CGIAR Initiative on Digital Innovation

Permanent URI for this collectionhttps://hdl.handle.net/10568/117892

Part of the CGIAR Action Area on Systems Transformation.

Primary CGIAR impact area: Poverty reduction, livelihoods and jobs

https://www.cgiar.org/initiative/25-harnessing-digital-technologies-for-timely-decision-making-across-food-land-and-water-systems/

Browse

Now showing 1 - 20 of 420

Parcela Agroclimática Digital: Un laboratorio en campo donde tecnología y agricultura se integran para transformar el agro.
(Infographic, 2025-07-01) Barrios, Camilo; Amaya, Alejandra; Arce, Daniela
La Parcela Agroclimática Digital, ubicada en el Campus Las Américas de la Alianza Bioversity y el CIAT en Palmira, Colombia, es un espacio experimental que combina soluciones de Internet de las Cosas (IoT), sensores de monitoreo en suelo, drones multiespectrales, imágenes satelitales de alta resolución (PlanetScope), y plataformas digitales de visualización en tiempo real, habilitadas por conectividad satelital (como Starlink). Esta infraestructura permite el monitoreo continuo y preciso de variables clave del cultivo, el suelo y el clima, como pH, NPK, temperatura, humedad, conductividad eléctrica y nivel freático, mediante redes de transmisión de datos como LoRaWAN. Más allá de capturar datos, este laboratorio en campo tiene como propósito generar información agroclimática procesada y contextualizada, que, al integrarse con técnicas avanzadas de modelación e inteligencia artificial, facilita la toma de decisiones informadas. De este modo, permite a agricultores, técnicos e investigadores optimizar las estrategias de manejo agronómico de acuerdo con las condiciones agroambientales específicas, contribuyendo a una agricultura más productiva, sostenible y resiliente frente a la variabilidad y el cambio climático. Entre sus beneficios clave se encuentran la optimización del uso de recursos (como el agua y los fertilizantes), la detección temprana de riesgos (plagas, enfermedades, estrés hídrico), la reducción de pérdidas de los cultivos, y el aumento de la eficiencia en la producción gracias a una mejor planificación del momento de ejecución de las labores agronómicas. Asimismo, la parcela permite alimentar sistemas de alerta temprana, realizar seguimiento fenológico, generar análisis comparativos entre parcelas o tratamientos, y desarrollar modelos agroclimáticos predictivos que anticipen impactos climáticos a corto plazo. Además de su enfoque técnico y aplicado, la Parcela Agroclimática Digital también cumple un propósito educativo y formativo, funcionando como un laboratorio de demostración para estudiantes, docentes y profesionales interesados en el uso de tecnologías avanzadas aplicadas al agro. Universidades y centros de formación pueden utilizar este espacio para desarrollar competencias en agricultura digital, agroclimatología y manejo de datos, fortaleciendo así capacidades en temas clave para el desarrollo agrícola sostenible. Gracias a su diseño adaptable y orientado a la toma de decisiones basada en evidencia, la Parcela tiene un alto potencial de escalabilidad hacia otros contextos, cultivos y regiones de América Latina, favoreciendo la transformación digital del agro, la seguridad alimentaria, el desarrollo rural y la contribución directa a los Objetivos de Desarrollo Sostenible (ODS).
Digital Agroclimatic Plot: An open-field laboratory where technology and agriculture come together to transform farming
(Infographic, 2025-07-01) Barrios, Camilo; Amaya, Alejandra; Arce, Daniela
The Digital Agroclimatic Plot, located at the Campus Las Américas of the Alliance Bioversity and CIAT in Palmira, Colombia, is an experimental space that integrates Internet of Things (IoT) solutions, soil monitoring sensors, multispectral drones, high-resolution satellite imagery (PlanetScope), and real-time data visualization platforms, enabled through satellite connectivity (such as Starlink). This infrastructure allows for continuous and precise monitoring of key variables related to crops, soil, and climate, such as pH, NPK, temperature, moisture, electrical conductivity, and water table levels, through data transmission networks like LoRaWAN. Beyond data collection, this field-based laboratory aims to generate processed and contextualized agroclimatic information, which, when combined with advanced modeling techniques and artificial intelligence, facilitates informed decision-making. In doing so, it enables farmers, technicians, and researchers to optimize agronomic management strategies according to the specific agro-environmental conditions at any given time, contributing to a more productive, sustainable, and resilient agriculture in the face of climate variability and change. Key benefits include the optimization of resource use (such as water and fertilizers), early detection of risks (pests, diseases, water stress), reduction of crop losses, and increased production efficiency through improved timing of agronomic interventions. The plot also supports early warning systems, phenological monitoring, comparative analysis between treatments or plots, and the development of predictive agroclimatic models to anticipate short-term climate impacts. In addition to its technical and applied functions, the Digital Agroclimatic Plot serves an educational and training purpose, acting as a demonstration laboratory for students, educators, and professionals interested in the use of advanced technologies applied to agriculture. Universities and training centers can use this space to develop competencies in digital agriculture, agroclimatology, and data management, thereby strengthening capacities in key areas for sustainable agricultural development. Thanks to its adaptable design and evidence-based decision-making orientation, the Plot has strong scalability potential across diverse contexts, crops, and regions in Latin America, promoting the digital transformation of agriculture, food security, rural development, and direct contributions to the Sustainable Development Goals (SDGs).
Kapiti Digital Twin project: SRUC-ILRI-Bodit
(Presentation, 2025-05-20) Salavati, M.; Kemp, Stephen J.; Dhulipala, Ram
Accuracy of farmer-generated yield estimations of common bean in decentralised on-farm trials in sub–Saharan Africa
(Journal Article, 2025-06-06) Nabateregga, Mabel; Dorado-Betancourt, Hugo; Ø Solberg, Svein; Van Etten Etten, Jacob; van Heerwaarden, Joost; Gregory, Theresia; De Sousa, Kaue
Improving agricultural productivity and resilience is essential to meet future food needs in sub-Saharan Africa under changing climate conditions. Achieving this will necessitate the development of high-yielding locally adapted crop varieties to mitigate the impacts of climate change. Despite advancements in crop improvement, varietal turnover in smallholder farms remains notably low. Continuous turnover of locally adapted varieties is essential, necessitating active dissemination of new varieties and withdrawal of obsolete ones across diverse target populations using participatory breeding approaches. A decentralised experimental approach, known as tricot, supported by citizen science, has proven effective in accelerating genotype selection while promoting inclusivity and diversity. However, the methodology has strongly relied on farmer-generated rankings, which provide relative performance insights but fall short in informing breeders with absolute yield data, limiting the ability to measure genetic gain or assess economic returns on breeding investments. To address this gap, we validated the accuracy of farmer-generated yield data for common bean (Phaseolus vulgaris L.), by comparing it with technician-generated volumes and researcher-generated absolute yield data. Results revealed strong cor relations between farmer and technician volumes (r = 0.96, p < 0.001). The mean difference in farmer-technician log-yield was close to zero, indicating significant agreement. We further developed a predictive model to estimate absolute yields using farmer showing minimal influence from intrinsic and extrinsic factors. Our findings demonstrate that farmer-generated yield data can reliably inform breeding decisions and support the accelerated turnover of improved varieties. Integrating such data into breeding programs offers a cost-effective and scalable pathway to enhance agricultural productivity and sustainability across smallholder systems in sub-Saharan Africa.
