农业食品经济

Advancing rice monitoring: EOSTAT workshop ignites innovation in Jakarta

12.12.2024

On December 3–4, 2024, Jakarta set the stage for a significant workshop organized by the Food and Agriculture Organization of the United Nations (FAO), in collaboration with Statistics Indonesia (BPS) and the Economic and Social Commission for Asia and the Pacific (ESCAP). Titled “Utilizing Earth Observation Data for Rice Data Calculation,” this hybrid event highlighted the transformative potential of Earth Observation (EO) data and machine learning technologies to revolutionize rice monitoring in Indonesia.   

As part of Indonesia's "One Rice Data Initiative", the workshop brought together global experts and keynote speakers to discuss integrating remote sensing and ground-truth data to modernize rice production statistics. Supported by FAO and ESCAP, this initiative aims to improve the timeliness, granularity and cost-effectiveness of agricultural statistics.   

Rice is fundamental to Indonesia’s food security, with 10.2 million hectares of paddy fields producing over 53 million tonnes of dry unhusked rice each year. However, the unique agricultural landscape, dominated by smallholder farming and diverse cropping cycles, presents challenges for accurately and efficiently collecting statistics on production. Modernizing these statistics with innovative EO-based methods is critical for supporting policy decisions that secure the nation's food supply. 

The FAO-ESCAP-BPS collaboration seeks to enhance the quality of rice data by integrating EO data with optimized field activities. This approach ensures that Indonesia’s rice statistics are not only more accurate and detailed but also generated more efficiently and cost-effectively.   

Impact of FAO's support   

FAO’s support to the One Rice Data Initiative has already delivered transformative results. Capacity-building efforts have enabled BPS to independently implement EO-based methodologies, increasing mapping accuracy from 67 to 95 percent in test areas using Sentinel-2 and Sentinel-1 data.   

Moreover, three analytical modules introduced by FAO have been successfully integrated into BPS’s operational rice monitoring workflow. These modules automatize the quality control of in-situ data based on spectral consistency and offer a robust solution to address gaps in EO data. As a result, BPS has scaled up the rice mapping exercise from one province in July 2024 to ten provinces by December 2024.   

Another of FAO's key contributions was the migration of the data pipeline and data cube from local machines to Indonesia's National Computing Centre. This transition considerably elevated the project's scalability and processing capabilities. 

Furthermore, in September 2024, FAO successfully submitted a data service application to the European Space Agency, securing the project a 1-year access to the Constellation of Small Satellites for Mediterranean Basin Observation (COSMO-SkyMed) high-resolution radar data (X-Band).

This data offers a spatial resolution of 1 to 3 meters and a temporal resolution of 1 to 2 image acquisitions per day, depending on the acquisition mode.   

The One Rice Data Initiative plans to use COSMO-SkyMed images in specific areas in Indonesia where the current model's accuracy requires improvement, greatly enhancing the project's precision and reliability.   

Transforming rice monitoring with capacity-building   

The recent two-day workshop in Jakarta showcased important achievements in 2024, particularly the integration of EO modules into BPS workflows. Validated through ground-truth data, this upgraded methodology has now been adopted as part of BPS’s official processes.

Presentation of the COSMO-SkyMed satellite images delivered by E-Geos.

New methodologies introduced during the workshop included crop yield modelling, cost-benefit analysis for evaluating EO data in rice statistics, and the use of COSMO-SkyMed high-resolution radar imagery. These advancements promise to refine accuracy and broader geographical application in Indonesia.   

The event featured prominent keynote speakers, including Dr. Jonathan Richetti, a research scientist at the Commonwealth Scientific and Industrial Research Organisation (CSIRO) specialized in crop modelling and EO applications for agricultural decision-making in Australia; Professor Jingfeng Huang, renowned scholar from the Zhejiang University of China, known for his expertise in agricultural remote sensing and rice yield estimation; and Jovino Lopez de Dios, Unit Head at the Philippine Rice Information System (PRiSM) and Supervising Science Research Specialist at the Philippine Rice Research Institute (PhilRice), where he drives digital transformation in rice monitoring systems.   

FAO was represented by Dr. Lorenzo De Simone, Geospatial Technical Adviser from the Agrifood Economics and Policy Division (ESA) and Leader of the EOSTAT program, who led discussions on advancing EO-based agricultural statistics globally, emphasizing the achievements of FAO’s technical cooperation with BPS, stating: "Through this collaboration, we were able to share best practices in utilizing Earth Observation data to (i) assess the suitability of the existing area frame for rice mapping, (ii) create Sentinel-1 and Sentinel-2 regularized data cubes using Microsoft Planetary Computing, and (iii) adopt machine learning methods for field survey quality control and enhancement."  

He further added, "These efforts have significantly improved the accuracy of the rice mapping solution under development, demonstrating the pivotal role of EO big data in modernizing national statistics systems."

Presentation of a global scale crop type mapping and crop yield forecasting solution. by Professor Huang Jingfeng, from Zhejiang University, representing the United Nations Big Data Global Hub China.

Last July, Gilberto Camara, Senior Research Fellow at Brazil's National Institute for Space Research (INPE), and Expert Consultant on the use of EO for the FAO-EOSTAT project, led a training session on the Satellite Image Time Series Analysis for EO Data Cubes (SITS) package, equipping the BPS team with critical skills to process time-series satellite imagery. This training was instrumental in developing the enhanced workflow showcased during the workshop.   

Next steps   

Building on the success of the December workshop, the One Rice initiative will continue advancing methodologies, expanding EO applications across Indonesia's agroecological zones, and strengthening national capacity for EO-based agricultural statistics. These efforts reaffirm FAO's commitment to supporting Indonesia's food security and agricultural modernization.