Edited by: Laura Anne Bristow, University of Southern Denmark, Denmark
Reviewed by: Frances Hopkins, Plymouth Marine Laboratory, United Kingdom; Colin Murrell, University of East Anglia, United Kingdom
†Present address: Cleo L. Davie-Martin, Terrestrial Ecology Section, Department of Biology, University of Copenhagen, Copenhagen, Denmark
This article was submitted to Marine Biogeochemistry, a section of the journal Frontiers in Marine Science
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Marine-derived volatile organic compounds (VOCs) influence global carbon cycling, atmospheric reactions, and climate. Yet, the biogenic production (sources) and consumption (sink) rates of marine VOCs are not well-constrained and are currently excluded from global chemical transport models. We directly measured the net biogenic production rates of seven VOCs (acetaldehyde, acetone, acetonitrile, dimethylsulfide, isoprene, methanethiol, and methanol) in surface seawater during four field campaigns in the North Atlantic Ocean that targeted different stages of the phytoplankton annual cycle. All of the VOCs exhibited strong seasonal trends, with generally positive rates during May (peak spring bloom) and lower, sometimes negative rates (net consumption), during November and/or March (the winter bloom minimum transition). Strong latitudinal gradients were identified for most VOCs during May and September, with greater production observed in the northern regions compared to the southern regions. These gradients reflect the interplay between high phytoplankton and bacterial productivity. During the bloom transition stages (March and September), acetaldehyde and acetone exhibited net production rates that bracketed zero, suggesting that biogenic production was either very low or indicative of a tightly coupled system with more complex underlying microbial VOC cycling. Our data provides the first direct evidence for widespread biogenic acetonitrile production and consumption in the surface ocean and the first net biogenic production rates for methanethiol in natural seawater.
Volatile organic compounds (VOCs), such as isoprene and acetaldehyde, have high vapor pressures (
Volatile organic compounds measured in algal cultures vary with the species employed and environmental conditions (
Marine-derived VOCs released to the atmosphere constitute a loss of photosynthetically fixed carbon. The contributions of biogenic VOCs to the marine carbon cycle are not well constrained and are often ignored in global chemical transport models. New information about the ecology of VOC production in the marine surface ocean is needed to advance models for which the underlying biological source terms for VOC air-sea exchange are unknown. Here, we present new approaches that couple proton-transfer-reaction time-of-flight mass spectrometry (PTR-ToF/MS) with incubation chambers to measure net rates of biogenic VOC production in natural seawater suspensions. Our methods build on a previously published approach (
The NAAMES study region in the Western North Atlantic Ocean covers a wide range of both temporal and spatial ecosystem variability (18–56°N and 71–37°W). We assessed biologically mediated VOC production during different stages of the annual phytoplankton bloom through field campaigns in November 2015, May 2016, September 2017, and March 2018 (
Cruise tracks and station locations (stars) in the Western North Atlantic Ocean during the four North Atlantic Aerosols and Marine Ecosystem Study (NAAMES) field campaigns. The main sampling intensives occurred along a north-to-south transect at approximately 40°W. During March 2018, we departed from San Juan, Puerto Rico instead of Woods Hole, MA, United States.
Natural seawater suspensions (100 mL) were incubated in 200 mL polycarbonate dynamic stripping chambers (
A PTR-TOF-1000 (IONICON Analytik GmbH, Innsbruck, Austria) was used to quantify the net biogenic production rates of seven target VOCs (acetaldehyde, acetone, acetonitrile, DMS, isoprene, methanethiol, and methanol) that are known to be present in marine environments and produced and/or consumed by marine plankton. PTR-ToF/MS data were acquired at 1 or 5 s intervals across the mass range 18–363 a.m.u. The PTR-ToF/MS can acquire data for only one chamber at a time and as such, a flow-through multiport valve (RJ Lee Group Inc., Monroeville, PA, United States) was used to sequentially switch between chambers and divert the headspace sample from the selected chamber into the PTR-ToF/MS inlet for 60 s each. Chambers were bubbled for 2 min prior to data acquisition to flush out the initial headspace and allow the bubbling to stabilize after activation of the air flow. Initiation of bubbling was also staggered by 1 min intervals so that each chamber was sampled at the same time periods relative to initial data acquisition. All sample lines between the incubators and the PTR-ToF/MS inlet were heated to approximately 60°C to prevent condensation within the tubing.
