Carbon monoxide in coastal waters and the open ocean

Carbon monoxide in coastal waters and the open ocean Dissertation Zur Erlangung des Doktorgrades der Mathematisch-Naturwissenschaftlichen Fakultät der Christian-Albrechts-Universität zu Kiel vorgelegt von Guanlin Li Kiel, 2023 Carbon monoxide in coastal waters and the open ocean Dissertation Zur Erlangung des Doktorgrades der Mathematisch-Naturwissenschaftlichen Fakultät der Christian-Albrechts-Universität zu Kiel vorgelegt von Guanlin Li Kiel, 2023 Erste Gutachterin: Prof. Dr. Hermann Bange Zweiter Gutachter: Prof. Dr. Anja Engel Tag der mündlichen Prüfung: 15.01.2024 Zum Druck genehmigt: III Eidesstattliche Erklärung Hiermit erkläre ich, dass ich die vorliegende Doktorarbeit selbständig und ohne unerlaubte Hilfe erstellt habe. Weder diese noch eine ähnliche Arbeit wurde an einer anderen Abteilung oder Hochschule im Rahmen eines Prüfungsverfahrens vorgelegt, veröffentlicht oder zur Veröffentlichung vorgelegt. Ferner versichere ich, dass die Arbeit unter Einhaltung der Regeln guter wissenschaftlicher Praxis der Deutschen Forschungsgemeinschaft entstanden ist. Kiel, den November 2023 gez. Guanlin Li IV Acknowledgements V Acknowledgements My deepest gratitude is first to Prof. Dr. Hermann W. Bange, for his constant encouragement, guidance, and positive feedback throughout my doctoral research and thesis. I would not have finished my PhD successfully without him. Also, I appreciate Prof. Dr. Anja Engel and PD Dr. Christa A. Marandino for being my ISOS/ FYORD committee members and for helping me understand the bigger picture in their scientific views. A huge ‘Thank you!’ goes to Dr. Damian L. Arévalo-Martínez, who offered me great help in all aspects and was always open to any further questions. Without his precious dataset and dedication, it would have been impossible for me to combine the different topics coherently in my doctoral thesis. I also thank Hanna and Riel for their help with fieldwork and lab work, pushing me to dive deep into my research. Special thanks to the crew of Littorina for making the BE cruises were smooth and successful during the past years. A big thanks to Mikołaj Słupiński, for helping my research with regards to the machine learning algorithms and model building. The workgroup “AG Bange” feels like a big family with a positive and friendly working atmosphere. I would also like to thank all my Chinese colleagues and friends. We had a good time together and their lovely company helped me relax from one after another tough research day. Thanks to the China Scholarship Council for the financial support. Last but not least, I would like to thank my dear parents for their love and great faith in me for many years. I love you with all my heart! Acknowledgements VI Manuscript contributions VII Manuscript contributions This dissertation is based on the following manuscripts: Manuscript I: Li, G., Arévalo-Martínez, D. L., Ingeniero, R. C. O., and Bange, H. W.: Carbon monoxide cycling in the Ria Formosa Lagoon (southern Portugal) during summer 2021, [preprint], https://doi.org/10.5194/egusphere-2023-771, 2023. Author contributions: GL and RCOI designed the study and participated in the fieldwork; GL performed the measurements and wrote the manuscript; DLAM, HWB, and RCOI contributed to the writing of the manuscript and discussion of the data. Manuscript II: Li, G., Arévalo-Martínez, D. L., Hepach, H., Engel, A., and Bange, H. W.: A seasonal study of dissolved carbon monoxide at the Boknis Eck Time Series Station in Eckernförde Bay (southwestern Baltic Sea), manuscript in preparation for Biogeosciences. Author contributions: GL and HWB designed the study; GL participated in the fieldwork and performed the carbon monoxide measurements; CDOM and DOC data processing were done by HH and AE; GL conducted further data analysis and wrote the article with contributions from all co-authors. Manuscript III: Li, G., Arévalo-Martínez, D. L., Słupiński, M., and Bange, H. W.: Global reconstruction of oceanic carbon monoxide emissions, manuscript in preparation for Global biogeochemical cycles. Author contributions: GL, DLAM, and HWB designed the study; DLAM and GL assembled the carbon monoxide supersaturation dataset; MS and GL developed the statistical mapping method and generated carbon monoxide emissions estimates. GL wrote the paper with input from MS, DLAM, and HWB. ZUSAMMENFASSUNG XIV vorgenommen, um die ozeanischen Emissionen mit Hilfe eines datengesteuerten maschinellen Lernansatzes zu rekonstruieren und so die Unsicherheit bei der Schätzung der globalen ozeanischen CO-Emissionen zu verringern. Diese Arbeit kann eine solide wissenschaftliche Grundlage für die Erstellung eines Modells des biogeochemischen Kreislaufs von CO im Ozean liefern. Die wichtigsten Ergebnisse sind die folgenden: In der ersten Studie wurden die ersten Messungen von gelöstem CO in der Lagune Ria Formosa, einem anthropogen beeinflussten System im Süden Portugals, durchgeführt. Die Konzentrationen von gelöstem CO in der Oberflächenschicht reichten von 0.16 bis 3.1 nmol L1 mit einer durchschnittlichen Konzentration von 0.75 ± 0.57 nmol L-1. Das COSättigungsverhältnis reichte von 1.7 bis 32.2, was darauf hindeutet, dass die Lagune im Mai 2021 als CO-Quelle für die Atmosphäre fungierte. Die geschätzte durchschnittliche Flussdichte zwischen Meer und Luft betrug 1.53 μmol m-2 d-1, was hauptsächlich auf die photochemische CO-Produktion zurückzuführen ist. Der mikrobielle Verbrauch machte 83 % der COProduktion aus, was darauf hindeutet, dass die resultierenden CO-Emissionen in die Atmosphäre durch den mikrobiellen Verbrauch in den Oberflächengewässern moduliert wurden. Die Ergebnisse eines Bestrahlungsexperiments mit Aquakulturabwässern zeigten, dass Aquakulturanlagen in der Lagune von Ria Formosa offenbar eine vernachlässigbare CO-Quelle für die Atmosphäre darstellen. In der zweiten Studie wurde die erste systematische saisonale Untersuchung des gelösten CO an der Zeitserienstation Boknis Eck in der Eckernförder Bucht (südwestliche Ostsee) durchgeführt. Die Konzentrationen von gelöstem CO lagen zwischen 0.20 und 1.89 nmol L-1 mit einem Mittelwert von 0.63 ± 0.38 nmol L-1, der im Frühjahr (0.70 nmol L-1) höher war als im Sommer (0.42 nmol L-1). Das CO-Sättigungsverhältnis schwankte zwischen 1.9 und 4.5 mit höheren Werten im Frühjahr und Sommer, was die Rolle dieses Gebiets als CO-Quelle für die Atmosphäre mit einer geschätzten durchschnittlichen Flussdichte von 3.66 μmol m-2 d-1 zwischen Meer und Luft belegt. Wir schätzen, dass die Ostsee etwa 0.15 % (6.1 Gg CO-C yr-1) der globalen ozeanischen CO-Emissionen beiträgt. In den Oberflächengewässern war CO stark mit Kieselalgen korreliert, während in der Bodenschicht hohe CO-Werte mit dem Auftreten von ZUSAMMENFASSUNG XV Hypoxie oder Anoxie verbunden waren. Wir vermuten, dass die CO-Akkumulation im Bodenwasser auf die Freisetzung von CO aus anoxischen Sedimenten und seine In-situProduktion in der darüber liegenden Wassersäule zurückzuführen ist. Mit Hilfe optischphotochemischer Modelle, die auf gemessenen CDOM-Absorptionskoeffizienten und spektralen CO-Quantenausbeuten basieren, schätzten wir die CO-Photoproduktion in dem Gebiet auf 20.1 Gg CO-C pro Jahr, was etwa 0.1 % des globalen Budgets entspricht. Wir lieferten wertvolle Erkenntnisse über die Umweltfaktoren, die die CO-Vertei lung in der Wassersäule der Ostsee bestimmen, und beleuchteten die jahreszeitliche Dynamik und die Wechselwirkungen mit anderen Variablen. In der dritten Studie wurden klimatologische CO-Emissionen aus dem Ozean durch Training von Machine-Learning-Modellen mit über 12,000 CO-Messungen aus dem Oberflächenozean rekonstruiert - die bisher größte Synthese. Die Vorhersage reproduzierte das COUngleichgewicht ( ∆ CO) relativ genau (r2 = 0.88), wobei der quadratische Fehler (root-meansquare error, RMSE) von ∆ CO 0.60 nmol L-1 beträgt. Die Karte zeigt Breitengradienten im beobachteten CO-Fluss und einen signifikanten globalen saisonalen Zyklus. Wir schätzten einen mittleren jährlichen CO-Fluss von 5.6 Tg CO-C yr−1. Darüber hinaus schätzten wir, dass 8.3 % des globalen ozeanischen CO-Flusses aus dem Südlichen Ozean stammen (0.46 ± 0.10 Tg CO-C yr−1). Diese Schätzung des ozeanischen Flusses stimmt mit dem in früheren Studien angegebenen Bereich überein, verringert aber die Unsicherheit um mehr als das Dreifache. ZUSAMMENFASSUNG XVI Table of Contents XVII Table of Contents ACKNOWLEDGEMENTS ....................................................................................... V MANUSCRIPT CONTRIBUTIONS ..................................................................... VII ABSTRACT ............................................................................................................... IX ZUSAMMENFASSUNG ........................................................................................ XIII TABLE OF CONTENTS ..................................................................................... XVII INTRODUCTION ....................................................................................................... 1 THESIS OUTLINE ................................................................................................... 33 CARBON MONOXIDE CYCLING IN THE RIA FORMOSA LAGOON (SOUTHERN PORTUGAL) DURING SUMMER 2021 ....................................... 35 A SEASONAL STUDY OF DISSOLVED CARBON MONOXIDE AT THE BOKNIS ECK TIME SERIES STATION IN ECKERNFÖRDE BAY (SOUTHWESTERN BALTIC SEA) ........................................................................ 67 GLOBAL RECONSTRUCTION OF OCEANIC CARBON MONOXIDE EMISSIONS ............................................................................................................. 101 CONCLUSIONS AND OUTLOOK ....................................................................... 133 Table of Contents XVIII Introduction 1 1 Introduction 1.1 Atmospheric carbon monoxide and its potential role in climate Greenhouse gases (GHGs) play a crucial role in the Earth's climate system and are closely linked to climate change. These gases trap heat in the atmosphere, significantly contributing to radiative forcing (Dickinson and Cicerone, 1986; Boucher et al., 2013; Kweku et al., 2017; Forster et al., 2021; Bange, 2022). With the increasing concentration of GHGs in the atmosphere, the consequences of climate change include rising global temperature, melting ice sheets and glaciers, sea-level rise, and more frequent and extreme weather events have been demonstrated (see e.g. Oppenheimer, 1998; Pollack et al., 1998; Dyurgerov et al., 2000; Soruco et al., 2009; Schiermeier, 2011; Zhai and Liu, 2012; Cazenave et al., 2014; Wunderling et al., 2020; Forster et al., 2023). Ocean acidification and deoxygenation can be associated with the GHG effect as well (Keeling et al., 2010; Pandolfi et al., 2011; Six et al., 2013; Rees et al., 2016). Carbon monoxide (CO), one of the most important prevalent trace gases in the troposphere, indirectly affects the concentrations of many chemically active gases in the atmosphere and thus is regarded as an indirect GHG (Evans and Puckrin, 1995; Thompson, 1992). The primary oxidant present in the troposphere is the hydroxyl radical (•OH), which is capable of oxidizing a variety of atmospheric species, like nitrogen compounds, carbon compounds, sulfur compounds, and halocarbons, through photochemical reactions (Thompson, 1992; Derwent, 1995; Moran and Miller, 2007; Seinfeld and Pandis, 2012). As the reaction between •OH and CO is ubiquitous (Derwent, 1995), an increase in the concentration of CO in the atmosphere will result in a decrease in the concentration of •OH in the troposphere, leading to a further reduction in the oxidation capacity of the troposphere. Over the past century, the concentration of OH in the atmosphere has decreased by approximately 20%, possibly due to Introduction 2 the increase in CO and CH4 (Thompson and Cicerone, 1986; Levine et al., 1985). From 1950 to 1985, the concentration of CO in the troposphere increased by 50%, resulting in a corresponding decrease of approximately 25% in the concentration of •OH in the troposphere (Levine et al., 1985) and the recent study by Lelieveld et al., 2016 shows that ~40% global tropospheric •OH is consumed by CO + •OH reaction. Moreover, CO oxidation in conjunction with the reduction of NOx influences the abundance of tropospheric ozone (O3) (Dignon and Hameed, 1985; Stubbins et al., 2006a; Ravishankara et al., 2009). Former studies demonstrated that in remote areas, approximately 20% to 40% of O3 in the atmosphere is generated due to the increase in CO. Conversely, in urban areas, about 10% to 20% of O3 is generated with the increase of CO (Crutzen, 1987; Cicerone, 1988). Thus, given the important role CO plays in atmospheric chemistry, a good understanding of CO is required for a full understanding of the anthropogenically and naturally driven changes as well as in trace gas budgets. CO has a comparably short atmospheric lifetime of 1–3 months (Zheng et al., 2019). Typical annual mean surface atmospheric CO mole fractions (Fig. 1.1.1) range from ~120 ppb in the Northern Hemisphere to ~40 ppb in the Southern Hemisphere (Petron et al., 2019). Furthermore, reconstructions of atmospheric CO mole fractions based on limited ice core samples in the Northern Hemisphere high latitudes suggest CO mole fractions of about 145 ppb in the 1950s, which rose by 10–15 ppb in the mid-1970s, and then declined by about 30 ppb to about 130 ppb by 2008 (Petrenko et al., 2013). Figure 1.1.2 presents the global concentration and trend of atmospheric CO in the past years. From the figure, it can be seen that there is a decreasing trend of atmospheric CO (NOAA ESRL, FTP:\ftp.cmdl.noaa.gov, 2023). Inversion-based analysis attributed the global CO decline during the past decade to decreases in anthropogenic and biomass-burning CO emissions despite a probable increase in atmospheric CO chemical production (Gaubert et al., 2017; Jiang et al., 2017; Zheng et al., 2019). Introduction 3 Figure 1.1.1: The mean surface atmospheric carbon monoxide mole fraction over the globe: climatology (Data from Measurement of Pollution in the Troposphere data. ftp://l5ftl01.larc.nasa.gov/MOPITT/) Figure 1.1.2: Three-dimensional representation of the latitudinal distribution of atmospheric carbon monoxide in the marine boundary layer. (Data from the Carbon Cycle cooperative air sampling network. The surface represents data smoothed in time and latitude. http://www.esrl.noaa.gov/gmd/ccgg/) Approximately over 1000 Tg CO–C yr-1 is introduced into the atmosphere annually (Zheng et al., 2019; Greening and Grinter, 2022) including both natural (CH4 and NMHC oxidation, oceanic emission, volcanoes, and wildfires) and anthropogenic sources (combustion of fossil fuels and biomass; oxidation of anthropogenic CH4 and NMHCs), with the anthropogenic sources currently constituting more than half of the total (Khalil and Rasmussen, 1984; Introduction 4 Petrenko et al., 2013). The natural sources of CO are due mainly to oxidation reactions. According to Holloway et al. (2000), approximately 80% of the CH4 in the troposphere is oxidized to CO, which accounts for 25-35% of the source of CO. The oxidation of NMHCs via reactions with •OH or O3 results in the production of CO. Although NMHCs are present in the troposphere at much lower levels than CH4, their oxidation rates are relatively high. In the vicinity of NMHCs sources, the amount of CO produced from the oxidation of NMHCs may even exceed that produced from the oxidation of CH4 (Lopez, 2003). Despite oceanic emission being a minor source of atmospheric CO contributing less than 1% to the natural sources of atmospheric CO (Conte et al., 2019; Zheng et al., 2019), it can contribute significantly to the atmospheric CO budget in remote areas such as the Arctic Ocean where the influence of other CO sources is marginal (Blomquist et al., 2012; Campen et al., 2023). The primary anthropogenic source of CO arises from the incomplete combustion/oxidation of fossil fuels, such as petroleum, natural gas, coal, and gasoline. Combustion of these fuels has been identified as a major contributor to atmospheric CO, according to various studies (Freyer, 1979; Jaffe, 1968; Seiler and Schmidt, 1984; Walsh, 1990; Hoesly et al., 2018; McDuffie et al., 2020) and over 60% of total CO is of anthropogenic origin after excluding the natural contribution of biomass burning, like naturally occurring wildfire events (Logan et al., 1981; Andreae and Merlet, 2001; WHO, 2004). Another significant anthropogenic source of CO is biomass combustion, particularly in the southern hemisphere (Andreae et al., 1996; Andreae and Merlet, 2001; Blake et al., 1999). Research revealed that the amount of CO generated from burning tropical rainforests and savannas is equivalent to that produced from fossil fuel combustion/oxidation (Granier et al., 1999; Levine, 1996; Newell et al., 1989). Reaction with •OH is the primary sink of CO (Ehhalt, 1999; Fichot and Miller, 2010; Khalil and Resmussen, 1990; Prinn et al., 1995) with a smaller contribution from dry deposition Introduction 5 (Conrad, 1995; Godde et al., 2000; Potter et al., 1996). About 90% of atmospheric CO reacts with •OH to form carbon dioxide (CO2) (Fichot and Miller, 2010; Petrenko et al., 2013). Biological uptake in soils is the primary dry deposition mechanism for CO. It was observed that greater uptake occurred when higher concentrations of organic matter were present in the soil (Inman et al., 1971; Cordero et al., 2019). The transport of CO into the stratosphere constitutes an additional sink for CO in the troposphere, as approximately 5% of tropospheric CO may be transported into the stratosphere (Bergamaschi et al., 2000; Seiler and Conrad, 1987; Taylor et al., 1996). 1.2 Cycling of carbon monoxide in the ocean Figure 1.2.1: Schematic diagram of the biogeochemical cycle of CO Swinnerton et al. (1968) conducted the first measurements of oceanic CO in the Mediterranean Sea and reported CO in the surface layer was supersaturated relative to the partial pressure of this indirect GHG in the atmosphere. Subsequent investigations confirmed the supersaturation of oceanic CO concentrations (Seiler and Junge, 1970; Lamontagne et al., 1971; Linnenbom et al., 1973; Ohta, 1997; Xie et al., 2002; Stubbins et al., 2006a) and supported the idea that the ocean is considered a source of atmospheric CO, especially in offshore areas far from land. Introduction 12 Subsequently, recent evidence suggested that photolysis of particulate organic matter (POM) can also produce a significant amount of CO in the water column (Stubbins et al., 2006b; Xie and Zafiriou, 2009; Song and Xie, 2017). The photoproduction rates of CO were linear in correlation with the POM enrichment factor. The CO produced by POM photochemistry accounted for 11–35% of CO produced by CDOM photodegradation. Indeed, the efficiency of POM photochemical produced is higher than that of CDOM (Xie and Zafiriou., 2009). 1.2.2.2 Dark production Organic matter in the ocean can also produce CO through thermodynamic decomposition. This underlying process of the so-called CO dark production is a potentially important source of marine CO. It’s a necessary term to explain the vertical distribution of CO. Kettle (1994; 2005) inferred the existence of the CO dark reaction through a CO biogeochemical model in the upper ocean. Xie et al. (2005) measured the average dark reaction rate of CO in Delaware Bay at 0.21 nmol L-1 h-1. Zhang et al. (2008) estimated the CO dark production in the global coastal waters to be 0.46–1.50 Tg CO–C yr-1 based on the database of the St. Lawrence Estuary in Canada, speculating the global oceanic CO dark production to be in the range from 4.87 to 15. 8 Tg CO– C yr-1, which accounted for 9.7%–31.6% of the global marine photoproduction. Xu et al. (2023) demonstrated dark production (0.43 ± 0.25 nmol L−1 d−1) accounted for one-fifth of the total production in the Eastern Indian Ocean. CO dark production exhibited strong seasonal and spatial variations. The rate of CO dark production is highest in summer, followed by autumn and spring, and lowest in winter (Zhang and Xie, 2012). However, the mechanism behind the dark production of CO has not been reported to date and potentially large uncertainties may exist in the current evaluation of CO dark production, necessitating further elucidation. 1.2.2.3 Other sources of CO in the ocean Despite the potential occurrence of this additional process, direct production of CO by biology is not a negligible source (Loewus and Delwicke, 1963; Pickwell et al., 1964; Chapman and Tocher, 1966; Junge et al., 1971). Seiler and Schmidt (1974) deduced the highest CO concentration in the eutrophic area of the Southern Ocean caused by biological production. Introduction 13 Song (2013) also reported that the change in CO concentration in the bottom ice is closely related to the biomass of ice algae. The CO in “brown ice” has a concentration of about four times that of the ocean column which is caused by the release of brown algae (Swinnerton and Lamontagne, 1974). Cyanobacteria and diatoms are the largest CO emitters and this source could account for up to 20% of the oceanic CO production (Gros et al., 2009; McLeod et al., 2021). Additionally, hydrothermal activity introduces a small amount of CO into the deep sea (Lilley et al., 1982; King and Weber, 2007) and Swinnerton et al. (1971) found that the CO in rainwater is saturated relative to the atmosphere, and measured the CO concentration in the rainwater in Hawaii and Washington as 3.5×10-7 g L-1. Similarly, Seiler and Schmidt (1974) argued for a CO concentration in the rainwater of 0.7×10-7 g L-1 implying that the rainwater is still a source of CO in the ocean although the mechanism of CO production in rainwater is uncertain. 