There is no specific justification for the claim that «
uncertainty associated with historical temperature reconstructions» is larger in TAR than in AR4, as no specific citation is given in the ISPM.
The ISPM overview states: «Natural climatic variability is now believed to be substantially larger than previously estimated, as is
the uncertainty associated with historical temperature reconstructions.»
The transient constraint has been looked at before of course, but efforts have been severely hampered by
the uncertainty associated with historical forcings — particularly aerosols, though other terms are also important (see here for an older discussion of this).
The transient constraint has been looked at before of course, but efforts have been severely hampered by
the uncertainty associated with historical forcings — particularly aerosols, though other terms are also important (see here for an older discussion of this).
Not exact matches
Previous large natural oscillations are important to examine: however, 1) our data isn't as good
with regards to external forcings or to
historical temperatures, making attribution more difficult, 2) to the extent that we have solar and volcanic data, and paleoclimate temperature records, they are indeed fairly consistent
with each other within their respective
uncertainties, and 3) most mechanisms of internal variability would have different fingerprints: eg, shifting of warmth from the oceans to the atmosphere (but we see warming in both), or simultaneous warming of the troposphere and stratosphere, or shifts in global temperature
associated with major ocean current shifts which for the most part haven't been seen.
In summary, our results show that in the CESM - LE, the range of
uncertainty in projected NAO trends and
associated influences on SAT and P over the next 30 years can be obtained to a large degree from the Gaussian statistics of NAO variability during the
historical period,
with some regional exceptions possibly
associated with AMOC variability.
If I had the time, I could formulate a Kalman Filter which I would prime
with the
historical data, and I could then project optimal estimates and
associated uncertainty bounds for the process forward (or backward) in time.