EFFECT OF CLIMATE VARIABILITY ON ESTIMATION OF GREENHOUSE PARAMETERS - USEFULNESS OF A PRE-INSTRUMENTAL TEMPERATURE RECORD

To estimate the strength of the greenhouse warming signal embedded in the noise of natural climate variability, a statistical methodology is applied to a functional model of the signal and stochastic models of the noise. The confidence of greenhouse parameter estimates, namely cl imate sensitivity, obtained from a record of global near-surface tempe rature is highly dependent on natural climate variability with a time scale of roughly a century. A pre-instrumental temperature record is n eeded to construct empirically a model for natural variability of cent ury time-scale which, in tum, is needed to determine the confidence of parameter estimates. However, for assumed models for climate variabil ity, data prior to 1950 does not significantly improve the confidence of estimates. In fact, addition of an accurate pre-instrumental record is not expected to have a large impact on climate sensitivity estimat es made with only the instrumental temperature record.