1.2.3 Sinks of carbon monoxide in the ocean 1.2.3.1 Microbial consumption Under the normal state of surface seawater, microbial consumption is the major sink (32±18 Tg CO–C yr-1), accounting for about 86% of CO removal in the marine environment (Zafiriou et al., 2003; Zhang et al., 2008; Greening and Grinter, 2022). The mechanism by which microorganisms consume CO remains to be found. However, it is known that Carboxydotrophs (e.g. Alphaproteobacteria, Gammaproteobacteria, Actinobacteria, and Bacilli) and Carboxydovores can oxidize CO in seawater. New molecular and isolation techniques, as well as genome sequencing, have greatly expanded our knowledge of the diversity of CO oxidizers (e.g. King and Weber, 2007). Typically, CO microbial consumption follows first-order reaction kinetics in seawater (Johnson and Bates, 1996; Zafiriou et al., 2003; Wang et al., 2015; Zhang et al., 2019; Sugai et al., 2020; Xu et al., 2023). [ #$ ] != [ #$ ] "'#$! (Eq. 1) Introduction 14 The [CO]t and [CO]0 are concentrations of CO at time t and 0, respectively. B is the CO microbial consumption rate constant (kbio), which is used to characterize the speed of CO microbial consumption in seawater. Three main methods are commonly utilized to determine kbio: dark incubation, 14CO tracer, and night loss. 1) Dark incubation method This simple method is to incubate whole-water samples in the dark. Since CO loss kinetics are thought to be first-order (Jones and Morita 1983, 1984; Johnson and Bates 1996; Ohta et al., 2000; Zafiriou et al. 2003; Xie et al., 2005), the absolute value of the exponential decay fitted to the exponential equation is kbio. A drawback of this method is that samples must be incubated in the dark to prevent interference from photochemical CO production, thereby ignoring any effects of light, such as potential inhibitory effects of solar ultraviolet radiation on bacterial activity (Herndl et al., 1993; Sommaruga et al., 1997; Tolli and Taylor, 2005). 2) 14CO tracer method For this method, 14CO is added to the seawater sample as a tracer to determine the trend of 14CO over time (time–series biological oxidation of 14C–CO to 14C–CO2) (Butler et al., 1987; Jones, 1991; Jones and Amador, 1993; Tolli, 2003; Tolli and Taylor, 2005). This method can be performed under lighted conditions, considering the potential inhibitory effects of solar ultraviolet radiation on bacterial activity. However, the CO concentration within the incubation mixture is typically 2 to 10-fold above oceanic values. The microbial consumption of CO at such high concentrations may change from a first-order reaction to a zero-order reaction (i.e., the reaction rate does not change with the substrate concentration). 3) Night loss method This method is based on monitoring the trends of CO concentration deducted by CO flux loss at night over time. The absolute value of the exponential term fitted to the exponential equation is kbio (Johnson and Bates, 1996; Ohta, 1997). Introduction 15 The kbio of CO in different areas is quite different (Table 1.2.2). Generally speaking, the kbio in an estuary area is higher than that on the continental shelf, and the kbio in the ocean is the lowest. Xie et al. (2005) recorded a maximum kbio of 2.44 h-1 in Delaware Bay, suggesting that microbial oxidation could complete a CO cycle in as little as 25 minutes within the bay. In contrast, the Northwest Atlantic Ocean exhibited a maximum kbio of 0.18 h⁻¹, indicating a potential cycle duration of up to 5.6 hours for CO in this marine area. Many factors affect kbio, such as bacterial species and abundance, the concentration of organic matter, temperature, salinity, pH, Chl-a, and nutrient concentration. Xie et al. (2005) found that when salinity is 0–19, kbio is increased with salinity; However, beyond 19, kbio decreases with the increase of salinity. The values of kbio are negatively correlated with salinity in the offshore area of China (Xie et al., 2005; Yang et al., 2010). Additionally, kbio has an evident positive correlation with Chl-a (Xie et al., 2005). Additionally, notable seasonal variations occur in the same area. Broadly, kbio tends to be higher during summer and lower during winter. Xie et al. (2009) could not determine the primary reason for this disparity and further research is needed. Introduction 16 Table 1.2.2: Microbial consumption rates in different study areas. Study area Season kbio (h-1) Method Reference Range Average Pacific Ocean – -0.021–0.12 0.036 Dark incubation Zafiriou et al., 2003 Pacific Ocean (equator) – – 0.13 Night loss method Ohta, 1997 Pacific Ocean (equator) Winter 0.00042–0.025 0.014 Dark incubation Zafiriou et al., 2003 South pacific (tropical) Spring – 0.038 Night loss method Johnson and Bates, 1996 Sargasso Sea Summer and autumn 0.0026–0.037 0.0084 14CO tracer method (Dark) Jones, 1991 Summer 0.019–0.028 0.023 14CO tracer method (Dark) Tolli and Taylor, 2005 0.0017–0.0025 – 14CO tracer method Sargasso Sea Spring 0.0083–0.017 0.01 14CO tracer method (Dark) Tolli and Taylor, 2005 0.0041–0.0062 – 14CO tracer method (Light) Eastern Caribbean Sea Spring and autumn 0.010–0.071 0.02 14CO tracer method (Dark) Jones and Amador, 1993 Open NW Atlantic Summer 0.071–0.18 0.099 Dark incubation Xie et al., 2005 Southern Ocean Summer – 0.004 Dark incubation Zafiriou et al., 2003 Offshore Beaufort Sea Autumn 0.011–0.025 0.02 Dark incubation Xie et al., 2005 Eastern Indian Ocean Autumn 0.016–0.113 0.045 Dark incubation Xu et al., 2023 Orinoco delta Spring and autumn 0.0092–0.5 0.069 14CO tracer method (Dark) Jones and Amador, 1993 Gulf of Paria Spring and autumn 0.053–0.33 0.1 E Caribbean Sea Spring and autumn 0.014–0.17 0.034 Vineyard Sound Winter 0.010–0.11 0.064 14CO tracer method (Dark) Tolli and Taylor, 2005 Seto Inland Sea Spring 0.38–0.51 0.44 Night loss method Ohta et al., 2000 0.35–0.48 0.41 Dark incubation method Coastal NW Atlantic Summer 0.096–0.91 0.33 Dark incubation Xie et al., 2005 Coastal Beaufort Sea Autumn 0.028–0.058 0.04 Amundsen Gulf Spring 0.10–0.43 0.23 Dark incubation Xie et al., 2009a Autumn 0.26–0.74 0.49 The South Yellow Sea and East China Sea November 0.15–2.14 0.80 Dark incubation Yang et al., 2010 The BS and the YS Autumn 0.104–0.310 0.20 Dark incubation Zhang et al., 2019 Introduction 17 Yaquina Bay All year round 0.0062–0.59 0.072 14CO tracer method (Dark) Butler et al., 1987 Delaware Bay Summer 0.26–2.44 1.11 Dark incubation Xie et al., 2005 Ise Bay Summer – 0.63 Night loss method Ohta et al., 2000 St. Lawrence Estuary May 0.053–1.01 0.14 Dark incubation Xie et al., 2009b; aZhang, 2008 July 0.081–0.60 0.27 October 0.035–0.71 0.13 December 0.024–0.38 0.032 Sagami Bay all year round 0.012–0.094 – Dark incubation Sugai et al., 2020 1.2.3.2 Sea–air exchange Sea–air fluxes of heat, momentum, moisture, and trace gases are critical components of our planet’s climate system. The ocean is a minor source of atmospheric CO, contributing about 0.4 to 0.8 % to the natural and anthropogenic sources of atmospheric CO (Conte et al., 2019; Zheng et al., 2019). The CO emissions estimated during the last several decades span two orders of magnitude (depending on different study areas), ranging from 3.7 to 600 Tg CO–C yr-1 (Swinnerton et al., 1971; Linnenbom et al., 1973; Seiler, 1974; Conrad et al., 1982; Bates et al., 1995; Zuo and Jones, 1995). With the spatial and temporal diversity of processes in controlling oceanic CO, this range of CO emissions reflects the scarcity of existing data. In recent years, with applications in numerical models and for estimating fluxes from larger datasets, investigations have found reliably that the amount of CO in the ocean diffusing into the atmosphere is 4~20 Tg CO–C yr-1 (Bates et al., 1995; Zafiriou et al., 2003; Stubbins et al., 2006a; Zhao et al., 2015; Park and Rhee, 2016; Conte et al., 2019; Zheng et al., 2019). However, CO oceanic emissions are still associated with a high degree of uncertainty and further research is needed. Introduction 18 References Andreae, M. O., Fishman, J., and Lindesay, J.: The Southern Tropical Atlantic Region Experiment (STARE): Transport and Atmospheric Chemistry near the Equator‐Atlantic (TRACE A) and Southern African Fire‐Atmosphere Research Initiative (SAFARI): An introduction. Journal of Geophysical Research: Atmospheres, 101(D19), 23519-23520, https://doi.org/10.1029/96JD01786, 1996. Andreae, M. O., and Merlet, P.: Emission of trace gases and aerosols from biomass burning. Global biogeochemical cycles, 15(4), 955-966, https://doi.org/10.1029/2000GB001382, 2001. Bange, H. W.: Non-CO2 greenhouse gases (N2O, CH4, CO) and the ocean. One Earth, 5(12), 1316-1318, https://doi.org/10.1016/j.oneear.2022.11.011, 2022. Bates, T. S., Kelly, K. C., Johnson, J. E., and Gammon, R. H.: Regional and seasonal variations in the flux of oceanic carbon monoxide to the atmosphere. Journal of Geophysical Research: Atmospheres, 100(D11), 23093-23101, https://doi.org/10.1029/95JD02737, 1995. Bergamaschi, P., Hein, R., Heimann, M., and Crutzen, P. J.: Inverse modeling of the global CO cycle: 1. Inversion of CO mixing ratios. Journal of Geophysical Research: Atmospheres, 105(D2), 1909-1927, https://doi.org/10.1029/1999JD900818, 2000. Blake, N. J., Blake, D. R., Wingenter, O. W., Sive, B. C., McKenzie, L. M., Lopez, J. P., ... and Rowland, F. S.: Influence of southern hemispheric biomass burning on midtropospheric distributions of nonmethane hydrocarbons and selected halocarbons over the remote South Pacific. Journal of Geophysical Research: Atmospheres, 104(D13), 16213-16232, https://doi.org/10.1029/1999JD900067, 1999. Blomquist, B. W., Fairall, C. W., Huebert, B. J., and Wilson, S. T.: Direct measurement of the oceanic carbon monoxide flux by eddy correlation, Atm. Measurement Techniques, 5, 3069-3075, 2012. Bourbonniere, R. A., Miller, W. L., and Zepp, R. G.: Distribution, flux, and photochemical production of carbon monoxide in a boreal beaver impoundment. Journal of Geophysical Research: Atmospheres, 102(D24), 29321-29329, https://doi.org/10.1029/97JD02234, 1997. Boucher, O., Randall, D., Artaxo, P., Bretherton, C., Feingold, G., Forster, P., Kerminen, V.-M., Kondo, Y. , Li ao, H ., an d L o hm ann , U .: Cl oud s a nd ae ros ols , i n: C l im ate c h an ge 20 13: t h e ph ysi cal s cie nce b a si s. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, 571–657, 2013. Bushaw, K. L., Zepp, R. G., Tarr, M. A., Schulz-Jander, D., Bourbonniere, R. A., Hodson, R. E., ... and Moran, M. A.: Photochemical release of biologically available nitrogen from aquatic dissolved organic matter. Nature, 381(6581), 404-407, https://doi.org/10.1038/381404a0, 1996. Introduction 19 Butler, J. H., Jones, R. D., Garber, J. H., and Gordon, L. I.: Seasonal distributions and turnover of reduced trace gases and hydroxylamine in Yaquina Bay, Oregon. Geochimica et Cosmochimica Acta, 51(3), 697706, https://doi.org/10.1016/0016-7037(87)90080-9, 1987. Butler, J. H., Pequegnat, J. E., Gordon, L. I., and Jones, R. D.: Cycling of methane, carbon monoxide, nitrous oxide, and hydroxylamine in a meromictic, coastal lagoon, Estuarine, Coastal and Shelf Science, 27(2), 181-203, https://doi.org/10.1016/0272-7714(88)90089-3, 1988. Campen, H. I., Arévalo-Martínez, D. L., and Bange, H. W.: Carbon monoxide (CO) cycling in the Fram Strait, Arctic Ocean, Biogeosciences, 20, 1371–1379, https://doi.org/10.5194/bg-20-1371-2023, 2023. Cazenave, A., Dieng, H. B., Meyssignac, B., Von Schuckmann, K., Decharme, B., and Berthier, E.: The rate of sea-level rise. Nature Climate Change, 4(5), 358–361, https://doi.org/10.1038/nclimate2159, 2014. Chapman, D. J., and Tocher, R. D.: Occurrence and production of carbon monoxide in some brown algae. Canadian Journal of Botany, 44(10), 1438-1442, https://doi.org/10.1139/b66-158, 1966. Cicerone, R. J. How has the atmospheric concentration of CO changed? In: F.S. Rowland and I.S.A. Isaksen (Editors), The Changing Atmosphere. John Wiley and Sons, pp. 49~61, 1988 Conte, L., Szopa, S., Séférian, R., and Bopp, L.: The oceanic cycle of carbon monoxide and its emissions to the atmosphere, Biogeosciences, 16, 881–902, https://doi.org/10.5194/bg-16-881-2019, 2019. Conrad, R.: Soil microbial processes and the cycling of atmospheric trace gases. Philosophical Transactions of the Royal Society of London. Series A: Physical and Engineering Sciences, 351(1696), 219-230, https://doi.org/10.1098/rsta.1995.0030, 1995. Conrad, R., Seiler, W., Bunse, G., and Giehl, H.: Carbon monoxide in seawater (Atlantic Ocean), J. Geophys. Res., 87, 8839, https://doi.org/10.1029/JC087iC11p08839, 1982. Cordero, P. R., Bayly, K., Man Leung, P., Huang, C., Islam, Z. F., Schittenhelm, R. B., ... and Greening, C.: Atmospheric carbon monoxide oxidation is a widespread mechanism supporting microbial survival. The ISME journal, 13(11), 2868-2881, https://doi.org/10.1038/s41396-019-0479-8, 2019. Crutzen, P. J. Role of the tropics in atmospheric chemistry. In: R.E. Dickinson (Editor), The Geophysiology of Amazonia. John Wiley, New York, 1987, 107~130. Day, D. A., and Faloona, I.: Carbon monoxide and chromophoric dissolved organic matter cycles in the shelf waters of the northern California upwelling system. Journal of Geophysical Research: Oceans, 114(C1), https://doi.org/10.1029/2007JC004590, 2009. Derwent, R. G.:Air chemistry and terrestrial gas emissions: a global perspective. Philosophical Transactions of the Royal Society of London. Series A: Physical and Engineering Sciences, 351(1696), 205-217, https://doi.org/10.1098/rsta.1995.0029, 1995. Introduction 20 Dickinson, R. E. and Cicerone, R. J.: Future global warming from atmospheric trace gases, Nature, 319, 109–115, https://doi.org/10.1038/319109a0, 1986. Dignon, J., and Hameed, S.: A model investigation of the impact of increases in anthropogenic NOx emissions between 1967 and 1980 on tropospheric ozone. Journal of Atmospheric Chemistry, 3, 491-506, https://doi.org/10.1007/BF00053873, 1985. Dyurgerov, M. B., and Meier, M. F.: Twentieth century climate change: evidence from small glaciers. Proceedings of the National Academy of Sciences, 97(4), 1406-1411, https://doi.org/10.1073/pnas.97.4.1406, 2000. Ehhalt, D. H.: Photooxidation of trace gases in the troposphere Plenary Lecture. Physical Chemistry Chemical Physics, 1(24), 5401-5408, https://doi.org/10.1039/A905097C, 1999. Erickson III, D. J.: Ocean to atmosphere carbon monoxide flux: Global inventory and climate implications. Global Biogeochemical Cycles, 3(4), 305-314, https://doi.org/10.1029/GB003i004p00305, 1989. Evans, W. F. J., and Puckrin, E.: An observation of the greenhouse radiation associated with carbon monoxide. Geophysical research letters, 22(8), 925-928, https://doi.org/10.1029/95GL00606, 1995. Fichot, C. G., and Miller, W. L.: An approach to quantify depth-resolved marine photochemical fluxes using remote sensing: Application to carbon monoxide (CO) photoproduction. Remote Sensing of Environment, 114(7), 1363-1377, https://doi.org/10.1016/j.rse.2010.01.019, 2010. Forster, P., Storelvmo, T., Armour, K., Collins, W., Dufresne, J.-L., Frame, D., Lunt, D.J., Mauritsen, T., Palmer, M.D., Watanabe, M., Wild, M., and Zhang, H.: The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity, in: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J.B.R., Maycock, T.K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou. B., Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 923–1054, https://doi.org/10.1017/9781009157896.009, 2021. Forster, P. M., Smith, C. J., Walsh, T., Lamb, W. F., Lamboll, R., Hauser, M., Ribes, A., Rosen, D., Gillett, N., Palmer, M. D., Rogelj, J., von Schuckmann, K., Seneviratne, S. I., Trewin, B., Zhang, X., Allen, M., Andrew, R., Birt, A., Borger, A., Boyer, T., Broersma, J. A., Cheng, L., Dentener, F., Friedlingstein, P., Gutiérrez, J. M., Gütschow, J., Hall, B., Ishii, M., Jenkins, S., Lan, X., Lee, J.-Y. , M or ic e, C . , K a dow, C ., Kennedy, J., Killick, R., Minx, J. C., Naik, V., Peters, G. P., Pirani, A., Pongratz, J., Schleussner, C.-F., Szopa, S., Thorne, P., Rohde, R., Rojas Corradi, M., Schumacher, D., Vose, R., Zickfeld, K., MassonDelmotte, V., and Zhai, P.: Indicators of Global Climate Change 2022: annual update of large-scale indicators of the state of the climate system and human influence, Earth Syst. Sci. Data, 15, 2295–2327, https://doi.org/10.5194/essd-15-2295-2023, 2023. Introduction 21 Freyer, H. D.: Atmospheric cycles of trace gases containing carbon. The Global Carbon Cycle, 101-28, 1979 Gao, H., and Zepp, R. G.: Factors influencing photoreactions of dissolved organic matter in a coastal river of the southeastern United States. Environmental Science & Technology, 32(19), 2940-2946, https://doi.org/10.1021/es9803660, 1998. Gaubert, B., Worden, H. M., Arellano, A. F. J., Emmons, L. K., Tilmes, S., Barré, J., ... and Edwards, D. P.: Chem ical feed bac k fro m d ecr easi ng c arbon m onox ide emis sions. G eoph ysi ca l Re search L et ter s, 44(19), 9985-9995, https://doi.org/10.1002/2017GL074987, 2017. Gnanadesikan, A.: Modeling the diurnal cycle of carbon monoxide: Sensitivity to physics, chemistry, biology, and optics. Journal of Geophysical Research: Oceans, 101(C5), 12177-12191, https://doi.org/10.1029/96JC00463, 1996. Granier, C., Mueller, J. F., Pétron, G., and Brasseur, G.: A three-dimensional study of the global CO budget. Chemosphere-Global Change Science, 1(1-3), 255-261, https://doi.org/10.1016/S14659972(99)00007-0, 1999. Greening, C., and Grinter, R.: Microbial oxidation of atmospheric trace gases. Nature Reviews Microbiology, 20(9), 513-528, https://doi.org/10.1038/s41579-022-00724-x, 2022. Gros, V., Peeken, I., Bluhm, K., Zöllner, E., Sarda-Esteve, R., and Bonsang, B.: Carbon monoxide emissions by phytoplankton: evidence from laboratory experiments, Environ. Chem., 6, 369, https://doi.org/10.1071/EN09020, 2009. Gödde, M., Meuser, K., and Conrad, R.: Hydrogen consumption and carbon monoxide production in soils with different properties. Biology and fertility of soils, 32, 129-134, https://doi.org/10.1007/s003740000226, 2000. Herndl, G. J., Müller-Niklas, G., and Frick, J.: Major role of ultraviolet-B in controlling bacterioplankton growth in the surface layer of the ocean. Nature, 361(6414), 717-719, https://doi.org/10.1038/361717a0, 1993. Hoesly, R. M., Smith, S. J., Feng, L., Klimont, Z., Janssens-Maenhout, G., Pitkanen, T., Seibert, J. J., Vu, L., Andres, R. J., Bolt, R. M., Bond, T. C., Dawidowski, L., Kholod, N., Kurokawa, J.-I., Li, M., Liu, L., Lu, Z., Moura, M. C. P., O'Rourke, P. R., and Zhang, Q.: Historical (1750–2014) anthropogenic emissions of reactive gases and aerosols from the Community Emissions Data System (CEDS), Geosci. Model Dev., 11, 369–408, https://doi.org/10.5194/gmd-11-369-2018, 2018. Holloway, T., Levy, H., and Kasibhatla, P.: Global distribution of carbon monoxide. Journal of Geophysical Research: Atmospheres, 105(D10), 12123-12147, https://doi.org/10.1029/1999JD901173, 2000. Introduction 28 carbon monoxide photoproduction, Deep-Sea Res. Pt. II, 53, 1695–1705, https://doi.org/10.1016/j.dsr2.2006.05.011, 2006b. Sugai, Y., Tsuchiya, K., Shimode, S., and Toda, T.: Photochemical Production and Biological Consumption of CO in the SML of Temperate Coastal Waters and Their Implications for Air‐Sea CO Exchange. Journal of Geophysical Research: Oceans, 125(4), e2019JC015505, https://doi.org/10.1029/2019JC015505, 2020. Swinnerton, J. W., Linnenbom, V. J., and Cheek, C. H.: A sensitive gas chromatographic method for determining carbon monoxide in seawater. Limnology and oceanography, 13(1), 193-195, https://doi.org/10.4319/lo.1968.13.1.0193, 1968. Swinnerton, J. W., Linnenbom, V. J., and Lamontagne, R. A.: The ocean: a natural source of carbon monoxide. Science, 167(3920), 984-986, https://doi.org/10.1126/science.167.3920.984, 1970. Swinnerton, J. W., and Lamontagne, R. A.: Carbon monoxide in the South Pacific Ocean 1. Tellus, 26(1‐ 2), 136-142, https://doi.org/10.1111/j.2153-3490.1974.tb01959.x, 1974. Swinnerton, J. W., Lamontagne, R. A., and Linnenbom, V. J.: Carbon monoxide in rainwater. Science, 172(3986), 943-945, https://doi.org/10.1126/science.172.3986.943, 1971. Taylor, J. A., Zimmerman, P. R., and Erickson III, D. J.: A 3-D modelling study of the sources and sinks of atmospheric carbon monoxide. Ecological modelling, 88(1-3), 53-71, https://doi.org/10.1016/03043800(95)00069-0, 1996. Thompson, A. M.: The oxidizing capacity of the Earth's atmosphere: Probable past and future changes. Science, 256(5060), 1157-1165, https://doi.org/10.1126/science.256.5060.1157, 1992. Thompson, A. M., and Cicerone, R. J.: Possible perturbations to atmospheric CO, CH4, and OH. Journal of Geophysical Research: Atmospheres, 91(D10), 10853-10864, https://doi.org/10.1029/JD091iD10p10853, 1986. Tolli, J. D.: Identity and dynamics of the microbial community responsible for carbon monoxide oxidation in marine environments. MASSACHUSETTS INST OF TECH CAMBRIDGE, 2003. Tolli, J. D., and Taylor, C. D.: Biological CO oxidation in the Sargasso Sea and in vineyard sound, Massachusetts. Limnology and oceanography, 50(4), 1205-1212, https://doi.org/10.4319/lo.2005.50.4.1205, 2005. Tran, S., Bonsang, B., Gros, V., Peeken, I., Sarda-Esteve, R., Bernhardt, A., and Belviso, S.: A survey of carbon monoxide and non-methane hydrocarbons in the Arctic Ocean during summer 2010, Biogeosciences, 10, 1909–1935, https://doi.org/10.5194/bg-10-1909-2013, 2013. Walsh M P. Motor vehicles and global warming. In: Leggett J, eds. Global Warming: The Greenpeace Introduction 29 Repor. Oxford: Oxford University Press, 1990, pp. 260-294. Wang, W. L., Yang, G. P., and Lu, X. L.: Carbon monoxide distribution and microbial consumption in the Southern Yellow Sea. Estuarine, Coastal and Shelf Science, 163, 125-133, https://doi.org/10.1016/j.ecss.2015.06.012, 2015. Wang, W. L., Peng, T., Lu, X. L., and Zhao, B. Z.: Diurnal, seasonal, and spatial variations and flux of carbon monoxide in Jiaozhou Bay, China, Marine Chemistry, 191, 1-8, https://doi.org/10.1016/j.marchem.2017.01.004, 2017. White, E. M., Kieber, D. J., Sherrard, J., Miller, W. L., and Mopper, K.: Carbon dioxide and carbon monoxide photoproduction quantum yields in the Delaware Estuary. Marine Chemistry, 118(1-2), 11-21, https://doi.org/10.1016/j.marchem.2009.10.001, 2010. WHO.: Report on Environmental Health Criteria 213, Carbon Monoxide (Second Edition), World Health Organization, Geneva, 2004. Wilson, D. F., Swinnerton, J. W., and Lamontagne, R. A.: Production of Carbon Monoxide and Gaseous Hydrocarbons in Seawater: Relation to Dissolved Organic Carbon, Science, 168, 1577– 1579, https://doi.org/10.1126/science.168.3939.1577, 1970. Wunderling, N., Willeit, M., Donges, J. F., and Winkelmann, R.: Global warming due to loss of large ice masses and Arctic summer sea ice. Nature Communications, 11(1), 5177, https://doi.org/10.1038/s41467-020-18934-3, 2020. Xie, H., Andrews, S. S., Martin, W. R., Miller, J., Ziolkowski, L., Taylor, C. D., and Zafiriou, O. C.: Val ida ted met hods for sampling and headspace analysis of carbon monoxide in seawater. Marine Chemistry, 77(2-3), 93-108, https://doi.org/10.1016/S0304-4203(01)00065-2, 2002. Xie, H., Bélanger, S., Song, G., Benner, R., Taalba, A., Blais, M., ... and Babin, M.: Photoproduction of ammonium in the southeastern Beaufort Sea and its biogeochemical implications, Biogeosciences, 9,3047-3061, doi: 1O.5194/bg-9-3047-2012, 2012. Xie, H., Zafiriou, O. C., Umile, T. P., and Kieber, D. J.: Biological consumption of carbon monoxide in Delaware Bay, NW Atlantic and Beaufort Sea, Mar. Ecol. Prog. Ser., 290, 1–14, 2005. Xie, H., Bélanger, S., Demers, S., Vincent, W. F., and Papakyriakou, T. N.: Photobiogeochemical cycling of carbon monoxide in the southeastern Beaufort Sea in spring and autumn. Limnology and Oceanography, 54(1), 234-249, https://doi.org/10.4319/lo.2009.54.1.0234, 2009a. Xie, H. and Zafiriou, O. C.: Evidence for significant photochemical production of carbon monoxide by particles in coastal and oligotrophic marine waters, Geophys. Res. Lett., 36, L23606, https://doi.org/10.1029/2009GL041158, 2009. Introduction 30 Xie, H., Zhang, Y., Lemarchand, K., and Poulin, P.: Microbial carbon monoxide uptake in the St. Lawrence estuarine system. Marine Ecology Progress Series, 389, 17-29, https://doi.org/10.3354/meps08175, 2009b. Xu, G. B., Xu, F., Ji, X., Zhang, J., Yan, S. B., Mao, S. H., and, Yang, G. P.: Carbon monoxide cycling in the Eastern Indian Ocean. Journal of Geophysical Research: Oceans, e2022JC019411, https://doi.org/10.1029/2022JC019411, 2023. Yang, G. P., Wang, W. L., Lu, X. L., and Ren, C. R.: Distribution, flux and biological consumption of carbon monoxide in the Southern Yellow Sea and the East China Sea, Marine Chemistry, 122(1‐4), 74– 82, https://doi.org/10.1016/j.marchem.2010.08.001, 2010. Yang, G. P., Ren, C. Y., Lu, X. L., Liu, C. Y., and Ding, H. B.: Distribution, flux, and photoproduction of carbon monoxide in the East China Sea and Yellow Sea in spring, J. Geophys. Res., 116, C02001, https://doi.org/10.1029/2010JC006300, 2011. Zafiriou, O. C.: Sunburnt organic matter: Biogeochemistry of light‐altered substrates. Limnology and Oceanography Bulletin, 11(4), 69-74, https://doi.org/10.1002/lob.200211469, 2002. Zafiriou, O. C., Andrews, S. S., and Wang, W.: Concordant estimates of oceanic carbon monoxide source and sink processes in the Pacific yield a balanced global “blue-water” CO budget: concordant oceanic CO budgets, Global Biogeochem. Cy., 17, 1015, https://doi.org/10.1029/2001GB001638, 2003. Zafiriou, O. C., Xie, H., Nelson, N. B., Najjar, R. G., and Wang, W.: Diel carbon monoxide cycling in the upper Sargasso Sea near Bermuda at the onset of spring and in midsummer, Limnol. Oceanogr., 53, 835–850, 2008. Zepp, R. G.: Solar UVR and aquatic carbon, nitrogen, sulfur and metals cycles. UV effects in aquatic organisms and ecosystems, 137-183, 2003. Ziolkowski, L. A., and Miller, W. L.: Variability of the apparent quantum efficiency of CO photoproduction in the Gulf of Maine and Northwest Atlantic. Marine chemistry, 105(3-4), 258-270, https://doi.org/10.1016/j.marchem.2007.02.004, 2007. Zhai, P. M., and Liu, J.: Extreme weather/climate events and disaster prevention and mitigation under global warming background. Engineering Sciences, 14(9), 55–63, 2012. Zhang, Y., and Xie, H.: The sources and sinks of carbon monoxide in the St. Lawrence estuarine system. Deep Sea Research Part II: Topical Studies in Oceanography, 81, 114-123, https://doi.org/10.1016/j.dsr2.2011.09.003, 2012. Zhang, Y., Xie, H., Fichot, C. G., and Chen, G.: Dark production of carbon monoxide (CO) from dissolved organic matter in the St. Lawrence estuarine system: Implication for the global coastal and blue water CO budgets, J. Geophys. Res., 113, C12020, https://doi.org/10.1029/2008JC004811, 2008. Introduction 31 Zhang, J., Wang, J., Zhuang, G. C., and Yang, G. P.: Carbon monoxide cycle in the Bohai Sea and the Yellow Sea: Spatial variability, sea‐air exchange, and biological consumption in autumn. Journal of Geophysical Research: Oceans, 124(6), 4248-4257, https://doi.org/10.1029/2018JC014864, 2019. Zhao, B. Z., Yang, G. P., Xie, H., Lu, X. L., and Yang, J.: Distribution, flux and photoproduction of carbon monoxide in the Bohai and Yellow Seas. Marine Chemistry, 168, 104-113, https://doi.org/10.1016/j.marchem.2014.11.006, 2015. Zheng, B., Chevallier, F., Yin, Y., Ciais, P., Fortems-Cheiney, A., Deeter, M. N., Parker, R. J., Wang, Y., Worden, H. M., and Zhao, Y.: Global atmospheric carbon monoxide budget 2000–2017 inferred from multi-species atmospheric inversions, Earth Syst. Sci. Data, 11, 1411-1436, 2019. Zhou, X., and Mopper, K.: Photochemical production of low-molecular-weight carbonyl compounds in seawater and surface microlayer and their air-sea exchange. Marine Chemistry, 56(3-4), 201-213, https://doi.org/10.1016/S0304-4203(96)00076-X, 1997. Zuo, Y. and Jones, R. D.: Formation of carbon monoxide by photolysis of dissolved marine organic material and its significance in the carbon cycling of the oceans, Naturwissenschaften, 82, 472– 474, 1995. Introduction 32 Thesis Outline 33 2 Thesis Outline The climatically relevant trace gas CO plays an important role in climate regulation. During the past decades, extensive field studies have been conducted to investigate the temporal distribution of CO and the controlling factors in the world’s oceans. However, lagoon systems in anthropogenically influenced coastal areas are considered a potentially significant source of CO emissions, despite the insufficient spatial sampling there. Moreover, the vast majority of studies focus on short-term investigations, and time-series studies are very sparse, which hinders our understanding of CO variability under changing seasonal conditions. Notably, just two geographical distributions of oceanic CO emission for use in global atmospheric chemistry models were derived from a single linear regression or NEMO-PISCES (Nucleus for European Modelling of the Ocean, Pelagic Interaction Scheme for Carbon and Ecosystem Studies) model estimation, respectively (Erickson, 1989; Conte., 2019). With large and increasingly controversial, uncertainties in its source strength, the oceanic natural source of atmospheric CO needs to be better characterized in terms of processes. Overall, to address these knowledge gaps, this PhD thesis is divided into the following three chapters, and each one addresses one underinvestigated scientific question: What are the sources and sinks of carbon monoxide in the lagoon system (Ria Formosa Lagoon) and is the aquaculture area a significant source of CO emissions? Chapter 2: The Ria Formosa Lagoon system, located in southern Portugal along the Northeast (NE) Atlantic Ocean coast, is a coastal area profoundly impacted by human activities. Its circulation is constrained, resembling an estuarine system, and experiences notable nutrient and organic matter inflows. This makes the Ria Formosa Lagoon system a potential source of atmospheric CO. In this chapter, we present the first in situ measurements of CO within this area, integrating these observations with incubation experiments conducted in waters directly Thesis Outline 34 affected by anthropogenic inputs. We elucidated an analysis of CO distribution, various CO sources, and sinks during the summer season. Moreover, we assess the potential influence of aquaculture activities on CO cycling within the region. What processes affect the seasonal variations and distributions of carbon monoxide in the water column at BE? Chapter 3: In this chapter, a monthly data set in 2020 of CO, as well as various environmental and phytoplanktonic parameters, in the water column at the well-established Boknis Eck timeseries station, in the southwest Baltic Sea, are presented. The seasonal variability and distribution of CO and the special processes affecting its marine cycling are identified. Is it possible to propose an updated global scale distribution of oceanic CO emissions, substantially reducing the uncertainty and enabling a tightening of the global budget of CO? Chapter 4: In this chapter, we employ an extensive dataset comprising over 10,000 sea-surface CO observations to reconstruct oceanic emissions through a data-driven machine-learning approach. We provide global predictions that significantly mitigate uncertainties when compared to prior estimates, facilitating a more precise determination of the global oceanic CO budget. This reconstruction unveils a vigorous seasonal cycle and, notably, presents the first estimation of CO emissions within the Southern Ocean. Carbon monoxide in the Ria Formosa Lagoon 35 3 Carbon monoxide cycling in the Ria Formosa Lagoon (southern Portugal) during summer 2021 Submitted as: Li, G., Arévalo-Martínez, D. L., Ingeniero, R. C. O., and Bange, H. W.: Carbon monoxide cycling in the Ria Formosa Lagoon (southern Portugal) during summer 2021, [preprint], https://doi.org/10.5194/egusphere-2023-771, 2023. Abstract. Carbon monoxide (CO) is an atmospheric trace gas that plays a crucial role in the oxidizing capacity of the Earth’s atmosphere. Moreover, it functions as an indirect greenhouse gas, influencing the lifetimes of potent greenhouse gases such as methane. Albeit being an overall source of atmospheric CO, the role of coastal regions in the marine cycling of CO and how its budget can be affected by anthropogenic activities, remain uncertain. Here, we present the first measurements of dissolved CO in the Ria Formosa Lagoon, an anthropogenically influenced system in southern Portugal. The dissolved CO concentrations in the surface layer ranged from 0.16 to 3.1 nmol L-1 with an average concentration of 0.75 ± 0.57 nmol L-1. The CO saturation ratio ranged from 1.7 to 32.2, indicating that the lagoon acted as a source of CO to the atmosphere in May 2021. The estimated average sea-to-air flux density was 1.53 μmol m-2 d-1, mainly fuelled by CO photochemical production. Microbial consumption accounted for 83 % of the CO production, suggesting that the resulting CO emissions to the atmosphere were modulated by microbial consumption in the surface waters. The results from an irradiation experiment with aquaculture effluent water indicated that aquaculture facilities in the Ria Formosa Lagoon seem to be a negligible source of atmospheric CO. Carbon monoxide in the Ria Formosa Lagoon 36 3.1 Introduction Carbon monoxide (CO) plays an important role in the oxidizing capacity of the Earth’s atmosphere. In the troposphere, CO rapidly reacts with hydroxyl radicals (OH), and in the presence of NOx, this reaction leads to the formation of ozone, a pollutant, and a greenhouse gas (Thompson, 1992). Moreover, due to its reactivity with OH, CO indirectly influences the lifetime of atmospheric methane, a potent greenhouse gas (GHG) (Canadell et al., 2021). Consequently, CO is an important indirect greenhouse gas with an effective radiative forcing comparable to nitrous oxide (Forster et al., 2021). The open and coastal oceans contribute up to 1 % of the global atmospheric CO sources (Zheng et al., 2019). Current estimates indicate that oceanic emissions of CO range between 4 Tg CO– C yr-1 and 20 Tg CO–C yr-1 (Stubbins et al., 2006; Park and Rhee, 2016; Conte et al., 2019; Zhang et al., 2019), while some earlier estimates (published before 1999) indicate oceanic CO emissions of up to 1200 Tg CO yr-1 (see overview in Zafiriou et al., 2003). Coastal areas appear to be ‘hot spots’ of CO emissions. Park and Rhee (2016) estimated that about 15–25 % of the global oceanic CO emissions can be attributed to coastal regions. The primary source of CO is its photochemical production via photo-oxidation of chromophoric (colored) dissolved organic matter (CDOM) in the open and coastal oceans (e.g., Zafiriou et al., 2003; Stubbins et al., 2006). Production by phytoplankton and the so-called ‘dark production’ are comparatively small but significant sources of oceanic CO (Zhang et al., 2008; Gros et al., 2009; McLeod et al., 2021). Microbial consumption is the major CO loss mechanism, accounting for about 80–90 % of the CO loss from the marine environment (Zafiriou et al., 2003; Moran and Miller, 2007; Greening and Grinter, 2022). Therefore, CO emissions to the atmosphere are a minor sink of oceanic CO (Zafiriou et al., 2003). Ria Formosa is a major lagoon system in southern Portugal (36°58’ N, 8°02’ W to 37°03’ N, 7°32’ W) (Newton and Mudge, 2003; Cravo et al., 2014). It is a mesotidal coastal lagoon with an average water depth of ~2 m. Water temperatures typically range from about 12 °C in the winter to about 27 °C in the summer (Newton and Mudge, 2003). Salinities range from 13 to Carbon monoxide in the Ria Formosa Lagoon 37 36.5 (Newton and Mudge, 2003). The lagoon is vertically well-mixed due to limited freshwater inputs and the predominance of a tidal-forced circulation pattern (Cravo et al., 2014). Being a typical coastal lagoon system, the water exchange with the adjacent shelf is restricted (Tett et al., 2003). In addition to nutrient inputs from aquaculture facilities, the Ria Formosa Lagoon receives significant nutrient inputs from agricultural runoff and sewage, particularly from the cities of Faro, Olhão, and Tavira (Newton and Mudge, 2005). A part of the Ria Formosa Lagoon system is a natural park established in 1987 (Aníbal et al., 2019). Ria Formosa also plays a vital role in the region’s tourism industry as well as supporting a fishery and aquaculture industry of national significance. Due to its restricted (estuarine-type) circulation and high nutrient and organic matter inputs, the Ria Formosa Lagoon system could be a potential source of atmospheric CO. The unique setting at the Ria Formosa lagoon system provides an excellent opportunity to investigate the dynamics of carbon monoxide in an anthropogenically-influenced coastal system. Additionally, eutrophication and algal bloom often occur in aquaculture ponds in Ria Formosa, which should be a non-negligible component of the RF–CO cycle process. How these environmental factors will affect CO production and emissions from the lagoon system (accounting for approximately 13% of the coastal areas worldwide) is unknown so far due to limited measurements and knowledge gaps regarding its sources and sinks. In this study, we present the first in situ observations of CO in the region, which we combined with incubation experiments in waters directly influenced by anthropogenic activities. The main objectives of this study were to i) elucidate the CO distribution and sea–air flux densities in the coastal lagoon system, ii) estimate the different CO sources and sinks in the lagoon system during summer, and iii) determine the potential impact of aquaculture activities on CO cycling in this region. 3.2 Materials and methods 3.2.1 Sample collection Discrete water samples for determining CO concentrations were collected using a Niskin bottle at approximately 1 m water depth from a small boat at 14 stations in the western part of the Ria Formosa Lagoon (Fig. 3.1). Samples were taken in the late afternoon (17:00~18:38 local time) Carbon monoxide in the Ria Formosa Lagoon 44 have been missed. Figure 3.2: Mean CO surface concentrations (± standard error estimate) at Stations 114 measured on 25 May 2021 (blue bars) and on 26 May 2021 (orange bars). The mean concentrations (i.e., the average from the two sampling days) of dissolved CO and water temperature, salinity, Chl-a, FDOM, pH, and nutrients in the Ria Formosa Lagoon are shown in Fig. 3.3 and listed in Table 3.1. Due to sea salt production facilities and aquaculture ponds, the study area’s salinity (mean salinity = 36.54) was higher than the seawater outside the lagoon. Chl-a concentrations (used here to estimate phytoplankton biomass) ranged from 0.02 to 0.50 μg L-1, with an average of 0.23 ± 0.16 μg L-1. In general, Chl-a concentrations decreased from the Ramalhete to the Faro–Olhão inlet (Aníbal et al., 2019), in correspondence to increasing distance from the nutrient-rich plumes from the WWTP. High Chl-a concentrations were recorded at Station 8, indicating higher phytoplankton biomass (0.50 μg L1). Carbon monoxide in the Ria Formosa Lagoon 45 Figure 3.3: Surface distributions of temperature (°C), salinity, chlorophyll a (μg L-1), pH, FDOM (QSU), CO concentrations (nmol L-1), nitrate (μmol L-1), and phosphate (μmol L-1) in the Ria Formosa Lagoon in May 2021. Carbon monoxide in the Ria Formosa Lagoon 46 Table 3.1: Mean environmental parameters in Ria Formosa Lagoon (mean = the arithmetical average of the data from 25 and 26 May 2021). Station COsurf (nmol L1) Chl-a (μg L-1) FDOM (QSU) Salinity pH T (°C) COatm (ppb) u10 (m s-1) NH4+ (μmol L-1 PO43- (μmol L-1) NO3- (μmol L-1) 1 1.62 0.40 12.19 36.13 8.05 21.73 — — 61.67 1.40 2.42 2 1.43 0.37 11.33 36.26 8.16 21.58 — — 31.38 1.78 1.59 3 0.61 0.37 9.51 36.58 8.18 21.63 — — ≤LOD 0.23 0.85 4 0.60 0.23 4.76 37.15 8.18 20.14 — — ≤LOD 0.24 1.03 5 0.40 0.20 2.26 36.69 8.17 18.02 — — ≤LOD 0.30 0.94 6 0.46 0.27 4.62 37.19 8.22 20.61 — — ≤LOD 0.20 2.36 7 0.59 0.38 5.13 37.17 8.21 20.78 — — ≤LOD 0.29 2.16 8 0.88 0.50 4.50 35.68 7.97 20.04 126 3.1 ≤LOD 0.30 0.35 9 0.51 0.04 0.16 36.41 8.12 15.57 98 3.3 ≤LOD 0.36 2.19 10 0.49 0.02 0.14 36.36 8.11 16.37 95 5.0 ≤LOD 0.50 3.31 11 0.65 0.05 0.32 36.38 8.12 15.85 93 4.8 ≤LOD ≤LOD 1.49 12 1.05 0.04 0.29 36.40 8.10 15.67 121 5.2 ≤LOD 0.80 3.19 13 0.34 0.14 2.34 36.62 8.18 19.28 — — ≤LOD 0.54 0.70 14 0.84 0.23 2.95 36.54 8.18 20.08 124 6.1 ≤LOD 0.54 1.00 To assess the spatial variability in the lagoon, we divided the sampling area into three regions according to FDOM vs. CO, Sal vs. T plots (Fig. 3.4): WWTP plume zone (Stations 1 to 7), Praia de Faro zone (Station 8), and Faro–Olhão inlet zone (Stations 9 to 14). The WWTP plume zone was characterized by high temperature, low salinity, high nutrient (NH4+ and PO43-) concentrations, and large amounts of FDOM. CO concentrations in this zone were strongly affected by environmental conditions with anthropogenic activities. Such effluents might be a source of FDOM, which could potentially enhance CO production (Zepp, 2003; Yang et al., 2011; Zhao et al., 2015). Contrarily, in the semi-diurnal tidal influenced Faro–Olhão inlet zone, a low-temperature, low-salinity, and low-biomass regime caused by “fresh” seawater entering the lagoon through the Faro–Olhão inlet could be observed. Consequently, close to this inlet, CO concentrations were comparatively low. Lastly, in the seagrass (Cymodecea and Zostera) region of the lagoon and in the Praia de Faro zone, where the highest phytoplankton biomass was observed, high CO concentrations were measured at Stations 8 and 14. This suggests a potential for direct biological CO production, which might be a significant source (Gros et al., 2009; McLeod et al., 2021). Besides, during the high tide, the water of the Faro channel caused Carbon monoxide in the Ria Formosa Lagoon 47 the WWTP thermal effluent plume to have a visible influence from the region of Montenegro through Esteiro Largo and to the West region through the Ramalhete (Aníbal et al., 2019). Figure 3.4: (A) mean salinity (filled light blue squares) vs mean temperature (filled orange dots) and (B) mean concentration of dissolved CO (filled blue dots) and mean FDOM (filled red squares) at all stations in the Ria Formosa Lagoon on 25 May (Day1) and 26 May 2021 (Day2). Indeed, a regression analysis revealed that CO concentrations were significantly correlated with FDOM (r = 0.416, p < 0.05, n = 28). Since no significant correlations were found between CO concentrations and pH, water temperature, salinity, and Chl-a, it appears that the distribution of dissolved CO in the Ria Formosa Lagoon was mainly driven by the photochemical production of CO via FDOM at the time of sampling. Nonetheless, a significant correlation was found between the CO and PO43concentrations (r = 0.860, p < 0.01, n = 14). 