Remote Sensing, Machine Learning and Socio-Economic Indicators for Climate-Smart Agriculture and Net Zero

Recent years have seen a significant increase in the availability of large datasets for agriculture applications. Remote sensing provides a powerful source of such data. Machine learning has expanded our capacity to derive useable knowledge from large datasets.

A growing number of methods have been developed, and made publicly available, to make use of these data and methods in support of climate-smart agriculture (CSA). In Africa, increasingly sophisticated crop classification, e.g. distinguishing monocrop from intercrop, has meant that remotely-sensed vegetation indices can be applied in bespoke ways to target a range of decisions, from national land use planning to crop management decisions. In the global north, a focus on Net Zero greenhouse gas emissions across societies, including in food systems, is changing the way technologies and data are being used in agriculture. Monitoring of flooding, irrigation, drought, crop health, livestock, soils, weather and, more recently, socio-economic conditions, means that a wide variety of decisions can be supported with remotely sensed data and ML techniques.

Transfer learning, where machine learning models are trained on large, diverse datasets in data-rich regions to be applied in data-scarce conditions, has opened up new possibilities for cost-effective ways to apply the above methods even where ground truth data are sparse. At the same time, socio-economic indicators from household and farm surveys, and resulting multicriteria analysis, provide a way to ensure that remotely-sensed data are relevant to lived experience.

This Research Topic focusses on applying these advances to diverse environments. Africa is a continent that has seen significant advances in the use of remote sensing and machine learning in recent years, and we expect that this will be reflected in the RT. The global north is likely to provide a contrast in terms of underlying agro-ecosystems, climates, data availability and technology. Studies that provide a holistic view of the role of CSA and or Net Zero agendas are especially welcome. Framing this view will be important: whilst CSA should encompass adaptation or resilience, increasing sustainable agricultural productivity and mitigation, Net Zero usually focusses primarily on the latter. However, CSA studies may focus on just one of these three elements providing they are placed within this broader context.

We welcome submissions from across the broad remit indicated in the title, for any regions of the globe. Specific areas where advances are needed include:

Development of indicators in support of CSA or Net Zero:
• Synergistic use of remote sensing with ground truth data (e.g. crop conditions, socio-economic data from householder or farmer surveys, land use) in support of CSA
• Assessing climate-smartness, or mitigation potential, using satellite data and/or machine learning
• Novel ways to generate socio-economic indicators using remotely-sensed data and/or machine learning
• Remote sensing and/or machine learning for mixed or livestock-based farming systems
• Transfer learning approaches that widen the domain of applicability of remotely-sensed data and/or machine learning methods, and deal with the challenges of data scarcity
• Use of remote sensing-derived crop and land use maps to support socio-economic objectives and agricultural practices/planning at local and regional scales.
• Approaches that compare and contrast remote sensing and/or machine learning in different geographies in order to deepen understanding and applicability.


Bespoke or co-developed use of satellite data, machine learning and modelling, such as:
• Approaches for capturing intercropping and/or other complex smallholder farming practices using remote sensing, to improve the accuracy of crop type maps and support more precise field-level decision-making
• Novel ways of combining remote sensing with crop modelling and/or machine learning to produce a more complete and/or accurate representation of conditions experienced on the ground
• Use of remote sensing within agro-advisories and publicly-funded initiatives, in support of resilient and adapted farming
• Novel ways of generating and using multi-scale data to support CSA / Net Zero at a range of scales from farm to country
• Advances in knowledge-guided machine learning that use satellite data in support of CSA / Net Zero
• Socio-economic indices in support of more complete characterizations of farming systems and/or greater uptake of remotely-sensed biophysical indices.