Unit 4 this week covering chapters 13,14, and 15. It is a lot of material to read and digest. Notes below are the things I highlighted in my book.

13-1 Project CLIMAP: Reconstructing the Last Glacial Maximum

• Climate Mapping and Prediction Project was a cooperative effort to reconstruct past climates started in the 1970. Reconstructed the Earth’s surface during the last glacial maximum
• The thing that stood out from the reconstruction were continent-sized ice sheets covering North America as far south as 37N latitude, Scandinavia down to 48N, and the Arctic margins of Eurasia.
  • Today ice sheets are 3% Earth’s surface and 10% of its land surface and Last glaciation ice sheets were 7% Earth’s surface and 25% of its land surface
• The North American ice sheet was the largest northern hemisphere ice sheet at 55% of the volume of ice in excess of the amount on Earth today.
• The reconstruction show 4C colder than today and the North Atlantic Ocean cooled; ice advanced past its current limit in the Southern Ocean
• Low latitude ocean temperature only slightly cooler than today’s and in some area even slightly warmer
• Over time other groups challenged and refined the CLIMAP reconstructions some things were wrong, but this was the starting point for other attempts to test against

13-2 How large were the ice sheets

• CLIMAP reconstruction published two papers: the original and one that portrayed the ice sheets at their max possible size
• Three highly controversial aspects of the reconstruction
• First was the lateral extent of the ice: Did the ice reach the ocean in higher latitude regions?
  • Some argued that there were less extensive because they were starved from moisture
  • C14 dating along the northern margins of the ice sheets confirmed that most were near or at their max limits 21,000 years ago
• Second: was maritime ice sheets that formed on shallow continental shelves with their bases lying below sea level. There were 2 locations: Barents Sea and Kara Sea
  • Partial accurate: evidence found for Barents Sea and only westmost of Kara Sea. Used glacial debris in sediment cores and depth surroundings showing moraine ridges.
• The third controversy was for the thickness and height of the ice sheet
  • as an ice sheet grows, it eventually weighs down the underlying bedrock by an amount equal to about 30% of its thickness.
  • CLIMAP showed high and thick sheets over North America and assumed near max extent and stiff ice that was that was frozen to the bedrock, which created steep ice sheet and great thickness.
  • Still controversial because the evidence does not tell the entire story. Sea level and coral reefs to estimate the size of ice sheets and rebounding. Results show values in between the two CLIMAP reconstructions

13-3 Glacial Dirt and Winds

• Water freezing and thawing in the cracks create breaks large chunks of rocks off
• Ice sheets carry and deposit debris to their margins when the ice melts creating moraines
• Running water from melting ice or local precipitation reworks the debris, extracting finer sediments and producing glacial outwash
• The wind organized the different grain into different sizes away from the ice margins (Large and medium stay in place while small get transported over short distances)
• Can use chemical analysis to see the source region of the dust and dust accumulated five times faster during the last glacial maximum and during several previous glaciations than it does today.
• Cores from the equatorial Atlantic Ocean reveal that dust was also deposited in that region at higher rates during glaciations.
• Glacial-age layers in ice cores from Antarctica contain more than ten times as much dust as interglacial layers
• One way of testing climate model simulations of the last glacial maximum is to examine the distribution of various kinds of debris carried by winds, ranging from desert sands to windblown silts to fine dust.
• Climate models can simulate the strength and direction of winds from the surface up to jet stream altitudes, the potential exists to compare model simulations with the observed patterns of windblown glacial debris.
• Current climate models can simulating the large-scale circulation of the atmosphere, but not good the smaller spatial scales needed to simulate dust transport. The models are less successful at simulating the processes that actually lift and transport silt and dust from Earth’s surface, such as local wind gusts along frontal systems or small-scale eddies of wind.

13-4 Project COHMAP: Data-Model Comparisons

• The Cooperative Holocene Mapping Project used a combined data-model approach to examine the last glacial maximum and the subsequent change to interglacial conditions.
  • Used 14C-dated records of lake levels and pollen in lake sediments in order to examine regional patterns.
• The first step in the COHMAP approach was to assemble records of the changing boundary conditions that have driven climate over the last 21,000 years.
• Ran model simulations of climate at intervals of several thousand years between the glacial maximum and the present to determine how changes in the major boundary conditions drove regional patterns of climate change.
  • focused on the role of orbital-scale changes in climate over intervals of thousands of years, rather than on shorter-term fluctuations superimposed on this gradual trend.
• The climate data produced as output from these model simulations were then tested against climate reconstructions based on 14C-dated records of pollen from lake cores and plankton shells from ocean sediment cores.
• Modern relationships between the abundances of species and climatic variables can be measured, quantified, and used to reconstruct past climates from fossil organisms.
• By comparing these fossil-based estimates of past climate with the changes simulated by the models, scientists can test the reliability of both approaches.

13-5 Pollen: An Indicator of Climate on the Continents

• Pollen is carried mainly by winds, and to a lesser extent by water and insects. The preserved pollen reflects the average composition of vegetation
• The pollen percentages are generally similar to those of the actual vegetation, although “overproducers” such as pine trees leave disproportionately large amounts of pollen compared to “under producers” such as maples. Climate scientists adjust for this disproportionate representation.
• These climatic controls can also be demonstrated by plotting pollen percentages against different combinations of seasonal and annual temperature and precipitation.
• The bottom layers of sediment in a 14C-dated core from Minnesota are late glacial in age, whereas the upper layers of mud record the postglacial climate of the present interglaciation 
• Most of the pollen in the older layers is from spruce trees, indicating conditions colder than today. An abrupt switch from spruce pollen to warm-adapted oak near 10,000 years ago indicates rapid warming in this region. Subsequent changes to maximum values of dry-adapted herb and grass pollen culminating near 6,000 years ago indicate a climate drier than today’s. Many hundreds of similar cores have been examined in North America, as well as additional hundreds in Europe and elsewhere in the world.
• Viewed together, these records provide a broad geographic perspective on the pattern of pollen distribution at the last glacial maximum and during the deglaciation.
• This larger map perspective can be compared with map patterns produced by model simulations.

13-6 Using Pollen for Data-Model comparison

• Data-model comparisons focus on the distribution of pollen at specific intervals in the past across geographic regions.
• These model-simulated pollen distributions result from several steps.
  • First, boundary conditions are chosen and used in model simulations
  • Then the model-derived estimates of temperature and precipitation are used to generate estimates of the percentage abundance of each type of pollen based on the modern relationship between climate and pollen.
  • Each estimate of annual precipitation and mean July temperature simulated for a specific grid box in the initial simulation yields a specific estimate of the percentage of oak, spruce, and prairie pollen for that particular location.
  • The map patterns of pollen abundance estimated in this way can then be compared directly with the map patterns based on pollen counts from lake cores.
• Comparisons of observed and modeled vegetation for times in the past with different boundary conditions can also be made using biome models.
  • different boundary conditions are used to drive a general circulation model simulation of a specific time slice.
  • Based on the temperature-precipitation output from the model, the possible range of major vegetation types that can occur in each region is narrowed to larger biome units.
• Both steps are based on observed modern relationships between vegetation and the environment.
• Because the first step in the biome method encompasses all of the major vegetation groupings on Earth, this approach can simulate changing patterns of vegetation on any continent.

