Evaluating paleochannels

Just what, if any, significance is there to the paleochannels from the Eocene on the west side of the Sierra Nevada? These have been held up as demonstrations of post-Eocene uplift of the range and demoted to insignificant artifacts of a landscape developed on metamorphic rock. Consider these conflicting statements from the abstracts of two recent papers:

Eocene paleochannels show lowest gradients parallel to the range axis, steepest ones perpendicular, and reaches with significant “uphill” gradients that rise in the paleo-downstream direction. Modern Sierran rivers lack this relationship. The azimuth-gradient relationships of paleochannels, especially the uphill gradients, require late Cenozoic tilting and uplift.-Wakabayashi, Geosphere, 2013

and the counterpoint:

The studies supporting recent tilting in the northern Sierra Nevada are inconclusive and rely on observations not unique to tectonic forcing. Indeed, much of the evidence based on the paleogradients of the Tertiary channels is consistent with an early trellis drainage network formed across alternating bands of resistant and weak lithologies. –Gabet, Am. J. Sci., 2014

Now to be transparent, GG has published the view that the drainages do support post-Eocene uplift, but that was then and this is now; given the work done in the past decade, reexamining this is worth some effort. (Hopefully sometime we’ll take a long look at the Gabet paper, which is a more comprehensive attempt to consider the surface geology of the Sierra).

First, the traditional interpretation, as summarized in the Wakabayashi paper, is that reaches of channels trending to the southwest are steeper than those trending to the southeast (or northwest), which are nearly flat. Assuming a uniform grade for the original river, these would be the product of tilting. Exactly how much tilt depends on a few things, but the general range is something over a kilometer of uplift at the crest of the northern Sierra.

There are two counterarguments made to this: one is that the river was not originally  at a uniform grade, the other that the deposits are so diachronous that correlations are meaningless (primarily made by Cassel and Graham in GSA Bull 2011 and repeated in several subsequent publications): you have no idea what grade the paleochannel had because it was cut at different times. (At least that is GG’s understanding; if in error, comment away).

So let’s do a thought experiment: Let’s presume, as Gabet argues, that the southwest-trending streams, in crossing structural strike, were steep and the channels trending to the SE or NW were in strike valleys and so had a low grade.  This seems reasonable and plausible given that the Auriferous Gravels, which form the basis of most of our knowledge of these deposits, contain gold (duh!) and the gold came from the metamorphic basement of the western part of the northern Sierra. So far, so good, no?

But we know of these channels not because of the bedrock thalweg, but because of the sediments that fill them. Are we going to accumulate sediments everywhere along rivers with variations in slope of more than a factor of ten? Well, we certainly could if the stream power somehow was equalized along the paleorivers, but that seems exceptionally fortuitous.

Let’s pretend that the Eocene South Fork of the Yuba is correctly connected [yes, this is disputed, but please, the point will tend to stand anyways] and consider how steeply we might deposit the Auriferous Gravels, especially the finer grained upper part (we will return to the lower, “Blue Gravels” in a moment). Cassel and Graham report paleogradients of 0.4 – 5.5 % from these rocks and argue that these are primary paleogradients.  Let’s take the steep westernmost reach of the paleo-Yuba then as the gradient at which we can deposit sediments: it falls 800m in 50 km, or a grade of 1.6%.  Above that we have gradients of maybe 100 m over the next 50 km, or 0.2%.  If we deposit sediment at that 1.6% grade along this reach, we should have an extra 1.4% upstream-thickening wedge, which would generate 700m of sediment at the upper end of this reach, and that thickness would have to extend up the next 50 km or so unless the river’s ability to accumulate sediment actually decreases upstream.  The maximum thickness of the Auriferous Gravels is about 300m.  So an interesting question would be to know the variation in thickness of the Auriferous Gravels as a function of position along an inferred paleodrainage. Also interesting would be the age of the sediments both vertically within a section and as a function of position along any ancestral channel; this is information we are unlikely to gain anytime soon as fossil localities are not uniformly widespread and use of detrital zircons is imperfect in the absence of a certain, continuing source of new zircons during the Eocene.

Now something like this could well be possible; sedimentation rate is presumably some function of sediment supply and stream carrying capacity.  It would be interesting to try to make these calculations to see what variations in these parameters would be consistent with the patterns of sedimentation seen.  But the point is that it is the accumulation of sediments, which is insensitive to the bedrock erodibility below, that is key here, so the existence of a trellis network with original variations in gradient must be consistent with the accumulation of sediments after its creation. (Another thing to look for would be to see what the channels do in crossing some of the plutonic rocks in the foothills; as the rock here is isotropic, we wouldn’t expect to see an azimuthally-dependent gradient to a channel).

OK, but the above was based on this business of connecting channels, which has been called into question. The argument is that the river channel at one location might have been cut at a much different time than that at another location, so simply following a thalweg–which is only preserved discontinuously anyways–might be traveling through time. This would tend to suggest that the channels were cut in relatively short order and then filled, since all the age constraints on these rocks place them in the Eocene. This would mean that we’d be overestimating the original grade of the river in using the thalweg, but by how much? And would there be any reason why this would be azimuthally dependent? If we chose to distribute the maximum 300m of sediment so as to minimize the grade, we would place the maximum thickness at the base of the steeper SW-trending reaches and feather it out to zero at the top, reducing, roughly, 1.6% grades down to 1.3% grades, still far higher than the low grades on the SE and NW trending river reaches.

