{"paper_id":"23560fea-7253-4014-860f-066b03136dac","body_text":"Posted on 24 Mar 2025 — CC-BY 4.0 — https://doi.org/10.22541/au.174283759.97027476/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary.\nShip waves induce methane ebullition in the littoral zone of a large\nlake\nOle Lessmann 1, Karla Martinez-Cruz 1, Lea Loraine Ropella 1, Nora Tutas 1, and Frank\nPeeters1\n1Environmental Physics, Limnological Institute, University of Konstanz\nMarch 24, 2025\nAbstract\nMethane (CH4) ebullition is a major emission pathway in lakes. However, ebullition remains difficult to assess, and data from\nlarge lakes are scarce. This study investigates ebullition in the littoral zone of a large lake, demonstrating that ship waves,\neven after traveling long distances, regularly trigger gas bubble releases from the sediments in shallow waters. Although these\nebullitions events occur only during the short time intervals of ship-wave passages, they lead to substantial CH 4 emissions.\nComparing several littoral sites, we assess the frequency and extent of ship-wave-induced ebullition events and contrast them\nwith ebullition caused by a storm event. Our findings reveal a strong cross-correlation between pressure drops due to ship waves\nand ebullition fluxes, highlighting ship-wave-induced pressure changes as a significant yet understudied driver of CH 4 emissions\nin lakes. With ship traffic on lakes steadily increasing, understanding ship-wave-induced ebullition is essential for improving\nestimates of lake-wide CH 4 emissions.\n1\n\nArticle type: Letter 1 \n 2 \nTitle: Ship waves induce methane ebullition in the littoral zone of a large lake 3 \n 4 \nRunning head: Ship waves induce methane ebullition 5 \n 6 \nAuthor names: Ole Lessmann1, Karla Martinez-Cruz1, Lea Loraine Ropella1, Nora Tutas1, Frank 7 \nPeeters1 8 \n 9 \nAffiliation: 1Environmental Physics, Limnological Institute, University of Konstanz, Mainaustr. 10 \n252, D-78464 Konstanz, Germany 11 \n 12 \nCorresponding author: Ole Lessmann, lessmann.ole@gmail.com 13 \n 14 \nAuthor Contribution Statement 15 \nOL: Conceptualization, Investigation, Formal analysis, Software, Visualization, Project 16 \nadministration and Writing (Original draft; Review and Editing). KMC: Conceptualization, 17 \nInvestigation and Writing (Review and Editing). LLR: Investigation, Formal analysis and Writing 18 \n(Review and Editing). NT: Investigation, Formal analysis and Writing (Review and Editing). FP: 19 \nConceptualization, Formal analysis, Software, Resources, Visualization, Project administration, 20 \nSupervision, Funding acquisition and writing (Review and Editing). 21 \n\nScientific Significance Statement  22 \nMethane (CH4) ebullition is a major pathway in global CH 4 emissions from lakes, yet it remains 23 \npoorly understood, particularly in large lakes where data are scarce. Based on continuous 24 \nrecordings of ebullition fluxes at several sites in the littoral zone of Lake Constance, we 25 \ndemonstrate that ship waves are an important driver of ebullition events. Ship -wave-induced 26 \nebullition events have short durations, but contribute a large fraction of overall CH 4 ebullition. 27 \nAlthough ebullition events due to ship waves occur only in shallow waters, they are a substantial 28 \nsource of basin-wide CH4 emissions almost on the same order as the diffusive CH 4 emissions. As 29 \nboating on inland waters continues to rise, our study highlights the need to account for ship-wave-30 \ninduced ebullition in future estimates. 31 \n 32 \nData Availability Statement 33 \nData will be made available in the Zenodo repository. 34 \n  35 \nAbstract  36 \nMethane ( CH4) ebullition is a major emission pathway in lakes. However, ebullition remains 37 \ndifficult to assess, and data from large lakes are scarce. This study investigates ebullition in the 38 \nlittoral zone of a large lake, demonstrating that ship waves, even after traveling long distances, 39 \nregularly trigger gas bubble releases from the sediments in shallow waters. Although these 40 \nebullitions events occur only during the short time intervals of ship -wave passages, they lead to 41 \nsubstantial CH4 emissions. Comparing several littoral sites, we assess the frequency and extent of 42 \nship-wave-induced ebullition events and contrast them with ebullition caused by a storm event. 