Site 1039 provided an excellent reference site, recovering a complete, 378-m thick sedimentary section that bottomed in gabbroic sills, as well as a complete section of high-quality logging-while drilling (LWD) data. Geochemical data at this site showed evidence of a seawater source deep in the section, implying flow of water through seafloor conduits (faults?) into the upper part of the crust and possibly explaining the low heat flow.
Sites 1040 and 1043 were drilled through the lowermost slope, penetrating the décollement at 371 mbsf and 151 mbsf, respectively, and recovering the underthrust sedimentary section to total depths of 661 mbsf and 282 mbsf, respectively. Site 1040 also bottomed in gabbroic sills. At Site 1043, LWD logging was successful through the entire sediment section (total depth of 482 mbsf), whereas at Site 1040 the LWD tools did not penetrate the décollement. The deformed sedimentary wedge above the décollement showed hardly any resemblance with the reference site, suggesting little or no frontal accretion. In addition, the underthrust sections at the two sites showed little if any loss of material relative to the reference site, based on lithology, biostratigraphy and magnetostratigraphy, geochemistry, and physical properties. The underthrust sections did, however, show significant volume change during subduction, with reductions in some lithologic units approaching two-thirds of the original volume. Detailed comparisons between the underthrust and reference sections are possible because of excellent LWD logs and detailed lithologic and physical properties measurements on cores.
Sites 1041 and 1042 were drilled farther upslope, through the sedimentary apron to total depths of 424 mbsf and 391 mbsf, respectively. LWD was successful at Site 1042 to a depth of 298 mbsf. Lithological, physical, and geochemical character of the apron material drilled at these sites shows close correspondence with the deformed wedge material at Sites 1040 and 1043. It is likely that the deformed wedge is composed largely of material from the sedimentary apron and very little from accreted trench sediment. Site 1041 yielded abundant gas hydrate, including several frozen sections where gas-hydrate cement was filling the pore space in volcanic ash layers. Site 1042 penetrated the high-velocity zone beneath the apron at a depth of about 300 mbsf, retrieving a middle Miocene breccia consisting of angular fragments of sandstone, basalt, red chert, and cumulate gabbro, cemented with calcite. These rocks were not derived by accretion from the Cocos Plate, but they closely resemble rocks exposed in coastal Costa Rica. It seems most likely that the high-velocity prism beneath the apron represents an extension of the rocks found onshore, and therefore, the amount of accretion is quite limited. Further analysis of benthic foraminifers post-cruise may clarify the vertical motion of the margin, with implications for either underplating or tectonic erosion.
Fluid conduits were in evidence at each of the four sites through the continental slope, based on geochemistry and LWD logs. Within the slope apron at Sites 1041 and 1042, localized minima of Cl and salinity, silica, and K lie in narrow zones, implying fluid flow from sources less saline than seawater. Such sources could be either deeper zones of clay dehydration or meteoric water. Some deep sources are needed, however, to explain local buildup of propane and butane, requiring formation temperatures approaching 100°C. The lower wedge Sites 1040 and 1043 each showed a major conduit within the wedge, and each showed that the décollement represents a significant fluid conduit as well. Fluid conduits are also supported by extremely low-density spikes in LWD logs over depth intervals of a few meters, within the wedge as well as along and just above the décollement zone.