Water Quality and Flood Monitoring Networks
Environmental monitoring is one of the most compelling applications for LoRa mesh networks: sensors can be deployed in remote, power-limited locations that are difficult or expensive to reach with traditional wired or cellular connectivity.
Water Quality Monitoring
Mesh-connected water quality sensors provide real-time data for rivers, lakes, reservoirs, and municipal water systems. Note that water-quality sensors (pH, turbidity, dissolved oxygen, conductivity) require regular calibration to stay accurate:
Parameters to Monitor
- pH - Atlas Scientific EZO pH circuit (~$46) plus probe (~$85), roughly $130 total (probe sold separately; prices as of 2026-06-08). pH outside roughly 6.5-8.5 (the EPA secondary drinking-water range) may warrant investigation rather than proving contamination on its own.
- Turbidity - Measures water clarity; spikes indicate sediment runoff or contamination events
- Dissolved Oxygen - Critical for aquatic life; low DO indicates algal bloom or organic pollution
- Conductivity/TDS - Total dissolved solids; elevated levels indicate industrial runoff or saltwater intrusion
- Water temperature - DS18B20 waterproof probe (~$5); temperature affects biological and chemical processes
- Water level - Ultrasonic or pressure transducer sensor; critical for flood warning systems
Hardware Architecture
A complete water quality node:
- RAK4631 base (ultra-low power nRF52840)
- Atlas Scientific EZO carrier board (I2C) for multi-parameter sensing - e.g. the Whitebox Tentacle carrier
- DS18B20 waterproof temperature probe
- IP68 fiberglass enclosure with cable glands for sensor probes
- 10W solar panel + 10Ah LiFePO4 battery. This supports autonomous operation across most of the year in favorable sun, but do not assume unconditional year-round operation: LiFePO4 cells must never be charged below 0°C (32°F) - charging a frozen LiFePO4 cell causes lithium plating and permanent damage - so a charge controller / BMS with a low-temperature charge cutoff is required for winter water-side deployments.
- Transmit interval: 15-30 minutes (low duty cycle saves power and channel bandwidth)
Flood Early Warning Systems
The National Weather Service (NWS), working with USGS streamgages, is the authoritative source for flood warnings. A community LoRa mesh is a supplementary, non-authoritative monitoring aid - it does not replace NWS/USGS warnings or official alerts. Mesh delivery is best-effort and unacknowledged; packets can drop. Use a local mesh to monitor your own property as an early local indicator, but make life-safety and evacuation decisions based on official NWS/USGS warnings and local emergency alerts.
As a supplement, a creek or river monitoring network may give downstream residents some local advance notice. Any lead time (sometimes cited as 30-120 minutes) is watershed-dependent and not a guaranteed property - flash-flood watersheds can give far less. Treat it as a best-effort early indicator, not a guaranteed warning window:
- Deploy water level sensors at upstream monitoring points (2-3 sensors per watershed)
- Set alert thresholds. The percentages below are illustrative only, not a safe universal recipe: trigger on rate-of-rise (e.g., cm per minute) as well as absolute level, set conservative early thresholds, and align cut points with the official NWS flood-stage definitions for that specific gauge (per-site calibration with the local NWS/USGS gage). A 90%-of-flood-stage "Emergency" trigger can fire too late to evacuate, especially in flash-flood-prone watersheds. Example illustrative levels: "Advisory" ~50% of flood stage, "Warning" ~75%, "Emergency" ~90% - but do not hard-code these as universal.
- Gateway node at the monitoring station forwards alerts to community mesh
- Room server stores alerts and delivers to all connected community members
- Integration with community alerting: Telegram bot, email, or siren activation
Case study framework: A network of 5 sensors along a 20-mile creek watershed, each transmitting hourly with a gateway node at the nearest road bridge, can give the downstream community a supplementary local indicator of rising water. This is an early local signal only - it does not replace NWS/USGS flood warnings, and residents should always treat official warnings and alerts as primary.
Data Management and Visualization
# Simple data pipeline for water monitoring:
# 1. Sensor node transmits JSON over LoRa mesh
# 2. Gateway node receives and publishes to MQTT
# 3. InfluxDB stores time-series data
# 4. Grafana displays dashboard with:
# - Current readings per sensor
# - Historical trend charts
# - Alert status indicators
# - Map overlay with sensor locations
# Sample query for a flood alert (threshold is illustrative only).
# Note: InfluxQL syntax shown below applies to InfluxDB 1.x;
# InfluxDB 2.x/3.x use Flux or SQL instead.
# SELECT last("level_cm") FROM water_sensors
# WHERE "location" = 'upstream_north'
# AND "level_cm" > 180 # illustrative alert threshold - calibrate per site