13-7 Model Simulation for Maximum Climates

• Glacial ice sheets are a critical boundary condition for simulations of glacial climate.
• Their central domes protruded upward as massive icy plateaus, redirecting the flow of air above the ice.
  • Climate model simulations suggest that a high-domed ice sheet over North America could have split the winter jet stream into two branches at the glacial maximum
• In modern winters, a single jet stream enters North America near the border between Canada and the United States.
  • Storms associated with this jet bring wet, rainy winters to the coasts of Oregon, Washington State, and British Columbia and snow to the higher terrain
  • During glacial times, the jet stream splits into a northern branch located along the northern flank of the ice sheet and a southern branch over the American Southwest.
• Ice sheets did not literally poke high enough into the atmosphere to block the flow of the jet stream.
  • The ice domes = 2 to 3 km, whereas winter jet streams flow = 10 to 15 km. But the ice blocked the lower-level atmospheric flow, and the disruption was propagated higher into the atmosphere.
  • This disruption, along with the tendency of jet streams to align their axis of flow along regions of strong temperature gradients at Earth’s surface, caused the split jet.
  • Model simulations using a high ice sheet split the jet to a much greater degree than do simulations based on lower-elevation ice
• A similar clockwise spiral of winds over the Scandinavian ice sheet brought cold, dry air southward into Europe. In addition, a strong upper-level jet stream crossed the Atlantic Ocean along latitudes between 45 and 50N and entered Europe south of the ice sheet.

13-8 Climate Changes Near the Northern Ice Sheets

• The most dramatic changes in climate at the glacial maximum were those in regions closest to and most directly influenced by the ice sheets.
  • Most of the climate changes simulated by the models are consistent with independent geologic evidence.
• The CLIMAP reconstruction of sea-surface temperatures based on the shells of planktic organisms shows the largest cooling occurring in the North Atlantic Ocean.
  • Frigid water and sea ice reached much farther south than today.
• The warm waters of the Gulf Stream and North Atlantic Drift flowed eastward toward Portugal instead of penetrating northeastward toward Scandinavia.
• The flow of cold winds off the North American ice sheet was an important cause of this glacial cooling of the North Atlantic Ocean.
  • Simulations using climate models that allow the ocean surface to react to the cold winds predict changes in sea-surface temperature similar to those estimated by CLIMAP.
  • In summer, the sea ice retreated to the north and the water warmed somewhat, but remained well below modern temperatures.
  • Later studies have indicated a larger summer retreat of sea ice and warmer temperatures than those in the CLIMAP reconstruction.
• Other large changes accompanied the North Atlantic cooling. A broad band of ice-rafted debris deposited in deep-ocean sediments near 50N latitude shows that icebergs broke off from continental ice sheets and drifted southward until encountering warm water and melting

13-9 Climate Change Far from the Northern Ice Sheets

• Farther from the northern hemisphere ice sheets, climate changes were less dramatic.
  • lower glacial concentrations of CO2 and CH4 were probably the major cause of the changes in climate
• Large changes occurred in the Antarctic, where CLIMAP estimated that the winter limit of sea ice expanded northward by several degrees of latitude in the far-southern Atlantic and Indian oceans. Later reconstructions have reduced these limits, but only slightly.
  • Associated with this shift in sea ice was a northward displacement of the region of strongest upwelling and highest surface-water productivity, but productivity decreased in regions nearer Antarctica where the cover of sea ice persisted longer during summer
• The record of climate change on arid southern hemisphere continents remains sparse. Expanded desert dunes in Australia suggest an even more arid climate and an intensification of the modern counterclockwise wind flow on that continent.
  
  • Caused by withdrawal of the ocean from a vast area just to its north and lower concentrations of greenhouse gases in the atmosphere.
• Climate in much of South America is heavily influenced by winds from nearby oceans. Most moisture from the Atlantic Ocean to the east is dropped in the Amazon rain forest and along the eastern flanks of the Andes.
• Along the Andes, where most lake-sediment records from South America exist, pollen data generally indicate drier conditions at the glacial maximum.
• This drying is probably the combined result of reduced extraction of water vapor from the cooler oceans, the lowering of sea level by 110–125 m, and the cooler temperatures resulting from lower CO2 and CH4 levels in the atmosphere.
• Pollen data from far southern latitudes indicate glacial climates wetter than those today west of the Andes, but drier to the east. Climate model simulations appear to account for these changes by showing a southward shift of the axis of westerly winds and moisture-bearing storms
• Because most of the tropics were more arid at the last glacial maximum, rain forest vegetation in both South America and Africa was probably less extensive than it is today.

13-10 Evidence for a Small Tropical Cooling

• The evidence for a small tropical cooling in the CLIMAP reconstruction was based on the small changes in planktic fauna and flora in low-latitude oceans.
• CLIMAP’s technique for reconstructing sea surface temperatures used the assumption that the distribution of species and assemblages of plankton is mainly determined by the temperature of the water in which they live.
• At higher northern latitudes during the glacial maximum, cold-adapted species moved into areas where warm-adapted species prevail today, indicating a large cooling in these regions.
• Across most low-latitude regions, however, the species that existed at the glacial maximum were not much different from the warm-adapted forms found there today.
• This lack of change in tropical plankton was the main reason CLIMAP concluded that ocean temperatures in the tropics cooled by an average of only 1.5C at the glacial maximum.

13-11 Evidence for a Large Tropical Cooling

• The most compelling evidence is the descent of the lower limit of mountain glaciers by 600 to 1,000 m throughout the tropics and middle latitudes.
  • This drop in the elevation of the ice line has been interpreted as requiring a cooling of 4–6C over tropical mountains.
• The lower limit of mountain glaciers today is determined mainly by temperature and secondarily by factors such as the amount of precipitation and the degree to which local mountain topography shelters the glaciers from direct sunlight.
• Glaciers exist today on tropical mountains higher than 5 km because the atmosphere cools by 6.5C or more per kilometer of elevation, resulting in subfreezing temperatures at higher elevations.
• Based on this relationship, the observed lowering of tropical mountain glaciers by 600 to 1,000 m during the glacial maximum would require a cooling of 4–6C.
• Additional evidence for larger glacial cooling comes from the descent of the upper tree limit and other specific kinds of vegetation high on tropical mountains.

