Home / Solutions / Water Management / Remote Water Monitoring

Water telemetry in South Africa — let the borehole text you, instead of driving out to check.

Boreholes, reservoirs, dams, pump stations and river gauges are where you need data and where grid power and signal are hardest to get. addanode monitors them with solar-powered water sensors on low-power radio, buffering through load shedding — so a remote asset reports its own level, flow and status, and you stop making the trip just to read a gauge. Live dam level monitoring means you see abstraction and overflow as they happen, not on the next site visit.

Talk to our Water Team Read the solar guide →
What we deliver

Off-grid sensing that runs for years.

What we sense

Reservoir and tank level, dam levels, borehole level, flow, pressure, pump status and basic water quality monitoring — the numbers that previously needed a site visit.

Solar & low-power

A modest solar panel and battery run a low-power node indefinitely, reporting over LoRaWAN — no grid, no airtime-hungry link.

Buffer & alert

Edge buffering through outages and poor signal, with alerts on a sudden level drop, a dry-run pump or an abnormal reading.

How we scope it

Sized for the site and the sun.

  • Match the reporting interval to the decision — it drives the power budget; reporting every few hours uses far less than every 15 minutes.
  • Size the battery for autonomy — several days with no sun, so the node rides out a cloudy spell.
  • Size the panel for the worst day — to recharge on a short, overcast winter day, not just a bright summer one.
  • Buffer at the edge — store readings locally and forward when the link returns, so a connectivity gap never becomes a data gap.
  • One dashboard — remote assets sit alongside the rest of your network on the addaNet platform, with the same alerting.

The discipline that matters: size for the worst week, not the average. A node sized to survive a cloudy winter spell runs for years; one sized to summer averages fails exactly when the weather turns.

What actually kills off-grid nodes — and how to avoid it.

Remote water nodes almost never fail because the sensor was wrong. They fail on sun, on antennas, and on small installation details that only reveal themselves a season later. This is the list we design against.

What goes wrongHow you noticeThe fix
Shading nobody saw in summerReports fine for months, then drops out on winter morningsCheck the sun path, not the site photo. A tree or a tank shadow at low winter sun is invisible on a midday visit
Battery chosen on capacity aloneCapacity collapses on cold nights, or the battery ages fast in a hot boxMatch the chemistry to the enclosure’s real temperature range — a steel box in the sun ages a battery faster than its cycle count does
Antenna inside the enclosure or below the tankSignal is fine at commissioning, marginal once the tank is fullMount outside and above the metal, and test in the worst state: tank full, lid closed, gate shut
Level referenced to no datumReadings look plausible but the decisions come out wrongRecord the datum and the sensor offset at installation. A level without a datum is a number, not a measurement
Dry-run detected from level aloneA pump burns out while the reading looks normalConfirm against pump status or current — a blocked suction gives you a healthy level and no flow
Gland and cable entry facing upReadings drift months later, corrosion at the terminalsGlands down, drip loop, and check the IP rating of the assembled gland — not just of the enclosure
Reporting interval set for the dashboardBattery life comes in at a fraction of the estimateReport on change with a heartbeat. Continuous polling is what empties batteries, not the sensing
No local bufferThe gap falls exactly across the event you needed to seeBuffer at the edge and backfill on reconnect. The readings that matter most are from the hours the link was down

Two of these cause most return visits. Winter shading is the first: a node commissioned in summer carries months of margin and then fails in the first overcast week of June — which is also when the site is hardest to reach. Antenna placement is the second, and it is deceptive, because the link gets tested on an empty tank with the lid open and the asset then spends its life full and closed. Both are prevented at survey, and neither is fixable from a desk.

None of these is a sensor fault, and none of them shows up on the day Remote nodes almost never fail because the sensor was wrong. They fail on sun, on antennas and on small installation details, and each of those reveals itself on its own schedule — some at the first event that mattered, some a season later. The two that cause most return visits are the two furthest to the right. At the first event Weeks later Months later First winter No local buffer The gap falls across Dry run inferred from level alone A pump burns out Antenna inside the enclosure Marginal once the tank Reporting interval set for the dashboard Battery life at a Gland and cable entry facing up Drift first, corrosion at Battery chosen on capacity alone Capacity collapses on Level referenced to no datum Plausible readings, Shading nobody saw in summer Fails in the first Winter shading is the expensive one: a node commissioned in summer carries months of margin, then fails when the site is hardest to reach.
None of these is a sensor fault Sun, antennas and installation details — each revealing itself on its own schedule. No local buffer At the first event · The gap falls across the event you needed to see Dry run inferred from level alone At the first event · A pump burns out while the level looks normal Antenna inside the enclosure Weeks later · Marginal once the tank is full Reporting interval set for the dashboard Weeks later · Battery life at a fraction of the estimate Gland and cable entry facing up Months later · Drift first, corrosion at the terminals later Battery chosen on capacity alone Months later · Capacity collapses on a cold night Level referenced to no datum Months later · Plausible readings, wrong decisions Shading nobody saw in summer First winter · Fails in the first overcast week — when the site is hardest to Winter shading is the expensive one.
The table above, arranged by when each mistake actually reveals itself. The two that cause most return visits are the two furthest to the right — and a node commissioned in summer carries months of margin before winter takes it away.
Frequently asked

Remote-monitoring questions.

