Methane Bubble Size Distributions, Flux, and Dissolution in a

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Article

Methane bubble size distributions, flux, and dissolution in a freshwater lake Kyle Delwiche, and Harold F. Hemond Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b04243 • Publication Date (Web): 08 Nov 2017 Downloaded from http://pubs.acs.org on November 12, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 17

Environmental Science & Technology

25 26 27

Introduction

28 29 30 31 32 33 34 35

Freshwater lakes and reservoirs are a significant source of methane to the atmosphere, and methane is a potent greenhouse gas with large implications for future climate change1-3. Ebullition is typically the dominant methane transport pathway compared with dissolution from sediments and transport through plant vascular tissue3, and while bubble compositions vary they are usually predominantly methane with a balance of nitrogen gas4-5. Since rising bubbles are subject to dissolution, only a portion of the methane released from the sediment as bubbles is directly emitted to the atmosphere. The remainder dissolves into the water column where it is available for biological incorporation into the lake food web6-8 or diffusion to the atmosphere.

36 37 38 39 40 41 42 43 44 45 46 47 48

Bubble dissolution rates depend strongly on bubble size and to a lesser extent on water chemistry, temperature, and dissolved gas concentrations9-11. Despite the importance of methane bubble size to dissolution and atmospheric emissions, relatively few estimates of lacustrine methane bubble size distributions exist12-14. Existing bubble size distribution measurements are typically limited to short periods in time or space, and therefore little is known about the spatial and temporal changes in methane bubble size distributions within a water body. Spatial variability could be produced, at least in part, by variations in sediment properties such as gradients in tensile and compressive strength which have been successfully used to model ebullition behavior15, and Young’s modulus and the tensile fracture toughness which have been found to affect the size and mobility of methane accumulations within sediment16. Temporal variability could result from systematic changes to the rate of methane generation (potentially due to seasonal changes in temperature and organic substrate deposition) which has also been shown to affect methane behavior in sediment17.

49 50

Given the limited knowledge about bubble size distributions, it is difficult to estimate the importance of bubble dissolution to the dissolved carbon budget in a lake. Some data suggest

ACS Paragon Plus Environment


51 52 53 54 55 56 57 58 59 60 61 62

that dissolution may contribute significantly to methane in the hypolimnion14, and insignificantly to methane in the epilimnion18. Recent modeling work suggests that bubble dissolution can be important to both hypolimnetic and epilimnetic methane accumulations at sufficiently high ebullition flux19. In all cases the bubble dissolution equations used to calculate contributions to dissolved methane were developed for marine environments, and while the velocity component of the model has been tested20, the mass transfer coefficient component has not been tested in freshwater conditions10. Since methane solubility is approximately 20% higher in freshwater versus saltwater21, bubble dissolution rates may vary accordingly. Additionally, surface-active compounds in the water can adsorb to the bubble interface and affect dissolution rates9. Microorganisms are known to produce surface-active compounds22-23, and therefore the wide variation in microorganism abundance within water bodies24-25 could potentially lead to additional differences in bubble dissolution rates between water bodies.

63 64 65 66 67 68 69 70

The purpose of this work is to characterize the temporal and spatial variability of methane bubble size distributions in Upper Mystic Lake, MA, to test the performance of a widely-used bubble dissolution model under the freshwater conditions of Upper Mystic Lake, and to quantify the importance of dissolving bubbles to dissolved methane levels within both the hypolimnion and epilimnion. Detailed bubble size distribution data were gathered using our recentlydeveloped bubble size sensors which can measure in-situ bubble volumes over long time periods26-27. By deploying multiple bubble size sensors at 2 depths within the lake for a multimonth period we are able to address the following questions:

71 72

1. What is the distribution of bubble sizes in this lake and how spatially and temporally variable are these sizes?

73 74

2. Do existing methane dissolution models accurately predict the evolution in bubble size distribution as bubbles rise through the water column in Upper Mystic Lake, MA?

