- The Global Overturning Circulation (GOC) is a major component of the global climate system. Understanding its behavior is pertinent to our prediction of climate change in the future. The lack of long-term observations of GOC in the modern instrumental era necessitates studies of GOC using paleoceanographic records. Of great interest to climate scientists is the history of GOC since the Last Glacial Maximum (LGM, ~22,000 years before present). This is because this period experiences the last major natural global warming event accompanied by an increase of atmospheric CO2 comparable to that since the industrial revolution, and because this period was characterized by centennial to millennial scale abrupt climate transitions, making it a potential analog of the present human-induced global warming.
The neodymium isotope composition (εNd) of authigenic phases in marine sediments has been used as a circulation proxy to reconstruct past change of GOC, but our incomplete understanding of the processes regulating the distribution of εNd in the ocean has so far hampered the application of this proxy, such that there is not yet a globally consistent description of past GOC history using this tracer.
In this dissertation, I develop new methods to extract authigenic phases from marine sediments for εNd analysis. The methods are designed specifically for sediments from the Pacific Ocean, which often contain disperse volcanic materials that usually cause significant contamination to extracted Nd when using existing methods developed for the other ocean regions. Using a suite of geochemical tools, I identify the sources of extracted Nd, and show that the methods are capable of extracting authigenic Fe-Mn oxyhydroxides with negligible contamination from detrital sediments including volcanic materials. However, when applying the methods to sediments from the Gulf of Alaska, I find that the authigenic εNd from modern core top samples are consistent more positive than the local bottom water, suggesting that in site diagenesis of volcanic materials affects the εNd of authigenic phases. Differences of εNd between authigenic phases and bottom water are widespread in the global ocean, thus questioning the traditional assumption that authigenic phases passively record bottom water εNd. I suggest that this difference is evidence that there exists a significant sedimentary source of Nd to the ocean, which is linked to the diagenesis of authigenic phases and detrital sediments. I propose a conceptual model to describe how authigenic phases acquire their εNd signatures through interactions with bottom water, pore water and detrital sediments. The two key factors controlling the interactions are the benthic exposure time and the reactivity of detrital sediments. I show that this conceptual model is applicable to the distributions of seawater and authigenic εNd in the Pacific Ocean.
With the new method and conceptual model, I generate two new authigenic εNd records using cores from the Gulf of Alaska at intermediate and abyssal depths, and reconstruct the Pacific circulation since the LGM. I synthesize the existing Pacific authigenic εNd records using a box model, in which I separate the non-conservative component from the preformed-conservative mixing component of εNd. Using the non-conservative component of εNd, I estimate the Pacific circulation rate, together with benthic radiocarbon (14C) records from the same sites. The results show that the transport of Southern Ocean deep waters to the Pacific reduced to ~half of today in the LGM, but increased later in the Northern Hemisphere stadial events to ~2-fold of today during the deglaciation. The inferred circulation changes imply accumulation of respired carbon in the glacial deep Pacific and its subsequent release during the deglaciation, supporting the conclusion that the Pacific circulation plays a key role in the global carbon cycle since the LGM.
Extending the conceptual model globally, I synthesize published authigenic εNd records from the global ocean since the LGM. The dominant modes of variability in these records are extracted using Principal Component Analysis. The first Principal Component (PC1) is related to the LGM-Holocene difference of εNd: the glacial deep ocean has consistently more positive εNd than the Holocene, and the magnitude of this increase was highest in the subpolar North Atlantic, decreasing toward the Southern Ocean and further so in the Pacific and Indian Oceans. This change is interpreted as the result of an increase of the preformed εNd of the deep water produced in the North Atlantic, linked to the southward shift of the deep water formation location in the LGM. The second Principal Component (PC2) is related to the abrupt changes during the Northern Hemisphere stadial events: in these events North Atlantic and North Pacific εNd converge. This mode of change is interpreted to be the consequence of strengthening of Southern Ocean overturning circulation, which overcomes the effect of the reduction of North Atlantic overturning, leading to overall stronger overturning of the global deep ocean. The PC1 and PC2 related mechanism are quantified using a global εNd box model. The reconstructed GOC since the LGM correlates to changes of atmospheric CO2 and ocean heat transport, giving key evidence to the long-standing hypothesis that links GOC to these climate variables.