Rice Paddies and Palm Oil Drive a New High-Resolution Map of Farm Emissions

Before you can cut agricultural emissions, you need a precise map of where they come from. Not a continental average, not a national total - a field-by-field accounting of which crops, grown where, under what management practices, are releasing greenhouse gases into the atmosphere. That map has been largely missing since the last serious attempt in 2000. A study published February 13 in Nature Climate Change tries to fill the gap.

An international team led by Cornell University professor Mario Herrero synthesized satellite imagery, ground measurements, hydrological models, and national inventory surveys to calculate global cropland emissions at roughly 10-kilometer resolution. The result is a dataset that breaks down emissions not just by country but by crop type, emission source, and region - giving policymakers a far more precise tool for targeting mitigation efforts.

2.5 Gigatons, and Where They Come From

The headline figure: croplands released greenhouse gases equivalent to 2.5 gigatons of carbon dioxide in 2020. That sounds large in isolation, but context matters. Croplands occupy about 12% of global land area and account for roughly 25% of total agricultural sector emissions. The other 75% comes from livestock, land-use change, and related activities - meaning crop farming is not the dominant agricultural emitter, but it is a significant and tractable one.

Geographically, East Asia and the Pacific contributed about half of total cropland emissions. South Asia, Europe, and Central Asia together added another 30%. That concentration reflects where rice production is heaviest - flooded paddies generate methane as organic matter decomposes without oxygen, and rice is grown at enormous scale across monsoon Asia.

The study tracked emissions across 46 crop classes, but four crops - rice, maize, oil palm, and wheat - accounted for nearly three-quarters of the total. Rice led with 43% of all cropland emissions. The sources differed by crop: drained peatlands used for palm oil production generated 35% of total cropland emissions, flooded rice paddies another 35%, and synthetic nitrogen fertilizers applied to high-productivity crops like wheat and maize made up most of the remainder.

A 25-Year Data Gap

The previous global cropland emissions map dated to 2000. In the intervening quarter-century, the agricultural sector expanded substantially, management practices shifted - particularly the spread of synthetic fertilizers and the conversion of tropical peatlands for palm cultivation - and remote sensing technology improved dramatically. The old map did not capture these changes, making it an unreliable foundation for current climate policy.

"This is an absolute global synthesis of all the information you need, by country, by production system, for calculating greenhouse gas emissions - it's been a significant undertaking," said Herrero.

The new dataset integrates multiple data streams that were not available or not standardized in 2000: high-resolution satellite land cover data, updated soil carbon inventories, gridded livestock distribution maps, improved hydrological models for estimating methane from waterlogged soils, and more complete national greenhouse gas inventory reports. Combining these into a consistent global dataset required resolving inconsistencies across different measurement methodologies - a challenge the team addressed by using standardized emission factors and cross-validation against independent datasets.

High Productivity, High Emissions - and What to Do About It

One complication the analysis highlights: regions that produce the most food are often the largest emitters. East Asia's dominance in the emissions map is inseparable from its dominance in global rice and wheat production. Simply reducing emissions in high-output regions risks undermining food security if production falls as a result.

The authors argue that mitigation planning must therefore account for productivity. The metric that matters is not emissions per hectare but emissions per unit of food produced - emissions intensity. A highly productive rice paddy that uses water management practices to reduce methane while maintaining yield is a better target than a low-yield paddy that emits less in absolute terms but contributes little to feeding people.

That framing points toward specific interventions. Alternate wetting and drying - a water management technique that periodically drains rice paddies rather than keeping them flooded continuously - can reduce methane emissions by 30 to 70% with modest yield penalties when properly managed. Similarly, reducing the area of tropical peatland converted to palm oil, or restoring already-drained peatlands, could cut a substantial share of that 35% peatland contribution. Precision fertilizer application, guided by soil sensors and crop models, can reduce nitrous oxide emissions from synthetic fertilizer without cutting yields.

Limitations the Authors Acknowledge

No global synthesis of this complexity is without uncertainty. Emission factors for flooded rice paddies vary widely depending on water temperature, organic matter inputs, and rice variety - and those factors are not uniformly measured across all producing regions. Peatland drainage estimates depend on satellite-based soil carbon maps that carry their own errors. Some emission sources, particularly from small-scale subsistence farming in sub-Saharan Africa and parts of South Asia, are poorly represented in the underlying data.

The 10-kilometer resolution, while a major improvement on earlier work, is still coarser than field-scale. Local mitigation decisions - whether a specific farm should change water management, for example - require finer data than this study provides. The value of the new map is at the regional and national policy scale, not the individual farm level.

That limitation does not diminish the practical importance of the work. National governments negotiating under the Paris Agreement need credible, comparable, consistent accounting of agricultural emissions to set and track mitigation targets. The new dataset, and the methodology behind it, offers a foundation that the 2000 map could not.