Geology and Earth ScienceCorrelation of Dansgaard-Oeschger Events in Glacial Cores to Continental Strata of the Southwest United States: Evidence of Global Impact
Sean Tvelia
During the last glacial maximum Greenland ice cores recorded variations of oxygen isotope ratios that have since revealed a series of rapid warming events followed by a gradual cooling. Researchers recognize twenty such events, known as Dansgaard- Oeschger events, occurring between 100kya to 14kya with a periodicity of roughly 1500 year. Little evidence of these warming events exists in continental strata which leads some to speculate whether or not these events were global events. Current research conducted on lake sediments in the southwest United States seems to indicate transgressive/regressive cycles also on the order of 1500 year cycles. This evidence, along with flora evidence, such as variation in pollen, seems to indicate that these rapid changes in climate as recorded in the Greenland ice core may have indeed been global events.
Most of our knowledge of Dansgaard- Oeschger events comes from three ice cores: The Greenland Ice Core Project (GRIP), The Greenland Ice Sheet Project 2 (GRISP2) and the West Antarctica Byrd Station (Roe & Steig, 2003). As can be seen in Figure1, these cores all show d18O variation occurring with a regularity of about 1500 years.
Since d18O ratios are directly related to temperature, the variation of d18O within ice cores represents the fluctuation of Earth’s climate (Ganopolski & Rahmstorf, 2001). In the three studied cores, d18O levels indicate a relatively fast warming trend, with mean temperature changes on the order of 10°C, over just a few decades (Roe & Steig, 2003). This rapid warming event, represented by the spike in d18O in figure 1, is followed by a gradual return to cooler temperatures. Twenty such variations have been recorded in the Greenland and Antarctic ice cores and are labeled 1-20 in figure 1.
With such great fluctuation in temperature over relatively short periods of time there have been many attempts to model processes that may trigger these events. Most of these models involve adjustments of the global thermohaline current due to variability in solar radiation, due to sub-Milankovitch cycles, and fresh water influx. This however suggests that Dansgaard-Oeschger were not simple polar anomalies and indeed have global implications.
Finding evidence of Dansgaard-Oeschger events within continental strata has been somewhat daunting. Glacial advances during successive cooling phases would have no doubt removed all traces of previous glacial retreats during the rapid warming phase therefore leaving behind little to no trace of the event. Therefore, continental correlation of these events relies heavily on the impact on local deposition due to climatic variation caused by possible Dansgaard-Oeschger events; specifically lacustrine deposition which are intrinsically tied to local climate variation.
Local climatic variation during the last glacial maximum is most directly related to the development of high pressure cells over glaciated regions. During the last glacial maximum high pressure cells that developed over the Laurentide Ice Sheet acted as a wall that forced the jet stream over progressively lower latitudes (Tchakerian & Lancaster, 2000). This, in turn, forced Pacific storm to track further south and led to increased precipitation in previously arid regions. This southward movement of the ice sheet also led to increased aeolian processes due to the sharp pressure gradient that existed between the cold glaciated regions and the warmer basin regions of the Southwest (Tchakerian & Lancaster, 2000).
Evidence of cyclical variation in deposition in response to Dansgaard-Oeschger events can be seen in the southwest United States and Northern Mexico. One such region, the Trans-Pecos Basin (figure 2) records fluctuating paleolake levels that seem to reflect Dansgaard-Oeschger events due to the geologic nature of the basin.
/Wilkins and Currey noted that “The Trans-Pecos closed basin is an internally drained and hydrographically closed region located between the Pecos River and the Rio Grande drainages in far west Texas and south central New Mexico” (Wilkins & Currey, 1996). The basin encompasses an area of over 20,000km2 and ranges in elevation from 2835ft in the north to 1087ft in the south (Wilkins & Currey, 1996). In 1948 P.B. King proposed that beach ridges along the northwest margin of this basin were evidence of Pleistocene lake levels; this area is now dubbed Lake King (Wilkins & Currey, 1996).