13-12 The Actual Cooling was Medium-Small

• the tropical cooling was neither as small as CLIMAP claimed, nor as large as the critics initially thought, but about midway between the two estimates.
  • One reason the CLIMAP estimates were too small is that plankton are less sensitive to changes in temperature at low latitudes than to changes in the availability of food.
  • In addition, the CLIMAP reconstruction for the Pacific Ocean was based on samples in which CaCO3 had been extensively dissolved on the seafloor, thereby altering the assemblages of foraminifera and coccoliths.
• Evidence obtained from the biochemical composition of plankton shells has been used to test the CLIMAP estimates. One technique is based on the relative abundance of complex organic molecules called alkenones that constitute small fractions of tiny plant plankton
  • The past abundances of these molecules can be measured in small CaCO3 plates (coccoliths) deposited in ocean sediments.
  • The relative amounts of two types of alkenone molecules are sensitive to temperature in the modern ocean and can be used to reconstruct past temperatures
• In a north-south transect of cores across the western Indian Ocean, the cooling indicated by the two methods is generally less than 2C, confirming the small CLIMAP cooling for most of this region. But at latitudes north of 15N, the alkenone method indicates a cooling of 2C or more, while many CLIMAP estimates indicate a warming.
  • Temperatures estimates based on Mg/Ca ratios measured in surface-water foraminifera have also been made for several regions, with a similar result: some areas are cooler than the CLIMAP estimates, and other areas agree.
• The large-cooling view also has its own problems. The drier glacial climate in most of the tropics would have steepened the lapse rate from its present 6.5C/km toward the 9.8C/km rate typical of very dry air.
• A steeper lapse rate in the drier glacial tropics could account for part of the discrepancy between the ocean and land evidence. The evidence from mountain glaciers is poorly dated.
• Only a handful of regions have 14C dates that closely constrain the glacial lowering of 600–1,000 m to the exact time of the maximum glaciation. Some glacial moraines initially thought to date from the glacial maximum have turned out to be 30,000 years old or earlier, a time of cooler but also wetter climates. Lower glacier limits at such times could have been caused at least in part by greater snowfall.

Chapter 14: Climate During and Since the Last Deglaciation

14-1 When did the Ice Sheets Melt

• The Milankovitch theory predicts that the orbital maximum in summer insolation near 10,000 years ago in the Northern Hemisphere should have caused high rates of ice melting.
  • At first, it might seem that the way to quantify the rate of deglacial ice melting would be to measure the gradual retreat of the ice sheet margins.
  • Radiocarbon dating of organic material found in, under, or atop hundreds of ice-deposited moraines shows that the retreat of the large ice sheet in North America began near 15,000 14C years ago, reached a midpoint near 10,000 14C years ago, and ended by 6,000 14C years ago
  • The timing of these retreats agrees with the Milankovitch theory.
• Knowing the area covered by the retreating ice is a good start, but a complete analysis requires that measurements be converted to ice volume.
  • To make this conversion, we need to know the thickness of the ice as it retreated.
• To complicate this calculation, the thickness of an ice sheet can be affected by the conditions in its basal layer.
• Portions of ice sheets that repeatedly slide on their basal layers are thin and relatively low in volume; areas frozen to their beds are thicker and larger in volume. Because of this uncertainty about thickness, records of changing ice area through time do not guarantee valid records of ice volume.

14-2 Coral Reefs and Rising Sea Level

• The best record of ice sheet melting comes from tropical coral reefs far from the polar ice sheets
• Because several coral species grow just below sea level, the current elevation of older reefs built by corals can be used as a measure of past sea level
• Coral reef measurements of lower sea level during the last glacial maximum and the subsequent deglaciation can be converted to a record of global ice volume, with each 1-meter rise of sea level equivalent to about 0.4 million km3 of ice.
• In the late 1980s, the marine geochemist Richard Fairbanks drilled and 14C-dated a series of now submerged coral reefs off Barbados, a Caribbean island.
  • These reefs yielded a history of sea level rise from its low extreme at the last glacial maximum to its higher position during the modern interglaciation.
  • The 14C-dated deglacial sea level curve in Barbados supports the Milankovitch theory in a general way: the middle of the deglaciation occurred near the insolation maximum 10,000 years ago, as expected. But the story is not that simple because the 14C dates on the corals do not represent their true ages.
• The Th/U shifts the timing of the middle part of the deglaciation back in time by about 2k years, and the earlier parts of the deglaciation by as much as 3.5k years.
• In contrast, the timing of the summer insolation signal remains fixed by the independent astronomical time scale. As a result, the major rise in sea level shifts back somewhat earlier in time compared to the insolation curve
• Does this earlier timing for the deglaciation invalidate the Milankovitch theory? Not really, Milankovitch chose summer as the critical season of insolation control of ice sheets, and the last deglaciation still occurred during a time when summer insolation was higher than it is now, although somewhat earlier in that interval than the Milankovitch theory predicts.

14-3 Rapid Early Deglaciation

• One important feature of the coral reef record is the fact that sea level did not rise smoothly throughout the deglaciation.
• Moderate sea level rose between 19,000 and 14,500 year, rose quickly between 14,500 and 13,000 years ago, after which the rate was considerably slower.
• One way to monitor the melting of individual ice sheets is to look for local pulses of meltwater delivery to the oceans. Because the δ18O values of the northern ice sheets were -30‰ to -35‰, whereas those in the surface ocean were near 0‰, major influxes of meltwater should be registered as distinct pulses of lower δ18O values in the shells of planktic organisms living in the ocean.
• Planktic foraminifera in the northeastern Norwegian Sea record a pulse of unusually negative δ18O values early in the deglaciation. 
• This evidence from marine sediments suggests that the moderately large marine ice sheet, with its base lying below sea level in the Barents Sea, was vulnerable to early collapse when summer insolation began to rise.
• A similar low-δ18O pulse in cores from the Gulf of Mexico indicates a short-term increase in meltwater flow down the Mississippi River from the southern margin of the North American ice sheet.
• North Atlantic sediment cores contain a distinctive layer of sediment deposited 17,000 to 14,500 years ago rich in ice-rafted sand grains but nearly barren of the planktic foraminifera and coccoliths normally found in that region.
• This layer is evidence of a large influx of icebergs to the North Atlantic Ocean early in the deglaciation. This evidence suggests that the major continental ice sheets lost a significant amount of their mass early in the deglaciation by calving icebergs to the ocean.

