**4. Relationship to past cold events in North America**

The first ocean water temperatures below freezing point were appearing in North America starting about 3.5 Ma [38]. Sea temperatures are currently estimated to be 7°C higher today than during the Last Glacial Maximum and the air temperature is 10°C higher. Drill cores from the Greenland ice cap and cores from sea bed sediment have been used to conclude that fluctuations between warm and cold periods during an ice age are caused by changes in the sea currents in the North Atlantic, particularly with relation to the process called the Thermohaline Circulation (THC). The current thinking by Oceanographers is that the changes in water temperature and salinity are probably the main triggers causing repeated glaciations in North America over the last 3.5 Ma [5, 70]. Krissek [71] noted late Cenozoic ice-rafting records from Leg 145 sites in the North Pacific Ocean, concluding that they started in late Miocene times and intensified during the late Pliocene times. Mudelsee and Raymo [72] suggested that the first glaciers had developed by 3.6 Ma B.P. based on patterns of globally distributed marine δ 18O records. Haug *et al*. [73] argued that a saline arctic halocline had developed in the North Pacific Ocean by 2.7 Ma B.P. with warm sea surfaces.

Since nowhere had a complete section showing the full sequence of glacial tills been found, Harris [38] summarized the dated evidence for glacial events and

#### *Causes and Mechanisms of Global Warming/Climate Change DOI: http://dx.doi.org/10.5772/intechopen.101416*

permafrost in North America to provide a guide as to when and where the known locations of the products of past cold events have been found. Methods of dating included tephrochronology, magnetostratigraphy, radiocarbon dating, and potassium-argon dating. The oldest dated deposits from North America included subglacial tuyas with glassy, fractured upper surfaces dated at 3.5 Ma B.P. from Wells Gray Park, British Columbia [74]. Cioppa *et al*. [75] reported that the lowest allow formation of the Kennedy Drift (Unit 1) in Southwest Alberta and Montana has Gauss normal polarization indicating that it is older than 2.6 Ma B.P. The first evidence of an ice cap consists of the Klondike gravels recently dated at 2.64 Ma B.P. with late Gauss magnetism [76]. This is believed to be the most extensive ice cap to have formed in Alaska while its magnetostratigraphy suggests that the increased amplitude of the 41 ka obliquity cycle had already developed [77]. This icecap is regarded as resulting from the warm sea surfaces at that time as well as being the source of the North Pacific ice-rated debris in the marine cores. This ice sheet also altered the course of the Yukon river [78]. About 2.8–2.4 Ma B.P., the Panama Gateway between North and South America closed up so the ENSO currents could not pass into the Atlantic Ocean to take part in warming the North Atlantic.

Altogether, 13 major cold events were recognized in the late Pliocene-Pleistocene record from North America, although some minor events were missed, and the dating of some of the evidence has been changed as a result of more recent work. The sequence started off with isolated events, followed by definite ice sheets separated by large interglacials that decrease in size until 1 Ma B.P. After 750 ka, the 100 ka eccentricity Milankovich cycle has controlled the frequency of the cold events, separated by short (11–20 ka) interglacials.

Barendregt and Irving [79] showed that each time there was a change in the direction of the Earth's magnetism, there was a significant change in the areas affected by glaciations. This implies that there were tectonic changes to the landscape at those times. In the case of the last change c. 1.0–0.8 ka B.P., the mountain range that had separated the Hudson Bay drainage from the Arctic drainage sank to form the present-day Arctic Islands. Meanwhile, the Western Cordillera tilted northwards so that the highest land lay in northern Mexico. This caused the incision of drainage causing the deep erosion of the Copper Canyon in northern Mexico and the Grand Canyon in the southwest United States.