Although current annual rainfall in the area (total- 280mm to 760mm) seldom reaches the floor of the basin, the Permian Bone Spring limestone provides significant groundwater to the region (Wilkins & Currey, 1996). According to Wilkins and Currey, Bjorklund determined the water table to be very near the surface of the basin which results in numerous visible depressions and sinks. (Wilkins & Currey, 1996)
Evidence of Lake King consists of “finely laminated centimeter scale couplets of olive green and gray” evaporate laminae which are composed of calcite, gypsum, and halite. Hussain et al. (1988) identified the olive layer as being gypsum rich while the darker gray layer is micritic and richer in organics (Wilkins & Currey, 1996).
This varve-like pairing of layers seems to indicate seasonal deposition of the sediment. Wilkins and Currey suggest that this was created when seasonal runoff increased calcium levels and flushed organic material into the lake, creating the dark gray micritic layer. Then when runoff decreased total alkalinity also decreased. During these periods the lake became enriched in sulfate due to underlying groundwater influx and gypsum deposition occurred (Wilkins, Currey, 96).
Throughout these varved lake sediments are layers of blackened dolomite that have been termed “black mats” during a previous study by Friedman (Wilkins, Currey, 1996). These layers are enriched in 18O and therefore suggest being formed during periods of increased evaporation or lowered precipitation (Wilkins, Currey, 1996). Friedman’s theory suggests that during these periods, dense brines forming along the margin of the lake would have sunk to the center and replaced the previously formed calcite with dolomite (Wilkins, Currey, 96). This model then suggests that the dolomite production would also coincide with shallow lake levels and therefore regressive periods.
When precipitation returned to the area, freshwater influx via runoff would have interrupt the normal thermohaline circulation and resulted in anoxic conditions at the base of the lake(Wilkins & Currey, 1996). Sulfides, produced through metabolic processes of anaerobic bacteria, then reduced iron oxides present in the lake and blackened the dolomite(Wilkins & Currey, 1996).
In this model the initial production of dolomite indicates dry, arid climates and regressive phases of the lake. Black mats, on the other hand, reflect transgressive phases of Lake King and must also reflect moist or humid climates. Along with
providing stratigraphic evidence of these climate shifts, black mats of the Trans-Pecos basin also contain enough organic carbon that may be radiometrically dated.
Radiocarbon dates of black mats reveal a similar pattern to Dansgaard-Oeschger events in Greenland ice cores. Figure 3 shows the correlation of black mat production ages with lake surface elevation as identified by shoreline. The lowstand, which is illustrated by dolomite production, was arbitrarily determined by the lowest visible shoreline (Wilkins & Currey, 1996).
Further evidence of the global impact of Dansgaard-Oeschger events can be seen along the Owens River system of the southwest United States.
Pictured in Figure 4, the Owens River system is fed by discharge from the Sierra Nevada Mountains. This system is made up of a series of closed basins, each of which receives overflow waters from the previous basin starting with Owens Lake. This creates a chain of pluvial lakes that, at times, extended into Death Valley (Tchakerian & Lancaster, 2000).
During the last glacial maximum, this system would have been fed by alpine glaciers in the Sierra Nevada Mountains and may have also received additional waters from the overflow of Lake Mono just north of the Owen River headwaters (Tchakerian & Lancaster, 2000).
Tchakerian & Lancaster further note that stratigraphic data indicate that Searles Lake levels went through significant oscillation between 100-25ka, with the greatest fluctuations occurring between 34 and 26ka. These fluctuations are marked by evaporate layers which would indicate the onset of an arid climate (Tchakerian & Lancaster, 2000). During this period Searles Lake appears to have had six major transgressive/regressive cylces (Figure 5) that can be correlated with Dansgaard-Oeschger event in the Greenland ice cores.
According to Tchakerian & Lancaster, similar oscillations have also been detected at Mono Lake, where each low lake level lasted between 1000 and 2000 years (Tchakerian & Lancaster, 2000). The cyclical nature and timing of these events do seem to indicate a correlation between arid/humid cycles in the southwest resulting from Dansgaard-Oeschger events recorded in the Greenland cores.
