The Enriquillo-Plantain Garden Fault (EPGF), one of two left-lateral transforms that define the Caribbean-North American plate boundary in Haiti, plunges beneath Lake Azuei in the eastern part of the country. In 2017, we acquired 220 km of sub-bottom and multichannel seismic reflection profiles (CHIRP and MCS) in a 1.2 km grid pattern across the lake. The two seismic methods achieved subbottom penetration of up to 15 m and 200 m, respectively. CHIRP and MCS data reveal folded turbidites across the expected extension of the EPGF fault zone at the south end of the lake, although direct evidence for faulting is lacking. Along the west side of the lake, however, MCS data image a broad NW-SE monoclinal fold whose geometry is compatible with an underlying SW-dipping blind thrust fault. CHIRP profiles image patches of soft- sediment deformations above the monoclinal fold; they also image a distinctive 11 m-deep paleoshoreline all around the lake that is uplifted by 1-2 m above that fold. These results are compatible with a scenario where some large slip event(s) on the presumed blind thrust occurred after formation of the paleoshoreline, locally uplifting the lakebed and causing liquefaction. Two short sediment cores sampled a layer correlatable to a reflection in the CHIRP profiles. That reflection extends laterally below the 11m shoreline, and thus predates its formation. On-going 14C dating of material from the two cores are expected to provide a maximum age for the shoreline, and thus for some hypothetical slip event(s) on the presumed blind thrust fault. The prominent character of the paleoshoreline suggests that it was stable, something best achieved if the lake level was controlled by the sill depth of an outlet. Presently, no such outlet exists and the lake level fluctuates. Pending results from radiometric dating, we propose that, earlier in the Holocene, the lake overflowed eastward into adjacent Lake Enriquillo along the narrow valley marking the extension of the EPGF fault zone - a valley that is presently blocked by an alluvial fan. Regardless of the relevance of that model, the uplifted shoreline implies a significant uplift rate on a structure extending up to 10 km north of the EPGF fault zone and striking oblique to it, confirming that transpressional tectonics is partitioned over an area at least as broad as Lake Azuei.

To a first order, the Caribbean plate converges obliquely at ~2 cm/yr toward the North American plate. This transpression is partly accommodated across the island of Hispaniola by the partitioning of motion between a fold-and-thrust belt trending NW-SE, and two E-W left-lateral fault systems located 150 km apart. The southern fault, the Enriquillo-Plantain Garden Fault (EPGF), is morphologically well expressed in western Haiti but its precise geometry in eastern Haiti is debatable. There, Lake Azuei stretches over 20 km in a direction parallel to the fold-and-thrust belt while its southern shoreline strikes EW, parallel to the expected trend of the EPGF. Because of a high sedimentation rate, the history of transpressional deformation should be captured in the lake stratigraphy and, accordingly, we acquired 220 km of multichannel seismic reflection (MCS) profiles across its surface. The survey followed a grid pattern with a spacing of 1.2 km and achieved a penetration of up to 300 m beneath lakebed. Interpretation of the dataset documents two major structures. First, the western side of the lake is occupied by a broad NW-trending monoclinal fold. This fold is cross-cut by a few NW-striking vertical (strike-slip) faults. We propose that this monocline is the surface expression of a SW-dipping blind thrust fault. The progressive steepening of the seismic horizons with depth suggests that it has been continuously active during the deposition of at least 300 m of sediments. The other major structure consists of a ~2 km-wide deformation zone that borders the EW-trending southern shore. This deformation zone is faintly imaged below a shallow gas front. We tentatively propose that it corresponds to a set of fault-propagation folds that are developing ahead of an EW, S-dipping oblique-slip fault. Such a model has been proposed already from three other independent studies involving GPS monitoring, seismological monitoring, and detailed field mapping. It is also supported by CHIRP profiles acquired concurrently with our MCS data and that document folding of the topmost turbidites but a lack of evidence for any stratigraphic offset across faults. Furthermore, a set of en echelon folds in that area are trending EW, while WNW-ESE fold axes would be expected instead above an EW vertical strike-slip fault.