Spatial Representativeness of Oceanic Proxies and Assimilation into Climate Models

Biases and uncertainties in climate models and proxies make their combined use through data assimilation (DA) challenging. The aims of this study are to assess the spatial representativeness of reconstructed sea surface temperature (SST) archives in the North Atlantic region from the Ocean2k synthesis database and to explore their potential for DA experiments over the past two millennia, using General Circulation Models (GCMs) and the Earth system model of intermediate complexity LOVECLIM. We explore the influence of spatial and time scale resolutions on simulated SSTs, and present several model-data scaling methods for an offline DA application with a particle filter. Results show that marine proxy locations are representative of a large oceanic area, with spatial length scales of several thousands of kilometers. Thus, they can be theoretically represented in climate models. However, proxy variability is several orders of magnitude higher than model variability. Model-data mismatch is highest for the coarser-grid resolution model LOVECLIM. The case studies presented highlight potentials and limitations of connecting paleo SSTs and climate models over multidecadal to centennial timescales.