So says Marcia Glaze Wyatt here:
"Raw data is adjusted, sometimes justifiably (yet still injecting uncertainty), yet sometimes, arguably not justifiably, adding more uncertainty!!! Raw data have all been changed – 20% of it changed 16 times in the last 2 and a half years. This plot shows NOT the average surface T trend between 1880 and 2010, but rather the trend of changes made in the temperature anomalies (1880 to 2010) between May 2008 and May 2015. Take the month of January for comparison b/n 1915 and 2000. In May of 2008, the difference b/n January temperature anomalies for those years was 0.39oC. As of May 2015 note, the difference is 0.52oC (almost a degree F). ... And while one assumes that good intentions motivate the adjustments, one thing is obvious: temperatures adjustments prior to 1950 have resulted in a substantial cooling of the early century (20th) and adjustments made after 1950 have substantially warmed the record; consequently, the trend of temperature increase has significantly steepened over the years – a product of data changes. Is this an accurate reflection of reality? Uncertainty..."
