Evaluation of the Temperature Trend and Climate Forcing in the Pre- and Post Periods of Satellite Data Assimilation - Satellite-based Applications on Climate Change - page 55

Fig. 4.2 Same as Fig. 4.1 except for the middle troposphere (500-700 hPa) (Adapted from Powell
and Xu ( 2011 ), Figure 3)

Table 4.1 Trend of global
mean temperature for
December-February units:
K/decade

NCEP/NCAR ERA40 MSU

100-70 mb (1979-2002) 0.92 0.16 0.47

100-70 mb (1958-1978) 0.24 0.18 n/a

700-500 mb (1979 - 2002) 0.015 0.064 0.090

700-500 mb (1958 - 1978)

0.021 0.044 n/a

4.4 Multiple Linear Regression Analysis

Generally, three methods are used to assess the impact of climate forcing changes
on the Earth: modeling, composite analysis, and statistical analysis. A previous
composite analysis likely indicates nonlinear relationships in the stratosphere and
troposphere (Powell and Xu 2010 , 2011 ); however, for a complicated climate
system, the composite analysis cannot identify which sources are contributing to
the observed temperature anomaly (TA) signal. Fortunately, the multiple linear
regression methodology is a good way to identify the linear signal from potential
multiple sources (Haigh 2003 ). Although the regression methodology can typically
only address linear relationships, it is a step to improve our understanding of the
relative source contributions. In other words, the regression method will overcome
to some degree the limitations in the composite analysis. However, one must keep
in mind that the regression approach also cannot make attributions for cause and
effect but only points to possible areas for investigation.