Observations of short periods where global mean temperatures have fallen, even though atmospheric concentrations of
GHGs have increased has raised questions as to the causal link between concentrations and warming (Plimer, 2009). A typical
way this can be determined is to check for a non-positive coefficient for the ordinary least squares regression of global
mean temperature against time for each 10-year period. For consistency with the previous analysis we used the same time
window as above from January 1950 during which GHGs have been consistently rising. There were 11 such (potentially
overlapping) time periods with falling temperatures between 1950 and 2010 (Fig. 1). The same bootstrap simulation
described above was performed including and excluding eCO2 forcing (Fig. 7b). When eCO2 forcing is included in the
analysis, a median of 15 periods of apparent falling decadal temperatures would be expected according to Model B and
10 periods according to Model E, and thus the 11 observed since 1950 are not unexpected. In fact, 11 is approximately
the 20th percentile of the simulated values from Model B in this case. However, when eCO2 is excluded from the analysis
and so does not influence global mean temperature, a median of 26 such periods are simulated from Model B. In fact, none
of the 100,000 simulations from Model E (excluding eCO2) in this instance produced 11 or fewer decadal periods of falling
temperatures. For Model F only 0.04 percent of the simulations produced 11 or fewer decadal periods of falling
temperatures.
Observations of short periods where global mean temperatures have fallen, even though atmospheric concentrations of
GHGs have increased has raised questions as to the causal link between concentrations and warming (Plimer, 2009). A typical
way this can be determined is to check for a non-positive coefficient for the ordinary least squares regression of global
mean temperature against time for each 10-year period. For consistency with the previous analysis we used the same time
window as above from January 1950 during which GHGs have been consistently rising. There were 11 such (potentially
overlapping) time periods with falling temperatures between 1950 and 2010 (Fig. 1). The same bootstrap simulation
described above was performed including and excluding eCO2 forcing (Fig. 7b). When eCO2 forcing is included in the
analysis, a median of 15 periods of apparent falling decadal temperatures would be expected according to Model B and
10 periods according to Model E, and thus the 11 observed since 1950 are not unexpected. In fact, 11 is approximately
the 20th percentile of the simulated values from Model B in this case. However, when eCO2 is excluded from the analysis
and so does not influence global mean temperature, a median of 26 such periods are simulated from Model B. In fact, none
of the 100,000 simulations from Model E (excluding eCO2) in this instance produced 11 or fewer decadal periods of falling
temperatures. For Model F only 0.04 percent of the simulations produced 11 or fewer decadal periods of falling
temperatures.
การแปล กรุณารอสักครู่..
![](//thimg.ilovetranslation.com/pic/loading_3.gif?v=b9814dd30c1d7c59_8619)