14-4 Mid-Deglacial Cooling: The Younger Dryas

• The slowing in the rate of deglacial ice melting after 13,000 years ago coincided with the onset of a cold climatic oscillation that is especially evident in records near the subpolar North Atlantic Ocean.
• Temperatures in this region had warmed part of the way toward interglacial levels, but this reversal brought back almost full glacial cold. The first evidence for this event came from pollen records in Europe, where a distinctive flowering plant Dryas moved south from the Arctic into Europe. This cold episode is called the Younger Dryas.
• Later work on sediments in the North Atlantic Ocean also detected a clear Younger Dryas imprint: a rapid oscillation in the regional extent of icy polar water.
• At the glacial maximum, polar water reached southward across the North Atlantic to 45N. The southern margin of this cold water was defined by the polar front, a zone of rapid transition to the more temperate waters to the south.
• Early in the deglaciation, near 15,000 years ago, the polar front had rotated back to the northwest around a hinge point near Newfoundland in eastern Canada. During this change, warm water had begun to flow northward along the European coast and to moderate climate enough that temperate species of trees had begun to advance northward from their full-glacial positions in far-southern Europe.
• Near 12,900 years ago the polar front abruptly advanced back to the south, almost reaching its full-glacial position. At the same time, with the North Atlantic Ocean again cooler, Arctic vegetation (including Dryas) returned to northern Europe. Later, near 11,700 years ago, the polar front abruptly retreated to the north, and forests began their final advance into north-central Europe.
• The Younger Dryas re-advance of the polar front represents a major reversal in Atlantic circulation.
• The estimated sea-surface cooling west of Ireland was at least 7C, close to the difference between fully glacial and typically interglacial extremes. A similar cooling has been estimated from changes in the fossil remains of temperature-sensitive insect populations in England.
• Ice cores from Greenland contain a remarkably detailed record of the Younger Dryas event.
• The geochemist Wally Broecker proposed changes in the path of meltwater flow from the North American ice sheet as the cause of the the Younger Dryas oscillation.
  • He suggested that the major meltwater route southward into the Gulf of Mexico was abruptly diverted eastward into the North Atlantic Ocean during the Younger Dryas. This diversion would have delivered a pulse of low salinity water that altered the circulation of the North
• Atlantic Ocean by lowering the density of its surface waters enough to prevent them from sinking and forming deep water.
  • Because ocean surface waters give off heat when deep water forms in the North Atlantic Ocean, cutting off this process could have cooled climate in the North Atlantic and in Europe.
• Subsequent investigations have raised questions about this hypothesis.
  • Examination of drainage patterns from large meltwater lakes near the southern ice sheet margin have failed to find convincing evidence of an unusual outflow of freshwater eastward to the Atlantic during the onset of the Younger Dryas event.
  • More recent studies, however, have found evidence for a strong outflow of meltwater from the northwestern Canadian ice sheet margin down the McKenzie River and into the Arctic Ocean at about 13,000 years ago. As a result, the possibility of a meltwater-outflow explanation of the Younger Dryas remains alive, although via a different pathway.
• Earlier claims that the Younger Dryas cooling was global in extent have proven incorrect. A small cooling evident in Antarctic ice cores that was once interpreted as correlative with the Younger Dryas event is not.
• Instead, at least part of the Antarctic was undergoing a slow warming throughout the Younger Dryas interval. The origin of the Younger Dryas cooling remains an enigma.

14-5 Positive Feedbacks to Deglacial Melting

• Most agree that rising summer insolation values due to Earth’s orbital tilt and precession initially set in motion the melting of the great northern hemisphere ice sheets.
• How did so small an insolation maximum melt so much ice so quickly? The first answer to this question is that positive feedbacks accelerated the loss of ice.
• Two widely acknowledged feedbacks would have been in action throughout the deglaciation, one of which was the rising concentration of CO2 in the atmosphere.
• As the ice sheets melted, the CO2 level rose, warming the planet through the greenhouse effect and adding to the initial effect of increasing summer insolation.
  • The CO2 increase led the sea level rise by an average of 1,000 years, too small an amount for CO2 to have been an independent driver of the ice sheets, but consistent with the idea that CO2 acted as positive feedback to ice melting.
• The change in surface albedo as the ice sheets melted would have been a second major source of feedback.
  • Large ice sheets with very high albedos reflected much of the incoming solar radiation back out to space, but the darker surfaces exposed as the ice melted would have absorbed most of the incoming radiation, except when snow-covered in winter.
• One noteworthy feature of the most recent deglaciation was the rapid early rate of sea level rise compared to the gradual increase in the summer insolation driver.
  • Part of this offset in timing was presumably due to the fact that a much larger amount of ice was available to melt earlier in the deglaciation than later.
• The rapid early rise in sea level between 19,000 and 14,000 also suggests that important positive feedbacks may have been in action during this time of rapid ice melting.
• The negative δ18O pulse in Norwegian Sea cores indicates that a substantial part of the marine ice sheet over the Barents Sea melted early in the deglaciation. Because its base lay below sea level, it was probably especially vulnerable to early destabilization.
• Many observations were found across the globe

14-6 Deglacial Lakes, Floods, and Sea Level Rise

• As the ice sheets melted back, the land in front of them remained depressed for some time, rather than immediately rebounding to its former level.
  • Into these depressions poured meltwater from the retreating ice sheets, forming proglacial lakes. Because of the large volumes of meltwater arriving each summer, the lakes frequently cut new channels and overflowed into other lakes and then into rivers that carried water to the ocean.
• The deglacial rise in sea level altered Earth’s surface on a very large scale.
  • Many regions of the world’s continental shelves had been exposed during the low sea level at the glacial maximum, and many continents or ocean islands had been linked by land connections.
• The lower level of the glacial ocean had also transformed smaller seas around the margins of the oceans, especially in the western Pacific.
• The present Yellow Sea was dry land, and other seas in the western Pacific were more isolated from the open ocean because sea level was lower. Rising sea level flooded these seas and rejoined them to the open ocean.
• Other Climate Changes During and After Deglaciation

• • Scientists have investigated two important changes during the late-deglacial and postglacial interval—the strength of north tropical monsoons and the warmth of summers in north polar latitudes.

14-7 Stronger, Then Weaker Monsoons

• Monsoons were strong near 10,000 years ago because of Earth’s orbital configuration. Summer insolation values over tropical and subtropical landmasses of the Northern Hemisphere at that time were 8% higher than they are today.
• According to Kutzbach’s orbital monsoon theory, increased insolation should have driven a stronger summer monsoon circulation and produced higher tropical lake levels near 10,000 years ago compared to those today.
• The COHMAP project compared climate model simulations against ground-truth geological data during this interval. Their simulations used summer insolation values from 9,000 years ago, when northern hemisphere ice sheets were greatly reduced in size. The model simulated stronger monsoons across the entire north-tropical region of North Africa, southern Arabia, and southern Asia. Geologic evidence matches this simulation.
• Although climate models and evidence from the geologic record confirm Kutzbach’s theory of stronger north-tropical monsoons 10,000 years ago, a closer look reveals a mismatch in amplitude.
  • The increase in rainfall in the models was small compared to the amounts indicated by geologic evidence from lake levels and pollen assemblages. Subsequent modeling efforts have explored positive feedback processes that could have amplified the small response simulated by the early models
• One important feedback is the increase in recycling of water vapor provided by evapotranspiration from vegetation. The initial increase in monsoonal rains in the early models was sufficient to allow trees to advance northward into grasslands, and grasses to move northward into deserts.
  • Recent models fixed this issue but one feedback not yet incorporated is increased recycling of moisture from small lakes and low swampy regions near rivers that are too small to be represented in modern climate model grid boxes.
• After reaching a peak near 10,000 years ago, summer insolation values at lower latitudes of the Northern Hemisphere have fallen continuously.
• By 6,000 years ago, summer insolation values in the northern tropics were still about 5% higher than the modern levels but were falling toward modern values.
• This slow decrease should have produced a corresponding decline in the strength of the tropical monsoons. Direct observations and 14C dates of lakes across North Africa confirm a major drop in water levels during the last 9,000 years. Today, lakes are lower than they were 9,000 years ago, and many have completely dried out.
• The short-term changes in lake levels represent a type of climate response that cannot be directly attributed to the smooth, gradual forcing provided by changes in summer insolation.

