Who would have thought that an obscure climate database would someday achieve science superstar status and help pave the way for a big-data revolution in tropical agriculture? As unlikely as it may seem, this was the destiny of a geek tool called WorldClim. Its story offers important lessons on the value of big data analytics and open-data policies for science-based innovation. And these lessons, in turn, should serve as a call to action for farmers, researchers, policymakers, and others committed to sustainable agricultural development.
Data in the limelight
Consisting of climate “grids” (the result of linking data gathered from thousands of rural weather stations with geographic locations across the tropics), WorldClim is freely available online and has been downloaded hundreds of thousands of times for mapping and modelling related to agriculture and the environment. The article in the International Journal of Climatology that officially launched WorldClim in 2005 is the second most frequently cited paper in the history of CGIAR, the global research partnership of which my organisation, CIAT, forms a part.
WorldClim has become a “granddaddy” among data resources, spawning a wide variety of other specialized tools, which have proved valuable for both anticipating challenges and targeting solutions. For example, experts on climate change can project how its impacts will affect the suitability of particular places for crops several decades from now. Similarly, if certain bean varieties are known to perform well at locations for which weather data are available, new tools using WorldClim can indicate with remarkable accuracy where else in the tropics these varieties are likely to thrive.
The power and versatility of WorldClim helped pave the way for a new open-access culture among CGIAR centers. All 15 of them have signed on to a new open-access policy, and CIAT is leading the design of a cross-centre platform with multiple partners that aims to extend the benefits of big data analytics more widely to smallholder farmers.
From analysis to action
One especially promising avenue toward this end is “site-specific agriculture” centering on a data-driven approach to agronomy. It helps farmers optimize crop yields by adjusting their management practices to subtle variations in growing conditions across sites and over time in a given area. Recommended practices are based on the use of big data tools to analyse large amounts of information collected with developing country partners, employing methods that range from drones and satellites to sensor networks and market information systems.
A big breakthrough for this work came a few years ago through a major project on agriculture and climate change in Colombia, which allowed CIAT and its national partners to apply the approach to major food crops like rice. A key challenge was to determine why average rice yields have dropped by a ton in the last 5 years and how to arrest the decline.
Through case studies, CIAT scientists showed which climate patterns have been associated with high and low yields in the past, and then linked the analysis of historical records with state-of-the-art seasonal weather forecasts. Next, they identified the weather patterns from previous years that best match the forecasts, checked to see which rice varieties and management practices did best in those years, and finally translated this information with local experts into recommendations for farmers.
Pleased with the results, Fedearroz (Colombia’s national rice growers association) adopted site-specific agriculture as a critical part of its effort of to make the sector more competitive by closing yield gaps, lowering production costs through more efficient use of inputs, and enhancing resilience in the face of climate change.
Four priorities for scaling up
In recognition of this work, a UN climate change initiative selected CIAT’s big data team to be one of the two winners of its Big Data Climate Challenge, which was awarded last year during the UN Climate Summit held in New York City. International recognition fueled interest in further exploring the potential of site-specific agriculture to raise the yields of rice and other crops. In a major step forward, the team gained World Bank support to share the approach with rice growers associations in Argentina, Nicaragua, Peru, and Uruguay.
What more will it take to scale up data-driven agronomy for achieving sustainable productivity growth? I would call attention to four priority actions.
- Strengthen two-way communication: New information and communications tools are needed for getting recommendations to farmers and for farmers to feed information into big data analytics platforms
- Leverage big data partnerships: Agriculture is a critical arena for fine-tuning big data approaches to address real-world problems – a tasks that requires teaming up heavy hitters in big data analytics with development agronomists.
- Build sustainable businesses around data-driven agronomy: We need to explore novel business models, in which farmers produce both grain and information, and every African village has a big-data promoter, providing data-informed agronomic services.
- Fund the democratization of big data: Smallholder farmers, who are already marginalized in many ways, risk having the information revolution pass them by, unless public and private donors invest in kickstarting big data analytics in tropical agriculture.
CIAT scientists often say that no one could have anticipated the many remarkable applications of WorldClim when it was first developed. In this same spirit of scientific serendipity, I urge everyone who has a stake in agricultural research for development to join us in creating the conditions for a big data revolution in agriculture.
Ruben Echeverría will facilitate a session at the Borlaug Dialogue held in Des Moines, Iowa this week entitled “Precision agriculture & big data: technologies for Resilience”. Follow #FoodPrize15 for the latest coverage of the event.