Tephrochronology and magnetostratigraphy both have their limitations. The first is only helpful if there are enough tephras that can readily be recognized that are widespread and have known ages. Magnetostratigraphy sorts the deposits by normal and reversed categories but it cannot prove which deposit is which within each group. In spite of this, it can be useful when determining the ages of a pile of different sediments using the known age sequence corresponding to the Quaternary deposits. Barendregt and Duk-Rodkin [80] produced maps of the distributions of normal and reversed tills corresponding to the four main magnetic zones in the period back to 2.8 Ma B.P., but they do not exhibit the former distribution of each former ice sheet.

During the last seven glaciations, each glacial cycle has spanned about 4 different marine temperature cycles (see **Figure 4**) but varies somewhat independently of them. The best indicator of the volume of ice present on land is provided by the changes in sea level (**Figure 12**). On land, the Wisconsin Glaciation began with the development of extremely cold winter temperatures in Siberia, northeast China, and Greenland with the cold air generating heavy snow accumulations over parts of eastern and northern North America [52, 56] corresponding to a reduction in sea level of c. 30–50 m, starting after 120 ka B.P. The growth of the ice centers continued during the Early Wisconsin glaciation for about 30 ka before a warmer period, probably due to the bipolar seesaw, heralded the commencement of the

#### **Figure 12.**

*Mean sea levels during the last 140 ka B.P. (http://rst.gsfc.nasa.gov/Sect16/Sect16\_2html).*

Middle Wisconsin Interglacial at about 75 ka B.P. (**Figure 12**). While the seawater warmed, the snow-covered ice caps reflected much of the incoming solar radiation back into space so that they only lost about half their ice mass. The next cooling of the seawater began around 70 ka B.P. with the commencement of the late Wisconsin Glaciation. Alternately cooling and warming seas resulted in fluctuating but alternating periods of glacial advances and retreats from the existing ice centers as the climate on land continued to cool. However, the southern Rocky Mountains south of the Peace River became ice-free by 59 ka B.P. and continued like that until 31 ka B.P. [81, 82]. It is thought that the climate there during that period was similar to today.

About 30 ka B.P., there was a major increase in the cold temperatures accompanied by a major drop in sea level of c.50 m that caused expansion of the ice sheets which advanced across the Canadian Prairies to the foothills of the Rocky Mountains. Concurrently in northeast China, there were extremely cold, dry conditions on the Tibetan Plateau, *e.g*., [45, 83–86] and very cold air traveled eastwards to northern Canada along paths II and III in **Figure 9**, probably over open water.

At this time, the Arctic Front was well south of the USA border and the main ice sheets and displacement of biogeographic zones extended far to the south. In order to have deglaciation, a vast amount of Arctic air had to be turned into the tropical air. The tilt of the Earth was changing causing increasing insolation in the Northern Hemisphere resulting in expansion of the air masses, which was greatest over land areas in the south. This applied extra pressure on the Arctic air mass over the northern lands which were relieved by the activation of the path I in **Figure 9** because the air over the Pacific Ocean did not warm up much due to the increased heat being absorbed in the water [52, 53]. Since the cold Arctic air crossed the open Pacific Ocean, it picked up both heat and moisture, only to deposit the moisture as snow in the mountain cirques beginning in 31 ka B.P. The air then crossed the continental divide and continued east downslope as a dry Chinook wind, starting the retreat of the local glaciers. By 25 ka, the cirques on the west side of the Cordillera were full and the glaciers started to descend to the valley floors. After 20 ka B.P., tremendous mounts of snow were deposited covering the southern Cordillera [87, 88] and formed a dome covering the mountains at latitude 54°, so relieving the pressure on the dwindling Arctic air mass and causing the Arctic Front to migrate north to about its present position. From 15 ka to 10 ka, the flow from the Arctic air mass decreased so that the last active glaciers ceased expansion about 10 ka B.P. in the Fraser Valley and on the east coast of Haida Gwai. By then, the Laurentide Ice Sheet had retreated from the Prairies to form a mass occupying Hudson Bay. It is likely that this pressure release also permitted the Scandinavian-Russian ice cap and the bulk of the ice in the Swiss Alps to retreat at the same time [89–91].
