Last month I took Reefline on its first real field test — three Atlas Scientific EZO sensors (temperature, pH, dissolved oxygen) deployed off a mooring at 4m depth on the west coast of Faial. The deployment lasted ten days. Here's what I learned.
What worked
The core logging pipeline held up. The Raspberry Pi Zero 2 W I was using as a logger ran for the full ten days on a 20,000 mAh battery pack with readings every 60 seconds. Reefline's timestamping uses GPS PPS for sub-second accuracy, which turned out to matter more than I expected — correlating events across sensors is much cleaner with tight timestamps.
The sync worked reliably when the mooring was within Wi-Fi range of shore. I set up a hotspot on the dock and configured Reefline to sync whenever a connection was available. About 80% of readings reached the base station in near-real-time; the rest batched and uploaded when the unit surfaced.
What didn't work
The pH sensor drifted over the ten days. Atlas Scientific recommends calibration every 3–6 months in stable lab conditions, but marine environments are harder on the membranes. By day 8, the pH readings were off by 0.15 units compared to a reference measurement. I'm adding a drift correction model in the next release.
The housing seal failed on one unit at day 6. The sensor kept reading (it's rated to 20m) but the connector moisture triggered some false readings. I'm switching to SubConn connectors for the next deployment.
Data from the deployment
Average surface temperature: 18.3°C (±0.4). pH range: 8.06–8.19, consistent with open Atlantic water. DO stayed between 6.8 and 8.1 mg/L across the tidal cycle — the lower end during the afternoon when photosynthesis in the water column is highest, which is the expected pattern.