Report compiled and edited by Bo Barker Jørgensen and Henrik Fossing

Table of content page 1.3.1 Spatial mapping of shallow gas using a low frequency multibeam echosounder 9 1.3.2 Towards quantification of shallow free gas in Baltic Sea sediments 10 1.3.3

BALTIC GAS Final scientific report

Reporting period:

January 1, 2009 – December 31, 2011

Report compiled and edited by

Bo Barker Jørgensen and Henrik Fossing

Table of content

page

1.3.1 Spatial mapping of shallow gas using a low frequency multibeam echosounder 9 1.3.2 Towards quantification of shallow free gas in Baltic Sea sediments 10 1.3.3 Rising of methane gas bubbles through the water column and pockmark distribution in

the Gdansk Basin

13

1.4.2 Holocene mud deposits and presence of free methane gas (an example from Aarhus Bay) 19 1.4.3 Distribution and temporal variability of dissolved methane in the water column of the

Baltic Sea

20 1.4.4 Continuous measurement of surface methane concentrations – ships of opportunity 22 1.4.5 The role of the inshore and littoral region for methane emissions from the Baltic Sea 24

1.5.1.2 Forecasting the impact of climate change on methane gas inventories 27 1.5.2 Environmental controls of gaseous methane production in the Baltic Sea (an example

from Aarhus Bay)

1 Executive Summary

BALTIC GAS is a research project funded by BONUS (i.e The Baltic Organisations Network for funding Science) that addresses methane in the Baltic Sea and its mutual coupling to climate change and eutrophication Through application of seismo-acoustic techniques and geochemical approaches BALTIC GAS mapped shallow gas in the Baltic Sea seabed and water column and analysed methane production, consumption, gas accumulation, and methane fluxes for a better understanding and quantitative synthesis of the dynamics and budget of methane in Baltic Sea

The BALTIC GAS research project brought together a multidisciplinary team of scientists from 12 research tions (see 4 BALTIC GAS Science team) with the goal to (1) quantify and map the distribution and flux of me-thane in the Baltic Sea, (2) analyse the controls on the relevant key biogeochemical processes, (3) integrate seismo-acoustic mapping with geochemical profiling, (4) model the dynamics of Baltic Sea methane in the past (Holocene period), present (transport-reaction models), and future (with predictive scenarios), and (5) identify hot-spots of gas and potential future methane emission in the Baltic Sea

institu-The research project applied modern advanced technology and novel combinations of approaches to pursue the listed goals i.a multibeam bathymetry and seismo-acoustic profiling to map gas distribution and escape struc-tures in combination with gravity coring Further methane ebullition was identified and analysed by acoustic flare imaging and sea surface emission by floating methane-gas flux chambers and “ferry box” monitoring Al-ready existing data were mined and combined with new observations to generate the first well-constrained methane budget of a coastal sea, to map gas appearance in the seabed and to generate a predictive model to understand and forecast methane fluxes as a function of environmental gradients, climate change, and contin-ued eutrophication

The BALTIC GAS research project was divided into 5 work packages (see also 3 Work package overview)

WP1: Project management, coordination and dissemination,

WP2: Data mining and GIS-mapping,

WP3: Gas and seismo-acoustic mapping,

WP4: Biogeochemistry,

WP5: Modelling and data integration,

The vast majority of the new and existing knowledge obtained during BALTIC GAS, however, was reached through a tight (interdisciplinary) cooperation between work packages

A short introduction that outlines the coordination of the BALTIC GAS research project is given below followed

by a presentation of the major outcome of the project Additional information may be read on the BALTIC GAS homepage (www.balticgas.au.dk, i.e Deliverable 1.1) where all BALTIC GAS Deliverables are accessible (except for two submitted manuscripts, i.e Deliverable 4.4 and 5.3)

1.1 BALTIC GAS main results

• A novel approach was developed for the monitoring of gas in the seabed Low frequency multibeam backscatter data provided unique mapping capabilities of the distribution and depth of free gas Com-

bined with geochemical analyses of deep sediment cores this has yielded new high-resolution maps of methane and gas distribution in selected areas of the Baltic Sea sediments

• A novel approach was developed to quantify gas in the seabed by a Parasound sediment echosounder using three individual wavelengths By this use of multichannel seismo-acoustics combined with ad-vanced data analysis it was possible to determine the gas volume in the sediment as well as the size of gas bubbles and the vertical extent of gassy sediment Such data are now used to verify model results on methane accumulation and cycling

• A novel application of a multibeam swath mapping system for sediment visualization was used to detect and quantify gas bubbles rising from the seabed A new cross-correlation technique similar to that used

in particle imaging velocimetry has now yielded impressive results with respect to unambiguous bubble detection and remote bubble rise velocimetry

• A detailed transect of seismic and geochemical data from non-gassy to gassy sediment in Aarhus Bay combined with reactive-transport modeling has now provided strong evidence that free gas bubbles in the Baltic Sea sediments migrate slowly upwards When approaching the sulfate zone the gas re-

dissolves and the methane is effectively broken down sub-bottom

• Hot-spots of methane outgassing from the sediment, often accompanied by pock-marks on the seafloor found by multibeam bathymetry, have now been detected and mapped in several areas of the Baltic Sea, in particular in the Polish and Russian sectors

• Long-term monitoring of methane in the surface water throughout the central Baltic Sea by a box” mounted on a ferry between Travemünde, Gdynia and Helsinki has revealed the seasonal dynamics and geographical distribution of methane Combined with a transect through the entire water column from the Bay of Bothnia to the Kattegat this has yielded a unique data set on methane in the Baltic Sea and on the source strength of this green-house gas to the atmosphere

“ferry-• Based on data mining and on new data an extensive database on methane and related parameters has been compiled and made publicly available through the BALTIC GAS homepage and the database, PAN-GAEA The data have also been used to develop new GIS-maps of the distribution of gas, the depth of the methane zone, and the subsurface methane fluxes in the central basins of the Baltic Sea

• Studies in the central basins were supplemented with detailed analyses of methane cycling in the dish archipelago Experiments indicated that 30-84% of the total methane flux in the sediment could be attributed to bubbles Yet, 98% of this methane was oxidized in the oxic water column, thus preventing emission to the atmosphere The remaining water-air flux was still 10-fold higher than in the central ba-sins

Swe-• Based on the large geophysical and geochemical data base compiled by BALTIC GAS, a transient tive-transport model was developed to understand the past and present methane cycle in the Baltic seabed and the accumulation of gas The model results now explain quantitatively how gas in the sea-bed is controlled by the thickness of Holocene mud which is the main modern source of methane

reac-• Model predictions of future methane fluxes and the potential for accelerating gas emissions from the seabed have shown a large robustness of the biogeochemical processes towards breaking down the me-thane This robustness could not have been predicted without the large amount of new data that could verify the model and has been a key result of the project The general model forecast is thus that the predicted temperature increase of 1-2 oC and salinity drop in the Baltic Sea, together with an unchanged level of eutrophication, is not expected to lead to a dramatic increase in the gas ebullition from the sed-iments during this century

Fig 1 Baltic Sea geographical areas

investi-gated during 15 BALTIC GAS cruises Aarhus

Bay (A), (B) Mecklenburg Bay, (C) Arkona

Basin, (D) Bornholm Basin, (E) Gdansk Bay,

(F) Baltic proper, (G) Gotland Deep, (H)

Both-nian Sea, (I) BothBoth-nian Bay), (J) Gulf of

Fin-land, (K) Himmerfjärden See also Table 2

1.2 Project management, research cruises, and data collection

A total of 52 scientist, post docs, Ph.D-students, master students, and technicians were engaged in BALTIC GAS They participated during the project period (1/1/2009 – 31/12 2011) in BALTIC GAS workshops (Table 1; Deliver-able 1.3), meetings between two or more BALTIC GAS institutions, Baltic Sea integrated seismo-acoustic training courses (see 5 Educational activities), (see 6 Stakeholder events and other related activities), conferences and stakeholder events (see 7 Meeting and conferences), and 15 cruises to the Baltic Sea (Deliverable 1.5) covering

in particular Aarhus Bay, Mecklenburg Bay, Arcona and Bornholm Basin, Gdansk Bay, Baltic proper, Gotland Deep, Gulf of Bothnia, Gulf of Finland, and Himmerfjärden (a Swedish fjord about 50 km SSW of Stockholm; see Table 2 and Fig 1) See also BALTIC GAS scientific Reports (Deliverable 1.2)