Agroclimatic monitoring system at Alliance’s experimental campus
(Infographic, 2025-05-23) Barrios, Camilo; Amaya, Alejandra; Arce, Daniela; Urdinola, Jaime
At the Alliance’s experimental campus, extreme climate conditions such as droughts, heavy rainfall, and high temperatures can signiﬁcantly affect both the management and the growth and development of crops. In this context of increasing climate variability, it is essential to have accessible and easy-to-use digital tools that support technical decision-making in the ﬁeld. These digital tools allow for more efficient planning and management of resources available to farmers. Satellite imagery is leveraged to enable regular monitoring of agroclimatic conditions and crop status without needing to be physically present in the ﬁelds. Speciﬁcally, high-resolution multispectral images, such as those provided by PlanetScope satellites, help detect water excesses or shortages in a timely manner and assess the health, growth, and development of vegetation. This not only optimizes agricultural management but also promotes more sustainable and resilient production in the face of climate change.
Sistema de monitoreo agroclimático del campus experimental de la Alianza Bioversity & CIAT
(Infographic, 2025-05-23) Barrios, Camilo; Amaya, Alejandra; Arce Gomez, Daniela; Urdinola, Jaime
En el campus experimental de la Alianza, las condiciones climáticas extremas, como sequías, precipitaciones intensas y temperaturas elevadas, pueden afectar signiﬁcativamente tanto el manejo como el crecimiento y desarrollo de los cultivos. En este contexto de creciente variabilidad climática, resulta fundamental disponer de herramientas digitales accesibles y fáciles de usar que apoyen la toma de decisiones técnicas en el campo. Estas herramientas digitales permiten planiﬁcar y gestionar de manera más eﬁciente el uso de los recursos disponibles para los agricultores. Para ello, se aprovecha el potencial de las imágenes satelitales, que facilitan el monitoreo periódico de las condiciones agroclimáticas y del estado de los cultivos sin necesidad de estar en las parcelas. Particularmente, las imágenes multiespectrales de alta resolución, como las proporcionadas por los satélites PlanetScope, permiten detectar a tiempo excesos o deﬁciencias hídricas, así como evaluar la salud, el crecimiento y el desarrollo de la vegetación. Esto no solo optimiza la gestión agrícola, sino que también impulsa una producción más sostenible y resiliente frente al cambio climático.
A multi-dimensional framework for responsible and socially inclusive digital innovation in food, water, and land systems
(Journal Article, 2025-04) Opola, Felix Ouko; Langan, Simon; Arulingam, Indika; Schumann, C.; Singaraju, N.; Joshi, Deepa; Ghosh, Surajit
Digital innovations can offer solutions to various food, water, and land systems challenges globally. However, there are concerns on the ethical and social inclusivity aspects of these innovations, particularly for marginalized groups of people in less industrialised countries. In this article, we describe the design and development of a digital inclusivity framework, which builds from a detailed synthesis of inclusivity in digital literature. Key insights from the review were collated into five dimensions: risk mitigation, accessibility, usability, benefits, and participation. These dimensions can be assessed by means of twenty-one concrete and measurable sub indicators. Our focus was to enable a more holistic approach to the usually technocentric design of digital innovations. The framework, including the associated indicators, lays the groundwork for the development of a digital inclusivity index, a tool for assessing and fostering the inclusivity of digital innovations in food, water, and land systems.
Building resilience through dynamic monitoring of shocks and enhanced access to near-real-time information using citizen science and crowdsourcing techniques: A report of a national stakeholder engagement workshop in Ethiopia
(Report, 2025-03-30) Shikuku, Kelvin Mashisia; Lepariyo, Watson; Gobu, Wako; Godana, Nura; Baraza, Meshack; Ochenje, Ibrahim; Banerjee, Rupsha R.
