USING GRAIN SIZE AND CARBONATE PRECIPITATION IN LOESS TERRACES TO RECONSTRUCT MASS WASTING EVENTS AND TRACE GROUNDWATER MOVEMENT

Loess hillslope terraces, interpreted alternately as anthropogenic or mass-wasting features, are a pervasive feature in Jiuzhaigou National Nature Reserve (JNNR). Typical terrace scarp stratigraphy is loess capped with a paleosol overlain by material consisting of a fine-grained matrix supporting gravel-size clasts. The purpose of this study is to quantify vertical composition changes across the transition from loess below to stony material above the paleosols, primarily through petrographic analysis of thin sections from 20 terrace scarp samples. For example, we determined the relative abundances of detrital mica and carbonate by randomly selecting a group of about 500 contiguous grains in each thin section and counting the number of detrital mica and carbonate crystals. Micromorphological interpretation focused on the presence or absence of Needle Fiber Calcite (NFC) and clasts >2 mm as well as the shapes of larger voids.

We find that grain size and the presence of NFC are depth dependent. Both the paleosol and the loess below it lack grains > 2mm whereas samples from above the paleosol average 12% clasts that are > 2mm. The latter are mostly metamorphic fragments rocks derived from local bedrock, which range from as little as 0.25 mm across up to boulder size. NFC is only found in samples near and above the paleosol. In contrast, we find no significant difference in the abundance of either detrital calcite or mica as a function of depth.

The presence of intact NFC bodies and gravel indicate the material above the paleosol is not simply unmodified loess. Furthermore, the presence of boulder and gravel-size grains [BS1] and matrix-supported gravel in the material above the paleosol are consistent with mass movement and formation of the terraces via natural processes. The NFC is fragile and therefore must have precipitated in-situ; furthermore, the longest fibers (up to 100 micron) require groundwater precipitation rather than a biogenic origin. Since the NFC only occurs near and above the paleosol, we hypothesize that the paleosol is influencing groundwater flow at mid-slope. Because the paleosol and NFC are present in all scarps of the terrace we studied, it is more likely that the emplacement of the NFC, paleosol, and large stones pre-date terrace formation, i.e., the terraces only formed after NFC was precipitated.