Based on models produced by Hostetler and Clark, control of glacial advance and retreat in most of the western United States during the last glacial maximum was dominated by changes in global temperature (Hostetler and Clark, 1997). In the Sierra Nevada Mountains, however, this same model shows that the advance and retreat of alpine glaciers was dominated by precipitation. This model fits well with Tchakerian & Lancaster’s model that suggests that displacement of the jet stream by the advancing ice sheet forced Pacific storm tracks further south. This, in turn, led to increased precipitation over the Sierra Nevadas, which in turn led to alpine glacial advance that fed the Owens River system.
Fluctuations of lake levels recorded in the stratigraphic record of these two localities along with the timing of the events imply correlation between events recorded in the Greenland and Antarctic ice cores. These correlations suggest that temperature fluctuation recorded by d18O in Greenland ice cores was not limited to the Polar Regions. Furthermore climatic modeling of the last glacial maximum also correlates with processes leading to transgressive/ regressive cycles in the study region.
References:
Ganopolski, A. and Rahmstorf, S., 2001: Rapid changes of glacial climate simulated in a coupled climate model. Nature, 409, 153-158
Hostetler, S. and Clark, P, 1997, Climatic Controls of Western U.S. Glaciers at the Last Glacial Maximum, Quaternary Science Review, v16, p. 505-511
Roe, G.H., Steig, E.J., 2004, Characterization of Millennial-Scale Climate Variability, Journal Climate, v17 no. 10, p.1929-1944
Tchakerian, V.P. and Lancaster, N., 2002, Late Quaternary Arid/Humid Cycles in the Mojave Desert and Western Great Basin of North America, Quaternary Science Reviews, v.21, p. 799-810
Wilkins, D.E. and Currey, D.R.,1997, Timing and Extent of Late Quaternary Paleolakes in the Trans-Pecos Closed Basin, West Texas and South Central New Mexico, Quaternary Research, v. 47 p. 306-315
During the last glacial maximum Greenland ice cores recorded variations of oxygen isotope ratios that have since revealed a series of rapid warming events followed by a gradual cooling. Researchers recognize twenty such events, known as Dansgaard- Oeschger events, occurring between 100kya to 14kya with a periodicity of roughly 1500 year. Little evidence of these warming events exists in continental strata which leads some to speculate whether or not these events were global events. Current research conducted on lake sediments in the southwest United States seems to indicate transgressive/regressive cycles also on the order of 1500 year cycles. This evidence, along with flora evidence, such as variation in pollen, seems to indicate that these rapid changes in climate as recorded in the Greenland ice core may have indeed been global events.
Most of our knowledge of Dansgaard- Oeschger events comes from three ice cores: The Greenland Ice Core Project (GRIP), The Greenland Ice Sheet Project 2 (GRISP2) and the West Antarctica Byrd Station (Roe & Steig, 2003). As can be seen in Figure1, these cores all show d18O variation occurring with a regularity of about 1500 years.
Since d18O ratios are directly related to temperature, the variation of d18O within ice cores represents the fluctuation of Earth’s climate (Ganopolski & Rahmstorf, 2001). In the three studied cores, d18O levels indicate a relatively fast warming trend, with mean temperature changes on the order of 10°C, over just a few decades (Roe & Steig, 2003). This rapid warming event, represented by the spike in d18O in figure 1, is followed by a gradual return to cooler temperatures. Twenty such variations have been recorded in the Greenland and Antarctic ice cores and are labeled 1-20 in figure 1.
With such great fluctuation in temperature over relatively short periods of time there have been many attempts to model processes that may trigger these events. Most of these models involve adjustments of the global thermohaline current due to variability in solar radiation, due to sub-Milankovitch cycles, and fresh water influx. This however suggests that Dansgaard-Oeschger were not simple polar anomalies and indeed have global implications.
