U.S. Department of Commerce | National Oceanic & Atmospheric Administration | NOAA Research

Skip navigation

Site navigation:

• About
  • Overview
  • Organization
  • Staff
  • Visiting ESRL
• Contact
  • Points of Contact
  • Staff List
• Research
  • Climate-Weather Connection
  • El Niño / La Niña
  • Polar Processes
  • Boundary Layer Applications
  • Publications
  • Field Programs
  • Observing Facillities
• Data
  • All
  • Gridded Climate
  • Cruise Data
  • Profiler Network Data & Image Library
  • Satellite
• Products
  • All
  • Publications
  • Map Room
  • Plotting & Analysis
  • Experimental Week-2 Forecasts
• Outreach
  • News
  • PSD Research Spotlight Articles
  • PSD Science Handouts
  • Education Resources
• Intranet

 

BOM MJO Forecast Methodology

Steps involved in computing the RMM1 & RMM2:

• Forecast anomalies of OLR, U850, & U200 were obtained by subtracting the hindcast climatology of the respective fields from the forecast.
• The hindcast climatologies were computed from a set of 26-year (1980-2005), once-a-month, 10 ensemble-member hindcasts. (Note: Currently, only 3 of the ensemble members are used to compute the climatologies; later on, all of the 10 members will be used.)
• The hindcast climatology on the forecast date were obtained by a linear interpolation of the present and next months' climatologies, as the hindcasts are available only for the first day of each month.
• The forecast anomalies were then averaged over latitudes (-16, 16).
• The forecast anomalies of the 3 fields were divided by the respective global standard deviations calculated from the NCEP reanalyses.
• The resulting anomalies were projected onto the first two EOF patterns, the latter being computed by Matt Wheeler from NCEP reanalyses.
• The two projection coefficient time series were further divided by the standard deviations of first two PC (also from NCEP reanalyses), to obtain the RMM12.
• To remove the interannual variability, a 120-day time mean of RMM12 is subtracted. (Note: At present, this 120-day mean is a forward time mean of one of the forecasts (the latest one). Later, the 120-day mean will be computed from observations or from the analyses used to initialize the model.)
• Plotted on the phase diagram are RMM12 of ten "ensemble" members (red curves) and the ensemble mean (blue curve). The ten ensemble members consist of the forecast of the most recent day and those of the previous 9 days. An 11-day running averaging was applied to each curve. The curves are plotted for up to 30-day lead time.
• We have a plan to add the "observed" RMM12 (i.e., from the analyses) to the phase diagram as soon as possible.