Shallow-Water Pockmark Formation in Temperate Estuaries: a Consideration of Origins in the Western Gulf of Maine with Special Focus on Belfast Bay

Marine Geology 225 (2006) 45–62 www.elsevier.com/locate/margeo

Shallow-water pockmark formation in temperate estuaries: A consideration of origins in the western gulf of Maine with special focus on Belfast Bay

Jeffrey N. Rogers a,*, Joseph T. Kelley b, Daniel F. Belknap b, Allen Gontz b, Walter A. Barnhardt c

a GeoSyntec Consultants, 289 Great Rd Suite 105, Acton, MA 01720, USA b Department of Earth Sciences, University of Maine, Orono, ME 04469-5711 USA c U.S. Geological Survey, Woods Hole, MA, USA Received 29 September 2004; received in revised form 15 July 2005; accepted 19 July 2005

Abstract

A systematic mapping program incorporating more than 5000 km of side scan sonar and seismic reflection tracklines in the western Gulf of Maine has identified more than 70 biogenic natural gas deposits, occupying 311 km2 in nearshore muddy embayments. Many of these embayments also contain pockmark fields, with some exhibiting geologically active characteristics including the observance of plumes of escaping fluids and sediment. Pockmarks, hemispherically shaped depressions of various size and depths, formed through fluid escape of gas and/or pore water, are sometimes found within or outside gas fields, although many gas fields lack pockmarks altogether. Although the origin of the natural gas remains unclear, if coastal environments at times of lower sea level were similar to the present, numerous lake, wetland, valley fill and estuarine sources of organic-rich material may have formed on the inner shelf. If these deposits survived transgression and remain buried, they are potential gas sources. Intensive mapping of the Belfast Bay pockmark field in 1998 produced the first nearly continuous side scan sonar mosaic of a Gulf of Maine pockmark field with a corresponding 3-dimensional geological model generated from seismic data. Statistical analysis of pockmark geometry, gas deposit loci, and subsurface evidence for gas-enhanced reflectors suggest that gas migration from deeper lateral sources along permeable subsurface strata may be the mechanism for pockmark formation in areas lacking gas- curtain seismic reflections. The coarse-grained transgressive ravinement unconformity between Pleistocene glacial-marine mud and Holocene mud may act as a conduit for distributing methane to the field’s margins. D 2005 Elsevier B.V. All rights reserved.

Keywords: methane; Belfast Bay; Maine; Quaternary; marine mud; sea-level change

1. Introduction

Pockmarks were first described as bconcave, crater- like depressions that occur in profusion on mud bot- toms across the Scotian ShelfQ (King and MacLean, * Corresponding author. Tel.: +1 978 263 9588; fax: +1 978 263 1970, p. 3141). They reach diameters of hundreds of 9594. meters and depths of tens of meters. Pockmarks are E-mail address: [email protected] (J.N. Rogers). recognized in a variety of continental margin settings

0025-3227/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.margeo.2005.07.011 46 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 all over the world (Hovland and Judd, 1988). Pockmark Fader, 1991; Kelley et al., 1994; Wever et al., 1998), fields are commonly found near deltas (Nelson et al., freshwater escape (Albert et al., 1998; Bussman and 1979) and areas of petroleum production (Hovland et Suess, 1998; Whiticar, 2002); and ice rafting (Paull et al., 1987, Uchupi et al., 1996), or tectonic activity al., 1999) were each invoked to explain the origin of (Field et al., 1982; Hasiotis et al., 1996; Vogt et al., pockmarks within geologically young sediments. In 1994). Shelf basins (Fader, 1991; Josenhans et al., some locations in both shallow and deep water, pock- 1978; McClennen, 1989), continental slopes and rises mark fields apparently lacking measurable quantities of (Orange et al., 1999; Paull et al., 2002) also host fields natural gas or escaping freshwater, were deemed of pockmarks. In all these locations, fluid escape is binactiveQ by Paull et al. (2002) and Ussler et al. (2003). invoked as the forcing mechanism for pockmark for- In discussions of pockmark origins it is essential to mation. Thermogenic and biogenic natural gas from distinguish between the origin of the fluid(s) and the petroleum occurrences, organic-rich sedimentary process by which the fluid(s) form and maintain a deposits and methane hydrates are probable sources pockmark. In coastal Maine (Fig. 1) neither of these for the fluids and buoyancy required to form pock- origins is established, and both thermogenic and bio- marks, although escaping groundwater (Whiticar, genic gas have been postulated to exist (Fleischer et al., 2002) and other more exotic mechanisms (ice rafting, 2001). Acoustic wipeout zones, which indicate the Paull et al., 1999; meteorites, Nelson et al., 1979) have presence of gas, are common in seismic reflection been suggested as causative agents. profiles of the northwestern Gulf of Maine, as are Pockmarks are also widespread in mid-latitude estu- pockmarks (Belknap et al., 1986, 2002; Kelley et al., aries (Fleischer et al., 2001; Missiaen et al., 2002; Gar- 1989b). However, a one-to-one correspondence be- cia-Gil et al., 2002; Ussler et al., 2003), especially in tween pockmarks and gas does not exist in coastal formerly glaciated regions (Scanlon and Knebel, 1989; Maine because biogenic gas fields often lack pock- Barnhardt and Kelley, 1995; Kelley et al., 1994; Fader, marks, and pockmarks without associated gas are also 1991; Whiticar, 2002). Biogenic gas (Albert et al., 1998; observed (Kelley et al., 1994; Gontz et al., 2002).

Fig. 1. The location of the study area in the western Gulf of Maine. Boxes locate other figures. (a) The area inside the box is enlarged in b; (b) the shoreline of the State of Maine with natural gas fields shaded green (gas only) and red (gas and pockmarks). J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 47

