AidData is leading a team of partners that has won a $4.74-million, four-year investment from the Bill & Melinda Gates Foundation for a project to spur the use of Earth Observation in how we study climate-sensitive agriculture across developing country geographies.
Named “GeoField” (Geospatial and Earth Observation for Impact Evaluation, Learning, and Development), the project will include impact evaluation case studies, new datasets, methods and code, the development of Communities of Practice, scientific papers, policy briefs, online tutorials and in-person trainings. GeoField will address the barriers to broader use and accelerate the existing efforts around Earth Observation (EO) for impact evaluations. EO encompasses remote-sensing technologies like images and data from satellites, airborne platforms like drones, as well as ground-based collection tools.
GeoField has three core partners: AidData, DevGlobal Partners, and Mercy Corps. The concept and plan for GeoField has been two years in the making, refined and honed in partnership with the Gates Foundation.
“Agronomists and environmental scientists have a long history of using EO to measure a range of biophysical conditions, while social scientists and program evaluators have a similarly long track record of estimating counterfactual comparisons via impact evaluation (IE),” said Ariel BenYishay, AidData’s Chief Economist, and one of GeoField’s Principal Investigators. “The challenge of climate change, especially as it relates to agricultural adaptation, is presenting us with an urgent call to bring together these two domains, to work together to integrate EO with impact evaluations. Combining the two will enable us to directly observe impacts, and to expand accessible IE outcome measures to include soil management, erosion control, irrigation and water management, livestock, and other practices. We’re tremendously grateful to the Foundation, and excited to get this ambitious, interdisciplinary project started. We believe it will have the power to help farmers while pushing the boundaries of scientific research. It will help aid organizations invest and measure impact more wisely.”
GeoField aims to show how EO can dramatically enlarge sample sizes and expand the geographic scope and time-coverage of evaluations, permitting researchers to detect changes in the biophysical conditions of crops and vegetation, livestock, and in human well-being that would be difficult if not impossible with solely ground-based observations.
“Multi-decade streams of data captured by EO satellites not only evidence climate change, they can inform the way forward,” said Matt Manning, Executive Director of DevGlobal Partners. “In many crises, satellites are our first eyes on the scene, providing the only option to survey farmers and communities affected by conflict or disaster. Satellites also image even the most remote communities, which means we can understand what’s working and not working for our most vulnerable neighbors.”
By expanding sample sizes, EO can power more statistically robust gender disaggregation than possible by surveys alone. Moreover, survey responses themselves contain errors and biases that exacerbate gender-related measurement challenges (plot size estimates are one common example), and EO can often provide independent validation of survey-based responses.
GeoField was designed with gender as a critical component. The project will ensure evaluations capture critical differences in how women and men are affected by agriculture interventions. For example, many evaluations rely on self-reported responses (including responses on a spouse’s agricultural practices and outcomes, in many cases). EO can offer independent, complementary measurements. The case studies supported under this investment will include tailored EO designs intended to reflect gender outcomes fully. For several of the studies, complementary small-scale ground data (including surveys, behavioral experiments, or focus groups) will also be used and cross-referenced to EO-based outcomes. Across the planned convenings and in assembling the community of practice, gender balance will be a priority. The GeoField team will also partner with local researchers and leaders who champion gender diversity.
Capitalizing on the revolution in Earth Observation
Advances in EO mean that more features of climate change are becoming measurable, including groundwater availability and usage, land degradation, and carbon sequestration. There are new fine scale remote-sensing platforms, including 10-cm satellite imagery from providers like Albedo, Landsat 9, and Microsoft’s Planetary Computer. These are coupled with emerging technologies, platforms, and interfaces, such as graphics processing units, open data sources (e.g., USGS Earth Explorer), and cloud platforms (e.g., Google Earth Engine), all opening up new doors for innovative research.
Despite this, technical and managerial barriers continue to hamper the incorporation of EO into evaluations. Hanna Camp, Mercy Corps’ Senior Adviser for Monitoring, Evaluation and Learning Technology, observed: “Even as the cost of imagery, skills, and computing power to perform EO analysis has fallen, in our experience EO knowledge remains relatively rare within implementing organizations. Humanitarian and international development organizations like Mercy Corps are increasingly focusing on climate outcomes and operating in difficult-to-survey contexts, but even where impact evaluations using EO show promising results, it can be hard to seize that opportunity if the methodology and analytic process remain opaque. That doesn’t mean everyone needs to learn how EO works, but it does mean that program designers and implementers need to have tools to determine when to use EO and when not, what to expect from it, how to properly plan and budget for it, and how to recruit the right help. Without that, experiments or pilots will often be undertaken by a highly-skilled consulting or academic research partner, without necessarily giving the implementing organization the tools they need.”
GeoField will support a new “Community of Communities” of Practice (CoP), building on existing networks of evaluation organizations and funders, EO scientists, policy and data organizations, and industry partners. Through GeoField, this community of evaluators, researchers, users, donors and other stakeholders will expand the use cases of EO, encourage more learning and cross-fertilization, and reduce the technical and managerial barriers to access and use EO across the spectrum of agriculture impact evaluations.
Every aspect of the selected case studies, starting from methods to peer-reviewed scholarly work, will be shared through various platforms, including a dedicated microsite on the AidData site, as well as GitHub and other repositories. The GeoField team will develop open-source methods, collect, and create both in-situ and remotely sensed data for each study, implement them through programming (Google Earth Engine, R, Python, etc.), and assess their ease-of-usability with prospective users via human-centered design. The team will provide critical datasets in an Analysis-Ready Cloud-Optimized (ARCO) data store for researchers and users. Each dataset will also be accessible via AidData’s GeoQuery platform.
GeoField will also work to condense process learnings from the case studies into usable guidance and tools for end-users seeking to benefit from—but not necessarily perform—impact evaluations using EO. This can include summary guidance on integrating evaluations using EO into more traditional monitoring and evaluation systems, example budgets, and template proposal language.
For more information, contact:
Brad Sagara, Director, Research & Learning – Resilience, Mercy Corps, email@example.com
Nicole Hunt, Communications Lead, DevGlobal Partners, Nicole.Hunt@dev.global
About the Partners
Housed at William & Mary's Global Research Institute, AidData connects decision-makers and researchers who have a shared interest in working together using granular data and innovative tools to solve pressing problems, precisely target resources, and use rigorous evidence to measure the impacts of policies and investments. We prioritize working in areas where our expertise can have an outsized impact—data-poor environments where there is an unmet need for better evidence and insights.
Spanning Africa, India, and the United States, DevGlobal Partners supports the world’s leading non-profits, companies, philanthropies, multilateral agencies, and government agencies on their innovation and sustainability initiatives. DevGlobal builds teams and brings organizations together to democratize technology, bridge divides and disrupt with intention. DevGlobal has extensive background in Earth Observation and geospatial technologies.
Mercy Corps is a global team of over 5,400 humanitarians working to create a world where everyone can prosper. In more than 40 countries affected by crisis, disaster, poverty and climate change we work alongside communities, local governments, forward-thinking corporations and social entrepreneurs to meet urgent needs and develop long-term solutions to make lasting change possible. Across its humanitarian and development program and research portfolio, Mercy Corps explores and tests solutions to advance climate-smart agriculture. Mercy Corps has a total operating budget of over $550M and last year reached over 50 million people. The GeoField team believes implementing organizations are key stakeholders in this project, and expects to expand the partnership to include up to four in-country implementing partners, in addition to potentially more university researchers and data suppliers.