Using sieving to deconstruct the regolith: A first attempt
Wednesday, May 8th, 2013
H.W. Nesbitt and colleagues published in 1996 and 1997 two papers that outlined a general model for the integrated role that regolith development in plutonic terranes, erosion, and sorting during riverine transportation play in producing quartzofeldspathic sediments. In this model, sorting of the eroded regolithic material produces a mud-rich component that is more weathered than a size-fractionated sandy component. In A-CN-K space, as a result of progressive weathering, the mud-rich and sandy components lie at opposite ends of rays extending from the A apex, the composition of kaolinite and gibbsite, downward to the plagioclase-K-feldspar join where they reflect as a result of the preferential incongruent alteration of plagioclase decreasing ratios of plagioclase to total feldspar. In order to test this model, I collected eleven corestone, four transition zone, and four saprock samples from a single site located within Japatul Valley, California. Each of the four saprock samples was divided into two parts. One part was labeled the bulk fraction, and the remaining part was sieved into the three following size fractions: >63 microns, 63 to 45 microns, and <45 microns. Each sample was then analyzed for its major element composition.
Thin section observations indicate that the corestone is a meta-granodiorite consisting of megacrysts of plagioclase and K-feldspar (~5 – 7.5 mm in size) embedded in an equigranular and finer grained matrix (~1 – 3 mm in size) of primarily plagioclase, quartz, biotite, and lesser amounts of hornblende. On an A-CN-K diagram (molar basis) corestone, transition zone, and bulk saprock samples define a very weak weathering trend oriented subparallel to the A-CN join. Utilizing Al as a reference frame element, mass balance calculations indicate that 19% (+10/-9) of the Ca mass, 15% (+9/-8) of the Na mass, 13% (+12/-10) of the Mg mass, and 47% (+11/-9) of the P mass was lost during the transformation of corestone to saprock. The loss of Ca and Na mass is attributed to the incongruent leaching of An-rich cores and zones within plagioclase, while losses in Mg mass are likely the result of the alteration of biotite. Leaching of apatite is responsible for the loss of P mass.
If erosion were to remove the regolith at the study site, then the >63 micron fraction would be sorted into sand and pebble sized material, while the 63-45 micron fraction would represent coarse silt. The <45 micron fraction would include medium and fine silt along with clay-sized materials. On an A-CN-K diagram, the >63 micron fractions plot just to the right of and overlaps the cluster of bulk saprock samples. This relationship is consistent with the fact that the >63 micron fractions made up between 90 – 95% of the bulk fractions from which they were sieved. In contrast, the 63-45 and < 45 micron fractions make up less than 5-10% of the bulk fractions from which they were derived, and are significantly poorer in K then are the bulk and the >63 micron fractions. This diminishment in K is likely due to the fact that most of the megacrystic K-feldspar was retained within the >63 micron fraction while plagioclase and its weathering products were retained within the 45-63 micron and <45 micron fractions. Significantly, the <45 micron fraction is more depleted in Ca and Na than is the 45-63 micron fraction, and both fractions plot about a linear trend oriented parallel to the A-CN join. In addition, XRD work suggests that <2 micron clay fraction is dominated by mainly gibbsite (Al(OH)3) with minor amounts of kaolinite (Al2Si2O5(OH)4) and even lesser amounts of illite. The projection of the trend defined by the 45-63 micron and <45 micron fractions downward from the A apex, the composition of kaolinite and gibbsite, to the plagioclase-K-feldspar tie line indicates a total feldspar composition dominated by plagioclase rather than K-feldspar as predicted by the model of Nesbitt and colleagues (plagioclase to total feldspar ratio = 0.9). These results are not predicted by the model of Nesbitt and colleagues, and are in need of additional study.