Go read the paper. But in my opinion, this is the money quote:
- NOAA adjustment procedure fails to address these issues. Instead, poorly sited station trends are adjusted sharply upward (not downward), and well sited stations are adjusted upward to match the already-adjusted poor stations. Well sited rural, non-airport stations show a warming nearly three times greater after NOAA adjustment is applied.
- What the compliant thermometers (Class 1&2) say: +.155°C/decade
- What the non-compliant thermometers (Class 3,4,5) say: +.248°C/decade
- What the NOAA final adjusted data says: +.309°C/decade
In other words, NOAA adjustments to raw data - adjustments that are in the direction OPPOSITE to what you would expect to account for Urban Heat Island effect - are responsible for fully half of the alleged warming in the US temperature record.
By the way, 1.5 degrees per century is the same as the rate of warming observed from 1700-1850, long before the SUV was invented. So the rate of warming in the US over the past 40 years is no different from the rate of warming before humans started burning fossil fuels in significant quantities.
So...if there's no unusual warming, then why do we have to make up a theory to explain it? Where, in short, is the evidence for an anthropogenic impact on temperature?
Ask the NOAA. It's in their adjustments.
Well done, Watts et al.
Now, not to put too fine a point on it, but some of us have been saying this - or at least, TRYING to say it - for years:
Variances between satellite temperature records – by far the most accurate and global means of measuring atmospheric temperature – and land-based thermometer records are largely the result of significant problems with the latter, which are affected inter alia by land-use changes, the Urban Heat Island (UHI) effect, station drop-out, inconsistent coverage, poor station siting, and manual alteration of data. The difference is manifested in a strong divergence between the satellite and land-based temperature record over the past six years. Since 2003, the satellite record shows temperatures dropping at a rate of 2.84ºC/century (UAH record) and 3.60ºC/century (RSS record), while the land-based temperature data maintained by the Goddard Institute of Space Studies shows a rate of decline of only 0.96ºC/century (GISS). The divergence between the satellite and the land-based temperature record is becoming more pronounced over time. It is curious, although perhaps not surprising, that despite the many problems with the quality of the data in the land-based temperature record, the proponents of the AGW thesis prefer the results it produces over those produced by satellite temperature measurements. All of these data, incidentally, are available on-line.
 See Ross R. McKitrick and Patrick J. Michaels, “Quantifying the influence of anthropogenic surface processes and inhomogeneities on gridded global climate data”, Journal of Geophysical Research, Vol. 112, D24S09, doi:10.1029/2007JD008465, 2007 [http://www.uoguelph.ca/~rmckitri/ research/jgr07/ M&M.JGRDec07.pdf].
 The number of global temperature measurement stations dropped from a high of 6000 in 1970 to roughly 2000 today. Most of the dropouts occurred at the fall of the Soviet Union, and most of the lost stations were rural. The result is a higher statistical emphasis on urban stations, exacerbating the contamination from UHI and land-use changes. See Joseph D’Aleo, “Recent Cooling and the Serious Global Data Base Issue” [http://icecap.us/images/uploads/Recent_Coolingand_the_Serious_Data_Integrity_issue.pdf].
 See the “How Not To Measure Temperature” series maintained by Anthony Watts at [http://wattsupwiththat.com/category/weather_stations/]. Recent examples include thermometers situated in front of air condition exhausts, near barbecues, and buried in Antarctic snow.
 A data posting error by the NASA GISS staff in autumn 2008 (in which September temperature data for Siberia was mistakenly copied over October data) led to renewed cries of impending doom by the AGW alarmists, until the data error was reported by a blogger, and subsequently correct by GISS. For a breakdown of the incident, see [http://wattsupwiththat.com/2008/11/12/corrected-nasa-gistemp-data-has-been-posted/].
 Steve Goddard and Anthony Watts, “GISS Divergence with satellite temperatures since the start of 2003”, 18 January 2009 [http://wattsupwiththat.com/2009/01/18/giss-divergence-with-satellite-temperatures-since-the-start-of-2003/].
 NASA GISS data are here [http://data.giss.nasa.gov/gistemp/]. RSS MSU/AMSU data re here [http://www.ssmi.com/msu/msu_data_description.html]. UAH AMSU data are here [http://datapool.nsstc.nasa.gov/]. Hadley Centre data are here [http://hadobs.metoffice.com/hadat/].
It's certainly nice to be vindicated. Anthony, Ross and company have earned a pat on the back.
P.S. It's amazing what you can do when you have actual MEASURED DATA to work with, isn't it? As opposed to the outputs of models tailored to your prejudices - or, as is so often the case these days, a load of crap you just made up?