TERMINAL MIOCENE ARRIVAL OF COLORADO RIVER SAND IN THE SALTON TROUGH, SOUTHERN CALIFORNIA: IMPLICATIONS FOR INITIATION OF THE LOWER COLORADO RIVER DRAINAGE

Two competing hypotheses have been proposed for initiation of the lower Colorado River (CR): upward-migrating headward erosion (predicts gradual expansion of catchments), or downward-propagating lake-overflow events (predicts rapid insertion of river discharge into the lowland region). New data from Split Mt. Gorge and Fish Creek in the SW Salton Trough, a late Cenozoic transtensional basin, provide insight into this problem. We measured a 1000-m-thick section, correlated in the upper 200 m with previous work by Johnson et al. (1983), and dated the strata with magneto- and biostratigraphy. The lower half of the section contains upper Miocene alluvial conglomerate and lower megabreccia overlain by locally derived (L-suite) marine turbidites and upper megabreccia. L-suite sandstone contains angular to subangular plutonic quartz and feldspar, plutonic lithic fragments, and abundant detrital biotite. CR sandstone in the Wind Caves Mbr (Pw) of the Latrania Fm contains abundant moderately to well rounded quartz grains with rare syntaxial overgrowths, and lithic fragments of chert, quartzite and siliceous to intermediate volcanic rocks. Reworked Cretaceous foraminifers are found in CR silt and mud. Abrupt first appearance of CR sand ~21 m above the base of Pw marks earliest development of the Pliocene Colorado delta (Winker, 1987; Winker and Kidwell, 1996). Above this level, CR sand dominates the section and L-suite sand is only found close to basement highs. This compositional change occurs between the top of C3An.1n (5.89 Ma) and base of Thvera (5.24 Ma) subchrons, and coincides closely with the Miocene-Pliocene boundary (5.33 Ma) defined by microfossils. Latest Miocene benthic foraminifers correlate with planktic foraminiferal zone N17 and calcareous nannoplankton zone NN11. Pliocene faunas lack Amphistegina. Sedimentation rate from the Mio-Pliocene boundary to the base of Thvera subchron is 0.74 ± 0.25 mm/yr (not decompacted). Using this rate, the base of CR sand is dated at 5.36 ± 0.06 Ma. Because the CR arrived in the Lake Mead area shortly after 5.51 Ma (House et al., this volume), we conclude that the river propagated ~400 km to the Salton Trough in ~100-200 k.y. The rapid pace of drainage integration supports a lake-overflow model, but does not rule out a marine-estuary interpretation for basal limestone of the Bouse Fm.