Atmospheric dynamics and internal processes in CFSv2 model during organization and intensification of BSISO

Abstract

The study finds out Climate Forecast System (CFS) version 2 (CFSv2T126 and CFSv2T382) models’ fidelity in capturing mean flow–eddy interaction, circulation–heating feedback and, the energy conversion processes during organization and intensification of Boreal Summer Intra-Seasonal Oscillation (BSISO) in the backdrop of the mechanism put forward by earlier observation-based study. Ten years of free-run is used to evaluate the models. CFSv2-T126 over-estimates the BSISO intensity from lag –24 and the over-estimation is even more for CFSv2-T382 compared to observation-based study. T126 model underestimates the mean kinetic energy (MKE) related to upper-level easterly wind and the under-estimation is notably more prominent for the strong events. The under-estimation is more for T382 model and the model has difficulty in capturing the upper-level MKE. However, both the model can capture the lower level MKE structure with a slight over-estimation for both the events. At the upper level, MKE to eddy kinetic energy (EKE) conversion for T126 model is significantly weak, and contrary to the observation-based study which shows a decrease in conversion in the subsequent lags. Nevertheless, T382 model shows an increase in MKE to EKE conversion for strong events which agrees with ERA analysis. CFSv2-T126 shows decreasing mean available potential energy (MAPE) to MKE conversion (CA process) for the strong events at the upper level in complete contrast to observation-based study. CFSv2-T382 model can capture the increasing MAPE to MKE conversion as BSISO approaches towards the organized and intense phase. For weak events, CA process is very weak in both models. Both models show a very weak vertical eddy momentum-vertical wind shear interaction (CK3 conversion) for strong events at the upper level. T126 model can capture the conversion for strong events at the lower level, but the magnitude is under-estimated. The process-based analyses of CFSv2 simulation bring out that the model has a significant deficiency in capturing the energy conversions and circulation–heating feedback processes. All these eventually led to lesser fidelity of the model in capturing the organization and intensification of BSISO.

Appendix

Derivation of individual CK terms (equation A2) from equation (A1).

$$ \frac{{\partial \overline{K}}}{\partial t} = \underbrace {{\overline{{\user2{V^{\prime}}\cdot\left( {{\varvec{V}}_{3}^{^{\prime}}\cdot\nabla_{3} } \right)\overline{\user2{V}}}} }}_{{{\text{CK}}}} - \underbrace {{\frac{R}{P}{ }\overline{T\omega } }}_{{{\text{CA}}}} - \underbrace {{\overline{{\varvec{V}_{3} }} \cdot \nabla_{3} \overline{K} - \overline{{{\varvec{V}}_{3}^{^{\prime}} \cdot\nabla_{3} \overline{K}}} }}_{{{\text{BK}}}} - \underbrace {{\overline{{\nabla_{3}\cdot \left( {{\varvec{V}}_{3} \Phi} \right)}} }}_{{{\text{BG}}}}\, +\, {\text{D}} .$$

(A1)

The CK terms in equation (A1) can be expanded as:

$$ \begin{aligned} {\text{CK}} & = \user2{V^{\prime}}\cdot \,\left(u^{\prime}\frac{\partial }{\partial x} +v^{\prime}\frac{\partial }{\partial y} + \omega ^{\prime}\frac{\partial }{\partial p}\right)\left(\hat{i}\overline{u} + \hat{j}\overline{v}\right) \\& = \, (\hat{i}u^{\prime} +\hat{j}v^{\prime})\cdot\left(\hat{i}u^{\prime}\frac{{\partial \overline{u}}}{\partial x} + \hat{j}u^{\prime}\frac{{\partial \overline{v}}}{\partial x} + \hat{i}v^{\prime}\frac{{\partial \overline{u}}}{\partial y} + \hat{j}v^{\prime}\frac{{\partial \overline{v}}}{\partial y} + \hat{i}\omega ^{\prime}\frac{{\partial \overline{u}}}{\partial p} + \hat{j}\omega ^{\prime}\frac{{\partial \overline{v}}}{\partial p}\right) \\ & = u^{{\prime}2}\frac{{\partial \overline{u}}}{\partial x} + u^{\prime} v^{\prime}\frac{{\partial \overline{u}}}{\partial y} +u^{\prime}\omega ^{\prime}\frac{{\partial \overline{u}}}{\partial p}+ v^{\prime} u^{\prime}\frac{{\partial \overline{v}}}{\partial x} +v^{{\prime}{2}} \frac{{\partial \overline{v}}}{\partial y} +v^{\prime} \omega ^{\prime}\frac{{\partial \overline{v}}}{\partial p} .\\ \end{aligned} .$$

(A2)

Rearranging the terms:

$$ {\text{CK}} = \underbrace {{u^{{\prime}2} \frac{{\partial \overline{u}}}{\partial x} + v^{{\prime}2} \frac{{\partial \overline{v}}}{\partial y} }}_{{{\text{CK1}}}} + \underbrace {{u^{\prime}v^{\prime}\frac{{\partial \overline{u}}}{\partial y} + u^{\prime}v^{\prime} \frac{{\partial \overline{v}}}{\partial x} }}_{{{\text{CK2}}}} + \underbrace {{u^{\prime}\omega ^{\prime}\frac{{\partial \overline{u}}}{\partial p} + v^{\prime}\omega ^{\prime}\frac{{\partial \overline{v}}}{\partial p} }}_{{{\text{CK3}}}}. $$

(A3)