As far as I can tell the principal events of the last 5 yrs pertaining to confidence in AGW are the following (please email me if anything is wrong or misleading below):
(1) Antarctica's measured surface temperatures haven't warmed in about 40 years*. The penguins are starting their march later in the spring. The southern hemisphere hasn't warmed much in 25y. The northern hemisphere warmed fast enough in the 1980's and the 90's to compensate so that the world average was consistent with GHG model projections, but only the arctic seems to be measurably warming since 2001.
(2) Since 2001, figures from the US National Climate Data Center show the world has warmed by .03 degrees, within the range of measurement error and thus consistent with no warming at all. Also since 2003 the oceans have cooled. Since 2001 atmospheric methane has mysteriously declined.
(3) A recalibration has brought the satellite data into better agreement with warming than it was in 2001. This is the only event of which I'm aware that might lead to increased confidence. However, even after the recalibration, it remains true that "none of the satellite data sets show warming in the tropical troposphere, and only one [of three] shows warming higher above the tropics." The tropics are not of fringe importance- they account for half the world's atmosphere, and "GHG models predict strong warming in the upper troposphere over the tropics." Also, according to Monckton, the agreement that does exist is dependent on treating as part of the trend a single potentially outlying event, the '98 El Nino. Finally, even after the recalibration, the satellite and radiosonde data still shows less warming at altitude than the surface measurements do, whereas the signature of GHG warming is more warming at altitude.
(4) The hockey stick graph, which was the flagship argument for AGW, featured in the 2001 summary, has been debunked, most recently by the USNAS. Today, if one were to draw a best guess graph going back far enough to give some perspective, it would look like the graph featured in UN reports in the 1990s, with a Medieval Optimum and a Maunder Minimum, in which the current run of warming doesn't seem remarkable. That a published paper on which one had relied so centrally was in error might seem cause for increased humility, as might discovering one's instruments were not correctly calibrated.
(5) There is a new appreciation of the importance of aerosols, which are believed to exert a cooling influence. Even direct effects are poorly understood, and indirect effects (such as influences on cloud formation!) are simply omitted from most of the models relied on by the IPCC. IPCC estimates that the Direct Radiative Effect of natural and anthropogenic aerosols is 3 times the estimated radiative forcing of anthropogenic CO2, with anthropogenic aerosols estimated to cancel 1/3 of anthropogenic GHG forcing. But different models of the effect of aerosols differ by an order of magnitude, and it doesn't seem to be ruled out either by theory or experiment that overall anthropogenic contribution (aerosols + GHGs) may be cooling.
(6) Since 2001 peer reviewed alternative theories of climate change have advanced strongly. In particular, evidence has been provided that fluctuations in cosmic ray backround explain climate moves over earth's history much better than atmospheric CO2 concentration (which seems to have been a trailing indicator over paleohistory). Recent lab experiments provide a mechanism by demonstrating that increased cosmic rays can cause increased clouds. These authors claim to naturally explain the recent lack of antarctic warming (since snow, like clouds, is white), and argue this hypothesis already better explains the data than GHG theories. Cosmic ray theories predict the temperature will start to cool in coming decades.
(7) "The third assessment report reported that solar activity was exceptionally high in the 20th century in the context of the last 400 years. Since then, new reconstructions have indicated it may be exceptionally high compared to the last 8,000 years." Solar activity was low at the Maunder minimum. The IPCC believes solar activity is not important, but indirect effects, "for which new evidence has emerged", are ignored. Other peer reviewed publications differ. Recent publications project that solar activity is likely to decline over the next several decades, and some peer reviewed articles predict advancing ice sheets within the 21st century. (I'm not endorsing this!)
(8) Over the last few years, Mars's icecaps have been melting, and observations also indicate global warming on Triton, Pluto, Neptune, and Jupiter. (Update-- the draft report itself has now been released, and I searched the entire pdf file, without finding any mention of any of these planets. )
(9) "The IPCC assumes temperature data are not contaminated by upward biases, ... many studies have shown that changes in land use and land cover can have large regional effects on the climate that are comparable in magnitude to temperature and precipitation changes observed over the last several decades ... the IPCC uses trend estimation techniques that likely overstate the statistical significance of observed changes, and the results of trend analysis often depend on the statistical model used."
Quotes above are from the Fraser Institute's Independent Summary for Policy Makers of the IPCC Fourth Assessment Report which reviews the draft 4th assessment report chapter by chapter and reproduces many graphs and data. Lord Monckton (former advisor to Thatcher) more briefly reviewed the draft 4th assessment report, here and comments "In the final draft report of the new report there is a change in tone. Though carbon dioxide in the air is increasing, global temperature is not."
That isn't the impression one gets from the official summary for policymakers. If there is valid reason why confidence in AGW has increased from 66% to 90% over the last 5 years, I would be very interested in learning of it. If there are any errors in the above, or if data is merely presented above in a misleading way, I would be very interested in hearing of them. If there isn't, if in fact evidence and events are going the other direction, it would be good if policymakers and the public were so informed.
The question arises, how can the IPCC summary be so at variance with the evidence. It is perhaps worth recalling that the IPCC summary is not peer reviewed, but more importantly no individual scientist puts her name on it and is accountable if it is wrong. Its my experience that scientists will tell journalists all kinds of baloney, but they are a lot more circumspect about what they will actually publish under their names in scientific journals.
*-- Claims that "Antartica's surface temperature had warmed by roughly 2.5 degrees C over the last half century" are consistent with this, because (as can be verified from Goddard Institute for Space Studies 2006 or Mears and Wentz 2006) measured surface temperature in Antarctica increased by roughly 2.5 degrees between 1950 and the early 1960's, and hasn't increased since.