315 Mean and Intraseasonal Variability Simulated by NCEP CFSv2 during Boreal Winter: Impact of Horizontal Resolution

Monday, 8 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
Shilpa Malviya, Indian Institute of Tropical Meteorology, Pune, India; and P. Mukhopadhyay, R. Phani, A. Dakate, and K. Salunke

This study demonstrates the fidelity of Climate Forecast System version 2(CFSv2) in three horizontal resolution: T62, T126 and T382, during boreal winter. As the Madden Julian Oscillation (MJO) is the major mode of variability during boreal winter, the emphasis of the study is on the fidelity of the models in capturing the MJO variability. CFSv2 shows moderate skill in simulating the intraseasonal oscillation and the skill is sensitive to the resolution of the model.

Boreal winter mean precipitation shows the tendency of the model to overdo the double ITCZ that increases with resolution. 20-100 day band pass filtered rainfall variances also indicate that the dominant variances are overestimated with increase in resolution. The characteristic eastward propagation is not captured by the model at all the resolutions. In an attempt to understand the limitation of the model, composites of specific humidity are analysed which show that the model's moisture profile is rather better represented at lower resolution than the higher ones. Analyses of relative humidity profile as a function of rainfall rates show that all the model resolutions fail to reproduce the lower level moistening prior to deep convection. Furthermore analysis of vertical velocity and moisture convergence bring out the deficiency of the model irrespective of resolutions which is it's inability to capture the shallow convection and thus the multiscale behaviour of MJO. This study therefore indicates that the fundamental moist convective processes do not improve in the model and neither its bias in capturing the intraseasonal variability vis-à-vis MJO just by increasing the resolution. This study can therefore be a basis for improving the moist convective processes with emphasis on the improvement of the shallow convection within boundary layer and also for the inclusion of resolution dependencies in parameterized processes.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner