Instructions on the interpretation of the forecast diagramms

At the example of a past temperature forecast we want to illustrate how the forecast diagrams are to be interpreted. Consider the forecast issued on 12 November 2009 for the week from 23 November (Monday) to 29 November (Sunday).

The height of the colored columns indicates the probability that the mean temperature of the forecast week (i.e. the 7-days-mean) will be colder than normal (blue), rather normal (green), or warmer than normal (red). In the example shown above, chances for a cold week are very low (only 2% probability). Temperatures in the central category, i.e. close to the climatological mean, are with 20% rather unlikely as well. It is much more probable (almost 80%) that the forecast week will be warmer than what is considered normal for the season.

But what do we exactly mean by a "warm", "normal", and "cool", week? These three categories always need to be interpreted with respect to climatology. They are defined such that each of these categories has been observed with the same frequency over the past 18 years, i.e. they are equiprobable (33.3% each).

The temperature value which is required to call a week "warm", "normal", or "cool", of course strongly depends on the season considered and needs to be determined individually for each forecast week. For the example shown above, i.e. for end of November, a week would be referred to as a ?warm week? if the mean temperature exceeds 4.5°C. A week would be called ?normal? if the average temperature is between 2 and 4.5°C, and it would be considered as a cold week if the mean temperature is lower than 2°C. In the forecast diagrams, these values are displayed on the lower axis. For the region ?Nord- und Ostschweizer Mittelland? (Swiss Plateau), the numbers refer to the average of the measurement stations in Basel, Berne and Zurich; for the region ?Suisse romande? (Western Switzerland) they refer to the station in Geneva, and for ?Sud delle Alpi? (south of the Alps) the refer to Lugano.

By the way, the forecast of this specific example has been a hit. A mean temperature of 8.0°C has been observed for the predicted week. This value is clearly in the upper category, which has been predicted to be the most likely category.

The interpretation of precipitation forecasts is analogous to that of temperature forecasts. Again, three climatologically equiprobable categories are considered ("dry", "normal", "moist"), which are now defined with respect to the weekly precipitation sum.

Finally, note that forecasts stretching beyond a few days into the future are by their nature associated with a high degree of uncertainty. It can always happen that unpredictable and chaotic weather events distort the expected mean tendency. The prediction uncertainty is indicated by the black error bars on each of the columns. We generally recommend being cautious when making decisions on the basis of these forecasts.

More information on the scientific background of extended range forecasts is provided here

Technical background