To improve weather forecasting for farmers in India, IBM is relying on AI3 min read . Updated: 05 Nov 2018, 09:01 AM IST
With climate change, forecasts are now pivotal for crop yields as farmers can't rely just on traditional knowledge
Six months back, government policy think tank NITI Aayog signed a deal with IBM India Pvt. Ltd to develop a prediction model for crop yields using artificial intelligence (AI).
As part of the first phase, the organizations are jointly developing this predictive model for 10 “aspirational" districts across the states of Assam, Bihar, Jharkhand, Madhya Pradesh, Maharashtra, Rajasthan and Uttar Pradesh.
While IBM will be using AI to develop the technological model for improving agricultural output and productivity for various crops and soil types for the identified districts, NITI Aayog will use the data insights generated through these AI models to help farmers and other stakeholders.
A large part of this agriculture platform has been developed by IBM researchers from India, according to Himanshu Goyal, India sales and alliances leader for The Weather Company, which IBM acquired three years back.
IBM has been focusing on agriculture globally too, points out Mary Glackin, head of science and government affairs at The Weather Company, which is the world’s largest private weather enterprise. The company provides up to 26 billion weather data and insights daily via Weather’s application programming interface(API) and its own digital products like The Weather Channel (www.weather.com) and Weather Underground(www.wunderground.com).
In the agricultural sector, The Weather Company adopts a “whole ecosystem" approach which includes suppliers, seed and fertilizer companies, farmers, lenders and distributors, according to Glackin. “We seek to provide a decision platform where we can bring together the information across these players to help everyone take informed decisions," she said in a recent interview.
Glackin insists that monsoon forecasts are important for Indian agriculture, especially as climate changes because farmers no longer can rely on just traditional knowledge.
But what exactly is this ‘Decision Platform’? Data is the first part of the platform, explains Glackin. It comprises weather data collected from remote sensing and satellite imagery and drones that are being used in some parts of the world.
The second big part is collating localized weather data, which could also be historical. This data helps IBM build an electronic field record to understand the historical context and current scenario. “If you are a lender, you may want to know the past record in that field or plot—risk of the particular farmer—what was grown there, for instance," explains Glackin. The third part is extracting insights from that data, which is where machine learning and AI are used.
The last piece is around “Decision Support"—exploring choices and making decisions. It’s here that AI (machine learning) helps in making decisions. “This is what we do in the weather world—we take forecasts from multiple sources and do machine learning on that to move it forward," says Glackin.
Other than agriculture, IBM also works with other sectors like aviation, energy trading (futures), energy management (platforms to correlate impact of storms to help companies pre-deploy equipment), fleet management (predicting the impact of weather and traffic) and renewables (for example how many sunny days in the context of solar power).
Weather forecasting is a lucrative market. Allied Market Research, which pegged the global weather forecasting services market value at $1.2 billion in 2016, has forecast it to more than double and touch $2.8 billion by 2023. Weather forecasting services are widely used in sectors like aviation, energy, agriculture, retail and media.
Major companies operating in the market, according to Acumen Research and Consulting include The Weather Company, AccuWeather Inc., Met Office, Skymet Weather Services Pvt. Ltd, BMT Group Ltd, Global Weather Corporation, Precision Weather, Fugro, ENAV S.p.A, and Skyview Systems Ltd.
Commenting on the stiff competition, Glackin said, “Our platform is our differentiation. We are not saying we are in the business of making the best crop forecast. We think we will make a good crop forecast. We are really here to create a platform that can be used by banks, insurance companies, suppliers, traders, etc."