3.3.4 — Agricultural Risk Intelligence — maturity: live

Crop Failure Analytics

Combining multi-spectral vegetation indices, soil moisture retrievals and yield-model assimilation to detect, quantify and forecast crop failure at field scale before harvest.

Auto-drafted reference page — content reviewed for shape, individual references not yet link-checked.

Governments that rely on commercial crop estimates—or worse, on anecdotal reporting from provincial offices—are perpetually behind the curve when a failure is unfolding. By the time a shortfall is declared, import tenders are late, food prices are already moving and social pressure is building. A sovereign constellation fusing optical NDVI/EVI time-series, SAR-derived soil moisture and thermal stress indicators can flag anomalous canopy decline weeks ahead of harvest, giving policymakers an actionable lead time that no ground survey can match.

The satellite stack works in layers. High-revisit multispectral imagery at 5–10m resolution tracks greenness trajectories against a climatological baseline; deviations beyond one standard deviation trigger automated anomaly flags. SAR backscatter fills the gap when cloud cover—common during the very monsoon and growing seasons that matter most—blocks optical sensors. These signals are ingested into a crop growth model (e.g. DSSAT or ORYZA) running on sovereign infrastructure, producing sub-national yield forecasts with confidence intervals updated every five days.

The operational outcome is a quantified failure estimate—hectares affected, expected yield deficit in tonnes, affected administrative units—delivered to the ministry of agriculture and the national food security authority simultaneously. That output drives strategic grain reserve drawdowns, targeted social-protection top-ups and World Food Programme coordination before a crisis becomes a famine. Owning the full data stack means the government can publish or withhold forecasts on its own schedule rather than waiting for a foreign vendor's subscription report.

What matters

A five-day revisit cycle at field scale closes the gap between ground surveys, which typically run once per season and miss in-season failure events.
SAR soil moisture retrievals remain unaffected by cloud cover, making the system operationally reliable during monsoon and tropical growing seasons when optical satellites go blind.
Yield-model assimilation translates raw spectral anomalies into tonne-level deficit estimates that treasury and procurement desks can act on directly.
National control over when and how forecasts are published prevents market manipulation by actors who would exploit early access to a foreign vendor's crop report.

Sovereignty score: 9/10 — Crop failure forecasts are a sovereign instrument of food security and political stability—no nation can afford to have them gated behind a foreign commercial subscription or a vendor's publication schedule.

Foreign vendors routinely embargo or delay granular sub-national crop reports for commercial reasons, leaving governments without actionable data at the precise moment they need to trigger emergency procurement or release strategic reserves.
Commodity markets react to crop forecast leaks; a government whose own estimates originate on foreign infrastructure loses control of information timing and exposes its sovereign grain procurement to front-running by better-informed traders.
Export-control and sanctions regimes can suspend commercial satellite data services at short notice—exactly the scenario that tends to coincide with regional conflict or diplomatic pressure, when food intelligence is most critical.
National yield models must be calibrated to local crop varieties, planting calendars and soil types; only a sovereign system can maintain the ground-truth training data and model parameters as proprietary national assets.

Reference architecture

Payload: Multispectral imager: 8 bands (Blue, Green, Red, Red-edge ×2, NIR, SWIR-1, SWIR-2), 5m GSD, 60km swath; secondary C-band SAR module, VV+VH polarisation, 20m IW mode for soil moisture retrieval
Bus class: ESPA-class microsat, 120–150kg, 600W payload power; multispectral and SAR payloads co-manifested on a shared bus or flown as complementary sub-constellations of 16U cubesats carrying optical-only payloads
Orbit: Sun-synchronous LEO at 520–560km; 18-satellite walker constellation (12 optical + 6 SAR) delivering 5-day full-coverage revisit nationally, 2-day revisit over priority breadbasket zones
Ground segment: 3-station national network with X-band high-rate downlink (320 Mbps) and S-band TT&C; primary station co-located with national meteorological service; SatNOGS UHF/VHF backup for housekeeping telemetry
Software & data
Latency
Cost band
Lead time

References

FAO Global Information and Early Warning System (GIEWS) methodology — GIEWS describes the multi-source remote sensing approach—NDVI anomaly, rainfall estimates, yield modelling—used for global crop monitoring and early warning of food crises.
ESA WorldCereal: global crop type and crop condition mapping — WorldCereal demonstrates continent-scale, field-level crop mapping using Sentinel-1 SAR and Sentinel-2 multispectral time-series, providing a validated baseline for national failure analytics systems.
USGS / NASA Famine Early Warning Systems Network (FEWS NET) remote sensing — FEWS NET integrates MODIS and Landsat vegetation anomaly products with rainfall and soil moisture data to produce sub-national food-security outlook reports used by governments and donors.
WMO Agricultural Meteorology Programme: crop monitoring and yield forecasting — WMO's AgMet programme documents best-practice integration of satellite-derived agroclimatic variables into national crop yield forecasting workflows.
IFPRI: Early warning systems and food price volatility — IFPRI research quantifies the link between timely crop failure detection and reduced food price volatility, showing that a 30-day earlier warning cuts emergency import costs by 15–25%.