Lebanon Wheat Area Falls 38% as Satellites Reveal Crisis-Era Planting Swings
A new national satellite study has reconstructed eight seasons of wheat planting in Lebanon and found a sector swinging sharply with farmer economics, policy support and access to land. Wheat area fell about 52% in 2021–22, recovered to the series high in 2022–23, then contracted by roughly 34% in 2023–24 and another 38% in 2024–25.
The sequence is more informative than any single percentage. It shows why a crop-area headline cannot be read as a harvest estimate—and why rainfall alone is an incomplete explanation of planting decisions in a country where fuel, fertilizer, credit, guaranteed purchasing and safe access to fields can change faster than the weather.
Eight seasons reveal a boom-bust planting cycle
Published on 21 October 2026 in npj Sustainable Agriculture, the peer-reviewed study is the first national, multi-year monitoring of Lebanon’s wheat area from the 2017–18 through 2024–25 seasons. The researchers used Sentinel-2 imagery, Google Earth Engine and a Random Forest classifier to identify wheat at 10-metre resolution.
Season or period | Change reported | What the study associates with it |
2017–18 to 2018–19 | About 28,000–30,000 hectares | Relatively stable pre-crisis baseline |
2019–20 | +11% | Seed support, government purchasing and private finance still operating |
2020–21 | +83% versus the prior three-season average | Food-security response and new domestic demand despite higher costs |
2021–22 | −52% | Subsidy withdrawal, credit stress and surging fuel and fertilizer costs |
2022–23 | Recovery to the series high | Some economic stabilization and improved relative returns |
2023–24 | −34% | Renewed insecurity and reduced field access |
2024–25 | −38% | Displacement and access constraints, especially in the south |
These are changes in estimated cultivated area, not tonnes of grain. The study does not estimate yield, grain quality, abandonment after sowing or farm profitability. A smaller area can sometimes produce more grain if yields improve; a larger area can disappoint if water, inputs, crop health or harvesting conditions deteriorate. Treating the map as a production balance would therefore overstate what the evidence proves.
Why raw satellite pixels were not enough
The classifier achieved 87% overall accuracy and correctly mapped an average 81% of surveyed wheat fields across the four seasons with ground observations. Yet the researchers found that simply counting pixels generally understated wheat area. The gap reached 4,272 hectares—almost 14% of cropped area—in one season.
To correct that bias, they used random stratified sampling for the four seasons with reference data. For seasons without field reference points, they fitted a calibration between pixel counts and sampled estimates, then applied a static average error rate of 14%. Re-testing those seasons with the largest observed error rate, 18%, did not change the direction or significance of the overall trajectory.
That makes the trend credible while leaving the exact level less certain in seasons without ground truth. It is an operationally important lesson: satellite systems can continue monitoring when surveys become difficult, but a national crop estimate still benefits enormously from regular field observations.
Rainfall did not explain the planting swings
The study found no statistically significant relationship between seasonal precipitation and wheat area. Even the drought-affected 2020–21 season had more wheat area than the year before. The authors point to a structural reason: more than 80% of Lebanese wheat plots receive supplementary irrigation.
This does not mean water ceased to matter. Rainfall and irrigation affect crop establishment and yield, while pumping itself depends on fuel, equipment, labour and groundwater access. The finding is narrower: over this eight-season series, precipitation did not explain how much land farmers chose to put under wheat as well as the shifting economic and access conditions did.
The 2021–22 contraction is the clearest example. The researchers calculate that Lebanese-pound production costs jumped 327% in a year. Diesel rose from LBP315,000 per 20 litres in 2021 to LBP841,000 in 2022, while the cited customs data show nitrogen fertilizer moving from about $385 a tonne in 2021 to $596 in 2022. Government purchasing support weakened and farmers faced cash and credit constraints. The area response was severe even though wheat remained strategically important.
The following season demonstrates the reverse mechanism. International wheat prices were high, some domestic prices stabilized through partial dollarization and the relative economics of planting improved. Wheat area rebounded to the highest level in the series, despite production costs remaining elevated.
Field access became the later constraint
For 2023–24 and 2024–25, the authors connect the renewed decline to conflict-related displacement, labour shortages and reduced access to fields, particularly in southern wheat-growing areas. That interpretation is consistent with the 2024 FAO/WFP crop assessment, which identified displacement, infrastructure damage and movement restrictions as constraints on winter-cereal sowing and agricultural logistics.
The satellite series alone cannot assign causation to each individual parcel. Crop rotation is common, particularly between wheat, potatoes and vegetables, and farmers respond differently to prices, risk and household needs. The strongest claim supported by the evidence is that the spatial losses coincided with documented economic and access shocks, while the magnitude and regional pattern were not explained by precipitation alone.
The 2026 rebound does not erase the structural exposure
A newer FAO GIEWS brief, dated 25 August 2026, estimates Lebanon’s total cereal production at about 160,000 tonnes after favourable rainfall helped crops recover from the previous year. Domestic wheat output improved. That evidence is compatible with the satellite study because it covers a later harvest and includes yield conditions, not just planted area through 2024–25.
Yet FAO still forecasts cereal import requirements of 1.3 million tonnes for the 2026–27 marketing year, around 4% above the five-year average. The practical conclusion is not that domestic wheat can replace imports quickly. It is that local planting capacity remains a useful resilience margin whose size can change sharply when policy, input finance, irrigation costs or physical access deteriorate.
What decision-makers should monitor next
Field-reference observations collected every season, so satellite area estimates do not depend on a static error assumption.
Yield and crop-quality surveys, which are needed to translate planted area into marketable grain.
Fuel, fertilizer and seed costs expressed in both dollars and farm-gate purchasing power.
Government procurement rules, payment timing and access to working capital for small farms.
Irrigation pumping conditions and water availability, even when rainfall does not determine planting area.
Safe field access, labour availability and logistics in the Bekaa, Akkar and southern production zones.
For traders and food-security planners, the early-warning value lies in combining these signals. A satellite-detected fall in wheat area is more consequential when paired with poor yield conditions or disrupted harvesting; a rebound is more durable when farmers can finance inputs and sell the crop through predictable channels.
Sources and methodology
Primary research: Ibrahim et al., “Wheat in Crisis: Variability of Wheat Area in Lebanon 2017–2025,” npj Sustainable Agriculture, published 21 October 2026. Agrilinkage independently compared the paper’s area-estimation method, accuracy tests, uncertainty discussion and socioeconomic interpretation.
Current production and trade context: FAO GIEWS Country Brief for Lebanon, reference date 25 August 2026. The FAO brief covers the 2026 harvest and import forecast, a later period than the satellite study.
This article is an Agrilinkage Research Desk analysis of published primary sources. Agrilinkage did not conduct interviews or field observations for this report. Percentages are the study authors’ national estimates unless otherwise stated.
Comments