Jonathan Harvey

Materials Contributed through SERC-hosted Projects

Other Contributions (2)

Glacier Fluctuations Since the Last Glacial Maximum in Southwest Alaska part of Vignettes:Vignette Collection
Glacier fluctuations since the last glacial maximum in southwest Alaska Introduction During the last glacial maximum (LGM), alpine glaciers in the western cordillera expanded, coalesced, and flowed onto newly-emergent continental shelves (Figure 1). During and after deglaciation, rising seas again submerged these shelves, obscuring the deposits left in the termination zones of these glaciers. Luckily for geologists, outlet glaciers also flowed into unglaciated lowlands on the landward side of the cordillera. The deposits associated with these land-terminating glaciers are relatively well-preserved, and can be used to reconstruct the extent of glaciation during the LGM, as well as the timing and style of deglaciation. Ongoing research in southwest Alaska provides an excellent learning example of the use of these geomorphic records to reconstruct landscape history. Geomorphic evidence of the last glacial maximum in southwest Alaska The Alaska Range is a rugged mountain range in south-central Alaska, home to Mt. McKinley (Denali), the highest peak in North America (6194 m). It is also the northernmost extent of the Cordilleran Ice Sheet, a continuous ice cap that formed along the Pacific coast of North America during the LGM. The western flank of the Alaska Range is an area of ongoing study of glacier dynamics during and following the LGM. Well-preserved moraines and associated glacial deposits record and permit reconstruction of the events that occurred there. The most prominent depositional glacial landforms in the area are the terminal moraines emplaced by piedmont glaciers during their late Wisconsin maximum extent. Over many hundreds of years, glacial till melting out of the ablation zone of the glaciers accumulated at their snouts to form broad, 1- to 2-km-wide hummocky ridges. Melting of ice blocks buried within the moraine resulted in the formation of hundreds of small kettle lakes on this landform. These moraines are visible in elevation data and satellite imagery and can be traced for hundreds of miles along the western flank of the Alaska Range (Figures 2, 3). Upstream from the terminal moraine, a series of similar landforms are present. These are recessional moraines, recording the long process of retreat following the LGM. These moraines were built by glaciers that either temporarily readvanced during overall retreat, or simply paused long enough to build a significant moraine. In many locations in southwest Alaska, at least four recessional moraine-building events, or stades, are recognized, often correlatable from valley to valley. Determining the age of glacial landforms How can we constrain the ages of the LGM and subsequent recessional stades? Because very little vegetation was present during the last glacial maximum, radiocarbon is rarely able to produce more than broadly-defined minimum and maximum ages. Optically-stimulated luminescence has the advantage of not requiring organic matter–it dates sand grains directly. However, OSL has not been thoroughly tested on glacial deposits in Alaska, and it only works on those grains that were sufficiently exposed to sunlight before deposition. Cosmogenic radionuclide (CRN) dating is a tool that is well-suited to determining the age of a glacial moraine. By measuring the levels of cosmogenic isotopes (isotopes produced by the interaction of high-energy particles from space with minerals on earth) in boulders on the crest of a moraine, geologists can determine how long the boulder has been exposed. However, workers must be careful to select only those boulders that appear to be relatively stable. Boulders that have shifted or rolled since deposition will have received less cosmogenic radiation, resulting in ages younger than the age of the landform. An additional source of uncertainty is shielding of cosmic rays by seasonal snowcover. This source of uncertainty can be reduced by sampling only those boulders that protrude well above the surrounding moraine surface. CRN Samples taken from boulders on the LGM terminal moraine on the western flank of the Alaska Range suggest that the moraine stabilized between 19-23 ka. The actual age of the LGM is likely toward the older end of this range, because the moraine boulders were probably subject to shifting for several hundred years after deposition. The recessional moraines are as yet undated. However, other forms of evidence provide insight. Radiocarbon dates from the base of cores taken from several kettle lakes along the shores of Lake Clark indicates that the glacier front had retreated into mountain valleys by 15 ka. Deglaciation was probably complete in southwest Alaska by around 11 ka. Neoglaciation Temperatures were at or above modern levels during the earliest Holocene, a period dubbed the Holocene Thermal Maxmimum. Most glaciers were probably smaller during this period than they are today. By the mid-Holocene (~5 ka), global temperatures were once again dropping, causing glaciers to advance beyond their early Holocene minima. This renewed glacial activity is referred to neoglaciation. On the western flank of the Alaska Range, 2-4 moraines of probable mid- to late-Holocene age are present. These relatively small moraines are overwhelmed and partially buried by a much larger, more voluminous moraine immediately downvalley of modern glacier termini (Figure 4) This younger moraine is often sharp-crested, unvegetated, and up to 100 m high. Field observations indicate that it is ice-cored, suggesting that its apparent volume may be transient. Dating of the older neoglacial moraines could best be accomplished with cosmogenic radionuclides. However, CRNs are not as effective for the younger, ice-cored moraine. Instead, its age is best constrained with lichenometry. Lichenometry is a method wherein the age of a moraine is estimated by comparing the largest lichens growing on moraine boulders with a known lichen growth rate. Though this method is somewhat imprecise (uncertainty ~ +/- 20%), lichen sizes on the youngest, ice-cored moraines suggest that they date to the culmination of the Little Ice Age (LIA) toward the end of the 19th century.

