blog

FAO’s Science Makes the Case for EO and AI in Resilient Agriculture

Written by Dave Gebhardt | Jan 14, 2026 4:09:56 PM

Global agriculture is in a difficult place. Production continues to climb, but the foundation that supports it is showing clear signs of stress. An FAO scientific assessment underscores this shift. Inputs are being pushed harder, with fertilizer use up about 40 percent and pesticide use doubled, and over 40 percent of rural populations now living in river basins where water scarcity is already a defining constraint.

Agriculture trade has expanded just as quickly. In the past five decades, the value of food and agricultural commodities has grown eightfold, binding national food systems more closely together. Even so, food insecurity persists. Hunger is increasing again, and more than 3 billion people cannot afford a healthy diet.

A World Changing Continuously, Not Occasionally

One of the central themes in the FAO white paper, Update on scientific findings on the interactions between agriculture, food systems and climate change, is the pace and character of environmental change.

Much of this change builds slowly and is easy to miss in a single season. Ocean warming has already lowered the maximum sustainable yield of key fish populations by a little over 4 percent in the last 80 years. Soil carbon also doesn’t follow one pattern; it changes with climate, soils and how land is managed. Coastal wetlands face rising seas, heavier storms, changing rainfall and steady development. 

Agriculture fits into this picture in a complicated way. It relies heavily on freshwater more than any other sector and it also accounts for over a quarter of the energy used in food production. Growing populations and changing diets place additional pressure on the resources agriculture depends on, and that pressure builds gradually rather than all at once. 

The FAO does not state this outright, yet the implication is clear: if agriculture is becoming more exposed, then visibility into that exposure must be continual.

Remote Sensing is Becoming the Backbone of that Visibility

The FAO assessment highlights a quiet but significant shift. The use of satellite data in national monitoring systems has expanded steadily. For many countries it is now the most consistent way to see how landscapes are evolving. These datasets support a wide range of national monitoring tasks, be it mapping land cover, tracking forest loss, estimating soil and biomass carbon, or watching changes in mangroves, croplands and coastal wetlands.

Slow shifts rarely show up from the ground. Regular satellite records give the seasonal and multi-year view needed to follow how coastlines, vegetation and restoration areas evolve.

AI Builds on What Continuous Observation Reveals

Where remote sensing offers consistency, artificial intelligence offers scale. The FAO assessment points to rapid growth in machine-learning applications that classify crops, estimate yields, identify degraded land, infer farming practices and detect stress signals before they appear visibly on the ground.

These models allow researchers and governments to analyse decades of imagery, integrate multiple sensors and pick up emerging patterns that would be impossible to detect manually. 

But the assessment also makes an important distinction. AI is not a shortcut; it is an amplifier, entirely dependent on the quality and stability of the underlying data.

Why Data Quality Still Holds Back Progress

FAO points out that many countries still don’t have the kind of consistent, well-calibrated agricultural data, including a mix of climate, soil conditions, biological cycles and management choices, needed for reliable analysis. Without that foundation, AI tools don’t perform as reliably as people hope.

Limited connectivity and computing capacity also create further obstacles, making advanced analytics difficult to scale.

Where stronger datasets do exist, AI can do considerably more. It can draw on long records and in-season observations to spot emerging stress, understand crop behaviour and support decision-support tools that link environmental conditions with management options.

The Takeaway? Resilience Begins with Measurement

Across chapters, the report makes one point repeatedly: environmental change doesn’t pause, and food systems are shaped by it constantly. Planning cannot rely on occasional snapshots.

FAO estimates that roughly 13 percent of food is lost between harvest and retail, much of it for reasons that only become clear when conditions are followed across seasons. That kind of visibility requires consistent records.

Remote sensing provides the long view. AI helps make sense of it. Field experience keeps it grounded. All three are part of the measurement base agriculture now depends on.

Where EarthDaily Fits Into This Measurement Shift

Agriculture has been central to EarthDaily’s work from the beginning, and the direction outlined in the FAO assessment closely matches the measurement principles we build around. The report highlights the need for consistent, calibrated, long-horizon visibility into land and crop conditions. 

EarthDaily’s systems are designed for that kind of continuity and for making environmental change easier to interpret at field, regional and global scales.

  • Data platform for consistent, comparable measurement: EarthOne is EarthDaily’s enterprise geospatial platform that brings calibrated, multi-source datasets into one environment and makes them usable at scale. It supports petabyte-level storage, integrates open and commercial Earth observation data and provides the computing needed to build and run analytics and AI models. For agricultural applications, this includes daily vegetation indices, soil-moisture indicators, land-surface temperature and long historical baselines, which feed into time series that show how conditions shift from season to season and over multiple years. This kind of stable, analytics-ready record reflects exactly the type of continuity the FAO identifies as essential for understanding long-term environmental change.
  • Constellation designed for daily, science-grade coverage: The upcoming EarthDaily constellation builds on this data foundation, adding a new layer of consistent, science-grade observation. The system will consist of 10 satellites, each carrying 22 spectral bands, all calibrated for measurement stability. Together, they will deliver radiometrically stable, global daily coverage designed to capture subtle shifts in vegetation, moisture and surface conditions. This combination of repeatability and spectral depth supports reliable change detection at the pace environmental shifts actually occur.
  • Agricultural solutions that turn measurement into insight: EarthDaily’s agricultural portfolio builds on a logical and proven sequence: consistent observation, structured analysis and usable insight. The data record begins with daily Earth observation, continues through AI-supported interpretation and ends with clear information that agribusinesses, insurers, lenders and traders can act on.


  • Daily monitoring and multi-year productivity histories give users a sense of how fields are developing across a season and how that compares with past performance. AI-based anomaly detection helps flag delayed emergence, drought stress or unusual growth patterns early. Insurers can rely on objective evidence for underwriting and claims; lenders use decades of field histories to understand production risk; traders follow acreage, emergence and in-season conditions using calibrated global baselines.

Across these applications, the idea is the same: consistent satellite data, strengthened by analytics, creates the level of visibility the FAO identifies as essential for resilience.

FAO doesn’t present technology as the fix, but as the means to see what needs fixing. With environmental pressures rising, that steady view is becoming as essential as production itself. 

The assessment points to a simple conclusion: agriculture resilience will depend on intelligence systems that can show how land, water and ecosystems are changing.