Based on one recent set of very shallow cores and till across the landscape and present seafloor (Kelley passive sensing, Ussler et al. (2003) concluded that and Belknap, 1991; Shipp et al., 1989, 1991; Kelley et neither groundwater nor natural gas escape was detect- al., 1998). Moraines stretch across many embayments able in the Belfast Bay, Maine field that was earlier (Shipp et al., 1987, Knebel, 1986; Knebel and Scanlon, considered to have an active pockmark formation (Kel- 1985; Miller, 1998) and probably acted as temporary ley et al., 1994). barriers to rising sea level in the Holocene (Barnhardt The mechanism of pockmark formation is compli- and Kelley, 1995; Shipp, 1989). Numerous lakes and cated and requires more study. In addition to ques- wetlands are framed by bedrock and glacial material tions of pockmark origins, Kelley et al. (1994) along the present coast. described other distinctive attributes of the Belfast The Presumpscot Formation glacial-marine sediment Bay field, such as acoustically reflective beyesQ in (Bloom, 1963) covers rock and till outcrops in most the center of some pockmarks, reflective circular coastal areas seaward of the late Pleistocene marine limit structures (bdark spotsQ) lacking significant bathymet- (Thompson and Borns, 1985; Dorion et al., 2001). ric relief, and linear chains including dozens of pock- Glacial-marine sediment often interfingers with coarse- marks and extending for more than a kilometer that grained, ice-proximal outwash deposits near its base were unexplained. (Thompson, 1982; Belknap and Shipp, 1991). The ma- In this paper we describe the geological setting in jority of the Presumpscot Fm. is a slightly sandy mud which natural gas and pockmarks exist along more than with ice-rafted dropstones. The upper part of the Pre- 400 km of inner continental shelf and estuaries in the sumpscot Fm. sometimes contains marine fossil mol- northwestern Gulf of Maine. Although we focus on lusks and barnacles, macroalgae and abundant organic Belfast Bay, Maine, we present geophysical, strati- matter, and shows little direct glacial influence (Fig. 2a) graphic, and sedimentological data from all along the (Stuiver and Borns, 1975; Retelle and Weddle, 2001). Maine coast to develop a model of pockmark formation Stratified glacial-marine muddy sediment is the most applicable to pockmarks in mid-latitude, formerly gla- common deposit along the inner shelf, and completely ciated settings. fills some bedrock basins with more than 50 m of sediment (Barnhardt et al., 1997). On land, thick, rela- 2. Geological Setting tively impermeable deposits of till and glacial-marine sediment fill many former river valleys, deranging con- The study area is in the northwestern Gulf of Maine, temporary drainage and ponding water into thousands of on the inner shelf and littoral of Maine, between the lakes and wetlands (Thompson and Borns, 1985; Tol- shoreline and the 100 m isobath (Fig. 1). This area is man et al., 1986). framed by Precambrian through Mesozoic bedrock, but The isostatic effects of glaciation have strongly in- primarily by Paleozoic igneous and metamorphic rocks fluenced sea-level during the late Quaternary (Belknap (Osberg et al., 1985). Differential erosion of the regional et al., 1987a,b; Belknap and Shipp, 1991; Kelley et al., bedrock, most recently by glaciation, shaped the coast 1992; Barnhardt et al., 1995, 1997). Marine incursion and inner shelf regions (Johnson, 1925; Barnhardt et al., accompanied deglaciation and glacial-marine sediment 1996a,b,c,d,e,f,g; Kelley, 1987; Kelley et al., 1998; crops out as eroding bluffs along the coast (Fig. 2; Uchupi, 2004). Thus, prominent headlands are sup- Thompson and Borns, 1985; Kelley and Dickson, ported by erosion-resistant rocks, while embayments 2000; Kelley, 2004). Retreat of the ice led to isostatic are carved from more easily eroded material. Peninsulas uplift, and sea level fell to approximately À60 m by of resistant rock continue offshore as chains of islands 10.8 ka (uncalibrated; Barnhardt et al., 1995, 1997; and shoals; and associated estuaries project seaward as Shipp et al., 1991; Shaw et al., 2002). Since then, sea deeper basins (Belknap et al., 1986; termed bNearshore level has risen at varying rates up to the present (Barn- BasinsQ in Kelley et al., 1989a,b, 1998). The landward hardt et al., 1995, 1997; Kelley et al., 1995; Gehrels et extension of estuaries is often a fresh/brackish water al., 1996). Between the lowstand and the present shore- lake separated from the sea by rapids or waterfalls line a transgressive unconformity with Holocene mud (Shipp, 1989; Hannum, 1997). There is frequently an or sand overlying glacial sediment is widely observed abrupt transition from a salt marsh to a freshwater (Belknap and Shipp, 1991; Barnhardt et al., 1997; wetland in many upper estuaries. (Kelley et al., 1988). Belknap et al., 2002; Kelley et al., 2003). Offshore of Late Wisconsinan glaciers retreated from the Maine beaches a sandy-muddy estuarine unit is recognized coast between 15 and 13 ka (Dorion et al., 2001; Retelle between the glacial-marine mud and modern shoreface and Weddle, 2001). They left heterogeneous deposits of sand (Barnhardt et al., 1997; Kelley et al., 2003, 2004). 48 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62

of high resolution seismic reflection data (ORE bboomerQ system; Raytheon RTT 1000a 3.5 kHz) and more than 3300 km of slant-range corrected, 100 kHz, side-scan sonar (EG and G SMS Model 260) data from many sites on the inner shelf over the past 20 yr. Much of this is published and synthe- sized in a series of Geographic Information System (GIS) maps of the inner shelf (Barnhardt et al., 1996a,b,c,d,e,f,g, 1997; Kelley et al., 1998, 2005). This geophysical survey covers most of the coastline in a reconnaissance fashion, and in many locations side-scan sonar mosaics with accompanying seismic reflection data exist (Barnhardt and Kelley, 1995; Dickson, 1999; Kelley et al., 1994, 2003; Rogers, 1999; Gontz, 2002; Gontz et al., 2002). More than 100 offshore vibracores and 1700 bottom samples were collected to ground-truth the remotely sensed data, and more than 50 submersible and remotely operated vehicle (ROV) dives were also made to examine details of the seafloor (Barnhardt et al., 1996a,b,c,d,e,f,g; Kelley et al., 1998). We draw upon all of these studies to reach generalizations regarding pockmarks and natural gas in our offshore region. All natural gas fields were mapped in detail by Barnhardt et al., 1996a,b,c,d,e,f,g based on seismic reflection records, and are compiled in Fig. 1. Earlier observations that concern natural gas and pockmarks in the Fig. 2. (a) An eroding outcrop of glacial-marine sediment in Jone- region include a study of seafloor sediment properties sport, Maine showing abundance of black organic matter in the upper and seismicity (Kelley et al., 1989b), reconnaissance part of the section. An unconformity separates freshwater peat from marine clay (dashed line); (b) freshwater peat deposits commonly rest observations of bays with pockmarks (Barnhardt and on glacial-marine mud, as in this eroding section in Lubec, ME. Kelley, 1995; Gontz et al., 2002), and detailed studies Location of photos shown in Fig. 1. of the Belfast Bay pockmark field (Scanlon and Kne- bel, 1989; Kelley et al., 1994; Rogers, 1999; Ussler et al., 2003; Gontz, 2002, 2005). In protected estuaries a sandy-muddy estuarine unit A side-scan sonar mosaic of the Belfast Bay pock- often rests between glacial-marine mud and Holocene mark field, Maine was created with analogue equip- mud (Kelley et al., 1992). The estuarine units also ment in 1989 (Kelley et al., 1994) and expanded in contain gravel, shells and wood fragments and range 1998 (Rogers, 1999) for this paper. The 1998 data were up to several meters in thickness with radiocarbon scanned and image-processing techniques were used to dates between 9.2 and 7.2 ka (Shipp, 1989; Kelley manipulate the digital sonar lines, which overlapped by et al., 1992; Barber, 1995; Barnhardt et al., 1995, 25% (Rogers, 1999). Differentially corrected global 1997; Knebel, 1986). Freshwater lake deposits be- positioning system (DGPS) geographic information tween the lowstand and modern shorelines have was annotated every minute onto each of the images been postulated (Shepard, 1963) and interpreted on by the sonar recording device and used to brubber the basis of seismic stratigraphic observations (Han- sheetQ the side scan record in a Geographic Information num, 1997), but never confirmed with core samples. System (Rogers, 1999). The NOAA hydrographic ship, R/V RUDE, gathered multibeam bathymetric data from 3. Methods the Belfast Bay field in 1998 with a Reson 8125 shallow water multibeam system operating at 455 To evaluate the regional Quaternary geology of the kHz (NOAA, 2003). We use an illustration from their Western Gulf of Maine, its sea-level history and mod- work, but we have not yet been able to directly use ern seafloor habitats, we gathered more than 5000 km those bathymetric data. J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 49

4. Results beneath pockmarks, allowing seismic penetration to bedrock (Fig. 4). 4.1. Regional gas and pockmark distribution Gas-bubble-rich zones typically begin 3–5 m below the seafloor (Barnhardt and Kelley, 1995, their Fig. 7), Natural gas can be recognized from characteristic but they are observed to top out at greater than 10 m seismic reflection signals. In small concentrations, gas depth as well. We have no direct evidence of their full bubbles scatter acoustic energy and enhance seismic thickness. In areas without pockmarks, gas bubbles may reflections (Fig 3a). We commonly observe this phe- extend to the surficial seismic reflection (Belknap et al., nomenon at the surface of what is interpreted as glacial- 1986, 1994). Gas is recognized in every muddy embay- marine sediment. The base of the Holocene sediment, ment examined along the Maine coast (Fig. 1). Despite which unconformably overlies the glacial-marine mate- extensive geophysical surveys, no gas or pockmarks rial, is commonly coarse-grained material with shell were observed in the sandy embayments of southern and wood fragments compared to uniformly muddy Maine (Fig. 1)(Kelley et al., 2003). We mapped gas at sediment above and below. Increasing concentrations water depths ranging to greater than 100 m, although of gas lead to areas of bacoustic turbidityQ and bacoustic most gas was observed shallower than 70 m. The spatial wipeoutQ (Schubel, 1974; Judd and Hovland, 1992; extent of mapped gas is 311 km2 within 70 specific Taylor, 1992), where acoustic energy is so scattered fields (Fig. 1), but this is a minimum estimate that is by gas bubbles that other reflectors are masked (Figs. constrained by the limits of existing seismic data. Large 3b, 4). The terms bcurtainQ or bblanketQ are also used, regions of muddy seafloor lack seismic observations, depending on the extent of the gas (Taylor, 1992; and the total area of gas and pockmark fields is probably Garcia-Gil et al., 2002). Gas is usually absent directly much larger than shown in Fig. 1.