Table 1 BALTIC GAS workshops organized during the project period: 1/1/ 2009 – 31/12 2011

Number of participants Bremen

Germany

February 4-6 2009 Max Planck Institute for Marine Microbiology

29 Warnemünde Ger-

of seismic data from a large area offshore Gotland The collected seismic data were loaded to seismic workstations by the data own-ers, the distribution of free gas was digitized, and the data com-piled at GEUS as basis for GIS-mapping carried out by Alfred We-gener Institute for Polar and Marine Research (see below and De-liverable 2.1, 2.2, and 2.3) Table 3 and Fig 2 gives an overview of the seismic lines recorded or mined from archived data during BAL-TIC GAS

Table 2 Cruises accomplished during BALTIC GAS (2009 – 2011) Investigations performed comprised i.a

seismo-acoustic measurements, sediment sampling and concomitant analyses to depict chemical and physical profiles, water column studies, and air-water flux measurements (see cruise reports for further details) Number of participating BALTIC GAS scientists and institutions are listed together with name of chief scientist

persons /institutions 2009

Gulf of Gdansk Vistula River mouth

Feb 20-27

8 pers/2 inst

Northern Baltic proper

Apr 21-25

Bothnian Sea and Bay

Jun 4-17

2 pers/2 inst

(i.e Vyborg Bay)

Jun 30 - July 2

seismo-acoustics sediment water column

Nikolay Pimenov

6 pers/1 inst

Bothnian Sea and Bay

Aug 28 - Sep 9

sediment water column

Nov 5-16

seismo-acoustics sediment

Zygmunt Klusek

14 pers/4 inst

Arco-na Basin Bornholm Deep

Stolpe Foredelta land Deep

Got-Nov 27 - Dec 17

seismo-acoustics sediment water column

Volker Brüchert

5 pers /2 inst

Gdansk Basin (i.e NW pers)

Jun 20-27

seismo-acoustics sediment water column

Vadim Sivkov

2 pers/2 inst

Arkona Basin Bornholm Basin Gotland Deep Bothnian Sea and Bay

Jul 31 - Aug 21

seismo-acoustics sediment water column

Sediment parameters were measured during 12 out of the 15 cruises and comprised a vast amount of both geochemical and physical observations in combination with sediment characterization and occasionally rate measurements of methanogensis, anoxic oxidation of methane, and sulphate reduction (see cruise reports for details: http://balticgas.au.dk/balticgasaudk/project/workingareasandcruises/ Deliverable 1.5) The number of parameters recorded differed between sediment cores but as a key parameter to BALTIC GAS methane (CH4) was measured in all sediment cores and sulfate (SO4

bio-) in most Thus sediment data submitted to the common database PANGAEA (http://pangaea.de/) comprised (when measured) (Deliveable 1.4):

2-A) Pore water chemistry: CH4, δ13CH4, SO4

, H2S, Cl-, Fe2+, Mn2+, NH4

2-+, PO4 3-, alkalinity, dissolved inorganic carbon (DIC), δ13DIC, acetate and other volatile fatty acids (VFA),

B) Solid phase chemistry: acid volatile sulfide (AVS), chromium reducible sulfur (CRS), ‘metals’, nutrients, Fe(solid phase), Pb-210, total nitrogen (TN), total carbon ( TC), C/N-ratio, total organic carbon (TOC),

δ13TOC,

C) Process rates: methanogensis, anoxic oxidation of methane, and sulphate reduction

D) Physical parameters: temperature, density, porosity

Sampling of the water column comprised CH4, δ13CH4, and H2S and was always accompanied (i.e initiated) by a conventional CTD cast The water column data were likewise submitted to PANGAEA

Fluxes of methane from the sediment to the bottom water and across the sea surface in coastal and open-sea Baltic waters were determined by modelling from concentration data and by direct flux measurement Sea-air exchange was quantified by data from an autonomous measurement system mounted on the ferry M/S FINN-MAID in November 2009 commuting regularly between Travemünde (Germany), Gdynia (Poland) and Helsinki (Finland) to measure methane and carbon dioxide concentration in the surface waters Direct sea-air fluxes of

Fig 2 Seismo-acoustic lines (i.e data) complied in

a common database by GEUS Black lines are

ar-chive data Red lines show seismo-acoustic data

measured during Baltic Gas

Table 3 Seismo-acoustic data (measured and archived)

complied in a common database by GEUS See Fig 2

Data source

Acoustic line length (km) Archive data University of Stockholm 2,700 Archive data Shirshov Institute of

Oceanology Atlantic Branch, Russian Academy of Sciences

18,300

Archive data Baltic Sea Research

Archive data The Geological Survey of

Denmark and Greenland (GEUS)

1,900 Archive data Department of Geoscienc-

es , University of Bremen 900 Acoustic data measured during Baltic Gas 4,600

methane were determined with floating chambers in near-shore areas of the Southern Stockholm archipelago,

in particular theHimmerfjärden estuary

The BALTIC GAS coordinators organized that the modellers received seismo-acoustic data and results from in situ sediment measurements on a regular basis and that data were exchanged between BALTIC GAS scientists, par-ticularly at BALTIC GAS workshops Here also new ideas, hypotheses, and theories were discussed based on the most recent findings and the modellers’ knowledge base’ was improved leading to the development of robust algorithms and models These models proved highly valuable in bringing the many point observations into a larger context and in confirming hypotheses concerning, e.g the transport-reactions models

1.3 Methane gas and seismo-acoustic mapping

During the BALTIC GAS research project seismo-acoustic surveying was the initial and most efficient method to find and map free methane gas in the sediment and water column In particular when combined with direct methane measurements in sediment cores and water column samples

In BALTIC GAS, acoustic monitoring of sediments was performed by use of a broad spectrum of acoustic niques and equipment i.a singlebeam echo-sounders with frequencies of 12, 38 and 200 kHz, low frequency multibeam echo-sounder (50 kHz ELAC), parasound sediment echo-sounder (4.2, 18,5 and 42.8 kHz) , Innomar sediment echo-sounder (5, 10 and 15 kHz), high resolution broadband chirp echo-sounder (1 – 10kHz), single-channel Boomer (2 – 4kHz), single-channel Sparker (1kHz), and multichannel Airgun seismics (200 Hz)

tech-The echo-sounder transmits high frequency sound waves down to the sea floor and further into the seabed Depending on the frequency, more or less of the energy is reflected at the sea floor, which enables a precise

Fig 3 A seismo-acoustic transect crossing methane gas saturated sediment in Bornholm Basin across a distance of

about 8 km (about 90 m water depth) between site 374200 (55 o 14.973N/ 15 o 26.147E) and site 374180 (55 o 20.329N/

15 o 26.237E) Methane gas bubbles efficiently absorb the acoustic energy and thus ‘blanks’ information from the lying sedimentary strata Yellow vertical lines show position and length of gravity cores sampled (see also Fig 11)

under-determination of the water depth with an accuracy of a few centimeters Lower frequency sound waves trate deeper into the sediment depending on the hardness of the seabed due to differences in mineralogy and other geological features The sound waves penetrate relatively easy into fine grained sediments as mud, silt, and clay, whereas penetration depths are very limited in sand, gravel and glacial till Thus, the seismo-acoustic data obtained give an acoustic cross section of the seabed where the sediments and sediment strata are seen by

pene-‘acoustic imagery’ as a vertical reflector pattern profile (Fig 3) Methane gas bubbles, however, efficiently absorb the acoustic energy and thus ‘blank’ information from the underlying sedimentary strata Hence by ‘acoustic imagery’, free methane gas is observed as a conspicuous, more or less homogeneous blanking on the seismic

‘picture’ or ‘scan’ (Fig 3)

During most BALTIC GAS cruises and at most stations studied hydroacoustic singlebeam echo-sounders were used as the standard tool for remote sensing of free methane gas in the seabed and water column However, during BALTIC GAS also new seismo-acoustic techniques were introduced and demonstrated as superior solu-tions for shallow gas mapping compared to singlebeam techniques as explained below

1.3.1 Spatial mapping of shallow gas using a low frequency multibeam echosounder

Bornholm Basin in the Baltic Sea (80m) hosting free methane gas was surveyed with low and high frequency multibeam acoustic equipment accompanied by standard sub-bottom profiling

The gathered multibeam backscatter data (Fig 4a) revealed distinct differences between areas with and without gas Compared to standard technique singlebeam data (Fig 4b) and geochemical analysis (Fig 4a, cs1 and cs2) BALTIC GAS scientist for the first time demonstrated a perfect match in regard to sensing free methane gas with

Fig 4 (a) Backscatter amplitude chart of EM120 with a transition zone between bluish/no gas and yellowish/shallow

gas areas; the inlet shows amplitude data gathered from the 95 kHz system not showing any transition, (b) SOUND sub-bottom data recorded along the blue and red line in (a) starting at 08:15 UTC The transition zone between shallow gas (right) and no shallow gas (left) plots exactly at the same time as seen in the multibeam data (a) On figure

PARA-(a) and (b) the blue and red line indicate the two sediment types ‘mud’ and ‘mud hosted with shallow gas’, respectively

this method (Deliverable 3.3) In contrast no data patterns were observed in the high frequency multibeam vey (Fig 4a insert) This emphasized the superior potential of our low frequency approach where the low fre-quency pulses not only penetrated the seafloor up to 10 m but the ‘acoustic gas front’ also mimicked the gas front observed form direct measurement in gravity cores Even small gas pockets clearly emerged as “bright spots” in the backscatter data on the very outer swath at 140° (Fig 4a, patch in northeasterly region) making the multibeam system a reliable tool for 2D wide-angle/spatial mapping of shallow gas

sur-The technique just introduced was further tested in the Botnian Sea (Fig 5) The respective survey shows more complex morphology with outcropping till on the seafloor and subbottom channels within the Holocene mud locally hosting pockets of shal-low gas The multibeam was run in parallel with the subbottom profiler Till, mud, and gas-bearing mud clearly plotted as different features in both da-tasets The till appeared as real bathymetric high (Fig 5), the mud caused deeper bathymetric meas-urements due to penetration; whereas the shallow gas within the mud caused a sudden bathymetric increase in the transition zone

Even though earlier studies demonstrated the feasibility of backscattering strength analysis in regard to sensing shallow gas, no multibeam studies exist revealing subbottom gas submerged several meters below the seafloor

in two dimensions Given the high sensitivity and the large coverage shown in our study we attribute low quency multibeam sounders a great potential in soft sediments in regard to spatial mapping of shallow gas, iden-tifications of individual gas pockets, and to locate buried objects

fre-1.3.2 Towards quantification of shallow free gas in Baltic Sea sediments

The presence of free gas bubbles introduces fundamental changes in the properties of sediments and their sponse to seismic sound waves While high frequency acoustic waves are strongly attenuated, lower frequency seismic waves are able to penetrate gas-charged sediment layers However, the speed in gas-bearing sediment is significantly reduced due to lower wet bulk density and modification of other elastic and sediment physical properties Thus by careful determination of interval velocity from raw multichannel seismic data, we were able

re-to estimate the amounts of free gas in the sediment

Recording the reflected seismic waves with an array of hydrophones/channels allowed an indirect measurement

of their velocity from the curvature of reflection hyperbolae (conventional interactive velocity analysis) In tion, we performed velocity analysis on pre-stack time migrated data, which, although time consuming and computationally intensive, allowed the determination of the velocity field over gassy areas more accurately and more extensively in space than hitherto done Depending on stratification (identifiable reflectors), the accuracy and resolution varied significantly In general, velocities dropped from about 1450 m/s in non-gassy fine-grained surficial sediments down to a few hundred m/s in the gas-charged zone Beneath the gas patches, in the post-glacial and glacial sediments, velocities again increased (>1500 m/s)

addi-Fig 5 Pseudo bathymetric presentation after application of

a slope filter (Botnian Sea) Red areas show outcropping till

seafloor, wheras blue and green data represent soundings

reflected from subbottom features like gas and submerged

till

To quantify the gas content based on the velocity field, we used Anderson & Hampton’s geoacoustic model (1980), which described the relationship between compressional wave velocity and the physical properties of gas-bearing marine sediments In the model, gas bubbles were assumed to be fully contained within the pore space, thus modifying its compressibility Taking the interval velocity values between reflectors, free gas content

in the pore volume could be estimated (Fig 6) Values of the free gas content at the test location in the holm Basin ranged from 0.1 to 2%, where sensitivity becomes reduced These numbers were basically in agree-ment with the modelling results

Born-When excited, gas bubbles in the sediment resonated at a fundamental frequency, which was mainly mined by the bubble size and physical properties of the surrounding medium As a result, acoustic behaviour

deter-Fig 6 High resolution multichannel seismics performed with a GI gun with a central low frequency of 200 Hz and a 50

m long streamer with 48 channels (seismo-acoustic transect GeoB10_044) (a) The interval velocity values between reflectors (m/s) showing significantly reduced velocities in the gas charged sediment (dark blue pixels in the white framed sediment section) compared to gas free sediments outside the frame Depth below surface is expressed in m/s

as the two-way travel time (TMT) i.e travel time from source and back to receiver The offset shows the distance (m) from the start of transect in the south to the endpoint in the north The vertical solid line shows the depth of the sea- floor (b) The interval velocity values between reflectors in the gas charged sediment (i.e white frame in panel (a)) (c) Free gas content estimated from ‘interval velocities’ up to 2% gas of the sediment pore volume (i.e gas replacing pore water).

was different below, at and above the resonance frequency but attenuation due to the scattering effects would

be strongest close to the resonance frequency

By imaging shallow gassy sediments at a broader frequency range, gas bubbles could be physically characterized from their acoustic response In the Bornholm Basin, gassy areas were surveyed with three frequencies of the Parasound sediment echo-sounder (4.2, 18.5 and 42.8 kHz) High reflection amplitudes from and strong signal attenuation beneath the gas front occurred at the lowest imaging frequency of the Parasound, although natural attenuation increased with frequency Accordingly, this effect could be attributed to bubble resonance behav-iour, which was not observed at the two higher frequencies (Fig 7) Based on the theoretical considerations of Anderson and Hampton (1980) and for typical sediment properties, bubble size distribution was likely to peak near a diameter of approx 2 mm (4.2 kHz) with the smallest bubbles larger than 0.2-0.4 mm (42.8 and 18.5 kHz, respectively)

These new results obtained by the BALTIC GAS project represent major scientific progress in the quantification

of gas distribution and gas volume in marine sediments based on geophysical analyses Using a diverse suite of seismic and acoustic equipment in parallel together with advanced methods of data processing and analysis, remote profiling measurements come within reach for routine gas quantification While larger uncertainties still exist and basic physical concepts still have to be developed and tested, the acquired results for gas content and bubble sizes seem to be in good agreements with evidence from biogeochemical measurements and modelling

Fig 7 Seismo-acoustic signal received from a Parasound sediment echo-sounder operated at frequencies of 4.2,

18.5, 42.8 kHz along an appox 700 m transect in Bornholm Basin from NW (left) to SE (right) Amplitudes at 4.2 kHz, close to the resonance frequency of about 2 mm bubble size, show scatter in the gas-charged layer and decrease beneath The can be considered as horizontal variation in this decrease can be considered as a measure of the gas content At the 18.5 and 42.8 kHz, above resonance frequency, the effect of gas is only revealed in generally lower amplitudes than values observed in adjacent gas-free sediments Depth below surface is expressed in m/s as the two- way travel time (TMT) i.e travel time from source and back to receiver The offset shows the distance (m) from the start of transect in the NW to the endpoint in the SE

1.3.3 Rising of methane gas bubbles through the water column and pockmark distribution in the Gdansk Basin State of the art multibeam seismo-acoustic techniques were used to remotely investigate gas bubbles rising through the water column BALTIC GAS scientists successfully deployed a prototype multibeam ecco-sounder that allowed us to image the rising of methane gas bubbles through the water column and to sense the respec-tive rise pattern of individual gas bubbles released from the sediment i.a from pock marks in the seafloor (Fig 8, Deliverable 3.3)

