Subseasonal to seasonal tropical forecasts using a Coupled Linear Inverse Model (C-LIM)

(Experimental NOAA/ESRL PSL and CIRES/U. of Colorado Forecast)

C-LIM VERSION 3.2beta: Interim update.
Resolution improved to 2x2 degree, replaced NCEP R1 winds with CFSR winds, and updates one day sooner.
STILL IN BETA; SOME PARTS INCOMPLETE.

Forecasts and verifications of Tropical Convection, Wind, and SST anomalies:

• Pentads 1-12
• Months 2-9
• Separation of tropical anomalies into coupled and uncoupled fields, 1982-present
• How we make these Tropical forecasts

Experimental forecasts of four key tropical fields, outgoing longwave radiation (OLR), 200 and 850 mb winds, and thermocline depth and sea surface temperature (SST); other variables may become available at a later date. Anomalies are averaged with a 5-day running mean and are relative to a 1982-2011 daily climatology (smoothed with a 31-day running mean). Forecast verification time is the central day of the forecast period. Projected skill is for the pattern correlation of forecast and verification throughout the entire tropical region.

Current initialization and a few selected forecast anomalies of OLR (W/m²), winds (m/s), 20 C isotherm depth (m, blue=positive), and SST (^oC):

One key benefit of the C-LIM is that the eigenvectors of its dynamical evolution operator separate into two distinct, but nonorthogonal, subspaces: one governing the nearly uncoupled subseasonal dynamics, and the other governing the strongly coupled longer term dynamics. This results in a dynamically-based filter that, unlike bandpass or EOF-based filtering, can distinguish between variability having similar spatial structures but very different time scales due to differences between coupled air-sea and internal atmospheric dynamics. We project the tropical data onto these two subspaces, producing a real-time clean split of OLR, wind, and SST anomalies into subseasonal variations including the MJO ("internal"), and longer term variations including ENSO ("coupled"). Hovmullers of the filtered datasets are updated daily; also available is an atlas of the C-LIM filtered data going back to 1982.

For a full description of the C-LIM methodology and a detailed analysis of the previous C-LIM version (v1.0), see:

Newman, M., P. D. Sardeshmukh, and C. Penland, 2009: How important is air-sea coupling in ENSO and MJO evolution? J. Climate, 22, 2958-2977.

Standard disclaimer: these forecasts are experimental. NOAA/PSL and CIRES/University of Colorado are not responsible for any loss occasioned by the use of these forecasts.