Improving NCMRWF Unified Model Rainfall Forecasts over India through IMERG-Calibrated IMDAA Reanalysis

Accurate rainfall forecasts from numerical weather prediction systems are critical for disaster risk management over the Indian subcontinent, yet the National Centre for Medium Range Weather Forecasting (NCMRWF) Unified Model (NCUM) operational forecasts continue to exhibit systematic biases that limit their regional utility. Since the Indian Monsoon Data Assimilation and Analysis (IMDAA) reanalysis shares the same dynamical core and physics as NCUM, IMDAA biases, once corrected against an observational reference, are expected to be transferable to NCUM forecasts, a hypothesis this study evaluates through independent forecast verification. This study calibrates IMDAA against Global Precipitation Measurement (GPM) Integrated Multi-satellitE Retrievals (IMERG) using four statistical bias correction (BC) methods over 1998 to 2017 and validates over 2018 to 2020 during the Indian Summer Monsoon. Raw IMDAA shows drizzle bias, underestimated variability, and near-absence of Very Heavy rainfall events (> 115.5 mm day−1). All methods reduce these biases, with non-parametric methods generally outperforming the parametric ones. Application to NCUM forecasts for June-September (JJAS) 2023 at 1, 3, and 5 day lead times improves the detection skill for Heavy (64.5 to 115.5 mm day−1) and Very Heavy rainfall events over India, supporting the transferability of the correction to operational forecasts.