The "Giant Wave" hypothesis for the deposition of chaotic gravels high on the coastal slopes of the Hawaiian Islands remains controversial, in part due to the scarcity of detailed submarine information regarding the Alika debris-avalanche deposits. In order to interpret the Alika debris avalanches, a comprehensive submarine investigation using high-resolution bathymetric maps, ocean-floor observations and samples from submersibles, piston cores, and improved GLORIA side-scan images, was carried out. The sequence of debris-avalanche events is based on bathymetric investigations, draped sediments on megablocks in the South Kona slide area, and turbidites recovered by a piston core sampling 135 km southwest of the island of Hawaii (PC-13). The chemical compositions of the draped sediments and turbidite layers in the piston core are very similar, but differ from the matrix and block facies of the South Kona slide. These compositions suggest that the source for the turbidite layers was probably lava erupted subaerially from Mauna Loa. This geological evidence suggests that the Alika debris-avalanche deposits are unlikely to have formed in a single event or in rapid succession. They can be redefined as three geologically separated landslide events: the Kealakekua rotational slump, the Alika 1 debris avalanche, and the Alika 2 debris avalanche. The Kealakekua rotational slump probably occurred after the South Kona slide complex and before the Alika debris avalanches and was emplaced slowly. The Alika 1 and 2 debris-avalanche deposits can be divided into subfacies: (1) the main body of the debris-avalanche deposits (Alika 1d and 2d), (2) megaturbidites (Alika 1t and 2t) accompanying the debris avalanches, and (3) levee-overflow sediments (Alika 2s) derived from the suspension and reworking of Alika 2d. Thus, complicated bathymetric characteristics in the Alika debris avalanche deposits can be interpreted as an overlapping of multiple slope failures. Therefore, the newly estimated volumes of the landslide deposits, their boundaries delineated by high-resolution digital elevation models, are smaller than those of previous estimations.
View full abstract