LFM Performance and Long-Wave Troughs

The LFM had some very large forecast errors during January and February 1982. Most forecasters in the Eastern U.S. were led down the garden path by a model now exhibiting schizoid behavior (steadily improving performance reverting to "gosh awful"). Many a Winter Storm Watch was issued only to be cancelled as the nonevent came closer. Recent verification of the 36-hour LFM 500mb progs by NMC showed a monthly mean algebraic height error of 100 meters centered over Kentucky for February 1982. (When the verification data is received from NMC, we will attach it to a future issue of the Outlook.) This meant that the LFM was overforecasting development in the Eastern U.S. A new product has recently become available which may help to diagnose and anticipate such errors in the future.