The International Livestock Research Institute, with funding support from the Supporting Pastoralism and Agriculture in Recurrent and Protracted Crises (SPARC) programme, has established and monitored sentinel sites in five counties in Kenya and in the Borena, Afder, and East Hararghe zones in Ethiopia. The work is implemented under the Drought Index-insurance for Resilience in the Sahel and Horn of Africa (DIRISHA) project. Through the DIRISHA project, transect sites have been constructed, markets monitored, and households surveyed by “contributors” recruited using a participatory approach that includes community members and local leaders. Crowdsourced data are gathered using the KAZNET smartphone application. Dissemination of near-real-time information is achieved by combining social learning and digital innovations. This report summarizes insights generated during a participatory multi-stakeholder workshop in Ethiopia aimed at sharing DIRISHA project learnings and exploring prospects for incorporating KAZNET into climate risk management and food system transformation initiatives in Ethiopia.
A FAIR and project-oriented template for open science data workflows
(Template, 2025-04-15) De Sousa, Kaue; Laporte, Marie-Angelique
CGIAR Research Initiative on Digital Innovation: Annual Technical Report 2024
(Report, 2025-04-15) CGIAR Initiative on Digital Innovation
Digital agro‑advisory tools in the global south: a behavioural analysis of impacts, and future directions
(Journal Article, 2025-03-27) Ofosu Ampong, Kingsley; Abera, Wuletawu; Mesfin, Tewodros; Abate, Tsion
Digital agro-advisory tools have emerged as a promising solution to address pressing challenges in agriculture, particuarly for smallholder farmers in the Global South. This study provides a comprehensive characterisation of these tools, examining their behavioural typologies and adoption rates. Our desk and contextual review identified several digital agro-advisory tools in Global South. Our analysis revealed that low-adoption tools primarily focus on information dis- semination and post-harvest loss reduction. Moderate-adoption tools emphasize financial inclusion and early warning systems. High-adoption tools, on the other hand, prioritise climate-smart agriculture, farmer empowerment, and collaborative platforms. In all, our characterisation revealed five behavioural typologies necessary for the adoption of digital agro-advisory tools in the Global South. Critically, we found that trust is the fundamental foundation that determines adoption and sustained use of digital agricultural tools in the Global South. To scale up the adoption of these tools, it is crucial to address key constraints such as digital literacy, infrastructure, and policy environments. Additionally, our review shows that inclusive design and requirement elicitation are essential to ensure that these tools are accessible and relevant to the needs of smallholder farmers. By investing in digital infrastructure, promoting digital literacy, and fostering collaboration between stakeholders, we can harness the transformative potential of digital agro-advisory tools to create a more sustainable and equitable agricultural future.
Recent drought prevalence in the Limpopo River Basin: insights from the digital twin platform
(Journal Article, 2025-02) Ghosh, Surajit; Vigneswaran, Kayathri; Dickens, Chris; Retief, H.; Garcia Andarcia, Mariangel
The Limpopo River Basin (LRB), a transboundary river basin extending over Botswana, Mozambique, South Africa, and Zimbabwe, is highly vulnerable to drought. This manuscript analyzes drought conditions in the LRB using Earth Observation (EO) datasets and key drought indices such as the Standardized Precipitation Index (SPI) and Vegetation Condition Index (VCI). The year 2023, marked by the El Niño phenomenon, exacerbated dry conditions, resulting in prolonged water shortages and reduced agricultural output. Approximately 37% of the basin has been experiencing drought since the 2023–2024 cropping season, impacting ecosystems and crop yields. The present manuscript presents a comprehensive analysis of drought conditions in the LRB and applications of the Digital Twin platform for the LRB to support resource allocation for agricultural planning. Integrating multiple near real-time datasets, the platform enables policymakers to visualize and analyze drought conditions, enhancing decision-making for sustainable resource management and food security in the basin.
Development and application of the FISHTRAC real-time remote monitoring tool for digital twinning of river basins in southern Africa
(Report, 2024-12-30) Kaiser-Reichel, A.; Burnett, M.; Dickens, Christopher; McNiel, T.; Retief, H.; Süßle, V.; Garcia Andarcia, Mariangel; O’Brien, G. C.
The natural world consists of various complex physical, biological and social systems that are connected and interact with each other. New technological developments are improving the ability of the managers of natural resources, to understand and contribute to the way we are developing using and sometime abusing our resources. Through the Digital Twin for management of water resources in the Limpopo River Basin we have an opportunity to integrate available sustainable environmental flows and water resource management technology into an integrated system, that will allow stakeholders of the Limpopo River and surrounding regions to understand, monitor and manage these resources for current and future generations. Fish are good ecological indicators and have been used for over 100 years by scientists to understand how ecosystems respond to changes in environmental conditions. The development of and use of water resources for agriculture, mining and industry and urban and peri-urban communities has affected the quality, flows and habitat of rivers. Scientists routinely use established biological methods or tools to evaluate the ecological consequences of changes, but these methods are usually reactive and used after impacts occur. The FISTRAC tool has been developed through the Limpopo River Digital Twin approach to allow stakeholders to use established fish behavioural monitoring methods in real-time to evaluate changes in river condition. The approach includes the integration of radio telemetry tagging and tracking methods with real-time monitoring approaches into an online web-based system. The FISHTRAC tool monitors the behaviour of tagged fish and water quality and flow variables in the real world. This is represented in real time on the Digital Twin systems. If pre-determined abnormalities in fish behaviour is observed and is correlated to changes in river flow or water quality the FISHTRAC Tool automatically evaluates the severity of the behavioural change and as such the environmental variable change, summarizes the information and alerts users to the information in real time. This tool developed for the Digital Twin has been tested in the Sabie and Crocodile Rivers in southern Africa and shows how new technology can be used to not only monitor ecosystems, but we can consider the biota of these ecosystem and use them to determine the consequence of how we use ecosystems in real to near-real time. The FISHTRAC tool has the potential to make a considerable contribution to the sustainable water resources in southern Africa through the Digital Twin system.
Real-time application of the PROBFLO framework risk approach as a part of the digital twin towards the implementation of environmental flows in the Limpopo Basin, southern Africa
(Report, 2024-12-30) O’Brien, G. C.; Wade, M.; Bembe, M. J.; Dickens, Christopher; McNiel, T.; Retief, H.; Silva, Paulo; Garcia Andarcia, Mariangel
The natural world consists of various complex physical, biological and social systems that are connected and interact with each other. New technological developments are improving the ability of the managers of our natural resources, to understand and contribute to the way we are developing using and sometime abusing our resources. Through the Digital Twin for management of water resources in the Limpopo River Basin we have an opportunity to integrate available sustainable environmental flows and water resource management technology into an integrated system, that will allow stakeholders of the Limpopo River and surrounding regions to understand, monitor and manage these resources for current and future generations. Environmental flow (e-flow) determination tools have been designed to provide stakeholders of rivers with the flow requirements needed to protect the river ecosystems. While useful as a development control these e-flows usually only represent a small part of the sustainable development problem that resource stakeholders face. Managers have the needs of people to consider and other stressors to manage including water pollution, habitat loss, barriers, alien invasive species and climate uncertainty. Modern PROBFLO e-flow determination applications in the Limpopo River Basin are holistic and considers flow and non-flow variables of change. The PROBFLO approach also considers the risk of natural, present, e-flow and future drought flow scenarios in the context of non-flow stressors to supporting, provisioning, regulatory and cultural services. This once off PROBFLO assessment has been useful but is not adaptable in its current form and while valuable the probability risk model was used once and is not available to managers of the water resources and people of the Limpopo Basin who consider a range of alternative development and protection scenarios. The real-time application of the PROBFLO framework risk approach for the Limpopo River through the Limpopo River Digital Twin solves this problem through the establishment of and testing on an internet based PROBFLO EFA application tool. The tool combines a range of software and the PROBFLO e-flow models into a single userfriendly tool that stakeholders can use to consider any past or present water use scenario for the Limpopo River. The PROBFLO EFA application tool was developed for the Balule River site on the Limpopo River and is available for roll out to the rest of the basin. The tool allows users to change the flow scenario and condition/state of any non-flow stressor including water pollution, habitat loss, barriers, alien invasive species etc. and evaluate the risk of this new scenario to the supporting, provisioning, regulatory and cultural services established in the PROBFLO assessment. Users can use the risk results to consider the risk of a range of unique scenarios including new development options, they can consider trade-off considerations between social and ecological endpoints and between locations throughout the basin. The PROBFLO EFA application tool is a useful component of the new Digital Twin for the Limpopo River basin and will contribute to sustainable water resource management in the region.
ICTforAg Learning Network
(Report, 2024) Angus, Dawn
ECOSat (Estimation of carbon offsets with satellites) - Final report
(Working Paper, 2024-12) Schulthess, Urs; Fonteyne, Simon; Gardeazabal Monsalve, Andrea
This study aimed to assess whether radar (Sentinel-1) and optical (Sentinel-2) satellite data could detect residue management practices and differentiate between conventional, minimal, and no tillage fields in Guanajuato, Mexico. The study used in-situ data collected by the CIMMYT-led MasAgro Guanajuato project, which tracks land preparation and crop management. Various tillage and residue indices were tested, including NDSVI, NDTI, and NDI5, based on Sentinel-2 bands. The conclusion suggests that most successful remote sensing applications for tillage detection and residue management rely on survey data. These data can then be used to train machine learning based algorithms.
Clarity tubes as effective citizen science tools for monitoring wastewater treatment works and rivers
(Journal Article, 2024-09) Graham, P. M.; Pattinson, N. B.; Lepheana, A. T.; Taylor, R. J.
Improved freshwater resource management requires the implementation of widespread, effective, and timely water quality monitoring. Conventional monitoring methods are often inhibited by financial, infrastructural, and human capacity limitations, especially in developing regions. This study aimed to validate the citizen-scientist-operated transparency or clarity tube (hereafter “clarity tube”) for measuring water clarity as a proxy for total suspended solids (TSS) concentration, a critical quality metric in river systems and wastewater treatment works (WWTW) effluent in Southern Africa. Clarity tubes provided a relatively accurate and precise proxy for TSS in riverine lotic systems and WWTW effluent, revealing significant inverse log- linear relationships between clarity and TSS with r 2 = 0.715 and 0.503, respectively. We demonstrate that clarity-derived estimates of TSS concentration (TSScde) can be used to estimate WWTW compliance with WWTW effluent TSS concentration regulations. The measurements can then be used to engage with WWTW management, potentially affecting WWTW performance. Overall, these findings demonstrate the usefulness of clarity tubes as low-cost, accessible, and easy-to-use citizen science tools for high spatial and temporal resolution water quality monitoring, not only in rivers in Southern Africa but also in WWTW effluent for estimating compliance, with strong global relevance to the sustainable development goals (SDGs).
Stakeholders Workshop on Farmer-Centric Digital Transformation of African Agriculture
(Report, 2024-12) Rupavatharam, Srikanth; Patil, Mukund; Gogumalla Pranuthi; Gumma, Murali Krishna; Kumar, Kishore G.; Kumar, Shalander; Jat, Mangi L.
The science of Digital Agriculture has been gaining prominence with the advent of fast-paced technological progress that can enhance last-mile delivery to smallholder farmers. Digital Agriculture offers a wide range of technology solutions for farmers, including smart farming, precision agriculture, data-driven decision support, extension systems, channels for improved market access, and financial services (Townsend et al., 2019). Digital Agriculture can contribute to increased productivity and profitability of farms and strengthen access to diverse marketing channels and resilience to climate change. However, in most cases, Digital Agriculture technologies have not been adequately accessed by smallholder farmers, women, and youth (CTA, 2019). The United Nations Secretary General’s strategy on the use of digital technologies to accelerate the achievement of SDGs identifies food security as a critical area that will be profoundly disrupted by advances in Digital Agriculture (United Nations, 2018). The CGIAR Research Initiative on Digital Innovation has achieved substantial progress, generating 207 results, including 101 knowledge products, 43 innovations, 18 capacity-sharing activities, four policy change results, and four innovation use results. The CGIAR focused on developing innovations across five Work Packages in 2024, such as online platforms for research and learning, digital twin systems, an index for assessing digital inclusiveness, and remote sensing analytics – we are on track to achieve all End of Initiative outcome goals in 2024.