43 \nOur findings reveal a strong cross -correlation between pressure drops due to ship waves and 44 \nebullition fluxes, highlighting ship -wave-induced pressure changes as a significant yet 45 \nunderstudied driver of CH4 emissions in lakes. With ship traffic on lakes steadily increasing, 46 \nunderstanding ship-wave-induced ebullition is essential for improving estimates of lake-wide CH4 47 \nemissions. 48 \n 49 \nKeywords: ebullition flux; methane emission; bubble traps; ship waves; Lake Constance  50 \n\nIntroduction  51 \nMethane ( CH4) is an important greenhouse gas, contributing strongly to human -made global 52 \nwarming and driving climate change (IPCC 2021). Freshwater systems (e.g., lakes, ponds, 53 \nreservoirs, and rivers) are responsible for approximately 36–82% of the natural and 16–44% of the 54 \ntotal global CH4 emissions (Rosentreter et al. 2021; Jackson et al. 2024). Although estimates of 55 \nglobal CH4 emissions from freshwater systems differ substantially between studies, lakes are 56 \nundoubtedly one of the largest sources (Rosentreter et al. 2021). 57 \nThe main CH4 emission pathways from lakes include ebullition, diffusion across the water -58 \natmosphere interface, plant -mediated fluxes, and release of stored CH4 during water column 59 \nturnover (Michmerhuizen et al. 1996; Bastviken et al. 2004; Encinas Fernández et al. 2014). 60 \nEbullition describes the release of bubbles from the sediment. A gas phase forms in the sediment 61 \nwhen the combined equilibrium partial pressures of all gases exceed the total ambient pressure. 62 \nThis process requires high rates of gas production within the sediment, typically driven by 63 \nmicrobial CH4 production. Bubble release occurs when the gas pressure is sufficient to break up 64 \nthe sediment matrix. Ebullition is often triggered by sudden drops in hydrostatic pressure, e.g., 65 \nduring water level regulation or pumping operations in reservoirs (Harrison et al. 2017; Encinas 66 \nFernández et al. 2020) and during water level drops caused by ship lock operations and ship 67 \npassages in rivers (Maeck et al. 2014). 68 \nCompared to deep-water conditions, CH4 production is higher in sediments of the shallow littoral 69 \ndue to warmer temperatures (Conrad 2023) and increased organic matter input (Praetzel et al. 70 \n2020). Additionally, the lower ambient pressure in shallow waters reduces the CH4 production 71 \nrequired for gas phase formation. As a result, ebullition is more common in shallow littoral zones 72 \nthan in deep profundal zones. Furthermore, CH4-rich bubbles rising from shallow waters lose less 73 \n\nCH4 to the ambient water during their brief ascent to the lake surface than those released from 74 \ngreater depths (McGinnis et al. 2006). Consequently, CH4 emissions to the atmosphere from 75 \nebullition fluxes are expected to be substantially larger in shallow than in deep-water regions. 76 \nEbullition is often the dominant CH4 emission pathway in lakes, contributing an estimated 56% of 77 \nglobal annual CH4 emissions from these systems (Johnson et al. 2022). However, the drivers of 78 \nebullition remain poorly understood (Savignano et al. 2024). Additionally, ebullition data from 79 \nlakes and especially from large and deep lakes are scarce (Bastviken et al. 2004; DelSontro et al. 80 \n2016; Praetzel et al. 2021). Therefore, understanding and quantifying ebullition in large lakes 81 \nremains a critical research gap.  82 \nIn this study, we investigate ebullition in the littoral zone of large pre -alpine Lake Constance and 83 \ndemonstrate that many of the observed ebullition events were associated with passages of ship 84 \nwaves. Even when the ship waves were generated at great distances from the shore, the pressure 85 \nchanges at the sediment surface associated with the passages of ship waves were typically sufficient 86 \nto trigger release of CH4-rich bubbles from the sediment in the littoral zone. We compare the 87 \nfrequency and magnitude of ship -wave-induced ebullition events at different littoral sites and for 88 \nship waves originating from the same route. Additionally, we present data showing ebullition 89 \ntriggered by wind waves and discuss how these events differ from ship-wave-induced ebullition.  90 \n 91 \nMethods 92 \nStudy site 93 \nMeasurements were conducted in Upper Lake Constance (9°30’ E, 47°30’ N), a large pre -alpine 94 \nlake in Europe (surface area ~470 km 2, maximum depth ~251 m). Instruments were deployed  at 95 \nthree littoral sites (Fig.1): At Obere Güll (OG) and Lipbachmündung (LM) from June 27 to July 6 96 \n\n2022 and at Bottighofen (BH) from July 11 to August 6, 2024 . Water depths at deployment were 97 \n2.0 m (OG), 1.7 m (LM) and 2.5 m (BH). Average temperatures during the measuring period were 98 \n20.2°C (OG), 21.6°C (LM) and 22.9°C (BH).  