What is a probability forecast?
|
|
When forecasting the weather, forecasters are always thinking about the uncertainty of the future weather.
Although we may give a deterministic forecast, like "Sunny tomorrow, high 25 degrees", we cannot be absolutely sure
that this is exactly what will happen tomorrow. Forecasting the weather in probabilistic terms is a way of expressing
uncertainty in the future weather. More importantly, when numerical probability values are assigned to the forecast,
it is a way of indicating the amount of uncertainty in the forecast. Rather than say "chance of showers tomorrow",
we can say "30% chance of rain tomorrow" when we think rain is not so likely, or we can say "80% chance of rain tomorrow"
when we think rain is likely. Assigning a numerical value – a probability – to the occurrence of a weather event
allows us to quantify the uncertainty.
A probability statement always estimates the likelihood of occurrence of a specific event, which must be fully defined
for the forecast to be verifiable. Events are defined by a time period and a specific location.
Probabilities are usually stated either as a percentage, which may range from 0 ("The event will definitely not occur")
to 100 ("The event will definitely occur"), or as a numerical value between 0 and 1. Values between 0 and 100 or
between 0 and 1 represent the different degrees of uncertainty.
Test your understanding of probability statements with the following multiple choice question:
Loading Questions
...
1. Which of the following probability forecasts does not give enough information to be verifiable and why?
Incorrect. Location and time of the event (rain) are specified, along with the probability estimate.
Incorrect. Location and time of the event are specified (Helsinki and "tomorrow", meaning all day.
This is called a "credible interval" forecast because the forecast is the percentage probability
that the maximum temperature will be in a specified interval.
Correct. Location and time are completely specified, but no probability is attached.
The user is left to wonder "How big a risk".
2. You see the forecast "Probability of rain for the Helsinki region tomorrow afternoon is 40%".
Which of the following interpretations is NOT a correct interpretation of the forecast?
This is a correct interpretation of this forecast.
This could happen, but it is not a correct interpretation of the probability forecast.
Unless stated otherwise, the probability statement applies to each point in the forecast area.
This is also a correct interpretation of the forecast and it is one method an experienced forecaster
might use to estimate a probability.
