«Chelsea Willett Advisor: Mark Brandon Second Reader: David Evans April 27, 2011 A Senior Thesis presented to the faculty of the Department of Geology ...»
History of long-term glacial erosion in the Patagonian
Advisor: Mark Brandon
Second Reader: David Evans
April 27, 2011
A Senior Thesis presented to the faculty of the Department of Geology and
Geophysics, Yale University, in partial fulfillment of the Bachelor's Degree.
In presenting this thesis in partial fulfillment of the Bachelor’s of Science
Degree from the Department of Geology and Geophysics, Yale University, I
agree that the department may make copies or post it on the department website. I further agree that extensive copying of this thesis is allowable only for scholarly purposes. It is understood, however, that any copying or publication of this thesis for commercial purposes of financial gain shall not be allowed without my written consent.
Chelsea D. Willett 27 April 2011 Willett 3 Abstract In the Patagonian Andes, home to the third largest ice field in the modern world, we investigate the influence of erosion via the glacial buzz saw on the mountain belt since its initial uplift ~25-20 Ma. It is thought that extensive glaciation began in the area 6 Ma or earlier. Plio-Pleistocene cooling sets up a natural experiment in Patagonia: was topography rapidly removed to establish the modern correlation between glacier extent and maximum elevations in only a few million years, or was this correlation developed on a longer timescale, implying that alpine glaciers were important erosive agents prior to the Pliocene? In the former case a major increase in erosion rates coincident with cooling would be expected, in the latter case, erosion rates would remain fairly steady from Pliocene to present. The eastern flank of Patagonia features a unique sequence of well-dated moraines and tills, each containing granitic cobbles derived from the Patagonian Batholith. Mean (U-Th)/He apatite cooling (AHe) ages from these deposits yield a record of the average erosion rate in the high Andes over time.
Preliminary results suggest that erosion rates have accelerated over the past 20 Ma, from as slow as 0.1 mm yr-1 to 0.45 mm yr-1. Additionally, LGM ice models predict that the ice divide sat west enough to transport material from the Jurassic-Cretaceous granite pluton on the west side of the mountain belt to the east to form the glacial deposits seen today. These results support the global cooling concept and indicate that rock transport crossed the modern continental divide, further developing a long-term erosional history of the area.
Introduction Glaciers are known to have a high erosion capacity: the sliding base of a glacier can erode as much as 1 m of bedrock material for every 1 km of sliding (Hallet et al., 1996). Glaciers can also be frozen based, where the ice/bedrock interface involves no sliding so erosion is minimal. Being a sliding base or frozen base glacier is determined by glacier thickness and ambient temperatures.
It is the highly erosive, sliding base glaciers that are thought to play a key role in shaping topography. These sliding base glaciers are found on mountains, at latitudes where the average yearly temperatures are low enough, but also at lower latitudes, at altitudes where the atmospheric lapse rate makes ambient temperature sufficiently low for enough of the year. Possible glacier localities are further limited by the constraint that, in order to build a glacier, sufficient precipitation is needed.
This configuration presents an interesting question: as a mountain belt grows, it rises to create an ideal location for glaciers, but the glaciers in turn erode the mountain belt. We strive to understand where this relationship balances, a theory called the “glacial buzz saw,” the idea that equilibrium line elevations (ELAs, or snowlines) control the maximum height of peaks (Reiners and Brandon, 2006). Glacier ELA is a contour that defines the meeting line of the zones of ablation and accumulation.
It has long been assumed that the glacier ELA controls the height of the Andes (Fig. 1). Note the transition from fast to slow erosion rates around 48ºS,
Figure 1: Variation in expected erosion rates in the Andes relative to past and present ELA (snowline). Between ~20ºS and 45ºS, fast erosion occurs, matching the modern snowline to the modern topography. South of 48ºS, glaciers transition to frozen base, decreasing the erosion rate (by Steve Porter, as presented in Broecker and Denton, 1990).
Is the assumption that ELA controls the Andes valid? My contribution to this investigation is to search for evidence of the glacial buzz saw in the Andes.
