FAO’s Earth Observation for Statistics (FAO-EOSTAT) project, an innovative initiative supported by the Agrifood Economics and Policy (ESA) and the Statistics (ESS) Divisions, leverages next-generation Earth Observation (EO) big data and Artificial Intelligence (AI) to enhance traditional statistical tools for producing land cover and land use statistics.

Launched in 2019, EOSTAT has expanded from its initial implementations in Senegal and Uganda to 25 countries, including Rwanda, Lesotho, Zimbabwe, Colombia, Indonesia and many others to make agrifood systems more resilient and work toward achieving SDG2.

EOSTAT aims to respond to the need for global national statistics offices (NSOs) to align with the big data revolution, underpinned by the 2030 Agenda and the requirements of the SDGs’ reporting framework. It relies on free Earth observation data, vegetation and climate modelling, field survey data and AI classification algorithms to build the capacity of countries. Its primary objective is to empower NSOs to produce their own accurate and standardised maps of land cover, land use and crop yield.

Earth Observation Data (FAO-EOSTAT) is a highly collaborative effort within FAO, working closely with various divisions, including the Forestry Division (NFO), ESS, Office of Emergencies and Resilience (OER), the Land and Water NSL, the Digital FAO and Agro-informatics Division (CSI) and OIN on Artificial Intelligence. It is also a collaboration with the UN Task Team on Earth Observations for Agricultural Statistics.

Lorenzo DeSimone (ESA), lead of the EOSTAT project, explained, "Earth Observation data is crucial for FAO in supporting countries for the overall sustainable development of agriculture and for the sustainable management of natural resources."

As Lorenzo detailed further, the benefits of the project include access to accurate and timely data, cost-effective monitoring, ensuring data is comparable across different regions and time periods, an enhanced capacity for disaster response, support for sustainable agriculture, and improvement of yield estimates.

Since its start, EOSTAT has been highly successful worldwide. For example, in Lesotho, since its implementation in 2021, EOSTAT has enabled Lesotho's Bureau of Statistics to produce annual standardised maps, significantly enhancing the country's ability to report on SDG Indicator 15.4.2 - Mountain Green Cover Index.

Prior to its implementation, there were no mechanisms to update the existing national land cover baseline from 2015, which made reporting on SDG 15.4.2 impossible, so this represents a significant improvement over the previous reliance on outdated manual field surveys.

Mokoena France, Acting Director of Lesotho Meteorological Services and UNFCCC Focal Point, highlighted that the innovative tool enables the government to track land cover changes in any place in the country at any time – noting that it is essential for efforts to protect and restore natural resources and take effective climate action in Lesotho.

In Indonesia, EOSTAT has also been a groundbreaking initiative. Under the One Rice Project led by the Ministry of National Development Planning (Bappenas), EOSTAT has revolutionised rice production monitoring. The integration of satellite data, enriched with data gathered on rice fields collected in the field ensures a comprehensive understanding of the rice crop's lifecycle across Indonesia's diverse agricultural landscapes.

Imam Machdi, Deputy Chief Statistician, BPS-Statistics Indonesia noted that the EOSTAT Indonesia project is instrumental in supporting the Indonesia One Rice Data initiative in fostering the implementation of sophisticated machine learning algorithms.

Cameroon provides another excellent recent example of EOSTAT's success, as it has modernised agricultural statistics through workshops and the adoption of the Crop Yield Mapper Application.

Here, the tool enables precise seasonal mapping of crop distribution and empowers crop yield forecasting with high spatial resolution. The integration of the Systems Approach to Land Use Sustainability (SALUS) application further enhances the capabilities of this innovative system.

To exemplify its truly global impact, a final example is visible in Colombia, where EOSTAT was launched this year. Similarly to the other regions, EOSTAT plans to improve the current methodology used in the country to map land cover and use, and differently, introduce a new solution for crop yield forecasting based on the integration of EO data with crop growth modelling.

Initial results in Colombia show higher accuracy due to recent training events and prototype map development.

"Overall, the integration of EO data into the agricultural statistical systems of EOSTAT countries significantly enhances their ability to produce reliable, timely, and actionable agricultural statistics." Lorenzo continued, "ultimately supporting better governance, food security, and sustainable development practices."

EOSTAT is an exciting and innovative approach, that exemplifies FAO's commitment to harnessing cutting-edge technology for sustainable development. By expanding its reach, EOSTAT will continue to bolster evidence-based agricultural policies, analyse trends in sustainability indicators, and develop tailored applications to address agriculture and food security challenges globally.