Fig. 3. Examples of gas in seismic records. As in earlier work (Barnhardt et al.,1997) Br is interpreted as bedrock, Gm as glacial-marine sediment, T as till, M as Holocene mud, Ng as natural gas in this and all other illustrations. (a) Small areas of gas are closely associated with enhanced acoustic reflectors within which gas apparently migrates up dip. Pockmarks exist above one of these reflectors. See Fig. 6 for location. (b) Gas enhances some reflectors in this section, and partly obscures the deeper subbottom returns. The question mark refers to the uncertainty in interpretation of an acoustically layered unit above glacial-marine sediment and at the base of the Holocene mud. Line is located in gas field 10 km to the east of Belfast Bay. 50 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62

Pockmarks typically occur within and around fields of natural gas (identified in seismic reflection profiles) (Fig. 6). Gas is not usually observed directly beneath a pockmark, allowing imaging of subbottom reflectors beneath pockmarks that are obscured by gas nearby (Fig. 6). This relationship suggests that the escape of the gas and associated pore waters formed the pockmark. Many large gas fields do not appear associated with pockmarks (Fig. 1), however, and some pockmarks have no gas near them (Figs. 3, 4; Kelley et al., 1994, their Fig. 3; Ussler et al., 2003). Side-scan sonar images of pockmark fields in a few cases reveal plumes of fluid and sediments that suggest active venting (Fig. 5)(Kelley et al., 1994; their Fig. 2). Anecdotal observations of bubble eruptions were reported to us by people who fish, similar to reports from other locations (Garcia-Gil et al., 2002). Ussler et al. (2003) suggest that the acoustic reflections may represent fish bladders. However, in our long experi- ence with this side-scan sonar system we have seen numerous fish schools, but we have never observed them in a distinctive plume structure rising from the seafloor. The plumes are densest at the seafloor over the Fig. 4. Side-scan sonar and simultaneously collected seismic reflec- pockmark, follow a parabolic curve into the water tion profiles over pockmark field in Belfast Bay. See Fig. 8 for column while gradually fading, and are associated location. Note that beneath pockmarks, deeper reflections are ob- with bclotsQ of scattered high reflectivity. We interpret served because of the lack of gas. bEyesQ appear where pockmarks extend down to glacial-marine surface. this as a gas and sediment plume (Kelley et al., 1994). Cores into gas-charged sediment always recover fine mud, and show expansion cracks upon opening (Barn- The most common regional occurrence for gas is hardt et al., 1997, their Fig. 10). Sediment correlated within a basin or paleovalley incised into the Presumps- with gas in seismic records is dark gray to black, unlike cot Fm., above dipping reflectors that we interpret to more oxidized overlying sediment (Barnhardt et al., represent the basal unconformity at the surface of gla- 1997). Cores of muddy sediment in inter-pockmark cial-marine sediment (Figs. 3, 4)(Belknap and Shipp, areas often contain alternating oxidized and reduced 1991; Barnhardt et al., 1997; Belknap et al., 2002). In beds, suggestive of intermittent sediment eruptions. seismic reflection surveys down the axis of elongate contemporary estuaries, natural gas usually occurs in 4.2. Belfast Bay elongate deposits that obscure the deepest basins (Fig. 1). These basins are typically separated by a seaward Belfast Bay is located in the northwest corner of deepening series of bedrock sills (Shipp, 1989; Belknap Penobscot Bay on the central Maine coast (Fig. 1) and et al., 1986, 1987, 1994), as are lakes that occur along was one of the first estuaries examined with seismic rivers upstream of the sea. In more open embayments, reflection methods (Ostericher, 1965). Whether it is gas is recognized in larger, more irregularly shaped similar to the many other pockmark fields in the region basins, similar to lakes in nearby terrestrial settings is unknown, but its pockmark field has been studied (Kelley et al., 1994; Gontz, 2002; Gontz et al., 2002; more than any other in Eastern North America (Scanlon Barnhardt and Kelley, 1995; Fader, 1991). In Somes and Knebel, 1989; Kelley et al., 1994; Rogers, 1999; Sound, a fjord-like basin cut into Mt. Desert Island, Ussler et al., 2003; Gontz, 2002, 2005). natural gas and pockmarks fill the region behind mor- Pockmarks cover the seafloor of this shallow muddy aines (Barnhardt and Kelley, 1995). At a time of lower- embayment (Scanlon and Knebel, 1989; Kelley et al., than-present sea level, these basins were also probably 2005; Rogers, 1999; Gontz, 2002)(Fig. 6). Water depths lakes like the many modern lakes on the island, which range from 10 to 70 m and in the southeastern corner of are impounded by moraines. the field the bathymetry is almost completely controlled J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 51

Fig. 5. Side-scan sonar observation of a pockmark experiencing an eruption of gas and sediment (black marks on record labeled bwater-column debrisQ; modified from Kelley et al., 1994). by edge-to-edge pockmarks (Fig. 7). Some of the plification was used for estimation of diameters. The world’s largest known pockmarks are located in this mean pockmark diameter was 47 m (median 43 m). A field, with diameters of nearly 300 m and relief of 20 histogram of frequency versus diameter shows that the to N30 m. A prominence at the south end of the embay- distribution of pockmark size is skewed towards larg- ment between the mainland and Islesboro Island (Kne- er diameter pockmarks (Fig. 8). Although the side- bel and Scanlon, 1986) may have cut the area off from scan sonar can image high-contrast features smaller ocean waters prior to the mid-Holocene. This promi- than 0.5 m, the apparent lack of small pockmarks (b3 nence, interpreted as an end moraine (Knebel and Scan- m) probably results from the fact that as the pock- lon, 1986), has a relief of 30 m and a depth below marks decrease in diameter they also decrease in present sea level of 45 m, 15 m above the lowstand depth, such that they resemble the flat seafloor and position (Barnhardt et al., 1997). Like Somes Sound, become harder to resolve because there are no high- this moraine may have impounded fresh water and reflective angles. The largest pockmarks are compos- formed a shallow lake or bog. To the east, the pockmark ite features formed by the coalescence of intermediate field is partly bounded by Isleboro Island as well as an size pockmarks. opening to East Penobscot Bay. This opening is shallow The bathymetric relief of pockmarks captured direct- and covered by Holocene mud overlying a flat-topped ly under the towfish were recorded in the field as a expanse of the Presumpscot Fm. cut by an unconformi- third-channel display of depth under the towfish. Of ty. We interpret this surface as the transgressive bluff-toe 2300 pockmarks, depth information was obtained for unconformity (Belknap et al., 2002) that marks the 11% of the population. A graph of recorded depths retreat path of bluffs of glacial-marine sediment that versus pockmark diameters displays a crudely linear formerly formed an isthmus between Sears Island and trend that flattens out toward larger sizes. An empirical northern Islesboro Island. This isthmus blocked Belfast fit for the data (Fig. 9a) was used to estimate the depths Bay from the main channel of East Penobscot Bay. for other pockmarks from the 1998 survey for which we did not have direct depth information. Using the 4.3. Belfast Bay pockmark statistics regression equation (Fig. 9a), we determined that the average pockmark depth is 6.4 m and the median depth With a nearly 100% coverage of the 7 km2 Belfast is 6 m. Bay pockmark field, approximately 2300 pockmarks Newly calculated depth values based on this re- were mapped in Figs. 6 and 7. Pockmark diameters gression were used to estimate pockmark volumes. were calculated based on the polygon areas as if the Previous work by Kelley et al. (1994) employed a pockmarks were completely circular. Most pockmarks conical model of an individual pockmark to calculate are nearly circular in planimetric view, and this sim- that approximately 9.9Â107 m3 of sediment and pore 52 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62

Fig. 6. Map of Belfast Bay pockmark field (from Rogers, 1999). Locations of other figures indicated. Area within dashed polygons, labeled bside scan sonar tracklinesQ in legend represent ensonified swaths. water was removed from Belfast Bay due to pockmark 4.4. Spatial distribution of pockmarks and natural gas activity. The new assessment of pockmark volume is in Belfast Bay based on a pockmark model that comprises 10% of a sphere (Rogers, 1999), and is 2.69Â107 m3, or less The distribution of pockmark sizes in Belfast Bay than one third of the previous estimate. The reason for is complex. Density calculations were performed on this large discrepancy was only partly the difference the pockmark field, involving the calculation of pock- between the two models of pockmark shape, but mark population per square kilometer using the cen- largely due to an error in the older GIS data (Rogers, troids of the pockmark polygon coverage. The 1999). centroids displayed a pattern indicating that the J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 53

Fig. 7. (a) NOAA Survey Vessel Rude’s multibeam, planview map of part of Belfast Bay field (NOAA, 2003); (b) oblique view of pockmark field. The arrow is in the same approximate position in each image. edges of the field and particularly the northeast region contains larger pockmarks. Another factor associated experiences a higher density (240–270 pockmarks/ with this spatial distribution is the thickness of Holo- km2) than the central section of the bay (120–150 cene sediment (Fig. 9b). Where the surface unit thick- pockmarks/km2)(Rogers, 1999). These high concen- ens to greater than about 20 m, fewer but larger tration areas consist of numerous small pockmarks, pockmarks occur. This relation was also observed in whereas the central area has a lower density, but other parts of Penobscot Bay (Gontz, 2002), on the

Fig. 8. Histogram of pockmark diameters versus frequency of occurrence in Belfast Bay (from Rogers, 1999). 54 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62

a 30 y = 0.3755x0.7432 R2 = 0.4786 25

15 Depth (m)

0 0 50 100 150 200 250 300 Diameter (m) b 60 y = 0.2248x R2 = 0.4546

20 Sediment Thickness (m)

0 0 50 100 150 200 250 Pockmark Diameter (m)

Fig. 9. (a) Pockmark diameter versus depth for 250 pockmarks that passed directly beneath the fathometer as side-scan data were gathered; note that there are many multiple values with diameters less than 50 m; (b) a graph of pockmark diameter versus thickness of Holocene mud shows a crude linear correlation that flattens out with larger diameter pockmarks.

Scotian Shelf (Josenhans et al., 1978) and in the rich zones had an average diameter of 67 m and an Bering Sea (Nelson et al., 1979). average depth of 9 m, while pockmarks outside this When the digitized pockmarks were overlain on the gas-rich zone had an average diameter of 37.5 m and an gas fields in the GIS (Fig. 6), 706 of 2300 (30%) of the average depth of 5 m. There are also locations where pockmarks occur directly over gas deposits capable of natural gas and relatively thick sedimentary deposits wiping out a seismic signal. Pockmarks above the gas- exist without pockmark formation. J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 55

Linear chains of similar-size pockmarks occur within depths. When these depths were compared to the a half-kilometer of and sub-parallel to till outcrops that corresponding Holocene thickness, 73% of the pock- trend north-northwest across the northwestern part of mark eyes were within 2–3 m vertical elevation of the the field (Figs. 6 and 7). The chains are mostly, but not surface of the Presumpscot Fm. (Kelley et al., 1994). all outside the area underlain by gas. To the north of the We believe the beyesQ are strong reflections from Pleis- till outcrops, chains of pockmarks trend from northwest tocene sediment which crops out at the base of many to northeast, but those closest to the till outcrops trend pockmarks. It was observed, from sonar data, that parallel to them. A seismic cross section of the field multiple swaths in the same location often revealed shows these chains occurring where a strong acoustic pockmarks that were eyed in one pass but not in reflector on the surface of glacial-marine sediment rises another. This inconsistency may have to do with the up over a bedrock or till shoal (Fig. 10). Gas appears to slant-range angle of the sonar beam. Eyes can be con- migrate along the base of the Holocene sediment to the cealed in acoustic shadows, and, thus, not observed. It northeast margins of the field. The linear pockmarks was also often difficult to see if a pockmark had an eye apparently form when the Holocene sediment thickness when the towfish passed directly over the feature, is about 7 m, and there is insufficient overburden to making it difficult to record depths of eyed pockmarks. contain the gas. To the north of the till outcrops inade- bDark spotsQ, or strongly reflecting, circular acoustic quate seismic reflection observations exist to associate features with little bathymetric relief were tentatively the linear chains with sub-bottom geology. interpreted by Kelley et al. (1994) as shallow subsur- Eyed pockmarks possess a strong acoustic reflector face gas deposits (their Fig. 2a). These features were at their base (Fig. 4)(Hovland and Judd, 1988; Fader, found only in a 1Â4 km area oriented with the long 1991; Kelley et al., 1994). Eighty percent of all the axis pointing North, towards Searsport Harbor, the 2nd mapped eyed-pockmarks are found in the western por- largest port in Maine. Remotely Operated Vehicle tion of the field. These features generally occur in (ROV) observations made around the features revealed Holocene sediment less than 20 m thick. Seismic pro- numerous cobbles in Holocene mud. Seismic profiles files show that most eyed pockmarks terminate at, or also showed strong acoustic diffractions associated with near, the erosional unconformity at the base of the the dark spots (Fig. 10). Although there is a designated Holocene (Scanlon and Knebel, 1989; Kelley et al., dredge spoil disposal site located several kilometers 1994). There are 19 eyed pockmarks that have recorded from the most remote dark spot, it is plausible that

Fig. 10. Seismic reflection profile PB 99-34 across Belfast Bay field. Acoustic units labeled as in Fig. 5. See Fig. 6 for location. 56 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 the spots result from the disposal of dredge spoils from Rogers (1999) observed gas reflections within a unit Searsport Harbor that did not reach the designated site. directly above the glacial-marine sediment, but below a reflector now interpreted as the transgressive ravine- 4.5. Belfast bay pockmark activity ment unconformity (Belknap et al., 2002). The most spatially extensive gas deposits in Belfast Bay are Pockmarks appear on an 1872 bathymetric chart of where Holocene mud is thickest, and this is where the Belfast Bay as a number of anomalous soundings on an largest pockmarks are located. This association strongly otherwise, flat seafloor. They were further noted as suggests that Holocene mud contributes gas to the btidal channelsQ in pioneering seismic reflection profiles sediment column. Both the Pleistocene and Holocene by Ostericher (1965). Comparison between the 1989 units have very low permeability, thus gas must escape and 1998 side-scan maps shows persistence of the long through localized failures or along permeable strata. chains of pockmarks (Rogers, 1999), but we were not Pockmarks located in areas removed from acoustically able to quantitatively compare the maps owing to imaged gas must receive a supply from a permeable inconsistencies in navigation and towing geometry. zone such as the sand and gravel lag immediately The 1989 map was made with LORAN-C, which was overlying the basal unconformity over the Pleistocene brubber-sheetedQ in a GIS to fit with known latitude/ glacial-marine sediment (Gontz et al., 2002). This lat- longitude positions. True positions are not known to eral transport also explains how eyed pockmarks extend within a 100 m diameter. Furthermore, no record was to the surface of glacial-marine sediment, requiring a kept of the layback of the towfish in 1989, compound- major source at the base of the Holocene sediment ing difficulties in comparison. column, but still upslope from a deeper Holocene gas Nevertheless, the pockmarks of Belfast Bay appear source. In addition, the basal supply would explain the to be actively forming features. They exhibit sharp flattening of pockmark diameter /depth ratios with size, boundaries in ROV observations and steep slope walls as the unconformity constrains deepening. If gas is of approximately 308 (Kelley et al., 1994). Submersible localized in the middle or upper part of the Holocene traverses demonstrated such extreme instability on the section, it is unclear how pockmarks are excavated to pockmark sidewalls, that dives were cancelled for safe- the base of the Holocene. ty reasons (Belknap and Shipp, 1991). Given their The Holocene sediments contain organic matter location near the mouth of a large river, the pockmark from which gas could readily be formed by microbial edges would be muted and slumped if they were not breakdown of organic matter (Martens et al., 1998). A maintained. core (PB-VC-93-06) collected by Barnhardt et al. During the 1989 survey, an image of a single, well- (1997) had gas-charged sediment below 2.8 m of the formed gas and sediment plume was captured on the seabed, and a sudden change in color from orange to sonar’s third channel (Fig. 5)(Kelley et al., 1994). In gray at 2.6 m was inferred to represent a switch from 1992, a local fisherman reported benormous numbers of oxidizing to reducing conditions. Kelley et al., (1994) gas bubblesQ erupting from the bay (Russell Coombs, speculated that the source of the organic material need- Islesboro Island, 1993, personal communication). In ed to produce the gas originates from buried bog and another report, a scuba diver observed that the water lake deposits that rest on top of the unconformity (Fig. over a pockmark bbubbled like soda waterQ for 30 min 11). This hypothesis is similar to that of Nelson et al. during a possible eruption (G. Vaughn, personal com- (1979) who found peaty mud deposited below a pock- munication). During the course of the 1998 survey, two mark field in Norton Sound, Alaska. Missiaen et al. possible plumes were observed emanating from pock- (2002) also suspected that peat was the source of gas marks. Neither plume was as distinct as the 1989 off the Belgian coast. More than 40 years ago, Shepard example but both were punctuated by a sudden rise noted: bIf sea level were considerably reduced, there of the sonar’s automatic bottom tracking directly over would be extensive lakes in the Gulf of Maine as well two separate pockmarks. as on the exposed shelf along the coast farther north and in the Gulf of St. LawrenceQ (Shepard, 1963, p.212). 5. Discussion These are areas with many mapped pockmark and gas fields (Fader, 1991). If modern processes and environ- Gas is most often recognized in Holocene mud, but ments are similar to those that occurred at times of it is also imaged as an enhanced acoustic reflector near lower sea level, then many organic-rich deposits of the contact between Holocene and Pleistocene glacial- lake, wetland and estuarine sediment may exist offshore marine sediment. In nine locations in Belfast Bay, today as sources of natural gas. J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 57

