Model insight into glacial–interglacial paleodust records

Abstract

We present a new method of analyzing model results to help identify the sensitivity of the location of different paleodust records to estimate local to regional scale variability of dust and climate variables. We use model simulations of global dust distribution from the last glacial maximum, preindustrial, current, and predicted future. The dust model has been previously shown to match available observational data for the current and last glacial maximum climate. Here, the model is compared to available source provenance data and is shown to agree with these limited observations. Using correlations and slopes across different time periods, the modeled relationships between deposition at specific observational sites and regional deposition and dustiness are shown. In addition, we evaluate the modeled relative slope of these cores to determine the location of paleodust sites that are especially easy to interpret as regional indicators of dustiness. Model predictions suggest that deposition in Antarctic ice cores is usually better than dust concentration to capture regional deposition and dustiness variability over glacial–interglacial time periods, in agreement with ice core interpretations. For Greenland, the model predicts a possible shift from dominantly wet deposition under modern conditions to dominantly dry deposition during glacial climate conditions indicating that deposition may be better suited to capture dustiness variability under LGM conditions in Greenland. The model also identifies specific regions that are not well covered by observations for glacial/interglacial or anthropocene dust variability. In addition, we evaluate the modeled relative slope of the location of these cores to determine regions that would provide ideal localities for pursuing records that would provide easily interpretable paleo-proxy records of regional dustiness.