Detecting water pollution - a sentinel sensor on conductivity and pH
The hardest thing about catching pollution is that it comes without warning - and a sample taken once a month will almost never hit that one moment when something flowed into the river. The people who protect watercourses, water intakes or valuable natural stretches know this well, and so they reach for a different mode of work: a sentinel sensor that measures non-stop and never sleeps.
There is something heartening in this role. Continuous measurement of conductivity and pH turns the water into a vigilant early-warning system: most inflows - sewage, run-off, leaks - leave a trace in these parameters before turbidity or smell become visible. A researcher who sets up such monitoring stops guessing and starts catching events red-handed, in time to establish the source and react. It is modern, very sensible water protection.
Why conductivity and pH in particular
These are sensitive indicators, fast and cheap to measure continuously:
- Electrolytic conductivity rises with the load of dissolved ions - a sudden spike is a classic signal of an inflow (sewage, salinity, run-off from roads/fields).
- pH reacts to many chemical discharges - a sharp deviation is a sign that something unusual is happening.
Neither of them will tell you what exactly flowed in - but both say excellently that it did and when, which is often enough to head into the field at the right moment.
How to set up sentinel monitoring
- Place the sensor at a strategic point - below a potential source, above an intake or a valuable stretch.
- Run continuous logging / telemetry so you can see the data live.
- Establish a baseline - the typical range of conductivity and pH for that water under various conditions.
- Define thresholds as deviations from the baseline, not “textbook” values - every water has its own norm.
- See to servicing and calibration, because a sentinel has to be trustworthy.
A threshold that catches the event
The whole value lies in the reaction time. A breach of the threshold (e.g. a sudden rise in conductivity by a set amount) is treated as a call: check, take a sample now, establish the source, notify whoever needs it. An event caught red-handed gives evidence and a chance to intervene; noticed a week later - guesswork at best.
| What you see in the data | Conductivity | pH | Most likely cause |
|---|---|---|---|
| A jump with no rain, outside working hours | sudden, sharp rise | usually unchanged or slightly down | sewage discharge / failure |
| A rise together with rain, building and fading slowly | moderate rise over hours to days | usually stable | surface run-off (road salt, fertilisers) |
| A winter rise correlated with road de-icing | clear rise, lasting for days | stable | road salt (chlorides) |
| A pH jump with no change in conductivity, during the day | stable | rises | intense photosynthesis (not a pollution event) |
| A sharp fall or rise in pH at any time of day | usually changes too | sharp deviation | chemical discharge |
These are patterns to recognise, not an automatic diagnosis - each needs confirmation by a sample and by context (time of day, weather, what lies upstream). Telling “photosynthesis” apart from “an event” matters most here, because a daytime rise in pH with no change in conductivity is normal, not an alarm.
The most common mistakes
- “Textbook” thresholds instead of a local baseline - false alarms or missed events.
- No calibration/servicing - a sentinel you cannot trust is worse than none.
- Poor placement - a sensor out of reach of the inflow will detect nothing.
- Reacting after the fact - without an immediate warning, the window to establish the source is lost.
How often
By definition continuous mode - this is not a campaign but constant supervision. The point comes from an uninterrupted series (the baseline) and from an immediate signal at the moment of deviation.
Sources and further reading
The basics of measuring conductivity and pH are described in standards (PN-EN 27888:1999, which adopts ISO 7888:1985, and PN-EN ISO 10523:2012, which adopts ISO 10523:2008), and the principles of water monitoring - in GIOŚ materials. A sentinel sensor is well complemented by periodic sampling for confirmation in the laboratory.
📊 See it live: Global Rivers & Lakes Watch - a public dashboard from three stations in three countries: the Maumee River (Ohio, USA), the river Main in Frankfurt and the LéXPLORE research platform on Lake Geneva. Water temperature, dissolved oxygen, pH, conductivity, turbidity and chlorophyll-a are fetched automatically (USGS Water Services, WSV Pegelonline, Eawag Datalakes), with correlation tiles and a station map. No login.
How to make it easier
Sentinel monitoring only makes sense when the data flows in live and when a deviation speaks up at once - otherwise the whole idea of “catching it red-handed” falls apart.
LimnoLog is made for exactly this:
- connectors / telemetry - the sensors connect so that the measurements flow in live to the station; alongside, you can pull in data from public APIs (rainfall from IMGW-PIB), because spikes often go hand in hand with run-off after rain;
- threshold alerts - you set guidelines around your own baseline, and when conductivity or pH breaches them, you get an e-mail notification - a signal that lets you head into the field before the trace disappears;
- the full series and the deviations are on a chart, ready to be shared via a link (e.g. for cooperating services) and exported as documentation of the event.
And since the same system covers, alongside conductivity and pH, turbidity, oxygen, level and any other indicator - your sentinel sees the water from many sides at once.
Early users get free access for now and keep their features for good. If you protect your watercourse or intake - take a look and try it.
See it in the LimnoLog app
Explore LimnoLog features