Glacial-marine mud was exposed during the late and freshwater organics from Eastern Penobscot Bay at Pleistocene lowstand. Contours for this surface, repre- À18 m in a similar setting (8 km northeast of Belfast senting paleobathymetry, generated from the DEM Bay), which gave a radiocarbon date of 7390F500 yr show that 2.9 km2 of a the 7.8 km2 central gas field BP (Core 211, W1306). in Belfast Bay was 10–20 m lower than the sea level The lack of complete accordance between pock- lowstand (À55 to À65 m; Barnhardt et al., 1995). If marks and gas could be explained by some combination there were no barriers isolating this region from the sea, of four scenarios: (1) most gas was released from these marine submergence would have prevented the forma- areas and production has ceased, leaving senescent tion of freshwater organic-rich deposits. However, the pockmarks with no associated gas, (2) gas is currently end moraine at the south end of the field area, discov- not concentrated enough to be detected as bubbles with ered by Knebel and Scanlon (1985), probably cut the the seismic profiler in areas with pockmarks, (3) gas is, area off from ocean waters. This moraine, standing 15 or was migrating along thin geologic contacts and m higher than the estimated lowstand position, could erupting to form pockmarks distant from the gas source, have cut the bay off from ocean waters, allowing and (4) gas escape is not the mechanism of pockmark formation of a shallow lake or freshwater bog, and formation. abundant organic deposition. We have insufficient seis- If gas production has ceased, it would be expected mic coverage to confirm or refute the completeness of that some, if not all, the pockmarks in this region this morainal blockage. Ostericher (1965) cored wood should be inactive, and filling or eroding and show

Fig. 11. Cartoon model of pockmark formation in coastal Maine. (A) Deglaciation with higher-than-present sea level and deposition of till and glacial-marine mud over bedrock; (B) lowstand of sea level with lakes and wetlands in till and glacial-marine mud-lined rock basins. These exposed areas are today under water. (C) Contemporary sea-level position with evolution of methane from buried organic matter. Gas and pore water migrates away from gas source along coarse-grained unit at contact between Pleistocene and Holocene mud deposits. Pockmarks then form where gas pressure exceeds the confining pressure of overburden. Some gas also erupts over high-pressure gas zones. (D) In some modern situations, gas escape from margins reduces overall field pressure and pockmark formation ceases and infilling begins. Alternatively, gas may continue to form, but diffuses upward into mud with insufficient pressure to form pockmarks. 58 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62

Fig. 11 (continued). signs of these processes. However, there is no evidence which contained abundant gas deeper than 2.5 m.In for this in the Belfast Bay pockmark field (it may be addition, 3 of 7 sampling locations included in the true in Eastern Penobscot Bay: Gontz et al., 2002). All Ussler et al. (2003) background survey were outside the pockmarks observed had sharp, distinct boundaries the mapped gas field, with 3 of the remaining 4 loca- that showed no sign of slumping as is expected on mud tions near the edge of the field. Their highest concen- slopes of ~308 (Booth et al., 1985). Fader (1991) and tration of methane is from one of the chain pockmarks Rogers (1999) were unable to find filled pockmarks. on the eastern edge of the field. If the gas is not concentrated enough to be detected Gas migration seems to best fit the observations of by seismic equipment, it is likely that no pockmarks or pockmarks with no apparent source of bubble-phase only small pockmarks would be produced, possibly as gas (Fig. 11c). It is clear from seismic data that gas- is observed in the northeastern portion of the field near enhanced reflectors are both a common feature and are the till ridges. However, areas where no gas was ob- associated with the pockmarks in Belfast Bay. Gas is served have pockmarks just as large and numerous as likely migrating into bubble-free regions through a regions with observable gas. Data collected by Ussler et permeable lag deposit atop the Pleistocene/Holocene al. (2003) suggests that methane gas is not abundant unconformity, and/or in minor bed planes in the Holo- enough anywhere in Belfast Bay to support the forma- cene sediment, in concentrations large enough to initi- tion of pockmarks. However, cores taken by Ussler et ate and maintain pockmark stability. As discussed al. (2003) were very short (averaging 1.1 m, maximum previously, this boundary is a coarse layer varying length of 1.9 m) and either directly in pockmark bot- from decimeters to 2 m thick in some places that is toms or on sidewalls, where seismic data already show much more permeable than either the stiff glacial-ma- that there is no obvious gas. Their cores are in direct rine mud or Holocene estuarine mud. The existence of contrast to the Barnhardt et al. (1997) cores, which were pockmark chains is explained by gas migrating along up to 5 m in length from interpockmark areas, and the Pleistocene/Holocene unconformity, and exiting J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 59 along the flanks of structural highs associated with till Many more fields probably exist within the largely ridges. Scanlon and Knebel (1989) suggested that the unexplored embayments of the north central Maine unconformity would make a good aquifer due to this coast. high permeability. However, it is difficult to imagine 2. Gas fields occur within incised valleys or basins the source of recharge and head for a hypothetical eroded into glacial-marine sediment, which in turn groundwater flow in this unit. We propose that this are typically framed by bedrock valleys. zone is a good conductor of gas, which migrates up- 3. Pockmarks are found both within and outside gas ward from a lateral, deeper source until a more perme- fields; many gas fields lack pockmarks. Gas migra- able vertical pathway is reached, or a pressure buildup tion from deeper lateral sources along permeable occurs that allows upward migration through crack strata may be the mechanism for pockmark forma- formation (Gardiner et al., 2003) and piping (Hovland tion in areas lacking gas-curtain seismic reflections. and Judd, 1988; Whiticar, 2002) through the overlying Breaching of the seal of a gas reservoir along its weakly stratified mud. It is plausible that where migrat- margin may explain why pockmarks do not occur in ing gas exits from beneath its Holocene mud cover, some fields. over time, pockmarks could merge and a large enough 4. Direct observations of pockmark activity through exit would exist to regularly vent gas at the margin of acoustic imaging as well as direct observation sug- the Holocene sediment. Under these circumstances, gas gests that fields are actively venting gas. This sug- might preferably be released from the margin of the gestion is supported by the observed steep sides field through this leak in the gas reservoir seal, and (~308) of pockmarks in regions with river-sediment pockmark formation would cease even though gas was input. still being produced (Fig. 11d). This scenario might 5. The origin of the natural gas is not known, but if explain the fields with acoustically detectable gas in coastal environments at times of lower sea level the sediment, but with no pockmarks (Fig. 1). were similar to the present, numerous lake, wetland, Presently, despite the assertions of Ussler et al. valley fill and estuarine sources of organic-rich ma- (2003) and Paull et al. (1999), there is no evidence to terial may have survived transgression and remain suggest that gas escape is not the mechanism of pock- buried offshore as potential gas sources. mark formation. A freshwater source is not apparent 6. After mapping and modeling beyedQ pockmarks in and is not evident in geochemical sampling (Ussler et Belfast Bay, we conclude that they are the result al., 2003). Geochemical samples also produced isotopic of pockmark excavation at or near the Pleistocene/ compositions of methane within the range of microbial Holocene unconformity. The beyeQ is a strong production, probably ruling out a thermogenic source seismic reflection from the lag materials and/or (Ussler et al., 2003). The crystalline bedrock geology is the stiff glacial-marine sediments at the bottom not a probable source of gas either (Kelley et al., 1994), of the pockmark that contrasts with the weaker although deep-sourced gas is reported from an area of reflection of the Holocene mud of the pockmark the Baltic Sea (Soderberg and Floden, 1992). Gas walls. hydrates are not capable of forming at depths as shallow 7. Previously identified bdark spotsQ found in Belfast as those in Belfast Bay (e.g., Kvenvolden, 1993). Tidal Bay (strongly reflective circular features with little or storm wave pumping of pore waters cannot be ruled bathymetric relief) and tentatively interpreted by out, but the escape of gas from these mechanisms Kelley et al. (1994) as shallow subsurface gas, ap- would likely occur first. Although there is as yet no pear in fact to be dredge spoil deposits derived from certain origin of the gas observed in the fields mapped a nearby industrial complex. in Maine, nor any triggering mechanism directly ob- 8. Earlier failure to detect gas in cores (Ussler et al. served, a preponderance of observations suggests that (2003)) probably resulted from (a) cores too short to early Holocene carbon-rich sediment is a source of the intersect the gas-rich zones (Barnhardt et al., 1997), gas that forms and maintains modern pockmarks (Fig. as well as (b) core locations inside pockmarks, 11 a, b). which are generally regions that are free of gas, and not a lack of methane in the sediment. 6. Conclusions We have presented evidence that demonstrates that methane is abundant in the western Gulf of Maine, 1. In the western Gulf of Maine, at least 311 km2 of and that gas venting and migration are processes seafloor is underlain by natural gas in 70 separate associated with pockmark formation. However, it is fields, as determined by acoustic mapping methods. clear that many questions remain. The origin and 60 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 sources of the gas remains to be confidently deter- Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dick- mined through direct sampling and analysis of the gas- son, S.M., 1996g. Surficial geology of the Maine inner continental shelf: Petit Manan to West Quoddy Head, Maine. rich zone; specific triggering mechanisms for gas re- Geologic Map No. 96-13, Maine Geol. Survey, Augusta, ME, lease also remain unknown. 1:100,000. Barnhardt, W.A., Belknap, D.F., Kelley, J.T., 1997. Stratigraphic Acknowledgement evolution of the inner continental shelf in response to late Qua- ternary relative sea-level change, northwestern Gulf of Maine. Geol. Soc. Amer. Bull. 109, 612–630. We thank the U.S. Minerals Management Service for Belknap, D.F., Shipp, R.C., 1991. Seismic stratigraphy of glacial- early support to map the Belfast Bay Field. We acknowl- marine units, Maine inner shelf. In: Anderson, J.B., Ashley, G.M. edge funding from NOAA-NESDIS’ Penobscot Bay (Eds.), Glacial-Marine Sedimentation; Paleoclimatic significance, Project and the Island Institute for continued mapping Geol. Soc. Am. Spec. Pap., vol. 261, pp. 137–157. of Belfast Bay and the Maine Sea Grant Program for Belknap, D.F., Shipp, R.C., Kelley, J.T., 1986. Depositional setting and Quaternary stratigraphy of the Sheepscot Estuary, Maine. research support to determine the activity of the gas Ge´ographie Physique et Quaternaire 40, 55–69. fields. We thank Steve Dickson for many useful discus- Belknap, D.F., Andersen, B.G., Anderson, R.S., Anderson, W.A., sions on this work. Kathy Scanlon and an anonymous Borns Jr., H.W., Jacobson Jr., G., Kelley, J.T., Shipp, R.C., reviewer provided many helpful comments as well. Smith, D.C., Stuckenrath Jr., R., Thompson, W.B., Tyler, D.A., 1987. Late Quaternary sea-level changes in Maine. In: Nummedal, D., Pilkey, O.H., Howard, J.D. (Eds.), Sea-Level References Fluctuation and Coastal Evolution, SEPM. Spec. Pub. vol. 41, pp. 71–85. Albert, D.B., Martens, C.S., Alperin, M.J., 1998. Biogeochemical Belknap, D.F., Shipp, R.C., Kelley, J.T., 1987. Quaternary stratigra- process controlling methane in gassy coastal sediments: Part 2. phy of representative Maine estuaries: initial examination by high Groundwater flow control of turbidity in Eckenforde Bay sedi- resolution seismic reflection profiling. In: FitzGerald, D.M., ments. Cont. Shelf Res. 18, 1771–1793. Rosen, P.S. (Eds.), Glaciated Coasts. Academic Press, San Barber, D.C., 1995. Holocene depositional history and modern sand Diego, pp. 177–207. budget of inner Saco Bay, Maine. Unpublished Masters Thesis, Belknap, D.F., Kraft, J.C., Dunn, R.K., 1994. Transgressive valley-fill University of Maine. 178 pp. lithosomes: Delaware and Maine. In: Boyd, R., Zaitlin, B.A., Barnhardt, W.A., Kelley, J.T., 1995. Carbonate accumulation on the Dalrymple, R. (Eds.), Incised Valley Fill Systems, SEPM Spec. inner continental shelf of Maine: a modern consequence of late Pub., vol. 51, pp. 303–320. Quaternary glaciation and sea-level change. J. Sediment. Res. Belknap, D.F., Kelley, J.T., Gontz, A.M., 2002. Evolution of the A65, 195–207. glaciated shelf and coastline of the northern Gulf of Maine, Barnhardt, W.A., Gehrels, W.R., Belknap, D.F., Kelley, J.T., 1995. USA. J. Coast. Res. Spec. Issue 36, 37–55. Late Quaternary relative sea-level change in the western Gulf of Bloom, A.L., 1963. Late-Pleistocene fluctuations of sea level and Maine: evidence for a migrating glacial forebulge. Geology 23, postglacial and crustal rebound in coastal Maine. Am. J. Sci. 261, 317–320. 862–879. Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dickson, Booth, J.S., Sangrey, D., Fugate, J., 1985. A nomogram for interpret- S.M., 1996a. Surficial geology of the Maine inner continental ing slope stability of fine-grained deposits in ancient and modern shelf: Piscataqua River to Biddeford Pool, Maine. Geol. Map No. marine environments. J. Sediment. Petrol. 55, 29–36. 96-7, Maine Geological Survey, Augusta, ME, 1:100,000. Bussman, I., Suess, E., 1998. Groundwater seepage in Ekenforde Bay Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dickson, (Western Baltic Sea): effect of methane and salinity distribution of S.M., 1996b. Surficial geology of the Maine inner continental the water column. Cont. Shelf Res. 18, 1795–1806. shelf: Ogunquit to Kennebec River, Maine. Geol. Map No. 96-8, Dickson, S.M., 1999. The role of storm-generated combined flows Maine Geol. Survey, Augusta, ME, 1:100,000. in shoreface and inner continental shelf sediment erosion, Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dickson, transport and deposition. Ph.D. Diss., Univ. of Maine, S.M., 1996c. Surficial geology of the Maine inner continental Orono. 179 pp. shelf: Cape Elizabeth to Pemaquid Point, Maine. Geologic Map Dorion, C.C., Balco, G.A., Kaplan, M.R., Kreutz, K.J., Wright, J.D., No. 96-9, Maine Geol. Survey, Augusta, ME, 1:100,000. Borns Jr., H.W., 2001. Stratigraphy, paleoceanography, chronolo- Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dickson, gy, and environment during deglaciation of eastern Maine. In: S.M., 1996d. Surficial geology of the Maine inner continental Weddle, T.K., Retelle, M.J. (Eds.), Deglacial History and Relative shelf: Boothbay Harbor to North Haven, Maine. Geologic Map Sea-Level Changes Northern New England and Adjacent Canada, No. 96-10, Maine Geol. Survey, Augusta, ME, 1:100,000. Geol. Soc. Am. Spec. Pap., vol. 351, pp. 215–242. Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dickson, Fader, G.B.J., 1991. Gas-related sedimentary features from the eastern S.M., 1996e. Surficial geology of the Maine inner continental Canadian continental shelf. Cont. Shelf Res. 11, 1123–1154. shelf: Rockland to Bar Harbor, Maine. Geologic Map No. 96-11, Field, M.E., Gardner, J.V., Jennings, A.E., Edwards, B.D., 1982. Maine Geological Survey, Augusta, ME, 1:100,000. Earthquake-induced sediment failures on a 0.258 slope, Klamath Barnhardt, W.A., Belknap, D.F., Kelley, A.R., Kelley, J.T., Dickson, River delta, California. Geology 10, 542–546. S.M., 1996f, Surficial geology of the Maine inner continental Fleischer, P., Orsi, T.H., Richardson, M.D., 2001. Distribution of free shelf: Mt. Desert Island to Jonesport, Maine, Geologic Map No. gas in marine sediments: a global overview. Geo Mar. Lett. 21, 96-12, Maine Geol. Survey, Augusta, ME, 1:100,000. 103–122. J.N. Rogers et al. / Marine Geology 225 (2006) 45–62 61

