Comments on “Testing the Fidelity of Methods Used in Proxy-Based Reconstructions of Past Climate”

Abstract

Mann et el. found that a version of the Regularized Expectation Maximization (RegEM) method to reconstruct the temperatures of the last millennium showed similar results to previous reconstructions in one of their earlier papers. They also tested the RegEM method in the surrogate climate of a simulation with the Climate System Model (CSM) and found no attenuation of the pseudoreconstructed centennial variability of the Northern Hemisphere mean temperature compared to the one simulated by the model. This is in contrast with the results by von Storch et al., who found, in a simulation with ECHO-G model, that the earlier Mann et al. method underestimates the centennial temperature variability of the Northern Hemisphere temperature. The newer paper by Mann et al. explains that this difference is in part due to the unrealistic character of the ECHO-G simulation. However, it is shown here that similar results to those of von Storch et al. are also found in an ECHO-G simulation that closely resembles the CSM simulation used by Mann et al. Therefore, it is argued here that this discrepancy could be related to other factors, probably to the use of a longer calibration period and to the difference between RegEM and the original method by Mann et al.