Arroyo Cutting in the Southwestern U.S. part of Vignettes:Vignette Collection
Introduction In the semiarid southwestern United States, water is extremely important. However, most streams in the region are ephemeral, meaning that they are dry for much of the year and flow only during flash floods and/or spring snowmelt. Many of these streams are confined to deep, steep-walled channels entrenched into alluvial valley bottoms. These channels are called arroyos, and can be found in drylands throughout the world (Figure 1). Most arroyos in the Southwest were cut between 1880 and 1910 AD. Prior to this historic episode of arroyo-cutting, settlers characterized valley bottoms as marshy wetlands or grasslands with shallow channels. During arroyo-cutting, observers described the upstream migration of steep headcuts during successive floods. This incision and subsequent widening of channels led to the loss of thousands of acres of farmland; destruction of roads, dams, and canals; and ultimately the abandonment of several communities. Figure 2 shows the striking example of Kanab, UT, where a ~30 m deep arroyo now runs along the town's western edge. Hypothesized causes of arroyo cutting For almost a century, geologists have debated the cause of historic arroyo cutting. Most hypotheses fit into one of three categories: (1) Intensive grazing and floodplain alteration associated with Anglo settlement led to increased runoff during storms, causing more erosive floods; (2) Arroyo cutting was the result of climate change. A wide range of climatic mechanisms have been suggested, ranging from fluctuations in groundwater levels to episodes of catastrophic flooding; and (3) Arroyo cutting and filling is an inherent behavior of semiarid drainage systems, wherein pulses of incision followed by widening and eventually aggradation migrate upstream in response to local slope anomalies. This last hypothesis is known as the 'complex response' hypothesis. In reality, each of the above hypotheses likely played some role in arroyo cutting. Despite the complex interactions of these possible driving mechanisms, the regional synchronicity of historic arroyo cutting has led many recent workers to conclude that the dominant driving mechanism must be one that operates across a whole region rather than on individual watersheds. As a result, much attention has been focused on the various climatic drivers proposed in hypothesis (2). Stratigraphic evidence Cutbanks along modern arroyos provide 'windows' into the hydrologic history (paleohydrology) of a stream. Alluvial deposits and stratigraphic relations exposed in these windows reveal that many cycles of arroyo cutting and filling have occurred throughout the Holocene (~10,000 BC-present). Evidence of these cycles lies in the form of aggradational sequences overlying and truncated by paleochannels (e.g. Figure 3). Usually, the aggradational "packages" represent the filling of an arroyo with sediment, and the paleochannels represent periods of arroyo cutting. Geologists initially relied on the relation of these aggradational packages to cultural artifacts of known age in order to establish their ages. Radiocarbon dating played a major role where no cultural features could be associated with alluvial deposits. Now, geologists are using optically-stimulated luminescence (OSL) dating where no cultural features or organic material can be found. All of the above tools can be used to constrain the timing of past arroyo cycles in order to establish their relation to hydrological and environmental changes. Regional Correlations Such chronostratigraphic studies have identified at least one other arroyo cutting event that was synchronously manifested in many drainages of the Southwest. This pre-historic arroyo cutting occurred around 1200 AD. The fact that this event took place in the absence of intensive grazing and floodplain modification suggests that it was driven by climatic changes. Interestingly, the pre-historic arroyo cutting also coincides roughly with the great Puebloan abandonment–when seeming successful, growing Puebloan cultures suddenly disappeared from the southern Colorado Plateau. Was prehistoric arroyo cutting related to this major cultural transition? It certainly seems plausible, as archeological evidence suggests that these burgeoning cultures relied on floodwater farming. If arroyos rapidly cut into the floodplains they were farming, irrigation would have become very difficult. These two major arroyo cutting events bound a period of channel aggradation from around 1300-1880 AD. Valley-fills of this age have been correlated across the Southwest (Figure 4). Hack (1942) dubbed this alluvial fill the 'Naha' alluvium. It generally fills arroyos that were cut into an older alluvial fill known as the 'Tsegi' alluvium, which encompasses several aggradational sequences that occurred before 1200 AD. Ongoing research is focused on identifying the timing of these older arroyo cutting-and-filling cycles. Some studies have identified upwards of 10 alluvial fills that predate the "Naha' alluvium, though existing age control is limited. Hypothesis Testing If arroyo cutting events are climate-driven, what specific mechanism(s) are at play? Many have suggested that episodes of frequent, high-magnitude flooding were the causal mechanism. How can we test this hypothesis? One way would be to examine the flood record for a stream for which an arroyo cut-and-fill record is known. Because stream gages have existed for only a short period of time, geologists use paleoflood hydrology to characterize pre-instrumental floods. Evidence of floods can be depositional (e.g. flood deposits, driftwood) or erosional (e.g. truncated alluvial fans, scour lines). Such evidence acts as a proxy for flood stage, which, combined with detailed surveying, can be used to calculate the paleodischarge. Additionally, the same geochronological tools described above can be used to determine the age of the flood. These tools allow construction of the flood history of a stream. If the reconstructed flood history for a stream suggests that arroyo cutting events coincide with episodes of frequent, high-magnitude flooding, it would suggest that that these episodes are the primary drivers of prehistoric and historic arroyo cutting events. If not, then geologists must come up with alternative explanations for these regionally synchronous channel changes.