13.4
Verification of Hindcasts Made by the Astronomical Climatic Prediction
John C. Freeman, Weather Research Center, Houston, TX; and J. F. Hasling
A forecast of the Cooling Degree-Days (CDD) for several cities was made for the months of June through September of the year 2000. To prepare for this forecast, a hindcast was made for the CDD for 13 cities for 15 years and Heating Degree Days for San Francisco by using the system of prediction called the Astronomical Climatic Prediction (ACP).
The ACP divides the years into active, moderate or inactive years according to criteria based on the orbit of the sun about the center of gravity of the solar system and criteria based on the (long term: 9 to 19 year) periodic properties of the moon about the earth. Active years have 80% warm or cold episodes in the Southern Oscillation, moderate years have 33% of such episodes and inactive years have 15%.
The active, moderate and inactive years are used to separate statistics of the monthly total of CDD for the months of June, July, August and September. For twenty years ending in 1998 and this record was used to make hindcasts of CDD.
The hindcasts were measured against climatology to find a skill score. If the climatological distribution was hindcast, the skill score was 0%.
Four of the years hindcast had skill ranging from -3.1% to +2.6% and seven of the years had skill ranging from 11.9% to 21.3% and three of the years had skills ranging from 27.8% to 33.7% when all the stations were considered. Industry sources tell us that the usual monthly forecast has an average skill of 5.0% to 10.0%.
The skill for each station was also analyzed, (except for Providence, RI that had a skill of 1%) the stations had skill greater than 9.3% and nine stations had skill greater than 15%. The highest skill for a station was Tucson, AZ with a skill of 27.4%.
Session 13, Seasonal Prediction (Parallel with Sessions 11 & 14)
Wednesday, 17 January 2001, 3:30 PM-4:30 PM
Previous paper