It is worth considering a different aspect of the problem, namely, the concentration of gold in the gravels. Now the Blue Gravels are the richest of the paleoplacer deposits; gold was found in the crevices in the underlying basement and so in many places tunnels were run under younger deposits to find the Blue Gravels and then follow them laterally; in this way several underground river channels were found and mapped.  The thing is, what are these gravels representing?

They are unquestionably far coarser than the materials above and are poorly dated, lacking the fine fossil deposits available at higher levels.  But there is the issue of the concentration of gold in them; how did that happen?  Most geologists presume that the gold was a primary sedimentary deposit very much like the placer gold found in the modern rivers (indeed, it is quite likely that the modern placer gold is mainly reworked from the Eocene deposits).  If this is true, then, where did all that gold come from?  An interesting approach was taken by Loen (Econ. Geol., 1992) in the form of a mass balance: if, on average, the rock that was eroding contains a fraction of gold F (g/m^3) and there is a drainage area of A for the river, then the total gold available to be in the placers would be G = F * A * E, where E is the amount of material eroded while sediment was accumulating. With some 2000 tons of placer and paleoplacer gold and something of the order of 10,000 km^2 surface area for the paleostreams depositing the gold and a gold concentration on the order of 2-3 ppb, we get E = 2 x 10^6 kg/ 2.5 x 10^-9 /2650 kg/m^3/10^10 m^2 = 30 m, but since a large part of the gold would be flushed downstream (Loen estimates half to 90%), the gold in the paleoplacers represents 60-300m of erosion with these numbers (Loen worked the other way and estimated the concentration of gold from an estimated erosion of 400-1400m of rock and found this could overlap with estimates of gold concentration of 0.3 -5.2 ppb).

So what? Well, that 60-300m of erosion needs to occur as the stream is aggrading those placer deposits, and as the Blue Gravels are the lion’s share of the gold, it means that the stream is processing material representing on the order of erosion of 100m of the surrounding landscape while depositing the Blue Gravels. What kind of time does this represent? Cecil et al. (GSA Bulll, 2006) argued for average erosion rates of 20-40 meters per million years from U-Th/He dates in the northern Sierra Nevada, which would mean that an individual stream segment might be accumulating the Blue Gravels for several million years in order to accumulate the gold found within them (If Loen’s analysis was correct, it would take the whole Eocene, but because of the way he did things, taking that “result” would in fact be simply taking his assumption).  Given the 20 million years of the Eocene and the ~100-150 km length that some river system must have had across the region, roughly 20 km of stream length must have been accumulating gold at any one time in the Blue Gravels. This seems plausible within the development of drainage envisioned by Cassel and Graham. Of course you can also do this by having the river exist for a long time.

Now there are some interesting pieces of information that are troubling.  Cassel and Graham argue for an eastward migration of sedimentation in the Eocene (sediments to the west have a flora dated as late early Eocene, sediments to the east have some zircon grains from the late Eocene).  It is hard to see how that is accomplished: why wouldn’t the western parts accumulate even more sediment as time went on? This leads us to consider the age constraint for the western gravels (well, ones in the You Bet-Chalk Bluffs-North Columbia swath), which is the Chalk Bluffs flora described by MacGinitie in 1941.  It isn’t clear what the basis for the age of these fossils really is: MacGinitie used an age by correlation with marine sediments of the Ione Formation to the west.  It seems that Jack Wolfe might have tried to date these rocks from the floral fossils, but the only indication of that is an assignment of these fossils to the late early Eocene as a line in a table. Since the fossils were collected from the whole of several sections, any difference in age through the section is lost and it could be that the age constraints are poorly known. This all matters because Malakoff, just east of North Columbia and about the same distance from the paleoshoreline as Chalk Bluffs itself, has a couple of late Eocene zircons. If all these dates are correct, then sedimentation was fairly rapid and exceptionally episodic. GG suspects this is hard to do.

There are a few other things that need to be explored.  In places, the remaining bedrock limits the extent of the Eocene rivers, so arguments for rivers shifting around have to respect those limits, and that might minimize ways to work around the apparent paleogradients. Many of the descriptions of the Eocene channels talk about a deep central channel some 100′ deep or so with a surrounding stretch of gentle terrain marked by terraces; do those 100′ deep channels have important meaning? Or should we focus on the wide and gentle terraces around them?

And the focus to date has been on the paleo-Yuba.  But there were other river systems, and they show many of the same characteristics of grade variations with azimuth; these channels to the south tend to have a greater exposure to plutonic rocks underneath and so might not be easily explained by trellises.

There’s lots more to discuss, but this suggests that we aren’t yet done arguing over these river channels.

Tags: , ,

Trackbacks / Pingbacks

  1. Return of the Young Sierra | The Grumpy Geophysicist - December 18, 2016

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: