CPC Experimental Long-Lead Outlook

Summary of Forecasts

March 1996

• Dynamical methods: The standard and the new Lamont Doherty Earth Observatory models predict continuing cold conditions through winter 1996-97.The NCEP coupled model calls for some weakening of the current cold SSTA east-central Pacific between now and boreal fall 1996. The Australian BMRC low order coupled model forecasts current cool Ni�o 3 to switch to warm for the second half of 96, further strengthening by late winter and spring 97. The Oxford coupled model calls for cool SST to continue for remainder of 96. The COLA coupled model predicts dissipation of cool SST conditions by summer 1996, warming to >1C by winter '96-97.

• Statistical methods: CDC's inverse modeling predicts dissipation of below normal east-central Pacific SST by summer '96, with possible positive anomalies emerging in the southeastern tropical Pacific. The Australian BMRC non-linear analogue predicts somewhat decreasing SOI by May 1996. UCLA's SSA-MEM predicts near to slightly below normal SST through '96, with near normal SOI. JPL/Utah State's CSSA/MARS "analog" system predicts a La Nina period peaking in early 1997. Univ. of British Columbia's neural net model predicts a warm episode for Ni�o 3 for fall-winter 96-'97. NCEP/CPC's CCA predicts current below normal Ni�o 3.4 (120�-170�W) SST dissipating by June and becoming moderately warm by Jan '97. NOAA's constructed analog predicts dissipation of cool Ni�o 3.4 SST by late spring 1996, warming to ~0.5�C for fall and winter 1996-97. NCEP's 4-model consolidated forecast projects cool SST through 1996, switching to warm after March 1997.

• Dynamical methods: The NCEP coupled model forecasts cool in northern tier for spring and summer '96.

• Statistical methods: NCEP/CPC's optimal climate normals scheme predicts warm Pacific Northwest, Northeast, Mid-Atlantic, cool Florida panhandle for Apr-May-Jun 96; warm Southwest and much of eastern half Jul-Aug-Sep. NCEP/CPC's CCA predicts positive anomalies parts of Texas, the Southwest, far West and Hawaii Apr-May-Jun, warm California, Mid-Atlantic, southern Alaska, Hawaii Jul-Aug-Sep. NCEP/CPC's screening multiple regression predicts above normal parts of southern tier and Maine, cool in Northwest and upper Great Lakes for Apr-May-Jun 96; above normal Southwest, part of Southeast, Maine, Montana in Jul-Aug-Sep. The Environment Canada CCA predicts cold near and west of Hudson Bay, warm British Columbia in Apr-May-Jun 96.

• Dynamical methods: The NCEP coupled model forecasts some enhanced wetness Great Lakes for Apr-May-Jun.

• Statistical methods: NCEP/CPC's optimal climate normals predicts wet in Pacific Northwest Apr-May-Jun '96; wet Michigan and Great Basin, dry North Carolina Jul-Aug-Sep. NCEP/CPC's CCA calls for wetness in Pacific Northwest, near normal Honolulu for Apr-May-Jun 96; dry central Alaska, wet Hilo for Jul-Aug-Sep. NCEP/CPC's screening multiple regression predicts dryness in south-central Texas and parts of eastern seaboard for Apr-May-Jun 96. The Environment Canada CCA calls for wet along east coast of Canada and central British Columbia, dry Saskatchewan for Apr-May-Jun 96.

• Dynamical methods: The Scripps/MPI ECHAM3.2 atmospheric GCM indicates drier than normal conditions in the Nordeste of northeast Brazil for boreal spring 1996.

• Statistical methods: The Univ. of Wisconsin multiple regression and discriminant analysis forecasts call for near-average rainfall for the Nordeste for Mar-Apr-May-Jun 96. The UK Meteorological Office multiple regression and discriminant analysis forecasts indicate dry to average rainfall for the Nordeste for Mar-Apr-May. NCEP/CPC's CCA predicts weak anomalies at most tropical Pacific islands this boreal spring, and some enhancement of rainfall at Johnston Island in summer.

• Statistical methods: The CSU LAD regression approach (see previous Bulletin issue) forecasts slightly negative Atlantic storm anomalies for most aspects of 96 tropical storm activity. The FSU Poisson and tropical/baroclinic approach (see previous issue) predicts approx. average 96 season for Atlantic hurricanes, somewhat below average for intense hurricanes. FSU's Bayesian probabilistic model (this issue) indicates 1.25 (~60% of normal) intense Atlantic hurricanes for 1996. The Australian BMRC SOI-based regression (see previous issue) calls for near to slightly below normal Australian region storm activity for the southern fall 1996 season.