The Andes present a natural experiment for answering this question. Initial uplift began at about 25 Ma to 20 Ma (Gregory-Wodzicki, 2000) and extensive glaciation began in the area at or earlier than 6 Ma (Thomson et al., 2010). They run a north-south transect of the Southern Hemisphere, crossing latitudes from nearly 10ºN down to 55ºS (Fig. 2). Between 35ºS and 55ºS, the study area for this project, virtually all of the region’s precipitation – 2 to 4 m yr-1 – falls on the western flank of the mountain belt, creating and maintaining large glaciers
is dry, thereby better preserving the glacial deposits, both moraines and tills. The final feature of the Andes is the fortuitous presence of magmatism throughout the Pliocene and Pleistocene, which brackets the glacial deposits between planar, datable layers.
Figure 2: This figure shows the latitudes spanned by the Andes in South America. (a) Maximum and mean elevation (thin line and gray area, respectively) for each degree latitude as well as modern snowline and Pleistocene low elevation. (b) Topography and bathymetry of western South America and the eastern Pacific Ocean. (c) Mean annual precipitation on a relief map of western South America (from Montgomery et al., 2001. See references for access to higher-resolution image) I use thermochronometric dating to estimate paleo-erosion rates and better understand the surface processes that sculpted the landscape long ago.
What moved where? How?
We observe that the moraines that lie east of the Andes contain 5-10%
which makes up the core of the Andes. The cobbles show evidence of being transported by glaciers, such as roundedness or striations in some cases. The batholith and the ice fields lie entirely west of the continental divide so mass transport between the western and eastern flanks is minimal. There is no modern feasible way for the granite to get from the source region to the deposit, that is, from the west side to the east side. This cannot always have been the case. The presence of these cobbles to the east tells us that the ancient Andes were different from the Andes of today.
The goal of this project is to reconstruct the landscape of the Patagonian Andes and understand the role that glaciation has played in defining mountain topography and mass transport in the region. This investigation involves observing the relative positions of the ice divide and the continental divide through the Last Glacial Maximum (LGM) and Greatest Patagonian Glaciation (GPG), measuring the cooling ages of batholith-sourced granites and calculating
Study Area Figure 3: The moraines relative to Lago Buenos Aires. Samples presented at this stage of the research were collected from Fenix I, Telken, and the Mercer.
Created by C. Willett using Google Earth. See appendix images A3-A5 for photographs of the three sampling sites.
The southern Andes present a series of glacial moraines and tills interspersed with extrusive igneous basalts. The study area is to the east and south of General Carrera-Lago Buenos Aires (LBA) and contains the Meseta Lago Buenos Aires, from roughly 45°S to 48°S and 70°W to 73°W (Fig. 3). This location is also due east of a tectonic triple junction, a divergent plate boundary that is being subducted beneath a third plate. The subduction of this plate boundary is thought to bring about the magmatism of the region (Lagabrielle et
Cycles of glaciation and deglaciation are interspersed with basaltic eruptions that create datable, planar boundaries that bracket the glacial deposits.
Previous studies have successfully dated a number of these flows, which jumpstarts this project by giving a time of deposition (Kaplan et al., 2004; Singer et al, 2004; Mercer and Sutter, 1982). Knowing the time of deposition is one if the key components in calculating lag time, described in Methods.
Additionally, the simple geology of this portion of the Andes makes it easier to determine when and where transport took place. The Patagonian batholith, which makes up the ridge crest and western flank of the Andes in the study area, is the only source of granitic rocks in the region. The tills and moraines in the study area contain granite cobbles as well as regionally metamorphosed rock associated with the batholith.
Methods Four main techniques were used in the reconstruction of the Patagonian Andes: digitizing and mapping ice extent, calculating flow lines, measuring age, and calculating paleo-erosion rates.