14-8 Warmer, Then Cooler North Polar Summers

• At the glacial maximum, the main controls on climate at high northern latitudes had been the regional cooling effects of the ice sheets and the global cooling caused by low CO2 and CH4 concentrations. As deglaciation proceeded, rising summer insolation values increasingly warmed land areas located far from the ice sheets and in time overcame the cooling effects of the shrinking ice sheets.
• The last deglaciation dramatically transformed the vegetation on the northern hemisphere continents.
  • In North America, cold-tolerant spruce trees retreated to northeastern Canada and warm-tolerant trees such as oak moved to mid-Atlantic states.
• The climate changes that occurred midway through the deglaciation produced unusual mixtures of plants called no-analog vegetation because no similar combination exists today.
• No-analog mixtures developed because each vegetation type responded to a different combination of environmental variables.
• The distributions of spruce and oak pollen simulated by climate models are compared with the observed distributions. Both sets of maps show the same large-scale northward relocation of spruce, and they agree on the existence of a mid-deglacial interval when spruce became rare throughout eastern North America.
• Both sets of maps also show a similar northward expansion of oak, but the model simulates more oak in the southeast during deglaciation than the pollen data show, a mismatch similar to that noted for the glacial maximum.
• Once CO2 values had risen to full interglacial levels, summer insolation values became the main variable controlling climate responses, particularly for vegetation.
• Some climate scientists refer to this warmer-than modern interval as the hypsithermal, but the time of greatest warmth actually varies widely from region to region, depending on when the nearby ice melted and its regional cooling effect was removed.
• Because far-northern regions are snow-covered for much of the year, this additional warming extended through most of the year and affected broad areas of northern Canada and Asia. The net effect of this vegetation albedo feedback almost doubled the initial insolation warming of high northern latitudes.
• Sea ice also contributed to these far-northern climate changes. High summer insolation caused the sea ice margin in the model to thin and retreat northward, and this change propagated into the rest of the yearly cycle, with delayed refreezing of seasonal sea ice in autumn, thinner and less extensive sea ice in winter, and earlier melting of sea ice in spring.
• During the last 6,000 years, Earth’s tilt has continued to slowly decrease and its precessional motion has moved the northern hemisphere summer solstice toward the aphelion position.
  • These combined orbital changes have produced a 5% decrease in summer insolation and a 5% increase in winter insolation at high latitudes for 6,000 years ago
  • As a result, summer temperatures have fallen significantly during the last several thousand years in several regions at high northern latitudes.
• Evidence of cooler summers comes from ice cores taken from small ice caps in several parts of the Arctic.
• Ice from the tiny Agassiz Ice Cap on Ellesmere Island, in far northern Canada, shows that summer melting episodes were far more frequent before 5,000 years ago than they have been since that time. This evidence supports a trend toward cooler summers.
• A lot of different evidence to show cooler North Polar summers 

Chapter 15 Millennial Oscillations of Climate

15-1 Oscillations Recorded in Greenland Ice Cores

• Long ice cores taken on Greenland in the 1970s recovered records spanning much of the last interglacial glacial cycle.
• Two signals from this record attracted particular interest—the δ18O composition of the ice and the concentration of dust.
  • Signals of δ18O in ice cores record changes in the composition of the 16O-rich water vapor that falls as snow and consolidates into ice.
  • Chemical analysis of the dust in Greenland ice shows that the main source region was northern Asia.
•  Both records show two distinct patterns.
  • One trend is the slow underlying change from low dust concentrations and relatively positive (low-16O) δ18O values at the top, to higher dust concentrations and more negative (high-16O) δ18O values in the middle part, and a return to positive δ18O values and little dust in the bottom part of the record.
  • The imprecise dating available at the time these features were discovered indicated that the upper section spans the current interglaciation, the middle part represents the last glacial interval, and the bottom section is part of the previous interglaciation.
  • The current and previous interglacial intervals were warmer (with more positive δ18O values) and relatively free of dust compared to the cold, dusty 16O-enriched glacial interval in the middle.
•  But these slower orbital-scale changes are to some extent masked by a far more prominent pattern— the rapid oscillations over much shorter intervals between high and low dust concentrations and between negative and positive δ18O values.
• Temperatures over Greenland during glacial intervals oscillated rapidly between extremely cold intervals called “stadials” and relatively mild intervals called “interstadials.” The amplitude of the temperature variation ranged between 8 and 16C, with a tendency toward a sawtoothed shape: fast warming and slower cooling. These oscillations are often referred to as Dansgaard-Oeschger oscillations.
• Each fluctuation toward more negative (glacial) δ18O values is matched by an abrupt increase in dust concentrations in the ice. As in the case of the δ18O changes, the range of variation in dust concentrations is a large fraction of the total difference between the glacial and interglacial values.
• Geochemical analysis of the dust shows that most of it comes from distant source regions in Asia, not from nearby North America. The size of the dust particles is larger in the cold intervals than in the milder ones, indicating that strong winds lifted and transported the dust when climate was very cold. The colder intervals also contain larger amounts of sea salt plucked from salty sea spray above the turbulent ocean during cold and windy intervals and carried to the ice.

15-2 Oscillations Recorded in North Atlantic Sediments

• In the mid-1980s, studies of rapidly deposited sediments in the North Atlantic Ocean first detected shorter climate oscillations.
• The marine geologist Hartmut Heinrich found episodes of unusually abundant ice rafting separated by as little as 7,000 years to as much as 12,000 years. These episodes are called Heinrich events.
• Later, the geologist Gerard Bond discovered even shorter-term (2,000–3,000 year) variations in two climatic indices:
  • The abundance of the one polar species of foraminifera as a fraction the total population and the relative amounts of the shells of foraminifera compared to the sand-sized ice-rafted grains.
  • As was the case for orbital-scale studies, higher percentages of the polar species of foraminifera and larger concentrations of ice-rafted debris are an indication of the presence of colder North Atlantic waters carrying larger numbers of icebergs
• Radiocarbon dating of the North Atlantic sediments back to 30,000 years ago indicates a correlation in time with the oscillations dated in Greenland ice by counting annual ice layers. Times of cold air correlate with times of cold ocean temperatures in the nearby ocean.
• Both records show a tendency toward a similar saw-toothed pattern: repeated slow drifts toward colder, more glacial conditions, followed by an influx of ice-rafted debris, and then an abrupt shift back to warmer conditions.
• Although the major Heinrich events show the largest ice-rafted influxes, the smaller Dansgaard-Oeschger oscillations also have lesser ice-rafted pulses.
  • Prior to 30,000 years ago, both the marine and ice-core dating methods become too uncertain to allow firm time correlations, but the very strong resemblance of the two signals suggests that the same close relationship persists.
• Was the relative increases in the amount of ice-rafted debris compared to the foraminifera were caused by greater delivery of ice-rafted debris, by reduced rates of deposition of foraminifera, or both? Other questions concerned the source or sources of the ice-rafted debris. 
• Radiocarbon dating of the CaCO3 shells of foraminifera in several of the younger ice-rafting layers indicated tenfold or larger increases in the rate of deposition of ice-rafted debris, compared to much smaller decreases in the rate of deposition of foraminifera. The major factor in these variations turned out to be the increased influxes of ice-rafted debris.
• Initial investigations of limestone fragments during the largest ice-rafting (Heinrich) events showed that rocks in and north of Hudson Bay in Canada were the major source of this debris. Further evidence supporting this conclusion came from geochemical (isotopic) analysis of ice-rafted mineral grains that pointed to bedrock sources north and east of Hudson Bay.
• During the fluctuations, ice-rafted debris came from a range of source regions, two of which left particularly distinctive evidence.
• Fragments of volcanic glass originated mainly from eruptions on Iceland. In addition, iron-stained quartz grains came from several sources around the North Atlantic where outcrops of Pangaean-age sandstone contain quartz grains stained red by iron oxidation during ancient monsoonal climates.
• Long sequences of ocean sediments recovered by the Ocean Drilling Program show that millennial fluctuations also occurred during previous glaciations, both during the 41,000-year cycles prior to 0.9 million years ago, and the subsequent ice sheet fluctuations at ,100,000 years.
• Because the millennial oscillations are so well developed in North Atlantic surface waters, scientists searched for evidence that this signal had penetrated into deep water formed in this region.
• The method exploited was the same one used to measure similar changes at orbital scales: more negative δ13C values in the CaCO3 shells of bottom-dwelling (benthic) foraminifera mark times when deep water from North Atlantic sources was replaced by bottom water formed in the Southern Ocean.
• The δ13C evidence showed that deep-water formation in the North Atlantic was rapid during warm interglacial climates, slower and shallower during glacial climates, and especially slow during these ice-rafting episodes. Apparently, the changes at the surface were felt in the deeper North Atlantic Ocean