99 \nWave measurements 100 \nHigh-resolution pressure time series (16 Hz) were measured using pressure sensors (PDCR 1730, 101 \nTeramess) with logger units developed at the University of Konstanz. The instruments were 102 \ndeployed ~0.5 m above the sediment. Time series of water surface elevations were derived from 103 \npressure time series by correcting for the wavelength -dependent attenuation of pressure 104 \ndisturbances with depth (Hofmann et al. 2008). Waves were identified applying a zero-up-crossing 105 \nalgorithm (IAHR 1989) to the detrended elevation time series. The wave field was characterized 106 \nby maximum wave height, H max, (wave height of the largest wave), and significant wave period, 107 \nTsig (mean wave period of the largest third of the waves) within 1-minute intervals. The passage of 108 \na surface wave trough is accompanied by a pressure decrease in the water column. For 109 \ncharacterising this pressure decrease, we defined P dec, the magnitude of the maximum pressure 110 \ndecrease at the sediment surface associated with the passage of the wave with maximum wave 111 \nheight within a 1-minute time interval. 112 \nEbullition flux measurements 113 \nEbullition fluxes from the sediment were measured using ebullition flux funnels (EFF) that were 114 \ncustom manufactured at the University of Konstanz. Gas bubbles released from the sediment are 115 \ncaught by an inverted funnel and directed into a cylindrical column where they displace the water. 116 \nThe pressure difference between the gas phase in the column and the surrounding water is recorded 117 \nat 1 Hz and is used to determine the gas volume of the captured bubbles. The change in gas volume 118 \nover time provides the ebullition flux, F eb. Attached to the EFFs we deployed loggers 119 \n\n(AQUAlogger® 520, Aquatec) for measuring ambient total pressure and temperature every 30 s. 120 \nThese data were used to convert the measured gas volumes to gas volumes at standard conditions 121 \n(1 atm, 20°C), assuming a molar volume of 24.0838 mol L -1. Ebullition fluxes were determined 122 \nfrom these time series by calculating the rate of gas -volume change per time employing a 30 s 123 \n(corresponding to 30 values) differential filter and smoothing the resulting time series of rates with 124 \na boxcar filter using a moving average of 3 minutes (corresponding to 180 values). The ebullition 125 \nflux was calculated by dividing the volume change per time by the cross -sectional area (0.65 m 2) 126 \nof the EFF. 127 \nExposure of littoral zones to ship waves 128 \nShoreline exposure to ship waves was estimated using a simple empirical model. Ship motion 129 \ncreates a characteristic wave pattern and the wave packets with the largest wave heights propagate 130 \nat an angle of approximately 35° relative to the direction of the ship track (Sorensen 1997). For 131 \neach location along the shoreline, the model determines whether and from which points along the 132 \nship’s route waves can reach the shore. The distance travelled by the ship waves was estimated as 133 \nthe shortest distance from their generation point along the ship’s route to the shore, without 134 \naccounting for diffraction effects. Ship -wave arrival times were determined from the distance 135 \ntravelled by the wave packets and their group velocity. The latter was estimated to be 41% of the 136 \nship speed at wave generation assuming propagation as deep-water waves (Crawford 1984) during 137 \nmost of the wave’s path. 138 \nThe shipping routes relevant to this study include the catamaran ferry route between Konstanz and 139 \nFriedrichshafen and the car ferry route between Konstanz and Meersburg (Fig. 1). During the study 140 \nperiod, the catamaran ferries operated hourly between 06:00 and 19:00 h on weekdays (Monday –141 \nFriday) and between 08:00 and 19:00 h on weekends (Saturday–Sunday). Additional evening trips 142 \n\nwere scheduled on Fridays and Saturdays at 20:00 and 22:30 h (westbound) and at 21:00 and 23:30 143 \nh (eastbound). According to the ship -wave-exposure model, waves from westbound catamaran 144 \nferries reached the stations BH, LM and OG after traveling approximately 2.4 km, 3.9 km and 8.8 145 \nkm respectively, while waves from eastbound catamaran ferries reached station LM after traveling 146 \n~4.6 km. The car ferries operated in intervals of 15–20 minutes during the day, and hourly at night. 147 \nOnly waves from southwestbound car ferries reached station OG after traveling ~4.4 km. 148 \n 149 \nResults  150 \nEbullition events and ship-wave passages 151 \nAt all stations, ebullition shows characteristic patterns of regularly occurring short -term events 152 \nwith relatively large ebullition fluxes (Fig. 2, 3). These ebullition events have a duration of several 153 \nminutes and are typically associated with the passage of ship waves at the measuring stations (Fig. 154 \n2). Ships and routes responsible for the waves reaching the different shore sites were identified 155 \nbased on the wave period and the predicted wave arrival time. Catamaran ferries, traveling at 32–156 \n34 km h-1, generate waves with periods typically exceeding 4 s (Fig. 2D, H), whereas car ferries, 157 \ntraveling at ~20 km h-1, generate waves with periods of 2.8–3.7 s (Fig. 2H). Observed and predicted 158 \nwave arrival times agreed well for all car and catamaran ferry routes (Figs. 2C, G). 159 \nEbullition events at station OG (Fig. 2E) were associated with waves from the southwestbound car 160 \nferry and the westbound catamaran ferry. However, the ebullition fluxes corresponding to 161 \ncatamaran ferry waves were substantially larger than those corresponding to car ferry waves. At 162 \nstation LM, the ebullition flux exhibited two peaks per hour (Fig. 2A), corresponding to the 163 \npassages of catamaran waves generated on the eastbound and westbound route. In contrast, the 164 \nebullition fluxes at stations OG and BH show only a single peak per hour due to catamaran waves 165 \n\n(Fig. 2E, 3), as these sites are exposed only to waves generated on the westbound route. Ebullition 166 \nevents not only reflect the hourly pattern in the schedule of the catamaran ferries, but also the 167 \ndifferences in these patterns between days (Fig. 3). 168 \nThe passages of surface waves are associated with a periodic pressure decrease at the sediment 169 \nsurface, which most likely is responsible for triggering ebullition. The pressure decrease at the 170 \nsediment surface, indicated by negative P dec, increases with H max but also depends on the wave 171 \nperiod, which determines the attenuation of the amplitude of pressure fluctuation with depth. This 172 \neffect is evident at LM at 09:40 h, when surface waves with large H max but small Tsig resulted in a 173 \nlow-magnitude Pdec, and, despite the large H max, the induced F eb was smaller compared to waves 174 \nwith larger Pdec (Fig. 2A–D). 175 \nCross-correlation between Pdec and Feb shows the most negative correlation coefficients at zero lag 176 \n(r = -0.11 (BH), r = -0.34 (LM), p < 0.01) confirming that ebullition fluxes are linked to the pressure 177 \ndecreases at the sediment surface (Fig. S1). Additional distinct peaks of negative correlation 178 \ncoefficients occur at lags of 1 hour and 24 hours at both stations, with further lags at approximately 179 \n21 and 42 minutes at station LM (Fig. S1). The patterns in the cross-correlations between Pdec and 180 \nFeb reflect the patterns in the auto -correlations of P dec and of F eb, which are consistent with the 181 \nhourly and diurnal patterns in the ship-wave arrival times. 182 \nThe mean ebullition flux during a catamaran -wave-induced ebullition event was estimated using 183 \nthe fluxes within a 7 -minute interval centered on the peak flux during the event.  On average, 184 \nebullition fluxes associated with catamaran waves were 0.017 mL m -2 s-1 for the westbound and 185 \n0.012 mL m-2 s-1 for the eastbound catamaran routes at station LM, and 0.025 mL m-2 s-1 at station 186 \nBH (Table 1). The average ebullition flux during times not associated with the passages of 187 \n\ncatamaran waves were 0.004 mL m-2 s-1 at LM (excluding a storm event, see Fig. 4) and 0.002 mL 188 \nm-2 s-1 at BH. 189 \nAdditional ebullition events 190 \nEbullition occurred not only during ship-wave passages but also during a strong wind event (wind 191 \nspeeds reaching 10 m s -1) on June 30, 2022, between 22:20 and 24:00 h. At station LM, Tsig was 192 \n~4.6 s, and Hmax reached 0.7 m during this event, leading to a Pdec of up to 0.21 dbar. Feb increased 193 \nin parallel with P dec reaching a maximum flux of 0.46 mL m-2 s-1 at 23:00 h (Fig. 4). Afterwards, 194 \nFeb decreased more rapidly than Pdec and ceased around 23:30 h. Pdec and Hmax remained elevated 195 \nuntil 24:00 h, indicating continued wave action that no longer triggered ebullition. The mean 196 \nebullition flux during the 100-minute duration of the storm was 0.090 mL m-2 s-1, over three times 197 \nlarger than that from catamaran waves. In contrast, surface waves and ebullition fluxes at site OG 198 \nremained unaffected by this storm event. 199 \nNot all surface waves associated with substantial pressure decrease caused peaks in the ebullition 200 \nflux (e.g., station OG around 9:20 h , Fig. 2E). Conversely, ebullition events also occurred 201 \nspontaneously without connections to surface waves (e.g., during the night, Fig. S2). 202 \n 203 \nDiscussion  204 \nThe patterns of ebullition events suggest that the pressure decrease during the passage of surface -205 \nwaves is the mechanism driving ebullition. Ebullition fluxes are typically larger the larger the 206 \npressure decrease, e.g., during the first half of the storm event and between sites. At station BH, 207 \nthe mean ebullition flux associated with the passage of catamaran waves and the corresponding 208 \npressure decrease are both significantly larger than at station LM (Fig. S3). However, the amount 209 \nof gas available in the sediment plays a role in limiting the ebullition flux. This is suggested by the 210 \n\nsubstantial reduction in ebullition fluxes during the second half of the storm event, likely due to 211 \nprior CH4 depletion from ebullition (Fig. 4). Hence, exhaustion of the CH4 pool in the sediment 212 \nresulting from previous ebullition and differences in CH4 production across seasons and sites, may 213 \ncontribute to the variability of ebullition fluxes. In the absence of pressure decreases from surface 214 \nwaves, the accumulation of CH4 in the sediment may generate sufficient gas pressure for triggering 215 \nebullition (Fig. S2). 216 \nEbullition fluxes associated with catamaran waves occur only during short time intervals, 217 \naccounting for ~13% of the day at LM and ~7% of the day at BH. Note that catamaran waves from 218 \nthe eastbound and westbound route reach LM, but only those from the westbound route reach BH. 219 \nDespite these short time intervals of impact, the daily mean ebullition flux associated with 220 \ncatamaran waves contributes on average 41% (LM) and 46% (BH) of the respective total daily 221 \nmean ebullition flux, because the average flux during the passages of catamaran waves is 3 (LM) 222 \nto 11 (BH) times higher than during the rest of the day (Table 1). At LM, the daily mean ebullition 223 \nflux for periods unaffected by catamaran waves is 60% larger than at BH, suggesting higher CH4 224 \nproduction at LM than at BH (Table 1). CH4 depletion cannot be responsible for the differences in 225 \ndaily mean ebullition fluxes between the sites, because the daily mean flux associated with 226 \ncatamaran waves is higher at LM than at BH. These results support the conclusion that the larger 227 \naverage ebullition flux at BH, compared to LM, results from differences in the trigger (Pdec) induced 228 \nby catamaran waves. 229 \nCH4 fluxes were estimated from ebullition fluxes, assuming 80% CH4 in the gas . This fraction, 230 \nmeasured in gas bubbles released from littoral sediments in Upper Lake Constance, aligns with 231 \nCH4 fractions in bubbles from shallow sediments reported in other studies (Maeck et al. 2014; 232 \nSepulveda-Jauregui et al. 2015).  Note that bubbles typically lose little CH 4 while rising from 233 \n\nshallow depths to the surface (McGinnis et al. 2006). Accordingly, the daily mean CH4 flux 234 \nassociated with catamaran waves from the westbound route was 3.3 mmol m -2 d-1 at LM and 4.8 235 \nmmol m-2 d-1 at BH and from the eastbound route 2.2 mmol m-2 d-1 at LM (Table 1). These values 236 \nare within the range of mean CH4 ebullition fluxes reported for the shallow zone (< 3 m depth) of 237 \nsmall lakes (1.1 mmol m-2 d-1) and ponds (4.6 mmol m-2 d-1) in Canada (DelSontro et al. 2016).  238 \nBasin-wide CH4 emissions due to catamaran-wave-induced ebullition were estimated based on the 239 \narea of the shallow water zone exposed to catamaran waves that may induce ebullition. Exposed 240 \nshore zones were identified by applying the ship -wave-exposure model to the catamaran routes. 