Multiforced statistical assessments of greenhouse-gas-induced surface air temperature change 1890–1985

Abstract

Based on univariate correlation and coherence analyses and considering the physical basis of the relationships, a simple multiforced (multiple) statistical concept is used which correlates observational climatic time series simultaneously with volcanic, solar, ENSO, and the anthropogenic greenhouse gases forcing. This is appropriate to remove some natural climate noise in the observed data and to evaluate the components (signals) possibly due to the anthropogenic greenhouse gas forcing (CO₂, or “equivalent” CO₂ implying additional gases) during industrial time. In this paper, we apply this technique to 100 global “box” data time series 1890–1985, of the surface air temperature, using observed data from Hansen and Lebedeff. The results are presented in terms of latitudinal-seasonal and regional trends, where the observed trend patterns are compared with the hypothetical signals (statistical assessments) possibly due to anthropogenic greenhouse forcing. These latter signals can be amplified to enable a comparison with corresponding results from general circulation model (GCM) CO₂ doubling experiments. These observed-statistical assessments lead to results which are, at least qualitatively and in respect to the zonal mean temperatures, very similar to some GCM experiments indicating the maximum CO₂ doubling signals (statistical assessment > 12 K) in the arctic winter. However, these signals are moderate in the tropics and in the Southern Hemisphere (global average 2.8–4.4 K). As far as the “industrial” signals are concerned (observed period) these signals are somewhat larger (maximum 7 K, global average 0.5–0.9 K) than the observed trends (maximum 5 K, global average 0.5 K). Phase shifts of cause and effect may amplify these signals but are very uncertain.