POAMA Rainfall Skill

Correlation and Brier Skill for Rainfall

Correlation and Brier scores for the POAMA seasonal forecast system's skill at predicting Australian rainfall are presented in the table below, for all months at a lead time of zero. For monthly average forecasts a lead time of zero means the forecast is the average over the first month of model time. For seasonal average forecasts lead time zero indicates the first three months from the forecast start date. Observed rainfall is derived from the National Climate Centre's gridded data, available to the public from SILO.

The correlation coefficient is defined as the ratio of the covariance of the sample populations to the product of their standard deviations. A value of 1.0 indicates a perfect fit between model and observed values. Correlation measures the strength of the linear association between two variables. The figures below show the correlation of POAMA's monthly average rainfall anomaly for each gridpoint with observed rainfall for the hindcast period 1980 until 2006.

The Brier score measures the mean square error of probability forecasts (Wilks, 2006). It is composed of three terms which measure reliability, resolution, and uncertainty respectively (Murphy, 1973). A Brier score of 1.0 indicates a perfect forecast, while 0.0 indicates no skill with respect to the reference forecast. The figures below show the Brier score of a forecast of the probability of above median rainfall based on the POAMA ensemble spread, using the climatological probability as reference.

Table Legend

CC ma: Correlation for monthly averages.
CC 3MRM: Correlation for three month running mean (seasonal average).
BSS Ma: Brier skill for probability of above median mean monthly rainfall.
BSS 3MRM: Brier skill for probability of above median three monthly running mean rainfall.

Month	CC Ma	CC 3MRM	Correlation legend	BSS Ma	BSS 3MRM	Brier skill legend
January
February
March
April
May
June
July
August
September
October
November
December

References

• Wilks, D. 2006: Statistical Methods in the Atmospheric Sciences, Academic Press, 592 pp.
• Murphy, A.H., 1973: A new vector partition of the probability score. J. Appl. Meteor., 12, 595-600.

 

 

For further information: email the POAMA group

© Copyright Commonwealth of Australia 2008
Centre for Australian Weather and Climate Research (CAWCR)
A partnership between the Australian Bureau of Meteorology and CSIRO.
Users of these Web pages are deemed to have read and accepted the conditions
described in the copyright declaration.