241 \nThe area of the shallow water zone exposed to catamaran waves that travelled up to 9 km (Fig. 1B) 242 \n— the largest wave travel distance for which we have confirmed ebullition measurements — and 243 \nassuming bubble release from 0.5 to 3.0 m water depth, is 4.36 km 2 for the eastbound route and 244 \n4.56 km2 for the westbound route. Using these areas and the daily mean CH4 fluxes from station 245 \nLM as lower limit, basin -wide daily mean CH4 emissions associated with catamaran waves are 246 \nestimated to be ~25 kmol d-1 or ~400 kg d-1. These represent about 43% of the daily mean diffusive 247 \nCH4 emission of 940 kg d-1 from Upper Lake Constance. The latter was estimated from the basin-248 \nwide mean CH4 surface concentration of 0.2 µmol L -1 in the northern basin of Upper Lake 249 \nConstance (Encinas Fernández et al. 2016) and a gas transfer velocity of 0.6 m d-1 for CH4 at 20°C, 250 \nassuming wind speeds below 4.2 m s -1 (Liss and Merlivat 1986; Schwarzenbach et al. 2003). The 251 \nestimation above demonstrates that CH4 ebullition induced by catamaran waves accounts for a 252 \nlarge fraction of basin -wide CH4 emissions. Note, that the basin -wide CH4 ebullition associated 253 \nwith catamaran waves is rather underestimated, because additional ebullition likely occurs from 254 \ndepths greater than 3 m and from other shoreline stretches. Also, CH4 ebullition induced by all ship 255 \nwaves must be larger than that by catamaran-waves alone. 256 \n\nDuring the strong wind event lasting less than 2.5 hours, ebullition caused a release of 18 mmol 257 \nCH4 m-2, which corresponds to the release by catamaran -wave-induced ebullition over 3.5 days. 258 \nHowever, wind events of similar magnitude are not very common in Lake Constance, i.e., between 259 \nMay and October, wind events with wind speeds exceeding 7 m s-1 only occur ~4 times per month. 260 \n 261 \nConclusions 262 \nAlthough ship-wave-induced ebullition events are limited to the shallow littoral and occur only 263 \nduring the short time intervals of ship -wave passages, the resulting daily mean ebullition fluxes 264 \nand the associated CH4 emissions can be substantial. Additionally, short and infrequent strong wind 265 \nevents can trigger large CH4 ebullition fluxes at wind -exposed shorelines. Hence, quantifying 266 \nannual lake-wide CH4 emissions associated with ebullition in large lakes is challenging not only 267 \ndue to the spatial heterogeneity of ebullition events but also because of their large temporal 268 \nvariability across different time scales.  269 \n\nAcknowledgements 270 \nWe thank the scientific workshop facilities at the University of Konstanz, especially Thomas 271 \nSchuchhardt, Oliver Hamburger, and Georg Heine for the construction and maintenance of the 272 \npressure loggers and the ebullition flux funnels. We wish to express our gratitude to Devin 273 \nJachtmann for dedicated data collection efforts, to Beatrix Rosenberg and Anton Pranger for 274 \ntechnical assistance in the field, and to Soo Hyun Im for  valuable support in manuscript 275 \npreparation. This manuscript benefitted from wind speed data provided by the German 276 \nMeteorological Service. The authors acknowledge using ChatGPT to correct grammatical errors. 277 \nThis research was conducted as part of the project SuBoLakes (Sustainable Boating on Lakes in 278 \nGermany), which was financially supported by the German Federal Environmental Foundation 279 \n(Deutsche Bundesstiftung Umwelt, DBU) under project No. 35825/01. 280 \n\n 281 \nFigure 1. Bathymetric maps of Lake Constance illustrating the location of study sites and ship routes. A) 282 \nOverview of Lake Constance, showing the locations of the three measurement stations: Obere Güll (OG), 283 \nBottighofen (BH) and Lipbachmündung (LM), as well as the ship tracks of the catamaran ferry and the car 284 \nferry. B) Enlarged section of Upper Lake Constance, displaying ship tracks (solid lines), the propagation of 285 \nship waves (dashed lines), and littoral zones exposed to waves generated on the catamaran routes. 286 \n\n\n 287 \nFigure 2. Time series of ~5 hrs of ebullition flux (A, E), pressure decrease at the sediment surface (B, F), 288 \nmaximum wave height (C, G), and significant wave period (D, H) on June 29, 2022, at stations OG and LM. 289 \nPredicted ship-wave arrival times are marked for the catamaran ferry (westbound and eastbound) and the 290 \ncar ferry (southwestbound). 291 \n\n\n 292 \nFigure 3. Ebullition flux pattern at station BH on four days with varying ship schedules. Predicted ship-293 \nwave arrival times are marked for the westbound catamaran ferry. A large peak reaching 0.175 mL m -2 s-1 294 \nis truncated around 18:00 h in panel A due to the axis scaling. 295 \n\n\n 296 \nFigure 4. Time series of 2.5 hrs of ebullition flux and pressure decrease at the sediment surface (inverted) 297 \n(A), maximum wave height (B), and significant wave period (C), overnight from June 30 to July 1, 2022, at 298 \nstation LM during a storm event.  299 \n\n\nTable 1 Ebullition fluxes and CH4 fluxes evaluated for different time periods with and without passages of 300 \ncatamaran waves at stations LM and BH. 301 \nTime period of evaluation Daily mean \nduration (hrs) \nAverage  \nebullition flux  \n(mL m-2 s-1) \nDaily mean \nebullition flux  \n(mL m-2 d-1) \nDaily mean CH4 \nebullition flux \n(mmol m-2 d-1) \n  LM BH LM BH LM BH LM BH \nall day  24 24 0.0047 0.0036 403.8 314.6 13.4 10.5 \npassages of  \ncatamaran waves \nwestbound 1.6 1.6 0.0171 0.0245 97.9 143.1 3.3 4.8 \neastbound 1.6 — 0.0117 — 66.8 — 2.2 — \nall day, excluding \npassages of \ncatamaran waves \n 20.8 22.4 0.0036 0.0022 267.4 175.9 8.9 5.8 \n  302 \n\nReferences 303 \nBastviken, D., J. Cole, M. Pace, and L. Tranvik. 2004. Methane emissions from lakes: Dependence 304 \nof lake characteristics, two regional assessments, and a global estimate. Global Biogeochem. 305 \nCycles 18: 1–12. doi:10.1029/2004GB002238 306 \nConrad, R. 2023. Complexity of temperature dependence in methanogenic microbial 307 \nenvironments. Front. Microbiol. 14. doi:10.3389/fmicb.2023.1232946 308 \nCrawford, F. S. 1984. Elementary derivation of the wake pattern of a boat. Am. J. Phys. 52: 782 –309 \n785. doi:10.1119/1.13550 310 \nDelSontro, T., L. Boutet, A. St-Pierre, P. A. del Giorgio, and Y. T. Prairie. 2016. Methane ebullition 311 \nand diffusion from northern ponds and lakes regulated by the interaction between temperature and 312 \nsystem productivity. Limnol. Oceanogr. 61: S62–S77. doi:10.1002/lno.10335 313 \nEncinas Fernández, J., H. Hofmann, and F. Peeters. 2020. Diurnal pumped -storage operation 314 \nminimizes methane ebullition fluxes from hydropower reservoirs. Water Resour. Res. 56: 315 \ne2020WR027221-e2020WR027221. doi:10.1029/2020WR027221 316 \nEncinas Fernández, J., F. Peeters, and H. Hofmann. 2014. Importance of the Autumn Overturn and 317 \nAnoxic Conditions in the Hypolimnion for the Annual Methane Emissions from a Temperate Lake. 318 \nEnviron. Sci. Technol. 48: 7297–7304. doi:10.1021/es4056164 319 \nEncinas Fernández, J., F. Peeters, and H. Hofmann. 2016. On the methane paradox: Transport from 320 \nshallow water zones rather than in situ methanogenesis is the major source of CH4 in the open 321 \nsurface water of lakes. J. Geophys. Res. Biogeosciences 121: 2717 –2726. 322 \ndoi:10.1002/2016JG003586 323 \n\nHarrison, J. A., B. R. Deemer, M. K. Birchfield, and M. T. O’Malley. 2017. Reservoir Water-Level 324 \nDrawdowns Accelerate and Amplify Methane Emission. Environ. Sci. Technol. 51: 1267 –1277. 325 \ndoi:10.1021/acs.est.6b03185 326 \nHofmann, H., A. Lorke, and F. Peeters. 2008. The relative importance of wind and ship waves in 327 \nthe littoral zone of a large lake. Limnol. Oceanogr. 53: 368–380. doi:10.4319/lo.2008.53.1.0368 328 \nIAHR working group on wave generation. 1989. List of Sea‐State Parameters. J. Waterw. Port, 329 \nCoastal, Ocean Eng. 115: 793–808. doi:10.1061/(ASCE)0733-950X(1989)115:6(793) 330 \nIPCC. 2021. Summary for Policymakers, p. 3–32. In Masson-Delmotte, V., P. Zhai, A. Pirani, S.L. 331 \nConnors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M. Huang, K. Leitzell, 332 \nE. Lonnoy, J.B.R. Matthews, T.K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, B. Zhou [eds.], 333 \nSummary for Policymakers. In: Climate Change 2021: The Physical Science Basis. Contribution 334 \nof Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate 335 \nChange. Cambridge University Press. doi:10.1017/9781009157896.001 336 \nJackson, R. B. and others. 2024. Human activities now fuel two-thirds of global methane emissions. 337 \nEnviron. Res. Lett. 19: 101002. doi:10.1088/1748-9326/ad6463 338 \nJohnson, M. S., E. Matthews, J. Du, V. Genovese, and D. Bastviken. 2022. Methane Emission 339 \nFrom Global Lakes: New Spatiotemporal Data and Observation -Driven Modeling of Methane 340 \nDynamics Indicates Lower Emissions. J. Geophys. Res. Biogeosciences 127: e2022JG006793 -341 \ne2022JG006793. doi:10.1029/2022JG006793 342 \nLiss, P. S., and L. Merlivat. 