shoaling the base of the oxycline, where maximum rates of CH4-oxidation occur. Any upward displacement of this redox-active zone could enhance the rate of CH4 consumption (and thus decrease CH4 concentrations) in the upper 85m. Indeed, over our time-series, we observed a statistically significant shoaling of the suboxic transition zone (i.e. depth where O2 < 5 µM) (-3.4 m y-1; Capelle et al., in prep.). Such shoaling of the hypoxic/suboxic boundary reduces the habitat for aerobic organisms, such as zooplankton, fish and benthic organisms, that supply CH4 to the upper 85 m indirectly through fecal pellets and sediment resuspension. Our results are therefore consistent with a reduction in CH4 concentrations through a shoaling of the CH4- 140

Figure 6.7 Time-series plot of mean CH4 in upper 85 m (a) and CH4 flux (b). Straight black line indicates line of best fit through the data.

minimum and reduced CH4 release from fecal pellets and sediment resuspension, though further work is needed to confirm this. Continued time-series monitoring is also required to determine whether the observed trend is part of a long-term decline in O2 in Saanich Inlet dating back to the

1960s (Capelle et al., in prep.), or if it is simply part of a natural, short-term cycle of variability, such as the El Niño Southern Oscillation, the Pacific Decadal Oscillation, or the large positive temperature anomaly (i.e. ‘The Blob’; Bond et al., 2015) that appeared across the Subarctic

Pacific in 2014.

141

6.5 Conclusion

Our time-series data represent the first thorough analysis of seasonal and interannual variability of CH4 concentrations in Saanich Inlet. We show that CH4 concentrations in Saanich

Inlet undergo seasonal variability, with the highest concentrations and sea-air fluxes during the early summer. Although we are unable to quantify different potential sources of CH4 in the upper

85m, sediment disturbance, production in anoxic microenvironments, DMS-cleavage, and advection from the deep basin CH4 maximum could all play a role. Reduced thermal stratification and enhanced deep mixing could contribute to reduced CH4 concentrations in the upper 85m during fall and winter. The persistent CH4 minimum just above the oxic-anoxic interface appears to be linked to resuspended sediments, which potentially act as a source of methanotrophs to the water column, thus restricting the upward transfer of CH4 from the deep

CH4 maximum. The increase in CH4 with depth in the deep basin suggests CH4 originates in sediments. This deep basin CH4 maximum also displayed seasonal variability, with clear links to the seasonal transition from oxic to anoxic conditions driven by deep basin renewal in the fall.

Inter-annual variability in deep basin renewal associated with El Niño appeared to influence CH4 concentrations in both the deep basin and upper 85m.

The compilation of repeated seasonal and inter-annual observations enables a more comprehensive assessment of CH4 variability (concentrations and sea-air fluxes) in Saanich Inlet, and provides a baseline for monitoring future changes in CH4 dynamics. Even over our relatively short time-series, we have observed significant declines in CH4 concentrations and sea- air fluxes in the upper 85m. This may be an unexpected response to de-oxygenation, possibly related to the ongoing decline in O2 and shoaling of the suboxic boundary in Saanich Inlet, which could reduce indirect CH4 supply and increases CH4 consumption near the upper 85 m zone. 142

This result is significant in light of the on-going de-oxygenation of Subarctic Pacific waters

(Crawford and Peña, 2013; Whitney et al., 2007), and given the relatively large contribution of shallow regions (~75% from continental shelves and estuaries) to global marine CH4 emissions

(Bange et al., 1994).

143

Chapter 7: Conclusion

Our understanding of marine CH4 and N2O cycling is presently limited in a number of important areas. Global inventories and sea-air fluxes are subject to significant uncertainty due to a lack of data from coastal regions, and the factors controlling N2O and CH4 concentrations in coastal environments are not well constrained. These shortcomings are significant given the relatively large CH4 and N2O emissions from coastal ocean systems, and the potential effects of ongoing OMZ expansion and warming on CH4 and N2O cycling. The research presented in this dissertation addresses these uncertainties, and contributes to an improved understanding of N2O and CH4 cycling in coastal waters of the Subarctic Pacific Ocean. Below, I summarize the important contributions of this work, and suggest areas of future research.

7.1 Research implications

Expanding the global inventory of N2O and CH4 measurements requires the analysis of large number of samples, including those obtained from regular time-series monitoring programs at sites around the world. The standard CH4 and N2O analysis methods used in most laboratories typically require manual processing of discrete samples, often with separate treatment of CH4 and N2O measurements. This is tedious, time-consuming, and prone to operator error. The development of high through-put automated methods is thus important to supporting large-scale measurement programs. The first contribution of this thesis was the development of such a method. Our automated purge and trap system coupled to a gas-chromatography mass- spectrometer (PT-GCMS) dramatically improved the efficiency of measuring dissolved CH4 and

N2O samples relative to manual analysis, while also minimizing the potential for human error during sample processing. These improvements were achieved without sacrificing precision, 144

sensitivity, or accuracy of our measurements, as shown by our extensive testing and calibration against internationally certified standards and inter-comparison with other conventional systems from leading laboratories in a number of countries.

The construction of an automated purge and trap system was important to achieve the seasonal and interannual resolution needed for this research. Using this improved analytical method, I was able to analyze over 4,000 discrete samples from coastal waters of the British

Columbia continental shelf and our time-series in Saanich Inlet. This represents a significant contribution to the existing global inventory of N2O and CH4 measurements. The automated purge and trap system could be also be integrated with other conventional CH4 and N2O detectors (e.g. flame-ionization detectors and electron-capture detectors), to facilitate the application of this method in laboratory groups who may not have access to a GC-MS. Increased use of high through-put methods would help improve data coverage in other ocean regions.

Using multi-year observations of N2O and CH4 concentration measurements from the

WCVI study area, we were able to estimate the relative significance of several biological, physical, and chemical controls on the concentrations of these gases in coastal BC waters. We found significant seasonal and inter-annual variability in CH4 and N2O distributions, which appeared to be driven by a few key factors. The seasonal and inter-annual resolution more accurately constrains mean annual concentration and sea-air flux budgets, as compared with other ocean regions with more limited data coverage. We found that O2 availability was one of the most important controls on distributions of N2O and CH4. Oxygen concentrations correlated strongly with CH4 and N2O, and defined the thresholds separating net production and consumption. Our research also showed that upwelling accelerates the delivery of CH4 and N2O to the surface, resulting in enhanced sea-air fluxes, as has been reported in other coastal 145

upwelling systems (Farías et al., 2015; Kock et al., 2008; Rehder et al., 2002; Wittke et al.,

2010). Our data also showed strong qualitative evidence that hydrothermal seeps provide a significant, and likely previously underestimated, quantity of CH4 in coastal shelf waters of southern BC. By comparison, N2O was likely derived from a combination of advection from the off-shelf N2O maximum and in situ water column production by nitrification. Using differences

- in N2O, NO3 , and O2 concentrations along isopycnals, we derived quantitative estimates of the supply of N2O from advection vs. in situ production, and showed that seasonal and inter-annual changes in O2 availability were related to variable N2O-yields from nitrification. These results were based on relatively simple models of water mass circulation, which ignored the effects of diapycnal mixing and sedimentary N2O sources. The processes could be significant and should be investigated in future studies.

In Saanich Inlet, our CH4 and N2O measurements were part of a larger collaborative research effort, comprising one of the few existing time-series measurements of both metagenomics and dissolved CH4 and N2O. The CH4 and N2O data showed pronounced seasonal and inter-annual variability, which had not been resolved previously in the limited N2O and CH4 measurements available from Saanich Inlet. Our data provided new insight into the sources, sinks, and drivers of dissolved gas variability across the strong redox gradients in Saanich Inlet.

Maximum near surface N2O concentrations in Saanich Inlet typically occurred during late summer, likely resulting from intensification and shoaling of the oxycline which would facilitate additional N2O supply from partial denitrification and increased N2O-yields from nitrification.

Deep basin N2O concentrations were supplied within renewal waters and declined almost immediately following renewal due to denitrification, reaching undetectable concentrations (< ~1 nM) within ~4 months. Our data enabled us to determine the O2 thresholds separating net N2O 146

production from consumption (~10 µM O2). Above this threshold, N2O is produced likely through a combination of nitrification and partial denitrification. The widespread detection of the

N2O reductase gene, nosZ, could help to explain the relatively low N2O concentrations in

Saanich Inlet relative to other open ocean OMZs. Metagenomic data also provided evidence for a distributed denitrification pathway, with some organisms containing genes for only one or a few steps of denitrification.

The time-series data from Saanich Inlet provided the first evidence of significant decadal- scale trends in N2O concentrations and sea-air flux associated with de-oxygenation, underscoring the value of time-series monitoring efforts. We combined our time-series observations with archive data from 1960-2007 to show that recent de-oxygenation is part of a long-term (and potentially accelerating) trend across the Subarctic Pacific. Changes in N2O, as observed in

Saanich Inlet, could be therefore be widespread throughout other marine systems experiencing

OMZ expansion, with potentially significant climate implications. Time-series monitoring is uniquely able to identify and quantify such longer-term changes, and to examine potential biogeochemical responses to environmental perturbations on various time-scales. Continued time-series monitoring is still required to document the trends over more significant time-scales, and to ensure that the observed trends are not part of short-term (decadal) cycles of variability.

As with N2O, CH4 concentrations in Saanich Inlet displayed pronounced seasonal and depth-dependent variability. Maximum near-surface (upper 85 m) CH4 concentrations occurred during spring and early summer, coincident with the spring bloom. In situ production

(potentially from methanogenesis within particles, and cycling of DMS/P), vertical transport from the deep-basin CH4 maximum, and lateral supply from bioturbated sediments are potential sources for this upper water column signature. The mid-depth (85 - 120 m) minimum in CH4 147

concentrations was associated with a region of intense CH4-oxidation, near the base of the oxycline. Methane oxidizers in this region could be transported with resuspended sediments that contribute to the turbidity maximum. In the deep basin, sedimentary sources likely dominate

CH4 supply, with concentrations decreasing following renewal with lower CH4 containing waters. On inter-annual time-scales, we found evidence for CH4 variability associated with weak/absent deep basin renewals during El Niño years. Our time-series observations also revealed a significant decline in CH4 concentrations in the upper 85 m between 2009 and 2015, coincident with decreasing O2 Concentrations. This somewhat unexpected observation may reflect the effects of a shoaling oxycline, which would reduce CH4 supply from sediment resuspension due to loss of suitable habitat for fish and other microorganisms, while increasing

CH4 oxidation in the upper 85 m due to a shoaling of the CH4-minimum. Continued monitoring is needed to determine whether the decline in CH4 is related to longer-term marine de- oxygenation, or if it is part of a shorter-term (e.g. decadal) cycle.

7.2 Future work

Although this research contributed to our understanding of CH4 and N2O cycling in coastal BC waters, it simultaneously highlighted a number of areas for further study. Below I discuss the areas that I feel most warrant future research consideration.

The PT-GCMS was instrumental in measuring large numbers of discrete samples. Sample throughput could be increased by adding a third 13-position selector valve, and this would be most beneficial during overnight runs when the current system sits idle for up to 10 hours. The

PT-GCMS is also only using a tiny portion of its analytical capacity. The detector is capable of measuring various volatiles of interest in marine settings, including CO2, which could be 148

measured by adding an automated headspace equilibration system. The PT-GCMS could also be used to measure isotopically labelled gases, provided the concentrations are above detection limits. These isotopic measurements would not likely have the analytical precision to detect changes in natural abundance ratios, but would facilitate tracer experiments with isotopically- labelled substrates. These future analytical developments will require both time and attention to details. It took us roughly 9 months to build the automated PT sampling system, and during the course of our analysis, we experienced a number of technical issues resulting in prolonged periods (months) of downtime for repair. Over time, with gained experience, these issues were resolved more expediently and the system has become significantly more robust. The method development also provided a significant opportunity to learn basic electronics and computer- based programming in the LabVIEW environment.

The N2O and CH4 concentration measurements used in this study have proven extremely helpful in identifying the primary controls on the cycling of these gases in our study regions.

However, the adoption of natural abundance isotopic measurements, and incubation-based rate measurements using isotopic tracers and selective inhibitors would provide a finer level of detail on the rates of supply and loss of these gases from different biological processes. The specific areas worth exploring include: i) rates of N2O production from bacterial nitrification, archaeal nitrification, nitrifier-denitrification, and partial denitrification across a range of O2 concentrations; ii) rates of N2O consumption across a range of O2 concentrations; iii) rates of

N2O and CH4 supply from sediments (including bubble plumes, bioturbation, and tidal disturbance) and anoxic microenvironments; iv) potential CH4 production from DMS-cleavage or methylphosphonate cleavage; v) potential N2O production from additional pathways such as

DNRA and Anammox. In addition to these rate measurements, continued exploration of meta- 149

genomic signatures would be useful. Specific targets would include the activity of N2O- reductase gene in well-oxygenated waters of Saanich Inlet and the size-fractionated abundance/activity of various N2O and CH4 related genes and proteins. For example, it would be interesting to quantify the numbers of particle-associated methanotrophs in the CH4 minimum zone of Saanich Inlet.

I did not attempt to measure natural isotope abundances during the course of this research. However, I attempted several selective inhibitor experiments to identify the N2O production from nitrification and partial denitrification, and tracer experiments to measure CH4 oxidation as well as CH4-production from DMS-cleavage. The data produced by these incubations were inconclusive, possibly due to a number of factors such as O2-contamination of samples, bottle-effects, or physiological stress induced on the microbial assemblages during transportation from the field to laboratory. However, these methods have been successful in other regions, and they should be adaptable for use at UBC. Indeed, a number of students in collaborating groups in the UBC Microbiology and Immunology department have successfully deployed some rate measurements of key processes of the N2O cycle in Saanich Inlet. The ongoing collaboration between several groups at UBC (Drs. Tortell, Hallam and Crowe) should be continued and strengthened into the future in order to gain a broader, cross-disciplinary perspectives on the biogeochemical cycles of low oxygen marine systems.

This research highlighted the value of making repeated measurements of N2O and CH4 to identify seasonal, inter-annual, and decadal-scale trends in the distributions of these gases.

Continued monitoring is needed to determine whether these trends are cyclical or part of long- term trends related to widespread de-oxygenation. Future monitoring efforts could improve upon the seasonal and spatial resolution in the WCVI study region, which only sampled depth- 150

resolved profiles from two transects during late spring and late summer. Increased spatial resolution in the surface layer could be achieved by adding continuous underway shipboard measurements using small cavity long-path length laser-absorption based detectors. Increased depth-resolution of concentration in the upper 75 m of Saanich Inlet would also be useful, as would improving the seasonal coverage and depth resolution of metagenomics samples in the upper 100 m of Saanich Inlet.

I would like to conclude by stating that I believe the most important area of future research is the continued monitoring of CH4 and N2O in our study regions, and I strongly encourage the establishment of new time-series stations monitoring N2O and CH4 covering a range of ocean biogeochemical provinces. Making careful, accurate measurements is crucial to understanding past, present, and future changes in the ocean environment. To paraphrase something I recently heard at an oceanography seminar - 'you can always repeat experiments, but you can’t re-measure the ocean once it’s changed'.

151

References

Altabet, M.A., Francois, R., Murray, D.W., Prell, W.L., 1995. Climate-related variations in

denitrification in the Arabian Sea from sediment 15N/14N ratios. Nature 373, 506–509.

Amouroux, D., 2002. Biogenic Gas (CH4, N2O, DMS) Emission to the Atmosphere from Near-

shore and Shelf Waters of the North-western Black Sea. Estuar. Coast. Shelf Sci. 54, 575–

587. doi:10.1006/ecss.2000.0666

Anderson, I.C., Poth, M., Homstead, J., Burdige, D., 1993. A comparison of NO and N2O

production by the autotrophic nitrifier Nitrosomonas europaea and the heterotrophic nitrifier

Alcaligenes faecalis. Appl. Environ. Microbiol. 59, 3525–33.

Anderson, J.J., Devol, A.H., 1973. Deep Water Renewal in Saanich Inlet, an Intermittently

Anoxic Basin. Estuar. Coast. Mar. Sci. 1, 1–10.

Archer, D., 2007. Methane hydrate stability and anthropogenic climate change. Biogeosciences

Discuss. 4, 993–1057. doi:10.5194/bgd-4-993-2007

Arévalo-Martínez, D.L., Kock, A., Löscher, C.R., Schmitz, R.A., Bange, H.W., 2015. Massive

nitrous oxide emissions from the tropical South Pacific Ocean. Nat. Geosci. 8, 530–533.

doi:10.1038/ngeo2469

Armstrong, F.A.J., Stearns, C.R., Strickland, J.D.H., 1967. The measurement of upwelling and

subsequent biological process by means of the Technicon Autoanalyzer® and associated

equipment. Deep Sea Res. Oceanogr. Abstr. 14, 381–389.

doi:http://dx.doi.org/10.1016/0011-7471(67)90082-4

Asher, E.C., Merzouk, A., Tortell, P.D., 2011. Fine-scale spatial and temporal variability of

surface water dimethylsufide (DMS) concentrations and sea-air fluxes in the NE Subarctic

Pacific. Mar. Chem. 126, 63–75. doi:10.1016/j.marchem.2011.03.009 152

Bakun, A., 1990. Global climate change and intensification of coastal ocean upwelling. Science

247, 198–201. doi:10.1126/science.247.4939.198

Bange, H., Rapsomanikis, S., Andraea, M.O., 1996. Nitrous oxide in coastal waters. Global

Biogeochem. Cycles 10, 197–207.

