COMPARISON OF STATISTICALLY OPTIMAL APPROACHES TO DETECTING ANTHROPOGENIC CLIMATE-CHANGE

Three statistically optimal approaches, which have been proposed for d etecting anthropogenic climate change, are intercompared. It is shown that the core of all three methods is identical. However, the differen t approaches help to better understand the properties of the optimal d etection. Also, the analysis allows us to examine the problems in impl ementing these optimal techniques in a common framework. An overview o f practical considerations necessary for applying such an optimal meth od for detection is given. Recent applications show that optimal metho ds present some basis for optimism toward progressively more significa nt detection of forced climate change. However, it is essential that g ood hypothesized signals and good information on climate variability b e obtained since erroneous variability, especially on the timescale of decades to centuries, can lead to erroneous conclusions.