3.3.2 Sea–air flux densities of CO The waters of the Ria Formosa Lagoon were consistently supersaturated with CO at all stations, with saturation ratios ranging from 1.7 to 32.2 (mean ± SD: 7.7 ± 5.9; Fig. 3.5a). This indicates that the lagoon was a source of atmospheric CO at the time of sampling, which concurs with the observations from other coastal waters (e.g., Liss et al., 2014). Carbon monoxide in the Ria Formosa Lagoon 48 Figure 3.5: (a) Concentrations of dissolved CO (COsurf; filled orange dots) and CO saturation ratios (α; filled light blue squares) at all stations in the Ria Formosa Lagoon on 25 May (Day1) and 26 May 2021 (Day2) and (b) flux densities of CO (filled blue dots) and wind speeds (u10; filled red squares) at Stations 8-12 and 14 on 26 May 2021 (Day2). The mean flux density of CO was estimated to be 1.53 ± 0.92 μmol m-2 d-1 for the Ria Formosa Lagoon. The sea-to-air flux was lower than those obtained from the North Atlantic Ocean (2.2 ± 1.5 μmol m-2 d-1) and the Northern Hemisphere Oceans (1.9 ± 1.3 μmol m-2 d-1) (Stubbins et al., 2006) but roughly in the same range of flux densities reported by Zhang et al. (2019) with an average of 1.76 μmol m-2 d-1 and by Zhao et al. (2015) with an average of 1.78 μmol m-2 d1. As shown in Fig. 3.5b, the flux densities of CO exhibited considerable spatial variability. Much of the variability was caused by the wind speed, as the highest flux densities were computed for Station 14 (2.87 μmol m-2 d-1) and coincided with the highest wind speed (6.1 m s-1), whereas the lowest flux densities were computed for Station 9 (0.53 μmol m-2 d-1) and coincided with low wind speeds (3.3 m s-1). However, at Station 8, which had the lowest wind speeds (3.1 m s-1), the flux density of CO (1.09 μmol m-2 d-1) was high due to the higher concentration of dissolved CO (0.88 nmol L-1) and higher saturation ratio of CO (8.2). Therefore, the large variability in the CO flux densities might also have been caused by spatial and temporal variations in anthropogenic and biogenic sources of CO. 3.3.3 Microbial CO consumption To investigate the microbial consumption of dissolved CO, we determined the CO consumption Carbon monoxide in the Ria Formosa Lagoon 49 rate using the dark incubation method (Zafiriou et al., 2003; Zhang et al., 2019; Sugai et al., 2020). CO concentrations decreased sharply with time during incubation, following the first‐ order kinetics (y = 1.5084e-0.403x, r2 = 0.9992) (Fig. 3.6). This is consistent with previous results, which reported an exponential decrease in CO concentrations with incubation time at locations such as the Delaware Bay, Northwest Atlantic, and East China Sea (Xie et al., 2005; Yang et al., 2010). The resulting CO consumption rate constant (Kbio) was 0.40 h-1, comparable to published CO consumption rate constants ranging from 0.02 h-1 to 1.11 h-1 (Xie et al., 2005). Figure 3.6: CO concentrations (± standard error estimate) versus incubation time during the dark incubation experiment. The relatively rapid turnover time (defined as 1/Kbio) of 2.5 h suggests that microbial CO consumption was the dominant removal pathway of dissolved CO in the Ria Formosa Lagoon system in May 2021, significantly modulating atmospheric CO emissions. Multiplying Kbio by the mean surface seawater CO concentration and the sampling depth (1 m) yields a rough estimate of the depth-integrated microbial consumption rate (Cbact) of 7.25 μmol m-2 d-1 in the surface layer for the study area. 3.3.4 Comparison of CO sources and sinks for Ria Formosa Lagoon in May 2021 The budget of CO for the surface layer (i.e., the sampling depth of 1 m) is driven by the balance of various sources and sinks, such as microbial consumption (Cbact), air–sea gas exchange (Fsea– air), photochemical production (Pphoto), dark production (Pdark) and production by phytoplankton (Pphyto): Carbon monoxide in the Ria Formosa Lagoon 50 Cbact + Fsea–air ≅ Pphoto + Pdark+ Pphyto (11) Cbact and Fsea–air have been estimated (Sections 4.2 and 4.3). Because the Ria Formosa Lagoon is usually well-mixed, we assumed that vertical mixing and advection were negligible at the time of our study. Pphyto can be roughly estimated with the CO production rates given in Gros et al. (2009) and the Chl-a concentrations measured here. It is known that diatoms can contribute an overall average of 53 % of the total phytoplankton biomass in the Ria Formosa Lagoon system (Pereira et al., 2007). Hence, we estimated Pphyto as: 6?@ABC K= K6 ( cd ) ×KChl−QK·Kh·SD7EF! (12) where h is the thickness of the sampling depth (set to 1 m), P(cd) is the maximum CO production rate given in Gros et al. (2009) (8.7×10-4 μg CO (μg chlorophyll)-1 h-1 for the diatom Chaetoceros debilis), Chl-a is the mean Chl-a concentration measured during our study in May 2021 (0.23 μg L-1) and the average daylight hours (tlight) in May is 14.20 h d-1 in this area. The resulting Pphyto was 5.38×10-5 μmol m-2 d-1. Pdark was roughly estimated according to the parameterization of Zhang et al. (2008) and Conte et al. (2019), in which Pdark is estimated from CDOM (represented by its absorption at 350 nm), temperature (T in °C), pH, and salinity as follows: 6G5'H =KQIGJ6(350)·T81 ·Kh·SG5'H (13) 34 ( T81 ×10: ) = −'%)(* ++0.494·UV −0.0257·B +41.9 (14) where the T CO is 1.78×10-2 nmol m L-1 h-1 (computed with the mean water temperature of 292.25 K, the mean pH of 8.14, and the mean salinity of 36.54 at the time of sampling), the Station 7 mean acdom (350) is 0.21 m-1 and the average dark hours (tdark) in this area were 9.80 h d-1 in May. The resulting mean Pdark is 3.66×10-5 µmol m-2 d-1. Finally, the CO photoproduction rate, Pphoto, was calculated as Cbact + Fsea–air - Pdark - Pphyto. The estimates of the production and consumption rates are given in Table 3.2. It is apparent that both Pphyto and Pdark were negligible compared to Pphoto (Table 3.2). Carbon monoxide in the Ria Formosa Lagoon 51 Table 3.2: Major sources and sinks (in µmol m-2 d-1) of CO in seawater in the Ria Formosa Lagoon in May 2021. Sources Photoproduction 8.78 Dark production 3.66×10-5 Production by phytoplankton 5.38×10-5 Sinks Sea-to-air flux density 1.53 Microbial consumption 7.25 The main sink of CO in Ria Formosa in May 2021 was the microbial consumption of CO (Cbact), which accounted for about 83% of the total CO loss terms (Cbact + Fsea–air). Our results align with the estimate of Zhang and Xie (2012), who reported that the microbial consumption of CO in the St. Lawrence Estuary accounted for 86–92% of the CO sink. Overall, the CO emissions to the atmosphere resulted from CO photoproduction, which, in turn, was modulated by microbial CO consumption. This supports the perception that microbial CO consumption is the dominant sink of oceanic CO, recycling approximately 80–90% of the CO in the open and coastal oceans before it is released into the atmosphere (e.g., Moran and Miller, 2007; Zafiriou et al., 2003). Furthermore, our estimates suggest that the CO budget in the Rio Formosa Lagoon system is balanced, with sources and sinks having roughly similar magnitudes. 3.3.5 CO production from aquaculture effluent The results of the 48h-incubation experiments with waters from the effluents of the aquaculture facility Estação Piloto de Piscicultura em Olhão (EPPO) are shown in Fig. 3.7. The samples exposed to light showed a linear increase in CO concentration over 48 h. The CO concentrations in the samples kept in the dark showed a steep decrease until 24 h and then continued to decrease only slightly until 48 h. This indicates that microbial CO consumption, probably restricted by CO dark production, led to a lower consumption state at the end of the dark incubation. Carbon monoxide in the Ria Formosa Lagoon 52 Figure 3.7: Temporal variability of dissolved concentrations of CO in dark-group (solid line) and lightgroup (dashed line) incubations with water from the effluents of the aquaculture facility Estação Piloto de Piscicultura em Olhão (EPPO). The absorption of CDOM during the light incubation experiments is shown in Fig. 3.8. Absorption by CDOM as a function of wavelength a(λ) (m-1) which is commonly modeled with an exponentially decreasing function (Shifrin, 1988) of the form: Q ( W ) = X×'#KL (15) where λ is a reference wavelength, A is amplitude, and S (nm-1) is the spectral slope parameter describing the relative steepness of the spectrum, understood here as a proxy for changes in the composition of CDOM (Carder et al., 1989). Figure 3.8: The CDOM absorption spectra for different groups (Initial 0, Light-24h-A, Light-24h-B, Light-48h-A, Light-48h-B) of the aquaculture photo-incubations (The curves corresponding to different groups are their fitted curves with the same color). The ratio of the spectral slope (SR) of two different spectral slopes, S280–295 and S350–400, were used to indicate the magnitude of the relative molecular weight of CDOM (Helms et al. 2008) and to identify the origin of the CDOM: the smaller the SR, the higher the proportion of Carbon monoxide in the Ria Formosa Lagoon 53 terrestrial CDOM, and vice versa (John et al., 2009; Helms et al., 2013; Zhu et al., 2017). Besides, Xu et al. (2023) pointed that there was a significant negative correlation between the apparent efficiency of CO photoproduction and SR. Ossola et al. (2022) demonstrated for the first time a precise mechanism for CO production: the aromatic methoxy groups can produce CO. Summarizing, using SR, the CO photoproduction capacity can be more accurately evaluated. The ranges of variation were 0.013 to 0.02 nm-1 and 0.015–0.036 nm-1 for S280–295 and S350–400, respectively, at different incubation times (Table 3.3). Further, we obtained SR mean values for different incubation times (SR-0h: 0.56; SR-24h: 0.52; SR-48h: 1.07). Variations indicate changes in CDOM degradation during light incubations, potentially due to alterations in the CDOM composition and the ratio of fulvic acids to humic acids (Carder et al., 1989). A lower SR indicates a predominant terrestrial input with a typically higher proportion of humic acid and a higher photochemical reactivity potential. CO is mainly produced by the photodegradation of specific compounds (e.g., humic acid), which are efficient CO producers (Ren et al., 2014). In the initial phase of the experiment (0–24 h), specific aromatic compounds were destroyed, and their concentrations decreased rapidly under the consumption of photodegradation; as the experimental time grew (24–48 h), the concentration of the second type of precursors without humic acid gradually dominated, making the CO photoproduction rate decrease continuously. Notably, the microbial degradation persisted, making it possible to show that the CO photoproduction rate/ net production rate decreased with increasing incubation time. Table 3.3: List of CDOM absorption spectra (S280-295 and S 350-400) and S R for different groups (Initial 0, Light-24h-A, Light-24h-B, Light-48h-A, Light-48h-B) of the aquaculture photo-incubations. S280-295 S 350-400 S R Initial 0 0.02 0.036 0.56 Light 24h-A 0.013 0.031 0.42 Light 24h-B 0.017 0.027 0.63 Light 48h-A 0.019 0.019 1 Light 48h-B 0.017 0.015 1.13 The resulting CO photoproduction rate and net production rate of 0.013 – 0.015 and 0.0038 – 0.0039 nmol L-1 h-1, respectively (Fig. 3.7) is the first CO photoproduction rate/ net production rate for an aquaculture effluent. It is extremely low compared to published CO photoproduction Carbon monoxide in the Ria Formosa Lagoon 60 Domingues, R. B., Guerra, C. C., Barbosa, A. B., and Galvão, H. M.: Will nutrient and light limitation prevent eutrophication in an anthropogenically-impacted coastal lagoon? Continental Shelf Research, 141, 11-25, https://doi.org/10.1016/j.csr.2017.05.003, 2017. Forster, P., Storelvmo, T., Armour, K., Collins, W., Dufresne, J.-L., Frame, D., Lunt, D.J., Mauritsen, T., Palmer, M.D., Watanabe, M., Wild, M., and Zhang, H.: The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity, in: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J.B.R., Maycock, T.K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou. B., Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 923–1054, https://doi.org/10.1017/9781009157896.009, 2021. Greening, C., and Grinter, R.: Microbial oxidation of atmospheric trace gases, Nature Reviews Microbiology, 20(9), 513-528, https://doi.org/10.1038/s41579-022-00724-x, 2022. Gros, V., Peeken, I., Bluhm, K., Zöllner, E., Sarda-Esteve, R., and Bonsang, B.: Carbon monoxide emissions by phytoplankton: evidence from laboratory experiments, Environmental chemistry, 6(5), 369379, https://doi.org/10.1071/EN09020, 2009. Helms, J. R., Stubbins, A., Perdue, E. M., Green, N. W., Chen, H., and Mopper, K.: Photochemical bleaching of oceanic dissolved organic matter and its effect on absorption spectral slope and fluorescence, Marine Chemistry, 155, 81–91, https://doi.org/10.1016/j. marchem.2013.05.015, 2013. Helms, J. R., Stubbins, A., Ritchie, J. D., Minor, E. C., Kieber, D. J., and Mopper, K.: Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter, Limnology and oceanography, 53(3), 955-969, https://doi.org/10.4319/lo.2008.53.3.0955, 2008. Højerslev, N.K., Aas, E.: Spectral light absorption by yellow substance in the Kattegat–Skagerrak area, Oceanologia 43, 39–60, 2001. IPCC, Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S. L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M. I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J. B. R., Maycock, T. K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou, B. (Eds.): Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK and New York, NY, USA, 2391 pp., 10.1017/9781009157896, 2021. Jones, R. D., and Amador, J. A.: Methane and carbon monoxide production, oxidation, and turnover times in the Caribbean Sea as influenced by the Orinoco River, Journal of Geophysical Research: Oceans, 98(C2), 2353-2359, https://doi.org/10.1029/92JC02769, 1993. John, R. H., Stubbins, A., Ritchie, J. D., Minor, E. C., Kieber, D. J., and Mopper, K.: Absorption spectral Carbon monoxide in the Ria Formosa Lagoon 61 slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnology & Oceanography, 54(3), 1023. https://doi.org/10.4319/lo.2009.54.3.1023, 2009. Kowalczuk, P., Olszewski, J., Darecki, M., Kaczmarek, S.: Empirical relationships between Coloured Dissolved Organic Matter (CDOM) absorption and apparent optical properties in Baltic Sea waters, Int. J. Remote Sens, https://doi.org/10.1080/01431160410001720270, 2003. Law, C. S., Sjoberg, T. N., and Ling, R. D.: Atmospheric emission and cycling of carbon monoxide in the Scheldt Estuary, Biogeochemistry, 59(1), 69-94, https://doi.org/10.1023/A:1015592128779, 2002. Li, J. L., Zhai, X., and Du, L.: Effect of nitrate on the photochemical production of carbonyl sulfide from surface seawater, Geophysical Research Letters, 49(13), e2021GL097051, https://doi.org/10.1029/2021GL097051, 2022. Liss, P. S., Marandino, C. A., Dahl, E. E., Helmig, D., Hintsa, E. J., Hughes, C., ... and Williams, J.: Short-lived trace gases in the surface ocean and the atmosphere, In Ocean-Atmosphere Interactions of Gases and Particles (pp. 1-54). Springer, Berlin, Heidelberg, 2014. Loiselle, S. A., Bracchini, L., Dattilo, A. M., Ricci, M., Tognazzi, A., Cózar, A., and Rossi, C.: The optical characterization of chromophoric dissolved organic matter using wavelength distribution of absorption spectral slopes, Limnology and oceanography, 54(2), 590-597, https://doi.org/10.4319/lo.2009.54.2.0590, 2009. McLeod, A. R., Brand, T., Campbell, C. N., Davidson, K., and Hatton, A. D.: Ultraviolet Radiation Drives Emission of Climate‐Relevant Gases from Marine Phytoplankton, Journal of Geophysical Research: Biogeosciences, 126(9), e2021JG006345, https://doi.org/10.1029/2021JG006345, 2021. Moran, M. A., and Miller, W. L.: Resourceful heterotrophs make the most of light in the coastal ocean, Nature Reviews Microbiology, 5(10), 792-800, https://doi.org/10.1038/nrmicro1746, 2007. Nakagawa, F., Tsunogai, U., Gamo, T., and Yoshida, N.: Stable isotopic compositions and fractionations of carbon monoxide at coastal and open ocean stations in the Pacific, Journal of Geophysical Research: Oceans, 109(C6), https://doi.org/10.1029/2001JC001108, 2004. Nelson, N.B., Siegel, D.A., Michaels, A.F.: Seasonal dynamics of colored dissolved material in the Sargasso Sea, Deep-Sea Res. 45, 931–957, https://doi.org/10.1016/S0967-0637(97)00106-4, 1998. Newton, A., and Mudge, S. M.: Temperature and salinity regimes in a shallow, mesotidal lagoon, the Ria Formosa, Portugal. Estuarine, Coastal and Shelf Science, 57(1-2), 73-85, https://doi.org/10.1016/S02727714(02)00332-3, 2003. Newton, A., and Mudge, S. M.: Lagoon-sea exchanges, nutrient dynamics and water quality management of the Ria Formosa (Portugal). Estuarine, coastal and shelf science, 62(3), 405-414, Carbon monoxide in the Ria Formosa Lagoon 62 https://doi.org/10.1016/j.ecss.2004.09.005,2005. Ossola, R., Gruseck, R., Houska, J., Manfrin, A., Vallieres, M., and McNeill, K.: Photochemical production of carbon monoxide from dissolved organic matter: Role of lignin methoxyarene functional groups. Environmental Science and Technology, 56(18), 13449–13460. https://doi.org/10.1021/acs.est.2c03762, 2022. Panofsky, H. A., and Dutton, J. A.: Amospheric turbulence: models and methods for engineering applications. A Wiley-Interscience publication, John Wiley and Sons, New York, 1984. Park, K., and Rhee, T. S.: Oceanic source strength of carbon monoxide on the basis of basin-wide observations in the Atlantic, Environmental Science: Processes and Impacts, 18(1), 104-114, https://doi.org/10.1039/c5em00546a, 2016. Petron, G., Crotwell, A.M., Crotwell, M.J., Dlugokencky, E., Madronich, E., Moglia, E., Neff, D., Thoning, F., Wolter, S. and Mund, J.W.: Atmospheric Carbon Monoxide Dry Air Mole Fractions from the NOAA GML Carbon Cycle Cooperative Global Air Sampling Network, 1988-2021, Version: 202207-28, https://doi.org/10.15138/33bv-s284, 2022. Pos, W. H., Riemer, D. D., and Zika, R.: Carbonyl sulfide (OCS) and carbon monoxide (CO) in natural waters: Evidence of a coupled production pathway, Mar. Chem., 62, 89-101, https://doi.org/10.1016/S0304-4203(98)00025-5, 1998. Pereira, M.G., Icely, J.D., Mudge, S.M., Newton, A. and Rodrigues, R.: Temporal and spatial variation of phytoplankton pigments of the Ria Formosa Lagoon, Southern Portugal, Environmental Forensics 8(3): 205-220, https://doi.org/10.1080/15275920701506151, 2007. Raymond, P. A., and Cole, J. J.: Gas exchange in rivers and estuaries: Choosing a gas transfer velocity, Estuaries, 24(2), 312-317, https://doi.org/10.2307/1352954, 2001. Ren, C., Yang, G., and Lu, X.: Autumn photoproduction of carbon monoxide in Jiaozhou Bay, China, Journal of Ocean University of China, 13(3), 428-436, https://doi.org/10.1007/s11802-014-2225-1, 2014. Schmidt, C. E., and Heikes, B. G.: Aqueous carbon monoxide cycling in a fjord-like estuary, Estuaries and coasts, 37(3), 751-762, https://doi.org/10.1007/s12237-013-9722-0, 2014. Shifrin, K. S.: Physical Optics of Ocean Water, transl. D. Oliver, AIP Transl, Ser, 1988. Stedmon, C.A., and Markager, S.: The optics of chromophoric dissolved organic matter (CDOM) in the Greenland Sea: analgorithm for differentiation between marine and terrestrially derived organic matter, Limnol. Oceanogr. 46, 2087–2093, https://doi.org/10.4319/lo.2001.46.8.2087, 2001. Stedmon, C.A., Markager, S., Kaas, H.: Optical properties and signatures of chromophoric dissolved organic matter (CDOM) in Danish coastal waters, Estuar. Coast. Shelf Sci.51, 267–278, Carbon monoxide in the Ria Formosa Lagoon 63 https://doi.org/10.1006/ecss.2000.0645, 2000. Stubbins, A., Uher, G., Kitidis, V., Law, C.S., Upstill-Goddard, R.C., Woodward, E.M.S.: The open-ocean source of atmospheric carbon monoxide, Deep Sea Res. Part II 53, 1685–1694, https://doi.org/10.1016/j.dsr2.2006.05.010, 2006. Stubbins, A., Uher, G., Law, C. S., Mopper, K., Robinson, C., and Upstill-Goddard, R. C.: Open-ocean carbon monoxide photoproduction, Deep Sea Research Part II: Topical Studies in Oceanography, 53(1416), 1695-1705, https://doi.org/10.1016/j.dsr2.2006.05.011, 2006. Sugai, Y., Tsuchiya, K., Shimode, S., and Toda, T.: Photochemical Production and Biological Consumption of CO in the SML of Temperate Coastal Waters and Their Implications for Air‐Sea CO Exchange, Journal of Geophysical Research: Oceans, 125(4), e2019JC015505, https://doi.org/10.1029/2019JC015505, 2020. Thompson, A. M.: The oxidizing capacity of the Earth’s atmosphere: Probable past and future changes, Science, 256(5060), 1157-1165, DOI: 10.1126/science.256.5060.1157, 1992. Tett, P., Gilpin, L., Svendsen, H., Erlandsson, C. P., Larsson, U., Kratzer, S., ... and Scory, S.: Eutrophication and some European waters of restricted exchange, Continental shelf research, 23(17-19), 1635-1671, https://doi.org/10.1016/j.csr.2003.06.013, 2003. Val ent ine , R . L. , a nd Z epp , R. G. : Fo rma tio n of ca rbo n mo nox ide fr om t he p hot ode gra dat ion of ter restrial dissolved organic carbon in natural waters, Environmental science and technology, 27(2), 409-412, 1993. Wang, W. L., Peng, T., Lu, X. L., and Zhao, B. Z.: Diurnal, seasonal, and spatial variations and flux of carbon monoxide in Jiaozhou Bay, China, Marine Chemistry, 191, 1-8, https://doi.org/10.1016/j.marchem.2017.01.004, 2017. Weiss, R. F., and Price, B. A.: Nitrous oxide solubility in water and seawater, Marine chemistry, 8(4), 347-359, https://doi.org/10.1016/0304-4203(80)90024-9, 1980. Wiesenburg, D. A., and Guinasso Jr, N. L.: Equilibrium solubilities of methane, carbon monoxide, and hydrogen in water and sea water, Journal of chemical and engineering data, 24(4), 356-360, 1979. Xie, H., Andrews, S. S., Martin, W. R., Miller, J., Ziolkowski, L., Taylor, C. D., and Zafiriou, O. C.: Val ida ted met hod s fo r sa mpl ing a nd h ead spa ce a nal ysi s of car bon mon oxi de i n se awat er, M ari ne Chemistry, 77(2-3), 93-108, https://doi.org/10.1016/S0304-4203(01)00065-2, 2002. Xie, H., Zafiriou, O. C., Umile, T. P., and Kieber, D. J.: Biological consumption of carbon monoxide in Delaware Bay, NW Atlantic and Beaufort Sea, Marine Ecology Progress Series, 290, 1–14. https://doi.org/10.3354/meps290001, 2005. Xie, H., and Zafiriou, O. C.: Evidence for significant photochemical production of carbon monoxide by particles in coastal and oligotrophic marine waters, Geophysical Research Letters, 36(23), Carbon monoxide in the Ria Formosa Lagoon 64 https://doi.org/10.1029/2009GL041158, 2009. Xu, G. B., Xu, F., Ji, X., Zhang, J., Yan, S. B., Mao, S. H., and, Yang, G. P.: Carbon monoxide cycling in the Eastern Indian Ocean. Journal of Geophysical Research: Oceans, e2022JC019411, https://doi.org/10.1029/2022JC019411, 2023. Yang, G. P., Wang, W. L., Lu, X. L., and Ren, C. R.: Distribution, flux and biological consumption of carbon monoxide in the Southern Yellow Sea and the East China Sea, Marine Chemistry, 122(1‐4), 74– 82. https://doi.org/10.1016/j.marchem.2010.08.001, 2010. Yang, G. P., Ren, C. Y., Lu, X. L., Liu, C. Y., and Ding, H. B.: Distribution, flux, and photoproduction of carbon monoxide in the East China Sea and Yellow Sea in spring, Journal of Geophysical Research: Oceans, 116(C2), https://doi.org/10.1029/2010JC006300, 2011. Zafiriou, O. C., Andrews, S. S., and Wang, W.: Concordant estimates of oceanic carbon monoxide source and sink processes in the Pacific yield a balanced global “blue‐water” CO budget, Global Biogeochemical Cycles, 17(1), https://doi.org/10.1029/2001GB001638, 2003. Zafiriou, O.C., Xie, H., Nelson, N.B., Najjar, R.G., and Wang, W.: Diel carbon monoxide cy cling in the upper Sargasso Sea near Bermuda at the onset of spring and in midsum-mer, Limnol. Oceanogr. 