1986. Air -Sea Gas Exchange Rates: Introduction and Synthesis, p. 343 \n113–127. In P. Buat -Ménard [ed.], The Role of Air -Sea Exchange in Geochemical Cycling. 344 \nSpringer Netherlands. 345 \n\nMaeck, A., H. Hofmann, and A. Lorke. 2014. Pumping methane out of aquatic sediments –346 \nebullition forcing mechanisms in an impounded river. Biogeosciences 11: 2925 –2938. 347 \ndoi:10.5194/bg-11-2925-2014 348 \nMcGinnis, D. F., J. Greinert, Y. Artemov, S. E. Beaubien, and A. Wüest. 2006. Fate of rising 349 \nmethane bubbles in stratified waters: How much methane reaches the atmosphere? J. Geophys. 350 \nRes. Ocean. 111: 1–15. doi:10.1029/2005JC003183 351 \nMichmerhuizen, C. M., R. G. Striegl, and M. E. McDonald. 1996. Potential methane emission from 352 \nnorth-temperate lakes following ice melt. Limnol. Oceanogr. 41: 985 –991. 353 \ndoi:10.4319/lo.1996.41.5.0985 354 \nPraetzel, L. S. E., N. Plenter, S. Schilling, M. Schmiedeskamp, G. Broll, and K. -H. Knorr. 2020. 355 \nOrganic matter and sediment properties determine in -lake variability of sediment CO2 and CH4 356 \nproduction and emissions of a small and shallow lake. Biogeosciences 17: 5057 –5078. 357 \ndoi:10.5194/bg-17-5057-2020 358 \nPraetzel, L. S. E., M. Schmiedeskamp, and K. -H. Knorr. 2021. Temperature and sediment 359 \nproperties drive spatiotemporal variability of methane ebullition in a small and shallow temperate 360 \nlake. Limnol. Oceanogr. 66: 2598–2610. doi:10.1002/lno.11775 361 \nRosentreter, J. A. and others. 2021. Half of global methane emissions come from highly variable 362 \naquatic ecosystem sources. Nat. Geosci. 14: 225–230. doi:10.1038/s41561-021-00715-2 363 \nSavignano, M. J., K. Ethan D., S. Laurence C., and M. and Engram. 2024. Geospatial Analysis of 364 \nAlaskan Lakes Indicates Wetland Fraction and Surface Water Area Are Useful Predictors of 365 \nMethane Ebullition. Ann. Am. Assoc. Geogr. 114: 299–313. doi:10.1080/24694452.2023.2277817 366 \n\nSchwarzenbach, R. P., P. M. Gschwend, and D. M. Imboden. 2003. Activity coefficient and 367 \nsolubility in water. Environ. Org. Chem. 2. 368 \nSepulveda-Jauregui, A., K. M. Walter Anthony, K. Martinez -Cruz, S. Greene, and F. Thalasso. 369 \n2015. Methane and carbon dioxide emissions from 40 lakes along a north–south latitudinal transect 370 \nin Alaska. Biogeosciences 12: 3197–3223. doi:10.5194/bg-12-3197-2015 371 \nSorensen, R. M. 1997. Prediction of Vessel-Generated Waves with Reference to Vessels Common 372 \nto the Upper Mississippi River System. U.S. Army Corps of Engineers, Upper Mississippi River - 373 \nIllinois Waterway System Navigation Study, ENV Report 4. 50 pp. 374 \n\n1 \n \nSupplementary Figures 1 \nSupplementary Figure S1 2 \n 3 \nSupplementary Figure S1 . Cross -correlation of the ebullition flux and the pressure decrease at the 4 \nsediment surface for stations BH and LM at daily (A, C) and hourly (B, D) timescales. Cross-correlations 5 \nare based on time series with 3 -minute averages consisting of a total of 12491 data points (BH) and 1656 6 \ndata points (LM).  7 \n\n\n2 \n \nSupplementary Figure S2 8 \n 9 \nSupplementary Figure S2. Time series of ebullition flux (A), pressure decrease at the sediment surface 10 \n(B), maximum wave height (C), and significant wave period (D), overnight from July 2 to July 3, 2022, at 11 \nstation LM. Ebullition peaks at 01:50 and 04:50 h are not associated with a pressure decrease suggesting 12 \nthat they occur spontaneously without a specific external trigger. In contrast, the other ebullition peaks, near 13 \nmidnight and from 08:00 h onwards, are associated with pressure decreases linked to passages of surface 14 \nwaves.  15 \n\n\n3 \n \nSupplementary Figure S3 16 \n 17 \nSupplementary Figure S3 . Maximum pressure decrease at the sediment surface (A) and maximum 18 \nebullition flux (B) during catamaran-wave-induced ebullition events at stations LM and BH for the eastbond 19 \nand westbound catamaran routes. Sample sizes are: LM east (n = 98), LM west (n = 95), and BH west (n = 20 \n350). The figure shows the median, upper and lower quartiles, minimum and maximum values (excluding 21 \noutliers), and individual outliers. Statistically significant differences between the datasets were determined 22 \nusing a two-sample t-test and are indicated by an asterisk (* p < 0.01). 23","source_license":"CC-BY-4.0","license_restricted":false}