The NADiRA consortium developed the processing chains required for satellite augmented detection of plot-level agronomic practices (e.g. plowing, flooding, harvesting), monitoring of crop response, and – with the support of novel, low-cost sensor networks and multi-model simulations – crop yield forecasting.
Corresponding workflows were connected into MANOBI’s agCelerant™ value chain orchestrator, which links agricultural smallholders to banks, insurance, input suppliers and agro-industry to cover the needs of the entire agricultural value chain continuum.
Simultaneously, the consortium developed a Geospatial Exploitation software Platform (GEP), capitalizing on technologies and expertise developed with the European Space Agency during the last decade. Delivered in 2019 as a pre-operational capability, implemented on Google Cloud, the GEP uses state of the art technology to organize, discover, collect and manage large quantities of geo-referenced data.
Its interfacing with agCelerant™ makes it the world’s first fully automated processing system providing concurrent detection of field-level agricultural events, monitoring of crop response, and production forecasts based on combined field and EO data streams, which are both essential for operational exploitation and scalability, particularly in heterogeneous smallholder agriculture.
The consortium deployed three pilots to monitor the crop seasons in 2018 and 2019 in Senegal for rice and groundnut and in Nigeria for sorghum. The pilots were deployed at the scale of 2.500 hectares for rice, 250 hectares for groundnut and 1.600 hectares for sorghum.
Despite this limited scale, the pilots have highlighted challenges arising from the concrete experience on the field, especially in the context of the smallholders’ agriculture: quality of field data, interpretation of the outcome of EO processing, sensitivity of the mechanistic crop models to the variability of field input data. The simultaneous performance of innovation and pilot deployment helped the consortium to assess the relevance of the developed services and to adjust and tune the algorithms as well as the interpretation and exploitation of the service products.
In the early stage of pilot deployment, agCelerantTM implemented automatic consistency checks of field data, based on Artificial Intelligence techniques, to solve the critical issue of data quality in field data streams.
The rice pilot in Senegal helped testing the operational preparedness of the business model on a limited scale (190 farmers), based on an agCelerant contract linking rice producers with a microfinance institution, an insurer and an industrial buyer.
The analysis of information produced in the successive pilots by the different sources of data (EO, IoT, Mobile2web, crop models) showed that no single source is reliable enough to be “served” individually to the stakeholders in the value chain. The outstanding value of the System must be derived from the “educated” interpretation of instantaneously available information by local agents.
Overall, the lessons learned from these pilots are extremely useful to improve the operational system prior to its first use in the upcoming business contracts.
MANOBI undertook early outreach actions to market agCelerant as a vehicle for the numerous innovations developed in the framework of NADiRA. Key partners from several African countries expressed strong interest in the orchestration of smallholder value chains development, including the Islamic Development Bank, in relation to its Regional Rice Value Chain Programme, and Nigeria’s Federal Ministry of Agriculture and Rural Development (FMARD). Several significant business opportunities were subsequently developed in 2019, which represent a positive input, based on a sound multi-scale analysis of market conditions, towards a viable business plan for NADiRA and the confirmation of NADiRA’s first commercial market by 2020.
