RSS reported an April Temperature anomaly of 0.110 C. That’s up relative to March’s value of -0.027. I’ve plotted all anomalies since Jan 1980, along with the multi-model mean for projections for the surface temperature. Some best fit curves and uncertainty intervals are provided for reference:

For those who asked, I extended the multi-model mean; the best fit is applied through the entire period illustrated. Note that it is 0.211 C/decade which is higher than the nominal trend of 0.2C/decade discussed in the IPCC AR4– also the multi-model mean corresponds to surface temperatures while RSS is for temperatures in the lower troposphere. Nevertheless, we can still compare– bearing in mind the granny-smith to golden delicious aspect of the comparison.

Currently, the trend observed trend since 1980 is roughly 3/4 that best fit trend computed based on the multi-model mean projections Also, for the multi-model mean falls outside and above the 2-σ uncertainty intervals computed either using ARMA(1,1) or Red corrected white noise. However, this comparison doesn’t include the uncertainty in our knowledge of the multi-model mean. (That rather small relative to uncertainties highlighted by the graph; I’ll try to remember to code that into graphs as the various agencies report their April anomalies.)

The temperature anomalies themselves fall inside the range expected. Using a baseline from 1980-1999 forces agreement of the average anomaly during that period. As that period is recent, models would have to be violently off for temperature anomalies themselves to fall outside that range. (This is why I tend to focus on trends which are not affected by choice of baseline.)

Still waiting for the other groups. It’s a rare month when all agencies move in the same direction, we’ll wait and see how many happened to show rises from March to April. 🙂