Calculation of 30-year climate normals

[matti]

Hi John. I see that NOAA released the new 1981-2010 climate normals this month. So, as I understand, the new normals will use a data period from 1981 to 2010, replacing the old period of 1971-2000. I understand that we want a large enough data set to give a realistic climate "normal," thus the choice of 30 years. But, instead of waiting to update the normals every 10 years, is there a reason they don't simply update the normals ever year?

I could see that recalculating the average every year wouldn't likely yield any material difference in the figures reported to the public. Thoughts? I hope my question makes sense.

I read the announcement here:
http://www.crh.noaa.gov/news/display_cmsstory.php?wfo=dlh&storyid=70807&source=0

 

[frnash]

… instead of waiting to update the normals every 10 years, is there a reason they don't simply update the normals ever[y] year? …

Certainly they have the computer power to do it. Recalculating "30 year normals" even on a daily basis would almost be a trivial exercise compared to the number of recalculations that are done each day on the atmospheric models using supercomputers.

I think it's really a trade off between the most absolutely current "normals" (probably of only minimal value) vs. a more stable set of "normals" to serve as a point of reference — one that isn't such a "moving target."

Hand off to John!

 

[jd]

Matti.

I'm sorry, but I really do not have an answer for you. I suspect that the answer might have to do with the amount of labor involved. While computers do all of the actual number crunching, humans must still do the quality control on the results.

I guess my question is why only use 30 years. It seems like a bit of an arbitrary number. Why not go to 50 years or even 60 years or more? I suppose the answer to that question lies in the fact that if they did that, then the climate change numbers would not be as dramatic and some climatologists might lose their jobs!

Just a thought.

Sorry I could not give you the answer for sure.

-John

 

[matti]

John, thanks.