Can reanalysis products with only surface variables assimilated capture Madden–Julian oscillation characteristics?

Abstract

The Madden–Julian Oscillation (MJO), as a dominant mode of tropical intraseasonal oscillation, plays an important role in the variability of global weather and climate. However, current state‐of‐the‐art atmospheric circulation models have difficulty in reproducing observed MJO characteristics when forced by observed daily sea surface temperature alone. An important practical question is how much data a model needs in assimilation to reproduce real MJO events? By analysing ERA‐20C and NOAA‐20CR reanalysis data, the authors tried to figure out whether a model could reproduce observed MJO events by assimilating the observed surface signal alone. The phase propagation and vertical structure associated with MJO were compared between the reanalysis data and observations during 1979–2010. A total skill score considering both temporal correlation and spatial standard deviation were defined. The result showed that both ERA‐20C and NOAA‐20CR could reproduce the observed MJO characteristics very well, with the former superior to the latter, regardless of MJO intensity. Thus, a minimum requirement for an operational atmospheric model for MJO prediction is the assimilation of the observed surface signals.

Simulation of monsoon intra‐seasonal oscillations in Geophysical Fluid Dynamics Laboratory models from Atmospheric Model Intercomparison Project integrations of Coupled Model Intercomparison Project phase 5

“…This presents prospects for further research which will be addressed in the future. A recent work suggested that a model could very well capture the evolution and vertical structure of ISV through assimilating the surface variables such as sea level pressure (Cui et al, 2020). The effect of surface data assimilation in improving the representation of MISO in an AGCM will be considered in the future study.…”

“…Unfortunately, poor simulation of the ISV is a fairly generic problem in current general circulation models (GCMs), starting from the atmospheric GCMs (AGCMs; Slingo et al, 1996; Rajendran, Nanjundiah, & Srinivasan, 2002; Waliser, Jin, et al, 2003a; Klingaman, Inness, Weller, & Slingo, 2008; Cui, Wang, Li, & Wu, 2020, among others), followed by atmosphere–ocean coupled GCMs (CGCMs; Fu, Wang, Li, & McCreary, 2003; Fu & Wang, 2004; Zheng, Waliser, Stern, & Jones, 2004; Rajendran & Kitoh, 2006; Seo, Schemm, Wang, & Kumar, 2007; Ajayamohan, Annamalai, Luo, Hafner, & Yamagata, 2011; Sharmila et al, 2013, among others), to the recent model intercomparisons (Lin et al, 2008; Neena, Waliser, & Jiang, 2017; Sabeerali et al, 2013; Sperber & Annamalai, 2008; Xavier, Duvel, & Doblas‐Reyes, 2008, among others). Recently, a number of international research groups carried out a set of climate runs for the Coupled Model Intercomparison Project phase 5 (CMIP5).…”

Interdecadal variability of intensity of the Madden–Julian oscillation

“…Different from shorter-period reanalysis products such as ERA-I and NCEP II reanalysis datasets, only surface observations were assimilated into the assimilation system in 20CR and ERA-20C. Nevertheless, the MJO structure and evolution derived from the 20CR and ERA-20C datasets appear reasonably well (Cui et al, 2020).…”

Enhanced winter and summer trend difference of Madden–Julian Oscillation intensity since 1871

“…To ensure that an MJO intensity index (described below) derived from the NOAA-20C is accurate in representing realistic MJO signals, a comparison is made first with the MJO index calculated using modern observations for a period of 1979(Cui et al, 2020. The observed daily OLR from NOAA (Liebmann and Smith, 1996) and zonal wind fields from National Centres for Environmental Prediction-National Centre for Atmospheric Research (NCEP-NCAR) Reanalysis I (Kalnay et al, 1996) are used for this purpose.…”