Skip to content

Weather

Status: planned - Not implemented · Tier: 2 · Instances: Single

TO-BE: Environmental data service that downloads weather observations from external APIs and local sensors, stores them in the database, and enables correlation of acoustic activity with environmental conditions.

1. The Problem / The Gap

  • No Environmental Context: Without weather data, acoustic recordings exist in isolation. It is impossible to correlate species activity with temperature, humidity, wind speed, or barometric pressure.
  • Manual Data Integration: Researchers must manually download weather data from third-party sources and align timestamps — a tedious and error-prone process.

2. User Benefit

  • Automated Enrichment: Weather data is automatically fetched and stored alongside recordings, enabling queries like "Show all BirdNET detections during rain events."
  • Behavioral Insights: Correlate dawn chorus timing with sunrise, bat activity with moonphase, or silence periods with approaching storms.

3. Core Responsibilities

Inputs

  • External APIs: Periodic polling of weather data providers (e.g., OpenMeteo, DWD Open Data) for the station's geographic coordinates.
  • Local Sensors (Future): Reads from BME680 (temperature, humidity, pressure, VOC) or SPS30 (particulate matter) via I²C/SPI if hardware is connected (see Future Ideas).

Processing

  • Data Normalization: Converts provider-specific formats to the canonical weather table schema.
  • Deduplication: Ensures no duplicate observations for the same timestamp and source.
  • TimescaleDB Partitioning: Inserts into the weather hypertable with time-based partitioning.

Outputs

  • Database Rows: INSERTs into weather table (TimescaleDB hypertable).
  • Redis Heartbeats: Via SilvaService base class (ADR-0019).

4. Operational Constraints & Rules

Aspect Value / Rule
Immutable Yes — config at startup, restart to reconfigure (ADR-0019)
DB Access Yes — writes weather table, reads system_config
Concurrency Async event loop, periodic fetch (e.g., every 5–15 min)
State Stateless — last-fetched timestamp derived from DB
Privileges Rootless (ADR-0007)
Resources Low — infrequent HTTP requests, minimal CPU
QoS Priority oom_score_adj=500 — expendable analysis service

5. Configuration & Environment

Variable / Mount Description Default / Example
SILVASONIC_WEATHER_PORT Internal health endpoint 9500
SILVASONIC_WEATHER_PROVIDER API provider openmeteo
SILVASONIC_WEATHER_LATITUDE Station latitude 51.9607
SILVASONIC_WEATHER_LONGITUDE Station longitude 7.6261
SILVASONIC_WEATHER_POLL_INTERVAL Seconds between fetches 900 (15 min)
SILVASONIC_WEATHER_VARIABLES OpenMeteo Hourly Vars temperature_2m,relative_humidity_2m,surface_pressure,wind_speed_10m,precipitation

6. Technology Stack

  • HTTP Client: httpx (async)
  • Data Provider: OpenMeteo Free API (no API key required)
  • Database: sqlalchemy (2.0+ async), asyncpg
  • Future Sensors: smbus2 or adafruit-circuitpython-bme680 for I²C sensors

7. Open Questions & Future Ideas

  • Multiple weather providers for fallback / cross-validation.
  • Local sensor support: BME680 for hyperlocal microclimate data (see Future Ideas).
  • Lightning detection: AS3935 sensor for storm proximity correlation.
  • Astronomy data: Moon phase and sun position from Suncalc for nocturnal activity analysis.

8. Out of Scope

  • Does NOT analyze audio (BirdNET / BatDetect's job).
  • Does NOT record audio (Recorder's job).
  • Does NOT provide a UI (Web-Interface's job).
  • Does NOT perform weather forecasting — only stores observations.

9. References