TEST OF STATISTICAL UNMIXING OF SEDIMENT POPULATIONS

An expanded array of statistical techniques has recently been applied to sedimentary geology. One tool, End Member Mixing Analysis has been used to differentiate environments and sediment sources using suites of sediments. End member mixing analysis assumes that 1) different sources and transport mechanisms create distinct sedimentary populations, these populations mix, and the relative contribution of different sources can be identified. However, no test has tracked the sediment to the inferred sources, demonstrating the mixing of different populations.

This study is the first test that relates sediments mixing along a drainage to the tributaries that provided the sediment. The study was conducted in the Dona Ana Mountains in New Mexico. 32 samples were collected from 3 tributaries. Two of the tributaries drained mixed sources. One tributary passed through beds of Permian limestones, sandstone, and shale. The samples were distinguishable based on the grain compositions.

Samples were collected from active arroyo floors and grain size distributions were similar, with mean sizes ranging from 0.5 mm to 1.5 mm. In general, the samples from the larger, mixed arroyos were both coarser and finer than the tributary in the Permian. Four end members were recognized. End member mixing distinguished the Permian source from other sources, with samples from the Permian Arroyo containing 30 to 95 percent of individual samples. The larger mixed arroyos contained two end members produced by depositional sorting of the sand during waning flow. Finally, the fourth end-member was matched to samples adjacent to a Holocene Terrace that had not been previously mapped.

End Member mixing can distinguish between sources, when the sediment properties are different. However, care must be taken not to over-interpret results as sorting of sediment during deposition can create end members that reflect those processes and not sources.