By Philip Eden
Last spring saw a flood of stories predicting drought and water shortages which never quite seemed to materialise as the doom-laden tones predicted.
May has been an extremely wet month in the UK
The preceding winter had been, we were told, "the fourth driest since records began".
Now the warnings are here again, alongside the hosepipe bans, and threats of even more drastic drought measures to come.
And the country's first drought order - for Sutton and East Surrey Water - starts on Saturday.
Yet to find a year drier than 2005 for most of the UK, there is no need to delve back into prehistory; 2003 fits the bill, as do 1997, 1996, 1991 and 1990.
And now, looking out into the southern England skies, it appears that rain clouds have taken up residence.
Current statistics show that May has already dumped more than double the normal amount of rain over England and Wales as a whole. In fact, we are well on course to register the wettest May since 1983, and it may turn out to be one of the half-dozen wettest since records began more than 300 years ago.
With statistics, it is said, you can prove anything. And weather is no exception, making it easy for water companies to define regions and timescales which suit their own agendas.
Tie this to the bandwagon of global warming, and you have a truly explosive news cocktail.
Let us go back to last summer for a graphic example of how statistics are selected to tell the stories which water companies and their ally, the Environment Agency, want us to hear.
The preceding winter and spring, we were told, had been exceptionally dry.
The period chosen to demonstrate the rains' recalcitrance ran from November 2004 to June 2005.
Wherever the sound of raindrops is absent, it seems, the sound of spin will quickly fill the void
During this period, rainfall was 77% of the long-term average for the same eight months. Not hugely down, perhaps; but a quarter of the expected total was missing, enough to make this November to June the driest since 1975-6.
But what happens when you decide to begin the period in October rather than November, including, logically, the first month of the "gathering season", when water begins to collect in reservoirs and aquifers after the summer?
October 2004 was wetter than usual. So rainfall for the nine-month period works out at 89% of the long-term average.
Just changing the measurement period reduces the shortfall from 23% to 11%.
A favourite phrase of the meteorological spin-meisters is "driest for xx years" - the bigger the value of xx, the greater the feeling of stress they can engender.
But go back to the data, and another interesting story of selective statistics emerges.
For the period from November 2004 to February 2005, xx equals 42 - satisfyingly remote for half the population, including virtually everyone left working in the media.
But by adding a month, citing November to March, xx falls to 29. If you go in the other direction and strip a month away, November to January is merely the driest for 16 years; much less ominous, much less newsworthy.
Frugality with water is no bad thing - it is perhaps something we will all have to adopt as the decades roll by
The present dry spell began in November 2004, so to make the statistics look more extreme, the water industry analysis compares the dry spell only to other 15-month periods that begin in November.
If you compare it with 15-month periods beginning in any month, 1995-97 wins hands down every time.
But on that occasion, below-average rainfall persisted from March 1995 to October 1997 inclusive - a period of 32 months, so we've got a long way to go to emulate that.
Another temporal trick concerns "return periods" - how long it is likely to be, given what we know about long-term weather statistics, before such a dry period turns up again.
The return period for the infamous eight months from November 2004 to June 2005 is seven years.
Throw in October, and it comes down to a distinctly un-newsworthy 2.5 years.
The same kind of trick can be pulled to a more limited extent with geography.
To say "the south of England has been exceptionally dry" is not really true.
For that, you need to define the area more closely; and so you do. Concentrate on a narrow band from south Wiltshire to west Kent, add south Cornwall, and you can legitimately claim that 2005 was the "driest year in a decade".
Look more locally still, to selected points in the Oxford/Reading/Basingstoke area, and you can even achieve a "driest year since 1921".
Wherever the sound of raindrops is absent, it seems, the sound of spin will quickly fill the void.
I am not arguing that parts of Britain have not had drier weather in recent months, even in the most recent few years, than they might expect.
In the south-east, some reservoirs are genuinely depleted, restrictions on use are inevitable in the short term.
Wet weather now will not be of much help in the needy counties of Kent, Surrey, Sussex and Hampshire.
You will only carry the general public with you... if you can demonstrate that the percentage of water lost through leakage is declining
The cumulative shortfall in this region over the last 17 months is 14 inches (36cm) of rain ; and an extra inch or two during a wet May will not be enough to compensate.
But has the weather been exceptionally dry enough that the water companies should act as if they are taken by surprise, as train operators apparently are sometimes when autumn leaves have the temerity to fall?
My suspicion is that at times the companies talk up the drought in order to distract attention from their own failings, and to persuade domestic consumers into frugality - a measure which helps ensure adequate supplies for the water companies' commercial customers.
Frugality with water is no bad thing. It is perhaps something we will all have to adopt as the decades roll by and a million or so extra homes roll into the South East of England.
But better planning and more investment in infrastructure would be no bad thing either.
My message to the water industry is this: you will only carry the general public with you during times of shortage like this if you can demonstrate that the percentage of water lost through leakage is declining, and you can only do that through proper investment in the infrastructure.
You should also make it clear that the domestic user is not being targeted unfairly, and that commercial and municipal customers suffer restrictions too.
And, please, stop the spinning.
Philip Eden is an independent meteorologist who forecast for many years on BBC Radio Five Live. He runs the websites climate-uk.com and weather-uk.com