53, 835– 850, https://doi.org/10.4319/lo.2008.53.2.0835, 2008. Zepp, R. G.: Solar UVR and aquatic carbon, nitrogen, sulfur and metals cycles, UV effects in aquatic organisms and ecosystems, 137-183, 2003. Zhang, Y., Xie, H., Fichot, C. G., and Chen, G.: Dark production of carbon monoxide (CO) from dissolved organic matter in the St. Lawrence estuarine system: Implication for the global coastal and blue water CO budgets, Journal of Geophysical Research: Oceans, 113(C12), https://doi.org/10.1029/2008JC004811, 2008. Zhang, Y., and Xie, H.: The sources and sinks of carbon monoxide in the St. Lawrence estuarine system, Deep Sea Research Part II: Topical Studies in Oceanography, 81, 114-123, https://doi.org/10.1016/j.dsr2.2011.09.003, 2012. Zhang, J., Wang, J., Zhuang, G. C., and Yang, G. P.: Carbon monoxide cycle in the Bohai Sea and the Yellow Sea: Spatial variability, sea‐air exchange, and biological consumption in autumn, Journal of Geophysical Research: Oceans, 124(6), 4248-4257, https://doi.org/10.1029/2018JC014864, 2019. Zhao, B. Z., Yang, G. P., Xie, H., Lu, X. L., and Yang, J.: Distribution, flux and photoproduction of carbon monoxide in the Bohai and Yellow Seas, Marine Chemistry, 168, 104-113, https://doi.org/10.1016/j.marchem.2014.11.006, 2015. Zheng, B., Chevallier, F., Yin, Y., Ciais, P., Fortems-Cheiney, A., Deeter, M. N., Parker, R. J., Wang, Y., Worden, H. M., and Zhao, Y.: Global atmospheric carbon monoxide budget 2000–2017 inferred from Carbon monoxide in the Ria Formosa Lagoon 65 multi-species atmospheric inversions, Earth Syst. Sci. Data, 11, 1411-1436, 10.5194/essd-11-1411-2019, https://doi.org/10.5194/essd-11-1411-2019, 2019. Zhu, W. Z., Yang, G. P., and Zhang, H. H.: Photochemical behavior of dissolved and colloidal organic matter in estuarine and oceanic waters. Science of the Total Environment, 607–608, 214–224. https://doi.org/10.1016/j.scitotenv.2017.06.163, 2017. Zuo, Y., and Jones, R. D.: Formation of carbon monoxide by photolysis of dissolved marine organic material and its significance in the carbon cycling of the oceans, Naturwissenschaften, 82(10), 472-474, https://doi.org/10.1007/BF01131598, 1995. Carbon monoxide in the Ria Formosa Lagoon 66 Carbon monoxide in the SW Baltic Sea 67 4 A seasonal study of dissolved carbon monoxide at the Boknis Eck Time Series Station in Eckernförde Bay (southwestern Baltic Sea) Manuscript in preparation for BG: Li, G., Arévalo-Martínez, D. L., Hepach, H., Engel, A., and Bange, H. W.: A seasonal study of dissolved carbon monoxide at the Boknis Eck Time Series Station in Eckernförde Bay (southwestern Baltic Sea). Abstract. Carbon monoxide (CO) is a crucial component of the global carbon cycle and is recognized as an indirect greenhouse gas. While the ocean is a perennial source of CO, uncertainties persist regarding the factors governing its sources and sinks, particularly in coastal, low-oxygen systems. We conducted the first systematic seasonal investigation into dissolved CO at the Boknis Eck Time Series Station in Eckernförde Bay (southwestern Baltic Sea). Dissolved CO concentrations ranged between 0.20 and 1.89 nmol L-1 with an average of 0.63 ± 0.38 nmol L-1 which were higher in spring (0.70 nmol L-1) than in summer (0.42 nmol L-1). CO saturation ratios varied between 1.9 to 4.5 with higher values in spring-summer, evidencing the role of this area as a CO source to the atmosphere, with an estimated average sea-to-air flux density of 3.66 μmol m-2 d-1. We estimated the Baltic Sea to contribute approximately 0.15% (6.06 Gg CO-C yr-1) of the global oceanic CO emission. In surface waters, CO was strongly correlated with diatoms, whereas on the bottom layer high CO values were associated with the occurrence of hypoxia or anoxia. We contend that the accumulation of CO in bottom waters was attributed to its release from anoxic sediment and its in-situ production within the overlying Carbon monoxide in the SW Baltic Sea 68 water column. Using optical–photochemical modeling based on measured CDOM absorption coefficient and spectral CO apparent quantum yields, we estimated a CO photoproduction of 20.05 Gg CO-C yr-1 in the area, constituting about 0.10% of the global budget. We provided valuable insights into the environmental factors driving CO distribution in the Baltic Sea's water column, illuminating its seasonal dynamics and interactions with other variables. 4.1 Introduction Carbon monoxide (CO) plays an important role in the oxidizing capacity of the Earth’s atmosphere: in the troposphere, CO reacts fast with hydroxyl radicals (OH) and, when nitrogen oxides (NOx) are present, it leads to the formation of ozone, which is a pollutant and a greenhouse gas (Thompson, 1992; Stubbins et al., 2006; Forster et al., 2021). Moreover, due to its reactivity with OH, CO indirectly influences the lifetime of atmospheric methane, a strong greenhouse gas (Evans and Puckrin, 1995; Taylor et al., 1996; Canadell et al., 2021). Emission estimates indicate that the marine source of CO ranges between 4 Tg CO–C yr-1 and 20 Tg CO– C yr-1 (Conte et al., 2019; Park and Rhee, 2016; Zheng et al., 2019), with coastal oceans contributing about 15-25 % to the overall oceanic emissions of CO (Park and Rhee, 2016). The major source of CO in the ocean is its photochemical production via the oxidation of colored dissolved organic matter (CDOM) under ultraviolet (UV) radiation in the upper layer of the open and coastal oceans (Xie and Zafiriou, 2009; Kitidis et al., 2011). Moreover, CO is also produced, albeit to a minor extent, by phytoplankton and by the so-called dark production (Loewus and Delwiche 1963; Seiler and Schmidt 1974; Gros et al., 2009; Zhang and Xie, 2012; McLeod et al., 2021). Once produced, CO is rapidly consumed by microbial uptake which accounts for about 86% of the CO sink in the marine environment (Zafiriou et al., 2003; Zhang et al., 2008; Greening and Grinter, 2022). A minor loss process is its release into the atmosphere (Zafiriou et al., 2003; Zhang et al., 2008). The Baltic Sea is a semi-enclosed marginal sea adjacent to the North Atlantic Ocean. It is characterized by limited horizontal and vertical water exchange due to the presence of shallow Carbon monoxide in the SW Baltic Sea 69 sills and salinity-driven stratification. Coastal and open ocean regions of the Baltic Sea are strongly affected by ongoing environmental perturbations such as warming, deoxygenation, and acidification, which are in turn influenced by different anthropogenic activities (e.g. agriculture, aquaculture, fisheries, industrial pollution, sewage discharge; Reckermann et al. (2022)). This study took place at the Boknis Eck Time Series Station (BE; www.bokniseck.de), which is located in the Eckernförde Bay (southwestern Baltic Sea), and is one of the oldest continuously operated coastal time-series stations worldwide (Lennartz et al., 2014). The hydrographic properties in the area are dominated by the balance between the inflow of saline waters from the North Sea and “fresher” waters from the Baltic proper, and there is a clear seasonality in biogeochemical properties which is driven by the development of seasonal stratification. Hence, BE is ideal for studying the biogeochemical dynamics of coastal ecosystems under the combined influence of natural and anthropogenic stressors, as well as the temporal variability of biogeochemical processes sensitive to changes in environmental oxygen concentrations. Here we report the first investigation of dissolved CO distribution and air-sea fluxes in the Baltic Sea. To this end, we measured dissolved and atmospheric CO on a monthly basis in 2020 at BE. The overarching goals of this study were to (i) decipher the seasonal variability of dissolved CO at BE, and (ii) identify the major processes affecting the marine cycling of CO in the area. 4.2 Study site description The sampling site at BE (54° 31.2′ N, 10° 02.5′ E) has a water depth of approximately 28 m (Fig. 4.1), and its water circulation is dominated by the inflow of waters from the North Sea through the Kattegat and the Great Belt. Thus, we consider this system to be representative of the biogeochemical setting of the southwestern Baltic Sea. Seasonal stratification, with warm, low salinity surface water overlaying colder and more saline deeper water masses, is caused by steep density gradients and is usually found from mid-March Carbon monoxide in the SW Baltic Sea 76 Figure 4.2: Monthly distributions of hydrological environmental parameters: temperature (A), salinity (B), dissolved oxygen (C), Chl-a (D), Phosphate (E), Nitrate (F), Nitrite (G), and Silicate (H) at BE in 2020. Please note that the blank areas are due to data gaps caused by cancellations (COVID-19) of the research cruises. Black dots (A) indicate monthly measurements of Secchi depth. 4.4.2 Variations of atmospheric CO The monthly atmospheric CO mixing ratios (xCOatm) at BE ranged between 158.43 and 230.54 ppb (200.71 ± 26.11 ppb), with most of the variability being explained by the seasonal contrast Carbon monoxide in the SW Baltic Sea 77 (Table 4.1). We observed the highest xCOatm in March and the lowest in July (Fig. 4.3). According to the 72h backward trajectories (Fig. S4.2A), air masses from Western Europe had been transported to BE. The higher concentrations of CO in this area were, therefore, strongly influenced by continental sources in March. In July (Fig. S4.2B), by contrast, an air mass from Iceland was influencing BE, with the lowest xCOatm. Figure 4.3: Monthly atmospheric CO mole fractions (COatm) at BE versus NOAA stations at HPB (Hohenpeissenberg, Germany), MHD (Mace Head, Ireland), and ICE (Storhofdi, Iceland) in 2020. Table 4.1: Monthly mean of surface layer dissolved CO concentrations (COsurf), wind speed (u10), atmospheric CO mixing ratios (COatm), CO saturation ratios (COsat), and CO flux density (Flux) at BE in 2020. COsurf nmol L-1 u10 m S-1 COatm ppb COsat - Flux µmol m-2 d-1 Jan 0.92 10 225.14 3.4 6.59 Mar 0.71 6.5 230.54 2.5 1.27 Apr 0.61 6 229.06 2.3 1.00 May 0.44 11 197.59 2.0 3.82 Jun 0.42 9 205.25 1.9 1.92 Jul 0.69 12 158.43 4.2 14.27 Aug 0.84 6 185.79 4.5 2.31 Sep 0.50 5 167.78 3.0 0.60 Oct 0.65 5.5 206.84 2.9 1.14 Despite the good agreement on seasonal variability trends with land-based time-series measurements by NOAA stations at HPB (Hohenpeissenberg, Germany) (Lat: 47.8011 °N; Lon: 11.0245 °E), MHD (Mace Head, Ireland) (Lat: 53.326 °N; Lon: 9.899 °W), and ICE (Storhofdi, Iceland) (Lat: 63.3998 °N; Lon: 20.2884 °W) (Petron et al., 2022), our values were on average Carbon monoxide in the SW Baltic Sea 78 72% higher than in all stations. This difference could be explained by the influence of ship traffic and other land-derived anthropogenic sources near BE (Holloway et al., 2000; Hoesly et al., 2018; Lopez, 2003; Lübbecke et al., 2019; Zheng et al., 2019). 4.4.3 Carbon monoxide surface saturation and Flux density The ocean’s surface is ubiquitously supersaturated with CO and its concentrations in surface waters have a clear seasonal pattern with higher values in spring-summer (e.g. Conte et al. 2019). In line with this, we observed a seasonal cycle of CO in surface waters at BE, with the highest saturation ratios in July–August (Fig. 4.4). The waters of the BE were supersaturated with CO all year and throughout the water column with saturation ratios ranging from 1.9 to 4.5 (mean ± SD: 3.0 ± 0.9) (Fig. 4.4). This indicates that the BE was a net source of atmospheric CO throughout the sampling period, which is in line with the observations from other coastal waters (see e.g. Day and Faloona 2009; Liss et al., 2014). Figure 4.4: Monthly surface layer concentrations of dissolved CO (filled purple squares), CO saturation ratio (α; filled green stars), wind speeds (u10; filled light blue dots), and flux densities of CO (filled gradual pink bars) at BE Series Station in 2020. The computed mean flux density (Fsea–air) of CO at BE was 3.66 μmol m-2 d-1 (Table 4.2), which is within the range of values (0.21 – 13.78 μmol m-2 d-1) reported in coastal areas and estuarine waters such as the Sargasso Sea, the California upwelling area (USA), and the Sagami Bay Carbon monoxide in the SW Baltic Sea 79 (Honshu, Japan) (Zafiriou et al., 2008; Day and Faloona 2009; Sugai et al., 2020). Contrarily, our value was higher than the flux densities reported by Zhang et al. (2019) with an average of 1.76 μmol m-2 d-1, and by Zhao et al. (2015) with an average of 1.78 μmol m-2 d-1. Table 4.2: Major source and sinks (in μmol m-2 d-1) of CO in seawater at BE in 2020. Source Photoproduction 12.11 Sinks Sea-to-air flux density 3.66 Microbial consumption 8.45 As both seawater and atmospheric CO showed seasonal variation, it is not surprising that the flux densities of CO varied seasonally (Fig. 4.4). Much of the variability was caused by the wind speed, as the highest flux densities were computed for July (14.27 μmol m-2 d-1) and coincided with highest wind speed (12.0 m s-1). Likewise, the lowest flux densities were computed in September (0.60 μmol m-2 d-1) coinciding with the lowest wind speeds (4.0 m s-1). In August which had wind speeds of 6.0 m s-1, the flux density of CO (2.31 μmol m-2 d-1) was high due to the higher saturation ratio of CO (4.5). We estimated the whole Baltic Sea CO emission was 6.06 Gg CO-C yr-1 on a yearly basis, based on its area of 3.77 × 105 km2, accounting for approximately 0.15% of the global oceanic CO emission of 4.0 Tg CO-C yr-1 (Conte et al., 2019). 4.4.4 Carbon monoxide concentration in the water column CO concentrations at BE showed a strong seasonal variability (Fig. 4.5E) and ranged from 0.20 nmol L-1 within the mixed-layer depth (10–15 m) in May and June to 1.89 nmol L-1 in 25 m water depth in September (Table 4.1). The overall mean dissolved CO concentration in the water column was 0.63 ± 0.38 nmol L-1 and higher concentration (0.70 nmol L-1) in spring than in summer (0.42 nmol L-1). The CO concentrations reported here are within the lower range of CO concentrations measured in shallow coastal and estuarine waters such as the Ishikari Bay (Hokkaido, Japan), the Jiaozhou Bay (China), the Scheldt Estuary (The Netherlands) and the Carbon monoxide in the SW Baltic Sea 80 Yaquina Bay (Oregon, USA) which range from 0.3 to 50 nmol L-1 (Butler et al., 1987; Law et al., 2002; Nakagawa et al, 2004; Wang et al., 2017). Figure 4.5: Monthly distributions of CO (E) and relevant photochemical parameters: DOC (A), aCDOM(350) (B), S275-295 (C), and SR (D) at BE in 2020. Please note that the blank areas are due to data gaps caused by cancellations (COVID-19) of the research cruises. Next, need to emphasize, the spectral slopes of the absorption spectrum (S) were obtained by nonlinear fitting of the absorption coefficient according to the equation from Stedmon et al. (2000). The ratio of the spectral slope (SR) of two different spectral slopes, S275–295 and S350–400, Carbon monoxide in the SW Baltic Sea 81 was used to indicate the magnitude of the relative molecular weight of CDOM (Helms et al. 2008) and to identify the origin of the CDOM: the smaller the SR, the higher the proportion of terrestrial CDOM, and vice versa (John et al., 2009; Helms et al., 2013). The combination of the ratios of the spectral slope of the shorter waveband (275–295 nm): S275–295, SR, aCDOM (350), solar radiation, and DOC could provide a more accurate estimate of the photochemical characteristics of CO in the area. Albeit, the S275–295 values were lower than those observed in the open ocean (0.04 nm−1) (Aurin and Mannino, 2012), suggesting that BE CDOM pool contained high-molecular-weight CDOM with higher terrestrial aromaticity than those found in the open ocean. However, based on a comparison of the average CDOM and SR (Helms et al., 2008; Chen et al., 2011; Shen et al., 2012; Yang et al., 2021), our CDOM (aCDOM (350): 0.28 ± 0.08 m-1) was in a low level while SR (1.35 ± 0.08) was the opposite. Oceanic CO is mainly produced photochemically via the reaction of UV light with CDOM (see e.g. Powers and Miller, 2015) and anthropogenic CDOM has been demonstrated to be more efficient at photochemically producing CO (Yang et al., 2011; Zhao et al.,2015; Ossola et al., 2022; Xu et al., 2023). Thus, at BE, CDOM was mostly of terrestrial origin and since the fluvial input is not particularly strong, the precursor substance (e.g. lignin) for CO is limiting its concentrations. Notably, the lower CO concentrations in the surface water than the subsurface or mixed-layer maximum (e.g. in July and October) could be attributed to the enhanced sea–air exchange, as the wind speeds were high at BE (Table 4.1). During the survey period, the distribution of aCDOM (350) and solar radiation in the surface layer did not have the same trend (Fig. 4.6). Surface aCDOM (350) generally presented a relatively average level and the fluctuation of solar radiation was noticeable. Although both factors can influence the distribution of CO in seawater, at BE, the dynamics of diatoms were better correlated with changes in the concentration of CO (see section 4.4.5). In the bottom layer, after the spring bloom in March, maximum values of aCDOM (350) and DOC were observed in April. Additionally, related to CO regeneration in anoxic water, the concentration of DOC increased again. For the CO, high concentrations were generally observed in the bottom layer with the maximum value of 1.89 nmol L−1 measured at 25 m depth in September 2020 (Fig. 4.5E and Fig. 4.6). In September 2021, CO had a similar vertical distribution (Fig. 4.7), under anoxic Carbon monoxide in the SW Baltic Sea 82 conditions, with the highest CO concentration (1.24 nmol L−1) in the bottom layer. Similarly, the average CO concentration in anoxic water at the Pettaquamscutt River lower pond was up to 2 nmol L-1 (Schmidt and Heikes, 2014). In the hypoxia/anoxia environment of the ocean, certain anaerobic microbes, including methanogenic archaea and acetogenic bacteria, produce CO as an intermediate during carbon dioxide fixation (Seravalli and Ragsdale 2000). Specifically, some methane-producing microorganisms produce CO during their metabolism because they need some auxiliary oxidants to help complete the metabolic process of methane (Conrad and Thauer 1983). Sediments in Eckernförde Bay receive large amounts of organic matter (Smetacek et al.,1987; Nittrouer et al., 1998) and are active sites of methane formation (Maltby et al., 2018; Ma et al., 2020; Perner et al., 2022). We, therefore, suggest that the accumulation of CO in the bottom water in September was caused by its release from the anoxic sediment (King 2007) and in-situ production in the overlying water column in combination with the pronounced water column stratification during autumn which prevented mixing of CO into the surface layer. However, no sediment samples were taken at that time and the mechanism of the anoxic production is unclear, which needs to be further investigated. Figure 4.6: Monthly distributions of CO (filled squares) vs solar radiation (SR, filled triangles), diatoms (filled rhombuses), aCDOM (350) (filled dots), Oxygen (filled stars), and DOC (filled hexagons) in the surface and bottom layer at BE in 2020. Carbon monoxide in the SW Baltic Sea 83 Figure 4.7: CO and dissolved oxygen distributions at BE in September 2020 and 2021. 4.4.5 Environmental controls on CO distribution at BE In this section, the correlation heatmap was used to analyze the environmental controls on CO distribution at BE (n = 48) (Fig. 4.8). The correlation revealed that nutrients in the area seem to have a significant contribution to the distribution of CO, especially in the photoproduction process. Nitrate (r = 0.65, p < 0.01), an excellent photosensitizer, is the main source of ∙OH in water and has been a hot topic of research for decades (McFall et al., 2018; Zhang et al., 2019) especially for organic matter that cannot absorb solar radiation for direct degradation (Li et al., 2022). Recently, Ossola et al. (2022) demonstrated for the first time a precise mechanism for CO production by CDOM photolysis: in Lignin, direct photolytic demethylation of aromatic methoxy groups to methanol, which oxidized by ∙OH to produce CO. Therefore, although nitrate is not directly involved in the photochemical production of CO, it can indirectly contribute to the photochemical production rate of CO through the production of ∙OH. For the other nutrients (e.g. silicate: r = 0.71, p < 0.01; phosphate: r = 0.82, p < 0.01), there was no valid evidence to directly affect the photodegradation of CDOM. Besides, salinity (r = 0.65, p < 0.01) and oxygen changes (r = -0.63, p < 0.01) were also important factors affecting the distribution of CO at BE and this could be linked to the processes above (benthic release, or production form available DOM accumulated at the bottom). Carbon monoxide in the SW Baltic Sea 84 Figure 4.8: Correlation analysis between CO concentrations and environmental parameters (n = 48). In 2020, during the spring bloom and minor bloom period in August, diatoms were in high abundance in the surface layer at the Boknis Eck station (Fig. 4.6). Moreover, a positive correlation was found between diatoms and CO in the surface layer (Fig. S3: r = 0.69, p < 0.01, n = 9). There seemed to be a co-occurrence in the seasonal cycle of diatom abundances and CO concentrations. Based on this data one can of course not prove conclusively that indeed diatoms are responsible, but it is certainly an indication that in systems with relatively low CDOM, this could play a more significant role as compared to waters in which CDOM is the primary source there is. Indeed, diatoms are known for significantly producing CO (Gros et al., 2009). 