Investigations were carried out in the Gdansk Basin by the Institute of Oceanology, Polish Academy of Science and the Atlantic Branch of the P.P.Shirshov Institute of Oceanology, Russian Academy of Sciences Acoustical surveys with multi-beam and side scan sonars were focused on mapping of pockmarks and detection of gas bub-bles released from the seafloor The presence of shallow gas in the Gdansk Basin area was manifested by differ-ent indications such as gas-saturated mud (including gas pockets), pockmarks, and gas outflow within pockmark (Fig 9) The total area covered by pockmarks in the Gdansk Basin was about 27 km2 (25.1 km2 in the Polish sector and 1.7 km2 in Russian sector, Table 4)

One area with pockmarks was located in the north-eastern part of the Gdansk Deep slope Seven pockmarks of various morphologies, typically elongated from the southwest to the northeast, were revealed in this area The horizontal length of the structures varied from 200 to 900 m, with a mean width of about 150-200 m and depths

of 1-3 m below the surrounding seafloor Apart from individual pockmarks, groups of 2-3 of these depressions were also observed Usually, pockmarks were surrounded by gassy mud or located at its periphery This distinct pockmark area was situated on a cross-section of different fracture zones with weakened zones of the sedimen-tary cover (supply channels, such as faults and furrows), which serve as a pathways for deep gas

Fig 8 (a) Successive echo-image frames recorded during water column imaging with SB3050 showing Rosette

(RWS) downcast, contact with gassy sediments, and induced bubble escape into the water column (b) “Beam-Slice” presentation with the x-axis representing the ping times in seconds where the y-axis is two-way-travel time [s] Hori- zontal features represent non-buoyant microbubbles (I) where to the right some ascending bubbles occur (II)

A relatively large single pockmark of 3 km length and 0.4 km wide elongated in south-north direction was covered and mapped in the Polish sector of the Gdansk Deep (54.738N/19.186E, center position) Additionally pockmarks were identified between 55.197N – 55.072N and 18.907E – 19.018E by use of broad banded echo-sounding Acoustically recognized pockmarks cross sections ranged from 20 to 200 m in diameter

dis-Fig 9 Distribution of gas outflow (at arrow), pockmarks and shallow gas in the Polish and Russian EEZ of Gdansk

Bay The gas outflow is also shown on Fig 10.

Table 4 Pockmarks in Gdansk Basin Polish and Russian EEZ

55.18N/18.94E Pockmark 1,5 55.36N/19.81E Pockmark 0.06

55.14N/18.99E Pockmark 15,0 55.36N/19.81E Pockmark 0.29

54.82N/18.84E Pockmark 1,4 55.36N/19.82E Pockmark 0.18

54.57N/19.16E Pockmark 6,7 55.35N/19.79E Pockmark 0.32

54.57N/19.16E Gas outflows 0,5 55.35N/19.78E Pockmark 0.49

55.32N/19.76E Pockmark 0.30 55.32N/19.74E Pockmark 0.07

An active gas outflow within pockmark (Fig.10) was documented in the southern part of the Gulf

of Gdansk, with the center positioned at the 54.571N/19.165E (Table 4) The size of the struc-ture was determined using a 12 kHz echo-sounder to be about 250 – 300 m Gas bubbles emanating from the sea floor at 80 m water depth were observed to ascent at least up to a water depth of 30 m An interesting feature of this pockmark was that the older and bigger low gas flux pockmark area confined the more active and deeper structure Using calibrated echo-sounders the radius of the raising gas bubbles was estimated to range from about 2 mm up to

In the Polish sector gas pockets, included in ‘shallow gas’ areas(Fig 9), were mostly localized in the area of the Gdansk Basin, especially in the vicinity of the Hel Peninsula Occurrence of such structures was associated with muddy sediments and high sedimentation rates of organic-rich matter at rates from 1.5 to over 2 mm per year Most of the gas generated in this area was mostly produced by bacteria in the Holocene sediments The total area in the Polish sector covered by gas-bearing sediments was about 440 km2

1.4 Sediment and water column biogeochemistry and physical characters

For an extensive quantification of methane concentrations in Baltic Sea sediments and in order to depict other chemical and physical profiles direct measurements were performed in the sediment (Deliverable 4.1) Based on the seimo-acoustic surveys targeted sediment sampling was done along transects reaching from sediments with deep or no ‘methane-reflection’ of the seismic signal (i.e non-gaseous sediment) to sediments with methane saturation and thus a sharp reflection (i.e gaseous sediment Fig 3 and Fig 11) Depending on stations and cruis-

es a variety of sampling equipment was applied, i.a gravity corer, Rumohr Lot corer, Frahm Lot corer, and ple corer An important part of the characterization of gas-bearing sediments was done both by a general core description (Deliverable 4.3) and by physical property studies (Deliverable 3.1) on cores obtained during an ex-tensive coring program of the Baltic Gas expeditions Multisensor core logging was used to estimate basic physi-cal properties of gas free and gas charged sediments The results (Fig.12) were used for interpretation of sedi-

multi-Fig 10 The acoustic transect through the gas outflow

within the pockmark showing gas bubbles emanating from

the sea floor (at 80 water depth) ascending up to a water

depth of about 30 m The image was obtained with a 12

kHz echosounder See location at Fig 9.

ment echo-sounder records From these data the thickness of the Holocene mud (deposits of the Littorina Sea from the past ca 9000 years) and of the older deposits from earlier Baltic Sea Stages can be estimated

Other ‘highlights’ from the sediment and water column studies as well as methane flux measurements across the sediment-water-atmosphere interfaces are presented below and on the BALTIC GAS home page

1.4.1 Methane in Baltic Sea sediments

Methane (CH4) was produced in great quantities in Baltic Sea sediments by methane-producing microorganism when organic matter was degraded through a process named methanogenesis However, sulfate-reduction dominated the upper sub-surface layers because sulfate reducing bacteria are energetically more effective in the degradation of organic matter than methane-producing microorganisms Therefore methanogenesis only took over deeper in the sediment, below the sulfate-methane transition (SMT) zone, where sulfate was exhausted or occurred at very low concentration (Fig 13)

In the Baltic Sea methane was continuously formed in the seabed and gas bubbles developed at sediment depths where the methane concentration exceeded saturation at ambient hydrostatic pressure However, by far most of the methane was effectively scavenged before it reached the sediment surface In the sub-surface sedi-ment, where there was no oxygen, sulfate was the oxidant for methane which was converted to carbon dioxide Most methane was oxidized at the depth to which sulfate penetrated also known as the sulfate-methane transi-tion (SMT)-zone (Fig 13) This microbially mediated anaerobic methane oxidation accounted for >90% of the entire methane flux in the sea floor and, therefore, played a critical role as a barrier against methane emission to the water column and further into the atmosphere

Fig 11 Methane concentrations profiles determined in sediment cores sampled along a transect in Bornholm Basin

crossing the methane gas saturated sediment shown on Fig 3 (a) Site 374200 (depth 93 m), (b) Site 374190 (depth 91 m), (c) Site 345175 (depth 93 m), and (d) Site 374170 (depth 93 m) Solid line and stipulated line show in situ CH 4 satu- ration and CH 4 saturation at 1 atm, respectively Methane is rapidly lost from the sediment core when brought on deck due to a pressure decrease Thus the scattered appearance of the CH 4 concentration profile at Site 374190 (b) – i.e sediment from the gas saturated sediment – is due to a significant loss of CH 4 before the sediment was subsampled At the Sites 374200 (a) and 345175 (c) the in situ CH 4 concentration was below saturation and not detected at all at Site

374170 (d)

Fig 12 Results of multisensor logging of gravity core 374200-06GC in Bornholm Basin (see also Fig 3 and Fig 11)

The deposits of the different Baltic Sea stages are separated by yellow lines and named in red letters The measured parameters are:" vp" - pwave velocity, "dwb" -wet bulk density, "vsh" - vane shear strength torsional moment , "con- ductivity" - electrical conductivity, "Water cont" - gravimetric bulk water content, "suszeptibility" - magnetic volume suszeptibility, "Ignition loss" - loss of ignition, "colorvalue H S V" - from core photo extracted color values of the HSV model A short sediment echosounder record (SES) is attached at the right side for comparison.

