How to hide the rise

Willis Eschenbach puts some perspective on the matter in Lies, Damned Lies, Statistics … and Graphs at Watts Up With That. His point is that all of the global climate change fuss is over changes in mean annual temperature changes of a degree or so over a person’s lifespan and this is insignificant when compared to variations in temperature we normally experience. He makes the point by plotting average monthly temperatures for each year of record.

The key is the scale of the graphs. Graphs of annual temperature means usually show a temperature anomaly range of four degrees or so. Graphs of monthly average temperature means show a temperature range of sixty degrees. Climate change is a quarter of the range on the annual graphs but less than two percent on the monthly. That means any trend in the climate is exaggerated by ten times or more in an annual mean versus a monthly mean.

It should be noted that plotting monthly means puts a lot more data on the graph as well. This tends to emphasize patterns that do exist, such as seasonal variations. Long term climate change is given a reference with these seasonal variations that does not exist in annual means graphs.

These issues of presentation do not change the fundamental problem of data quality, sparse data, and the methods for determining a statistical aggregate to analyze over time. The question about just what is meant by ‘average temperature’ is not well defined and that tends to confuse the picture as well.

Comments are closed.