Inside Meghdoot, the app delivering weather-based agricultural advice to farmers across India

Top image sources: Meghdoot / CIMMYT

Since 2020, farmers across nearly all of India have been able to access to localized (district-level) weather and agricultural advice through a smartphone app. The Meghdoot app, which had been downloaded almost 300,000 times by the end of 2023, offers multiple advantages over delivery formats such as SMS or radio, and serves as a platform to continually deliver improved services to India’s farmers.

“Farmers don’t just need weather forecasts but agronomic recommendations which are interpreted for a weather forecast” said Ram Dhulipala, senior scientist at the International Livestock Research Institute (ILRI)

In 2008, multiple government agencies in India collaborated to create the District-level Agrometeorological Advisory Service (DAAS). The Indian Institute of Tropical Meteorology (IITM), the Indian Council for Agricultural Research (ICAR) and state universities set up specialized units called Agro-Met Field Units (AMFUs) under this program that develop crop-specific agro-advisories twice a week at the district level.

The goal of the Meghdoot app was to help more farmers to access critical and use this information. Through a simple interface, farmers can see the latest advisories alongside weather data and forecasts from the India Meteorological Department (IMD), targeted to their crops and locations, and available in 12 languages.

Meghdoot was initially developed by the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) with funding from the Ministry of Earth Sciences, Government of India under the Monsoon Mission program. The app and the backend platform that supports it were institutionalized under the Agro-advisory Services (AAS) division at IMD. The app and the platform were well received by farmers and hence enhancements to the app and the dashboards was later collaboratively developed with AAS and undertaken with support from the CGIAR Platform on Climate Change and Food Security (CCAFS) and later the CGIAR Initiative on Digital Innovation. As part of the project, the DAAS teams were also trained in crop simulation modelling and data analytics to develop more location-specific advisories.

Containing more information than a text message, easier to use than a website, and always available unlike a radio show, the app also creates a two-way flow of information. It allows the AAS division at IMD to collect usage statistics and source farmer feedback to make improvements. In 2023, some of the enhancements to the app included an offline mode, automated notifications when new advisories are available, integration to share information on social media, and more detailed analytics for policymakers to understand usage trends.

In the near future, artificial intelligence may be used to automate the development of advisories and weather summaries. “Provisioning such weather informed agro-advisory services poses huge challenges” said Dhulipala, “but this is where fusing decision support tools with large language models can offer transformative solutions.”

A pilot project that coupled the use of random forest machine learning with OpenAI to generate district and crop specific advisories in three districts has shown promising results. Successful of demonstration and scaling of this pilot can prove transformational for the DAAS services in India with potential opportunities to replicate these efforts globally.

To learn more about Meghdoot and other AI- and data-enhanced approaches to agricultural extension, join our webinar on 17 May. Register here: https://bit.ly/3wtNPF1