Volume 38, Issue 7 p. 2994-3012
RESEARCH ARTICLE

Hindcast skill improvement in Climate Forecast System (CFSv2) using modified cloud scheme

Samir Pokhrel

Corresponding Author

Samir Pokhrel

Indian Institute of Tropical Meteorology, Pune, India

Correspondence

Samir Pokhrel, Indian Institute of Tropical Meteorology, NCL Post, Pashan, Pune 411008, India.

Email: [email protected]

Search for more papers by this author
Anupam Hazra

Anupam Hazra

Indian Institute of Tropical Meteorology, Pune, India

Search for more papers by this author
Hemantkumar S. Chaudhari

Hemantkumar S. Chaudhari

Indian Institute of Tropical Meteorology, Pune, India

Search for more papers by this author
Subodh K. Saha

Subodh K. Saha

Indian Institute of Tropical Meteorology, Pune, India

Search for more papers by this author
Febin Paulose

Febin Paulose

Cochin University of Science and Technology, Kochi, India

Search for more papers by this author
Sujith Krishna

Sujith Krishna

Indian Institute of Tropical Meteorology, Pune, India

Search for more papers by this author
Phani Murli Krishna

Phani Murli Krishna

Indian Institute of Tropical Meteorology, Pune, India

Search for more papers by this author
Suryachandra A. Rao

Suryachandra A. Rao

Indian Institute of Tropical Meteorology, Pune, India

Search for more papers by this author
First published: 25 March 2018
Citations: 13

Abstract

Two sets of CFSv2 retrospective forecast experiments are performed to check the model's fidelity for operational forecast usage for the prediction of Indian summer monsoon rainfall (ISMR). The first experiment (Exp1) is identical to the present operational mode of the model. The second experiment (Exp2) includes major changes in terms of the different cumulus parameterization scheme, modified cloud microphysics scheme and the variable critical relative humidity. These changes have already shown enhancement in the seasonal viability of the model in the free-run mode. This study has carried out exclusive hindcast experiments by combining the above mentioned major changes. There is a marked improvement in the spatial distribution of the precipitation and the amplitude of the annual cycle of ISMR. The underestimation of the peak of the annual cycle of ISMR in Exp1 is enhanced by 23% in Exp2. Because of better simulations of clouds and tropospheric temperature gradient, the point of maximum precipitation has migrated northwards from equator (Exp1) to 20°N (Exp2). These improvements also impress upon all the other aspect of the ocean–atmosphere coupled interaction, namely planetary-scale Hadley circulation, air–sea interactions and most of the facets of monsoon teleconnections. The skill of extended Indian monsoon rainfall region (65°–95°E, 5°–35 N) has increased from 0.50 in Exp1 to 0.67 in Exp2 and the same holds true for other regions as well. The skill of Niño3.4 index enhances from 0.58 in Exp1 to 0.67 in Exp2. The dynamical wind shear based monsoon performance indices also show the surge in the skill score. The significant improvement of seasonal skill scores across all the variables clearly shows the dynamical consistency and at the same time establishes the superiority of the Exp2 for seasonal forecast. This work will add new dimension to develop a new genre of monsoon forecasting model.