Paleoclimate proxy records from regions sensitive to humidity/aridity extremes provide valuable insights into natural forcing mechanisms underlying long-term climate variability in the wider region. One such area is Northwest Australia, where the Australian monsoon impacts its northernmost fringes, which are bordered by the Great Sandy Desert inland. Marine sediments from the Australian Northwest Shelf record fluvial run-off and aeolian dust input during the wet and dry seasons. The location is therefore ideal for investigating long-term variability in the Australian monsoon and Northwest Australian dust fluxes over orbital timescales. However, there are few continuous, high-resolution paleoclimate records from the Australian Northwest Shelf spanning the Early Pleistocene, and there is ambiguous orbital phasing even among Late Pleistocene paleoclimate records from the region. Here, we present geochemical and environmental magnetic proxy records of CaCO3 and dust-flux variability spanning 2.9 to 1.6 Myr ago from International Ocean Discovery Program Expedition 356 Site U1464 on the Australian Northwest Shelf. We establish a new, orbitally-tuned chronology for Site U1464, and observe strong obliquity variability (41 kyr and 54 kyr periodicities) but almost no precession signal in our dust records. We propose that the 41 kyr cycle in Northwest Australian dust fluxes could be a linear response to the East Asian winter monsoon (EAWM) and/or summer inter-tropical insolation gradient (SITIG), whereas the 54 kyr cyclicity might be a non-linear response to obliquity amplitude modulation via the SITIG effect on cross-equatorial flows.

Global ice volume (sea level) and deep-sea temperature are key measures of Earth’s climatic state. We synthesize evidence for multi-centennial to millennial ice-volume and deep-sea temperature variations over the past 40 million years, which encompass the early glaciation of Antarctica at ~34 million years ago (Ma), the end of the Middle Miocene Climate Optimum, and the descent into bipolar glaciation from ~3.4 Ma. We compare different sea-level and deep-water temperature reconstructions to build a resource for validating long-term numerical model-based approaches. We present: (a) a new template synthesis of ice-volume and deep-sea temperature variations for the past 5.3 million years; (b) an extended template for the interval between 5.3 and 40 Ma; and (c) a discussion of uncertainties and limitations. We highlight key issues associated with glacial state changes in the geological record from 40 Ma to present that require attention in further research. These include offsets between calibration-sensitive versus thermodynamically guided deep-sea paleothermometry proxy measurements; a conundrum related to the magnitudes of sea-level and deep-sea temperature change at the Eocene-Oligocene transition at 34 Ma; a discrepancy in deep-sea temperature levels during the Middle Miocene; and a hitherto unquantified non-linear reduction of glacial deep-sea temperatures through the past 3.4 million years toward a near-freezing deep-sea temperature asymptote, while sea level stepped down in a more uniform manner. Uncertainties in proxy-based reconstructions hinder further distinction of “reality” among reconstructions. It seems more promising to further narrow this using three-dimensional ice-sheet models with realistic ice-climate-ocean-topography-lithosphere coupling, as computational capacities improve.