Forecasts never that straight forward 

Posted: 8:22 am Tuesday, May 1st, 2018

By Kirk Mellish

We forecasters usually give just a single number for the high and low temperature forecast, and the precipitation chance, wind speed, precipitation amount etc.

In reality meteorologists are aware that there is in fact a range of possible outcomes for these forecast parameters each day, much like the forecast hurricane “cone of uncertainty”.

And like with the cone of uncertainty, that range of uncertainty grows with time. So it’s usually a small range on day one and usually a large one at days 7-10-15.

Screen Shot 2018-05-01 at 8.03.54 AM

Screen Shot 2018-05-01 at 8.05.03 AM

There are various ways to show this uncertainty range on charts using model ENSEMBLE temperature output from numerical weather prediction (NWP) data.

Above is the WxBell chart showing the ECMWF model output, it is called a “box and whisker” chart in statistical analysis and visualization.

Below from weather.us is the same model output visualized in a slightly different way, but still designed to show the variance or range of possible outcomes in the shading:

Screen Shot 2018-05-01 at 7.57.19 AM

Screen Shot 2018-05-01 at 7.55.54 AM

This applies to precipitation as well, there is always A RANGE in the “chance” or “likelihood” of precipitation AND the amount:

Screen Shot 2018-05-01 at 7.56.56 AM

Note in this example the rain/thundershower chance on Saturday is 30-80%.

My 5-day forecast is always available online here.

MY forecast is OFTEN different from other forecasters and from any given model or APP.

Follow me on Twitter @MellishMeterWSB.

0 comments