A COASTAL TRANSITION ZONE EXAMPLE:  GEORGIA CD5 PRECIP

     Central Georgia Climate Division 5 has shown the largest observed precipitation swings in a state with perhaps the most severe experience of the 1997-present Southeast US drought.  A more detailed version of this January 2000 beginning precipitation prediction chart was presented and also handed out August 8, 2000 at the AASC Annual Meeting in Logan, Utah.  DynaPred's presenter indicated that substantial relief for three months July-August-September should be forthcoming.  Falling mainly in August and September, actual observed three months accumulation (12.79 in.) deviated from the sum of the predicted monthly averages by 0.00 inches total--admittedly unusually accurate at midrange.  This example is especially noteworthy because the September observed precipitation included two hurricanes that helped bring 7.5-9.0 inches to the three Georgia CDs closest to the coast.  Adjoining CD#5 received 5.6 inches in September 2000 with the pattern and cumulative measurements well captured by the DP model.  The last updated observed chart point is December 2006 (200612).  CY2003 prediction was described at bottom of this page beforehand and the Georgia CD#5 line chart is clickable or link through the above navigation button.  Source data is courtesy National Climatic Data Center's state climate divisions series, also known as the 'cirs' or TD9640 series. 
     20040210.  With 2004 being into the fifth year of blind prediction, we note from the chart that 2004 predicts generally below average for all but the July-August months coinciding with the 30 year average peaks.
     20050211.  With 2005 being into the sixth year of blind prediction, we note CY2005 predicts to be generally below average while retaining overall seasonal precipitation pattern.  Similarly, into the seventh year, CY2006 predicts wet summer and tropical storm period fall after first half year expected dry.
     20070115.  Completion of 2006 and posting of updated data from 200001-200612 represents 7 years forward of blind prediction as first presented at the AASC Annual Meeting in Logan, Utah.  Georgia Climate Division 5 precipitation will be retired with this 200612 posting.

ARCHIVAL comments:

     Calendar Year 2001 predicted to be volatile with wet highs and quite dry lows, but turned out to be average accumulation with a decidedly dry background tone.  Beginning with the hydrologic year in October, measured precipitation began following the longer term dynamics of Prediction2, notably low as characteristic of CY2002.

     CY 2002 predicts dry: P1=36.27, P2=38.73, (P1+P2)/2=37.5 inches.  Since prediction began 200001, nearby climate dynamics attuned Prediction 1 had indicated above average accumulations.  Similarly, Prediction2, attuned to dynamics of whole period of record has indicated lower than average precipitation.  Since October 2001, the actual observed has more closely tracked this lower precipitation path.  In 2002, both P1 and P2 predict lower than the 45.34 inch average of years 1970-99.  Months 9-12 precipitation exceeded the 1970-99 average by almost 6.5 inches.  Annual precipitation then totaled  45.98 inches, or 0.64 inches above the 30 year average.

     Please click on Georgia CD5 on the above tool bar to see updated chart through last available month with cumulative precipitation shown for CYs 2000, 2001, cumulative beginning to June (200206), and CY 2002 through June and December 2002 (200212). 

     CY2003 predicts to be 89% of the 1970-99 average of 45.34 inches. The individual predictions are PNear = 40.20, PWhole = 40.45, and (PN +PW)/2 = 40.33 in.  Predicted distribution is dry first 4 months, wet middle 4 months, dry last 4 months, with the nearby climate dynamics and whole range climate dynamics tracking well the first 2/3 of 2003.  In September and October, the longer term dynamics would tend to run above the 30 year average, implying some major storms extending the wet summer accumulations but with a dry undertone before a dry November and December.  Most recent data point entry is January 2004 (200401).

     NOTE:  The post 200301 predictions are an extension beyond the announced 3 year accuracy range.  During the 2000-2002 time period DynaPred's Average Predictor was 2% under the observed accumulation and 1970-99 average was 7% over actual observed.  Prior to the wet last 1/3 of 2002, these percentages were more in the range DynaPred Average 1% and 30 year Average 10%.  Similarly, with 2004 being the fifth year of blind prediction, we note from the chart that 2004 predicts generally below average for all but the July-August months coinciding with the 30 year average mid-summer peaks (comment 20040210).