Constraining the role of age uncertainty in event relationships in data-model comparison

Abstract:

Many paleoclimatological questions regarding relationships between system parts depend on the ability to establish the precise timing of events. Further understanding of such relationships can result from data-model comparisons. Whereas model output is time certain, paleo data often contain large chronological uncertainties that hinder the equivalence of the data-model comparison. To constrain the role of age uncertainty in assessing the relationship between events within a data-model comparison framework, we demonstrate a case study of simulating age uncertainty in model output. To determine whether mineral dust was higher during megadroughts in the southwest United States, we compare paleo records of time-uncertain dust in lake sediment and time-certain droughts from tree rings. To better understand the mechanisms of this process, we compare these results with dust emissions and deposition and soil moisture from the time-certain output of the Community Earth System Model Last Millennium Ensemble. The dust-drought and data-model comparison is then repeated after incorporating age uncertainty into the model output. In this case study, incorporating age uncertainty in the analysis is the difference between a positive and negative result. Paleoclimatological answers to questions that depend on the precise timing of events in time uncertain records need a careful evaluation of age uncertainty, and we propose that our simple approach may be generally useful in paleo data-model intercomparison.