Garcia-Gil, S., Vilas, F., Garcia-Garcia, A., 2002. Shallow gas fea- Warping, and Sea-level Change. Maine Geol. Sur. Bull., vol. 40, tures in incised-valley fills (Ria de Vigo, NW Spain): a case study. pp. 157–204. Cont. Shelf Res. 22, 2302–2315. Kelley, J.T., Dickson, S.M., Belknap, D.F., Stuckenrath Jr., R., 1992. Gardiner, B.S., Boudreau, B.P., Johnson, B.D., 2003. Growth of disc- Sea-level change and late Quarternary sediment accumulation on shaped bubbles in sediments. Geochim. Cosmochim. Acta 67, the southern Maine inner continental shelf. In: Fletcher, C., Weh- 1485–1494. miller, J. (Eds.), Quarternary Coasts of the United States: Marine Gehrels, W.R., Belknap, D.F., Kelley, J.T., 1996. Integrated high- and Lacustrine Systems, Soc. Econ. Paleontol. Mineral., Spec. precision analyses of Holocene relative sea-level changes; les- Publ., vol. 48, pp. 23–34. sons from the coast of Maine. Geol. Soc. Amer. Bull. 1089, Kelley, J.T., Dickson, S.M., Belknap, D.F., Barnhardt, W.A., Hender- 1073–1088. son, M., 1994. Giant sea-bed pockmarks: evidence for gas escape A.M. Gontz, 2002. Evolution of seabed pockmarks in Penobscot Bay, from Belfast Bay, Maine. Geology 22, 59–62. Maine. MS Thesis, University of Maine, Orono, 118 pp. Kelley, J.T., Gehrels, W.R., Belknap, D.F., 1995. Late Holocene A.M. Gontz, 2005. Sources and implications of shallow subsurface relative sea-level rise and the geologic development of tidal methane in estuarine environments. Unpublished d. D. University marshes at Wells, Maine, U.S.A. J. Coast. Res. 11, 136–153. of Maine, Orono, ME, 308 pp. Kelley, J.T., Barnhardt, W.A., Belknap, D.F., Dickson, S.M., Kelley, Gontz, A.M., Belknap, D.F., Kelley, J.T., 2002. Seafloor features and A.R., 1998. The seafloor revealed. Maine Geol. Sur. Open File characteristics of the Black Ledges area, Penobscot Bay, Maine, Rept. 96-6 Augusta, ME, 55 pp. USA. J. Coast. Res. Spec. Issue 36, 333–339. Kelley, J.T., Dickson, S.M., Belknap, D.F., Barnhardt, W.A., Barber, Hannum, M.B., 1997. Late Quaternary Evolution of the Kennebec D.C., 2003. Sand volume and distribution on the paraglacial inner and Damariscotta River Estuaries, Maine, M.S. Thesis, Univ. continental shelf of the northwestern Gulf of Maine. J. Coast. Res. Maine, Orono, 98 pp. 19, 41–56. Hasiotis, T., Papatheodorou, G., Kastanos, N., Ferentinos, G., Kelley, J.T., Barber, D.C., Belknap, D.F., FitzGerald, D.M., van 1996. A pockmark field in the Patras Gulf (Greece) and its Heteren, S., Dickson, S.M., 2005. Sand budgets at geological, activation during the 14/7/93 seismic event. Mar. Geol. 130, historical and contemporary time scales for a developed beach 333–344. system, Saco Bay, Maine, USA. Mar. Geol. 214, 117–142. Hovland, M.A., Judd, A.G., 1988. Seabed Pockmarks and Seepages. King, L.H., MacLean, B., 1970. Pockmarks on the Scotian Shelf. Graham and Trotman, London, p. 144. Geol. Soc. Amer Bull. 81, 3141–3148. Hovland, M.A., Talbot, M.R., Qvale, H., Olaussen, S., Aasberg, L., Knebel, H.J., 1986. Holocene depositional history of a large glaciated 1987. Methane-related carbonate cements in pockmarks of the estuary, Penobscot Bay, Maine. Mar. Geol. 73, 215–236. North Sea. J. Sediment. Petrol. 57, 881–892. Knebel, H.J., Scanlon, K.M., 1985. Sedimentary framework of Pe- Johnson, D., 1925. In: Facsimile (Ed.), The New England–Acadian nobscot Bay. Maine. Mar. Geol. 65, 305–324. Shoreline. Hafner Pub. Co, 1967, New York. 608 pp. Kvenvolden, K.A., 1993. A primer on gas hydrate. In: Howell, D.G. Josenhans, H.J., King, L.H., Fader, G.B.J., 1978. A side-scan sonar (Ed.), The Future of Energy Gases, U.S. Geol. Sur. Prof. Pap., mosaic of pockmarks on the Scotian Shelf. Can. J. Earth Sci. 15, vol. 1750, pp. 279–291. 831–840. Martens, C.S., Albert, D.B., Alperin, M.J., 1998. Biogeochemical Judd, A.G., Hovland, M., 1992. The evidence of shallow gas in processes controlling methane in gassy sediments: Part 1. A marine sediments. Cont. Shelf Res. 12, 1081–1095. model coupling organic matter flux to gas production, oxidation Kelley, J.T., 1987. An inventory of coastal environments and classi- and transport. Cont. Shelf Res. 18, 1741–1770. fication of Maine’s glaciated shoreline. In: FitzGerald, D.M., McClennen, C.E., 1989. Microtopography and surficial sediment Rosen, P.S. (Eds.), Glaciated Coasts. Academic Press, San patterns in the central Gulf of Maine: a 3.5 kHz survey and Diego, pp. 151–176. interpretation. Mar. Geol. 89, 69–85. Kelley, J.T., 2004. Coastal Bluffs of New England. In: Hampton, M.A., Miller, G.T., 1998. Deglaciation of Wells Embayment, Maine: inter- Griggs, G.B. (Eds.), Coastal Cliff Erosion: Status and Trends. pretation from seismic and side-scan sonar data. M.S. Thesis, U. S. Geological Survey Prof. Pap., vol. 1693, pp. 95–106. Dept. Geol. Sci., Univ. Maine, Orono, 210 pp. http://pubs.usgs.gov/pp/pp1693/. Missiaen, T., Murphy, S., Loncke, L., Henriet, J.P., 2002. Very high- Kelley, J.T., Belknap, 1991. Physiography, surficial sediments and resolution seismic mapping of shallow gas in the Belgian coastal Quaternary stratigraphy of the inner continental shelf and near- zone. Cont. Shelf Res. 22, 2291–2301. shore region of central Maine, United States of America. Cont. Nelson, H., Thor, D.R., Sandstrom, M.W., Kvenvolden, K.A., Shelf Res. 11, 1265–1283. 1979. Modern biogenic gas-generated craters (sea-floor Kelley, J.T., Dickson, S.M., 2000. Low-cost bluff-stability mapping in bpockmarksQ) on the Bering Shelf, Alaska. Geol. Soc. Amer. coastal Maine: providing geological hazard information without Bull. 90, 1144–1152. alarming the public. Environ. Geosci. 7, 46–56. NOAA, 2003, The NOAA ship RUDE; multibeam imagery from Kelley, J.T., Belknap, D.F., Jacobson Jr., G.L., Jacobson, H.A., 1988. Belfast Bay. http://www.pmc.noaa.gov/ru/belfast.htm. The morphology and origin of salt marshes along the glaciated Orange, D.L., Greene, H.G., Reed, D., Martin, J.B., McHugh, C.M., coastline of Maine, USA. J. Coast. Res. 4, 649–665. Ryan, W.B.F., Maher, N., Stakes, D., Barry, J., 1999. Wide- Kelley, J.T., Belknap, D.F., Shipp, R.C., 1989a. Sedimentary frame- spread fluid expulsion on a translational continental margin: mud work of the southern Maine inner continental shelf: influence of volcanoes, fault zones, headless canyons, and organic-rich sub- glaciation and sea-level change. Mar. Geol. 90, 139–147. strate in Monterey Bay, California. Soc. Geol. Am. Bull. 111, Kelley, J.T., Belknap, D.F., Shipp, R.C., Miller, S.B., 1989b. An 992–1009. 1999. investigation of neotectonic activity in coastal Maine by Osberg, P.H., Hussy, A.M. II, Boone, G.M., 1985. Bedrock Geo- seismic reflection methods. In: Anderson, W.A., Borns, H.W. logic Map of Maine: Maine Geol. Sur. Augusta, Maine. (Eds.), Neotectonics of Maine, Studies in Seismicity, Crustal 1:500,000. 62 J.N. Rogers et al. / Marine Geology 225 (2006) 45–62