Bange, H.W., 2008. Gaseous Nitrogen Compounds (NO, N2O, N2, NH3) in the Ocean, in:

Nitrogen in the Marine Environment. Elsevier, pp. 51–94. doi:10.1016/B978-0-12-372522-

6.00002-5

Bange, H.W., 2006. Nitrous oxide and methane in European coastal waters. Estuar. Coast. Shelf

Sci. 70, 361–374. doi:10.1016/j.ecss.2006.05.042

Bange, H.W., Andreae, M.O., Lal, S., Law, C.S., Naqvi, S.W.A., Patra, P.K., Rixen, T., Upstill-

Goddard, R.C., 2001. Nitrous oxide emissions from the Arabian Sea: A synthesis. Atmos.

Chem. Phys. Discuss. 1, 167–192. doi:10.5194/acpd-1-167-2001

Bange, H.W., Bartell, U.H., Rapsomanikis, S., Andreae, M.O., 1994. Methane in the Baltic and

North Seas and a reassessment of the marine emissions of methane. Global Biogeochem.

Cycles 8, 465–480.

Bange, H.W., Bergmann, K., Hansen, H.P., Kock, A., Koppe, R., Malien, F., Ostrau, C., 2010.

Dissolved methane during hypoxic events at the Boknis Eck time series station

(Eckernforde Bay, SW Baltic Sea). Biogeosciences 7, 1279–1284.

Bange, H.W., Dahlke, S., Rmaesh, R., Meyer-Reil, L.-A., Rapsomanikis, S., Andreae, M.O.,

1998. Seasonal Study of Methane and Nitrous Oxide in the Coastal Waters of the Southern

Baltic Sea. Estuar. Coast. Shelf Sci. 47, 807–817.

Barwell-Clarke, J., Whitney, F. a, 1996. Institute of Ocean Sciences nutrient methods and

analysis. Can. Tech. Rep. Hydrogr. Ocean Sci. 49. 153

Beal, E.J., House, C.H., Orphan, V.J., 2009. Manganese- and iron-dependent marine methane

oxidation. Science (80-. ). 325, 184–187. doi:10.1126/science.1169984

Betlach, M.R., Tiedje, J.M., 1981. Kinetic explanation for accumulation of nitrite, nitric oxide,

and nitrous oxide during bacterial denitrification. Appl. Environ. Microbiol. 42, 1074–1084.

doi:Article

Bianchi, D., Dunne, J.P., Sarmiento, J.L., Galbraith, E.D., 2012. Data-based estimates of

suboxia, denitrification, and N2O production in the ocean and their sensitivities to dissolved

O2. Glob. Biogeochem. Cycles 26, GB2009. doi:10.1029/2011gb004209

Boles, J.R., Clark, J.F., Leifer, I., Washburn, L., 2001. Temporal variation in natural methane

seep rate due to tides, Coal Oil Point area, California. J. Geophys. Res.

doi:10.1029/2000JC000774

Bond, N.A., Cronin, M.F., Freeland, H., Mantua, N., 2015. Causes and impacts of the 2014

warm anomaly in the NE Pacific. Geophys. Res. Lett. 42, 3414–3420.

doi:10.1002/2015GL063306.Received

Bourbonnais, A., Lehmann, M.F., Hamme, R.C., Manning, C.C., Juniper, S.K., 2013. Nitrate

elimination and regeneration as evidenced by dissolved inorganic nitrogen isotopes in

Saanich Inlet, a seasonally anoxic fjord. Mar. Chem. 157, 194–207.

doi:10.1016/j.marchem.2013.09.006

Bussmann, I., 2005. Methane release through resuspension of littoral sediment. Biogeochemistry

74, 283–302. doi:10.1007/s

Bylhouwer, B., Ianson, D., Kohfeld, K., 2013. Changes in the onset and intensity of wind-driven

upwelling and downwelling along the North American Pacific coast. J. Geophys. Res.

Ocean. 118, 2565–2580. doi:10.1002/jgrc.20194 154

Capelle, D.W., Dacey, J.W., Tortell, P.D., 2015. An automated, high through-put method for

accurate and precise measurements of dissolved nitrous-oxide and methane concentrations

in natural waters. Limnol. Oceanogr. Methods 13, 345–355. doi:10.1002/lom3.10029

Capelle, D.W., Tortell, P.D., 2016. Factors controlling methane and nitrous-oxide variability in

the southern British Columbia coastal upwelling system. Mar. Chem. 179, 56–67.

doi:10.1016/j.marchem.2016.01.011

Careri, M., Mori, G., Musci, M., Viaroli, P., 1999. Evaluation of dynamic headspace and purge-

and-trap techniques for the high-resolution gas chromatography analysis of nitrous oxide in

seawater. J. Chromatogr. A 848, 327–335.

Carter, L., 1973. Surficial Sediments of Barkley Sound and the Adjacent Continental Shelf, West

Coast Vancouver Island. Can. J. Earth Sci. 10, 441–459.

Casciotti, K.L., Buchwald, C., 2012. Insights on the marine microbial nitrogen cycle from

isotopic approaches to nitrification. Front. Microbiol. 3, 1–14.

doi:10.3389/fmicb.2012.00356

Castro-González, M., Farías, L., 2004. N2O cycling at the core of the oxygen minimum zone off

northern Chile. Mar. Ecol. Ser. 280, 1–11.

Chow, C.E.T., Winget, D.M., White, R.A., Hallam, S.J., Suttle, C.A., 2015. Combining genomic

sequencing methods to explore viral diversity and reveal potential virus-host interactions.

Front. Microbiol. 6, 1–15. doi:10.3389/fmicb.2015.00265

Cline, J.D., 1969. Spectrophotometric Determination of Hydrogen Sulfide in Natural Waters.

Limnol. Oceanogr. 14, 454–458.

Codispoti, L.A., 2010. Interesting times for marine N2O. Science (80-. ). 327, 1339–40.

doi:10.1126/science.1184945 155

Codispoti, L.A., Brandes, J., Christensen, J.P., Devol, A.H., 2001. The oceanic fixed nitrogen

and nitrous oxide budgets: Moving targets as we enter the anthropocene? Sci. Mar. 65, 85–

105.

Cohen, Y., 1978. Consumption of dissolved nitrous oxide in an anoxic basin, Saanich Inlet,

British Columbia. Nature 272, 235–237.

Cohen, Y., Gordon, I., 1979. Nitrous oxide production in the Ocean. J. Geophys. Res. 84.

Cornejo, M., Farías, L., 2012. Meridional variability of the vertical structure and air–sea fluxes

of N2O off central Chile (30–40°S). Prog. Oceanogr. 92-95, 33–42.

doi:10.1016/j.pocean.2011.07.016

Costa, E., Pérez, J., Kreft, J.U., 2006. Why is metabolic labour divided in nitrification? Trends

Microbiol. 14, 213–219. doi:10.1016/j.tim.2006.03.006

Crawford, W.R., Peña, M.A., 2013. Declining Oxygen on the British Columbia Continental

Shelf. Atmosphere-Ocean 51, 88–103. doi:10.1080/07055900.2012.753028

Cunnold, D.M., Steele, L.P., Fraser, P.J., Simmonds, P.G., Prinn, R.G., Weiss, R.F., Porter,

L.W., O’Doherty, S., Langenfelds, R.L., Krummel, P.B., Wang, H.J., Emmons, L., Tie,

X.X., Dlugokencky, E.J., 2002. In situ measurements of atmospheric methane at

GAGE/AGAGE sites during 1985–2000 and resulting source inferences. J. Geophys. Res.

107, 98. doi:10.1029/2001JD001226

Currie, L.A., 1968. Limits for qualitative detection and quantitative determination. Application

to radiochemistry. Anal. Chem. 40, 586–593. doi:10.1021/ac60259a007

Cynar, F.J., Yayanos, A.A., 1992. The distribution of methane in the upper water of the Southern

California Bight. J. Geophys. Res. 97, 11,211–269,285.

Dalsgaard, T., Stewart, F.J., Thamdrup, B., De Brabandere, L., Revsbech, N.P., Ulloa, O., 156

Canfield, D.E., DeLong, E.F., 2014. Oxygen at Nanomolar Levels Reversibly Suppresses

Process Rates and Gene Expression in Anammox and Denitrification in the Oxygen

Minimum Zone off Northern Chile. MBio 5, 1–14. doi:10.1128/mBio.01966-14.Editor

Damm, E., Helmke, E., Thoms, S., Schauer, U., 2010. Methane production in aerobic

oligotrophic surface water in the central Arctic Ocean. Biogeosciences 7, 1099–1108.

De Angelis, M. a., Lee, C., 1994. Methane production during zooplankton grazing on marine

phytoplankton. Limnol. Oceanogr. 39, 1298–1308.

De Wilde, H.P.J., De Bie, M.J.M., 2000. Nitrous oxide in the Schelde estuary: Production by

nitrification and emission to the atmosphere. Mar. Chem. 69, 203–216. doi:10.1016/S0304-

4203(99)00106-1

Devol, A.H., 1983. Methane oxidation rates in the anaerobic sediments of Saanich Inlet. Limnol.

Oceanogr. 28, 738–742. doi:10.4319/lo.1983.28.4.0738

Devol, A.H., Anderson, J.J., Kuivila, K., Murray, J.W., 1984. A model for coupled sulfate

reduction and methane oxidation in the sediments of Saanich Inlet. Geochim. Cosmochim.

Acta 48, 993–1004. doi:10.1016/0016-7037(84)90191-1

Dickens, G.R., Neil, J.R.O., Rea, D.K., Owen, R.M., 1995. Dissociation of oceanic methane

hydrate as a cause of the carbon isotope excursion at the end of the Paleocene.

Paleoceanography 10, 965–971. doi:10.1029/95PA02087

Dore, J.E., Karl, D.M., 1996. Nitrification in the euphotic zone as a source for nitrite, nitrate, and

nitrous oxide at Station ALOHA. Limnol. Oceanogr. 41, 1619–1628.

doi:10.4319/lo.1996.41.8.1619

Edgcomb, V.P., Pachiadaki, M., 2014. Ciliates along oxyclines of permanently stratified marine

water columns. J. Eukaryot. Microbiol. 61, 434–445. doi:10.1111/jeu.12122 157

Ekeberg, D., Ogner, G., Fongen, M., Joner, J., Wickstrom, T., 2004. Determination of CH4, CO2

and N2O in air samples and soil atmosphere by gas chromatography mass spectrometry,

GC-MS. J. Environ. Monit. 6, 621–623.

Elkins, J.W., 1980. Determination of dissolved nitrous oxide in aquatic systems by gas

chromatography using electtron-capture detection and multiple phase equilibration. Anal.

Chem. 52, 263–267.

Ettwig, K.F., Butler, M.K., Le Paslier, D., Pelletier, E., Mangenot, S., Kuypers, M.M.M.,

Schreiber, F., Dutilh, B.E., Zedelius, J., de Beer, D., Gloerich, J., Wessels, H.J.C.T., van

Alen, T., Luesken, F., Wu, M.L., van de Pas-Schoonen, K.T., Op den Camp, H.J.M.,

Janssen-Megens, E.M., Francoijs, K.-J., Stunnenberg, H., Weissenbach, J., Jetten, M.S.M.,

Strous, M., 2010. Nitrite-driven anaerobic methane oxidation by oxygenic bacteria. Nature

464, 543–548. doi:10.1038/nature08883

Falkowski, P., Algeo, T., Codispoti, L., 2011. Ocean Deoxygenation: Past, Present, and Future.

Eos (Washington. DC). 92, 409–420.

Farías, L., Besoain, V., García-Loyola, S., 2015. Presence of nitrous oxide hotspots in the coastal

upwelling area off central Chile: an analysis of temporal variability based on ten years of a

biogeochemical time series. Environ. Res. Lett. 10, 044017. doi:10.1088/1748-

9326/10/4/044017

Farías, L., Castro-González, M., Cornejo, M., Charpentier, J., Faúndez, J., Boontanon, N.,

Yoshida, N., 2009. Denitrification and nitrous oxide cycling within the upper oxycline of

the eastern tropical South Pacific oxygen minimum zone. Limnol. Oceanogr. 54, 132–144.

doi:10.4319/lo.2009.54.1.0132

Farias, L., Fernandez, C., Faundez, J., Cornejo, M., Alcaman, M.E., 2009. 158

Chemolithoautotrophic production mediating the cycling of the greenhouse gases N2O and

CH4 in an upwelling ecosystem. Biogeosciences 6, 3053–3069.

Florez-Leiva, L., Damm, E., Farías, L., 2013. Methane production induced by dimethylsulfide in

surface water of an upwelling ecosystem. Prog. Oceanogr. 112-113, 38–48.

doi:10.1016/j.pocean.2013.03.005

Foreman, M., 2000. Seasonal current simulations for the western continental margin of

Vancouver Island. J. Geophys. Res. 105, 19665–19698.

Frame, C., Deal, E., Nevison, C.D., Casciotti, K.L., 2014. N2O production in the eastern South

Atlantic: Analysis of N2O stable isotopic and concentration data. Global Biogeochem.

Cycles 28, 1262–1278. doi:10.1002/2013GB004790.Received

Frame, C.H., Casciotti, K.L., 2010. Biogeochemical controls and isotopic signatures of nitrous

oxide production by a marine ammonia-oxidizing bacterium. Biogeosciences 7, 2695–2709.

doi:10.5194/bg-7-2695-2010

Francis, C. a, Roberts, K.J., Beman, J.M., Santoro, A.E., Oakley, B.B., 2005. Ubiquity and

diversity of ammonia-oxidizing archaea in water columns and sediments of the ocean. Proc.

Natl. Acad. Sci. U. S. A. 102, 14683–8. doi:10.1073/pnas.0506625102

Freing, A., Bange, H.W., Wallace, D.W.R., 2012. Global oceanic production of nitrous oxide.

Philos. Trans. R. Soc. B Biol. Sci. 367, 1245–1255. doi:10.1098/rstb.2011.0360

Ganesh, S., Bristow, L.A., Larsen, M., Sarode, N., Thamdrup, B., Stewart, F.J., 2015. Size-

fraction partitioning of community gene transcription and nitrogen metabolism in a marine

oxygen minimum zone. ISME J. 9, 1–15. doi:10.1038/ismej.2015.44

Gao, H., Schreiber, F., Collins, G., Jensen, M.M., Kostka, J.E., Lavik, G., de Beer, D., Zhou, H.,

Kuypers, M.M.M., 2009. Aerobic denitrification in permeable Wadden Sea sediments. 159

ISME J. 4, 417–426. doi:10.1038/ismej.2009.127

Garcia, H.E., Locarnini, R.A., Boyer, T.P., Antonov, J.I., Baranova, O.K., Zweng, M.M.,

Johnson, D.R., 2009. World Ocean Atlas 2009, Volume 3: Dissolved Oxygen, Apparent

Oxygen Utilization, and Oxygen Saturation, S. Levitus, Eds. NOAA Atlas NESDIS 70. U.S.

Government Printing Office, Washington, D.C.

Gargett, A., Stucchi, D., Whitney, F., 2003. Physical processes associated with high primary

production in Saanich Inlet, British Columbia. Estuar. Coast. Shelf Sci. 56, 1141–1156.

Goreau, T.J., Kaplan, W.A., Wofsy, S.C., McElroy, M.B., Valois, F.W., Watson, S.W., 1980.

Production of NO2 and N2O by nitrifying bacteria at reduced concentrations of oxygen.

Appl. Environ. Microbiol. 40, 526–532.

Grant, N.J., Whiticar, M.J., 2002. Stable carbon isotopic evidence for methane oxidation in

plumes above Hydrate Ridge, Cascadia Oregon Margin. Global Biogeochem. Cycles 16,

1124. doi:10.1029/2001GB001851

Grundle, D.S., Juniper, S.K., 2011. Nitrification from the lower euphotic zone to the sub-oxic

waters of a highly productive British Columbia fjord. Mar. Chem. 126, 173–181.

doi:10.1016/j.marchem.2011.06.001

Grundle, D.S., Maranger, R., Juniper, S.K., 2012. Upper Water Column Nitrous Oxide

Distributions in the Northeast Subarctic Pacific Ocean. Atmosphere-Ocean 50, 475–486.

doi:10.1080/07055900.2012.727779

Hawley, A.K., Brewer, H.M., Norbeck, A.D., Pa a-Toli, L., Hallam, S.J., 2014. Metaproteomics

reveals differential modes of metabolic coupling among ubiquitous oxygen minimum zone

microbes. Proc. Natl. Acad. Sci. 111, 11395–11400. doi:10.1073/pnas.1322132111

Heeschen, K.U., Collier, R.W., de Angelis, M.A., Suess, E., Rehder, G., Linke, P., 160

Klinkhammer, G.P., 2005. Methane sources, distributions, and fluxes from cold vent sites at

Hydrate Ridge, Cascadia Margin. Global Biogeochem. Cycles 19, GB2016.

doi:10.1029/2004GB002266

Herlinveaux, R.H., 1962. Oceanography of Saanich Inlet in Vancouver Island, British Columbia.

J. Fish. Res. Board Canada 19, 1–37.

Holmes, E.M., Sansone, F.J., Rust, T.M., Popp, B.N., 2000. Methane production, consumption,

and air-sea exchange in the open ocean: An evaluation based on carbon isotopic ratios.

Global Biogeochem. Cycles 14, 1–10.

Holmes, R.M., Aminot, A., Kerouel, R., Hooker, B.A., Peterson, B.J., 1999. A simple and

precise method for measuring ammonium in marine and freshwater ecosystems. Can. J.

Fish. Aquat. Sci. 56, 1801–1808. doi:10.1139/cjfas-56-10-1801

Hooper, A.B., Terry, K.R., 1973. Specific inhibitors of ammonia oxidation in Nitrosomonas. J.

Bacteriol. 115, 480–485.

Ianson, D., Feely, R.A., Sabine, C.L., Juranek, L.W., 2009. Features of coastal upwelling regions

that determine net air-sea CO2 flux. J. Oceanogr. 65, 677–687. doi:10.1007/s10872-009-

0059-z

IPCC, 2013. Technical Summary, in: Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V.,

Midgley, P.M. (Eds.), Climate Change 2013: The Physical Science Basis. Contribution of

Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate

Change. doi:10.1017/CBO9781107415324

Jayakumar, A., O’Mullan, G.D., Naqvi, S.W.A., Ward, B.B., 2009. Denitrifying bacterial

community composition changes associated with stages of denitrification in oxygen

minimum zones. Microb. Ecol. 58, 350–362. doi:10.1007/s00248-009-9487-y 161

Jia, Z., Liu, T., Xia, X., Xia, N., 2016. Effect of particle size and composition of suspended

sediment on denitrification in river water. Sci. Total Environ. 541, 934–40.

doi:10.1016/j.scitotenv.2015.10.012

Judd, A., Hovland, M., Dimitrov, L., 2002. The geological methane budget at continental

margins and its influence on climate change. Geofluids 2, 109–126.

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, S., Saha, S.,

White, G., Zhu, Y., Leetmaa, A., Reynolds, R., Chelliah, M., Ebisuzaki, W., Higgins, W.,

Janowiak, J., Mo, K.C., Ropelewski, C., Wang, J., Jenne, R., Joseph, D., Iredell, M., Saha,

S., White, G., Woollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J.,

Mo, K.C., Ropelewski, C., Wang, J., Leetmaa, A., Reynolds, R., Jenne, R., Joseph, D.,

1996. The NCEP/NCAR 40-Year Reanalysis Project. Bull. Am. Meteorol. Soc.

doi:http://dx.doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2

Karl, D.M., Beversdorf, L., Björkman, K.M., Church, M.J., Martinez, A., Delong, E.F., 2008.

Aerobic production of methane in the sea. Nat. Geosci. 1, 473–478. doi:10.1038/ngeo234

Karl, D.M., Lukas, R., 1996. The Hawaii Ocean Time-series (HOT) program: Background,

rationale and field implementation. Deep Sea Res. Part II Top. Stud. Oceanogr. 43, 129–

156. doi:10.1016/0967-0645(96)00005-7

Katz, T., Yahel, G., Reidenbach, M.A., Tunnicliffe, V., Herut, B., Crusius, J., Whitney, F.,

Snelgrove, P.V.R., Lazar, B., 2012. Resuspension by fish facilitates the transport and

redistribution of coastal sediments. Limnol. Oceanogr. 57, 945–958.

doi:10.4319/lo.2012.57.4.0945

Keeling, R.E., Körtzinger, A., Gruber, N., 2010. Ocean deoxygenation in a warming world. Ann.

Rev. Mar. Sci. 2, 199–229. doi:10.1146/annurev.marine.010908.163855 162

Kock, A., Arévalo-Martínez, D.L., Löscher, C.R., Bange, H.W., 2015. Differences between

coastal and open ocean distributions of N2O in the oxygen minimum zone off Peru.

Biogeosciences Discuss. 12, 10167–10193. doi:10.5194/bgd-12-10167-2015

Kock, A., Gebhardt, S., Bange, H.W., 2008. Methane emissions from the upwelling area off

Mauritania (NW Africa). Biogeosciences 5, 1119–1125. doi:10.5194/bg-5-1119-2008

Kohfeld, K.E., Le Quere, C., Harrison, S.P., Anderson, R.F., 2005. Role of marine biology in

glaciat-interglacial CO2 cycles. Science (80-. ). 308, 74–78. doi:10.1126/science.1105375

Koike, I., Hattori, a, 1978. Denitrification and ammonia formation in anaerobic coastal

sediments. Appl. Environ. Microbiol. 35, 278–82.

Konwar, K.M., Hanson, N.W., Bhatia, M.P., Kim, D., Wu, S.J., Hahn, A.S., Morgan-Lang, C.,

Cheung, H.K., Hallam, S.J., 2015. MetaPathways v2.5: Quantitative functional, taxonomic

and usability improvements. Bioinformatics 31, 3345–3347.

doi:10.1093/bioinformatics/btv361

Konwar, K.M., Hanson, N.W., Pagé, A.P., Hallam, S.J., 2013. MetaPathways: a modular

pipeline for constructing pathway/genome databases from environmental sequence

information. BMC Bioinformatics 14, 202. doi:10.1186/1471-2105-14-202

Labonté, J.M., Hallam, S.J., Suttle, C.A., 2015. Previously unknown evolutionary groups

dominate the ssDNA gokushoviruses in oxic and anoxic waters of a coastal marine

environment. Front. Microbiol. 6, 1–10. doi:10.3389/fmicb.2015.00315

Lam, P., Kuypers, M.M.M., 2011. Microbial Nitrogen Cycling Processes in Oxygen Minimum

Zones. Ann. Rev. Mar. Sci. 3, 317–345. doi:10.1146/annurev-marine-120709-142814

Lam, P., Lavik, G., Jensen, M.M., van de Vossenberg, J., Schmid, M., Woebken, D., Gutiérrez,

D., Amann, R., Jetten, M.S.M., Kuypers, M.M.M., 2009. Revising the nitrogen cycle in the 163

Peruvian oxygen minimum zone. Proc. Natl. Acad. Sci. U. S. A. 106, 4752–4757.

Lavik, G., Stuhrmann, T., Bruchert, V., Van der Plas, A., Mohrholz, V., Lam, P., 2009.

Detoxification of sulphidic African shelf waters by blooming chemolithotrophs. Nature 457,

581. doi:10.1017/CBO9781107415324.004

Leifer, I., Boles, J., 2005. Measurement of marine hydrocarbon seep flow through fractured rock

and unconsolidated sediment. Mar. Pet. Geol. 22, 551–568.

doi:10.1016/j.marpetgeo.2004.10.026

Lennartz, S., 2013. Time Series Analysis and Development of a 1D Box Model to Simulate

Trace Gas Concentrations at the Time Series Station Boknis Eck, Baltic Sea. Technische

Universitat Braunschweig.

Lilley, M.D., Baross, J.A., Gordon, L.I., 1982. Dissolved hydrogen and methane in Saanieh Inlet,

British Columbia. Deep Sea Res. 29.

Liu, T., Xia, X., Liu, S., Mou, X., Qiu, Y., 2013. Acceleration of denitrification in turbid rivers

due to denitrification occurring on suspended sediment in oxic waters. Environ. Sci.

Technol. 47, 4053–4061. doi:10.1021/es304504m

Loulergue, L., Schilt, A., Spahni, R., Masson-Delmotte, V., Blunier, T., Lemieux, B., Barnola,

J.-M., Raynaud, D., Stocker, T.F., Chappellaz, J., 2008. Orbital and millennial-scale

features of atmospheric CH4 over the past 800,000 years. Nature 453, 383–386.

doi:10.1038/nature06950

Lueker, T.J., Walker, S.J., Vollmer, M.K., Keeling, R.F., Nevison, C.D., Weiss, R.F., 2003.

Coastal upwelling air-sea fluxes revealed in atmospheric observations of O2, N2, CO2 and

N2O. Geophys. Res. Lett. 30, 1–4. doi:10.1029/2002GL016615

Manning, C.C., Hamme, R.C., Bourbonnais, A., 2010. Impact of deep-water renewal events on 164

fixed nitrogen loss from seasonally-anoxic Saanich Inlet. Mar. Chem. 122, 1–10.

doi:10.1016/j.marchem.2010.08.002

Martens-Habbena, W., Berube, P.M., Urakawa, H., de la Torre, J.R., Stahl, D. a, 2009. Ammonia

oxidation kinetics determine niche separation of nitrifying Archaea and Bacteria. Nature

461, 976–9. doi:10.1038/nature08465

Marty, D.G., 1993. Methanogenic bacteria in seawater. Limnol. Oceanogr. 38, 452–456.

