Wednesday 19 June 2013

the art of prediction

The relative humidity was over ninety per cent last night, and the Systems Administrator said that the temperature need only drop a couple of degrees overnight for there to be fog in the morning. Sure enough, a thin veil of grey cloud hung between me and the wind turbine on the next farm, when I pulled up the bathroom blind.

That was about the most reliable forecast I received for today.  Yesterday, when I checked the Met Office five day Wednesday forecast for Colchester, temperatures were set to rise over twenty five Celsius, and I got excited about the prospect of lunch in the garden, sitting in the shade of the parasol and looking at the roses (which are still not fully out and must be running a good month late).  By lunchtime yesterday the forecast for today was for rain all day, and this morning the forecast maximum temperature for Wednesday had slipped to a modest high somewhere in the low twenties, with rain forecast in the afternoon, which never materialised.

The newspaper reports on the conclusions of the weather summit down in Exeter have been odd, to say the least, with several running front page headlines on how we can expect a decade of washout summers.  The weather man I heard interviewed on the radio yesterday evening said they hadn't reached any firm predictions, merely discussed some possible scenarios, one of which was for a series of poor summers due to the jet stream being in the wrong place.  All they agreed was where future research efforts should be directed.  I'm not even convinced their existing models can forecast the jet stream, given they can't forecast the weather for north Essex twenty-four hours ahead.

In my City days I did notice that most economic and company forecasts seemed to consist of taking whatever the trend had been for the past two or three years, and extrapolating it into the future. The economists and analysts would have denied that they were doing that, and had elaborate models, none of which seemed to work particularly well.  Academic research (which I can't reference as I should in a proper essay, because I can't remember who did it, but I expect I heard it somewhere on the BBC) confirms that most economic forecasts are slightly worse than useless. That is, their chance of being wrong is a little higher than if the forecasts were invented by a non-economist.  The researchers put this down to the economists having cognitive bias based on their existing theories which made them unwilling to consider new data that conflicted with those theories.  I suspect it is simply in human nature to believe that whatever has been happening recently is the norm, and will persist into the future, be it a banking boom, a slump, or wet summers.

Going to release the chickens from their house into their run I was hit by the blast of muggy air, but above me  a skylark was singing over the lettuce field, while in the distance I could hear a cuckoo, and when later I went out I disturbed a flock of goldfinches in the lane.  You are never very far removed from human activity here, what with the lettuce farm, the military ranges, the ever-present rumble of traffic from one direction or another, and the rumble of planes overhead, but the wildlife carries on regardless.  I sprinkled a can of water on the greenhouse floor as the heat of the sun intensified, to try and cool it down for the robins, but the babies were looking rather seedy when I went out just now to water the plants in there.  I fear that in this case the wildlife would have done better to stick to the wild, and nest in a bush like the other birds, or at least choose the shed instead of the greenhouse.

No comments:

Post a Comment