This paper is related because it gives convincing data that the current
warming is not a natural event, as opposed to the MWP, which *had* to be
natural:

http://www.pnas.org/cgi/content/full/94/16/8370?ck=nck

"The probability of the observed [modern] changes occurring through natural,
as opposed to anthropogenic, causes appears to be exceedingly small. First,
although a major ice age causes a larger temperature change than has happened
so far in response to CO2, the temperature increase that occurred between 1920
and 1990 would have taken more than 2,000 years even at the historically rapid
rate of the last deglaciation. During deglaciation, transient warming in the
North Atlantic after a Heinrich event was faster (35), but there is no
evidence for a Heinrich event during the last few centuries. Second, internal
climate oscillations, shifts in storm tracks, and the like obviously can
change local and regional climate by much larger amounts than observed in the
hemispheric averages. These regional fluctuations appear to be superimposed on
the general global trend. Additionally, such internal oscillations produce
warm and cool regions that interchange over decade to century time scales (32,
36), but whose effects largely cancel in hemispheric averages. Third, while
there is reasonable evidence for greater climate variability during the
Holocene than has been observed during the period where instrumental data are
available (37, 38), there is no evidence in the statistics that a major
unidentified source of natural variation is present during the instrumental
record. Such a source would have to mimic, perversely, either solar irradiance
changes or the changes in atmospheric CO2 to cause the observed temperature
changes and to be mistaken for them. Similarly, while mindful of the many
caveats on data quality, spatial coverage, etc. given in ref. 1, the
appearance of possible leap-year artifacts at a level below 10 mK in the
residuals suggests that the data cannot be as untrustworthy as is occasionally
implied. The residual temperature variation remaining once the known effects
of precession, solar irradiance changes, and atmospheric CO2 concentration are
removed bound unknown effects to about 200 mK peak-to peak in the hemispheric
average series during the last century.



Consider the null hypothesis that the observed temperature fluctuations and
atmospheric CO2 levels are independent: The probability that the hemispheric
temperatures would fluctuate purely by chance in such a way to produce the
observed coherences with CO2 is exceedingly low. Given that the records
encompass more than a century, the probability is so low that one would not
expect to see such an event by chance during the age of the earth. The
probability of the observed coherence between atmospheric CO2 and changes in
the timing of the seasons shown in figure 13 of ref. 2 without a causal
connection is similarly low. Consequently one must strongly reject the
hypothesis of independence between atmospheric CO2 and temperature. The
alternative hypothesis, that increasing levels of atmospheric CO2 plus a
slight change in solar irradiance are causally responsible for the observed
changes in temperature, in contrast, results in test statistics that are
ordinary in every way. Because major changes in climate as a response to human
use of fossil fuels have been predicted for more than a century (39, 40),
their detection can hardly be considered surprising. "