Tercile and quintile probabilities explained
How below-normal, near-normal and above-normal probabilities are derived from ensemble members and hindcasts, and how to read the tercile summary map.
Categories from the hindcast
A tercile forecast divides the model’s possible outcomes into three equally likely bins: below normal, near normal and above normal. The boundaries are set from the hindcast: for each start month, lead and grid cell, all qualifying hindcast members from 1993 to 2016 are pooled and the 33rd and 67th percentiles are taken as thresholds. Quintiles work the same way with the 20th and 80th percentiles, and the median exceedance uses the 50th.
Counting members
The probability of a category is the share of forecast members that land in it. If 30 of 50 members are above the upper threshold, the upper-tercile probability is 60%. Lower categories use a strict “less than” and upper categories a strict “greater than”, so a member exactly on a threshold counts as near normal. The three tercile probabilities for a cell always sum to 100%.
Reading the tercile summary
The tercile summary map condenses the three probabilities into one colour. A cell is coloured for above or below normal only when that outer tercile exceeds 40% and also exceeds the opposite outer tercile. Everything else is neutral. Neutral therefore covers two different situations: the model genuinely favours near-normal conditions, or the members are split with no clear lean. Inspect the individual tercile probabilities when the distinction matters.
The baseline is not zero
A tercile probability of 33% or a quintile probability of 20% means the forecast is no different from climatology. The legend marks this baseline. A 45% chance of an above-normal tercile is a modest lean; a 70% chance is a strong one for a seasonal forecast. Probabilities near 100% are rare and usually appear only in the tropics or in SST.
How this site derives them
Probabilities here are computed locally from individual forecast and hindcast members downloaded from C3S, following the documented C3S method with empirical quantiles. Member counts follow C3S guidance on qualifying members for each centre. The implementation is faithful to the published method but is not guaranteed to reproduce every pixel of the charts on the Copernicus site. See the C3S product descriptions for the reference definitions, and why a probability is not forecast skill before treating any percentage as confidence.