The extended 'pause' appears to have finally rattled the scientists at NOAA. They seemingly chose to robustly adjust thousands of past monthly temperature observations to make sure the embarrassing pause would be no longer.
The adjacent chart depicts the global annual temperature anomalies computed from the monthly NOAA dataset reported in 2014 (orange columns); and the red columns represent the new annual anomalies after NOAA's massive 2015 revisions.
By slightly adjusting down the past reported temperatures for years 2000-2003, and then by adjusting up years 2005-2014, NOAA was able to almost double the warming trend (i.e. slope of linear trend-line) for the 15 years ending in 2014.
To put the NOAA adjustments in scientific perspective, the chart also includes the satellite reported annual temperature anomalies for the same periods. Note for the satellite columns that only 3 of the 15 years (2012, 2013 and 2014) had minor adjustments of 0.01 each. As a result, the satellite temperature trend-lines remain flat for both 2014 and 2015.
One can assume that the RSS scientists avoided the temptation to massively adjust the satellite temperature dataset just to achieve desired political/activist objectives; thus, the 'pause' in the lower atmosphere continues and empirical science was not tarnished.
Unfortunately, NOAA has succumbed, allowing the bureaucracy to put the political cart before the scientific horse - the AGW hypothesis clearly predicts that the lower atmosphere has to warm first and at a faster pace than surface temperatures. The empirical evidence of the pause has not supported the prediction of a severely warming atmosphere for almost two decades now.
At this point in time, the satellite empirical evidence is significantly more trustworthy as it does not suffer from the constant monthly historical revisions that both NOAA and NASA perform on their respective temperature datasets.
Notes: Calendar year (annual) anomalies were computed using Excel from the RSS satellite 2015 year-end monthly dataset and NOAA's 2015 year-end monthly dataset.