The uplifted marine terrace of last interglacial (stage 5e) age at Cave Hill, Barbados, has been investigated with views toward terrace architecture, developmental processes, and sea level history. Methods include stratigraphic analysis of new deep exposures, precise 230Th geochronology, and geomorphic mapping. Cave Hill was central to earlier studies that led to important ideas on reef and terrace evolution but with which we find significant disagreement. We present a new model of terrace evolution that emphasizes the role of marine erosion, deposition of carbonate cover during the full eustatic cycle rather than only at highstand, and principal reef development during transgression rather than at highstand by keep-up rather than catch-up growth. The new model and geochronology contribute to an improved understanding of surficial processes during emergence of uplifting coral coasts and of global sea level changes in the last interglacial.
Transgression in stage 5e at Cave Hill was accompanied by progressive marine erosion of a terrace floor and receding seacliff and deposition of a seaward-thickening reefal wedge on the floor and above limestones of stage 6 and 7 ages. The wedge contains a diachronous basal Acropora palmata fringe reef. This is overlain by an in-place A. cervicornis reef that aggraded progressively during sea level rise. The transgressive phase took place between 136 ka (or earlier) and ca. 128 ka. During highstand between ca. 127 and 120 ka, the floor and seacliff continued to recede landward. Re gression began at or after ca. 120 ka, and sea level fell ≥37 m below the highstand level by ca. 115 ka. In stage 5c, the seaward front of the last interglacial terrace was eroded landward an uncertain distance. The sea level record at Cave Hill has implications for timing and constituent events of the penultimate glacial, the last interglacial highstand, and the fall in sea level at the transition from stage 5e to stage 5d.
Shoreline angles, which are isochronous linear geomorphic features, are the most accurate markers of sustained highstand levels. Highstand levels and uplift rates interpreted from A. palmata in Barbados are less accurate and lower because the coral grew mainly in transgression. The coral-derived data therefore include uncertainties of depths of growth and collapse.