Nellipudi et al. (2026) Seasonal prediction of summer monsoon rainfall over homogeneous regions of india: evolution of monsoon mission coupled climate models

Identification

• Journal: Climate Dynamics
• Year: 2026
• Date: 2026-01-06
• Authors: Nanaji Rao Nellipudi, Suryachandra A. Rao, Deepeshkumar Jain, Prasanth A. Pillai, Ankur Srivastava, Maheswar Pradhan
• DOI: 10.1007/s00382-025-08004-z

Research Groups

• Indian Institute of Tropical Meteorology (IITM), Ministry of Earth Sciences, Pune, India
• National Centre for Medium Range Weather Forecasting (NCMRWF), Ministry of Earth Sciences, Noida, UP, India

Short Summary

This study evaluates the seasonal prediction skill of the upgraded Monsoon Mission Coupled Forecast System version 2 (MMCFSv2) for summer monsoon rainfall across five homogeneous regions of India, demonstrating improved skill over its predecessor (MMCFSv1) in most regions, except Northwest India, and analyzing the underlying teleconnections with Indo-Pacific Sea Surface Temperatures.

Objective

• To assess the skill of MMCFSv2 in simulating Indian Summer Monsoon Rainfall (ISMR) across five homogeneous regions of India and evaluate its improvement over MMCFSv1.
• To explore differences in teleconnections between regional rainfall and tropical Indo-Pacific Sea Surface Temperatures (SSTs) in both models.
• To identify mechanisms responsible for improved regional ISMR prediction in MMCFSv2 and discuss remaining limitations for future model development.

Study Configuration

• Spatial Scale: All-India Region (AIR) and five homogeneous regions of India: Northeast India (NEI), Northwest India (NWI), Central Northeast India (CNEI), West Central India (WCI), and South Peninsula India (SPI). Model horizontal resolution is approximately 38 kilometers. Observational precipitation data (GPCP) is at 1.0° x 1.0° resolution, and SST data (ERSST V5) is at 2.0° x 2.0° resolution.
• Temporal Scale: Hindcast simulations for a 25-year period (1998–2022) for monthly (June, July, August, September) and seasonal (JJAS) mean rainfall. Seasonal forecasts are generated with a two-month lead time, initialized in April.

Methodology and Data

• Models used:
  • Monsoon Mission Coupled Forecast System version 2 (MMCFSv2): Features a Semi-Lagrangian dynamical core for the atmosphere, MOM6 as the ocean component (1440 × 1080 horizontal grids, 75 vertical levels), a New Simplified Arakawa-Schubert (SAS) cumulus parameterization, CICE5 sea-ice model, and NOAA Environmental Modeling System (NEMS) framework as the coupler.
  • Monsoon Mission Coupled Forecast System version 1 (MMCFSv1): Utilizes an Eulerian dynamical core for the atmosphere, MOM4p0 as the ocean component (720 × 410 horizontal grids, 40 vertical levels), and the Simplified Arakawa-Schubert (SAS) scheme for cumulus convection.
  • Both models share a horizontal resolution of approximately 38 kilometers and 64 atmospheric vertical levels, and use a four-layer NOAH land surface model.
• Data sources:
  • Precipitation: Global Precipitation Climatology Project (GPCP) daily data (1.0° x 1.0° resolution) for validation.
  • Sea Surface Temperature (SST): Extended Reconstructed SST (ERSST V5) from NCEI-NOAA (2.0° x 2.0° resolution) for validation and teleconnection analysis.
  • Initial Conditions: NCEP-Climate Forecast System Reanalysis (CFSR).
  • Homogeneous Regions: Boundaries identified using India Meteorological Department (IMD) reports and previous studies.
  • Statistical Metrics: Bias, Standard Deviation, Coefficient of Variation, Root Mean Square Error (RMSE), Anomaly Correlation Coefficient (skill), and Receiver Operating Characteristic (ROC) curves (Area Under the Curve, AUC).

Main Results

• MMCFSv2 significantly improves seasonal mean precipitation skill over the All-India Region (AIR) to 0.72, compared to 0.54 for MMCFSv1.
• MMCFSv2 shows improved seasonal mean precipitation skill in Central Northeast India (CNEI) (0.38 vs 0.09), West Central India (WCI) (0.66 vs 0.60), South Peninsula India (SPI) (0.55 vs 0.45), and a reduction in negative skill over Northeast India (NEI) (-0.18 vs -0.42).
• MMCFSv2 exhibits lower seasonal mean precipitation skill in Northwest India (NWI) (0.21) compared to MMCFSv1 (0.53).
• MMCFSv2 reduces the overall dry bias over the Indian landmass by 3.4% during the JJAS seasonal mean and generally yields lower RMSE values for June, July, August, and the JJAS season compared to MMCFSv1.
• Rainfall over homogeneous regions is strongly teleconnected to Indo-Pacific SSTs; MMCFSv2 realistically represents SST variability over the Southern Indian Ocean (SIO) and tropical Pacific (Niño 3.4, El Niño Modoki Index, Indian Ocean Dipole regions), leading to improved teleconnections with rainfall in most regions.
• MMCFSv2 fails to adequately capture Arabian Sea (AS) SSTs, which is identified as the primary reason for its reduced predictive skill over the NWI region.
• Despite improvements, MMCFSv2 still underestimates rainfall in southwestern WCI and northern SPI, shows a high July and August dry bias, and struggles with extremes along the Western Ghats and deficiencies over NWI.
• MMCFSv2 demonstrates an improved ability to simulate rainfall distribution during extreme (excess/deficient) and normal monsoon years compared to MMCFSv1, although further improvements are needed for grid-to-grid variability in precipitation extremes.

Contributions

• First comprehensive evaluation of the upgraded Monsoon Mission Coupled Forecast System version 2 (MMCFSv2) for seasonal Indian Summer Monsoon Rainfall (ISMR) prediction across five homogeneous regions of India.
• Detailed quantification of performance improvements of MMCFSv2 over its predecessor, MMCFSv1, across various statistical metrics and spatial/temporal scales.
• Identification of specific model advancements (new cumulus scheme, higher resolution ocean model, improved sea-ice model, Semi-Lagrangian atmospheric core, better initial conditions) contributing to enhanced rainfall-SST teleconnections.
• Pinpointing regional strengths and weaknesses of MMCFSv2, particularly its improved skill in most regions but persistent challenges in Northwest India due to inadequate Arabian Sea SST simulation, guiding future model development.
• Extended analysis period (1998–2022) providing insights into recent decadal changes and variability in the Indian monsoon system.

Funding

The Indian Institute of Tropical Meteorology (IITM), a research institute under the Ministry of Earth Sciences, Government of India, fully funded this research.

Citation

@article{Nellipudi2026Seasonal,
  author = {Nellipudi, Nanaji Rao and Rao, Suryachandra A. and Jain, Deepeshkumar and Pillai, Prasanth A. and Srivastava, Ankur and Pradhan, Maheswar},
  title = {Seasonal prediction of summer monsoon rainfall over homogeneous regions of india: evolution of monsoon mission coupled climate models},
  journal = {Climate Dynamics},
  year = {2026},
  doi = {10.1007/s00382-025-08004-z},
  url = {https://doi.org/10.1007/s00382-025-08004-z}
}