For how to size solar so it survives winter, see our solar-powered remote monitoring guide.

Yes. A monitoring node draws only a few watt-hours a day, so a small solar panel and battery run it indefinitely, off-grid — there's no grid connection to install and no load shedding to survive.
A properly sized node keeps running because the battery carries several days of autonomy and the panel is sized to recharge on a poor-sun day. The edge device also buffers readings and forwards them when the link returns, so nothing is lost.
Borehole and reservoir levels, dam levels (with abstraction and overflow alerts), pump status, flow and pressure, river gauges and remote tanks — anything remote and low-power, reporting over LoRaWAN so distance from the grid and from a tower isn't a barrier.
For a rural borehole you want one node that covers pump run/dry-run status, flow, pressure and tank level, with alarms — running on solar and battery (no grid to install), buffering data through outages, and reporting over LoRaWAN or 4G depending on what reaches the site. Submersible pressure transducers, inline flow meters and ultrasonic/level sensors are the proven sensor set. The deciding factor is rarely the sensor — it's whether it survives off-grid and stays connected from a remote location, and whether one local provider installs, supports and integrates it (SMS/WhatsApp/email alerts, dashboard or SCADA) so a failed pump triggers an alert instead of a dry reservoir. That is exactly how addanode builds remote borehole monitoring.
With a private LoRaWAN gateway placed at the nearest point that has power and backhaul — sensors reach it over kilometres with no mobile operator involved — or satellite backhaul where a site is truly beyond everything. The sensors themselves run for years on batteries or solar, so no coverage doesn't have to mean no data.

Water telemetry hardware we supply and integrate

We are vendor-neutral on the sensor and supply the class that fits the site; the table gives the typical specifications we quote against. Existing meters and transmitters are read, not replaced.

Device classTypical specification (exact model quoted per site)Comms / powerWhere it goes
Ultrasonic or radar level sensor0–10 m (ultrasonic) or 0–30 m (radar), ±0.25% of range, IP67/68LoRaWAN node or 4G RTU; battery or solarReservoirs, tanks, wet wells, borehole static level
Submersible level transmitter0–50 m water column, 4–20 mA, vented cableWired to node or RTUBorehole dynamic water level
Electromagnetic flow meterDN50–DN600, ±0.5% of reading, pulse or Modbus outputRead by node/RTU; mains or battery-powered meterBulk lines, zone inputs, abstraction metering
Ultrasonic clamp-on flow meterDN50–DN1000, no pipe cut, ±1–2%4G RTU with pulse/ModbusExisting pipelines where cutting is not an option
Pressure transmitter0–16 or 0–25 bar, 4–20 mA, ±0.25%LoRaWAN nodePump discharge, pipeline profile, DMA inlet
Pump status and currentRun/stop contact, 0–100 A CT clamp, dry-run tripLoRaWAN node or RTU digital/analogue inputsBorehole and booster pumps
LoRaWAN sensor nodeIP67, 4–20 mA / pulse / Modbus inputs, multi-year battery, reporting every 1–15 minPrivate LoRaWAN gateway with 4G backhaulBeyond mobile coverage; farms, wellfields
4G RTU / edge gatewayModbus RTU/TCP, 4–20 mA, digital I/O, local buffer for load-shedding windowsLTE Cat-1/Cat-M1; solar 20–50 W with batteryPump stations, reservoirs on the grid edge

Radio equipment is ICASA type-approved; LoRaWAN runs in the 868 MHz band per the South African plan (see our LoRaWAN guide).

Where water telemetry is deployed in South Africa

Site typeFirst measurementLink
Borehole (farm, mine, municipal wellfield)Abstraction flow + dynamic level + pump currentsite-by-site map
Reservoir or elevated tankLevel + inflow/outflow + overflowzone balance
Booster or transfer pump stationPressure + flow + pump run and currentpressure management
District metered area inletFlow + pressure, night-flow analysisDMA design
Irrigation schemeAbstraction against licence + soil moistureautomated irrigation
Off-grid environmental stationLevel, rainfall, quality probes on solarcatchment monitoring

Named with permission: Coca-Cola Beverages South Africa, where we monitor plant water on the addaNet platform. Other South African deployments — municipal and private borehole schemes, treatment works, farms and industrial sites — are described as representative cases without client names. We deliver from Johannesburg to every province, and hand over the data path as yours.

Have a remote water asset you can't easily reach?

Tell us what you need to monitor and where. We'll size a solar-powered, off-grid node to run reliably through South African winters and cloudy spells.

Book a Discovery Call Back to Water →