The fundamentals of PTR-ToF/MS technology have been described elsewhere (
Background concentrations of VOCs arising from the experimental setup were determined by monitoring the VOCs evolved from autoclaved, bubbled artificial seawater (ASW, i.e., “VOC-free” seawater) at different temperatures (
Flow chart showing the calculation steps to obtain net VOC production rates. The flowchart is illustrated with hypothetical data. For example, in step 1, a background acetone concentration of 1.77 ppbv was calculated for a surface seawater sample collected at 15°C (yellow shading). Temperature-dependent background concentrations determined from measurements of autoclaved and bubbled (“VOC-free”) artificial seawater (blue line) are subtracted from sample concentrations (green line) and abiotic control concentrations (gray line) to account for signals arising from the experimental setup. VOC production rates from abiotic controls are subtracted from the sample VOC production rates to account for the physical removal (“stripping”) of VOCs caused by bubbling, yielding the net biogenic production rate (yellow arrow).
Linear temperature regressions used to derive background concentrations (ppbv) of target VOCs.
VOC | Linear regression for predicting background concentrations | LOD at 15°C (ppbv) |
Acetaldehyde | [Ald] = 0.0890 × T + 0.436 ( |
1.8 |
Acetone | [Ace] = 0.297 × T – 2.75 ( |
1.7 |
Acetonitrile | [MeCN] = 0.00937 × T – 0.0276 ( |
0.11 |
Dimethylsulfide | [DMS] = 0.0477 × T – 0.0762 ( |
0.64 |
Isoprene | [Isop] = 0.0542 × T + 0.235 ( |
1.0 |
Methanethiol | [MeSH] = 0.0109 × T – 0.0291 ( |
0.13 |
Methanol | [MeOH] = 0.246 × T + 3.20 ( |
6.9 |
Production rates (PR, nmol L–1 h–1) for samples and abiotic controls were calculated as:
where
Sample replicates (where
Gas standards containing ∼1 ppm of acetaldehyde, acetone, DMS, and isoprene were obtained from Matheson Tri-Gas (Newark, CA, United States). Serial dilutions were performed with synthetic air that had passed through a dynamic stripping chamber filled with 100 mL of ASW media (to mimic the viscosity and relative humidity produced during the bubbling of natural seawater samples). The full dilution series (0–160 ppbv) was used to test the dynamic linear range of the instrument approximately once every fortnight. Two-point calibration curves (0 and 80–120 ppbv) were run daily to calculate multiplier factors for each target VOC from the inverse slope of the linear regression between measured and expected concentrations (
R Studio version 1.1.463 (Boston, MA, United States) was used for all statistical analyses, including one-way analysis of variance (ANOVA) to assess seasonal trends in net VOC production rates and pairwise
Additional physical, chemical, and biological measurements and the underway ship data collected across all four NAAMES campaigns [e.g., net primary production (NPP), dissolved organic carbon (DOC), photosynthetically active radiation (PAR), chlorophyll-a etc.] are publically available at: doi:
Here we report net rates of biogenic VOC production from natural seawater communities (including phytoplankton and bacterioplankton). Positive net rates indicate that the biogenic production or release of VOCs by the population as a whole (directly via metabolism or indirectly due to diffusion or cell lysis) are greater than the biological losses, which may be due to microbial consumption (assimilation into biomass for growth and/or microbial oxidation to carbon dioxide for energy) or adsorption/diffusive processes. Negative net rates mean that the loss term outpaced the production term. Our abiotic controls account for physical processes leading to VOC production (e.g., photodegradation of CDOM) and losses (e.g., removal via experimental bubbling). It is also important to keep in mind that the magnitude of the net biogenic production rate is not necessarily in agreement with the magnitude of biological cycling or directly comparable between different VOCs. For example, a small, positive net production rate could be indicative of slow biogenic production or a highly active biological VOC cycle, in which the biogenic production rate is essentially equivalent to the biological loss rate. We refer to the latter phenomenon as “tightly coupled,” which results in a net biogenic production rate close to zero. Large deviations from zero in the net production rates indicate that the system is uncoupled (i.e., biogenic VOC production is > or < biological consumption). It is also important to note that negative net production rates do not preclude VOC release to the atmosphere via physical and abiotic drivers. In later sections, we explore the coupled state of VOC production in the North Atlantic Ocean and environmental factors that influence the balance between the biological VOC production and consumption terms.
