Importance of groundwater in propagating downward integration of the 6-5 Ma Colorado River system: Geochemistry of springs, travertines, and lacustrine carbonates of the Grand Canyon region over the past 12 Ma

We applied multiple geochemical tracers (⁸⁷Sr/⁸⁶Sr, [Sr], δ¹³C, and δ¹⁸O) to waters and carbonates of the lower Colorado River system to evaluate its paleohydrology over the past 12 Ma. Modern springs in Grand Canyon reflect mixing of deeply derived (endogenic) fluids with meteoric (epigenic) recharge. Travertine (<1 Ma) and speleothems (2-4 Ma) yield ⁸⁷Sr/⁸⁶Sr and δ¹³C and δ¹⁸O values that overlap with associated water values, providing justification for use of carbonates as a proxy for the waters from which they were deposited. The Hualapai Limestone (12-6 Ma) and Bouse Formation (5.6-4.8 Ma) record paleohydrology immediately prior to and during integration of the Colorado River. The Hualapai Limestone was deposited from 12 Ma (new ash age) to 6 Ma; carbonates thicken eastward to ~210 m toward the Grand Wash fault, suggesting that deposition was synchronous with fault slip. A fanningdip geometry is suggested by correlation of ashes between subbasins using tephrochronology. New detrital-zircon ages are consistent with the "Muddy Creek constraint," which posits that Grand Wash Trough was internally drained prior to 6 Ma, with limited or no Colorado Plateau detritus, and that Grand Wash basin was sedimentologically distinct from Gregg and Temple basins until after 6 Ma. New isotopic data from Hualapai Limestone of Grand Wash basin show values and ranges of ⁸⁷Sr/⁸⁶Sr, δ¹³C, and δ¹⁸O that are similar to Grand Canyon springs and travertines, suggesting a long-lived springfed lake/marsh system sourced from western Colorado Plateau groundwater. Progressive up-section decrease in ⁸⁷Sr/⁸⁶Sr and δ¹³C and increase in δ¹⁸O in the uppermost 50 m of the Hualapai Limestone indicate an increase in meteoric water relative to endogenic inputs, which we interpret to record progressively increased input of high-elevation Colorado Plateau groundwater from ca. 8 to 6 Ma. Grand Wash, Hualapai, Gregg, and Temple basins, although potentially connected by groundwater, were hydrochemically distinct basins before ca. 6 Ma. The ⁸⁷Sr/⁸⁶Sr, δ¹³C, and δ¹⁸O chemostratigraphic trends are compatible with a model for downward integration of Hualapai basins by groundwater sapping and lake spillover. The Bouse Limestone (5.6-4.8 Ma) was also deposited in several hydrochemically distinct basins separated by bedrock divides. Northern Bouse basins (Cottonwood, Mojave, Havasu) have carbonate chemistry that is nonmarine. The ⁸⁷Sr/⁸⁶Sr data suggest that water in these basins was derived from mixing of high-⁸⁷Sr/⁸⁶Sr Lake Hualapai waters with lower-⁸⁷Sr/⁸⁶Sr, first-arriving "Colorado River" waters. Covariation trends of δ¹³C and δ¹⁸O suggest that newly integrated Grand Wash, Gregg, and Temple basin waters were integrated downward to the Cottonwood and Mojave basins at ca. 5-6 Ma. Southern, potentially younger Bouse basins are distinct hydrochemically from each other, which suggests incomplete mixing during continued downward integration of internally drained basins. Bouse carbonates display a southward trend toward less radiogenic ⁸⁷Sr/⁸⁶Sr values, higher [Sr], and heavier δ¹⁸O that we attribute to an increased proportion of Colorado River water through time plus increased evaporation from north to south. The δ¹³C and δ¹⁸O trends suggest alternating closed and open systems in progressively lower (southern) basins. We interpret existing data to permit the interpretation that the southernmost Blythe basin may have had intermittent mixing with marine water based on δ¹³C and δ¹⁸O covariation trends, sedimentology, and paleontology. [Sr] versus ⁸⁷Sr/⁸⁶Sr modeling suggests that southern Blythe basin ⁸⁷Sr/⁸⁶Sr values of ~0.710-0.711 could be produced by 25%-75% seawater mixed with river water (depending on [Sr] assumptions) in a delta- marine estuary system. We suggest several refinements to the "lake fill-and-spill" downward integration model for the Colorado River: (1) Lake Hualapai was fed by western Colorado Plateau groundwater from 12 to 8 Ma; (2) highelevation Colorado Plateau groundwater was progressively introduced to Lake Hualapai from ca. 8 to 6 Ma, (3) Colorado River water arrived at ca. 5-6 Ma; and (4) the combined inputs led to downward integration by a combination of groundwater sapping and sequential lake spillover that first delivered Colorado Plateau water and detritus to the Salton Trough at ca. 5.3 Ma. We propose that the groundwater sapping mechanism strongly influenced lake evolution of the Hualapai and Bouse Limestones and that groundwater flow from the Colorado Plateau to Grand Wash Trough led to Colorado River integration.