I. Mapping Ice Extent In order to understand the relative positioning of the modern continental divide, which is the Chile-Argentina border north of 52°S, I studied Hulton’s ice model, which estimates ice extent and thickness in all of Patagonia at the Last
Using Didger 4, I georeferenced the Hulton image to a reference map of South America, created with the General Mapping Tool (GTM) in Universal Transverse Mercator (UTM WGS 84). Ten points of reference were used for an accurate fit. In a new layer, I traced each 250-m ice elevation contour using a polyline and assigned the elevation value in meters as the primary polyline attribute. This new layer of the digitized, georeferenced ice sheet was then saved as a shapefile and exported to ArcMap 10.
Within ArcMap, the shape files were added to a 2-D geologic map of the research area. Creation of an interpolated 3-D ice surface is in progress.
Incorporating these surfaces into modern 2-D and 3-D maps of the region answers two questions: if and when. First, it allows us to determine if the ice sheet was ever large enough and positioned properly to move material over and east of the modern continental divide. Second, it gives a time period in which transport from the batholith on the western side to the glacial deposits on the eastern side of the continental divide could occur.
II. Calculating Ice Flow To estimate the flow of ice in the ice sheet, I exported the contour lines from Methods I as a shapefile from Didger and imported it into Matlab, I then aided Dr. Brandon as he completed the bulk of the following steps. First, all the contour lines were converted into points with elevation as an attribute for each.
Second, we set up a set of points along the perimeter of the grid that were lower
that in hand, we made a set of points evenly distributed along the ice sheet outline (the mask contour). With this completed dataset, we simply calculated streamlines, a Matlab command that calculates lines that follow the maximum gradient at each point, creating a path of steepest descent. These lines were the plotted with the mask contour and the whole shape placed over a map of southern South America in ArcMap (Fig. 8).
III. Measuring Age The (U-Th)/He method of dating apatite-containing rocks, like the granite of the Patagonian Batholith, is based on the accumulation of 4He produced by the α-decay of the parent 238U, 235U, and 232Th isotopes. Radiogenic 4He readily diffuses out of the apatite until it cools to ~70ºC, after which diffusion essentially stops. The concentration of 4He and the parent isotopes can be used to calculate a cooling age. Measurements are by first degassing of the crystal by heating and gas-source mass spectrometry to measure 4He, followed by inductively-coupled plasma mass spectrometry on the same crystal to measure U and Th (Reiners and Brandon, 2006).
Because the relation between temperature and depth is well understood below the earth’s surface, the cooling age offers an insight into how far below the surface the diffusion of 4He stopped and the clock started. Assuming a simple exhumation path, this length scale and time scale can be used to estimate a
Figure 4: Apatite-containing rock cools below its closure temperature (Tc, range of a few degrees because of Partial Retention Zone (see Reiners and Brandon 2006)) at time tc as it is exhumed to the surface. Once at the surface, at time te, cobbles are picked up by glaciers and ice sheets and deposited at time td. Lag time simply integrates between tc and td and mainly represents the exhumation time (C. Willett, after Bernet et al, 2009).
Samples were collected from three glacial deposits near Lago Buenos Aires in Provincia Santa Cruz, Argentina. Cobbles were selected from these deposits because each has bracketing basalt flows that have been dated. The deposits are: the Fenix I moraine, deposited during the Last Glacial Maximum at
15.6±1.1 ka (Kaplan et al., 2004); Telken moraine, with a depositional age of 760±14 to 984±35 ka (Singer et al, 2004); and the oldest preserved till, here called Mercer, with a depositional age of 7 to 5 Ma (Mercer and Sutter, 1982).
We employ a method wherein we date multiple granite cobbles from one source region to establish tighter bounds on the igneous rock source area. Each granite cobble weighed 6 to 10 pounds. In the rock lab at Yale University, cobbles were coarsely crushed on a BICO Chipmunk Crusher, finely ground on a disc mill (BICO Braun Pulverizer, type UA), and sieved on a 32-mesh of 500 microns using a Ro-Tap. Samples were then bagged and shipped to the University of California Santa Cruz where Keith Ma and I completed the mineral separations. First, the samples were separated by density on a shaker table, which separates the densest roughly 20% of the sample from the rest. After an overnight soak in hydrogen peroxide, samples were dried in a 40ºC oven and prepared for magnetic separation on the Franz Isodynamic Separator.