The Vision of a Digital Public Infrastructure for Agriculture
(Book Chapter, 2024-12-23) Dhulipala, Ram; Mehrotra, Nipun; Kanitkar, Ajit
The book chapter is published on page no 105-121, by Ram Dhulipala, Nipun Mehrotra and Ajit Kanitkar. The authors published a policy brief in 2023 titled “The Vision of a Digital Public Infrastructure for Agriculture” for the T20/G20, which was subsequently selected out of 300+ briefs by the Government of India to be published as a detailed book chapter in 2024 in the G20 Compendium. The chapter emphasizes the dual challenge in agriculture of enhancing smallholder productivity and incomes while ensuring environmental sustainability. Digital technologies, when implemented responsibly, can address these issues by transforming agrifood systems. A conceptual approach termed "Digital Public Infrastructure for Agriculture" (DPI4A) focuses on equitable development in G20 nations, aligning with the IDEA framework from India. Key elements include ethical safeguards, public-private partnerships, and governance mechanisms. The chapter highlights G20 leadership in fostering partnerships, global pilots, and frameworks for collaboration to address climate change and equity challenges in agrifood systems.
A scalable crop yield estimation framework based on remote sensing of solar-induced chlorophyll fluorescence (SIF)
(Journal Article, 2024-04) Kira, Oz; Wen, Jiaming; Han, Jimei; McDonald, Andrew J.; Barrett, Christopher B.; Ortiz-Bobea, Ariel; Liu, Yanyan; You, Liangzhi; Mueller, Nathaniel D.; Sun, Ying
Projected increases in food demand driven by population growth coupled with heightened agricultural vulnerability to climate change jointly pose severe threats to global food security in the coming decades, especially for developing nations. By providing real-time and low-cost observations, satellite remote sensing has been widely employed to estimate crop yield across various scales. Most such efforts are based on statistical approaches that require large amounts of ground measurements for model training/calibration, which may be challenging to obtain on a large scale in developing countries that are most food-insecure and climate-vulnerable. In this paper, we develop a generalizable framework that is mechanism-guided and practically parsimonious for crop yield estimation. We then apply this framework to estimate crop yield for two crops (corn and wheat) in two contrasting regions, the US Corn Belt US-CB, and India's Indo–Gangetic plain Wheat Belt IGP-WB, respectively. This framework is based on the mechanistic light reactions (MLR) model utilizing remotely sensed solar-induced chlorophyll fluorescence (SIF) as a major input. We compared the performance of MLR to two commonly used machine learning (ML) algorithms: artificial neural network and random forest. We found that MLR-SIF has comparable performance to ML algorithms in US-CB, where abundant and high-quality ground measurements of crop yield are routinely available (for model calibration). In IGP-WB, MLR-SIF significantly outperforms ML algorithms. These results demonstrate the potential advantage of MLR-SIF for yield estimation in developing countries where ground truth data is limited in quantity and quality. In addition, high-resolution and crop-specific satellite SIF is crucial for accurate yield estimation. Therefore, harnessing the mechanism-guided MLR-SIF and rapidly growing satellite SIF measurements (with high resolution and crop-specificity) hold promise to enhance food security in developing countries towards more effective responses to food crises, agricultural policies, and more efficient commodity pricing.

Browse

Recent Submissions