Masson, D., 2002. Deep Water Renewal in the Strait of Georgia. Estuar. Coast. Shelf Sci. 54,

115–126. doi:10.1006/ecss.2001.0833

Masson, D., Cummins, P.F., 1999. Numerical Simulations of a Buoyancy-Driven Coastal

Countercurrent off Vancouver Island. J. Phys. Oceanogr. doi:10.1175/1520-

0485(1999)029<0418:NSOABD>2.0.CO;2

Matabos, M., Piechaud, N., De Montigny, F., Sarradin, P.-M., Sarrazin, J., 2015. The VENUS

cabled observatory as a method to observe fish behaviour and species assemblages in a

hypoxic fjord, Saanich Inlet (British Columbia, Canada). Can. J. Fish. Aquat. Sci. 72, 1–13.

doi:10.1139/cjfas-2013-0611

Mau, S., Valentine, D.L., Clark, J.F., Reed, J., Camilli, R., Washburn, L., 2007. Dissolved

methane distributions and air-sea flux in the plume of a massive seep field, Coal Oil Point,

California. Geophys. Res. Lett. 34, 2–6. doi:10.1029/2007GL031344

Merbt, S.N., Stahl, D.A., Casamayor, E.O., Mart??, E., Nicol, G.W., Prosser, J.I., 2012.

Differential photoinhibition of bacterial and archaeal ammonia oxidation. FEMS Microbiol.

Lett. 327, 41–46. doi:10.1111/j.1574-6968.2011.02457.x

Metcalf, W.W., Griffin, B.M., Cicchillo, R.M., Gao, J., Janga, S.C., Cooke, H.A., Circello, B.T.,

Evans, B.S., Martens-Habbena, W., Stahl, D.A., van der Donk, W.A., 2012. Synthesis of 165

Methylphosphonic Acid by Marine Microbes: A Source for Methane in the Aerobic Ocean.

Science (80-. ). 337, 1104–1107. doi:10.1126/science.1219875

Middelburg, J., Nieuwenhuize, J., Iversen, N., Hogh, N., de Wilde, H., Helder, W., Seifert, R.,

Christof, O., 2002. Methane distribution in European tidal estuaries. Biogeochemistry 59,

95–119.

Musenze, R., Grinham, A., 2013. Methane and nitrous oxide emissions from coastal, estuarine

and freshwater systems in subtropical Australia. EGU Gen. … 15, 6725.

Naqvi, S.W.A., Bange, H.W., Farías, L., Monteiro, P.M.S., Scranton, M.I., Zhang, J., 2010.

Marine hypoxia/anoxia as a source of CH4 and N2O. Biogeosciences 7, 2159–2190.

doi:10.5194/bg-7-2159-2010

Naqvi, S.W.A., Jayakumar, D.A., Narvekar, P. V., Naik, H., Sarma, V.V.S.S., D’Souza, W.,

Joseph, S., George, M.D., 2000. Increased marine production of N2O due to intensifying

anoxia on the Indian continental shelf. Nature 408, 346–349.

Naqvi, S.W.A., Naik, H., Jayakumar, D.A., Pratihary, A., Narvenkar, G., Kurian, S., Agnihotri,

R., Shailaja, M.S., and Narvekar, P.., 2009. Seasonal anoxia over the western Indian

continental shelf, in: Indian Ocean: Biogeochemical Processes and Ecological Variability,

Edited by: Wiggert, J. D., Hood, R. R., Naqvi, S. W. A., Brink, K. H., and Smith, S. L.,

Geophyical Monograph Series, 185, AGU, Washington, D.C.

Naqvi, S.W.A., Yoshinari, T., Jayakumar, D.A., Altabet, M.A., Narvekar, P. V, Devol, A.H.,

Brandes, J.A., Codispoti, L.A., 1998. Budgetary and biogeochemical implications of N2O

isotope signatures in the Arabian Sea. Nature 394, 462–464. doi:10.1038/28828

Nauhaus, K., Boetius, A., Kruger, M., Widdel, F., 2002. In vitro demonstration of anaerobic

oxidation of methane couples to sulfate reduction in sediment from a marine gas hydrate 166

area. Environ. Microbiol. 4, 296–305.

Nemcek, N., Ianson, D., Tortell, P.D., 2008. A high-resolution survey of DMS, CO2, and O2/Ar

distributions in productive coastal waters. Global Biogeochem. Cycles 22, 1–13.

doi:10.1029/2006GB002879

Nevison, C., Butler, J.H., Elkins, J.W., 2003. Global distribution of N2O and the ΔN2O-AOU

yield in the subsurface ocean. Global Biogeochem. Cycles 17, 1–18.

doi:10.1029/2003GB002068

Nevison, C.D., Lueker, T.J., Weiss, R.F., 2004. Quantifying the nitrous oxide source from

coastal upwelling. Global Biogeochem. Cycles 18, n/a–n/a. doi:10.1029/2003GB002110

Nightingale, P.D., Malin, G., Law, C.S., Watson, A.J., Liss, P.S., Liddicoat, M.I., Boutin, J.,

Upstill-Goddard, R.C., 2000. In situ evaluation of air-sea gas exchange parameterizations

using novel conservative and volatile tracers. Global Biogeochem. Cycles.

doi:10.1029/1999GB900091

O’Keefe, A., Deacon, D. a. G., 1988. Cavity ring-down optical spectrometer for absorption

measurements using pulsed laser sources. Rev. Sci. Instrum. 59, 2544.

doi:10.1063/1.1139895

Oremland, R.S., 1979. Methanogenic activity in plankton samples and fish intestines: A

mechanism for in situ methanogenesis in oceanic surface waters. Limnol. Oceanogr. 24,

1136–1141.

Orsi, W., Song, Y.C., Hallam, S., Edgcomb, V., 2012. Effect of oxygen minimum zone

formation on communities of marine protists. ISME J. 6, 1586–601.

doi:10.1038/ismej.2012.7

Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., 167

Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M.,

Lipenkov, V.Y., Lorius, C., Pepin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate

and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica.

Nature 399, 429–436. doi:10.1038/20859

Pierotti, B.D., Rasmussen, R.A., 1980. Nitrous oxide measurements in the eastern tropical

Pacific Ocean. Tellus 32, 56–72.

Portmann, R.W., Daniel, J.S., Ravishankara, A.R., 2012. Stratospheric ozone depletion due to

nitrous oxide: influences of other gases. Philos. Trans. R. Soc. B Biol. Sci. 367, 1256–1264.

doi:10.1098/rstb.2011.0377

Poth, M., Focht, D.D., 1985. N Kinetic Analysis of N(2)O Production by Nitrosomonas

europaea: an Examination of Nitrifier Denitrification. Appl. Environ. Microbiol. 49, 1134–

41.

Prinn, R., Cunnold, D., Alyea, F., Rasmussen, R., Simmonds, P., 1990. Atmospheric emissions

and trends of nitrous oxide deduced from 10 years of ALE-GAGE data. J. Geophys. Res.

95, 18369–18385.

Punshon, S., Moore, R.M., 2004a. Nitrous oxide production and consumption in a eutrophic

coastal embayment. Mar. Chem. 91, 37–51. doi:10.1016/j.marchem.2004.04.003

Punshon, S., Moore, R.M., 2004b. A stable isotope technique for measuring production and

consumption rates of nitrous oxide in coastal waters. Mar. Chem. 86, 159–168.

doi:10.1016/j.marchem.2003.11.003

Raghoebarsing, A., Pol, a, van de Pas-Schoonen, K., Smolders, A., Ettwig, K., Rijpstra, W.,

Schouten, S., Damste, J., Op den Camp, H., Jetten, M., Strous, M., 2006. A microbial

consortium couples anaerobic methane oxidation to denitrification RID B-9395-2011 RID 168

F-5114-2011 RID B-8834-2011. Nature 440, 918–921. doi:10.1038/nature04617

Reeburgh, W.S., 2007. Oceanic Methane Biogeochemistry. Chem. Rev. 107, 486–513.

