Abstract ID: 063
Potential increase in MJO predictability under global warming
Lead Author: Danni Du
University of Colorado, Boulder, United States of America
Keywords: MJO, predictability, global warming, Weighted Permutation Entropy
Abstract: The Madden-Julian Oscillation (MJO) is the leading source of predictability in our climate system on the subseasonal time scale. In this study, we explore and explain the increasing MJO predictability during the past century. We use RMMI to represent MJO.
First, we will show the increasing MJO predictability trend we observed from model ensemble forecasts and reanalysis data. Following the traditional method of using model ensemble forecasts and evaluating with the bivariate anomaly correlation coefficient, we obtained a significant positive trend in MJO predictability for the past century. We then analyzed the MJO in ECMWF coupled climate reanalysis for the 20th century (CERA-20C) using the Weighted Permutation Entropy (WPE) method, which has been proven as a useful tool in analyzing predictability. The higher the WPE, the lower the predictability. We witnessed a consistent decreasing trend in WPE among all 10 CERA-20C ensemble members, which reflects a robust, increasing trend in the MJO predictability.
Then, we will present the MJO predictability change in CESM2 and CESM2-WACCM historical runs using the WPE method. Most historical runs are with a WPE changing trend within the spread of the trends estimated from the control run; however, the distribution of the WPE trends in historical runs shifts to the negative side compared to the distribution calculated from the control run. This suggests that the increasing MJO predictability we observed in the past century is likely caused by the internal climate variability and the external forcing (the global warming).
Next, we will present the MJO predictability change in CESM2 and CESM2-WACCM future projections under the ssp585 scenario. With a much stronger global warming forcing, the distribution of the WPE trends shifts even more to the negative side than the distribution calculated from the control run, which further supports the assumption that global warming can increase the MJO predictability.
Finally, we will explain why there is such an increase in MJO predictability. In both reanalysis data and CESEM2/CESM2-WACCM ssp585 future projection, we noticed that, within a range of 10 days, the sequential amplifying/weakening of RMM1, RMM2 and MJO amplitude, and the organized eastward propagation occur more and more frequently. These regular patterns make the MJO more predictable.
Co-authors:
Aneesh C. Subramanian (University of Colorado, Boulder, CO, USA)
Weiqing Han (University of Colorado, Boulder, CO, USA)
Jeffrey B. Weiss (University of Colorado, Boulder, CO, USA)
William Eric Chapman (NCAR, Boulder, CO, USA)
