Because of the relatively brief observational record, the climate dynamics that drive multiyear to centennial hydroclimate variability are not adequately characterized and understood. Paleoclimate reconstructions based on data assimilation (DA) optimally fuse paleoclimate proxies with the dynamical constraints of climate models, thus providing a coherent dynamical picture of the past. DA is therefore an important new tool for elucidating the mechanisms of hydroclimate variability over the last several millennia. But DA has so far remained untested for global hydroclimate reconstructions. Here we explore whether or not DA can be used to skillfully reconstruct global hydroclimate variability along with the driving climate dynamics. Through a set of idealized pseudoproxy experiments, we find that an established DA reconstruction approach can in principle be used to reconstruct hydroclimate at both annual and seasonal timescales. We find that the skill of such reconstructions is generally highest near the proxy sites. This set of reconstruction experiments is specifically designed to estimate a realistic upper bound for the skill of this DA approach. Importantly, this experimental framework allows us to see where and for what variables the reconstruction approach may never achieve high skill. In particular for tree rings, we find that hydroclimate reconstructions depend critically on moisture-sensitive trees, while temperature reconstructions depend critically on temperature-sensitive trees. Real-world DA-based reconstructions will therefore likely require a spatial mixture of temperature- A nd moisture-sensitive trees to reconstruct both temperature and hydroclimate variables. Additionally, we illustrate how DA can be used to elucidate the dynamical mechanisms of drought with two examples: Tropical drivers of multiyear droughts in the North American Southwest and in equatorial East Africa. This work thus provides a foundation for future DA-based hydroclimate reconstructions using real-proxy networks while also highlighting the utility of this important tool for hydroclimate research.
Bibliographical noteFunding Information:
Acknowledgements. We acknowledge the CESM1 (CAM5) Last Millennium Ensemble Community Project and the supercomputing resources provided by NSF/CISL/Yellowstone. This work was supported by the NOAA Climate and Global Change Postdoctoral Fellowship Program administered by UCAR’s Visiting Scientist Programs. This work was also supported in part by the National Science Foundation under grants AGS-1243204, AGS-1401400, AGS-1602581, and AGS-1602920. LDEO contribution number 8153.
© Author(s) 2017.