146 Skill of GEFS Medium-Range Reforecasts for Rossby Wave Breaking Associated with Extreme Precipitation in the Central and Eastern U.S

Monday, 23 January 2017
4E (Washington State Convention Center )
Benjamin J. Moore, SUNY, Albany, NY; and L. F. Bosart and D. Keyser

Skill of 1–16-day forecasts from the NOAA global ensemble forecast system (GEFS) reforecast dataset is assessed for a 1985–2015 cool-season (September–May) climatology of Rossby wave breaking (RWB) events over North America associated with extreme precipitation in the central and eastern U.S.  It is found that the skill of 250-hPa geopotential height forecasts over the U.S. for RWB events is on average lower than that for the GEFS-based climatology, particularly at lead times beyond 5 days. Furthermore, as lead time increases the GEFS tends to increasingly underforecast the amplitude and overforecast the phase speed of potential vorticity (PV) streamers (i.e., elongated troughs) linked to RWB, indicating a tendency to underforecast the strength of RWB.  Consistent with the errors in PV streamer amplitude and phase speed, the GEFS tends to underforecast moisture flux and precipitation amounts downstream of PV streamers.

Composites of 250-hPa geopotential height error and spread anomalies, calculated relative to the GEFS climatology, for forecasts initialized 6 days prior to RWB onset demonstrate the growth and downstream propagation of anomalously high error and spread across the North Pacific and North America in conjunction with the formation of a well-defined Rossby wave train.  In the composites, error and spread anomalies propagate with the group velocity of the wave train.  Moreover, error and spread increase markedly over North America in conjunction with RWB.

The RWB cases are stratified according to the root-mean-square error for 250-hPa geopotential height forecasts over the U.S. initialized 6 days prior to RWB onset, and the bottom and top quintiles of the distribution, designated as “high-skill” and “low-skill” cases, respectively, are compared. Reanalysis-based composites indicate that, relative to the high-skill cases, the low-skill cases feature a higher-amplitude Rossby wave train across the North Pacific and, moreover, greater poleward moisture flux and stronger diabatic ridge amplification over the eastern North Pacific. GEFS-based composites for the low-skill cases indicate that the ridge amplification is significantly underforecasted, culminating in a forecast wave pattern over North America that is too low in amplitude and too progressive. These errors in the forecast wave pattern directly correspond to large errors in the forecast precipitation over the central and eastern U.S.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner