15 JULY 2007 NOTES AND CORRESPONDENCE 3699 Reply

Mann et al. (2005, hereafter MRWA05) examined the performance of the Regularized Expectation Maximization (RegEM) climate field reconstruction (CFR) technique, which has been favored in recent proxybased climate reconstruction by Mann and coworkers (Mann and Rutherford 2002; Rutherford et al. 2003, 2005; MRWA05; Zhang and Mann 2005). MRWA05 demonstrated that the method yields skillful reconstructions using synthetic proxy (“pseudoproxy”) datasets with similar attributes to actual proxy networks, derived from a simulation of the past millennium using the National Center for Atmospheric Research (NCAR) Climate System Model version 1.4 (CSM 1.4) coupled model. They consequently concluded that surface temperature reconstructions resulting from the application of this method to actual proxy data (e.g., Rutherford et al. 2005) are likely to yield realistic reconstructions within estimated uncertainties. Zorita et al. (2007, hereafter ZVS07) appear to challenge the findings of MRWA05, though they have not, as best as we can discern, demonstrated fault with any of the methods or calculations of that study. As outlined below, we reject—by explicit demonstration—the claims made by ZVS07. Von Storch et al. (2006, herafter VS06) argued that real-world proxy-based CFR reconstructions are likely to underestimate low-frequency variability, based on experiments using pseudoproxy networks derived from a millennial simulation of the GKSS ECHAM and the global Hamburg Ocean Primitive Equation (ECHO-G) coupled model. Yet they considered only the more primitive EOF-based CFR method of Mann et al. (1998, hereafter MBH98), long since supplanted by the RegEM CFR approach in all recent work by Mann and collaborators. Wahl et al. (2006) have moreover challenged the VS06 calculations, arguing that that they suffer from an artifact of an originally undisclosed procedure used by VS06 in which data were detrended prior to calibration. Wahl et al. (2006) note that such a procedure a priori removes the primary pattern of lowfrequency variability from the surface temperature data. ZVS07 argue that the skillful results of the MRWA05 experiments are somehow an artifact of the use of a long, combined nineteenth–twentieth century (1856– 1980) rather than a short twentieth-century-only (1900– 80) calibration interval. More indirectly, they suggest that behavior peculiar to the specific (NCAR CSM 1.4) model simulation used by MRWA05 and not common to their own (GKSS ECHO-G) simulations might be involved in the skillful MRWA05 outcome. Additionally, ZVS07 suggest (and argue more strenuously in VS06) that the results presented by MWRA05 may represent an artifact of an inappropriate model for the Corresponding author address: Michael Mann, Department of Meteorology, The Pennsylvania State University, 503 Walker Building, University Park, PA 16802-5013. E-mail: [email protected] 15 JULY 2007 N O T E S A N D C O R R E S P O N D E N C E 3699