Rates of net biogenic VOC production in surface waters were determined across wide spatial and temporal scales in the North Atlantic Ocean.
Box and whisker plots showing the range, median, and interquartile range of net VOC production rates for each of four field campaigns that targeted different stages of the annual phytoplankton bloom. Outliers displayed as independent points reflect measurements greater than 1.5 times the interquartile range above and below the median. The number of measurements for each VOC in each month is displayed at the top of each box. These measurements were made from a total of 14, 56, 50, and 74 experiments in November, March, May, and September, respectively. Rates of VOC production that showed no statistical difference (
Measurements of net production rates were rare during November for most targeted VOCs, as well as for acetaldehyde and acetone during March, because the sample VOC concentrations were lower than background concentrations in >50% of measurements within the 1 h incubation period. These rates were deemed below the limits of detection for our system and not calculated. In addition, we only report herein measurements carried out using the same experimental protocols. Because November was the first of the NAAMES campaigns and the first opportunity to employ our incubation system at sea, we experimented with different bubbling rates and light conditions. Thus, those experimental data are not considered, causing the coverage of net VOC production rates reported during November to be greatly reduced (14 sample measurements) compared to the other months (50–74 sample measurements).
One-way ANOVA revealed statistically significant seasonal differences in net biogenic production rates for all targeted VOCs (
Our four field campaigns covered wide spatial gradients in the North Atlantic Ocean from 18 to 56°N and 37 to 71°W.
Location of measurements and relative net production rates for target VOCs during the four field campaigns in the North Atlantic Ocean. Net production rates colored purple were negative, yellow were positive, or shown as black crosses when below the experimental limit of detection (n.d.). The size of the bubbles represents the absolute magnitude of the net production rate normalized to the maximum absolute net production rate for each VOC (which were 14.4, 80.2, 1.42, 43.3, 26.3, 7.86, and 859 nmol L− 1 h− 1 for acetaldehyde, acetone, acetonitrile, DMS, isoprene, methanethiol, and methanol, respectively).
A number of spatial patterns emerge in the sign and magnitude of net VOC production rates. Net acetaldehyde production was higher in May at more northerly stations, whereas we primarily observed negative net rates in the easterly and southern locations. Acetone, DMS, and isoprene net production rates were also generally higher in the more northern latitudes during May. These three compounds were generally below the detection limits of our system at more southerly locations. Trends in net acetonitrile production were more strongly determined by seasonal differences than geography, except during September, where net production was positive toward coastal shelf waters and negative in the open ocean to the east. Net methanethiol production was more spatially consistent (largely positive net rates), except during the west-to-east transect in March, where negative production rates were observed. There was no apparent spatial variability in net methanol production. Next, we examine trends in areas of spatial overlap along the west-to-east longitudinal and north-to-south latitudinal gradients.
Spearman’s rank outcomes (
VOC | Longitudinal correlations |
Latitudinal correlations |
||||||||
November |
March |
September |
May |
September |
||||||
Acetaldehyde | −0.20 | 0.9 | 0.60 | 0.4 | −0.38 | 0.3 | ||||
Acetone | − | − | − | − | − | − | −0.23 | 0.4 | ||
Acetonitrile | 0.20 | 0.5 | −0.21 | 0.6 | ||||||
Dimethylsulfide | − | − | 0.52 | 0.2 | ||||||
Isoprene | −0.40 | 0.8 | − | − | −0.13 | 0.7 | ||||
Methanethiol | 0.40 | 0.2 | 0.11 | 0.7 | ||||||
Methanol | 0.08 | 0.9 | − | − | −0.42 | 0.1 |
During the September field campaign, we occupied the northernmost station (S6 at 53.376°N, −39.542°W) for 97 h and performed round-the-clock measurements to investigate diel variability in VOC production rates.
Timeseries showing diel variability in net VOC production rates during a 97-h occupation at the northernmost station (53.376°N, –39.542°W) during the September field campaign. Net production rates were normalized to the highest absolute net production rate observed for each target VOC during the long-term station. The horizontal dotted line represents a net production rate of zero and the gray shading indicates periods of darkness (∼21:00 to 08:00 UTC). The bottom panel shows the photosynthetically active radiation (PAR, μmol photons m− 2 s− 1) and seawater temperature (Tseawater, °C), also normalized to their highest values (1657 μmol photons m− 2 s− 1 and 12.1°C, respectively).
Net methanol production was very high during day 2 at station S6, with a peak during midday that was, in fact, the highest net methanol production rate recorded during the four NAAMES campaigns. During days 3 and 5, net methanol production appeared to reach a steady net production rate of approximately 100 nmol L–1 h–1, but net production increased again during nighttime on day 4. Acetone exhibited strong net production during the night and again in the early morning of day 2, but these higher production rates were not repeated throughout subsequent days on station. Instead, net acetone production stabilized near zero. Acetonitrile displayed less diel variability than the other target VOCs, with generally small negative net production rates that gradually decreased over the 5 days. Trends in diel isoprene cycling are difficult to interpret due to the lack of consistent measurements above detection limit, but it appears that net isoprene production rates decreased gradually during days 1–3, with higher afternoon production on days 3 and 5.
Net biogenic production rates for seven target VOCs varied widely across time and space in the North Atlantic Ocean, but also exhibited patterns that reveal the ecological interplay between phytoplankton emission and bacterioplankton consumption processes. Net production rates for the seven VOCs were sometimes negative, sometimes positive, and sometimes balanced, such that they were very close to zero. Thus, the plankton community in the surface ocean can act as both a net source and net sink of VOCs to/from the atmosphere, which is consistent with previous production measurements on a subset of the VOCs we measured: acetaldehyde, acetone, DMS, and methanol (
Significant seasonal variations in net production rates were observed for all targeted VOCs (
The drivers of spatial variability in VOC production are intertwined and reflect contributions from a variety of factors, such as temperature, light intensity, water mass circulation, and shifts in community composition, to name a few. The spatial gradients within each field campaign can also help reveal the relationship between VOC production and the seasonality of the massive phytoplankton bloom as we transited from north to south (or vice versa).
In this section, we review the biological context for the net production and/or consumption rates (i.e., biological cycling) of each targeted VOC and relate these activities to the temporal (seasonal/diel) and spatial gradients encountered during the NAAMES project.
Biogenic acetaldehyde production was initially inferred from observations above phytoplankton cultures (
Despite methodological differences, our finding that net microbial consumption of acetaldehyde often outcompetes its production is similar to previous observations (
It is tempting to consider that specific members of the phytoplankton community may impact the balance of the acetaldehyde cycle. The relative abundances of pico- and nano-eukaryotic cells dominated during May, and the latter is significantly and positively correlated with net acetaldehyde production (
At the northernmost station during September, net acetaldehyde production rates peaked between dawn and midday, and returned to a more tightly coupled state with rates close to zero during the night, particularly during the latter half of our long-term occupation. Because biogenic acetaldehyde production increases with light intensity and photochemical processes (both abiotic and biological), net acetaldehyde production likely outpaces its consumption during the first half of the daylight hours (
Numerous studies provide evidence of biogenic marine acetone production, by both phytoplankton (direct, light-dependent production) (
The strong north-to-south gradient we observed in net acetone production rates during May can be used to develop hypotheses about the factors controlling acetone production in the North Atlantic. Rates of net acetone production were always positive in May, ranging from a mean of 3.2 nmol L–1 h–1 in the south (station 5) to 37.3 nmol L–1 h–1 in the northern latitudes (stations 0). Despite the higher production rates, seawater concentrations of acetone were, in fact, lower in the northern region (
The lower net acetone production rates observed in the southern stations during May are associated with lower chlorophyll-a concentrations and higher seawater acetone and DOC (including amino acid substrates) concentrations compared to the northern stations (
The role of acetonitrile in the surface ocean has been largely overlooked. However, atmospheric acetonitrile concentrations (whose sources are dominated by terrestrial biomass burning) have been shown to rapidly decline in the marine boundary layer, leading to suppositions of a biogenic oceanic sink for acetonitrile (
Net acetonitrile production rates exhibited statistically significant seasonal trends, whereby observations were highest and positive during May and November, and lowest and most often negative during the bloom transition months of March and September. We postulate that acetonitrile is produced through bacterial metabolism of DOC, including protein-rich substrates. Interestingly, and in keeping with this protein-degradation hypothesis, rates of net acetonitrile production were higher nearer the coast than in the open ocean during September. Additionally, net acetonitrile production was more positive toward the more nutrient rich northern regions during May and September (
Dimethylsulfide is a well-known, climate-active gas implicated in the formation of cloud condensation nuclei. It is the most prevalent and well-studied marine-derived sulfur compound and the ocean represents a major source of DMS (biogenic sulfur) to the atmosphere. In marine ecosystems, the ubiquitous precursor, dimethylsulfoniopropionate (DMSP), is produced and/or released by phytoplankton and bacteria, for regulation of osmotic pressure and protection against multiple stressors (e.g., light, temperature, nitrogen limitation, viral lysis, and grazing by zooplankton) (
Our data shows that biogenic production of DMS dominates over microbial consumption, particularly during the peak phytoplankton bloom in May and in the northern regions (
Light-dependent biogenic production of isoprene has been previously observed in cultures of marine phytoplankton (
We speculate that the colder waters and phytoplankton bloom conditions (e.g., high chlorophyll-a concentrations and actively growing biomass – see
It was recently revealed that the abundant marine bacterium,
Our data provide direct evidence of biogenic methanethiol production by natural seawater populations, throughout all stages of the phytoplankton annual cycle. To the best of our knowledge, there is currently no evidence for direct methanethiol production by phytoplankton and we theorize that the net production we observe is most likely due to microbial degradation of DMSP. We observed higher net biogenic methanethiol production during November than in March and May. During March and September, we observed a longitudinal gradient whereby net biogenic methanethiol production increased away from the coast toward the open ocean. We also observed a latitudinal gradient during September, with higher net methanethiol production toward the north. Concentrations of DOC, of which DMSP is a large component (
The foundations for biological methanol cycling have been well-documented in recent years. Modeling and field measurements provide indications of a large biogenic source of methanol (
The data we present here suggest that methanol production rates were dependent on phytoplankton production, with methanol consumption causing the system to become a methanol sink with negative production rates during November and March. Low phytoplankton abundance, growth rates, and productivity during November (
We present net biogenic production rates for seven VOCs in natural seawater communities during four field campaigns in the North Atlantic Ocean that targeted different stages of the annual phytoplankton biomass cycle. We observed seasonal and spatial trends related to bloom dynamics, particularly as we transitioned from north-to-south during May and September. These findings highlight the roles of plankton communities and interactions among phytoplankton and bacteria in determining VOC accumulation and production rates. We provide the first direct evidence for biogenic acetonitrile production and consumption, and postulate mechanisms that could explain acetonitrile as a product and substrate of carbon-cycling activity of bacterioplankton. Our data illuminate multi-domain interactions among marine phytoplankton and bacterioplankton that contribute to the cycling of VOCs in the surface ocean and indicate their potential to modulate VOC release to the atmosphere. Further work is warranted to gain a more complete picture of the mechanisms governing VOC cycling in marine surface waters, the nutrient regimes or conditions that determine which carbon sources are preferentially consumed, and the magnitude of the bulk VOC pool cycled among plankton and transferred across the sea-air interface.
The datasets presented in this study can be found in online repositories. Pre-processed PTR-ToF/MS and ancillary datasets from the four NAAMES campaigns are publicly available at:
CD-M carried out on-board measurements and performed all data analysis and preparation of the first manuscript draft. WP and KH helped with on-board measurements. KH and SG conceived the original idea. MB lead the North Atlantic Aerosols and Marine Ecosystems Study (NAAMES) and was in charge of all field campaigns and project management. All the authors reviewed and contributed to the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We would like to thank Armin Wisthaler, for his thoughtful comments on an earlier version of the manuscript, and the NAAMES science team and crew aboard the
The Supplementary Material for this article can be found online at: