> Automatically verified official update: This page reproduces an official government communication from Ministry of Earth Sciences without adding legal conclusions. Always follow the linked primary source where exact wording matters.
Since 2021, the India Meteorological Department (IMD), under the Ministry of Earth Sciences (MoES), has adopted an advanced Multi-Model Ensemble (MME) approach for long-range forecasting. This has significantly improved the accuracy of long-range forecasts, with all forecasts issued since then remaining within the error limits. For example, operational monsoon forecasts since 2021 have been within the forecast error limits, with an average absolute error of 2.2% of the Long Period Average (LPA) over the five years (2021-2025). The year-wise details of the accuracy achieved by the IMD's Long-Range Forecast system during the last ten years are provided in the table below:
All India Monsoon Rainfall
Year
Actual Rainfall (% of LPA)
First stage Forecast issued in April
(Model error ± 5% of LPA)
Second stage Forecast issued in May
(Model error ± 4% of LPA)
2015
86
93
88
2016
97
106
106
2017
95
96
98
2018
91
97
97
2019
110
96
96=
2020
109
100
102
2021
99
98
101
2022
106
99
103
2023
95
96
96
2024
108
106
106
2025
108
105
106
During the last ten years, the forecast error exceeded 10% of the LPA only once, in 2019, when the absolute forecast error was 14% of the LPA. Since the adoption of the MME strategy in 2021, the India Meteorological Department has significantly improved the accuracy of its long-range forecasts, achieving a mean absolute forecast error of only 2.2% of the LPA.
The India Meteorological Department, under the Ministry, has progressively upgraded its LRF system by adopting an advanced MME forecasting approach based on coupled dynamical climate models. The MME system combines forecasts from multiple climate models, thereby reducing uncertainties associated with individual models and improving the reliability and skill of seasonal forecasts. The adoption of the MME forecasting system has led to a significant improvement in the reliability of seasonal monsoon forecasts. During the period 2021–2025, the average absolute error of the first-stage forecast was 3.1% of the LPA, while the average absolute error of the second-stage forecast further reduced to 2.2% of the LPA. In comparison, the average absolute error during 2016–2020 was 7.8% of the LPA. Thus, the average absolute error has reduced from 7.8% to 2.2% of the LPA, indicating a substantial enhancement in the reliability and skill of the seasonal prediction system following the adoption and operational refinement of the MME forecasting system. The MME system has also enhanced the consistency of forecasts by improving the representation of large-scale climate drivers such as the El Nino–Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD), leading to more reliable seasonal monsoon predictions.
The Government has taken several measures to further improve long-range monsoon forecasting and prediction of extreme weather events in the country. These include:
The above measures are being implemented uniformly across the country for the benefit of all States and Union Territories. Therefore, no State-wise allocation or distribution of these measures is maintained.
This information was given by the Minister of State (Independent Charge) for Earth Sciences Dr. Jitendra Singh in a written reply in Lok Sabha today.
NKR/JKP
Official source
[Read the original PIB release](https://pib.gov.in/PressReleaseIframePage.aspx?PRID=2290919)
Source attribution: Press Information Bureau, Government of India.