Notes on NWP and the History of Long Range Forecasting

Don Gilman (ret) and Huug van den Dool, NOAA/NCEP/CPC

Legend written by Namias.

• Very little NWP-influence on operational seasonal forecasts (at NMC/NCEP) prior to 1995.

• Beginning in 1995 the coupled ocean/atmosphere model routinely became a tool for the seasonal forecast. NWP NCP

The precursors of NWP ( NCP) are:

1. Pioneering integrations by Miyakoda, Gilchrist, Shukla and others. 1970s, 1980s2. (linear, steady state) Global models describing Rossby wave dispersion from long-lived tropical sources of heat and vorticity. Explained Teleconnections?? 1980 onward3. AMIP runs from 1989 onward (Gates, PCMDI)4. An ocean data assimilation system (Leetmaa et al) for initial conditions of the coupled system..Missing: a proper land initial condition

• The two-tiered approach to ‘Coupled’ Modeling (still widespread)

• Invented to deal with (very) large biases• Can the ‘anomaly’ be salvaged when the

mean is that far off?? (history of bias correction)

• Summer 2004: NCEP is going ‘1-tiered’

Defensive approaches with a profound impact:

Coupled Models at NCEP• Ji, Leetmaa and Kumar (CMP12/14; 1995

• Kanamitsu et al (SFM; 2001

• Saha et al (CFS; 2004

• CMP and SFM had only Pacific Tropical ocean and were 2-tiered. CFS is global, 1-tiered, and at T62L64 (atmosphere), 1/3rd degreeL40 (ocean) resolution. CFS has 2500 years worth of hindcasts.

Issues of skill and ‘consolidation’

• What is the skill of NCP?• What is the skill of NCP vis-à-vis other

methods (like CCA)• How to combine Methods A, B and C?• The pioneering role of seasonal prediction in

hindcasts and ‘a-priori’ skill requirements• Will the shorter range NWP follow with

hindcasts?

Skill assessment of CFS.

OFFicial Forecast(element, lead, location, initial month) =

a * A +b * B +c * C +

…Honest hindcast required 1950-present.

Covariance (A,B), (A,C), (B,C), (A, obs), (B, obs), (C, obs) allows solution for a, b, c (element, lead, location, initial month)

Conclusions

• NWP techniques have entered the arena of seasonal forecasts, from 1995 onward, and are expanding.

• Skill remains modest (predictability?)• Empirical techniques have not yet been retired • Reanalysis (as such NWP related) has boosted

empirical work thru availability of global data sets (1948-present). (Reanalysis was a revolution for empirical work).