Rehder, G., Collier, R.U., Heeschen, K.U., Kosro, P.M., Barth, J.A., Suess, E., 2002. Enhanced

marine CH4 emissions to the atmosphere off Oregon caused by coastal upwelling. Global

Biogeochem. Cycles 16. doi:10.1029/2000GB001391

Rehder, G., Leifer, I., Brewer, P.G., Friederich, G., Peltzer, E.T., 2009. Controls on methane

bubble dissolution inside and outside the hydrate stability field from open ocean field

experiments and numerical modeling. Mar. Chem. 114, 19–30.

doi:10.1016/j.marchem.2009.03.004

Reuer, M.K., Barnett, B. a., Bender, M.L., Falkowski, P.G., Hendricks, M.B., 2007. New

estimates of Southern Ocean biological production rates from O2/Ar ratios and the triple

isotope composition of O2. Deep Sea Res. Part I Oceanogr. Res. Pap. 54, 951–974.

doi:10.1016/j.dsr.2007.02.007

Riedel, M., Spence, G.D., Chapman, N.R., Hyndman, R.D., 2002. Seismic investigations of a

vent field associated with gas hydrates, offshore Vancouver Island. J. Geophys. Res. 107,

1–16. doi:10.1029/2001JB000269

Roegner, G.C., Needoba, J. a., Baptista, A.M., 2011. Coastal upwelling supplies oxygen-depleted

water to the Columbia River estuary. PLoS One 6. doi:10.1371/journal.pone.0018672

Rönner, U., 1983. Distribution, production and consumption of nitrous oxide in the Baltic Sea.

Geochim. Cosmochim. Acta 47, 2179–2188. doi:10.1016/0016-7037(83)90041-8

Roux, S., Hawley, A.K., Beltran, M.T., Scofeld, M., Schwientek, P., Stepanauskas, R., Woyke,

T., Hallam, S.J., Sullivan, M.B., 2014. Ecology and evolution of viruses infecting

uncultivated SUP05 bacteria as revealed by single-cell- and meta-genomics. Elife 2014, 1– 169

20. doi:10.7554/eLife.03125.001

Sancetta, C., 1989. Sediment transport by fecal pellets in British Columbia fjords. Mar. Geol. 89,

331–346.

Sancetta, C., Calvert, S.E., 1988. The annual cycle of sedimentation in Saanich inlet, British

Columbia: implications for the interpretation of diatom fossil assemblages. Deep Sea Res.

35, 71–90. doi:10.1016/0198-0149(88)90058-1

Sansone, F., Popp, B., Gasc, A., Graham, A.W., Rust, T.M., 2001. Highly elevated methane in

the eastern tropical North Pacific and associated isotopically enriched fluxes to the

atmosphere. Geophys. Res. Lett. 28, 4567–4570.

Sansone, F.J., Holmes, M.E., Popp, B.N., 1999. Methane stable isotopic ratios and

concentrations as indocators of methane dynamics in estuaries. Global Biogeochem. Cycles

13, 463–474. doi:10.1029/1999GB900012

Santoro, A.E., Buchwald, C., McIlvin, M.R., Casciotti, K.L., 2011. Isotopic signature of N2O

produced by marine ammonia-oxidizing archaea. Science 333, 1282–5.

doi:10.1126/science.1208239

Santos, F.J., Galceran, M.T., 2003. Modern developments in gas chromatography–mass

spectrometry-based environmental analysis. J. Chromatogr. A 1000, 125–151.

doi:10.1016/S0021-9673(03)00305-4

Schmale, O., Haeckel, M., McGinnis, D.F., 2011. Response of the Black Sea methane budget to

massive short-term submarine inputs of methane. Biogeosciences 8, 911–918.

doi:10.5194/bg-8-911-2011

Schmitz, R.A., Koenneke, M., Bange, H.W., LaRoche, J., Kock, A., Loescher, C.R., 2012.

Production of oceanic nitrous oxide by ammonia-oxidizing archaea. Biogeosciences 170

Discuss. 9, 2095–2122. doi:10.5194/bgd-9-2095-2012

Schneider von Deimling, J., Rehder, G., Greinert, J., McGinnnis, D.F., Boetius, A., Linke, P.,

2011. Quantification of seep-related methane gas emissions at Tommeliten, North Sea.

Cont. Shelf Res. 31, 867–878. doi:10.1016/j.csr.2011.02.012

Seitzinger, S.P., Kroeze, C., Styles, V., 2000. Global distribution of N2O emissions from aquatic

systems: natural emissions and anthropogenic effects. Chemosphere 2, 267–279.

Shakhova, N., Semiletov, I., Leifer, I., Sergienko, V., Salyuk, A., Kosmach, D., Chernykh, D.,

Stubbs, C., Nicolsky, D., Tumskoy, V., Gustafsson, Ö., 2013. Ebullition and storm-induced

methane release from the East Siberian Arctic Shelf. Nat. Geosci. 7, 64–70.

doi:10.1038/ngeo2007

Shakhova, N., Semiletov, I., Salyuk, A., Yusupov, V., Kosmach, D., Gustafsson, Ö., 2010.

Extensive Methane Venting to the Atmosphere from Sediments of the East Siberian Arctic

Shelf. Sci. 327 , 1246–1250. doi:10.1126/science.1182221

Sich, I., Russow, R., 1999. 15N Analysis of nitric oxide and nitrous oxide by cryotrap

enrichment using a gas chromatograph quadrupole mass spectrometer and its application to

(15)N-tracer investigations of NO/N(2)O formation in soil. Rapid Commun. Mass

Spectrom. 13, 1325–1328. doi:10.1002/(SICI)1097-0231(19990715)13:13<1325::AID-

RCM622>3.0.CO;2-K

Solomon, E.A., Kastner, M., MacDonald, I.R., Leifer, I., 2009. Considerable methane fluxes to

the atmosphere from hydrocarbon seeps in the Gulf of Mexico. Nat. Geosci 2, 561–565.

Sowers, T., 2006. Late-Quaternary atmospheric CH4 isotopre record suggests marine clathrates

are stable. Science (80-. ). 311, 838–840.

Sowers, T., Alley, R.B., Jubenville, J., 2003. Ice core records of atmospheric N2O covering the 171

last 106,000 years. Science 301, 945–948. doi:10.1126/science.1085293

Steinle, L., Graves, C. a., Treude, T., Ferré, B., Biastoch, A., Bussmann, I., Berndt, C., Krastel,

S., James, R.H., Behrens, E., Böning, C.W., Greinert, J., Sapart, C.-J., Scheinert, M.,

Sommer, S., Lehmann, M.F., Niemann, H., 2015. Water column methanotrophy controlled

by a rapid oceanographic switch. Nat. Geosci. 8, 1–6. doi:10.1038/ngeo2420

Stieglmeier, M., Mooshammer, M., Kitzler, B., Wanek, W., Zechmeister-Boltenstern, S.,

Richter, A., Schleper, C., 2014. Aerobic nitrous oxide production through N-nitrosating

hybrid formation in ammonia-oxidizing archaea. ISME J. 8, 1135–46.

doi:10.1038/ismej.2013.220

Stramma, L., Johnson, G.C., Sprintall, J., Mohrholz, V., 2008. Expanding oxygen-minimum

zones in the tropical oceans. Science 320, 655–658. doi:10.1126/science.1153847

Stramma, L., Schmidtko, S., Levin, L.A., Johnson, G.C., 2010. Ocean oxygen minima

expansions and their biological impacts. Deep Sea Res. Part I Oceanogr. Res. Pap. 57, 587–

595. doi:10.1016/j.dsr.2010.01.005

Suess, E., Torres, M.E., Bohrmann, G., Collier, R.W., Greinert, J., Linke, P., Rehder, G., Trehu,

a., Wallmann, K., Winckler, G., Zuleger, E., 1999. Gas hydrate destabilization: Enhanced

dewatering, benthic material turnover and large methane plumes at the Cascadia convergent

margin. Earth Planet. Sci. Lett. 170, 1–15. doi:10.1016/S0012-821X(99)00092-8

Sutka, R.L., Ostrom, N.E., Ostrom, P.H., Breznak, J.A., Pitt, A.J., Li, F., Gandhi, H., 2006.

Distinguishing Nitrous Oxide Production from Nitrification and Denitrification on the Basis

of Isotopomer Abundances Distinguishing Nitrous Oxide Production from Nitrification and

Denitrification on the Basis of Isotopomer Abundances. Appl. Environ. Microbiol. 72, 638–

644. doi:10.1128/AEM.72.1.638 172

Swinnerton, J., Linnenbom, V., 1967. Determination of the C1 to C4 hydrocarbons in sea water

by gas chromatography. J. Chromatogr. Sci. 5, 570–573.

Thomson, R.E., Krassovski, M. V., 2010. Poleward reach of the California Undercurrent

extension. J. Geophys. Res. Ocean. 115, 1–9. doi:10.1029/2010JC006280

Timothy, D.A., Soon, M.Y.., Calvert, S.E., 2003. Settling fluxes in Saanich and Jervis Inlets,

British Columbia, Canada: sources and seasonal patterns. Prog. Oceanogr. 59, 31–73.

doi:10.1016/j.pocean.2003.07.001

Timothy, D.A., Soon, M.Y.S., 2001. Primary production and deep-water oxygen content of two

British Columbian fjords. Mar. Chem. 73, 37–51. doi:10.1016/S0304-4203(00)00071-2

Tortell, P.D., Merzouk, A., Ianson, D., Pawlowicz, R., Yelland, D.R., 2012. Influence of regional

climate forcing on surface water pCO2,O2/Ar and dimethylsulfide (DMS) along the

southern British Columbia coast. Cont. Shelf Res. 47, 119–132.

doi:10.1016/j.csr.2012.07.007

Tryon, M.D., Brown, K.M., Torres, M.E., Tréhu, A.M., McManus, J., Collier, R.W., 1999.

Measurements of transience and downward fluid flow near episodic methane gas vents,

Hydrate Ridge, Cascadia. Geology 27, 1075–1078. doi:10.1130/0091-

7613(1999)027<1075:MOTADF>2.3.CO;2

Upstill-Goddard, R.C., Rees, a. P., Owens, N.J.P., 1996. Simultaneous high-precision

measurements of methane and nitrous oxide in water and seawater by single phase

equilibration gas chromatography. Deep Sea Res. Part I Oceanogr. Res. Pap. 43, 1669–

1682. doi:10.1016/S0967-0637(96)00074-X

Valentine, D.L., 2011. Emerging Topics in Marine Methane Biogeochemistry. Ann. Rev. Mar.

Sci. 3, 147–171. doi:10.1146/annurev-marine-120709-142734 173

van der Maarel, M.J.E.C., Sprenger, W., Haanstra, R., Forney, L.J., 1999. Detection of

methanogenic archaea in seawater particles and the digestive tract of a marine ¢ sh species.

FEMS Mircrobiology Lett. 173, 189–194. van Kessel, M.A.H.J., Speth, D.R., Albertsen, M., Nielsen, P.H., Op den Camp, H.J.M., Kartal,

B., Jetten, M.S.M., Lücker, S., 2015. Complete nitrification by a single microorganism.

Nature 528, 555–559. doi:10.1038/nature16459

Waldichuk, M., 1957. Physical Oceanography of the Strait of Georgia, British Columbia. J. Fish.

Res. Board Canada 14, 321–486. doi:10.1139/f57-013

Walsh, D.A., Hallam, S.J., 2011. Bacterial Community Structure and Dynamics in a Seasonally

Anoxic Fjord: Saanich Inlet, British Columbia BT - onlinelibrary.wiley.com, in: de Bruijn,

F.J. (Ed.), Handbook of Molecular Microbial Ecology II. John Wiley & Sons, Inc, pp. 253–

267. doi:10.1002/9781118010549.ch2

Walsh, D.A., Zaikova, E., Howes, C.G., Song, Y.C., Wright, J.J., Tringe, S.G., Tortell, P.D.,

Hallam, S.J., 2009. Metagenome of a versatile chemolithoautotroph from expanding

oceanic dead zones. Science 326, 578–582. doi:10.1126/science.1175309

Walter, S., Bange, H.W., Breitenbach, U., Wallace, D.W.R., 2006. Nitrous oxide in the North

Atlantic Ocean. Biogeosciences 3, 607–619.

Wang, D., Gouhier, T.C., Menge, B.A., Ganguly, A.R., 2015. Intensification and spatial

homogenization of coastal upwelling under climate change. Nature 518, 390–394.

doi:10.1038/nature14235

Ward, B., 2011. Measurement and distribution of nitrification rates in the oceans. Methods

Enzym. 486, 307–323. doi:10.1016/B978-0-12-381294-0.00013-4

Ward, B.B., Kilpatrick, K.A., 1990. Relationship between substrate concentration and oxidation 174

of ammonium and methane in a stratified water column. Cont. Shelf Res. 10.

Ward, B.B., Kilpatrick, K.A., Wopat, A.E., Minnich, E.C., Lidstrom, M.E., 1989. Methane

oxidation in Saanich Inlet during summer stratification. Cont. Shelf Res. 9, 65–75.

doi:10.1016/0278-4343(89)90083-6

Weiss, R., Price, B., 1980. Nitrous Oxide Solubility in Water and Seawater. Mar. Chem. 8, 347–

359.

Welsh, D., Castadelli, G., Bartoli, M., Poli, D., Careri, M., de Wit, R., Viaroli, P., 2001.

Denitrification in an intertidal seagrass meadow, a comparison of 15 N-isotope and

acetylene-block techniques: dissimilatory nitrate reduction to ammonia as a source of N 2

O? Mar. Biol. 139, 1029–1036. doi:10.1007/s002270100672

Whitney, F.A., 2015. Anomalous winter winds decrease 2014 transition zone productivity in the

NE Pacifi. Geophys. Res. Lett. 42, 428–431. doi:10.1002/2014GL062634.

Whitney, F.A., Freeland, H.J., Robert, M., 2007. Persistently declining oxygen levels in the

interior waters of the eastern subarctic Pacific. Prog. Oceanogr. 75, 179–199.

doi:10.1016/j.pocean.2007.08.007

Wiesenburg, D.A., Guinasso, N.L., 1979. Equilibrium solubilities of methane, carbon monoxide,

and hydrogen in water and sea water. J. Chem. Eng. Data 24, 356–360.

doi:10.1021/je60083a006

Wittke, F., Kock, A., Bange, H.W., 2010. Nitrous oxide emissions from the upwelling area off

Mauritania (NW Africa). Geophys. Res. Lett. 37. doi:10.1029/2010GL042442

Wolff, E., Spahni, R., 2007. Methane and nitrous oxide in the ice core record. Philos. Trans. A.

Math. Phys. Eng. Sci. 365, 1775–92. doi:10.1098/rsta.2007.2044

Wright, J.J., Konwar, K.M., Hallam, S.J., 2012. Microbial ecology of expanding oxygen 175

minimum zones. Nat. Rev. Microbiol. 10, 381–394. doi:10.1038/nrmicro2778

Yahel, G., Yahel, R., Katz, T., Lazar, B., Herut, B., Tunnicliffe, V., 2008. Fish activity: A major

mechanism for sediment resuspension and organic matter remineralization in coastal marine

sediments. Mar. Ecol. Prog. Ser. 372, 195–209. doi:10.3354/meps07688

Yamagishi, H., Yoshida, N., Toyoda, S., Popp, B.N., Westley, M.B., Watanabe, S., 2005.

Contributions of denitrification and mixing on the distribution of nitrous oxide in the North

Pacific. Geophys. Res. Lett. 32, 1–4. doi:10.1029/2004GL021458

Yoshida, N., Morimoto, H., Hirano, M., Koike, I., Matsuo, S., Wada, E., Saino, T., Hattori, A.,

1989. Nitrification rates and 15N abundances of N2O and NO3- in the western North

Pacific. Nature 342, 895–897. doi:10.1038/342895a0

Yoshida, O., Inoue, H.Y., Watanabe, S., Suzuki, K., Noriki, S., 2011. Dissolved methane

distribution in the South Pacific and the Southern Ocean in austral summer. J. Geophys.

Res. 116, C07008. doi:10.1029/2009JC006089

Yvon-Lewis, S.A., Hu, L., Kessler, J., 2011. Methane flux to the atmosphere from the Deepwater

Horizon oil disaster. Geophys. Res. Lett. 38. doi:10.1029/2010GL045928

Zaikova, E., Walsh, D.A., Stilwell, C.P., Mohn, W.W., Tortell, P.D., Hallam, S.J., 2010.

Microbial community dynamics in a seasonally anoxic fjord: Saanich Inlet, British

Columbia. Environ. Microbiol. 12, 172–91. doi:10.1111/j.1462-2920.2009.02058.x

Zamora, L.M., Oschlies, A., 2014. Surface nitrification: A major uncertainty in marine N2O

emissions. Geophys. Res. Lett. 41, 4247–4253. doi:10.1002/2014GL060556

Zamora, L.M., Oschlies, A., Bange, H.W., Huebert, K.B., Craig, J.D., Kock, A., Löscher, C.R.,

2012. Nitrous oxide dynamics in low oxygen regions of the Pacific: Insights from the

MEMENTO database. Biogeosciences 9, 5007–5022. doi:10.5194/bg-9-5007-2012 176

Zhan, L., Chen, L., 2009. Distributions of N2O and its air‐ sea fluxes in seawater along cruise

tracks between 30° S–67° S and in Prydz Bay, Antarctica. J. Geophys. Res. Ocean. 114,

C03019. doi:10.1029/2007JC004406

177

Appendix

Supplementary Figures:

Figure A.1 N2O, NO2-+NO3- and O2 profiles from LC04 and LC11 for each cruise.