|
|
|
Introduction
In 2006, I set up a private weather station in Mosfiloti. Since then, it
has acquired vast amounts of data, including rainfall which is relevant
to this website. If we look at the map below, Mosfiloti is just in the
emerald green area to the west of Larnaca and should therefore receive
an average annual rainfall of somewhere between 400 and 500 mm. The
problem is that this map is hardly accurate. I therefore decided to try
and interpret the precipitation data into something more interesting.
However, the four years in question have not received anything like an
average rainfall.
Generally speaking, the average rainfall at any locality is determined
by measuring the quantity over many years. Climatologists often use 30
years. The chart below shows that the variations over a long period of
time are actually very great, although the average becomes relatively
constant. It is evident that I cannot determine an average from four
years of poor rainfall! So, is there any way that I can do so?
Method
I am not a mathematical wizard but I have a slight grasp of a technique
called modelling. This consists of taking whatever suitable data are
available and manipulating them in such a way that they are likely to
resemble the required information, much as a sculptor modelling clay can
make it look like the bust of his subject. In this case, the data I had
available were several decades of the rainfall in the major towns of
Cyprus plus Troodos and Platres. However, from the map, it can be seen
at the rainfall in each place varies very considerably. The nearest town
to Mosfiloti is Larnaca which has a much lower rainfall. I therefore
tried the technique called interpolation. This consists of taking each
set of data and weighting it according to the distance, as the crow
flies, from Mosfiloti. I restricted this to a radius of 60 km. This
seemed to give a reasonable approximation, but many other corrections
were required to cover altitude, topography, vegetation, sea breezes,
katabatic winds and so on. The average rainfall has diminished by about
50 mm over the past 50 years, as may be judged from the above graph.
This is due to climate change, of partially indeterminate cause. I
therefore split the difference and subtracted 25 mm from all the data to
compensate for this. It is emphasised that this is purely
theoretical and cannot be considered as an accurate representation of
the monthly averages in our village. That having been said, it is my
belief that it is probably accurate to within ± 10 to 15%. The secret of
good climate modelling is to compare the theoretical results with the
practical results. Unfortunately, I do not have sufficient practical
results as a control. The chart below shows the short-term correlation
between the theoretical and real monthly averages over the 3 1/2
winters. Note that the pluvial year in Cyprus is always from October to
September. Nevertheless, the overall shape of the real averages
correlates reasonably well with the theoretical, making it look hopeful.
Of course, as more data become available, then it will be possible to
refine the algorithms used in the modelling.
The second chart below shows the annual rainfall over the same 3 1/2 years,
by year. This shows the great variability between each year. It must be
noted that the 2009/2010 column starts at 31 January, which is the reason
for the arrow, because the pluvial year is certainly not finished. To date,
it is on track to be approximately an average year, although this may vary
as time goes on.
The last chart shows the calculated average per month (black with red
border) and every month in the 3 1/2 year period with measured results. This
looks very chaotic because of the semi-random nature of any single month's
rainfall. However, there are a number of points which are worth mentioning.
The 2006 to 7 winter was characterised by two months of reasonable rainfall
shown in magenta. September 2009 (blue) was exceptionally high and
contributed more to the 2009 to 10 winter, in practice. Above all, December
2009 and January 2010 were "good months", shown in purple. It can easily be
seen that the year 2007 to 8 were about across-the-board, May being the best
month.
Conclusion
Simple modelling has enabled a first try at determining average rainfall per
month and per year, with minimal input data. It is believed that this is
reasonably accurate, although it cannot be guaranteed.
It is important to note that these calculations apply only to Mosfiloti and
cannot be used in other localities. Interestingly, the average rainfall for
the whole island is usually within a few percent of the measured
precipitation for the village. This has been observed on to three occasions,
particularly when there has been heavy rainfall in the Paphos area, while it
has been particularly dry in the Famagusta area, such as occurred in this
last winter. This is why the dams in the west have filled up far more
quickly than those in the east of the island.
|