Our award-winning reporting has moved

Context provides news and analysis on three of the world’s most critical issues:

climate change, the impact of technology on society, and inclusive economies.

Savvy sowing for Indian monsoon-dependent farmers

by Jerome Bossuet | @JeromeandAlina | International Crops Research Institute for the Semi-Arid Tropics (ICRISAT)
Tuesday, 29 August 2017 11:46 GMT

Farmers using the Sowing App in Kurnool district, Andhra Pradesh, India. S.Punna /ICRISAT

Image Caption and Rights Information

* Any views expressed in this opinion piece are those of the author and not of Thomson Reuters Foundation.

Rain fed farmers often lack advice before the rainy season starts about what crop to plant, and when to sow seeds

Groundnut farmer Bheemanna lives on few acres in the dryland region of Devanakonda, in Kurnool district, Andhra Pradesh State in India. Every year, around June- July, Bheemanna like millions of other monsoon-dependent smallholder farmers, tries to assess when the best time to sow their groundnut seeds is and how much area they should sow.

The onset of rains coming from Indian Ocean, overall rainfall quantity and recurrent dry spells vary a lot from one year to another in semi-arid tropical India. Rain fed farmers, who represent over 57 percent of India’s cultivated area and support 40 percent of the Indian population, often lack advice before the rainy season starts about what crop to plant, and when to sow seeds.

With increasingly erratic monsoon patterns over the past years, farmers feel even more insecure. For them, weather forecast information is the most important information, after commodity prices, in planning their crops. But most often they don’t have reliable predictions.

Some farmers would follow weather predictions from village elders interpreting signs like stars, wind or dragonfly behaviour, based on traditional astrological almanacs. The Indian Meteorological Department (IMD) does provide seasonal forecasts a few months before the rainy season but these forecasts cover huge areas and cannot be used on a local scale as farmers need.

Such weather predictions can become a very sensitive issue as witnessed a few weeks ago in Maharashtra. Some farmers there blamed the meteorologists for misguiding them and threatened legal action. Back in April, IMD had predicted a normal rainfall so farmers had sown in early June following heavy pre-monsoon rains. But in some parts of Maharashtra the rains were followed by over three weeks of dry weather resulting in farmers fearing for this year’s harvest and their livelihoods.

Timely advice

Scientists from the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) with the support of Microsoft and Indian weather institutes have developed real time farming decision tools that could lift a great part of climate risks for millions rainfed farmers in India.  

As part of the Andhra Pradesh Primary Sector Mission known as Rythu Kosam, a Sowing App  helped Bheemanna during the 2016 cropping season decide when it was best to sow his groundnut seeds. He received a series of agro advisories by text in his local language throughout the cropping season.

Using artificial intelligence and crop modelling tools, researchers could define for each village a moisture adequacy index through computation of local soil data, like the soil water retention capacity, historical and actual climate data and 5 days weather forecasts. Such index helps define the most suitable sowing date.

Bheemanna was advised to wait a month before planting when most of his peers had already planted their seeds. He also received advice on land preparation, sowing depth and density, water conservation tillage techniques like broad bed and furrow and micro nutrient fertilisation. Bheemanna harvested 1.5 tons of groundnut per hectare, much higher than the meagre yields of his peers.

The 150 farmers involved in this pilot gained an average yield increase of 30 percent . This mobile application is now scaled up in 13 districts of Andhra Pradesh.

Another approach explored by ICRISAT researchers is to build a decision tree using a mix of weather forecasts (seasonal, but also two-week and five-day forecasts) together with crop farming simulation to evaluate the different options farmers have in terms of crops, sowing time and farming practices to minimize climate risks.

This Intelligent agricultural Systems Advisory Tool (ISAT), initiated through the CCAFS Climate Smart Village initiative, is now scaled up with support from Sehgal Foundation.  After initial crop modelling validation and training, local extension agents and progressive farmers receive a whole range of cultivation advice text messages in local language, issued automatically every week, through the crop season.

Rainfall predictions are updated every seven days thanks to a network of automatic weather stations as well as rain gauge readings by lead farmers in each pilot village. The type of text advice sent to the farmer is selected automatically depending on the weekly rainfall amount.

This ISAT decision tool is based on four years testing with 450 farmers from semi-arid Anantapur district, which is a major area for groundnut farming in Andhra Pradesh. Here, farmers see 7 to 9 years out of ten as being very climate risky. 70 years of weather data show that when summer rains are below 300mm, this often leads to dry spells over 3-4 weeks during the period end June-August, which is very detrimental to the crops.  

Based on the seasonal climate forecast, farmers are advised to go for groundnut or shift from groundnut to more drought tolerant crops like sorghum or foxtail millet, after which they are advised on the optimal sowing date and management.

Farmer Subbubu for instance shifted to blackgram from cotton cultivation based on the advisory which resulted in a 20 percent increase in yields.. After using this advisory service for three years and feeling the benefits, farmers are ready to pay for the mobile application.

An exciting national development is India’s investment in weather forecast big data computing capacity. IMD is already able to scale down the rainfall predictions to district level and by next year, down to 330m grid range.

With better rain predictions for each Indian village and the advances of machine learning, this sowing decision toolbox will provide even more tailored and timely crop advice, for each monsoon season, to eager drylands farmers. Bheemanna and Subbubu are ready to pay for this precious service as they saw its value in their own fields.