The frost report

Please note the lower case in the title: that should give you a hint this is nothing to do with Sir David of that ilk.  So should the fact that this is on a blog post, not a ouija board.

Since there are a few folk who don't appear to like graphs  about the weather (or anything else) I have decided to put them at the bottom of the post and some burble about them at the top.  That might induce a bit discipline in my writing since I'll have to explain things in words rather then letting the pictures do the business.

My starting point was to think about Ground frosts (temp below 2^oC) and air frosts (temp below 0^oC).  As I said in the May Weather report there are more days per month in which there are ground frosts than air frosts.  It was interesting that when I graphed the two series of average frost days per month (see Chart 1 below)  the two series were of a very similar shape and both fitted very well to a 4th order polynomial.  It will be interesting to see how June 2016 fits into the pattern as having several days of Easterlies pumping warm wet air from the coast will have reduced the number of frost opportunities more than somewhat.

My next mental rumbling was about when the frost lifted.  This was stirred up by noticing that:

• on 12 June that a puddle on the Plain was still iced over at 10AM; and
• on 13 June shaded grass in Queanbeyan was still frosted at about 10:30.

So I wondered when the temperature rose above 2^oC on a day when it had dipped below that temperature.  I was able to conjure up two measures for this, one showing the latest time at which this occurred  and the other the average time of the barrier breaking.   This is covered in Chart 2 below.  Again the two series are very consistent and fit well to a polynomial: if one was to be picky it could be said that days in April and June warm up a tad quicker than one might expect but since I can't explain that, I won't take the idea any further.

There are no frosts for January or February in my time series (which starts in November 2013) although there are several frost-days (in some years 9 or 10) in those months in the longer term series (starting January 1993) compiled by another recorder.  Plotting those occurrences by year generates a graph that looks like a particularly nasty stage of Le Tour (albeit not with a mountain top finish) so isn't in itself an indicator of climate change.

I have done some further fiddling about on the pattern on temperature change through the day but have yet to work out how to validate what I have done (and, if I've done it correctly, what it means) so watch this space.  I have now thought some more about this and the outcome is Chart 3.  I won't go into the detail of how I calculated it  because:

• life is too short to waste the amount ofyour time that would be needed; and 
• it is possible that I have processed the data so much all meaning has been destroyed

but what I think it shows is a period of strong warming (increase >1^oC per hour) from about 8am to noon and a period of strong cooling (decrease >1^oC per hour) from about 5pm to 10pm.  There is a narrow 'transition phase' of weak warming of about an hour either side of the strong warming and the rest is weak cooling.

I also repeated the exercise restricting it to days on which a frost occurred.  The major difference I picked up from the graph was a longer period of strong cooling in the evening extending into the early morning.  However, I couldn't get the numbers to support that.

Chart 1.

 Chart 2 

Chart 3