I wanted to have a reference for any posts that talk about temperature and temperature departure (aka: temperature anomaly). This post shows an example of the calculation of a “temperature departure”.
Temperature departure is referring to the difference between temperature in the given month and the average temperature of that month over a range of years. For example, as shown in the table below, the average temperature for the month of November 2012 in the USA was 44 F. I got this from NCDC time series data which I downloaded a couple of weeks ago. You can calculate an average November temperature as well if you choose different average periods. Using the NCDC data, I calculated the 1981-2010 and 1900-1999 average November data, also shown in the table below.
T (F) TD 81-10 (F) TD 00-99 (F)
November 2012 44.05 +1.28 +2.03
November 1981-2010 42.77 N/A N/A
November 1900-1999 42.01 N/A N/A
How does November 2012 compare to past Novembers? Well, it doesn’t make much sense to choose just any November unless you have a specific reason, such as comparing November 2012 to November 2011 so you can remember why you wore a sweater on Thanksgiving last year and didn’t have to this year. From a climate standpoint, it’s more informative to choose a LOT of Novembers and average those November temperatures together. How many is enough? A good place to start is 30 years. This is the length of time that the climate system typically needs to average out natural variations in temperature, such as effects of El Nino-Southern Oscillation and the corresponding La Nina, cooling from a large volcanic eruptions like in 1991-1992, or a less famous (than El Nino) form of variability that affects regional or global temperatures. NCDC uses the last century as a time length for calculating the “average” temperature. HPRCC and NCDC also use the 30 year average period between 1981-2010 and call this the “Climate Normal”. Both are completely valid as long as you understand what the temperature in any particular month is being compared to. The temperature departure for November 2012 (TD in the table) is calculated as T in November 2012 (44.05 F) minus the average November temperature. Comparison to 1981-2010 gives 44.05 F minus 42.77 F = +1.28 F. Similarly, comparison of November 2012 to 1900-1999 gives 44.05 F minus 42.01 F = +2.03 F. Hence the USA had a +2 F temperature departure in November 2012, as pointed out in a different post.
A discussion of temperature departures is peppered with specific statements of time and space scales. In November 2012 then, nothing changes about the average temperature (44 F), but if we compare this to the climate normal of 1981-2010, we (most likely) get a different departure than we would compared to 1900-1999. This is clear in the table. More to the point, temperature departure is a relative metric and temperature is an absolute metric. Temperature departures are usually more informative since it’s hard to just remember if it’s a little cooler or warmer any particular month, but temperature itself is of course more intuitive.