75 76

3. How important is bubble dissolution to accumulated methane in the hypolimnion, or to the airwater flux of dissolved methane from the epilimnion?

77 78

Materials and Procedures

79 80 81 82 83 84 85 86 87

Bubble size distributions were measured using a bubble size sensor and data processing algorithms described in Delwiche et al.26 and Delwiche and Hemond27. Briefly, the sensor uses infrared LEDS, phototransistors, and a microcontroller to optically detect and measure bubbles rising through an inverted glass funnel stem. The sensor electronics are encased in a custom water-proof housing, and a custom surface data buoy holds batteries and electronics capable of powering the sensor via a data cable for up to 4 weeks between battery replacements. Raw sensor data is downloaded at the data buoy and post-processing algorithms convert the data to estimated bubble volumes. The sensor also includes optional attachments to increase the number of intercepted bubbles and to sample their bulk methane content.

88

Sensor Deployment


Page 2 of 17

Page 3 of 17

89 90 91 92 93 94 95 96 97 98 99 100 101 102 103


Eight bubble size sensors were deployed in Upper Mystic Lake, MA from July-November of 2016. Upper Mystic Lake is a dimictic, eutrophic kettle lake near Boston, MA with a surface area of 0.5 km2, an average depth of 15 m, a maximum depth of 24 m, and a history of methane ebullition28-29. The sensor deployment site shown in Figure 1 was chosen because of its mild bathymetry gradients and observed history of ebullition in preliminary field work. Using the deployment system described in Delwiche and Hemond27, we positioned four sensors at 16 m depth (sensors 1, 2, 3, and 4), and four sensors at 3 m depth (sensors 6, 7, 8, and 9) within the field site. Initial deployments prior to July had more sensors in the vicinity of sensor 3, but field results showed that collected bubble volumes were consistently lower around sensor 3 and much higher around sensor7, so in July sensors 1, 6, and 4 were moved southward to capture more bubble data. Sensor 3 remained in place and continued to see negligible ebullition for the remainder of the deployment window. We believe the lack of measured ebullition represents a real phenomenon because the sensor continued to record data consistent with particulate matter floating past the detectors (indicating the sensor was working), and we did not collect gas in the gas collection chamber.

104

105 106 107

Figure 1: Upper Mystic Lake field site near Boston, Massachusetts showing the sensor deployment locations (sensor IDs visible in inlay). Bathymetric lines drawn at 3.05 m (10 foot) increments.

108

Data collection and sampling



Page 4 of 17

109 110 111 112 113 114 115 116 117 118

We visited sensors every 1-2 weeks to download data and collect gas accumulated in the gas collection chambers. Collected gas was stored in wetted glass syringes and analyzed within 24 hours on a thermal conductivity gas chromatograph. Methane results were calibrated using commercial standards of 10%, 50%, and 100% methane, and nitrogen results were calibrated against 10%, 50% and atmospheric nitrogen concentrations. We periodically measured temperature and dissolved oxygen using a Hydrolab MiniSonde probe, and gathered water samples for dissolved methane analysis using a sampling device described in Peterson30. Dissolved gases were extracted into a helium-filled headspace by shaking samples for 20 minutes and then analyzing the headspace gas on a Shimadzu 2014 gas chromatograph using a flame ionization detector.

119 120 121 122 123 124 125

To quantify potential changes in bulk methane content within the gas collection chambers in between sampling intervals, we incubated a known quantity of methane in a trap deployed at 3 m depth and modified to block any incoming bubbles. Multiple measurements of the change in gas volume and methane content over 1-2 week periods allowed us to modify measured methane contents in the other traps to account for this modest in-trap dissolution and/or oxidation. In addition to the modified trap at 3 m, we also ran incubation trials in T3 (at 16 m) when it became apparent that T3 was deployed in a location with negligible ebullition flux.

126 127

Bubble dissolution model

128

Bubble dissolution rates are calculated using the following mass transfer equation: 𝑑𝑀𝑖 4𝜋𝑟 2 = −𝐾𝐿𝑖 (𝐻𝑖 𝑃𝑖 − 𝐶𝑖 ) 𝑑𝑧 𝑣𝑏

129

(1)

130 131 132 133 134 135 136 137 138 139

where i is the gas species, M is the number of moles of gas i in the bubble, z is vertical distance, KL is the gas transfer coefficient, H is the Henry's constant, P is the partial gas pressure within the bubble, C is the dissolved gas content in the water, r is bubble radius, and ʋb is the bubble rise velocity. Many approximations for the bubble gas transfer coefficient and the bubble rise velocity exist31, and the equations used in this study are those included in the model from McGinnis et al.10 The bubble dissolution model was written in Matlab, integrated using the Euler method, and run with Upper Mystic Lake dissolved oxygen, dissolved methane, and temperature profiles (profiles in S.I.). Dissolved nitrogen was assumed to be in equilibrium with the atmosphere because of twice-annual water column turnover.

140 141

Results and Discussion

142

Bubble size and flux characteristics

143

Spatial heterogeneity in bubble size distributions


Page 5 of 17


144 145 146 147 148

During the 2016 field season sensors were fully operational from August through midNovember, excluding times during power loss or sensor malfunction. Of the 8 sensors deployed, 7 recorded a substantial number of ebullition events throughout the season (see Figure 2 for number of bubbles recorded per sensor). We believe that sensor 3, which only recorded 4 bubbles all season, was deployed in a low flux area as described in Materials and Methods.

149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167

Previous studies of methane bubble size distributions have either included distributions at a single location in a water body, or distributions measured over a short period of time. These previous studies have therefore left open the question of spatial heterogeneity in bubble size distributions within a water body. The 2016 field data shows that there are statistical differences between some distributions even over the relatively small and bathymetrically homogeneous sampling area chosen for this study (Figure 2), indicating that localized sediment structure may influence ebullition. This finding is consistent with recent modelling work by Scandella et al., 2011 on Upper Mystic Lake showing that ebullition behavior can be predicted using a modelling parameter that incorporates the sum of the vertical gradients in tensile and compressive strength15. Additionally, work by Algar and Boudreau, 2010 has shown that sediment parameters such as Young’s modulus and the tensile fracture toughness partially govern the size of methane gas deposits within the sediment, though more work is necessary to understand how sediment physical characteristics influence bubble size distributions16. While we do not have localized sediment structure data to potentially explain our observed variability, our data suggests that it may not be appropriate to use a single bubble size distribution to estimate dissolution rates over an entire water body. We also note that several size distributions had a double-peak structure with one peak between 3-4 mm diameter, and a peak around the mean bubble diameter. Further work is necessary to determine prevalence and cause of this double peak distribution structure.

168 169



170 171 172 173 174 175 176

Figure 2: Bubble size distributions from each sensor deployed during 2016, including mean bubble diameters as measured at 16 m or 3 m, Sauter Mean Diameters (SMD), and the number of bubbles observed (n). The letter in the bottom right corner shows statistically different distributions (comparisons are only made between sensors at the same deployment depth, 16 m or 3 m).

177

Temporal heterogeneity in bubble size

178 179 180 181 182 183

As shown in Figure 2, bubble size distributions vary spatially even over relatively short distances. The question remains as to whether size distributions vary temporally. Although our bubble size data only span 3 months, within this limited context we saw no evidence of systematic changes to the bubble size distributions (using linear regression analysis for bubble size versus time). Future longer-term studies could address whether there are seasonal trends in ebullition as bubble flux, organic matter deposition, and water column mixing vary.

184 185

We did find that average daily bubble size was positively (though non-linearly) correlated with daily ebullition flux (Figure 3). This indicates that higher ebullition flux is attributed both


Page 6 of 17

Page 7 of 17

186 187 188 189 190 191 192 193 194 195 196


to more bubbles and to increased bubble size. The natural logarithm of daily ebullition flux was a significant predictor for average daily bubble size (p

Methane Bubble Size Distributions, Flux, and Dissolution in a

Recommend Documents