Modulation of mid-Atlantic tropical cyclone landfalls by the Madden-Julian Oscillation

Abstract

Tropical cyclone (TC) landfalls over the U.S. mid-Atlantic region are very infrequent. However, when they do occur, the resulting human and material losses can be severe, as was the case with Hurricane Sandy in 2012. Therefore, it is important to predict these land-falling events as far in advance as possible. In this study, we investigate the relationship between mid-Atlantic TC landfalls and the Madden-Julian Oscillation (MJO), which is the dominant source of climate variability in the tropics on intraseasonal time scales. This is largely accomplished by using a high-atmospheric-resolution ensemble prediction system based on the European Centre for Medium-Range Weather Forecasts (ECMWF) operational model (Project Minerva) to compile the statistics of these rare events, and the velocity potential MJO (VPM) index to define the phase and amplitude of the MJO. We find that at longer lead times (between 14 and 7 days prior to landfall), statistically significant peak landfall probabilities are present during MJO phases 1 and 7 and, to some extent, phase 8. This result is largely supported by observational data. At shorter lead times (between 6-day lead and landfall), phase 1 is strongly favored with some contribution from phases 2 and 3. These findings suggest a potential for extended-range predictions of the mid-Atlantic TC landfall risk based on the phase of the MJO.

Keywords

• A high-atmospheric-resolution seasonal prediction system is used to investigate the link between mid-Atlantic TC landfalls and the MJO
• Between 14- and 7-day lead times, TC landfalls are more likely to occur during MJO phases 1 and 7, which is supported by observational data
• In the short range (between 6-day lead and landfall), MJO phase 1 is strongly favored with some contribution from phases 2 and 3