Fig 13 Concentrations of dissolved methane (CH 4 ) and sulfate (SO 4 2- ) in pore waters from Station 011 (Mecklenburg Bay) obtained during RV Merian cruise Jul 31 - Aug 21, 2010 In situ CH 4 concentration at 40 m water depth (10 ‰ salinity, 9.3 o C) and CH 4 saturation at 1 atm (on deck) are shown on left figure Expanded figure (right) show the SO 4 2-

and CH 4 flux gradients, blue and red solid lines, respectively At this station the upward CH 4 flux (red arrow) of 430 µmol m -2 d -1 is balanced by the downward SO 4 2- flux (blue arrow) of 460 µmol m -2 d -1 when CH 4 is oxidized (con- sumed) in the sulfate-methane transition zone (yellow box) by reduction of SO 4

2-.

Fig 14 GIS-map of the spatial distribution of the sulfate-methane transition zone’s (SMTZ) depth (m) in Baltic Sea

sediments Observations of concomitant presence of both sulfate and methane were predominantly done in muddy sediments with a SMTZ median value of about 0.35 m caused by a high content of particulate organic matter and thus increased production of methane Signatures of SMTZ-depth show the performed observations.

In the close vicinity to the SMT-zone, concentration gradients of sulfate and methane were steep and well fined and the sulfate flux down to this interface balanced the methane flux from the deep sub-surface (Fig 13) From the perspective of biogeochemical analysis and sampling techniques, the depth of the SMT-zone was de-fined by pore water studies, whereas determinations of methane fluxes from sediments into the water column

de-or atmosphere were much mde-ore demanding Therefde-ore, the depth of the SMT-zone provided a robust proxy – in terms analytical accuracy and data availability –for identification of regions at the Baltic Sea seafloor where high

or low methane production as well as methane fluxes into the water column were to be expected Figure 14 shows GIS-maps of the spatial distribu-tion of the SMT depth with respect to sediment types; bedrock, hard bottom, hard clay, mud, and sand (Deliverable 2.2 and 2.3) As expected the analyses revealed that shallow SMT-zone depths were predominantly observed within muddy sedi-ments with median values of about 0.35 m due to the high content of particulate organic matter and thus increased sulfate reduction and production of methane Additionally Fig 15 shows a compilation

of pore water methane fluxes to the sediment face based on the various sediment surveys con-ducted during Baltic Gas (Deliverable 4.2) The highest benthic flux rates were measured in the inshore areas of Himmerfjärden followed by the central Gotland Basin and the Arkona Basin Low rates, with the exception of a seep site in the Both-nian Bay, were measured in the northern Baltic

sur-In conclusion BALTIC GAS scientists observed (with very few exceptions) that free methane in the Baltic Sea in general was restricted to Holocene marine mud areas and that a minimum threshold thickness of mud was re-quired before free methane gas was observed in the seabed (see 1.5 Modelling methane dynamics in the Baltic Sea below) Further, detailed sediment studies in combination with seismo-acoustic investigations at a variety of locations in the Baltic Sea showed that the Holocene mud deposits in general were thinner than the threshold thickness for bubble formation and that the existing areas with free methane could be characterized as geologi-cal sediment traps

1.4.2 Holocene mud deposits and presence of free methane gas (an example from Aarhus Bay)

Based on extensive acoustic survey and sediment sampling programs performed during previous projects, i.a METROL (EU 5th Framework), Aarhus Bay sediments were proven ideal to BALTIC GAS scientist for a detailed study on the control mechanisms of methane accumulation in Baltic Sea sediment and their relation to Holocene mud thickness Seismic studies had previously shown accumulation of free CH4 gas in the central area of the bay

Fig 15 Compilation of the diffusive methane flux rates

towards the sediment surface determined during BALTIC

GAS Fluxes were calculated form the methane

concentra-tions gradients in the pore water samples (see Fig 13).

where more than 4-5 m thick homogenous mud had accumulated The lithology suggests that most CH4 is formed in the Holocene mud with little contribution from deeper layers of organic-poor glacial clay1

Thirteen 3-7 m long sediment cores were collected in October 2009 and May 2010 by gravity coring at very close distances of 20-200 m along a 600 m transect crossing from gas-free into gas rich sediment (Fig 16) The Holo-

cene mud thickness increased gradually along the transect and the measured pore water gradients CH4 and SO4

increased in steepness (Fig 17 Deliverable 4.4) The SMT-zone shifted up closer to the sediment surface when moving from the gas-free into the gas-rich area with a SO4

/CH4 flux ratio close to 1 and thus in accordance to the theoretical value We extrapolated the depth trend of organic carbon mineraliza-tion rates deep down into the methane zone to estimate the total depth-integrated rates of methanogenesis From these results

2-we conclude that the thickness of the ic-rich Holocene mud layer, and thus the sedimentation rate, was the main parameter controlling the initiation of sub-surface methane accumulation and free gas formation (Fig 18) The relationship between these factors is, however, non-linear due to a positive feedback whereby a small upward displacement of the SMT exposes sediment with more reactive organic mat-ter to methanogenesis and thereby enhances the overall methane production A higher sedimentation rate has a similar effect by increasing the burial of reactive organic matter down below the SMT where it strongly stimu-lates methanogenesis Due to the positive feedback, the SMT is further shifted upwards and the methane fluxes are increased by the transition from non-gassy to gassy sediment This mechanism of free gas formation in Baltic Sea sediments were further confirmed through sediment modeling as explained below (see 1.5.2 Environmental controls of gaseous methane production in the Baltic Sea (an example from Aarhus Bay below)

organ-1.4.3 Distribution and temporal variability of dissolved methane in the water column of the Baltic Sea

The distribution of dissolved methane in the water column of the Baltic Sea was extensively investigated based

on analysis of data gathered prior to or during the BALTIC GAS project by partner IOW A strong correlation tween the vertical density stratification, the distribution of oxygen, hydrogen sulfide, and methane was identi-fied (Fig 19) A widespread release of methane from the seafloor was indicated by methane concentrations increasing with water depth The deep basins in the central Baltic Sea showed the strongest methane enrich-ments in stagnant anoxic water bodies, with a pronounced decrease towards the pelagic redox-cline and only slightly elevated surface water concentrations In general, the low-salinity basins in the northern part of the Baltic Sea were characterized by lower water column methane concentrations and with surface water saturation values close to the atmospheric equilibrium (Fig 19)

1

for further details see: see Jensen, J.B., and o Bennike (2009) Geological setting as background for methane distribution in

Holo-Fig 16 Seismic profile of Aarhus Bay sediment showing the

Hol-ocene mud layer (56º N 6.81’, 10º E 24.71’ to 56º N 6.64’, 10º E

25.21’) The top of the free gas layer (= free gas depth, FGD) is

shown by the upper dashed line Below the FGD the presence of

free gas in the Holocene mud results in acoustic blanking and

concealment of underlying sediments The base of the mud layer

in the gassy sediments is extrapolated from the slope in the non–

gassy sediments (lower dashed line) The sampling stations and

penetration depth of the gravity cores are indicated.

Fig 17 CH 4 and SO 4

profiles along the transect shown on Fig 16 Stations M21 – M26 are in the gas-free area with M26 at the transition and M27 – M30 are in the gassy sediment area The gray line represents the position of the SMT (defined at equi-molar concentrations of CH 4 and SO 4

Fig 19 Oxygen (b) and methane (c) concentration along two transects across the Baltic Sea sampled in summer 2008

Hydrogen sulfide was converted into negative oxygen equivalents The insertion in a1 displays the location of the two transects (note red and green color code), insertion a2 shows the bathymetry of the Baltic Sea and the location of the main basins (K, Kattegat; BS, Belt Sea; AB, Arkona Basin; BB, Bornholm Basin; WGB, Western Gotland Basin; EGB, Eastern Gotland Basin; A, Åland Sea; BOS, Bothnian Sea; BOB, Bothnian Bay; GF, Gulf of Finland) The extension of the individual basins is also indicated at the top of the oxygen section The data obtained from the red (station 3075 to 3041) and green transect (station 3005to 3095) are displayed on the left and right side in Figure 1bc, with the stations labeled at the top of Figure 1b for better orientation Modified from Schmale et al (2010 2 )

Based on the comprehensive analysis represented in this basin-wide data set, more detailed investigations of the water column were performed The strong link between enhanced methane concentrations and oxygen defi-ciency was demonstrated by vertical profiles from fixed locations at stations with frequent oxic/anoxic shifts of the bottom water sampled various times (Fig 20) The mechanism of this fast buildup of a dissolved methane pool in the water column is still under investigation, and demonstrates the sensitivity of the methane cycle to changes in ventilation and to the extent of hypoxic and anoxic areas in the Baltic Sea

1.4.4 Continuous measurement of surface methane concentrations – ships of opportunity

Within the framework of Baltic Gas, the partner IOW developed and operated a system which allows the uous measurement of methane and carbon dioxide concentrations in surface waters autonomously using ships

contin-of opportunity (Fig 21; Deliverable 4.2) The analytical setup consists contin-of a methane and carbon dioxide analyzer based on off-axis integrated cavity output spectroscopy (ICOS) coupled to an established equilibrator setup

2

Schmale, O, J Schneider v Deimling., W Gülzow, G Nausch, J Waniek, G Rehder (2010) The distribution of methane in the water column of the Baltic Sea Geophysical Research Letters, 37, L12604,

Fig 20 Methane concentrations (dots) and hydrographic parameters at a station in the central Bornholm Basin in

De-cember 2009 (left), and August 2010 (right, with high resolution sampling of the lower 5m) Note jump in methane scale Bottom waters were characterized by inflow of oxygenated waters at the bottom in December and anoxic conditions in summer, in conjunction with an increase of dissolved methane concentrations from 20 to 80 nM over this period of time

Fig 21 Schematic of a system for the

contin-uous measurement of CH4 and pCO2 in face waters using off-axis ICOS The system

sur-is installed on the ferry Finnmaid run by lines

Finn-The analyzer used a highly specific infrared band laser with a set of strongly reflective mirrors to obtain an tive laser path length of several kilometers This allowed us to detect methane and carbon dioxide with high precision (better 0.1%) and frequency The system was installed in November 2009 on the cargo ship Finnmaid (Finnlines) that commutes regularly in the Baltic Sea between Travemünde (Germany), Gdynia (Poland) and Hel-sinki (Finnland)

Figure 22 shows the first complete year of tion (2010), with more than 300 days of operation, allowing hitherto unrivaled insights into the spatio-temporal development of sea-air disequilibria and fluxes for methane in a marginal sea, and the analy-sis and identification of the controlling parameters Surface methane saturations with general minimum values from December to February and maximum values during August till September showed great seasonal differences in shallow regions like the Mecklenburger Bight (103-507%) compared to deeper regions like the Gotland Basin (96-161%) Parameters influencing methane supersaturation and emission to the atmosphere, like temperature, wind and mixed layer depth, as well as processes, like upwelling, mixing of the water column, and sedimentary methane emissions, were investigated

opera-Highest methane fluxes were observed during the autumn and winter period The annual interaction of cation and mixed layer depth was found to be a key parameter for methane fluxes in deeper regions like Gulf of Finland or Bornholm Basis Methane fluxes from shallow regions like the Mecklenburger Bight are controlled by sedimentary production and consumption of methane, wind events and the temperature induced change of the solubility of methane in the surface water

stratifi-1.4.5 The role of the inshore and littoral region for methane emissions from the Baltic Sea

Investigations of the inshore coastal fluxes from the sediment and to the atmosphere focused primarily on the southern Stockholm archipelago with the eutrophied Himmerfjärden (Deliverable 4.2) In addition to water col-umn methane concentration measurements, air measurements and floating methane-gas flux chambers (for the first time Lagrangian) were deployed in the coastal regions of Swedish Baltic waters From 2009 to 2011, an as-semblage of 69 chambers was used for direct flux measurements between 0.5 m and 75 meter depth

Sea-to-air fluxes determined at water depths from 3 to 75 m in June 2011 and October 2011 ranged from 0.01 to 0.12 mmol CH4 m-2 d-1 with an average of 0.06 ± 0.007 mmol CH4 m-2 d-1 (Fig 23) Methane fluxes decreased slightly with water depth The highest flux was obtained from additional 24-hour measurements at the edge of densely vegetated shore areas in only 0.5 meter water depth Here the fluxes were as high as 0.57 mmol m-2 d-1 Bubble shield experiments at four shallow sites in depths less than 5 meters were conducted to separate diffu-sive and bubble fluxes These experiments indicated that between 30% and 84% of the total flux could be at-tributed to bubbles Inshore measurements in the eutrophic inner Himmerfjärden revealed clear methane sur-face maxima, which are likely due to discharge of methane from a local sewage treatment These concentrations

Fig 22 Methane surface concentrations between Lübeck

and Helsinki in 2010 along all transects passing close to

the west of east of the Island of Gotland color-coded for

each individual month

Fig 23 Sea-to-air methane fluxes (mmol m -2 d -1 ) at three localities Himmerfjärden, Hållsviken, and Tvären, respectively

in the southern Stockholm archipelago Chamber derived measurement were performed in June and October 2011.

were significantly higher than concentrations measured in bottom waters over a whole summer-fall measuring campaign and suggest that a significant part of the methane in this area is not derived from benthic emissions, but of sewage origin

Of particular interest was the finding that the efficiency of methane oxidation above the deep anoxic basins of the archipelago sea was very high The deep water in these basins had methane concentrations as high as 644

nM, but more than 98% of this methane was oxidized at the chemocline and in the oxic water column above resulting in very low emissions to the atmosphere

Based on our data, we conclude that the inshore zone has methane emissions that are an order of magnitude higher than in the open waters of the Baltic Sea Of these emissions, the littoral zone with water depths less than

8 meters emits a significant part of methane in the form of bubbles Since the littoral area is the most critically affected zone by nutrient runoff and groundwater discharge, future work must concentrate on the littoral to improve our predictions for future methane emissions

Himmerfjärden

Hållsviken

Tvären

1.5 Modelling methane dynamics in the Baltic Sea

Shallow seismic data were important basic information for locating free methane in the Baltic Sea sediments In combination with physical/chemical parameters measured in sediment cores – in particular methane and sulfate – BALTIC GAS scientists at Utrecht University (NL) established models which were able to couple a large array of user-defined geochemical reactions to transport processes which affected aqueous and/or solid species The modeling performed for BALTIC GAS was based on the Biogeochemical Reaction Network Simulator (BRNS) de-veloped by Regnier and co-workers and made available to the public through the BALTIC GAS homepage at

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp5-deliverable-1/ (see also Deliverable 5.1) Here, the user can access the model and define which chemical species, reactions, transport processes, and spatio-temporal domain is to be used Although the model can be adapted to variable boundary conditions, grids, and highly-complex reaction equations, these features require additional manipulation of the model code which can

be performed by the scientists involved in the project Several examples of sediment modeling for the dynamics

of organic matter, sulfate and methane are shown below

1.5.1 Climate-related effects on past and future methane dynamics

A transient reactive-transport model (Deliverable 5.1) was developed to study the evolution of the benthic thane and sulfur cycles at the millennial and centennial timescales The overarching goal of the research was (1)

me-to reconstruct the evolution of methane turnover rates as a result of long-term changes in climate conditions over glacial-interglacial cycles (Holocene period, 104 year timescale) and (2) to predict whether future climate-dependent changes in temperature and ventilation of the Baltic Sea, combined with continued organic carbon loading, could enhance methane gas production and release from the seabed (Deliverable 5.2)

1.5.1.1 Hindcasting methane dynamics during the Holocene period

The model was used to track the development of the methane geochemistry following the deposition and radation of organic rich sediments This process was initiated 8,000 years ago when the Baltic Sea changed from

deg-a freshwdeg-ater system to deg-a brdeg-ackish system due to the rising sedeg-a level deg-and deg-a connection to the North Sedeg-a (the rina Sea stage of the Baltic Sea) By simulating the sedimentary history of the methane cycle since its inception, the required timescales for the development of a methanogenic zone and for free gas formation in Baltic Sea sediments was reconstructed Figure 24 shows the benthic biogeochemical dynamics near the center of the Ar-

Litto-Fig 24 Development in particulate organic carbon (POC, i.e organic matter deposition; brown line), sulfate (SO 4 2- ; blue line), dissolved methane (CH 4(aq) ; red line), and free methane gas (gray shaded curve) in Baltic Sea sediment (Arcona Basin) starting 8,000 years before present (BP) when the environment gradually turned brackish The maximum depth penetration of POC (vertical brown line) equals the thickness of the Holocene mud layer An example from the Arkona

kona Basin as an example Simulations revealed that sulfate diffusion and sulfate-reduction controlled the fate

of organic matter during the first 3 kyr of the Littorina Sea stage Thereafter, organic carbon degradation ceeded the rate at which sulfate was transported to the deeper sediment layers and methanogenesis occurred Almost concomitantly, anaerobic oxidation of methane began to consume the sulfate diffusing down from above the methanogenic zone and also the residual sulfate pool within the glacial sediment below the muddy layer Consequently, the sulfate-methane transition shoaled upwards towards the sediment-water interface A further

ex-3 kyr until 1.6 kyr BP were required for the dissolved methane to reach the in situ solubility limit and form free methane gas Over the last 2 kyrs, the gas volume fraction increased to reach a contemporary concentration of about 5 % by volume Repeating such simulations at selected locations in the basin has also allowed to delineate the zones where aqueous and gaseous methane are present and to construct a basin-scale methane budget 1.5.1.2 Forecasting the impact of climate change on methane gas inventories

A reactive-transport model, which accounts for the effect of climate change on the productivity, bottom-water temperature and salinity of the Baltic Sea, has been applied to forecast the evolution of the seafloor methane gas inventory (Fig 25) Full transient simulations were performed for the period 2010-2110, using boundary conditions extracted from a 3-dimensional ecosystem model of the Baltic Sea3 forced by a regional dataset of

greenhouse gas emissions (IPCC scenario A1B) The results obtained for the Arhus Bay transect reveal that the temperature rise of circa 1.8 de-grees predicted for the period 2010-2100 could trigger a significant increase in gaseous methane inventory, move the gas front closer to the ment-water interface and lead to the formation of gas at stations where there currently is none (sta-tion M26) Similar results have been obtained in shallow sediments of the Bothnian Bay, where gas production is enhanced by the combined effects

of temperature and decrease in bottom-water sulfate induced by the freshening of Baltic Sea Altogether, these factors could favor methane release from the seabed, although this remains essentially unknown In the example below from Aarhus Bay, model results reveal that if gas mi-grates upwards through the sediment, most (if not all) of this gaseous methane is concurrently re-dissolved and oxidized during transit towards the sediment-water interface This gas movement should theoretically occur if the gas pressure ex-ceeds the pore throat entry pressure but is not sufficiently high to fracture the sediment No gas fractures were observed in the sediment, and neither was gas escape into the water column The exact mechanisms of gas advection and dissolu-

3

Neumann, T (2010) Journal of Marine Systems, 81, 213-24

Fig 25 Concentration profiles of sulfate, methane and

methane gas at 4 stations in Aarhus Bay with increasing

mud thickness denoted by the horizontal line (see also Fig

16 and 17) The stations are approximately 50 m apart The

sulfate-methane transition is indicated by the gray shaded

band The top panels represent the present day (steady

state) situation, and the lower panels are those where the

model is run for 100 yr imposing a +0.018 o C yr -1 change in

temperature in the bottom water.

tion remain uncertain and prediction of how gas migration will respond to future environmental and climate changes (e.g through the onset of sediment fracturing) remains similarly uncertain (Deliveable 3.2) Sensitivity studies show that the methane flux (aqueous + gaseous) to the water column forecasted for the year 2100 is highly dependent to the controlling processes, with high advection rates and/or slow dissolution rates promot-ing the propensity for methane escape (results not shown) The model results represent the first data-

supported predictions of future methane fluxes in the Baltic Sea Yet, further research in this area is essential for

a more accurate forecasting of the role of Baltic Sea sediments to green-house gas emission and thereby to mate-induced warming

cli-1.5.2 Environmental controls on gaseous methane production in the Baltic Sea (an example from Aarhus Bay) The methane dynamics in the Baltic Sea are closely related to the deposition and build–up of an organic–rich marine mud layer which began around 8 kyr ago as a result of rising sea level and brackish-water inundation of the Baltic Sea that we know of today This Holocene mud overlays organic–poor silty sediments deposited under freshwater of glacial origin Because of the uneven topography at the upper fringe of the freshwater sediment, the thickness of the overlying marine mud deposits is often highly variable Numerous seismic observations throughout the Baltic Sea reveal that the formation of methane gas only occurs once a critical mud thickness is surpassed As an example, a seismic profile from Aarhus Bay at the entrance to the Baltic Sea is shown in Fig 16 and illustrates that the appearance of free methane gas in this case occurs where the mud layer exceeds about

10 m Yet, this depth is not fixed, but varies over Aarhus Bay and over the Baltic Sea in general

Reactive-transport modeling was applied to (1) identify the main controls of methanogenesis and gas formation in the seabed and (2) derive a mechanistic explanation for the abrupt appearance of gas when a critical mud thickness

is reached The study area covered a mud lens in Aarhus Bay (Denmark), sampled for concentrations and rates at

7 stations along a transect characterized by increasingly thicker Holocene mud (Fig 16)

Numerical simulations show that the main trigger for gas formation is the bulk sediment accumulation rate ciated with increasing mud thickness High accumulation rates dilute the organic material deposited on the sea floor with inorganic material and lead to a more rapid burial of reactive organic matter down into the methano-

asso-genic zone, with resulting higher rates of genesis as well as gas production This is illustrated

methano-in Fig 25 (upper panel), where the sedimentation rate increases from 110 cm kyr-1 at station M25 to

152 cm kyr-1 at station M28 and where gas forms when the mud thickness becomes larger than ̴10

m (station M27) The model captures also the tion of the gas layer, the so called Free Gas Depth (FGD) corresponding to the uppermost depth where gas first occurs, by allowing methane gas to advect upwards through the sediment If the gas did not move but instead remained in situ, a hypo-thetical simulation from site M29 shows that the FGD would not rise above 850 cm and the simulat-

posi-ed sulfate penetration depth would be about one meter deeper than observed (Fig 26a,b) Gas ad-vection is accompanied by gas dissolution in the

Fig 26 Simulated (curves) and measured (symbols) pore

water concentrations and free gas volumes at station M29

(see Fig 17) (a) Without allowing for gas advection

through the sediment, (b) with gas advection, (c) δ13 C

isotopic distributions of dissolved inorganic carbon without

gas advection and (d) δ13 C isotopic distributions of

dis-solved inorganic carbon with gas advection The gray band

indicates the sulfate-methane transition zone (SMTZ)

pre-dicted by the model

zone of Anaerobic Oxidation of Methane (AOM) Since this process consumes dissolved methane and is the prime cause for bringing the methane concentration down below saturation, AOM can be likened to a geochem-ical barrier for gas escape Modeling of stable carbon isotope distributions support further the hypothesis that methane gas advection and dissolution occur in the AOM zone (Fig 26d) Without this mechanism, the AOM rates would be significantly lower and would lead to simulated δ13C isotopic distributions that were significantly heavier (less negative) than the measured values (Fig 26c) The suite of data and model results are nevertheless consistent with the idea that all the methane transported by diffusion and gas migration is ultimately consumed

by AOM, and consequently only minor or no methane currently escapes to the ocean–atmosphere (Deliverable 5.3)

1.5.3 Regionalization and budgeting of methane cycle

In sediments of the Baltic Sea high concentrations of methane (CH4) were observed by biogeochemical as well as geophysical investigations Biogeochemical investigations were based on sediment and pore water sampling at selected sites and subsequent chemical and microbiological analyses This provided detailed information about the production and fate of biogenic methane, generated by microbially mediated processes, as well as the po-tential release of this greenhouse gas into the water column Geophysical methods like shallow seismic surveys provided new information about the spatial distribution of free gas (methane gas bubbles) in sediments Data derived by biogeochemical or geophysical studies provided very detailed information for selected sites as well as along survey lines Nevertheless, the spatial coverage of these studies was – due to the time consuming and costly techniques – rather sparse For considerations of large scale spatial patterns and budgets, a combina-tion of elaborate, site specific measurements with geophysical data on forcing factors which are available with sufficient spatial coverage, are required This supports the computation of methane budgets as well as identifi-cation of regions where high or low methane concentrations are expected

For spatial modeling, forcing factors like the accumulation rate of particulate organic matter (POC), the content, bottom water concentrations of e.g sulfate and oxygen, bathymetry, slope, morphological units, bot-tom water temperature, and current speed as well as indicators for methane formation like pockmarks were considered All data were projected and combined applying the Lambert azimuthal equal-area projection and using similar grid size By statistical analysis we compared the former mentioned parameters for regions where free gas was observed with regions in the surrounding where free gas was not observed (Fig 27) The spatial modeling was applied for the different sub-regions of the Baltic Sea For each region, a set of factors was derived that are likely to contribute to the formation of free gas These factors were iteratively improved and applied to compute predictive maps about the spatial distribution and the total area of free gas in sediments of the Baltic Sea (Fig 28)

POC-Fig 27 (A) Locations in the Baltic Sea where free gas were observed in surface sediments (red polygons) For spatial

analysis POC (i.e particulate organic carbon) accumulation rates (B) or oxygen concentrations in bottom water (C) were considered as forcing factors From statistical analysis of forcing factors related to the formation of free gas in surface sediments weighting coefficients were derived.