Finding evidence of Dansgaard-Oeschger events within continental strata has been somewhat daunting. Glacial advances during successive cooling phases would have no doubt removed all traces of previous glacial retreats during the rapid warming phase therefore leaving behind little to no trace of the event. Therefore, continental correlation of these events relies heavily on the impact on local deposition due to climatic variation caused by possible Dansgaard-Oeschger events; specifically lacustrine deposition which are intrinsically tied to local climate variation.
Local climatic variation during the last glacial maximum is most directly related to the development of high pressure cells over glaciated regions. During the last glacial maximum high pressure cells that developed over the Laurentide Ice Sheet acted as a wall that forced the jet stream over progressively lower latitudes (Tchakerian & Lancaster, 2000). This, in turn, forced Pacific storm to track further south and led to increased precipitation in previously arid regions. This southward movement of the ice sheet also led to increased aeolian processes due to the sharp pressure gradient that existed between the cold glaciated regions and the warmer basin regions of the Southwest (Tchakerian & Lancaster, 2000).
Evidence of cyclical variation in deposition in response to Dansgaard-Oeschger events can be seen in the southwest United States and Northern Mexico. One such region, the Trans-Pecos Basin (figure 2) records fluctuating paleolake levels that seem to reflect Dansgaard-Oeschger events due to the geologic nature of the basin.
/Wilkins and Currey noted that “The Trans-Pecos closed basin is an internally drained and hydrographically closed region located between the Pecos River and the Rio Grande drainages in far west Texas and south central New Mexico” (Wilkins & Currey, 1996). The basin encompasses an area of over 20,000km2 and ranges in elevation from 2835ft in the north to 1087ft in the south (Wilkins & Currey, 1996). In 1948 P.B. King proposed that beach ridges along the northwest margin of this basin were evidence of Pleistocene lake levels; this area is now dubbed Lake King (Wilkins & Currey, 1996).
Although current annual rainfall in the area (total- 280mm to 760mm) seldom reaches the floor of the basin, the Permian Bone Spring limestone provides significant groundwater to the region (Wilkins & Currey, 1996). According to Wilkins and Currey, Bjorklund determined the water table to be very near the surface of the basin which results in numerous visible depressions and sinks. (Wilkins & Currey, 1996)
Evidence of Lake King consists of “finely laminated centimeter scale couplets of olive green and gray” evaporate laminae which are composed of calcite, gypsum, and halite. Hussain et al. (1988) identified the olive layer as being gypsum rich while the darker gray layer is micritic and richer in organics (Wilkins & Currey, 1996).
This varve-like pairing of layers seems to indicate seasonal deposition of the sediment. Wilkins and Currey suggest that this was created when seasonal runoff increased calcium levels and flushed organic material into the lake, creating the dark gray micritic layer. Then when runoff decreased total alkalinity also decreased. During these periods the lake became enriched in sulfate due to underlying groundwater influx and gypsum deposition occurred (Wilkins, Currey, 96).
Throughout these varved lake sediments are layers of blackened dolomite that have been termed “black mats” during a previous study by Friedman (Wilkins, Currey, 1996). These layers are enriched in 18O and therefore suggest being formed during periods of increased evaporation or lowered precipitation (Wilkins, Currey, 1996). Friedman’s theory suggests that during these periods, dense brines forming along the margin of the lake would have sunk to the center and replaced the previously formed calcite with dolomite (Wilkins, Currey, 96). This model then suggests that the dolomite production would also coincide with shallow lake levels and therefore regressive periods.
When precipitation returned to the area, freshwater influx via runoff would have interrupt the normal thermohaline circulation and resulted in anoxic conditions at the base of the lake(Wilkins & Currey, 1996). Sulfides, produced through metabolic processes of anaerobic bacteria, then reduced iron oxides present in the lake and blackened the dolomite(Wilkins & Currey, 1996).
In this model the initial production of dolomite indicates dry, arid climates and regressive phases of the lake. Black mats, on the other hand, reflect transgressive phases of Lake King and must also reflect moist or humid climates. Along with
providing stratigraphic evidence of these climate shifts, black mats of the Trans-Pecos basin also contain enough organic carbon that may be radiometrically dated.
Radiocarbon dates of black mats reveal a similar pattern to Dansgaard-Oeschger events in Greenland ice cores. Figure 3 shows the correlation of black mat production ages with lake surface elevation as identified by shoreline. The lowstand, which is illustrated by dolomite production, was arbitrarily determined by the lowest visible shoreline (Wilkins & Currey, 1996).
Further evidence of the global impact of Dansgaard-Oeschger events can be seen along the Owens River system of the southwest United States.
Pictured in Figure 4, the Owens River system is fed by discharge from the Sierra Nevada Mountains. This system is made up of a series of closed basins, each of which receives overflow waters from the previous basin starting with Owens Lake. This creates a chain of pluvial lakes that, at times, extended into Death Valley (Tchakerian & Lancaster, 2000).
During the last glacial maximum, this system would have been fed by alpine glaciers in the Sierra Nevada Mountains and may have also received additional waters from the overflow of Lake Mono just north of the Owen River headwaters (Tchakerian & Lancaster, 2000).
Tchakerian & Lancaster further note that stratigraphic data indicate that Searles Lake levels went through significant oscillation between 100-25ka, with the greatest fluctuations occurring between 34 and 26ka. These fluctuations are marked by evaporate layers which would indicate the onset of an arid climate (Tchakerian & Lancaster, 2000). During this period Searles Lake appears to have had six major transgressive/regressive cylces (Figure 5) that can be correlated with Dansgaard-Oeschger event in the Greenland ice cores.
According to Tchakerian & Lancaster, similar oscillations have also been detected at Mono Lake, where each low lake level lasted between 1000 and 2000 years (Tchakerian & Lancaster, 2000). The cyclical nature and timing of these events do seem to indicate a correlation between arid/humid cycles in the southwest resulting from Dansgaard-Oeschger events recorded in the Greenland cores.
Based on models produced by Hostetler and Clark, control of glacial advance and retreat in most of the western United States during the last glacial maximum was dominated by changes in global temperature (Hostetler and Clark, 1997). In the Sierra Nevada Mountains, however, this same model shows that the advance and retreat of alpine glaciers was dominated by precipitation. This model fits well with Tchakerian & Lancaster’s model that suggests that displacement of the jet stream by the advancing ice sheet forced Pacific storm tracks further south. This, in turn, led to increased precipitation over the Sierra Nevadas, which in turn led to alpine glacial advance that fed the Owens River system.
Fluctuations of lake levels recorded in the stratigraphic record of these two localities along with the timing of the events imply correlation between events recorded in the Greenland and Antarctic ice cores. These correlations suggest that temperature fluctuation recorded by d18O in Greenland ice cores was not limited to the Polar Regions. Furthermore climatic modeling of the last glacial maximum also correlates with processes leading to transgressive/ regressive cycles in the study region.
References:
Ganopolski, A. and Rahmstorf, S., 2001: Rapid changes of glacial climate simulated in a coupled climate model. Nature, 409, 153-158
Hostetler, S. and Clark, P, 1997, Climatic Controls of Western U.S. Glaciers at the Last Glacial Maximum, Quaternary Science Review, v16, p. 505-511
Roe, G.H., Steig, E.J., 2004, Characterization of Millennial-Scale Climate Variability, Journal Climate, v17 no. 10, p.1929-1944
Tchakerian, V.P. and Lancaster, N., 2002, Late Quaternary Arid/Humid Cycles in the Mojave Desert and Western Great Basin of North America, Quaternary Science Reviews, v.21, p. 799-810
Wilkins, D.E. and Currey, D.R.,1997, Timing and Extent of Late Quaternary Paleolakes in the Trans-Pecos Closed Basin, West Texas and South Central New Mexico, Quaternary Research, v. 47 p. 306-315
