Most FAQs
1. How to Read PEAC Experimental Forecasts?
Let's take an example of November 1996's forecast for Chuuk (figure below). Forecasts
for various 3-month seasons over the coming year are shown with an abbreviation using the
first letter in each month (example: "SON" refers to September-October-November 1996).
The actual forecasts are indicated by the gray bars, and give a forecast of the
"standard anomaly" (units are shown on the left axis) expected for rainfall in the season
indicated, compared to the long-term average. Standard anomalies are a statistical measure
which can be converted to an expected range of rainfall when the historical average and
range of actual rainfall are known. A positive
(negative) standard anomaly is associated with above (below) average rainfall. Higher standard anomalies
indicate higher degree anomal ranfall values relative to historical averages. The line indicate skill levels
(the statistical "correlation coefficient", in units on the right axis) for the various forecasts. Generally
speaking, rainfall forecasts are thought to be of use if the skill value is greater than 0.3,
with higher skills corresponding to greater expected accuracy of the forecasts.
Thus from the figure, we can
see that the rainfall in Chuuk is expected to be slightly above average in coming year, and
historical experiments indicate such forecasts are useful since skills in most 3-month periods are basically
have values of 0.3 or greater.
Pacific ENSO Applications Center
Last Revised: October 23, 2006
peac@noaa.gov
Back to Home Page