Loading…
Since adopting the Multi-Model Ensemble (MME) approach in 2021, IMD's mean absolute long-range monsoon forecast error has fallen from 7.8% of the LPA (2016-2020) to 2.2% (2021-2025)
In the last ten years, forecast error crossed 10% of the LPA only once — in 2019, when it hit 14%
The government cites Mission Mausam's radar, AWS and AI/ML expansion as the next lever for further accuracy gains
LPA is the average all-India southwest monsoon rainfall computed over a fixed climatological base period, against which every year's actual rainfall is expressed as a percentage — 96-104% of LPA is classed 'normal'. Per World Meteorological Organization (WMO) norms, India updates this baseline every 10 years: it was 89 cm (1951-2000 base) until 2018, then 88 cm (1961-2010 base), and is currently 87 cm (1971-2020 base).
Simple Analogy: A rolling 30-year climate 'par score' that rainfall each year is graded against.
National agency for weather and monsoon forecasting; issues the two-stage (April, May) long-range forecast
₹2,000-crore mission (Cabinet-approved September 2024, launched January 2025) expanding Doppler Weather Radars, Automatic Weather Stations and AI/ML forecasting tools — the measure this reply cites for further accuracy gains
Large-scale climate drivers whose improved representation inside the MME system is credited with reducing forecast error
GS Paper 3 > Science & Technology > Disaster Management (Forecasting)
Recurring — asked most monsoon sessions
A forecasting method combining outputs of multiple coupled global climate models to reduce individual-model uncertainty
The benchmark average rainfall over a fixed climatological period against which a season's actual rainfall is measured