For years, Adalberto Martinez has cherished the dream of producing fruits and vegetables. However, there are many obstacles, such as the American embargo impeding importation of pesticides, the shortage of fuel, the virtual non-existence of wells and farm machinery, and especially the increasing unpredictability of rainfall.
Instead, the agronomy engineer has planted aloe on the four hectares he cultivates. “This plant requires little water and is very popular in the pharmaceutical industry. But it doesn’t feed Cubans! Climate change is making life more difficult for farmers. During dry spells, floods spoil our crops, and in the middle of the rainy season, we endure major drought events,” said Mr. Martinez, who belongs to a cooperative of 500 farmers in Matanzas province, one of the regions hardest hit by recent drought events in Cuba.
This year, however, Adalberto Martinez decided to take a risk by planting black beans, a local favourite, alongside his enormous aloe plants. Shortages of this legume are common on the island due to low national production. He is taking this risk because he is taking part in a pilot project run by the Cuban Meteorological Institute (INSMET). “Thanks to the precipitation report I’ve been receiving on my cell phone for the past year, I know three months in advance how much rain will fall on my land every day!” he said with a big smile. “The black beans I planted this spring would usually be harvested at the end of the rainy season in December. However, if I find out this summer that my plants could be spoiled because there will be several days without rain at the end of November, I’ll have several weeks to find fuel and a tractor to harvest my crop before the drought. That changes everything!”
Funded by Canada’s International Development Research Centre (IDRC), the Building resilience to droughts project, or SIN-Sequía (Without drought), has been run for almost two years by INSMET scientists in collaboration with Cuban universities and the Water Centre for the Humid Tropics of Latin America and the Caribbean. “We use several computer systems to predict daily precipitation over a period of 15 days, a month or even six months in advance, at a resolution of 1 kilometre,” emphasized Arnoldo Bezanilla, a researcher at INSMET’s Centre for Atmospheric Physics. “This gives farmers, authorities and the tourism industry several months to prepare. It’s something that’s already being done in Europe and the United States, but it’s new in Latin America.”
The project leader for SIN-Sequía explained that these long-term weather predictions are then processed by the Cuban Institute of Meteorology for outreach and dissemination. “Our data feeds the government’s drought and fire indices, tropical storm warnings, and the INSMET Institute of Meteorology reports on a web platform.”
Niliám Fernández Rosado, assistant technical director of the Matanzas Provincial Meteorological Centre, emphasized that other information formats were developed in order to reach as many people as possible. “Farmers have asked us to send them our forecasts through WhatsApp messaging because they have easier access to cellular data than to a computer connected to the Internet. These messages also serve as evidence for their insurers in order to obtain compensation when their crops are spoiled by an abnormal level of precipitation.”
In order to identify the information and data needs of the most vulnerable populations, such as women and the elderly, field interviews are currently being carried out in the municipalities of Ciego de Avila and Matanzas, two provinces where the agricultural and tourism sectors have been hard hit by drought in recent years.
Reading the future
How do you generate a high-resolution precipitation forecast months in advance? You need to use a large amount of data describing the state of weather and the hydrological cycle in the area under study. “Our first effort was to create a national database containing information recorded since 1961 at our 630 weather stations. It wasn’t easy because the data was scattered all over the country, sometimes in paper format,” explained Abel Centella, the INSMET researcher who carried out the work.
INSMET mathematician Maibys Sierra Lorenzo explained that artificial intelligence (AI) helped improve the new database. “Since we cover the whole country, there were areas where there were no direct observations, and sometimes satellite or radar measurements were not numerous or accurate enough to be useful to us. With the use of AI, we were able to combine more data to compensate for this lack of information and make the entire grid more accurate.”
The new data and an abundance of other information, notably taken from international databases, are then processed by digital weather- and climate-prediction models, which are computer programs that perform complex calculations to transpose the data onto a three-dimensional grid. As with a digital camera, the image is more precise, with more pixels. To produce a high-definition forecast, we need to divide the atmosphere into as many pixels as possible, meaning into small cubes representing 1 to 10 km on each side, which requires very powerful and stable machines.
“During the pandemic, new AI-based computer programs appeared in Europe and the United States. We decided to integrate them into our research to make the grids produced by current models even more accurate. The advantage of these AI-based models was that they didn’t require computers as powerful as those using current computer programs,” noted Maibys Sierra Lorenzo.
“This work allowed for the creation of very precise forecasts a month in advance,” explained Arnoldo Bezanilla. “Since April 2022, we have been providing daily precipitation forecasts for the following month. After that month, we take the actual observations and compare them with our forecasts. Currently, we have a precision rate of 70%, but we’d like to reach 80% or more.”
Maibys Sierra Lorenzo is convinced that these models will also be useful outside Cuba. “Several countries in the region are interested. It is important for us that the data be accessible to all, free of charge.”
Top photo: Agronomy engineer Adalberto Martinez cultivates aloe in Jovellanos, Matanzas province. Credit: Nancy Caouette
The initiative described in this article and production of this report were made possible with support from Canada’s International Development Research Centre.
This article was originally published in French in the March 2024 issue of the magazine Québec Science.