sd = standard deviation. A measure of the spread or range in an average (mean) of a group. In a group of model predictions that follow a normal distribution, approximately 67% of the models’ predictions fall within 1 standard deviation of the mean.
One example of a model that calculates sd is the NBM (model blend). The NBM is an advanced, statistical outcome-weighted model based on a group of over 20 other models.
The NBM calculates the mean (average) temperature of the group of models that comprise it. It also calculates the spread or range of the models that comprise it, expressed as a standard deviation. A higher spread (standard deviation) indicates higher uncertainty.
As an example, you can think of a forecast mean high temperature of 64º with a standard deviation of 2.0º as being a forecast high of 64º± 2º. That means that the temperature could be 66º or 62º as a likely range with 67% of the models forecasting somewhere in that range. It could also be outside that range, with 33% of the models forecasting outside that range.
With NBM high temperatures, a sd of 1.5º is typical. Anything more than an sd of 3º suggests increased uncertainty and lower confidence in the mean value.