1.6 Deliverables

WP1.1: BALTIC GAS web-page

www.balticgas.au.dk(Bo Barker Jørgensen, Henrik Fossing)

WP1.2: Scientific reports (Y1, Y2, final)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp1-deliverable-no-2-scientific- reports/

(Bo Barker Jørgensen, Henrik Fossing)

WP1.3: BALTIC GAS Workshops and meetings (reports)

meetings-reports/ (Bo Barker Jørgensen, Henrik Fossing)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp1-deliverable-no-3-baltic-gas-workshops-and-WP1.4: Submission of data to a common database

common-database/ (Bo Barker Jørgensen, Henrik Fossing)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp1-deliverable-no-4-submission-of-data-to-a-Fig 28 Spatial distributing of free gas (modeled) in Baltic Sea (except Gulf of Bothnia and Finland) showing the

proba-bility to find free gas in surface sediments The prediction is based on analyses of parameters like particulate organic carbon (POC) accumulation, O 2 and SO 4 2- concentration in bottom water as well sediment type, observed within known free gas areas The data set was factorized to obtain a prediction for the occurrence of free gas areas within the Baltic Sea This procedure was optimized by comparison of the similarity of the spatial distribution of known free gas and predicted free gas areas Based on this comparison predicted probability levels of gas occurrence (low, medium-low, medium, high) were assigned

WP1.5: Research cruises

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp1-deliverable-no-5-research-cruises/

(Bo Barker Jørgensen, Henrik Fossing)

WP2.1: GIS-map of mined data

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp2-deliverable-no-1-gis-map-of-mined-data/ (Jørn B Jensen, Bo Barker Jørgensen)

WP2.2: GIS-map of methane flux and distribution in sediments

distribution-in-sediments/ (Michael Schlüter, Torben Gentz, Roi Martinez, Jørn B Jensen, Laura Lapham)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp2-deliverable-no-2-gis-map-of-methane-flux-and-WP2.3: GIS-map of hot-spots of present and future CH4-emission

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp2-deliverable-no-3-gis-map-of-hot-spots-of-present-and-future-ch4-emmission/ (Michael Schlüter, Torben Gentz, Roi Martinez)

WP3.1: Mapping of shallow gas and physical characterisation of gas-bearing sediments

physical-characterisation-of-gas-bearing-sediments/ (Jørn B Jensen, Rudolf Endler)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp3-deliverable-no-1-mapping-of-shallow-gas-and-WP3.2: Identification of zones of potential weakness

potential-weakness/ (Gregor Rehder, B.B Jørgensen)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp3-deliverable-no-2-identification-of-zones-of-WP3.3: Detection and monitoring of methane ebullition

methane-ebullition/ (Jens Schneider v Deimling, Wanda Gülzow, Marina Ulyanova, Zygmunt Klusek, Gregor Rehder)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp3-deliverable-no-3-detection-and-monitoring-of-WP4.1: Methane distributions and breakdown

breakdown/ (Timothy G Ferdelman, Volker Brüchert, Sabine Flury, Henrik Fossing, Bo Barker Jørgensen, Laura Lapham, Nikolay Pimenov, Maja Reinholdsson, Nguyen M Thang)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp4-deliverable-no-1-methane-distributions-and-WP4.2: Methane emission through sediment-water and sea-air interfaces

sediment-water-and-sea-air-interfaces/ (Volker Brüchert, Timothy G Ferdelman, Henrik Fossing, Wanda Gülzow, Laura Lapham, Gregor Rehder, Nguyen Thanh Manh, Jens Schneider von Deimling, Torben Gentz, Michael Schlüter)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp4-deliverable-no-2-methane-emission-through-WP4.3: Holocene evolution of the Baltic Sea ecosystem

baltic-sea-ecosystem/ (Daniel Conley, Maja Reinholdsson, Conny Lenz, Lovisa Zillén)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp4-deliverable-no-3-holocene-evolution-of-the-WP4.4: Submitted MS on: Sulphur and methane biogeochemistry

Flury, S., A.W Dale, H Røy, H Fossing, J.B Jensen, B.B Jørgensen (submitted) Methane fluxes and shallow gas formation controlled by Holocene mud thickness in Baltic Sea sediments Geochimica et Cosmochimica Acta WP5.1: Transport/ reaction models reg methane and sulphur dynamics

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp5-deliverable-1/ (José Mogollón and Pierre nier)

Reg-WP5.2: Predictive model and climate change scenarios

change-scenarios/ (Pierre Regnier and Andy Dale)

http://balticgas.au.dk/balticgasaudk/project/deliverables/wp5-deliverable-no-2-predictive-model-and-climate-WP5.3: Submitted MS on: Integration gas, acoustics and biogeochemistry

Dale, A.W., S Flury, P Regnier, H Røy, H Fossing, B.B Jørgensen (submitted) Coupling between sis, anaerobic oxidation of methane and δ13C distributions in gassy sediments from the Baltic Sea (Aarhus Bay) Geochimica et Cosmochimica Acta

methanogene-2 Further research and exploitation of the results

2.1 Further research

BALTIC GAS has generated a large and high-quality dataset for the distribution and dynamics of methane, bly the most comprehensive methane dataset for any marginal sea This has been possible through the acquisi-tion of new data during the many research cruises and through the mining of existing data The information was compiled in GIS maps that provide geographical overviews and information on the parameters controlling me-thane accumulation and turnover Such GIS maps are based on our current understanding of statistical and causal relationships between sediment properties, water quality, and microbial processes The GIS maps should

proba-be considered as dynamic, however, and will proba-be improved as further data proba-become available or as the algorithms for the calculation of derived properties are adjusted

Researchers of the BALTIC GAS project have combined seismo-acoustic mapping of gas distribution and mentology with biogeochemical point analyses An important verification and quality control of the developed GIS maps is therefore a further targeted core sampling and analysis to check the properties predicted by the GIS algorithms Key areas for such verification include the Bornholm Basin in which the latest interpretation phase has shown that it is possible to map late Holocene subunits like the Medieval Warm Period and the Little Ice Age Future research would be able to focus on such Holocene time intervals and investigate methane production and flux in these intervals

sedi-Due to the limited capacity of BALTIC GAS, the project focused on the main sedimentary basins in the Baltic Sea where most methane is supposedly generated The coastal zone is, however, much more dynamic and is also much more sensitive to local effects such as sewage outlets (e.g Himmerfjärden) or river outlets (e.g Bay of Gdansk) Our data from the Swedish archipelago indicate that the coastal zone methane emission to the atmos-phere may be ten-fold higher than in the central Baltic Sea where most studies were done Information on the source strength of the coastal zone for the overall emission of methane is therefore needed in order to develop

a methane budget for the entire Baltic Sea In this respect data are largely lacking for the Bothnian Gulf (i.e Bothnian Sea and Bay) and in the Gulf of Finland compared to the southern part of the Baltic Sea Future moni-toring of methane requires the installation of more ship-based and coast-based continuous measurement sys-tems to extend areal and temporal coverage

As a part of BALTIC GAS we searched for point sources of gas ebullition and for sediment structures (pockmarks)

as indicators for such ebullition The project reached the conclusion that pockmark areas are rather few and that