Roy Spencer’s response to Kevin Trenberth, April 26, 2009 [see also a version of this post on Roy’s weblog]
Kevin Trenberth begins his blog post with, “I saw Roy Spencer’s comment for the first time and it is not correct”, but I see no specific refutation of any of my points in my blog posting.
To support my comments, here are the global-average CERES ERBE-like ES-4 Edition 2 radiative flux anomalies for reflected solar (1st graph) and outgoing longwave radiation (OLR, 2nd graph)…these are daily running 91-day averages:
Clearly, the long-term “trend” during 2000 through 2008 was in the reflected solar (SW), not OLR (LW).
What is important for global warming or cooling is the sum of the global SW and LW, shown in the following graph (note I have flipped the y-axis, to correspond to the sense of the plot Kevin and John Fasullo showed in their Science Perspectives article):
What Kevin DOES discuss then is the anomalous drop in OLR around the beginning of 2008. He should recognize that there is a very simple explanation for it: global-average temperatures were quite low at that time, as seen in the next graph:
The expected change in OLR with temperature is about 3.2 Watts per sq. meter per degree C. This temperature plot shows a fall of about 0.4 deg. C from early 2007 to early 2008, which should cause a reduction in OLR by about (0.4 x 3.2 ), or about 1.3 Watts per sq. meter. As seen in the LW plot above, there was indeed a fall of about 1 Watt per sq. meter. To the extent that the drop in OLR with cooling was not quite as much as might be expected could be due to a small positive feedback in high clouds and/or water vapor. These are just rough estimates, anyway.
In our new paper accepted for publication in JGR, we show that this 2007-08 cooling event was due to a temporary increase in low cloud cover, evidence of which is clearly seen in the form of a large spike in reflected sunlight in the first plot, above. The OLR event is completely consistent with the resulting drop in global-average temperatures.