LoRaWAN Forest Wind Monitoring System

The true test of a well-architected IoT node is its modularity. Originally designed for indoor air quality monitoring in schools, my generic LoRaWAN hardware platform was requested by the Department of Forestry for a completely different application: investigating canopy windspeed and aerodynamics.

In collaboration with fellow engineer Keegan Hull, we adapted the core hardware and firmware to interface with cup anemometers, deploying a fleet of six autonomous devices deep within the mountainous Coetzenburg forest. Because the end-users were non-technical forestry researchers, we architected an end-to-end data pipeline that abstracted away the complexities of LoRaWAN telemetry, delivering clean, actionable data directly to the stakeholders over a continuous two-year experiment.

1. Hardware adaptation & Collaborative firmware

Adapting an indoor environmental sensor for outdoor wind measurement required specific hardware interfacing and firmware modifications. While the core Atmel and RFM95 LoRa architecture was based on my original design, I collaborated closely with Keegan Hull to execute the adaptations.

• Sensor Integration: We modified the hardware to interface securely with an external cup anemometer, ensuring the node could supply the necessary power and read the data lines without compromising the overall power budget of the device.
• Firmware Development: Working collaboratively on the embedded C/C++ firmware, we programmed hardware interrupts and timers to accurately count the physical pulses generated by the anemometer. The firmware calculated the minutely wind speeds before formatting the data payload for efficient LoRaWAN transmission.
• Power & Environmental Resilience: The enclosure was adapted to ensure the external anemometer wiring remained weatherproof. The system was completely solar-powered, with the solar charge controllers and batteries optimised to survive the shaded, varying light conditions of the forest canopy, resulting in a successful two-year continuous deployment.

2. Coetzenburg forest canopy deployment

Deploying hardware in a dense, mountainous forest introduces severe challenges for RF transmission, solar harvesting, and mechanical mounting.

• Strategic Mounting: The six devices were physically deployed and secured high up in the trees of the Coetzenburg forest. The mechanical setup had to be non-destructive to the trees while remaining rigid enough to ensure the anemometers captured true canopy wind patterns.
• RF Penetration: Transmitting 868MHz LoRa signals through dense, wet biomass (tree canopies) significantly attenuates signal strength. The network was carefully deployed to ensure the nodes maintained a reliable link to the gateways despite the environmental RF obstruction.

3. Data engineering for non-technical users

Hardware is only useful if the resulting data is accessible. The researchers in the Department of Forestry required reliable time-series data to analyse micro-climates but did not have the technical background to manage LoRaWAN payload decryption, network servers, or raw JSON parsing.

• Abstracting Complexity: I took ownership of the full data collection pipeline, bridging the gap between the raw RF telemetry and the end-user.
• Visualisation & Automated Delivery: By leveraging custom webhooks, we hosted a dedicated, user-friendly platform on ThingSpeak, allowing researchers to visualise real-time wind patterns across the mountain.
• Stakeholder Enablement: Additionally, the backend was configured to compile the decrypted payloads into clean, downloadable CSV files containing precise, minutely wind speed records for each of the six devices. This allowed the forestry researchers to bypass the technical complexities of IoT infrastructure entirely and focus on their environmental analysis.