For more information about Confidence Intervals and its interpretation, click the link below

About::Confidence IntervalsFor more information about Confidence Intervals and its interpretation, click the link below

About::Confidence Intervals
Plot the largest differences at the FAZ level among multiple runs to
explore variations in runs with different random seeds.

The confidence interval is a measure of uncertainty in the forecast and there are many possible uncertainty sources (e.g., inaccuracies in the input data, the ways in which the model represents reality, the randomness in those processes).

To calculate the confidence interval, model results are first compared with known outcomes, which in this case are 2017 data. The uncertainty captured in those comparisons is then extended into the future to arrive at a set of confidence intervals for a forecast year.

An 80% confidence interval represents a 1-in-10 chance that the actual result will fall below the interval and a 1-in-10 chance that it will fall above. The median value is the best forecast point.