4.4.6 Comparison of surface CO production/consumption rates at BE The surface CO photoproduction rate over all the relevant wavelength range (290 – 490 nm) was calculated as (Xie et al., 2009; Conte et al., 2019): 6Z[\PFJ!J = ∫ ^ ( W ) ×Q8Q1R(W) ST" =T" ×X^_(W)\W (10) where Q is the global spectral solar irradiance (photons m-2 d-1 nm-1), the spectra were modeled using the SMARTS2 (Simple Model of the Atmospheric Radiative Transfer of Sunshine, version 2.9.5) (Gueymard, 2001). AQY (Apparent quantum yield in moles of CO produced per moles of photons absorbed by CDOM). We computed the spectral variation in AQY by taking Carbon monoxide in the SW Baltic Sea 85 the average of three published parameterizations: Zafiriou et al. (2003), Ziolkowski and Miller (2007), and White et al. (2010) (Fig. 4.9A). CDOM absorption coefficient mean in the surface layer at BE in 2020 is shown in Fig. 4.9B. Thus, we estimated the Prodphoto in the area to be 12.11 μmol m-2 d-1 (Table 4.2). This result reported here is at the lower end of CO photoproduction rates estimated in shallow coastal and estuarine waters such as the Amundsen Gulf (Canada), the Jiaozhou Bay (China), and the Mississippi–Atchafalaya River (Mexico) which range from 6.0 to 45.8 μmol m-2 d-1 (Xie et al., 2009; Ren et al., 2014; Powers and Miiller. 2015). Based on the area of the survey area (~3.77 × 105 km2), the annual CO photoproduction in the Baltic Sea was 20.05 Gg CO-C yr-1, accounting for about 0.10% of the global oceanic CO photoproduction (19.1 Tg CO-C yr-1) (Conte et al., 2019). Figure 4.9: CO apparent quantum yield (AQY in mol of CO per mol of photons) (A) and CDOM absorption coefficient mean in the surface layer at BE in 2020 (aCDOM in m−1) (B) as a function of the wavelength (nm). We estimated a microbial consumption rate (Cbact) was estimated to be 8.45 μmol m-2 d-1 (Prodphoto - Fsea–air) (Table 2). The main sink of CO at BE in 2020 was the Cbact, which accounted for about 70% of the total CO loss terms in the surface layer. Our result was lower than the estimate of Zhang and Xie (2012), who reported that the microbial consumption of CO in the St. Lawrence Estuary accounted for 86–92% of the CO sink. Here, high winds (see section 4.3) were observed, and the percentage of air-sea gas exchange loss became higher. Summarizing this data, microbial CO consumption was still the dominant removal pathway of dissolved CO in the surface layer at BE thus significantly modulating atmospheric CO emissions. Overall, the source and sink at BE were basically balanced with roughly similar magnitudes. Carbon monoxide in the SW Baltic Sea 92 Carbon monoxide in the SW Baltic Sea 93 References Aurin, D., and Mannino, A.: A database for developing global ocean color algorithms for colored dissolved organic material, CDOM spectral slope, and dissolved organic carbon. The Oceanography Society, Glasgow, UK, 2012. Bange, H. W., Bergmann, K., Hansen, H. P., Kock, A., Koppe, R., Malien, F., and Ostrau, C.: Dissolved methane during hypoxic events at the Boknis Eck time series station (Eckernförde Bay, SW Baltic Sea), Biogeosciences, 7, 1279–1284, https://doi.org/10.5194/bg-7-1279-2010, 2010. Becker, S., Tebben, J., Coffinet, S., Wiltshire, K., Iversen, M. H., Harder, T., Hinrichs, K. U., and Hehemann, J. H.: Laminarin is a major molecule in the marine carbon cycle, P. Natl. Acad. Sci. USA, 117, 6599–6607, https://doi.org/10.1073/pnas.1917001117, 2020. Butler, J. H., Jones, R. D., Garber, J. H., and Gordon, L. I.: Seasonal distributions and turnover of reduced trace gases and hydroxylamine in Yaquina Bay, Oregon, Geochimica et Cosmochimica Acta, 51(3), 697706, https://doi.org/10.1016/0016-7037(87)90080-9, 1987. Campen, H. I., Arévalo-Martínez, D. L., and Bange, H. W.: Carbon monoxide (CO) cycling in the Fram Strait, Arctic Ocean, Biogeosciences, 20, 1371–1379, https://doi.org/10.5194/bg-20-1371-2023, 2023. Canadell, J. G., Monteiro, P. M., Costa, M. H., Da Cunha, L. C., Cox, P. M., Alexey, V., ... and Lebehot, A. D.: Global carbon and other biogeochemical cycles and feedbacks, https://doi.org/10.1017/9781009157896.007, 2021 Carstensen, J., Andersen, J. H., Gustafsson, B. G., and Conley, D. J.: Deoxygenation of the Baltic Sea during the last century. Proceedings of the National Academy of Sciences, 111(15), 5628-5633, https://doi.org/10.1073/pnas.1323156111, 2014. Chen, H., Zheng, B., Song, Y., and Qin, Y.: Correlation between molecular absorption spectral slope ratios and fluorescence humification indices in characterizing CDOM. Aquatic Sciences, 73, 103-112, https://doi.org/10.1007/s00027-010-0164-5, 2011. Conley, D. J., Carstensen, J., Aigars, J., Axe, P., Bonsdorff, E., Eremina, T., ... and Zillén, L.: Hypoxia is increasing in the coastal zone of the Baltic Sea, Environmental Science and Technology, 45(16), 67776783, https://doi.org/10.1021/es201212r, 2011. Conrad, R., Seiler, W., Bunse, G., and Giehl, H.: Carbon monoxide in seawater (Atlantic Ocean), J. Geophys. Res., 87, 8839, https://doi.org/10.1029/JC087iC11p08839, 1982. Conrad, R., and Thauer, R. K.: Carbon monoxide production by Methanobacterium thermoautotrophicum, FEMS microbiology letters, 20(2), 229-232, https://doi.org/10.1111/j.15746968.1983.tb00122.x, 1983. Carbon monoxide in the SW Baltic Sea 94 Conte, L., Szopa, S., Séférian, R., and Bopp, L.: The oceanic cycle of carbon monoxide and its emissions to the atmosphere, Biogeosciences, 16, 881–902, https://doi.org/10.5194/bg-16-881-2019, 2019. Dale, A. W., Bertics, V. J., Treude, T., Sommer, S., and Wallmann, K.: Modeling benthic–pelagic nutrient exchange processes and porewater distributions in a seasonally hypoxic sediment: evidence for massive phosphate release by Beggiatoa?, Biogeosciences, 10, 629–651, https://doi.org/10.5194/bg-10-629-2013, 2013. Day, D. A. and Faloona, I.: Carbon monoxide and chromophoric dissolved organic matter cycles in the shelf waters of the northern California upwelling system. Journal of Geophysical Research: Oceans, 114(C1), https://doi.org/10.1029/2007JC004590, 2009. Draxler, R. R., and Rolph, G. D.: HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY Website. Silver Spring, MD: NOAA Air Resources Laboratory. ready. arl. noaa. gov/HYSPLIT. php, 2013. Dreshchinskii, A., and Engel, A.:Seasonal variations of the sea surface microlayer at the Boknis Eck Times Series Station (Baltic Sea), Journal of Plankton Research, 39(6), 943-961, https://doi.org/10.1093/plankt/fbx055, 2017 Evans, W.F.J., Puckrin, E.: An observation of the greenhouse radiation associated with carbon monoxide, Geophys. Res. Lett. 22 (8), 925–928, https://doi.org/10.1029/95GL00606, 1995. Forster, P., Storelvmo, T., Armour, K., Collins, W., Dufresne, J.-L., Frame, D., Lunt, D.J., Mauritsen, T., Palmer, M.D., Watanabe, M., Wild, M., and Zhang, H.: The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity, in: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J.B.R., Maycock, T.K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou. B., Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 923–1054, https://doi.org/10.1017/9781009157896.009, 2021. Gindorf, S., Bange, H. W., Booge, D., and Kock, A.: Seasonal study of the small-scale variability in dissolved methane in the western Kiel Bight (Baltic Sea) during the European heatwave in 2018, Biogeosciences, 19, 4993–5006, https://doi.org/10.5194/bg-19-4993-2022, 2022. Greening, C. and Grinter, R.: Microbial oxidation of atmospheric trace gases, Nature Rev. Microbiol., 20, 513-528, https://doi.org/10.1038/s41579-022-00724-x, 2022. Gros, V., Peeken, I., Bluhm, K., Zöllner, E., Sarda-Esteve, R., and Bonsang, B.: Carbon monoxide emissions by phytoplankton: evidence from laboratory experiments, Environmental chemistry, 6(5), 369379, https://doi.org/10.1071/EN09020, 2009. Gueymard, C. A.: Parameterized transmittance model for direct beam and circumsolar spectral irradiance, Carbon monoxide in the SW Baltic Sea 95 Solar Energy, 71(5), 325-346, https://doi.org/10.1016/S0038-092X(01)00054-8, 2001. Hansen, H. P., Giesenhagen, H. C., and Behrends, G.: Seasonal and long-term control of bottom water oxygen deficiency in a stratified shallow-coastal system, ICES J. Mar. Sci. Suppl., 56, 65–71, https://doi.org/10.1006/jmsc.1999.0629, 1999. Helms, J. R., Stubbins, A., Perdue, E. M., Green, N. W., Chen, H., and Mopper, K.: Photochemical bleaching of oceanic dissolved organic matter and its effect on absorption spectral slope and fluorescence, Marine Chemistry, 155, 81–91, https://doi.org/10.1016/j. marchem.2013.05.015, 2013. Helms, J. R., Stubbins, A., Ritchie, J. D., Minor, E. C., Kieber, D. J., and Mopper, K.: Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter, Limnology and oceanography, 53(3), 955-969, https://doi.org/10.4319/lo.2008.53.3.0955, 2008. Hoesly, R. M., Smith, S. J., Feng, L., Klimont, Z., Janssens-Maenhout, G., Pitkanen, T., Seibert, J. J., Vu, L., Andres, R. J., Bolt, R. M., Bond, T. C., Dawidowski, L., Kholod, N., Kurokawa, J.-I., Li, M., Liu, L., Lu, Z., Moura, M. C. P., O'Rourke, P. R., and Zhang, Q.: Historical (1750–2014) anthropogenic emissions of reactive gases and aerosols from the Community Emissions Data System (CEDS), Geosci. Model Dev., 11, 369–408, https://doi.org/10.5194/gmd-11-369-2018, 2018. Holloway, T., Levy, H., and Kasibhatla, P.: Global distribution of carbon monoxide, Journal of Geophysical Research: Atmospheres, 105(D10), 12123-12147, https://doi.org/10.1029/1999JD901173, 2000. Hoppe, H. G., Giesenhagen, H. C., Koppe, R., Hansen, H. P., and Gocke, K.: Impact of change in climate and policy from 1988 to 2007 on environmental and microbial variables at the time series station Boknis Eck, Baltic Sea, Biogeosciences, 10, 4529–4546, https://doi.org/10.5194/bg-10-4529-2013, 2013. John, R. H., Stubbins, A., Ritchie, J. D., Minor, E. C., Kieber, D. J., and Mopper, K.: Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnology and Oceanography, 54(3), 1023. https://doi.org/10.4319/lo.2009.54.3.1023, 2009. King, G. M.: Microbial carbon monoxide consumption in salt marsh sediments, FEMS Microbiology Ecology, 59(1), 2-9, https://doi.org/10.1111/j.1574-6941.2006.00215.x, 2007. Kitidis, V., Laverock, B., McNeill, L. C., Beesley, A., Cummings, D., Tait, K., ... and Widdicombe, S.: Impact of ocean acidification on benthic and water column ammonia oxidation, Geophysical Research Letters, 38(21), https://doi.org/10.1029/2011GL049095, 2011. Law, C. S., Sjoberg, T. N., and Ling, R. D.: Atmospheric emission and cycling of carbon monoxide in the Scheldt Estuary, Biogeochemistry, 59(1), 69-94, https://doi.org/10.1023/A:1015592128779, 2002. Carbon monoxide in the SW Baltic Sea 96 Lennartz, S. T., Lehmann, A., Herrford, J., Malien, F., Hansen, H.-P., Bi est er, H ., and Ba ng e, H. W.: Long-term trends at the Boknis Eck time series station (Baltic Sea), 1957–2013: does climate change counteract the decline in eutrophication?, Biogeosciences, 11, 6323–6339, https://doi.org/10.5194/bg11-6323-2014, 2014. Li, J. L., Zhai, X., and Du, L.: Effect of nitrate on the photochemical production of carbonyl sulfide from surface seawater. Geophysical Research Letters, 49(13), e2021GL097051, https://doi.org/10.1029/2021GL097051, 2022. Liblik, T. and Lips, U.: Stratification has strengthened in the Baltic Sea–an analysis of 35 years of observational data, Frontiers in Earth Science, 7, 174, https://doi.org/10.3389/feart.2019.00174, 2019. Liss, P. S., Marandino, C. A., Dahl, E. E., Helmig, D., Hintsa, E. J., Hughes, C., ... and Williams, J.: Short-lived trace gases in the surface ocean and the atmosphere, In Ocean-Atmosphere Interactions of Gases and Particles (pp. 1-54). Springer, Berlin, Heidelberg, 2014. Loewus, M. W. and Delwiche, C. C.: Carbon monoxide production by algae. Plant Physiology, 38(4), 371, https://doi.org/10.1104/pp.38.4.371, 1963. Loginova, A. N., Thomsen, S., and Engel, A.: Chromophoric and fluorescent dissolved organic matter in and above the oxygen minimum zone off Peru, Journal of Geophysical Research: Oceans, 121(11), 79737990, https://doi.org/10.1002/2016JC011906, 2016. Lopez, J. D. P.: Seasonality and global growth trends of carbon monoxide during 1995—2001, 2003. Lübbecke, E., Lübbecke, M. E., & Möhring, R. H.: Ship traffic optimization for the Kiel Canal. Operations Research, 67(3), 791-812, https://doi.org/10.1287/opre.2018.1814, 2019. Ma, X., Sun, M., Lennartz, S. T., and Bange, H. W.: A decade of methane measurements at the Boknis Eck Time Series Station in Eckernförde Bay (southwestern Baltic Sea), Biogeosciences, 17, 3427–3438, https://doi.org/10.5194/bg-17-3427-2020, 2020. Maltby, J., Steinle, L., Löscher, C. R., Bange, H. W., Fischer, M. A., Schmidt, M., and Treude, T.: Microbial methanogenesis in the sulfate-reducing zone of sediments in the Eckernförde Bay, SW Baltic Sea, Biogeosciences, 15, 137–157, https://doi.org/10.5194/bg-15-137-2018, 2018. McFall, A. S., Edwards, K. C., and Anastasio, C.: Nitrate photochemistry at the air–ice interface and in other ice reservoirs. Environmental science & technology, 52(10), 5710-5717, https://doi.org/10.1021/acs.est.8b00095, 2018. McLeod, A. R., Brand, T., Campbell, C. N., Davidson, K., and Hatton, A. D.: Ultraviolet Radiation Drives Emission of Climate‐Relevant Gases from Marine Phytoplankton, Journal of Geophysical Research: Biogeosciences, 126(9), e2021JG006345, https://doi.org/10.1029/2021JG006345, 2021. Nakagawa, F., Tsunogai, U., Gamo, T., and Yoshida, N.: Stable isotopic compositions and fractionations Carbon monoxide in the SW Baltic Sea 97 of carbon monoxide at coastal and open ocean stations in the Pacific, Journal of Geophysical Research: Oceans, 109(C6), https://doi.org/10.1029/2001JC001108, 2004. Nittrouer, C. A., Lopez, G. R., Wright, L. D., Bentley, S. J., D’Andrea, A. F., Friedrichs, C. T., ... and Sommerfield, C. K.: Oceanographic processes and the preservation of sedimentary structure in Eckernförde Bay, Baltic Sea, Continental Shelf Research, 18(14-15), 1689-1714, https://doi.org/10.1016/S0278-4343(98)00054-5, 1998. Ossola, R., Gruseck, R., Houska, J., Manfrin, A., Vallieres, M., and McNeill, K.: Photochemical Production of Carbon Monoxide from Dissolved Organic Matter: Role of Lignin Methoxyarene Functional Groups, Environmental Science & Technology, 56(18), 13449-13460, https://doi.org/10.1021/acs.est.2c03762, 2022. Park, K. and Rhee, T. S.: Oceanic source strength of carbon monoxide on the basis of basin-wide observations in the Atlantic, Environmental Science-Processes and Impacts, 18, 104-114, 10.1039/c5em00546a, https://doi.org/10.1039/C5EM00546A, 2016. Perner, M., Wallmann, K., Adam-Beyer, N., Hepach, H., Laufer-Meiser, K., Böhnke, S., ... and Scholz, F.: Environmental changes affect the microbial release of hydrogen sulfide and methane from sediments at Boknis Eck (SW Baltic Sea), Frontiers in Microbiology, 13, Art-Nr, https://doi.org/10.3389/fmicb.2022.1096062, 2022. Petron, G., Crotwell, A. M., M.J. Crotwell, M. J., Dlugokencky, E., Madronich, M., Moglia, E., Neff, D., Thoning, K., Wolter, S., Mund, J. W.: Atmospheric Carbon Monoxide Dry Air Mole Fractions from the NOAA GML Carbon Cycle Cooperative Global Air Sampling Network, 1988-2021, Version: 2022-0728, https://doi.org/10.15138/33bv-s284, 2022. Piontek, J., Meeske, C., Hassenrück, C., Engel, A., and Jürgens, K.: Organic matter availability drives the spatial variation in the community composition and activity of Antarctic marine bacterioplankton. Environmental Microbiology, 24(9), 4030-4048, https://doi.org/10.1111/1462-2920.16087, 2022. Pisani, O., Yamashita, Y., and Jaffé, R.: Photo-dissolution of flocculent, detrital material in aquatic environments: Contributions to the dissolved organic matter pool, Water Research, 45(13), 3836-3844, https://doi.org/10.1016/j.watres.2011.04.035, 2011. Powers, L. C., and Miller, W. L.: Photochemical production of CO and CO2 in the Northern Gulf of Mexico: Estimates and challenges for quantifying the impact of photochemistry on carbon cycles, Marine Chemistry, 171, 21-35, https://doi.org/10.1016/j.marchem.2015.02.004, 2015. Raymond, P. A. and Cole, J. J.: Gas exchange in rivers and estuaries: Choosing a gas transfer velocity, Estuaries, 24(2), 312-317, https://doi.org/10.2307/1352954, 2001. Ren, C., Yang, G., and Lu, X.: Autumn photoproduction of carbon monoxide in Jiaozhou Bay, China. Journal of Ocean University of China, 13, 428-436, https://doi.org/10.1007/s11802-014-2225-1, 2014. Carbon monoxide in the SW Baltic Sea 98 Reckermann, M., Omstedt, A., Soomere, T., Aigars, J., Akhtar, N., Bełdowska, M., Bełdowski, J., Cronin, T., Czub, M., Eero, M., Hyytiäinen, K. P., Jalkanen, J.-P. , Ki essling, A. , Kjells trö m, E ., Ku liń ski, K ., Larsén, X. G., McCrackin, M., Meier, H. E. M., Oberbeckmann, S., Parnell, K., Pons-Seres de Brauwer, C., Poska, A., Saarinen, J., Szymczycha, B., Undeman, E., Wörman, A., and Zorita, E.: Human impacts and their interactions in the Baltic Sea region, Earth System Dynamics, 13, 1–80, https://doi.org/10.5194/esd-13-1-2022, 2022. Riggsbee, J. A., Orr, C. H., Leech, D. M., Doyle, M. W., and Wetzel, R. G.: Suspended sediments in river ecosystems: Photochemical sources of dissolved organic carbon, dissolved organic nitrogen, and adsorptive removal of dissolved iron, Journal of Geophysical Research: Biogeosciences, 113(G3), https://doi.org/10.1029/2007JG000654, 2008. Schmidt, C. E., and Heikes, B. G.: Aqueous carbon monoxide cycling in a fjord-like estuary, Estuaries and coasts, 37(3), 751-762, https://doi.org/10.1007/s12237-013-9722-0, 2014. Seiler, W. and Schmidt, U.: Dissolved nonconservative gases in seawater, The sea, edited by: Goldberg, ED, 5, 219-243, ISSN: 1559-2723, 1974. Seravalli, J., and Ragsdale, S. W.: Channeling of carbon monoxide during anaerobic carbon dioxide fixation, Biochemistry, 39(6), 1274-1277, https://doi.org/10.1021/bi991812e, 2000. Shen, Y., Fichot, C. G., and Benner, R.: Floodplain influence on dissolved organic matter composition and export from the Mississippi—Atchafalaya River system to the Gulf of Mexico. Limnology and Oceanography, 57(4), 1149-1160, https://doi.org/10.4319/lo.2012.57.4.1149, 2012. Smetacek, V., Bodungen, B. V., Bölter, M., Bröckel, K. V., Dawson, R., Knoppers, B., ... and Zeitzschel, B.: The pelagic system. Seawater-Sediment Interactions in Coastal Waters: An Interdisciplinary Approach, 32-68, 1987. Stedmon, C. A., Markager, S., and Kaas, H.: Optical properties and signatures of chromophoric dissolved organic matter (CDOM) in Danish coastal waters, Estuarine, Coastal and Shelf Science, 51(2), 267–278. https://doi.org/10.1006/ecss.2000.0645, 2000. Sugai, Y., Tsuchiya, K., Shimode, S., and Toda, T.: Photochemical Production and Biological Consumption of CO in the SML of Temperate Coastal Waters and Their Implications for Air‐Sea CO Exchange, Journal of Geophysical Research: Oceans, 125(4), e2019JC015505, https://doi.org/10.1029/2019JC015505, 2020. Stubbins, A., Uher, G., Kitidis, V., Law, C., Upstill-Goddard, R.C., Woodward, E.M.S.: The open-ocean source of atmospheric carbon monoxide, Deep Sea Res. 53, 1685e1694, https://doi.org/10.1016/j.dsr2.2006.05.010, 2006. Taylor, J. A., Zimmerman, P. R., and Erickson, D. J.: A 3-D modelling study of the sources and sinks of atmospheric carbon monoxide, Ecol. Model., 88, 53–71, https://doi.org/10.1016/0304-3800(95)00069-0, Carbon monoxide in the SW Baltic Sea 99 1996. Thompson, A. M.: The oxidizing capacity of the earth’s atmosphere – probable past and future changes, Science, 256, 1157–1165, 1992. Van H euk ele m, L. , a nd T homas, C. S.: Computer-assisted high-performance liquid chromatography method development with applications to the isolation and analysis of phytoplankton pigments, Journal of Chromatography A, 910(1), 31-49, https://doi.org/10.1016/S0378-4347(00)00603-4, 2001. Wang, W. L., Peng, T., Lu, X. L., and Zhao, B. Z.: Diurnal, seasonal, and spatial variations and flux of carbon monoxide in Jiaozhou Bay, China, Marine Chemistry, 191, 1-8, https://doi.org/10.1016/j.marchem.2017.01.004, 2017. Weiss, R. F. and Price, B. A.: Nitrous oxide solubility in water and seawater, Marine chemistry, 8(4), 347-359, https://doi.org/10.1016/0304-4203(80)90024-9, 1980. White, E. M., Kieber, D. J., Sherrard, J., Miller, W. L., and Mopper, K.: Carbon dioxide and carbon monoxide photoproduction quantum yields in the Delaware Estuary, Marine Chemistry, 118(1-2), 11-21, https://doi.org/10.1016/j.marchem.2009.10.001, 2010. Wiesenburg, D. A. and Guinasso Jr, N. L.: Equilibrium solubilities of methane, carbon monoxide, and hydrogen in water and sea water, Journal of chemical and engineering data, 24(4), 356-360 1979. Xie, H., Andrews, S. S., Martin, W. R., Miller, J., Ziolkowski, L., Taylor, C. D., and Zafiriou, O. C.: Val ida ted met hod s fo r sa mpl ing a nd h ead spa ce a nal ysi s of car bon mon oxi de i n se awat er, M ari ne Chemistry, 77(2-3), 93-108, https://doi.org/10.1016/S0304-4203(01)00065-2, 2002. Xie, H., Bélanger, S., Demers, S., Vincent, W. F., and Papakyriakou, T. N.: Photobiogeochemical cycling of carbon monoxide in the southeastern Beaufort Sea in spring and autumn, Limnology and Oceanography, 54(1), 234-249, https://doi.org/10.4319/lo.2009.54.1.0234, 2009. Xie, H. and Zafiriou, O. C.: Evidence for significant photochemical production of carbon monoxide by particles in coastal and oligotrophic marine waters, Geophysical Research Letters, 36(23), https://doi.org/10.1029/2009GL041158, 2009. Xu, G. B., Xu, F., Ji, X., Zhang, J., Yan, S. B., Mao, S. H., and, Yang, G. P.: Carbon monoxide cycling in the Eastern Indian Ocean. Journal of Geophysical Research: Oceans, e2022JC019411, https://doi.org/10.1029/2022JC019411, 2023. Yang, G. P., Ren, C. Y., Lu, X. L., Liu, C. Y., & Ding, H. B.: Distribution, flux, and photoproduction of carbon monoxide in the East China Sea and Yellow Sea in spring, Journal of Geophysical Research: Oceans, 116(C2), https://doi.org/10.1029/2010JC006300, 2011. Yang, L., Zhang, J., and Yang, G. P.: Mixing behavior, biological and photolytic degradation of dissolved Carbon monoxide in the SW Baltic Sea 100 organic matter in the East China Sea and the Yellow Sea. Science of the Total Environment, 762, 143164, https://doi.org/10.1016/j.scitotenv.2020.143164, 2021. Zafiriou, O. C., Andrews, S. S., & Wang, W.: Concordant estimates of oceanic carbon monoxide source and sink processes in the Pacific yield a balanced global “blue‐water” CO budget, Global Biogeochemical Cycles, 17(1), https://doi.org/10.1029/2001GB001638, 2003. Zafiriou, O.C., Xie, H., Nelson, N.B., Najjar, R.G., Wang, W.: Diel carbon monoxide cycling in the upper Sargasso Sea near Bermuda at the onset of spring and in midsummer, Limnol. Oceanogr. 53, 835–850, https://doi.org/10.4319/lo.2008.53.2.0835, 2008. Zhang, J., Wang, J., Zhuang, G. C., and Yang, G. P.: Carbon monoxide cycle in the Bohai Sea and the Yellow Sea: Spatial variability, sea‐air exchange, and biological consumption in autumn, Journal of Geophysical Research: Oceans, 124(6), 4248-4257, https://doi.org/10.1029/2018JC014864, 2019. Zhang, Y., Bao, F., Li, M., Chen, C., and Zhao, J.: Nitrate-enhanced oxidation of SO2 on mineral dust: a vital role of a proton. Environmental Science & Technology, 53(17), 10139-10145, https://doi.org/10.1021/acs.est.9b01921, 2019. Zhang, Y., Xie, H., Fichot, C. G., and Chen, G.: Dark production of carbon monoxide (CO) from dissolved organic matter in the St. Lawrence estuarine system: Implication for the global coastal and blue water CO budgets, Journal of Geophysical Research: Oceans, 113(C12), https://doi.org/10.1029/2008JC004811, 2008. Zhang, Y., and Xie, H.: The sources and sinks of carbon monoxide in the St. Lawrence estuarine system, Deep Sea Research Part II: Topical Studies in Oceanography, 81, 114-123, https://doi.org/10.1016/j.dsr2.2011.09.003, 2012. Zhao, B. Z., Yang, G. P., Xie, H., Lu, X. L., and Yang, J.: Distribution, flux and photoproduction of carbon monoxide in the Bohai and Yellow Seas, Marine Chemistry, 168, 104-113, https://doi.org/10.1016/j.marchem.2014.11.006, 2015. Zhao, Y., Schlundt, C., Booge, D., and Bange, H. W.: A decade of dimethyl sulfide (DMS), dimethylsulfoniopropionate (DMSP) and dimethyl sulfoxide (DMSO) measurements in the southwestern Baltic Sea, Biogeosciences, 18, 2161–2179, https://doi.org/10.5194/bg-18-2161-2021, 2021. Zheng, B., Chevallier, F., Yin, Y., Ciais, P., Fortems-Cheiney, A., Deeter, M. N., ... and Zhao, Y: Global atmospheric carbon monoxide budget 2000–2017 inferred from multi-species atmospheric inversions, Earth System Science Data, 11(3), 1411-1436.https://doi.org/10.5194/essd-11-1411-2019, 2019. Ziolkowski, L. A. and Miller, W. L.: Variability of the apparent quantum efficiency of CO photoproduction in the Gulf of Maine and Northwest Atlantic, Mar. Chem., 105, 258–270, https://doi.org/10.1016/j.marchem.2007.02.004, 2007. Global reconstruction of oceanic carbon monoxide emissions 101 5 Global reconstruction of oceanic carbon monoxide emissions Manuscript in preparation for GBC: Li, G., Arévalo-Martínez, D. L., Słupiński, M., and Bange, H. W.: Global reconstruction of oceanic carbon monoxide emissions. Abstract: The assessment of the global oceanic emission estimate of the indirect greenhouse gas carbon monoxide (CO) is limited by the uncertainty of the oceanic CO fluxes to the atmosphere, because of the sparse sampling of dissolved CO in the marine environment. Here, we reconstruct climatological CO emissions from the ocean by training machine-learning models with over 12,000 CO measurements from the surface ocean—the largest synthesis to date. The prediction reproduced CO disequilibrium (∆CO) relatively faithfully (r2 = 0.88), with a root-mean-square error (RMSE) of 0.60 nmol L-1. The map of CO fluxes shows latitudinal gradients and a significant global seasonal cycle. We estimate an annual mean CO flux of 5.6 Tg CO-C yr−1, 8% of which occurs in the Southern Ocean (0.46 ± 0.10 Tg CO-C yr−1). This ocean flux estimate is in line with the ranges reported in former studies but reduces its uncertainty by more than three-fold. Global reconstruction of oceanic carbon monoxide emissions 108 (Zafiriou et al., 2003; Xie et al., 2009a; Day and Faloona 2009; Stubbins et al., 2011; Liss et al., 2014; Park and Rhee, 2016; Conte et al. 2019), we found that all the surface ocean is supersaturated with CO, including productive upwelling regions, the coastal areas, the oligotrophic gyres, and the northern part of the Southern Ocean (Fig. 5.2C). In general, the climatological mean values of ∆CO are highest at low latitudes, tend to decrease mid-latitudes, and with a significant increase at high latitudes. ∆CO exhibits elevated levels in the equatorial regions of both the Atlantic and Pacific Oceans, with a more pronounced peak observed (> 3.5 nmol L-1) in the Atlantic Ocean. However, there is no pronounced maximum in the eastern equatorial Pacific Ocean with its highly productive area where surface waters are enriched in nutrients by the equatorial upwelling (Wyrtki, 1981). Another aspect of our reconstruction is the capture of the narrow bands of strong supersaturation observed at about 40° S latitude including the boundary upwelling systems (e.g. southeast coast of South America). This latitude cuts on the lower boundary of the Humboldt Current and as such it could be associated with an upwelling system. Furthermore, based on the latitude and the fact that this band is consistent throughout all basins, this could be more related to the subtropical front (STF) that runs between the lower end of the subtropical gyres and the Antarctic circumpolar current. Based on the Fast Repetition Rate fluorometry photosynthetic measurement (Westwood et al., 2011), STF is pointed out to be the most productive region (Kerkar et al., 2020), which contributed to some of the high ∆CO was predicted, even exceeding 3 nmol L-1 there. Moreover, in the Southern Ocean, the biological production of CO reaches more than 3 mmol m−2 yr−1 (Conte et al., 2019). Bates et al. (1995) reported their high CO surface values (up to 4.7 nmol L−1), and high CO concentrations could be due to the release of organic matter during ice melting (Belzile et al., 2000) with the slower microbial CO uptake rates (0.09 d−1) in cold waters (Zafiriou et al., 2003) which cause the high ∆CO predicted. Compared to the integrated patterns, ∆CO presents clear minima in the centre of the subtropical gyres. In these oligotrophic gyres, Chl-a and CDOM are low, which prevents high CO photoproduction (Zafiriou et al., 2003; Stubbins et al., 2006; Ziolkowski and Miller, 2007). The main driving force for changing CO in these regions is the water side, the relatively low Global reconstruction of oceanic carbon monoxide emissions 109 concentration of CO in the surface layer triggers the lower ∆CO. 5.3.2 CO emissions Figure 3A presents the spatial patterns of emissions resulting from our model. The reconstructed CO air-sea flux shows that all oceanic regions are net sources of CO for the atmosphere, in particular, in the tropics, coastal upwelling systems, and subpolar regions. These patterns closely mirror those of ∆CO but with a relative amplification in regions of high winds, such as the mid to high latitudes. We estimated a global annual mean sea-air CO flux of 1.27 ± 0.54 mmol m−2 yr−1. The spatial pattern of emissions follows the one of the ∆CO. The strongest emissions are simulated in the equatorial region, especially in the Caribbean Sea and the east Pacific Ocean on both sides of the Equator. Figure 5.3: Oceanic CO emissions simulated by our model. Panel (A) presents the spatial distribution of annual mean emissions. The flux is in mmol m−2 yr−1 and is directed toward the atmosphere. Panel (B) presents the mean seasonal variation with latitude (mmol m-2 yr-1). Most importantly, emissions are stronger in the Southern Hemisphere than in the Northern Hemisphere (mean sea–air CO flux of 1.68 and 1.03 mmol m−2 yr−1, respectively), which could be attributable to the larger surface area covered by oceans, and relatively higher wind speed value. Combined with previous studies, sea–air CO flux values increase from the open ocean to the shelf areas (Conrad, 1982; Bates et al., 1995; Bourbonniere et al., 1997; Stubbins et al., 2006a; Zafiriou et al., 2008; Day and Faloona 2009; Yang et al., 2010; Liss et al., 2014; Zhang et al., 2019; Sugai et al., 2020). Coastal region (<200 m) emissions are stronger than in open ocean (mean CO emission of 1.47 and 1.26 mmol m−2 yr−1, respectively) (Depth data from Global reconstruction of oceanic carbon monoxide emissions 110 NOAA National Centers for Environmental Information, 2022, Fig. S5.4). However, coastal oceans just contribute about 3.60 % (0.2 Tg CO-C yr−1) to the overall oceanic emissions of CO. Annual mean emissions are reduced to around 0.5 mmol m−2 yr−1 in the centre of the subtropical gyres (e.g. Atlantic gyre, Indian Ocean gyre). In the Southern Ocean area (60°–69° S latitude in this study), where the average wind speed exceeded 10 m s-1 (Fig. S5.5), the average flux is up to 1.92 mmol m−2 yr−1. High emissions are also reached locally along the southwest coast of South America and Africa. Seasonal changes in the flux field are seen in Fig. 5.3B, and they are attributed to a combination of the effects of seasonal changes in water temperatures, ∆CO as well and changes in wind speeds. Emissions at the Equator (20° S–20° N) are roughly constant throughout the year (around 1.4 mmol m−2 yr−1) and between 20° N to 30°N, values are stabilized around 0.9 mmol m−2 yr−1. Emissions at intermediate latitudes (30° to 60°) vary seasonally from 0.48 mmol m−2 yr−1 to the highest values encountered (up to 2.27 mmol m−2 yr−1). At high latitude areas, seasonal variations are more pronounced caused by the temperature gradient, high wind speed conditions, and ice melting. Overall, the mean global ocean emission flux (Fig. 5.4C) has a small seasonal variability. The order of oceanic CO flux densities, sorted from largest to smallest, is Atlantic (1.55 mmol m−2 yr−1), Pacific (1.21 mmol m−2 yr−1), and Indian Ocean (1.12 mmol m−2 yr−1). Fluxes are notably more pronounced in the southern hemisphere compared to the northern hemisphere within the same oceans. The sea–air CO flux is governed by the complex interactions of various oceanographic and meteorological processes. To gain insight into regional differences in the influences of these processes, the mean flux per area in different months is shown in Fig. 5.4. The seasonal amplitude of the fluxes for the temperate North and South Atlantic Ocean (Fig. 4B) is more intense than those for the North and South Pacific Ocean (Fig. 5.4A), respectively. Besides, the mid-latitude South Atlantic Ocean is the most intense CO source area, and most mid-latitude regions of the Atlantic, Indian, and Pacific Oceans all change from a weak source in the respective summer seasons to a strong source in the respective winter seasons. Conversely, even though the seasonality is most evident in the north Indian Ocean (north of 20° N), with Global reconstruction of oceanic carbon monoxide emissions 111 the highest emission in summer where the summer monsoon winds force upwelling coincident with a high gas transfer rate (Takahashi et al., 2002). On the other hand, the sea-to-air CO flux over the temperate South Atlantic (1.65 mmol m−2 yr−1) is higher than the corresponding regions of the Indian Ocean (1.45 mmol m−2 yr−1 for 20°–60° S) and Pacific (1.04 mmol m−2 yr−1 for 20°–60° S). The equatorial Pacific and Atlantic show little seasonal variability except the equatorial Indian Ocean in CO flux. Figure 5.4: The monthly mean values for sea-air CO flux in the three major ocean basins: (A) Pacific, (B) Atlantic, (C) Indian, and Global Oceans. Average values in each climatic zone are plotted against the month (1 = January, 2 = February, …, and 12 = December). Finally, our global oceanic emission of CO is 5.6 Tg CO-C yr−1 and the Pacific Ocean alone contributes 43.6% (2.42 Tg CO-C yr−1). We trace most of the uncertainty in our estimate to the flux formulation, the windspeed, and ∆CO, in order of importance. The root-mean-square error (RMSE) of ∆CO is 0.60 nmol L-1 and the gas transfer coefficient (kCO), varies from 2.11 to 67.38 m s-1 (Wanninkhof, 1992; Nightingale et al., 2000; Edson et al., 2011; Butterworth and Miller, 2016) with an area-weighted global mean of 13.00 m s-1. Using the global ocean area of 361 × 106 km2 and ignoring ancillary data sets inherent uncertainties. To propagate these sources of uncertainty into our flux calculation, the random error for the global flux is estimated Global reconstruction of oceanic carbon monoxide emissions 112 to be about ± 2.96 Tg CO-C yr−1. From Pacific Ocean data and Atlantic Ocean data, Bates et al. (1995) and Park and Rhee (2016) pointed to a range of 3–11 Tg CO-C yr−1 and 1–12 Tg CO-C yr−1, respectively. The result simulated with our model falls well into the range of previous estimations based on extrapolations from in situ oceanic CO measurements but with reduced uncertainty. Figure 5.5: Oceanic CO emissions simulated by NEMO-PISCES of Conte et al. (2019) (A), present the spatial distribution of annual mean emissions. All fluxes are in mmol m−2 yr−1 and are directed toward the atmosphere. (B) Comparison of in situ oceanic CO flux density simulated with different models with those measured in the surface ocean (± standard error estimate) (Dots: longitudinal mean observed CO flux densities; Solid lines: longitudinal mean simulated CO flux densities). (C) Comparison of oceanic global CO emission simulated with different models with those measured in the surface ocean (± standard error estimate). The CO emissions produced by Conte et al. (2019) are currently used as spatial distribution by global chemistry-climate models (Fig. 5.5). In agreement with our simulation, all oceanic regions are sources of CO for the atmosphere. The emission pattern for Conte presents maxima around the equatorial zone and minima in the centre of the oligotrophic gyres. On an annual basis, Conte’s global emission is 4.0 Tg CO-C, which is roughly in the same range as ours. However, for the same given radiation in our model, a multitude of possible values for the ∆CO Global reconstruction of oceanic carbon monoxide emissions 113 is controlled by different features and not only by the light intensity. Additionally, compared with Conte’s emissions, we reconstructed a pronounced different spatial distribution. Besides, in the NEMO-PISCES (Nucleus for European Modelling of the Ocean, Pelagic Interaction Scheme for Carbon and Ecosystem Studies) model, the atmospheric CO mole fraction over the ocean is considered to be spatially constant and fixed to 90 ppb, and the RMSE (1.80 nmol L1) associated with the comparison with in-situ CO concentrations is triple higher than ours, which results in a flux estimation with great uncertainty. Additionally, Conte et al. (2019) stated that a relatively clear underestimation of the simulated CO concentrations at high latitudes significantly downplayed flux estimates in the region. Last but not least, Xu et al. (2023) pointed out that NEMO-PISCES simulated values are generally lower than their Eastern Indian Ocean data and that the parameterization of CDOM and the high variability of microbial processes uncertainties may have limited the model. By contrast, the observed data from 60° N to 50° S are better reproduced by our model than the NEMO-PISCES (Fig 5.5B), we demonstrate again that biases are behind previous underestimates of CO fluxes in the Southern Hemisphere, especially in the Southern Ocean area (south of 60° S latitude). By complementing the Southern Ocean observations in the late spring and winter, which led to an estimate of 0.46 ± 0.10 Tg CO-C yr−1 for the Southern Ocean oceanic CO emissions from our model. Such evidence reaffirmed that CO emissions should not drop to negligible values. This result highlights the limitations of previous approaches based on direct extrapolation of sparse measurements, compared to the model we use here. In conclusion, the values reached by us are relatively more accurate than those of the formers. Moreover, the choice of algorithm used in this study is subjective, given the availability of numerous other options. We selected the random forest regressor due to its suitability for the continuous regression task at hand, its relatively low computational cost, and its interpretability. Given the primary sources of uncertainty arising from limited and sparse observations, utilizing more complex techniques did not significantly improve results, considering various trade-offs. As for limitations, the model does not account for inter-annual variability in ocean physics and biogeochemistry. This omission could restrict the evaluation of simulated ∆CO against in situ measurements, as the data often cannot be averaged to represent daily means due to their Global reconstruction of oceanic carbon monoxide emissions 114 decades-long span and intermittent availability. Additionally, the contributions of rivers are not reflected in the model, and due to the lack of ancillary data (e.g. aCDOM320: Fig. S5.6, wind speed: Fig. S5.5) for the Arctic Ocean, our reconstruction, ∆CO does not cover that polar region, which may prevent a more accurate representation of the global budget of oceanic CO emissions. 5.4 Conclusions It is the first time that such a dataset of ∆CO has been gathered and that a machine learning model is used to predict global ∆CO distribution. Despite the scarcity of measurements regarding space and time, our model reproduces the observed ∆CO reasonably well. A spatially 1-degree global map from decades of observations reveals an intensely variable distribution of oceanic CO emissions, with the contribution of 5.6 Tg CO-C yr−1, and estimated the annual budget of CO in the Southern Ocean for the first time. Our emissions are quantitatively much closer than those published by Conte et al. (2019) and in better agreement in terms of spatial distribution with the known processes controlling the oceanic CO. The reduced uncertainty of our oceanic CO emissions and its ability to capture seasonal variations make it ideal for use as a prior in atmospheric inversions. This would translate into improved estimates of the overall atmospheric CO budget. Besides, the implementation of our oceanic CO flux module within an Earth system model will enable the exploration of potential Earth system feedbacks associated with the oceanic CO cycle. 5.5 Supplemental material Detailed methods for ∆ CO prediction In this work, we are interested in the supervised learning approach (Bishop, 2006; Murphy, 2012). Given a set of N training examples of the form {(x1, y1), ..., (xN, yN)} such that xi is the feature vector of the i-th example and yi (∆CO prediction) is its response variable, a learning algorithm seeks a function f: X → Y, where X is the input space and Y is the output space. The function f is an element of some space of possible functions F - called the hypothesis space. Global reconstruction of oceanic carbon monoxide emissions 115 Usually, we find f using a loss function l: X×Y → ℝ such that f is defined as returning the y value that gives the lowest loss value. Imputing missing value Because we are dealing with missing values, we first work on the imputation procedure. We use the IterativeImputer class from the package scikit-learn, which models each feature with missing values as a function of other features, and uses that estimate for imputation. It follows an iterated round-robin procedure: At each step, a feature column is designated as output y and the other feature columns are treated as input x. A regressor is fit on (x, y) for known y. Then, the regressor is used to predict the missing values of y. We repeat the procedure for each feature in an iterative fashion, and then repeat for a fixed number of imputation rounds - three in our case. The results of the final imputation round are returned. Ancillary data selection Some features may lead to overfitting or bring unnecessary noise to the model. So, we prefer to use a subset of features, rather than put every possible measurement and hope that eventually the model will learn something useful. We might be tempted to try every possible subset of variables; however, if there are k potential independent variables, then there are 2k distinct subsets of them to be tested. For example, suppose that you have 10 candidate-independent variables. In that case, the number of subsets to be tested is 210, which is 1024, and if you have 20 candidate variables, the number is 220, which is more than one million. That is why we are interested in some kind of score allowing us to find the most informative subset of features. Let (X, Y) be a pair of random variables with values in the space X × Y. If the joint distribution is P(X, Y) and your marginal distributions are PX and PY, the mutual information (Ross, 2014) is defined as Global reconstruction of oceanic carbon monoxide emissions 116 8 ( 9;; ) =</0(= ( 2,4 ) ∥=2⨂=4) (S1) where DKL is the Kullback-Leibler divergence. In the case of jointly continuous random variables, it is computed by a double integral: 8 ( 9;; ) = ∬ = ( 2,4 ) (&,B)log/(6 ( ",$ )( 7,8 ) 6" ( 7 ) 6$ ( 8 ) )4&4B (S2) where P(X, Y) is now the joint probability density function of X and Y, and PX and PY are the marginal probability density functions of X and Y respectively. Intuitively, mutual information measures the information that X and Y share, that is, how much knowing one of these variables reduces uncertainty about the other. From Eq. S2, if X and Y are independent, then knowing X does not give any information about Y and vice versa, so their mutual information is zero. We can also observe that if X is a deterministic function of Y and Y is a deterministic function of X, then all the information conveyed by X is shared with Y. As a result, in this case, the mutual information is the same as the uncertainty. We cannot compute this value directly not knowing those distributions. However, we performed an estimation of those values using the method described by (Kraskov et al., 2004) and implemented it in scikitlearn package. For each feature set Sk we pick k features with the highest mutual information between this feature and response. Models used We can divide regressors into two general categories - linear and nonlinear. The former is widely used in all kinds of applications ranging from econometrics to natural sciences. They are easy to interpret and quick to fit. However, they are not suitable for more complicated relationships, for example, y = sin(x). One of the most basic non-linear regressors is the decision tree (Breiman, 2017). Its main idea is to partition the space into smaller regions. We then partition the subdivisions again - this is called recursive