The role of warm conveyor belts for medium-range forecast error growth

Warm Conveyor Belts (WCBs) are moist air streams that ascend from the warm sector of extratropical cyclones into the upper troposphere. Cloud-condensational processes along their ascent lead to the release of latent heat which further accelerates the ascending motion and results in cross-isentropic transport of lower-tropospheric air with low values of potential vorticity (PV) to the upper troposphere. This diabatically enhanced outflow intensifies the PV-gradient along the dynamical tropopause, accelerates the jet stream and ultimately modulates the upper-tropopsheric Rossby wave pattern through ridge amplification and downstream development. In numerical weather prediction (NWP) models, it is crucial to correctly represent this interaction of WCBs with the large-scale flow, as mis-representations thereof can lead to poor forecast skill in downstream regions. At the same time, this multi-scale interaction poses a major challenge for global NWP-models, as the governing moist-dynamical processes in WCBs are parameterized and introduce uncertainty into the forecast.

In this study, we investigate the role of WCBs for the error growth in medium-range forecasts over the North Atlantic in a systematic way. This is done by analyzing the relationship between commonly used error metrics of variables that characterize the large-scale flow and imprints of WCBs detected with trajectory analysis in three years of operational ECMWF ensemble forecasts.

Forecasts with high WCB activity are on average characterized by an amplified Rossby wave pattern and anticyclonic flow anomalies downstream. We find that the forecast skill is generally reduced when the WCB-activity is high, and that the WCB-activity is particularly increased around forecast times when the error growth is largest. To establish a relationship between the occurrence of WCBs and error growth, composites of normalized forecast error fields centered on WCB objects are computed. The composites reveal that anomalously high forecast errors associated with mis-representations of the Rossby wave pattern emerge along and downstream of WCB ascent and outflow. While there is considerable case-to-case variability in the mid-tropospheric error patterns, the structure of the errors associated with the upper-level jet stream is robust.

The WCB-relative errors are largest when the WCB-event occurs on forecast days 3-5, whereas WCBs at earlier leadtimes have a substantially reduced impact on the forecast errors, pointing towards an amplification of pre-existing errors by WCBs. Future work will attempt to further distinguish between situations in which WCBs amplify forecast errors, and in which WCBs are the source of forecast errors.