Connect with us

Support The NewsHawks

Some smallholder farmers are now using AI to predict rains

News

Zimbabwe smallholder farmers use AI and Open-Source Intelligence to battle climate change

“This season, when Accuweather predicted low rainfall, we made preparations for irrigation. This has given us advantage over the shifting seasons and extremely low rainfall in 2024.”

Published

on

BY NATHAN GUMA

IN a remote farm in Ndiri, Concession in Mazowe District, 73 kilometres North of Zimbabwe’s capital, Harare, Darlington Kupara, a smallholder farmer walks through his lush green tobacco field.

While his tobacco crop is lively, it is surrounded by parched and wilting crops, which have been affected by a long dry spell.

“My tobacco plant has been faring better than maize. This is because I have been switching to irrigation whenever prediction tools show there will be little to no rain. I however don’t have irrigation equipment in the maize field,” he told The NewsHawks in early January.

Kupara is one of the few farmers using open-source data and artificial intelligence to predict rainfall patterns to beat climate change, which has caused a drastic shift in seasons.

Open-source data refers to datasets that are freely accessible to the public. This means anyone can access, use, modify, and distribute the data without restrictions.

He has been accessing weather data through Meta AI, a large language model (LLM), where he inputs his location to receive tailored weather forecasts.

For instance, a prompt for the amount of rain to expect for a particular day on WhatsApp’s Meta AI directs him to open-source websites such as accuweather or timeanddate.com among others.

The AI interface then simplifies complex weather data into easy-to-understand interactive graphs. These usually show amount of rainfall to expect, humidity and cloud cover among others. 

Innovation

For Kupara, necessity has been the mother of invention. He is using AI and open-source data to predict rainfall patterns, which has helped him to save his crops where others have failed.

Kupara stands besides his tobacco which has fared well compared to other farmers in the region. He has been using open source intelligence to predict rain.

“I cultivated tobacco and 10 hectares of maize and have been using my phone to predict rain. I have been using google weather data mostly. All I need is data and a smartphone to be able to access some of the data. It is free and readily accessible,” he said.

Whenever open-source websites predict prolonged low rainfall, Kupara has irrigation equipment ready, which has saved his tobacco crops.

Despite their accuracy, Kupara says some of the open-source websites have been misleading. For instance, in November when the initial rains came, accuweather, an open-source website predicted optimum rainfall. 

However, there was above-normal rainfall, which washed away seeds which he had planted for the season. 

“In November, I used accuweather.com, and it told me that there was going to be normal rain. I immediately hired a planter and covered the entire 10ha swath of land with maize,” he said. 

“Part of the maize seed was washed in the rain. This is why you see random parts of the field with crops of differing heights.”

While he is happy that his crops are faring better in the shifting seasons, his yield has been diminishing due to climate change-induced low levels.

“Last year was better. It was a bit better as there was more rain. I planted soya beans and maize. All the maize seed germinated. Rains went when it had already produced a cob. I managed to get two to three tonnes per hectare. But this year is a much different story,” he says.

“The rains were late. And, if you look at what I have produced so far. It is a bit disheartening. But the previous one was better. I think this has been caused by a shift in seasons.”

Maxwell Matotoba, another farmer from Ndiri has also been using various types of open-source websites to access weather data, which has been giving him complete weather data showing the amount and intensity of rainfall to expect. 

“I got to know AI and open-source tools through our extension officers. Right now, I have been using accuweather. There are several of them, and some of them are not always reliable. So, it all depends on which tool you use. The other one is called Arabia, timeanddate.com,” Matotoba says.

“These are important as they tell you the exact amount of rain to expect and intensity, which helps protect against crop damage. In terms of accuracy, for this year I have been using accuweather.com. It has been more accurate. Though rain is no longer consistent, it would have predicted the correct amount of rainfall to expect - no matter how little it is.

“This season, when Accuweather predicted low rainfall, we made preparations for irrigation. This has given us an advantage over the shifting seasons and extremely low rainfall in 2024.”

Digital Gap and unreliable information

Farmers under the Zimbabwe Smallholder Organic Farmers Forum (Zimsoff) say it has been difficult for them to predict the erratic rainfall from mainstream weather data.

This is because most of them are moving away from radio and television as a main source of weather information.

“Our farmers have been relying on information from the Met Dept, the information is mostly to those who are privileged to have radios and access to print media. At times the information is misleading as it mostly does not come as predicted, there should be other ways to improve how the Meteorological department works like asking for feedback from the grassroots,” said Ngoni Chikowe, a technical officer at Zimsoff in Mutoko.

“The Met Department does not give accurate rain amount and duration of pour. For example, they do not give information on a specific area. For instance, they can give weather information for Marondera, and not the exact place where rain will occur.”

Chikowe says this has seen most farmers in Mutoko not planting for the 2024-25 season, as they have lost seeds they have planted twice.

“No one gives them seed to replant, so they have just given up. It has also become more impossible to replant as the season is almost halfway with little to no hope of a harvest,” Chikowe said.

Reliability of AI and open-source data

However, Zimbabwe’s Meteorological Services Department (MSD) says farmers should get reliable information, as some sources have misleading information. 

“There have been so many players. But then, you must get back to the source and see how they gather their information. We have been forecasting for over 100 years and climate information collection for over 150 years. Zimbabwe was once a breadbasket because of all the information they got from the stations,” said James Ngoma, MSD head of weather services.

Ngoma says his department has been working with agricultural extension officers nationwide to disseminate weather data to farmers so they can have reliable information.

“We have been working with the Department of Agricultural, Technical and Extension Services (Agritex), which works directly with farmers to disseminate climate and information. The officers work with the farmers and inform them on how best they can be productive. So, farmers would have seen the Met Services when they see extension officers.

“In reaching communities, we have one of the most devolved sectors. We have district offices which are responsible for gathering data that farmers in those regions then use. These know exactly what is happening at every station. So, we get all the information right from the source.”

Ngoma says that the MSD has digital stations, which can record weather data - particularly rainfall, in real-time.

“There used to be the old stations using Stevenson's Screen and the popular wet and dry bulb thermometer. All those are being transformed into digital. The digital stations transmit weather information in real-time. So, in real-time, you can see the temperature trends in all places.

“We also have over 100 automatic stations across Zimbabwe that give real-time weather, adding parameters such as solar radiance, wind speed, direction and rainfall using an automatic rain gauge. We can pick if intense rainfall happened and when it decreased.” 

Ngoma says that the MSD now has a sound radar network, for data collection, which has been augmenting volunteer data collectors.

“We are one of the few countries to have done so in Africa. We are even better in terms of our radars compared to South Africa. And, we are also the best in terms of coverage because we have covered the entire country using our radar system.

“And added to that we have personnel that is over 200, that can give you data as it happens, down to the grassroots level. We work with Agritex and the Civil Protection Unit (CPU). We have also inserted rain gauges in schools. We have over 600 rain gauges in schools.”

For Kupara, open-source data will be a crucial part of his farming as climate change continues to wreak havoc.

“Most big farmers have been doing it. I believe that as technology advances, we are just going to be at par with them,” he says.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *


Advertisement




Popular