87th AMS Annual Meeting

Wednesday, 17 January 2007: 8:30 AM
The Madden-Julian Oscillation and subseasonal prediction
206A (Henry B. Gonzalez Convention Center)
Mathew A. Barlow, University of Massachusetts, Lowell, MA; and D. Salstein
MJO-related predictability of daily summer rainfall at individual stations is assessed for three regions: the US Gulf Coast, Mexico, and Central America. Predictability of extreme events is also examined, particularly in terms of the role of MJO modulation of hurricanes on daily rainfall in the three regions.

Predictability of daily station rainfall is assessed at lead times from 1 to 31 days using both a retroactively calculated MJO index and a real-time, operationally calculated MJO index. As a lower bound on predictability, a very simple prediction scheme is utilized based on assuming that the MJO propagates in a perfectly regular fashion. Statistical significance is evaluated via Monte Carlo analysis. Preliminary results for the stations most strongly influenced by the MJO suggest statistically significant forecast skill extending beyond 3 weeks using the retroactive MJO index and extending to at least 2 weeks using the real-time MJO index.

Extreme daily rainfall events in some parts of these areas are also strongly linked to the phase of the MJO. Hurricane and tropical storm activity in the region is also linked to the phase of the MJO and to extreme daily rainfall events. The predictability of extreme daily events is assessed both directly in terms of MJO influence and indirectly in terms of MJO influence on hurricanes.

Supplementary URL: