Tuesday, 3 June 2014

Some slightly longer views on the weather at Carwoola

I was looking at the information from my weather station recently and noticed that I had close to six months of data in there.  So that is now in my database also.  Then I noticed a dataset from my previous weather station before it went to weather station morgue.  So that can be played with as well.

I am not sure how far I will pursue this but I will start off with some temperature data.

The first graph shows maximum and minimum temperatures since November last year.
I have applied polynomial trends to remove the effects of day to day variation.  The correlation coefficients are not significant, but given the level of day to day variation are not trivial.   To no surprise whatsoever the trends show it was warmer in Summer than Autumn!  At a more detailed level:

  • The apex of the trend for maximum occurs in mid-January while the minimum continues to rise for another month or so before dropping off.
  • There has been much comment about the warm patch in mid-late May.  It is interesting that this is reflected in the kick-up at the end of the trend line.
The second graph illustrates a comparison between the periods 12 November to 31 May.  As the original data resembled a cross section of the Alps it wasn't possible to see what was going on and I decided to apply a 7 period moving average to remove some of the 'noise'.  (Note I chose 7 periods as something that seemed a modest period of time: I'm not suggesting that a natural phenomenon actually works on a weekly cycle.)  Here is the outcome.
Until the end of February the two series appear almost perfectly out of synchronisation.  When one rises the other falls and vice versa.  It then seems that the series for 2012-13 show essentially a warmer Autumn (possibly due to the rainier -  and thus cloudier - weather) in 2013 -14 until warm later May kicks in.

No comments: