Extended-Range Predictability of Midlatitude Weather Patterns due to Recurving Tropical Cyclones

Many errors in extended-range forecasts of midlatitude weather patterns have been found to originate from the tropics. Tropical sources of errors are related to convection and the organization of convection by the Madden-Julian Oscillation (MJO) and convectively-coupled equatorial waves. The role of tropical convection as a source of errors in midlatitude forecasts has been predominantly examined during winter periods when the midlatitude jet stream is strong and the jet shifts equatorward such that there are dynamic and thermodynamic linkages between the tropics and midlatitudes.

Based on several case studies, tropical cyclones that recurve from the tropics into the midlatitudes have been related to high-impact weather over downstream midlatitude regions. Furthermore, recurving tropical cyclones have also been identified as sources of errors and uncertainty in midlatitude forecasts. Although the midlatitude jet stream shifts poleward and is weaker in summer, a recurving tropical cyclone increases the linkage between the tropics and midlatitudes. Tropical convection and related impacts are brought to higher latitudes and upper-level outflow from the tropical cyclone can merge with and increase the intensity of the midlatitude jet stream. The purpose of the present study is to extend case studies of tropical cyclone impacts on predictability over the midlatitudes and establish the overall significance of such impacts. In this context, predictability is examined based on spread among ensemble members.

Ensemble and deterministic extended-range predictions of synoptic-scale fields from the National Centers for Environmental Prediction Global Ensemble Forecast System (NCEP/GEFS) Reforecast-2 are used to examine predictability in forecasts over the Northern Hemisphere midlatitudes. The once-daily (at 0000 UTC) Re-forecast-2 data are available from 1985 through the present and contain a deterministic forecast, control forecast and 10 perturbation members. This data set allows for a general examination of predictability because it is produced with a fixed model and uses re-analysis data for initial conditions. Representative amplitudes and spatial patterns in predictability are established in association with recurving tropical cyclones. Using the long period of forecasts provided in the reforecast data set, climatological values of predictability and uncertainty are computed. The forecast attributes associated with recurving tropical cyclones are statistically compared to climatological values to establish that recurving tropical cyclones are associated with periods of reduced predictability in downstream midlatitude weather patterns. Furthermore, the tendency for predictability to be reduced relative to specific regions or flow patterns is examined.