15-3 Detecting and Dating Other millennia Oscillations

• The verification of similar millennial oscillations in both Greenland ice and North Atlantic sediments set off an energetic search for similar fluctuations in other regions.
• Scientists who took part in this search faced two major questions at the outset:
  • Is the climatic archive that is being examined capable of recording such brief oscillations? How accurately can the oscillations be dated?
• Resolution of climate signals varies from archive to archive and from region to region.
  • The ideal archive is one that allows resolution of annual changes and provides a record stretching well back into the last glaciation or preferably beyond.
  • Few archives combine these characteristics, but many archives that reach far back in time can resolve climate changes at a resolution of tens to hundreds of years.
  • This resolution is sufficient for detecting the presence of millennial oscillations, even if inadequate to define their full amplitude.
• The second problem is determining how the oscillations detected in other archives correlate in time with those found in Greenland ice and North Atlantic sediments.
• The 14C method used to date most continental records has analytical uncertainties of several thousand years for most glacial-age material.
  • Because these dating errors grow comparable in size to the length of the oscillations prior to 30,000 years ago, it is more often than not impossible to determine how the observed oscillations actually correlate.
• It is much easier to show that millennial-scale oscillations are present or absent in a given climate record, whatever their exact ages and correlations to other records.
• In a few such cases, the millennial oscillations match the pattern of the changes in Greenland and the North Atlantic so closely that little doubt exists that we are looking at the same oscillations, even though the dating is not very accurate. But even in these cases, small leads or lags between the signals could exist.

15-4 Oscillations Elsewhere in the Northern Hemisphere

• Other regions of the Northern Hemisphere also show millennial-scale oscillations, including western Europe.
  • European soils were richer in organic material during warmer episodes but almost free of organic carbon during colder oscillations. The  pollen records that show orbital scale changes from interglacial forests to glacial tundra also reveal short fluctuations within glacial intervals from full tundra to mixed grass steppe and forest vegetation.
• Although the correlations of the European records with the Greenland δ18O fluctuations are not particularly obvious, 14C dating of the younger fluctuations indicates that cold-adapted vegetation occurred in Europe during times of colder air over Greenland and colder temperatures in the North Atlantic Ocean.
• Short-term fluctuations have also been discovered in the glacial sections of windblown loess deposits in China.
• Climate scientists have also found millennial oscillations in regions far from the North Atlantic Ocean.
• Other indications of millennial-scale oscillations in western North America come from fluctuations of glacial Lake Bonneville in Utah, where the younger 14C-dated lake level maxima appear to correlate with major ice-rafting events in the North Atlantic. (a lot of examples in book)
• Cave calcite deposits from the northern and southern tropics also show clear millennial-scale oscillations that are very accurately dated by the Th/U method.
  • At orbital time scales, the δ18O variations in these deposits show out-of-phase responses between the hemispheres at the 23,000-year period of precession.
  • This behavior reflects out-of-phase summer insolation forcing of low-latitude monsoons in the two hemispheres, resulting in migrations of the intertropical convergence zone (ITCZ) across tropical regions.
• At the millennial scale, southern hemisphere δ18O oscillations recorded in Brazilian speleothems are somewhat reduced in amplitude compared to those in the northern tropics and are out of phase with those in the north. Brazil is warm and wet when the north is cold and dry. This pattern appears to be the result of short-term ITCZ migrations from more northerly positions to more southerly ones. Southward movement of the ITCZ leaves the northern tropics and subtropics drier and cooler, but makes the southern tropics and subtropics warmer and wetter.

15-5 Oscillations in Antarctica

• The presence of millennial oscillations with similar timing across much of the Northern Hemisphere raised the obvious question whether or not they were global in extent, and ice cores from Antarctica will tell us.
  • Antarctic ice does show short-term δ18O (and other) oscillations, but they are much smaller in size than those in Greenland ice and more symmetrical in shape compared to the saw toothed oscillations in the north.
  • Unfortunately, Antarctic ice lacks the annual layering needed for sufficiently accurate dating to determine the relative phasing with the changes in the Northern Hemisphere, but both Antarctic and Greenland ice contain a common signal that can be used to correlate the two records very closely in a relative sense: millennial-scale changes in atmospheric methane concentrations.
• With the two trends synchronized based on the methane variations, the temperature oscillations over Antarctica turn out to be roughly opposite in timing to those in the north. In general, slow warming trends in Antarctica were underway when Greenland ice recorded the coldest northern temperatures, and the fastest rates of warming in Greenland occurred when Antarctica had already reached maximum warmth and was beginning to cool. The two regions seem to be nearly, but not precisely, out of phase
• This offset in timing is confirmed by an ocean core from near Portugal, where the δ18O signal recorded in planktic foraminifera follows the pattern of North Atlantic oscillations, but the δ18O changes in benthic foraminifera more closely resemble the oscillations found in Antarctic ice.
• The fact that both signals are present in the same core allows scientists to compare their relative timing, even without high-resolution radiocarbon dating. Again, the northern and southern signals are offset in timing.
• Antarctic ice cores also show millennial-scale changes in atmospheric CO2 concentration of as much as 20–25 ppm.
  • The CO2 signal is out of phase with the Dansgaard-Oeschger oscillations in the Northern Hemisphere, and roughly in phase with the temperature changes in Antarctica. The small CO2 lag relative to that of Antarctic temperature may result from errors in correcting for differences between the ages of the gases (CO2) and the solid ice.

15-6 Are the Oscillation Periodic?