partitioning - until we get to the fragments of the space, allowing us to fit simple models to them. Regression trees use the tree to represent the recursive partition. Each of the terminal nodes, or leaves, of the tree represents a cell of the partition and has attached to Global reconstruction of oceanic carbon monoxide emissions 117 it a simple model which applies in that cell only. A point x belongs to a leaf if x falls in the corresponding cell of the partition. Decision trees, even though in some cases they are more powerful than linear models, are still considered weak regressors. However, we can combine a collection of weak learners into a more powerful model. There are two main classes of ensemble learning methods, namely bagging and boosting, although ML algorithms can be a combination of both with certain variations (Friedman, 2001; Bishop, 2006; Murphy, 2012). The bagging method builds models in parallel using a random subset of data (sampling with replacement) and aggregates predictions of all models. The boosting method builds models in sequence using the entire data, with each model improving the previous model error - CatBoost (Dorogush et al., 2018), LightGBM (Ke et al., 2017), and XGBoost (Chen and Guestrin, 2016) are all variations of gradient boosting algorithms. CatBoost is an open-source gradient boosting algorithm, with its name coined from “Category” and “Boosting”. CatBoost builds symmetric (balanced) trees, unlike XG-Boost and LightGBM. In each step, the leaves of the previous tree are split using the same condition. The feature-split pair that accounts for the lowest loss is selected and used for all the level nodes. Even though LightGBM and XGBoost are both asymmetric trees, LightGBM grows leaf-wise, while XGBoost grows level-wise. To put it simply, we can think of LightGBM as the selective growth of the tree, resulting in smaller and faster models compared to XGBoost. Contrary to classic boosting algorithms, which are prone to overfitting on small or noisy datasets, CatBoost uses the concept of ordered boosting, a permutation-driven approach to train the model on a subset of data while calculating residuals on another subset. In CatBoost, a greedy method (Cormen et al., 2009) is used such that a list of possible candidates for featuresplit pairs is assigned to the leaf as the split and the split that results in the smallest penalty is selected. In LightGBM, Gradient-Based One-Side Sampling (GOSS) keeps all data instances with large gradients and performs random sampling for data instances with small gradients. Data points with a larger magnitude of gradients have higher errors and might be important for finding the optimal split point, whereas data points with smaller gradient values would be important for keeping accuracy for learned decision trees. By incorporating this sampling Global reconstruction of oceanic carbon monoxide emissions 124 Figure S5.3: Monthly average fractional sea ice cover of a grid cell climatology was obtained from the Global Modeling and Assimilation Office (GMAO): M2TMNXOCN (V5.12.4). Figure S5.4: Coastal areas (seawater depth < 200 m) (data from NOAA National Centers for Environmental Information, 2022). Figure S5.5: Monthly average 10m windspeed over ice-free ocean (Remote Sensing Systems Version7 Microwave Radiometer Data: 2016). Global reconstruction of oceanic carbon monoxide emissions 125 Figure S5.6: Monthly average aCDOM320 over ice-free ocean (MODIS 2018; Johannessen et al., 2003; Fichot and Miller. 2010). Global reconstruction of oceanic carbon monoxide emissions 126 References Arévalo-Martínez, D. L., Beyer, M., Krumbholz, M., Piller, I., Kock, A., Steinhoff, T., Körtzinger, A., and Bange, H. W.: A new method for continuous measurements of oceanic and atmospheric N2O, CO and CO2: performance of off-axis integrated cavity output spectroscopy (OA-ICOS) coupled to nondispersive infrared detection (NDIR), Ocean Sci., 9, 1071–1087, https://doi.org/10.5194/os-9-1071-2013, 2013. Bates, T. S., Kelly, K. C., Johnson, J. E., and Gammon, R. H.: Regional and seasonal variations in the flux of oceanic carbon monoxide to the atmosphere, J. Geophys. Res., 100, 23093, https://doi.org/10.1029/95JD02737, 1995. Belzile, C., Johannessen, S. C., Gosselin, M., Demers, S., and Miller, W. L.: Ultraviolet attenuation by dissolved and particulate constituents of first-year ice during late spring in an Arctic polynya, Limnol. Oceanogr., 45, 1265–1273, 2000. Bishop, C. M.: Pattern recognition and machine learning. Information science and statistics. New York: Springer, 2006. Blomquist, B. W., Fairall, C. W., Huebert, B. J., and Wilson, S. T.: Direct measurement of the oceanic carbon monoxide flux by eddy correlation, Atmos. Meas. Tech., 5, 3069–3075, https://doi.org/10.5194/amt-5-3069-2012, 2012. Bourbonniere, R. A., Miller, W. L., and Zepp, R. G.: Distribution, flux, and photochemical production of carbon monoxide in a boreal beaver impoundment. Journal of Geophysical Research: Atmospheres, 102(D24), 29321-29329, https://doi.org/10.1029/97JD02234, 1997. Breiman, L.: Random forests, Mach. Learn., 45, 5–32, 2001. Butterworth, B. J., and Miller, S. D.: Air‐sea exchange of carbon dioxide in the Southern Ocean and Antarctic marginal ice zone. Geophysical Research Letters, 43(13), 7223-7230. https://doi.org/10.1002/2016GL069581, 2016. Campen, H. I., Arévalo-Martínez, D. L., and Bange, H. W.: Carbon monoxide (CO) cycling in the Fram Strait, Arctic Ocean, Biogeosciences, 20, 1371–1379, https://doi.org/10.5194/bg-20-1371-2023, 2023. Canadell, J. G., Monteiro, P. M., Costa, M. H., Da Cunha, L. C., Cox, P. M., Alexey, V., ... and Lebehot, A. D.: Global carbon and other biogeochemical cycles and feedbacks, https://doi.org/10.1017/9781009157896.007, 2021 Cicerone, R. J.: How has the atmospheric concentration of CO changed, 1988. Conte, L., Szopa, S., Séférian, R., and Bopp, L.: The oceanic cycle of carbon monoxide and its emissions Global reconstruction of oceanic carbon monoxide emissions 127 to the atmosphere, Biogeosciences, 16, 881–902, https://doi.org/10.5194/bg-16-881-2019, 2019. Conrad, R., Seiler, W., Bunse, G., and Giehl, H.: Carbon monoxide in seawater (Atlantic Ocean), J. Geophys. Res., 87, 8839, https://doi.org/10.1029/JC087iC11p08839, 1982. Daniel, J. S., and Solomon, S.: On the climate forcing of carbon monoxide. Journal of Geophysical Research: Atmospheres, 103(D11), 13249-13260, https://doi.org/10.1029/98JD00822, 1998. Day, D. A. and Faloona, I.: Carbon monoxide and chromophoric dissolved organic matter cycles in the shelf waters of the northern California upwelling system. Journal of Geophysical Research: Oceans, 114(C1), https://doi.org/10.1029/2007JC004590, 2009. Edson, J.B., Fairall, C.W., Bariteau, L., Zappa, C.J., Cifuentes-Lorenzen, A., Mcgillis, W.R., Pezoa, S., Hare, J.E., Helmig, D.: Direct covariance measurement of CO2 gas transfer velocity during the 2008 Southern Ocean Gas Exchange Experiment: wind speed dependency. J. Geophys. Res. 116 (C00F10). http://dx.doi.org/10.1029/ 2011JC007022, 2011. Erickson, D. J.: Ocean to atmosphere carbon monoxide flux: Global inventory and climate implications, Global Biogeochem. Cy., 3, 305–314, https://doi.org/10.1029/GB003i004p00305, 1989. Fichot, C. G., and Miller, W. L.: An approach to quantify depth-resolved marine photochemical fluxes using remote sensing: Application to carbon monoxide (CO) photoproduction. Remote Sensing of Environment, 114(7), 1363-1377, https://doi.org/10.1016/j.rse.2010.01.019, 2010. Forster, P., Storelvmo, T., Armour, K., Collins, W., Dufresne, J.-L., Frame, D., Lunt, D.J., Mauritsen, T., Palmer, M.D., Watanabe, M., Wild, M., and Zhang, H.: The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity, in: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J.B.R., Maycock, T.K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou. B., Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 923–1054, https://doi.org/10.1017/9781009157896.009, 2021. Friedman, J. H.: Greedy function approximation: a gradient boosting machine. Annals of statistics, 11891232, https://www.jstor.org/stable/2699986, 2001. Garcia, H. E., Weathers, K., Paver, C. R., Smolyar, I., Boyer, T. P., Locarnini, R. A., Zweng, M. M., Mishonov, A. V., Baranova, O. K., Seidov, D., and Reagan, J. R.: World Ocean Atlas 2018, Volume 3: Dissolved Oxygen, Apparent Oxygen Utilization, and Oxygen Saturation. A. Mishonov Technical Ed.; NOAA Atlas NESDIS 83, 38pp, https://archimer.ifremer.fr/doc/00651/76337, 2018. Garcia, H. E., Weathers, K., Paver, C. R., Smolyar, I., Boyer, T. P., Locarnini, R. A., Zweng, M. M., Mishonov, A. V., Baranova, O. K., Seidov, D., and Reagan, J. R.: 2018. World Ocean Atlas 2018, Volume 4: Dissolved Inorganic Nutrients (phosphate, nitrate and nitrate+nitrite, silicate). A. Mishonov Technical Ed.; NOAA Atlas NESDIS 84, 35pp, https://archimer.ifremer.fr/doc/00651/76336, 2018. Global reconstruction of oceanic carbon monoxide emissions 128 Gaubert, B., Worden, H. M., Arellano, A. F. J., Emmons, L. K., Tilmes, S., Barré, J., ... and Edwards, D. P.: Chemic al f eed ba ck f rom decre asi ng c ar bon mono xide e mi ssi on s. G eophys ica l Resea rch Let ters, 44 , 9985–9995. https://doi.org/10.1002/ 2017GL074987, 2017. Global Modeling and Assimilation Office (GMAO), MERRA-2 tavgM_2d_ocn_Nx: 2d, Monthly mean, Time-Averaged, Single-Level, Assimilation, Ocean Surface Diagnostics V5.12.4, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GES DISC), Accessed: [09.2023], 10.5067/4IASLIDL8EEC, 2015. Greening, C. and Grinter, R.: Microbial oxidation of atmospheric trace gases, Nature Rev. Microbiol., 20, 513-528, https://doi.org/10.1038/s41579-022-00724-x, 2022.Thompson, A. M.: The oxidizing capacity of the earth’s atmosphere – probable past and future changes, Science, 256, 1157–1165, 1992. Johannessen, S. C., Miller, W. L., and Cullen, J. J.: Calculation of UV attenuation and colored dissolved organic matter absorption spectra from measurements of ocean color, Journal of Geophysical Research: Oceans, 108(C9), https://doi.org/10.1029/2000JC000514, 2003. Kerkar, A. U., Tripathy, S. C., Minu, P., Baranval, N., Sabu, P., Patra, S., ... and Sarkar, A.: Variability in primary productivity and bio-optical properties in the Indian sector of the Southern Ocean during an austral summer. Polar Biology, 43, 1469-1492, https://doi.org/10.1007/s00300-020-02722-2, 2020. Kwon, Y. S., Baek, S. H., Lim, Y. K., Pyo, J., Ligaray, M., Park, Y., and Cho, K. H.: Monitoring coastal chlorophyll-a concentrations in coastal areas using machine learning models. Water, 10(8), 1020, https://doi.org/10.3390/w10081020, 2018. Linnenbom, V. J., Swinnerton, J. W., and Lamontagne, R. A.: The ocean as a source for atmospheric carbon monoxide, J. Geophys. Res., 78, 5333–5340, https://doi.org/10.1029/JC078i024p05333, 1973. Liss, P. S., Marandino, C. A., Dahl, E. E., Helmig, D., Hintsa, E. J., Hughes, C., ... and Williams, J.: Short-lived trace gases in the surface ocean and the atmosphere, In Ocean-Atmosphere Interactions of Gases and Particles (pp. 1-54). Springer, Berlin, Heidelberg, 2014. Locarnini, M. M., Mishonov, A. V., Baranova, O. K., Boyer, T. P., Zweng, M. M., Garcia, H. E., Reagan, J. R., Seidov, D., Weathers, K. W., Paver, C. R., Smolyar, I.: World Ocean Atlas 2018, Volume 1: Temperature. NOAA Atlas NESDIS 81, 52pp. https://archimer.ifremer.fr/doc/00651/76338, 2018. Moran, M. A. and Miller, W. L.: Resourceful heterotrophs make the most of light in the coastal ocean, Nat. Rev. Microbiol., 5, 792– 800, 2007. Murphy, K. P.: Machine learning: a probabilistic perspective. Adaptive computation and machine learning series. MIT Press, Cambridge, MA, 2012. NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology Processing Group. Global reconstruction of oceanic carbon monoxide emissions 129 Moderate-resolution Imaging Spectroradiometer (MODIS) Aqua Inherent Optical Properties Data; 2018 Reprocessing. NASA OB. DAAC, Greenbelt, MD, USA. https://doi: 10.5067/AQUA/MODIS/L3M/IOP/2018. Accessed on 07/20/2022. NOAA National Centers for Environmental Information. 2022: ETOPO 2022 15 Arc-Second Global Relief Model. NOAA National Centers for Environmental Information. DOI: 10.25921/fd45-gt74. Accessed Accessed on 05/08/2023. Nightingale, P.D., Malin, G., Law, C.S., Watson, A.J., Liss, P.S., Liddicoat, M.I., Boutin, J., UpstillGoddard, R.C.: In situ evaluation of air-sea gas exchange parameterizations using novel conservative and volatile tracers. Global Biogeochem Cycles 14 (1), 373–387. https://doi.org/10.1029/1999 gb900091, 2000. Park, K. and Rhee, T. S.: Source characterization of carbon monoxide and ozone over the Northwestern Pacific in summer 2012, Atmospheric Environment, 111, 151–160, https://doi.org/10.1016/j.atmosenv.2015.04.015, 2015. Park, K. and Rhee, T. S.: Oceanic source strength of carbon monoxide on the basis of basin-wide observations in the Atlantic, Environmental Science-Processes and Impacts, 18, 104-114, 10.1039/c5em00546a, https://doi.org/10.1039/C5EM00546A, 2016. Petron, G., Crotwell, A. M., Crotwell, M. J., Dlugokencky, E., Madronich, M., Moglia, E., Neff, D, Thoning, K., Wolter, S., Mund, J. W.: Atmospheric Carbon Monoxide Dry Air Mole Fractions from the NOAA GML Carbon Cycle Cooperative Global Air Sampling Network, 1988-2022, Version: 2023-0828, https://doi.org/10.15138/33bv-s284, 2023. Pfeifroth, U., Kothe, S., Müller, R., Trentmann, J., Hollmann, R., Fuchs, P., Kaiser, J., and Werscheck, M.: Surface Radiation Data Set - Heliosat (SARAH) - Edition 2.1, Satellite Application Facility on Climate Monitoring, DOI:10.5676/EUM_SAF_CM/SARAH/V002_01, https://doi.org/10.5676/EUM_SAF_CM/SARAH/V002_01, 2019. Remote Sensing Systems, 2016, updated 2021. Monthly Mean Wind Speed Data Set on a 1-degree grid made from Remote Sensing Systems Version-7 Microwave Radiometer Data, V0701, 07/20/2022. Santa Rosa, CA, USA. Available at www.remss.com. Seiler, W.: The influence of the biosphere on the atmospheric CO and H2 cycles, in: Environmental Biogeochemistry and Geomicrobiology, edited by: Krumbein, W. E., vol. 3, Ann Arbor Science Publishers, Ann. Arbor, 773–810, 1978. Sherwen, T., Chance, R. J., Tinel, L., Ellis, D., Evans, M. J., and Carpenter, L. J.: A machine-learningbased global sea-surface iodide distribution, Earth Syst. Sci. Data, 11, 1239–1262, https://doi.org/10.5194/essd-11-1239-2019, 2019. Stubbins, A., Uher, G., Law, C. S., Mopper, K., Robinson, C., and Upstill-Goddard, R. C.: Open-ocean Global reconstruction of oceanic carbon monoxide emissions 130 carbon monoxide photoproduction, Deep Sea Research Part II: Topical Studies in Oceanography, 53(1416), 1695-1705, http://dx.doi.org/10.1016/j.dsr2.2006.05.011, 2006. Stubbins, A., Uher, G., Kitidis, V., Law, C., Upstill-Goddard, R.C., Woodward, E.M.S.: The open-ocean source of atmospheric carbon monoxide, Deep Sea Res. 53, 1685e1694, https://doi.org/10.1016/j.dsr2.2006.05.010, 2006 Stubbins, A., Law, C. S., Uher, G., and Upstill-Goddard, R. C.: Carbon monoxide apparent quantum yields and photoproduction in the Tyne estuary, Biogeosciences, 8, 703–713, https://doi.org/10.5194/bg8-703-2011, 2011. Sugai, Y., Tsuchiya, K., Shimode, S., and Toda, T.: Photochemical Production and Biological Consumption of CO in the SML of Temperate Coastal Waters and Their Implications for Air‐Sea CO Exchange, Journal of Geophysical Research: Oceans, 125(4), e2019JC015505, https://doi.org/10.1029/2019JC015505, 2020. Takahashi, T., Sutherland, S. C., Sweeney, C., Poisson, A., Metzl, N., Tilbrook, B., ... and Nojiri, Y.: Global sea–air CO2 flux based on climatological surface ocean pCO2, and seasonal biological and temperature effects. Deep Sea Research Part II: Topical Studies in Oceanography, 49(9-10), 1601-1622, https://doi.org/10.1016/S0967-0645(02)00003-6, 2002. Tao, F., Huang, Y., Hungate, B. A., ... and Luo Y.: Microbial carbon use efficiency promotes global soil carbon storage, Nature, https://doi.org/10.1038/s41586-023-06042-3, 2023. Underwood, K. L., Rizzo, D. M., Hanley, J. P., Sterle, G., Harpold, A., Adler, T., ... and Perdrial, J. N.: Machine‐learning reveals equifinality in drivers of stream DOC concentration at continental scales. Water Resources Research, e2021WR030551, https://doi.org/10.1029/2021WR030551, 2023. Watson Gregg and Cecile Rousseaux (2017), NASA Ocean Biogeochemical Model assimilating satellite chlorophyll data global monthly VR2017, edited by Watson Gregg and Cecile Rousseaux, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GES DISC), 10.5067/BHCFDIICIOU5, Accessed: 07/20/2022 Weber, T., Wiseman, N.A. and Kock, A.: Global ocean methane emissions dominated by shallow coastal waters, Nat Commun 10, 4584, https://doi.org/10.1038/s41467-019-12541-7, 2019. Westwood, K. J., Griffiths, F. B., Webb, J. P., Wright, S. W.: Primary production in the Sub-Antarctic and Polar Frontal zones south of Tasmania, Australia; SAZ-Sense survey, 2007. Deep Res Part II 58(21– 22):2162–2178, https://doi.org/10.1016/j. dsr2.2011.05.017, 2011. Wyrtki, K.: An Estimate of Equatorial Upwelling in the Pacific, J. Phys. Oceanogr., 11, 1205–1214, https://doi.org/10.1175/15200485(1981)011<1205: AEOEUI>2.0.CO;2, 1981. Xie, H., Bélanger, S., Demers, S., Vincent, W. F., and Papakyriakou, T. N.: Photobiogeochemical cycling Global reconstruction of oceanic carbon monoxide emissions 131 of carbon monoxide in the southeastern Beaufort Sea in spring and autumn, Limnology and Oceanography, 54(1), 234-249, https://doi.org/10.4319/lo.2009.54.1.0234, 2009. Xu, G. B., Xu, F., Ji, X., Zhang, J., Yan, S. B., Mao, S. H., and Yang, G. P.: Carbon monoxide cycling in the Eastern Indian Ocean. Journal of Geophysical Research: Oceans, 128, e2022JC019411, https://doi.org/10.1029/2022JC019411, 2023. Yang, G. P., Wang, W. L., Lu, X. L., and Ren, C. R.: Distribution, flux and biological consumption of carbon monoxide in the Southern Yellow Sea and the East China Sea, Marine Chemistry, 122(1‐4), 74– 82. https://doi.org/10.1016/j.marchem.2010.08.001, 2010. Yang, S., Chang, B. X., Warner, M. J., Weber, T. S., Bourbonnais, A. M., Santoro, A. E., ... and Bianchi, D.: Global reconstruction reduces the uncertainty of oceanic nitrous oxide emissions and reveals a vigorous seasonal cycle. Proceedings of the National Academy of Sciences, 117(22), 11954-11960, https://doi.org/10.1073/pnas.1921914117, 2020. Zafiriou, O. C., Andrews, S. S., and Wang, W.: Concordant estimates of oceanic carbon monoxide source and sink processes in the Pacific yield a balanced global “blue‐water” CO budget, Global Biogeochemical Cycles, 17(1), https://doi.org/10.1029/2001GB001638, 2003. Zafiriou, O. C., Xie, H., Nelson, N. B., Najjar, R. G., Wang, W.: Diel carbon monoxide cycling in the upper Sargasso Sea near Bermuda at the onset of spring and in midsummer, Limnol. Oceanogr. 53, 835– 850, https://doi.org/10.4319/lo.2008.53.2.0835, 2008. Zeng, J., Matsunaga, T., Saigusa, N., Shirai, T., Nakaoka, S. I., and Tan, Z. H.: Evaluation of three machine learning models for surface ocean CO 2 mapping. Ocean Science, 13(2), 303-313, https://doi.org/10.5194/os-13-303-2017, 2017. Zhang, J., Wang, J., Zhuang, G. C., and Yang, G. P.: Carbon monoxide cycle in the Bohai Sea and the Yellow Sea: Spatial variability, sea‐air exchange, and biological consumption in autumn, Journal of Geophysical Research: Oceans, 124, 4248–4257, https://doi.org/10.1029/2018JC014864, 2019. Zheng, B., Chevallier, F., Yin, Y., Ciais, P., Fortems-Cheiney, A., Deeter, M. N., ... and Zhao, Y: Global atmospheric carbon monoxide budget 2000–2017 inferred from multi-species atmospheric inversions, Earth System Science Data, 11(3), 1411-1436.https://doi.org/10.5194/essd-11-1411-2019, 2019. Ziolkowski, L. A., and Miller, W. L.: Variability of the apparent quantum efficiency of CO photoproduction in the Gulf of Maine and Northwest Atlantic, Marine Chemistry, 105(3-4), 258-270, https://doi.org/10.1016/j.marchem.2007.02.004, 2007. Zweng, M. M., Reagan, J. R., Seidov, D., Boyer, T. P., Locarnini, R. A., Garcia, H. E., Mishonov, A. V., Baranova, O. K., Weathers, K., Paver, C. R., and Smolyar, I.: World Ocean Atlas 2018, Volume 2: Salinity. A. Mishonov Technical Ed.; NOAA Atlas NESDIS 82, 50pp, https://archimer.ifremer.fr/doc/00651/76339, 2019. Global reconstruction of oceanic carbon monoxide emissions 132 Conclusions and Outlook 133 6 Conclusions and Outlook Covering measurements from a lagoon system, a coastal time-series, and modelling the global CO distribution in the ocean, the results presented in the thesis offer fresh insights into the understanding the ocean cycling of the climate-relevant trace gas CO. Accurately characterizing its variabilities and the controlling factors in the water column is key to quantifying CO emissions, estimating its impact on the climate and, in turn, gaining some clues of future CO trends. Each chapter addresses one research question outlined in this thesis and, subsequently, the provide responses to the questions posed in this thesis, along with corresponding further study suggestions: 1) What are the sources and sinks of carbon monoxide in the lagoon system (Ria Formosa Lagoon) and is the aquaculture area a significant source of CO emissions? Chapter 2: The first dissolved CO surface data measured in the Ria Formosa Lagoon system is reported in this study. CO cycling in the Ria Formosa Lagoon system is aligned with observations from other coastal regions. The waters within the lagoon exhibited CO supersaturation, establishing Ria Formosa as a source of atmospheric CO. Dissolved CO concentrations were closely correlated with FDOM, pointing to photochemical processes as the major contributors to CO concentrations in Ria Formosa. Notably, potential CO sources like phytoplankton and dark production were found to have a minor importance. The primary sink for CO was identified as microbial consumption, overshadowing comparatively minor emissions to the atmosphere. Additionally, the production of CO from effluent waters near the study area originating from an aquaculture facility was minimal. This implies that aquaculture facilities in this region do not make a substantial contribution to atmospheric CO levels.