I get a lot of questions about post-1980 proxies and I find it a very interesting question. One would expect that 1998 – the "warmest" year of the millennium – and the 1990s – the "warmest decade" of the millennium would leave a loud signal in a valid proxy. I’m going to start discussing some proxies which I have on hand, starting with a ring width series from the Twisted Tree Heartrot Hill (TTHH) site in northern Canada ending in 1999 and where the proxy shows no strong signal in the 1990s or 1998 as shown below:
Figure 1. Site Chronology, Twisted Tree Heartrot Hill. Source, D’Arrigo et al .
TTHH is a classic northern treeline site, used in Jacoby and d’Arrigo’s northern treeline composite, which in turn is used in several of the classic multiproxy studies (Jones et al. , Jones and Mann ) and is used as an individual proxy in MBH98. In MM03, we pointed out that MBH98 used an obsolete version of this series, while an updated version to 1992 was then available. D’Arrigo  updates this site to 1999.
The site chronology obviously does not have a loud 1998 or loud 1990s. The authors propose that the relationship of ring width to temperature is an upside-down U-shaped quadratic. We mention this in MM05 (E&E) but re-iterate it here because the implications of this proposal are profound for multiproxy reconstructions – which hypothesize a linear relationship between ring width and temperature. Obviously, if the relationship is an upside-down U, knowledge of the ring width does not enable you to decide if you are on the warm leg or the cold leg of the reconstruction. D’Arrigo et al  unaccountably fail to make this observation.
D’Arrigo, R. D., R. K. Kaufmann, N. Davi, G. C. Jacoby, C. Laskowski, R. B. Myneni, and P. Cherubini (2004),Thresholds for warming-induced growth decline at elevational tree line in the Yukon Territory, Canada, Global Biogeochem. Cycles,18, GB3021, doi:10.1029/2004GB002249. http://cybele.bu.edu/download/manuscripts/darrigo01.pdf