
How
we attribute different contributions to an observed change in expected temperatures
nExpected
temperatures are not a random variable.
nTreat
this as a standard regression problem, with contributions from model-simulated patterns
of response
to different external drivers as the unknown parameters.
nConventional
likelihoods for uncertainty estimates.
nComplications
arising from noise in model-simulated patterns: use a Total Least Squares or
Errors in Variables approach.