A rotation procedure to improve seasonally varying Empirical Orthogonal Function bases for MJO indices

Abstract

Various indices have been defined to characterize the phase and amplitude of the Madden-Julian oscillation (MJO). One widely used index is the Outgoing Longwave Radiation (OLR) based MJO index (OMI), which is calculated using the spatial pattern of 30-96-day eastward-filtered OLR. The EOFs used to calculate the OMI in observations are prone to degeneracy and exhibit oscillations on the order of 10-20 days, despite initial filtering of the OLR. We propose a simple modification to the OMI that involves aligning the EOFs between neighboring days while retaining the spatial pattern described by the EOFs. This rotation method is implemented as a postprocessing step of the current OMI calculation and cleanly removes the spurious oscillations and degeneracy issues seen in the standard method. A similar rotation procedure can be implemented in calculations of other MJO indices.

Potential predictability of the Madden-Julian Oscillation in a superparameterizated model

“…To isolate the MJO signal from our forecast runs, we use a version of the OLR-based MJO index (OMI), first developed in Kiladis et al (2014) and modified in Weidman et al (2022). The OMI is an EOF-based index that uses the two leading principal components (PCs) of 30-96-day filtered daily tropical OLR to track the strength and propagation of the MJO through time.…”

“…This is to remove high and low frequency signals beyond that of the MJO. The filtered/smoothed OLR is then projected onto a set of seasonally-varying OLR-based EOFs calculated from an independent 40-year SPCAM run, following the EOF rotation procedure in Weidman et al (2022).…”

Potential Predictability of the Madden‐Julian Oscillation in a Superparameterized Model

“…SPCAM is part of the Community Earth System Model project, which is supported primarily by the National Science Foundation (NSF). The ROMI data and analysis code for this work can be found at Weidman (2023). The EOFs for the ROMI were calculated using the mjoindices Python package published in Hoffmann et al (2021).…”