Jan 2, 2013

December Snow Forecast Results & New Locations Added to the Trends

Recently I added three new locations to the Trend output. (thanks for the inquiries gentlemen, you know who you are)

Dayton, OH
Davenport, IA
Seattle, WA

These locations have now been added to the snow forecast outputs as well. Also, with the close of December I've put together the results of the December snow forecast. They can be accessed by clicking on the "Results" link next to the "December" map link on the bottom right side of the maps webpage linked below.

The result columns can be sorted by clicking on the table headers. The Error slider will list the stations based on the criteria selected.

For anyone who checks the results out, I would be very interested in your thoughts in regards to accuracy, presentation, etc. For example, is six inches too much of a difference between forecast and actual to give the forecast a passing grade? How many Days With Snow from actual would be considered a fail? Is there another category of analysis I should be presenting? This use of the theory is a bunch of crock and should be halted at once.

All feedback/ideas are welcome.

Something to note with precipitation forecasts for Jan and Feb. Something I call the 'cycle ratio' of the algorithm updates Jan-11. Consider it a new initialization of the modeled precip for future cycles. What this actually means is that the numbers that are forecast now will likely change on Jan-11. The next 'cycle ratio' update occurs Mar-6.