During the December 2007, SEA-CALIPSO experiment we collected seismic reflection profiles offshore of Montserrat. Off the east coast, we imaged deep fans of volcaniclastic debris from three volcanoes progressively active from ∼2 Ma to present. Near-shelf sedimentation rates of 8–9 cm/ka are approximated following cessation of local volcanic activity. The fans were deposited on sediments with apparent dips towards the ESE-trending Montserrat-Havers fault system (MHFS) in southern Montserrat. The MHFS encloses the Soufrière Hills Volcano, has elevated crustal blocks at Roche's Bluff, St. Georges Hill, and Garibaldi Hill, and extends off the west coast. Off the west coast, the N-dip of two faults supports a N-dip interpretation for a major component of MHFS, the Belham Valley fault. We propose that local deformation is affected by stress redistributions consistent with a right-stepping, sinistral en-echelon fault system, but the interplay of transtension and magmatism has resulted in complex and evolving stress regimes.
1. SEA-CALIPSO Reflection Profiling and the Geologic/Tectonic Setting
 The December 2007, SEA-CALIPSO experiment (Seismic Experiment with Airgun-source – Caribbean Andesitic Lava Island Precision Seismo-geodetic Observatory) at Montserrat, Lesser Antilles, was an onshore-offshore seismic study of the crust and magmatic system under Montserrat and the Soufrière Hills volcano (SHV) [Paulatto et al., 2010; Shalev et al., 2010; Voight et al., 2010]. The experiment included a 48 channel, 600 m streamer, and a 2600 in3 airgun in a seismic reflection survey that explored local submarine deposits and faults and expanded knowledge based on previous seismic and bathymetric studies [e.g., Feuillet et al., 2001, 2002]. We present key results from our survey and discuss their implications on the local tectonic and volcanic interactions.
 On and west of SHV, young andesitic domes (<300 ka) and structurally uplifted areas [Harford et al., 2002] are aligned due to normal faulting as part of the extensional Montserrat-Havers fault system (MHFS) [Feuillet et al., 2010]. The MHFS includes an ESE-trending lineament interpreted as the Belham Valley fault (BVF) [Harford et al., 2002] (Figure 1b). Normal faulting continues to the SE of Montserrat with a right step to the Bouillante-Montserrat fault system (BMF) (Figure 1b). Extension with ∼N–S trend is prevalent in the region, which Feuillet et al.  suggest is accommodated as oblique shear along a series of en echelon, mainly NE-dipping normal fault systems including the BMF, MHFS, and the Redonda fault system (RFS) (Figure 1b). We propose that these systems also accommodate minor local shear that has resulted in rotation of older sediments and deformation of the footwall of the BVF and related faults.
2. Data Collection and Processing
 As part of the SEA-CALIPSO experiment, the RRS James Cook collected offshore reflection profiles (Figure 1b). The eight GI airgun array was fired every 60 s with streamer signals sampled every 2 ms. The shot interval was 139 m. The data were bandpass filtered between 4 Hz and 64 Hz, stacked, and migrated using sediment velocities from semblance analysis. A low source frequency and long shot interval were selected to maximize the onshore tomography data and were thus less optimal for the reflection study. Six sections are shown in Figures 2–4 (red profile lines, Figure 1b). Time to depth conversions used an average sediment velocity of 2200 m/sec [Paulatto et al., 2010].
3. Volcaniclastic Fans and Faulting Off the Eastern Shelf
 The reflection profiles off the eastern shelf are dominated by chaotic sediment packages that are, interpreted as accumulations of volcaniclastic debris. In the profiles, debris is visible as eastward tapering lenses that extend offshore from Silver Hills and Centre Hills (Figures 1b and 2; Lines 7 and 9). The debris from Silver Hills (Line 7) extends ∼10 km from the shelf and overlies strata that step down towards Montserrat. From Centre Hills as well (Line 9), debris onlaps layered sediments that dip westward, from 1.5 s two-way travel time (twt) at km 12 to 2.2 s twt at km 5 (Figure 2). The apparent dip (Line 9) and downward fault-step pattern (Line 7) towards the island are consistent with either island subsidence or with rotation on the hanging wall of the MHFS faults (Figure 1b).
 From Centre Hills the debris lenses have accumulated in stacks, the largest being ∼10 km long (Figure 2, Line 9). The lenses are onlapped by and alternate with sub-horizontal strata. We interpret the lenses as submarine fans from emplacements of volcaniclastic flows, caused largely by lava dome collapses and deposited over several hundred ka. Coarse submarine fans have formed in this way during the current volcanism, producing tapering units extending as much as 8 km offshore [Le Friant et al., 2009, 2010].
 Where fans are overlain by flat-lying sediment, sedimentation rates can be estimated. We recognize that these rates could be influenced by multiple factors, including minor flows, remobilized deposits, and tephra fallout, but assume that flat sediment accumulation fundmentally marks the cessation of the main volcanic source. Off Silver Hills the largest fan is covered by ∼80 m of flat sediments. Silver Hills became extinct at ∼1 Ma [Harford et al., 2002], yielding a sedimentation rate of ∼8 cm/ka. Off Centre Hills the sediments are ∼44 m thick. The Centre Hills became extinct at ∼500 ka suggesting a rate of ∼9 cm/ka. Le Friant et al.  report hemipelagic sedimentation rates at a site 60 km offshore as 1–3 cm/ka. At a site ∼16 km offshore Trofimovs et al.  report a rate of 4–7 cm/ka. Our higher rates are consistent with a persistent, near-shore sediment source.
 The N–S reflection profiles off the eastern shelf cut across the debris fans. Approximately 1–2 km offshore (Figure 3, Line 2) a mounded feature is visible just below the sea floor between km 6 and km 11. SHV flow deposits 4–5 km off the shelf are indicated by the mound on Line 23 between ∼kms 10 and 20. A 6 km-wide channel is visible in Line 2 at km 11 to 15, which we identify as an embayment associated with a previously described gravity flow [Le Friant et al., 2004] (Figure 1b). The southern slope of the embayment is consistent with the fault scarp north of Roche's Bluff (Figure 1b) [Feuillet et al., 2010] which subsequently has been modified by landsliding.
4. Exposed and Buried Faults Off the Western Shelf
 North-dipping fault scarps offset the ocean floor on profiles approximately 6 and 14 km off the west coast (Figures 1b and 4, Lines 21 and 15). The western scarp offset at km 10 of Line 15 is at least 40 m. The MHFS fault scarp and S-tilted footwall block are clear features as well on nearby profile gwa058 of Feuillet et al. . To the north several faults break the ocean floor, indicating recent activity. A normal fault-bounded step in the morphology near km 16 on Line 15 is associated with the Redonda fault system (RFS) (Figure 1b). Further north, buried scarps have created basins of folded, syn-rift sediments. Beyond km 23, faulting is buried by ∼100 m of flat, basin-filling sediments.
 In the south, both profiles reveal complex footwall deformation. At km 4–10 of Line 15 a series of piggy-back basins have formed on the tilted footwall, which is an ascending slope over ∼1300 m of elevation. On the elevated footwall of Line 21 (∼km 1), a small graben is infilled by subhorizontal sediments (Figure 4).
 At km 3 on Line 21, sediments appear to dip to the south and onlap onto a major N-dipping normal fault (Figure 4). This fault is not the same as the principal scarp fault of Line 15, but is en-echelon and south of it by ∼2 km (Figure 1b) [Feuillet et al., 2010, Figure 2]. The fault may coincide with an along-strike projection of the Belham Valley fault (BVF, Figure 1b).
5. Tectonic Synthesis and Relation to Onshore Deformation and Volcanism
 A new tectonic interpretation of southern Montserrat can be proposed by integrating our data with older studies and the work of Feuillet et al. . We agree with the regional model of Feuillet et al.  that the major fault systems (RFS, MHFS, and BMF) are mainly normal and arranged in a right-stepping en echelon structure, and also that on the large scale, this section of the arc accommodates regional left lateral shear. We disagree with the interpretation of the onshore features in Feuillet et al.  and add a discussion of the SHV feeder dike in relation to the complexity of local tectonic and magmatic stresses.
Feuillet et al.  re-introduce the old idea [e.g., Rea, 1974] that Garibaldi and St. Georges Hills are volcanic cones, and they suggested the hills were fed by vents in a fissure parallel to the BVF. Notably, the field evidence does not support this hypothesis. St. Georges, Garibaldi, and Richmond Hills are composed mainly of block-and-ash-flow and pumice-and-ash flow deposits, and epiclastic beds [Harford et al., 2002]. The deposit lithofacies are not compatible with vent or near-vent andesite eruption products but most likely have been sourced from SHV. Notably, while the lithofacies are similar to those of modern lava-dome flank environments where original bedding dips <8°, exposed beds on Garibaldi Hill have dips up to 50° [Harford et al., 2002]. In addition, 3-D tomography [Shalev et al., 2010] indicates low P-wave velocity material under St. Georges and Garibaldi Hills, quite unlike the high velocity cores under SHV and Centre Hills. Thus we conclude that the morphology of these hills is not primary; they are fault-bounded tectonic uplifts that have been deformed by (mostly) normal faults, and deposits have been tilted beyond the sedimentary slope limits.
 A closely related issue is the BVF, which, in contrast to Feuillet et al.  but following Harford et al. , we interpret as a N-dipping normal fault. This is consistent with our interpretation of Garibaldi and St. Georges Hills as tectonic uplifts; in addition, these onshore blocks seem analogous to the prominent elevated footwalls seen offshore on Lines 15 and 21 (Figure 4) (cf. profiles gwa 055 and 058 of Feuillet et al. [2010, Figure 3]). The N-dipping fault in our marine profile Line 21 (Figure 4, ∼km 3), is aligned with the BVF as an along-strike projection, and the north-dip on the offshore profile supports a similar interpretation for the BVF. Finally, 3-D tomography [Shalev et al., 2010] indicates that the contact of the P-wave velocity anomaly boundary under St. Georges Hill dips (very roughly) 50°N.
 Regionally, southern Montserrat is part of a transtensional regime with extensional overprinting. Locally, southern Montserrat includes a right-step between the MHFS and the BMF, en echelon normal fault systems in sinistral slip; thus uplift may have been encouraged by a minor contractional component [Deng et al., 1986; Cunningham and Mann, 2007].
 The interplay of transtention, contraction, and magmatism has resulted in evolving stress regimes around SHV and the BVF and related faults, as demonstrated by Miller et al. . Estimates of the orientation of the SHV feeder dike include a NNE trend based on 1995 seismic data [Miller et al., 2010], and a NNW trend based on 1997 ground tilt data [Hautmann et al., 2009]. A way of reconciling these data is to have a dike oriented ∼N, which is compatible with a northerly trend of magma-transport-related hybrid earthquakes under SHV [Miller et al., 2010, Figure 2]. A northerly dike trend implies a local E–W axis of minimum compression, which is approximately orthogonal to the MHFS-BMF regional trend of N–S extension [Feuillet et al., 2001, 2010]. Stresses appear to have been subjected to strong local reorientation. This could be due to some combination of varying pressures caused by the propagation of dikes and evolution of the magma storage system [e.g., Gudmundsson, 2006] and to being within a complex fault step [Oglesby, 2005].
 Based on marine reflection profiles and related onshore data, we conclude that on Montserrat the interplay among local faulting, volcanism, and stresses is very complex. The regional transtensional system of en echelon faults [cf. Feuillet et al., 2010] has influenced volcanism, while the local fault step suggests both a component of compression and complicated, evolving stress regimes and fault movements.
 We thank the NSF CD, Geophysics, and I & F Programs, UK NERC, BGS, Discovery Channel TV, and British FCO for funding SEA-CALIPSO. RSJS was supported by a NERC advanced grant. We thank volunteers and crew of RRS James Cook cruise JC19, and Fusion Petroleum Technologies, Inc., for aid in data processing. Special thanks to R. Reichlin of NSF for support of streamer costs, and Brandy Murphy and Lee Ellett of Scripps for excellence in collecting streamer data. The paper was improved with comments from S. Henrys, M. Jancin, J. Rowland, and two anonymous reviewers.