Executive Summary

The 2023–2024 El Niño event imposed a significant climate debt on key Arabica-producing regions. While agronomic conditions have partially recovered, our multi-temporal analysis of Sentinel-2 NDVI and MODIS Land Surface Temperature data reveals that moisture deficit stress persists across an estimated 340,000 hectares of prime Arabica-growing terrain in Brazil's Cerrado Mineiro and Sul de Minas sub-regions.

Key finding: Vegetation stress indices remain 1.8–2.3 standard deviations below the 2015–2023 baseline across the most climatically impacted zones — a signal historically associated with 12–18% yield suppression in the following harvest cycle.

Satellite Data Sources & Methodology

This analysis integrates four satellite data streams captured between August 2024 and February 2025:

Regional Analysis

Brazil — Minas Gerais

The Cerrado Mineiro PDO and Sul de Minas regions show the most pronounced post-El Niño stress fingerprint. NDVI anomaly maps computed against the 2015–2023 climatological baseline show persistent below-normal vegetation greenness in the critical August–November flowering and cherry-set window.

Sub-Region NDVI Anomaly (σ) Affected Area (ha) Risk Rating
Cerrado Mineiro−2.1187,000HIGH
Sul de Minas−1.8153,000ELEVATED
Chapada de Minas−1.268,000MODERATE
Matas de Minas−0.729,000LOW-MODERATE

Colombia — Huila & Nariño

The Colombian picture is more complex. Huila — responsible for approximately 14% of Colombian export volumes — shows bimodal stress patterns consistent with ITCZ displacement. The mitaca (secondary harvest) cycle appears particularly impacted, with flowering failure indicators present across the northern slopes of the Macizo Colombiano.

Futures Market Implications

Integrating these geospatial stress signals with our probabilistic yield model produces a Q3 2025 supply outlook that diverges materially from current CONAB and ICO consensus estimates. Our model indicates:

Disclaimer: This intelligence brief is provided for informational purposes only and does not constitute financial advice. Geospatial signals are probabilistic indicators subject to uncertainty. Always consult qualified commodity advisors before making trading decisions.

Methodology Notes

All satellite composites are cloud-masked using the Sentinel-2 Scene Classification Layer (SCL) and MODIS MOD35 cloud mask. Anomaly calculations use a pixel-level z-score approach against the 2015–2023 growing-season baseline. Spatial aggregation to administrative boundaries uses the IBGE/DANE municipal polygon datasets.

All Intelligence Briefs Get Full Access