💡 Software-only environmental intelligence layer via satellite and public API data

[github-actions[bot]]

Summary

Provide hardware-free environmental context for every apiary by integrating satellite NDVI (vegetation health proxy for bloom estimation), growing degree day accumulation, soil moisture, and regional phenological stage data from NASA, NOAA, and USDA public APIs. This data feeds into the recommendation engine to improve seasonal timing accuracy without requiring any sensor hardware investment.

Market Signal

BroodMinder's 2026 roadmap adds soil moisture and solar irradiance measurements, but only for hardware customers ($150+/hive). HiveSense integrates BLE sensors but has no environmental intelligence without hardware. BeeSentry (ESA-backed) uses satellite imagery for commercial pollination monitoring. The gap: no consumer beekeeping app provides satellite-derived environmental intelligence. NASA Landsat (30m resolution) and ESA Sentinel-2 (10m) NDVI data are freely available via Google Earth Engine or USGS EarthExplorer APIs. This is the democratization play — environmental intelligence for every beekeeper, not just those who can afford sensor hardware.

User Signal

The PRD requires that integrations "measurably alter recommendation priority, not just display data" (Technical Success criteria). The architecture defines a telemetry ingestion pipeline via Pub/Sub. Existing idea #365 (climate-adaptive bloom forecasting) validates the concept but focuses narrowly on bloom/nectar prediction — this proposal is broader, covering vegetation health, moisture stress, growing degree day accumulation, and frost risk as a unified environmental layer that feeds the recommendation engine.

Technical Opportunity

The architecture's ExternalSignal adapter pattern and Pub/Sub topic infrastructure are purpose-built for ingesting external data sources. Google Earth Engine (for NDVI) runs natively on GCP. NOAA's ACIS and PRISM APIs provide gridded climate data including GDD accumulation. These can feed the recommendation engine's context assembly step (Cloud Memorystore pre-computation) to adjust seasonal timing recommendations per-apiary. The apiaries table already stores lat/lng coordinates needed for spatial queries.

Assessment

Dimension	Score	Rationale
Feasibility	high	All data sources are free public APIs. GCP-native (Earth Engine). Architecture already supports external signal ingestion.
Impact	high	Replaces the static "regional calendar" approach with location-specific, season-responsive environmental awareness for ALL users — not just sensor owners.
Urgency	med	Most valuable when the recommendation engine is operational. Should be designed now for integration during the recommendation engine build.

Adversarial Review

Strongest objection: Satellite NDVI is a coarse proxy for bloom status — it measures vegetation greenness, not nectar availability. The correlation between NDVI and actual nectar flow is uncertain and varies by flora type.

Rebuttal: NDVI is not a nectar-flow predictor — it is a vegetation phenology proxy that, combined with GDD and historical bloom timing, produces better bloom window estimates than calendar date alone. Academic research on pollinator habitat assessment uses NDVI routinely. Confidence scoring ensures the recommendation engine communicates uncertainty transparently. This replaces guesswork with data-informed estimation, not certainty claims.

Suggested Next Step

Build a proof-of-concept NDVI + GDD adapter for 3 test apiaries across different USDA zones. Compare bloom window predictions against historical inspection records and known bloom timing to validate correlation strength before integrating into the recommendation engine.