This selective review of the Quaternary paleoclimate of the South American summer monsoon (SASM) domain presents viewpoints regarding a range of key issues in the field, many of which are unresolved and some of which are controversial. (1) El Nino-Southern Oscillation variability, while the most important global-scale mode of interannual climate variation, is insufficient to explain most of the variation of tropical South American climate observed in both the instrumental and the paleoclimate records. (2) Significant climate variation in tropical South America occurs on seasonal to orbital (i.e. multi-millennial) time scales as a result of sea-surface temperature (SST) variation and ocean -atmosphere interactions of the tropical Atlantic. (3) Decadal-scale climate variability, linked with this tropical Atlantic variability, has been a persistent characteristic of climate in tropical South America for at least the past half millennium, and likely, far beyond. (4) Centennial-to-millennial climate events in tropical South America were of longer duration and, perhaps, larger amplitude than any observed in the instrumental period, which is little more than a century long in tropical South America. These were superimposed upon both precession-paced insolation changes that caused significant variation in SASM precipitation and eccentricity-paced global glacial boundary conditions that caused significant changes in the tropical South American moisture balance. As a result, river sediment and water discharge increased and decreased across tropical South America, lake levels rose and fell, paleolakes arose and disappeared on the Altiplano, glaciers waxed and waned in the tropical Andes, and the tropical rainforest underwent significant changes in composition and extent. To further evaluate climate forcing over the last glacial cycle (similar to 125 ka), we developed a climate forcing model that combines summer insolation forcing and a proxy for North Atlantic SST forcing to reconstruct long-term precipitation variation in the SASM domain. The success of this model reinforces our confidence in assigning causation to observed reconstructions of precipitation. In addition, we propose a critical correction for speleothem stable oxygen isotopic ratios, which are among the most significant of paleoclimate proxies in tropical South America for reconstruction of variation of paleo-precipitation (or SASM intensity). However, it is already well known that any particular delta O-18 value observed in speleothem carbonate is affected by two processes that have nothing to do with changes in precipitation amount-the influence of temperature on carbonate-water isotopic fractionation in the cave and the influence of changing delta O-18 of seawater. Quantitatively accounting for both "artifacts" can significantly alter the interpretations of speleothem records. In tropical South America, both adjustments act in the same direction and have the tendency to increase the true amplitude of the paleo-hydrologic signal (but by different amounts in glacial and inter-glacial stages). These corrections have even graver implications for the interpretation of tropical Northern Hemisphere speleothem records (e.g. Chinese speleothems) where the combined adjustments tend to decrease or even eliminate the "true" signal amplitude. (C) 2015 Elsevier Ltd. All rights reserved.
