Frozen Troughs and Burst Pipes: Winter Water Monitoring Without Walking the Fields

Winter stockmanship on an outlying farm is largely a water problem. Stock can handle cold. What they cannot handle is no water, and in a hard frost a trough goes from fine to solid faster than most people expect, particularly a shallow one on an exposed ridge. Meanwhile at the other end of the system, a pipe that splits under a track or in an unheated building will quietly run for days, and you find out from the water bill or from the field that has turned into a bog. Both of those are the same kind of problem: something you can only confirm by physically going and looking, at the time of year when going and looking means a frozen quad run before daylight. And both are now cheaply detectable remotely. Last updated: November 2026. What is actually worth monitoring Not everything. Four things carry almost all the value. Water temperature at the trough. The honest early warning. A trough heading towards zero is a trough that will be solid by morning, and knowing that at nine in the evening is worth a great deal more than discovering it at eight the next day. A temperature sensor in the water tells you the state of every trough from the kitchen. Trough or tank level. This catches the other failure: a ballcock stuck shut, a supply that has frozen further back, or a trough somebody knocked. Stock drink it down, nothing refills it, and the level simply keeps falling. A level reading that does not recover overnight is an unambiguous fault. Flow on the main. This is the burst-pipe detector, and it is the one that saves the most money. Normal winter flow on a stock supply is intermittent: stock drink, the troughs refill, flow stops. Continuous flow that never stops is a leak, and nothing else. A flow sensor that alerts on sustained flow over a set period will find a burst within hours instead of within a billing cycle. Temperature in buildings with plumbing. The dairy, the pump house, the sprayer store, the water heater in an unheated shed. Anywhere a pipe can split. One temperature sensor per building with a threshold alert at two or three degrees gives you a night's warning to do something. Why this is a sensor job, not a camera job Worth being clear about, because we sell cameras and this is not a camera application. A camera at a trough tells you that stock came and went. It does not tell you the water temperature, the level, or whether the supply is flowing. You could in principle look at a gauge through a lens, and it is a fragile, power-hungry way to read a number that a sensor can report directly for years on a battery. This is what LoRaWAN is for, and winter water is close to its ideal use case. The sensors are tiny, battery-powered, and send a handful of bytes a few times an hour, which is why they run for years rather than days. The range is in kilometres rather than metres, so one gateway at the farmstead covers outlying fields that have no power and no signal worth speaking of. And there is no SIM and no data plan per device, which matters when you want fifteen of them rather than two. Our guide to LoRaWAN for the land covers how the network goes together, and choosing the right network for a large estate sets it against 4G and HaLow. The practical division of labour on a winter farm: sensors tell you the state of things, cameras tell you what happened. Both report to a phone, and they are complementary rather than competing. Your solar 4G cameras still watch the gate and the fuel tank. The cold weather problem that catches people out If you are running anything battery-powered through a British winter, two things deserve attention. Battery chemistry slows down when it is cold. A sensor rated for years of life is rated at room temperature. In a hard frost, capacity drops temporarily and a marginal battery will report a low voltage or miss a transmission at exactly the moment you most want it reporting. Specify the cold-rated option where the sensor sits fully exposed, and replace batteries in autumn rather than waiting for them to fail in February. Solar stops being a power source in December. This is the one that undoes off-grid installs. A solar-powered device sized on an annual average will run flat in the darkest weeks, which is precisely when a frost alert matters most. Our post on whether solar CCTV gets through a UK winter works through sizing for the worst month rather than the average one, and the same arithmetic applies to a solar-fed gateway. For the sensors themselves, a primary battery with a multi-year life is usually the simpler answer than a panel the size of a beer mat. The gateway is the part that needs real power and a backhaul, so put it where there is a supply: the farmhouse, the workshop, a building with mains. Sensors then need nothing at all. Setting the alerts so you can trust them The failure mode of any monitoring project is alerts nobody believes, and that happens within a fortnight if the thresholds are wrong. Alert on trend, not on a single reading. One cold reading at four in the morning is weather. A trough that has dropped steadily for three hours and is approaching zero is a problem. Set the rule so it fires on the second, not the first. Give yourself a night, not a notification at the moment of failure. A frost alert at two degrees and falling is actionable. An alert at zero is a report. Flow alerts need a duration. Sustained flow for thirty or sixty minutes with no stock movement, not simply flow. Separate urgent from informational. A burst main and a frost warning should reach your phone. A battery at forty per cent should appear in a weekly summary and nowhere else. Decide who the alert goes to, especially where there is relief cover or a contract shepherd. An alert sent only to a phone that is in a drawer on a Sunday is not a system. What it changes on the ground The honest version: this does not stop troughs freezing. Nothing stops troughs freezing. What it changes is the amount of driving. On an outlying block with ten troughs, winter checking…