The idea behind Precision Agriculture is about putting an end to speculation and instinct, even though it may have been acquired over many decades of hard toiling in the fields. Arguably that’s why “gut-feel” in business decision-making is on its way out too and is being replaced by hard data and analysis. Studies reveal that AI-based sowing advisories can increase agricultural yield b y as much as 30%. Encouraged by similar findings, NITI Ayog was prompted to start a pilot project in seven states. 

The AI mechanism is fed with historic climatic data of the last 30 years based on which the Moisture Adequacy Index (MAI) is calculated to determine if the sowing period has been rightly identified. MAI is a standard measure that calculates rainfall and correlates with soil moisture and the crop. Daily rainfall data can be used to predict real-time MAI, and in the same manner, the predictive tools can forecast future MAI as well. 

Typically, monsoon months are June, July, and August. And farmers start sowing from June onwards, but with changing weather patterns, this may not always be the ideal approach. Even city dwellers blissfully confined in their concrete jungles understand that crops require moisture and the sun for growth. But for a farmer that’s his very existence and even a slight deviation can cause misery. While they do take day-to-day decisions that impact costs, efficiency, and yield, but it may be worthwhile to see how critical these decisions are and influenced by weather conditions. 

• Crop growth. It’s a no-brainer that appropriate amounts of moisture, light, and temperature are required to maximize growth. Historical data analyzed in great detail can enable farmers to take vital decisions that have high-cost implications - how much to irrigate, when, or even, should they be doing it at all for a particular crop? 
• Fertilizer – timing. When is it going to be dry enough to apply fertilizers? They are frightfully expensive after all, and if they drain away due to rainfall, it could cripple the farmer monetarily. 
• Pest & Disease Control. Certain weather conditions are conducive to pests that destroy crops. Forecast guidance incorporated into pest and disease modelling can help determine when (or whether) to use pest controls, and if the wind conditions are favourable. 
• Field-workability is the availability of days that are suitable for fieldwork – once again, it is dependent on soil moisture and temperature.  

Fasal Salah – Advising the farmer

It’s an informative free-to-use app that helps farmers by providing real-time personalized crop advisory services based on current and forecasted weather. It was developed by BKC WeatherSys Ltd., one of India’s earliest private sector meteorology and environmental technology companies. The app is created by agronomists, and very thoughtfully, they have enabled audio-advisory features which are of great help to semi-literate farmers.

Detailed advice on the parameters mentioned above – weather forecast, including temperature, humidity, wind speed, direction, and rainfall, can be provided at least ten days in advance. The estimates are granular and customized at the village level. They also provide real-time information on the possibility of pest attacks and diseases on crops and issue advisories on how these issues can be tackled.

This app can also assist the crop insurance system, which is often burdened with the twin challenge of delayed yield data submission and data quality. While issuing out various advisories, the information is already being captured, which adds a high degree of accuracy as well. 

Currently, about one lakh Indian farmers are using this app.

AI in weather forecasting

Traditional models of weather forecasting are based on statistical measures of numeric models, and it does not give answers in the binary. While that may be good enough for most people but the airline industry or agriculture, require something far more precise. 

The data is collected from a myriad of sources such as deep space satellites, weather balloons, radar systems, newscasting weather warnings, and even IoT sensors. But with massively increased data sets, we move into the domain of Artificial Intelligence, and it is truly an asset which can augment predictions with far greater accuracy, 

The AI-based predictions use machine learning algorithms that process complex data sets in a very short period using mathematical models. The model can navigate through vast data sets transmitted from weather satellites and relay stations to deliver short-term forecasts or long-term predictions. For instance, Google’s AI forecast tool is based on the UNET Convolutional Network that makes it possible for researchers to generate accurate rainfall predictions six hours before the actual precipitation. Imagine the kind of applications that can be delivered to avert natural disasters to save thousands of lives. 

Like every other sector, talent is the game-changer. To be fair, it’s not easy for Meteorological Departments (world over) to attract this kind of talent, and that’s why we see so many global tech titans operating in this space. It will be interesting to observe how the new business model shapes up. Will it work on the lines of PPP, or will private players take it over completely? 

But for now, we should take pride that companies such as BKC WeatherSys (Fasal Salah app) are feverishly working towards addressing Sustainable Development Goals – Zero Hunger, Reduced Inequalities & Sustainable Communities.