Do Multi-Model Ensembles Improve Reconstruction Skill In Paleoclimate Data Assimilation?

Reconstructing past climates remains a difficult task because pre-instrumental observational networks are composed of geographically sparse and noisy paleoclimate proxy records that require statistical techniques to inform complete climate fields. Traditionally, instrumental or climate model statistical relationships are used to spread information from proxy measurements to other locations and to other climate variables. Here ensembles drawn from single climate models and from combinations of multiple climate models are used to reconstruct temperature variability over the last millennium in idealized experiments. We find that reconstructions derived from multi-model ensembles produce lower error than reconstructions from single-model ensembles when reconstructing independent model and instrumental data. Specifically, we find the largest decreases in error over regions far from proxy locations that are often associated with large uncertainties in model physics, such as mid- and high-latitude ocean and sea-ice regions. Furthermore, we find that multi-model ensemble reconstructions outperform single-model reconstructions that use covariance localization. We propose that multi-model ensembles could be used to improve paleoclimate reconstructions in time periods beyond the last millennium and for climate variables other than air temperature, such as drought metrics or sea ice variables.Plain Language Summary Understanding past climate variability is important for contextualizing climate change as well as for testing the ability of climate models to simulate the climate system before global warming. However, reconstructing past climate variability remains a complex task because pre-instrumental paleoclimate proxy records, such as tree rings, corals, ice cores, and sediment cores, are geographically sparse and are not perfect recorders of climate information. Exactly how to extrapolate information from paleoclimate proxies to other locations and climate variables remains an outstanding issue. Traditionally, information about how one location varies with another location or variable (covariance) is derived from one climate model or from one instrumental data source. Here we find that reconstructions using covariance estimated from combinations of multiple climate models produce less error than reconstructions that use just one climate model.