Ostericher, C., 1965. Bottom and subbottom investigations of Penob- Shipp, R.C., Staples, S.A., Ward, L.G., 1987. Controls and zonation scot Bay Maine, 1959. U.S. Naval Oceanographic Office Tech. of geomorphology along a glaciated coast, Gouldsboro Bay, Rept. 173, 177 pp. Maine. In: FitzGerald, D.M., Rosen, P.S. (Eds.), Glaciated Coasts. Paull, C.K., Ussler III, W., Borowski, W.S., 1999. Freshwater ice Academic Press, San Diego, pp. 209–231. rafting: an additional mechanism for the formation of some high- Soderberg, P., Floden, 1992. Gas seepages, gas eruptions and degas- latitude submarine pockmarks. Geo Mar. Lett. 19, 164–168. sing structures in the seafloor along the Stromma tectonic linea- Paull, C.K., Ussler III, W., Maher, N., Greene, H.G., Rehder, G., ment in the crystalline Stockholm Archipelago, east Sweden. Lorenson, T., Lee, H., 2002. Pockmarks off Big Sur, California. Cont. Shelf Res. 12, 1157–1171. Mar. Geol. 181, 323–335. Stuiver, M., Borns Jr., H.W., 1975. Late Quaternary marine invasion Retelle, M.J., Weddle, T.K., 2001. Deglaciation and relative sea- in Maine: its chronology and associated crustal movement. Geol. level chronology, Casco Bay Lowland and lower Androscoggin Soc. Amer. Bull. 86, 99–104. River Valley, Maine. In: Weddle, T.K., Retelle, M.J. (Eds.), Taylor, D.I., 1992. Nearshore shallow gas around the U.K. coast. Deglacial History and Relative Sea-Level Changes Northern Cont. Shelf Res. 12, 1135–1144. New England and Adjacent Canada. Geol. Soc. Am. Spec. Pap., Thompson, W.B., 1982. Recession of the late Wisconsinan ice sheet vol. 351, pp. 191–214. in coastal Maine. In: Larson, G.J., Stone, B.D. (Eds.), Late Rogers, J.N., 1999. Mapping of subaqueous, gas-related pockmarks in Wisconsinan Glaciation of New England. Kendall/Hunt Pub. Belfast Bay, Maine using GIS and remote sensing techniques: MS Co, Dubuque, pp. 211–228. Thesis, Dept. Geol. Sciences, Univ. of Maine, Orono, ME, 139 pp. Thompson, W.B., Borns, Jr., H.W., 1985. Surficial Geologic Map of Scanlon, K.M., Knebel, H.J., 1989. Pockmarks on the floor of Pe- Maine. Maine Geol. Sur., Augusta, Maine 1:500,000. nobscot Bay, Maine. Geo Mar. Lett. 9, 53–58. Tolman, A., Kelley, J., LePage, C., 1986. Buried Valleys in South- Schubel, J.R., 1974. Gas bubbles and the acoustically impenetra- western Maine. Geol. Soc. Am. 18, 72 (Abstracts with Program). ble, or turbid, character of some estuarine sediments. In: Uchupi, E., 2004. The Stellwagen Bank region off eastern Massachu- Kaplan, I.R. (Ed.), Natural Gases in Marine Sediments. Plenum setts: a Wisconsin glaciated Cenozoic sand banks/delta? Mar. Press, NY, pp. 275–298. Geol. 204, 325–347. Shaw, J., Gareau, P., Courtney, R.C., 2002. Palaeogeography of Uchupi, E., Swift, S.A., Ross, D.A., 1996. Gas venting and late Atlantic Canada 13–0 kyr. Quat. Sci. Rev. 21, 1861–1878. Quaternary sedimentation in the Persian (Arabian) Gulf. Mar. Shepard, F.P., 1963. Submarine Geology, 2nd ed. Harper and Row, Geol. 129, 237–269. NY, p. 557. Ussler III, W., Paull, C.K., Boucher, J., Friederich, G.E., Thomas, R.C. Shipp, 1989. Late Quaternary sea-level fluctuations and geologic D.J., 2003. Submarine pockmarks: a case study from Belfast Bay, evolution of four embayments and adjacent inner shelf along the Maine. Mar. Geol. 202, 175–192. northwestern Gulf of Maine. Ph.D. Thesis, School of Marine Vogt, P.R., Crane, K., Sundvor, E., Max, M.D., Pfirman, S.L., 1994. Sciences, Univ. of Maine, 832 pp. Methane-generated (?) pockmarks on young, thickly sedimented Shipp, R.C., Belknap, D.F., Kelley, J.T., 1989. A submerged shoreline oceanic crust in the Arctic: Vestnesa Ridge, Fram Strait. Geology on the inner shelf off the Maine coast, Northwestern Gulf of 22, 255–258. Maine. In: Tucker, R.D., Marvinney, R.G. (Eds.), Studies in Wever, Th.F., Abegg, F., Fiedler, H., Fechner, G., Stender, I.H., 1998. Maine Geology, Quaternary Geology, vol. 5. Maine Geol. Sur, Shallow gas in the muddy sediments of Eckernforde Bay, Ger- Augusta, pp. 11–28. many. Cont. Shelf Res. 18, 1715–1739. Shipp, R.C., Belknap, D.F., Kelley, J.T., 1991. Seismic–stratigraphic Whiticar, M.J., 2002. Diagenetic relationships of methanogenesis, and geomorphic evidence for a post-glacial sea-level lowstand in nutrients, acoustic turbidity, pockmarks and freshwater seepages the northern Gulf of Maine. J. Coast. Res. 7, 341–364. in Eckernforde Bay. Mar. Geol. 182, 29–53.