From climate monitoring to carbon accounting: Remote sensing powers SDG-aligned agriculture

Among the most significant recommendations is the development of multi-scale monitoring networks that seamlessly integrate ground-based sensors, UAV observations and satellite data. Such networks would enable more complete tracking of respiration dynamics across temporal and spatial scales, capturing both short-term events and long-term trends.

Rising global temperatures, intensified extreme weather, and shifts in land-use practices are reshaping the way croplands and grasslands breathe out carbon, challenging long-held assumptions about food system sustainability and climate mitigation strategies, according to a new review published in Agriculture.

The study, “Remote Sensing-Based Advances in Climate Change Impacts on Agricultural Ecosystem Respiration,” analyses how Earth-observing technology and machine learning are reshaping carbon monitoring between 2021 and 2025. The paper evaluates 222 peer-reviewed publications and unveils the emergence of a data-rich, technologically enhanced era of agricultural carbon science.

Remote sensing and AI emerge as key tools in tracking carbon loss from farmlands and grasslands

The report identifies agriculture as a critical component of the global carbon cycle, responsible for substantial emissions through soil processes, microbial respiration, and vegetation activity. As climate change alters temperature patterns, precipitation cycles and freeze–thaw rhythms, agricultural respiration becomes increasingly variable and difficult to predict using conventional monitoring tools. The authors argue that traditional ground-based measurements, while accurate, lack the scale and temporal coverage necessary to capture rapid climate-driven shifts across diverse agricultural landscapes.

The study documents a dramatic surge in the use of remote sensing technologies, including satellite-based thermal sensors, hyperspectral imaging, microwave instruments and UAV-mounted systems, to monitor carbon fluxes over large areas. These tools provide continuous data on vegetation growth, soil moisture, canopy temperature and stress conditions, enabling more precise modelling of respiration at regional and global scales.

Machine learning, identified as the largest and fastest-growing research cluster in the bibliometric analysis, is transforming this field. Algorithms ranging from random forests to neural networks are now essential for integrating multi-source remote sensing data, downscaling global climate datasets and predicting carbon fluxes with improved accuracy. The authors report that AI-driven models reduce uncertainty in respiration estimates by up to 40 percent compared to traditional inversion methods.

Research hotspots that dominate recent publications include permafrost agriculture, alpine grassland carbon fluxes, drought-related respiration anomalies and eddy covariance applications. These areas reflect growing concern over climate-sensitive regions where warming accelerates soil carbon loss and destabilizes long-term carbon storage capacity. Permafrost and alpine systems show particularly strong climatic responses, with respiration increases of 15 to 25 percent in degraded zones.

The review finds that agricultural grasslands, especially those with heavy grazing or mowing pressure, have become a focal point for evaluating carbon stability. Land management strategies such as grazing prohibition demonstrated measurable benefits, reducing respiration and enabling soils to accumulate more carbon. These findings highlight the link between ecological management and climate adaptation in agricultural systems.

Farmland ecosystems also experience climate-driven respiration pulses linked to freeze–thaw cycles and drought episodes. Conservation tillage practices were shown to decrease soil respiration by 15 to 25 percent, suggesting that management choices play a major role in balancing productivity with carbon retention. Optimized irrigation and crop management strategies further stabilize seasonal respiration fluctuations by reducing moisture stress.

The paper traces the evolution of the field across three phases. Between 2021 and 2022, research was characterized by traditional flux monitoring and foundational observational work. From 2023 onward, rapid growth occurred as remote sensing tools became more widely deployed. By 2025, the field had entered a stage defined by integration, where AI, multi-source data assimilation and advanced process-based modelling began working in combination rather than isolation.

This technological shift has profound implications. Remote sensing offers a consistent, scalable foundation for monitoring carbon changes across vast agricultural regions, while AI enhances the interpretive power of these datasets. Together, they provide a more complete picture of respiration dynamics and their links to climate change, management practices and ecosystem resilience.

China and the United States lead global research as collaboration networks expand

The bibliometric analysis highlights a dramatic rise in global scientific output on agricultural respiration, with China and the United States emerging as dominant contributors. China accounted for 135 publications during the review period, reflecting significant national investment in carbon neutrality initiatives and satellite technology. The United States produced 60 publications and remains a key innovator through institutions such as NASA and the Department of Energy, which support long-term flux tower networks and high-resolution remote sensing missions.

International collaboration networks expanded significantly between 2021 and 2025, forming dual-core research hubs centered in East Asia and North America. European partners, including France, Norway and Scotland, played critical roles by advancing modelling frameworks and contributing to cross-regional ecosystem comparisons. The Chinese Academy of Sciences was identified as the most influential institution, with strong co-authorship links that extend across continents.

Keyword clustering revealed 13 major research directions shaping the global discourse. Alongside machine learning and carbon flux modelling, clusters included permafrost degradation, alpine meadow carbon dynamics, drought stress responses, satellite-based NPP estimation, and eddy covariance calibration techniques. These clusters represent interdisciplinary convergence of ecology, atmospheric science, agronomy and data science.

The burst keyword analysis further showed how research priorities evolved over time. Early work focused heavily on soil respiration and moisture interactions. From 2023 onward, the emphasis shifted toward multi-process coupling, integrating vegetation productivity, water-use efficiency, thermal anomalies and land-use transitions. By 2025, emerging themes centered on intelligent modelling and the creation of cross-scale monitoring frameworks that combine ground stations, UAVs and satellites.

Grassland ecosystems were among the most intensely studied systems due to their sensitivity to climatic variation and management intensity. The paper highlights that mowing and grazing influence carbon dynamics through vegetation structure, soil temperature, and microbial activity. Grassland degradation, particularly in alpine and semi-arid regions, accelerated carbon losses as warming intensified respiration rates. The authors emphasize that restoring degraded grasslands through targeted management can significantly improve carbon sequestration capacity.

Permafrost agricultural systems, though less widespread, showed some of the strongest climate responses due to thaw-induced carbon release. Freeze–thaw cycles destabilized soil carbon pools, generating respiration surges that remote sensing tools increasingly captured through thermal and microwave observations. These findings underline the importance of monitoring high-latitude agricultural zones as global warming accelerates.

The authors stress that the rapid progression of this field reflects both technological advancement and growing urgency as agriculture faces mounting environmental pressures. The rising frequency of droughts, heatwaves and extreme seasonal transitions intensifies respiration variability, increasing the need for continuous monitoring and early-warning capabilities.

Future directions call for integrated monitoring networks and intelligent modelling systems

The study outlines several critical research and policy priorities aimed at improving accuracy, reducing uncertainty and strengthening global climate resilience. Among the most significant recommendations is the development of multi-scale monitoring networks that seamlessly integrate ground-based sensors, UAV observations and satellite data. Such networks would enable more complete tracking of respiration dynamics across temporal and spatial scales, capturing both short-term events and long-term trends.

The authors advocate for the expansion of grey-box models, which combine the mechanistic insights of process-based models with the predictive strengths of machine learning. These hybrid models can reduce uncertainty by leveraging the strengths of each approach while minimizing their weaknesses. With agricultural respiration influenced by a complex interplay of temperature, moisture, vegetation structure and human activity, grey-box frameworks offer a path toward more realistic carbon simulations.

Another major recommendation involves strengthening research on extreme climate impacts. As global warming drives more frequent and severe heatwaves, droughts and freeze–thaw cycles, understanding how these events reshape respiration dynamics is essential for climate adaptation planning. Remote sensing is poised to play a central role given its ability to detect rapid changes that ground-based systems may miss.

The study also calls for more attention to management-driven mitigation strategies. Practices such as conservation tillage, optimized irrigation, grazing management and revegetation programs have measurable impacts on respiration and soil carbon storage. Quantifying these impacts through remote sensing provides evidence-based guidance for farmers and policymakers seeking to balance production goals with carbon reduction commitments.