• Before looking at specific causes, a key issue worth considering is whether or not the observed millennial variations are cyclic.
• The Dansgaard-Oeschger oscillations have been often referred to as “Dansgaard-Oeschger cycles,” and the longer-term variations in polar foraminifera and ice-rafted debris are sometimes called “Bond cycles,” after the geologist Gerard Bond who first detected them
• The case for cyclic behavior is very weak because the difference in time between successive peaks or valleys in these oscillations ranges from as little as 1,000 years to as much as 9,000 years. No strong evidence of a preferred timing within that range is evident.
• These oscillations have the characteristic form of red noise.
  • The word noise conveys the idea that the fluctuations are random and unpredictable, rather than cyclic and predictable. The term red refers to a characteristic behavior in which the longer duration oscillations are generally larger in size than the shorter-term oscillations.
• Some scientists claim to have found a 1,500-year cycle in ice sheet and ocean records, but in general the evidence is sparse.
  • Time-series analysis of the Greenland ice core δ18O record indicates that only a very small fraction of the observed variations falls within a band centered near 1,500 years. Most of the power at or near 1,500 years appears to come from three δ18O variations that occurred from 30,000 to 35,000 years ago, but the 1,500-year period is absent or weak in the rest of the record.
• The glacial geologist Richard Alley has proposed that millennial oscillations could fall somewhere between the extremes of cyclic behavior and red noise.
  • His term for this intermediate behavior, stochastic resonance, is a composite of two concepts that seem in conflict with each other because the word stochastic means random, whereas resonance implies cyclicity.
• In his view, resonance is evident in the fact that oscillations at a cycle near 1,500 years do appear at some times in some records. At other times, however, the climate system skips past individual oscillations at 1,500 years because of interference from the effects of random noise.
  • When this happens, northern hemisphere climate continues to drift slowly toward colder conditions until it eventually responds to one of the subsequent multiples of the 1,500-year cycle and shifts abruptly back to warmer climates.
  • Because these abrupt transitions occur over a wide range of intervals from 1,500 to 9,000 years, they appear to be random, but could actually be an expression of an irregular cyclic behavior. In any case, noise plays an important role in this concept.

15-7 Are the Oscillations Forced by the Sun?

• Most of the energy that Earth receives from the Sun arrives at visible or near-visible wavelengths. Unfortunately, direct records of past changes in this incoming solar radiation do not extend back into the last glaciation.
• Instead, several proxies provide scientists with other potential measures of solar fluctuations. The difference between the ages derived by counting tree rings and those derived by 14C dating of the same rings are one such index.
• Shorter-term changes in age offsets are also apparent within the last 10,000 years. These discrepancies may also reflect changes in the rate of production of 14C atoms in Earth’s atmosphere, although in this case the main cause is thought to be changes in emissions from the Sun, rather than the overprint of Earth’s magnetic shielding.
  • Particles streaming from the Sun deflect some of the incoming cosmic rays (protons) that would otherwise enter Earth’s atmosphere. Changes in the amount of solar deflection over hundreds of years could alter the 14C production rate in the atmosphere and explain the short-term differences in ages derived from the two dating methods..
• One reconstruction of the timing of mountain glacier advances over the last 7,000 years does show a hint of a response near 2,100 years, but the other reconstruction does not. A hint of changes at intervals near 2,600 years occurs in the sea salt signal in Greenland ice, but no such response occurs in the ice-rafting signals in the nearby ocean.
  • This evidence refutes the existence of a cyclic millennial scale link between the Sun and Earth’s climate, although it does not eliminate the possibility of a noncyclic link
• Another proposed solar proxy is the 10Be isotope, which is produced by collisions with cosmic particles.
  • Ice-core records show large millennial-scale changes in 10Be concentration during the interval from 40,000 to 25,000 years ago, with an obvious correlation between 10Be maxima and δ18O minima that are indicative of colder climates. The variations in 10Be concentration are actually an artifact of climatic changes. During cold oscillations, accumulation rates on ice sheets fall. With slower ice accumulation, the concentration of 10Be in the ice increases even if its rate of production in the atmosphere remains constant. Similarly, intervals of lower 10Be concentrations during warm climates (δ18O maxima) reflect dilution of the 10Be signal by more rapidly accumulating ice. As a result, these older 10Be fluctuations cannot be used to infer a connection to solar deflection of cosmic rays.
• Changes in 10Be concentration in ice cores have also occurred over the last 5,000 years, and they have been compared against changes in 14C production determined from the difference between tree ring counts and 14C analyses.
• In making this comparison, allowance has to be made for differences in how the two signals are registered by the climate system.
  • In effect, the 10Be signal must be smoothed in such a way as to mimic the more complicated processes that alter the 14C record.
  • The very good match of the two signals suggests that both the 10Be and the 14C signals may have varied in response to changes in solar deflection of the incoming cosmic rays.
  • But another possible explanation is that both oscillations are a response to changes in the internal operation of the climate system such as the circulation of the ocean..
• The evidence argues against a strong cyclical effect of solar variability on climate. A link between random variations in Sun strength and climate at shorter time scales remains a possibility.

15-8 Are the Oscillations Caused by Natural Ice Sheet Instabilities?

• Because millennial fluctuations only occur when ice sheets are present in the Northern Hemisphere, some kind of causal link to the ice sheets must exist.
• Two mechanisms that favor an independent initiator role for ice sheets focus on the massive episodic releases of icebergs to the North Atlantic Ocean.
• Along the marine margins of ice sheets, ice flows over bedrock with irregular bumps and depressions. The bottom layers of ice scrape against higher-standing areas called bedrock pinning points, and the resulting friction slows the flow of ice. The bottom layers of ice can also freeze to the bedrock and slow the flow even more.
• Ocean water can produce the opposite effect: because the ice margins float in seawater, changes in sea level can lift the ice off its pinning points.
• One idea is that the very slow natural release of small amounts of heat from Earth’s interior can melt the lower ice layers along ice margins.
• A second idea, proposed by the glacial geologist Doug MacAyeal, focuses on a different kind of interaction between ice margins and bedrock.
• Both of these hypotheses are consistent with the timing of the observed ice-rafting episodes in the North Atlantic Ocean compared to the prevailing regional climate. The largest episodes occurred when the air and ocean were very cold, rather than during the warmer intervals that would seem more likely to cause fast melting and ice margin collapse. This evidence argues against local warming as the driver of the iceberg pulses but allows for mechanisms based on internal ice sheet instabilities
• The composition of the debris deposited by icebergs in the North Atlantic shows that it came from many distinct source regions. But why would so many ice sheet margins be simultaneously releasing icebergs?
  • One possible link is sea level. If one ice margin surges and sends icebergs into the ocean, the melting icebergs raise sea level. A rise in sea level could then destabilize ice shelves on other coastal margins off their bedrock pinning points and cause them to discharge some of their ice into the ocean.
• This explanation raises the issue of whether or not the millennial-scale rises in sea level were large enough to link the ice sheets.
• Variations in sea level based on coral reefs along the slowly uplifting coast of New Guinea suggest that sea level changes could have been as large as 10–15 meters during major millennial oscillations. In addition, marine δ18O signals recorded in the shells of benthic foraminifera from the deep tropical Pacific Ocean show millennial-scale variations of 0.1‰ or slightly larger during these intervals. The coral reef and δ18O evidence both point to sea level changes of 10 meters or more during the largest millennial oscillations.
• The massive North American ice sheet is the best candidate for causing sea level rises large enough to trigger reactions in the other ice sheets during major ice-rafting events.
• Changes in size of the northern ice sheets during the smaller and shorter millennial oscillations probably produced sea level fluctuations of no more than a few meters. It is more difficult to see how sea level changes this small could have provided a link among the northern ice sheet margins. Yet several ice margins contributed ice-rafted debris during the smaller oscillations as if they were linked.
• Another possibility is that the ice streams delivering icebergs to the Atlantic Ocean pulled enough ice out of the interior of the North American ice sheet to alter atmospheric circulation. Over orbital time scales, climate responds to splitting of the jet stream because of changes in the elevation of the North American ice sheet. Some climate scientists have suggested that this explanation might also apply to shorter millennial-scale circulation shifts in response to ice volume changes of 1% to 10%.
  • For this explanation to be viable, atmospheric circulation would have to have been extremely sensitive to small changes in the elevation of the North American ice sheet, probably requiring the existence of a critical threshold height.
• If northern ice sheets are the initiator of millennial oscillations, then the question is: How were millennial scale oscillations transmitted to other regions? The evidence summarized to date suggests an answer.
  • Discharges of icebergs and meltwater into the North Atlantic would lower its salinity and enhance formation of sea ice, which would have acted as a very powerful positive climate feedback and cooled high northern latitudes. From this northern “center of action,” the transmission of the oscillations through the rest of the climate system may then have occurred by means of rapid-acting atmospheric processes.
  • In southern China and in India, well-dated δ18O trends in cave deposits show millennial intervals of weakened north-tropical monsoons and cooler/drier climate that correlate closely with changes in the North Atlantic. This close timing and rapid response must have happened through an atmospheric link. As part of this response, the northern margin of the intertropical convergence zone (ITCZ) was pushed south.
• Southward ITCZ displacements were also recorded in the Southern Hemisphere, but in this case with climatic effects that were the reverse of those in the north.
• Well-dated speleothem δ18O trends in Brazil show rapid millennial responses, but toward warmer and wetter climates. This opposing response makes sense if we look at the large-scale pattern of dislocation of the ITCZ.
• The same southward ITCZ shifts that weakened the north-tropical monsoons and produced drier and cooler climates also displaced the southern ITCZ margin to the south and produced warmer and wetter climates along its southern margin in the southern hemisphere tropics and subtropics.
• An important remaining question is whether the millennial-scale warming episodes registered in the southern ITCZ margin in Brazil are responsible for the warm intervals in the Antarctic. The Antarctic warm intervals are fewer in number and smoother in form than the oscillations in the north, but this may result from the fact that the ocean-dominated climate system around Antarctica smoothes out the signals arriving from the north and mutes the weaker oscillations.
• At this point, the evidence at hand seems consistent with a scenario in which signals originating in the North Atlantic region reached all the way south to Antarctica, thereby producing a global response.

15-9 Are the Oscillations Caused by Interhemispheric Climate Instabilities?

• An alternative possibility is that the millennial oscillations were the product of slower-acting interhemispheric exchanges of heat.
  • In this view, changes in the amount of northward redistribution of heat by the Atlantic Ocean produce a response pattern called the bipolar seesaw.
• In most oceans today, the typical pattern of heat transport removes excess heat from the warm tropics and carries it toward the cold poles. The Indian and Pacific Oceans both follow this pattern, but the Atlantic Ocean does not. Instead, heat from the South Atlantic Ocean crosses the equator and moves into the high latitudes of the North Atlantic Ocean.
• The marine geologist Tom Crowley first proposed that changes in the northward transport of heat in the surface Atlantic Ocean could explain the bipolar seesaw pattern. Greater cross-equatorial transport of heat would warm the North Atlantic Ocean but leave the Southern Ocean cold. Weaker transport would cool the North Atlantic Ocean but leave more heat in the Southern Ocean
• The geochemist Wally Broecker proposed that changes in the rate of freshwater and/or iceberg fluxes from northern ice sheets could control the amount of deep water formed in the North Atlantic Ocean.
  • During times of large influxes, the low salinity produced in the surface Atlantic layers could have reduced the formation of deep water and the amount of ocean heat “pulled” northward into the Atlantic, leaving the Southern Ocean warmer.
  • Although many experiments with simplified ocean models have supported the idea that freshwater discharges can suppress deep-water formation and alter surface heat patterns, the models that incorporate the most complete array of processes tend to indicate weaker effects.
  • Physical oceanographers have also criticized this explanation by noting that the large-scale circulation of the surface ocean is largely driven by winds, not by rates of deep-water formation.
• Another problem with the hemispheric seesaw hypothesis is the lack of an explanation for the sudden onset of ice-rafted events in the North Atlantic Ocean. Shallow waters cannot be responsible for melting the ice margins because they were already very cold before and during these episodes. As noted in Section 15-5, the deep-water circulation in the Atlantic does have a southern (Antarctic) timing, but these changes occur at depths too far removed from the surface to have played any role in melting the ice sheet margins. The possibility that heat delivered northward in shallow subsurface Antarctic waters melted the ice has been explored but not convincingly demonstrated.

15-10 What Role Did the Greenhouse Gases Play?

• Because greenhouse gases play a major role in climatic changes at tectonic and orbital time scales, their behavior during millennial oscillations is worth considering.
  • How large were the millennial-scale oscillations in carbon dioxide and methane, and what kind of role did they play in the millennial oscillations?
• Methane concentrations in ice cores show clear millennial oscillations at amplitudes as large as 200 ppb.
  • Because these changes lag slightly behind the δ18O fluctuations in the north, they appear to have been the result of the northern oscillations and consequently a source of positive feedback, rather than a driver of those changes.
• Carbon dioxide varied by as much as 20–25 ppm over millennial intervals. These CO2 changes are roughly in phase with the Antarctic temperature responses, which probably indicates a causal link to deep-water formation and sea-ice extent in the Southern Ocean, both of which can alter CO2 concentrations in the atmosphere.
  • Because the timing of these CO2 variations roughly matches the temperature changes in the Southern Hemisphere, they provide positive feedback, but because they are opposite in sense to the temperature oscillations in the Northern Hemisphere, they act as negative feedback

15-11 Implication for Future Climate

• Because millennial oscillations occur much faster than orbital-scale changes, they have the potential to have a more immediate impact on our climatic future. 
• Some scientists have speculated that a natural millennial-scale warming is underway now.
  • If this view is correct, it means that the observed warming of the last century or so could, in part, be the result of natural processes rather than human activities.
• This claim is unjustified for several reasons.
  • Because millennial oscillations are either completely random or at best quasi-periodic, their present and future course cannot be predicted with any confidence.
  • More critically, the largest oscillations have occurred only during glacial climates, whereas the changes during the past 8,000 years of warm interglacial climate have been small and local in scale.
  • This observation argues against natural oscillations playing a major role in recent or future climate changes.
• Of course, the Greenland and Antarctic ice sheets are still in place, and they remain susceptible to some degree of melting in the warmer climate of the future.
  • Because ice sheets played an important role in the glacial-age millennial oscillations, partial melting of the present ice sheets because of human activities could conceivably trigger changes in the climate system in the future, although likely at a much smaller scale.
  • In this case, however, the ultimate cause of these future changes will still be human activities, not natural variations within the climate system