How to calculate application uniformity for a pivot
Application uniformity is the number that tells you whether a pivot is putting water where it's supposed to go, evenly across every sector, or whether some ground is getting soaked while other ground goes dry. It's the same question a dry spot or a ring of over-irrigation around a tower is asking you to answer, just expressed as a single percentage instead of a problem you spot driving the circuit.
The catch-can test
The standard field method is still the catch-can test, and most district engineers learned it the same way: you set a line of catch cans on a radial line from the pivot point out to the last tower, usually one can every few feet near the center point and wider spacing as you move out, since the same angular sweep covers more ground area the farther out you go. Run the pivot through the line (either a full revolution or a controlled pass, depending on your setup), then collect the cans and measure what each one caught.
From there you calculate Christiansen's uniformity coefficient, the formula most people mean when they say "CU":
CU = 100 × (1 − (average absolute deviation from the mean catch) ÷ (mean catch))
A CU above 85 is good uniformity for a pivot. Below 70, you have a real problem: a worn or plugged nozzle package, a failed pressure regulator, or an end gun that's not matching the rest of the circuit.
Some engineers prefer distribution uniformity instead, specifically the low-quarter DU: average the catch from the lowest 25% of cans, divide by the average catch across all cans, multiply by 100. DU_lq runs more conservative than CU because it weights the worst-performing part of the pattern instead of averaging everything together, which matters if you're trying to catch a dry spot rather than just describe the pattern as a whole.
Why the pivot application rate isn't flat
First-time catch-can testers usually trip on the same thing: application rate isn't supposed to be the same at every nozzle. Because the area swept increases as you move out from the pivot point, the nozzle package near the end of the span has to deliver more water per minute than the one near the center just to apply the same depth. Nozzle sizing is designed around that geometry already. A catch-can test measures how closely the actual catch matches that design curve, can by can, against the depth the geometry calls for at each point along the line.
That's also why a uniformity problem on one pivot doesn't look like a uniformity problem on another. A plugged nozzle three towers out shows up as a much bigger dry wedge than the same plug near the pivot point, because of how much ground that one nozzle is responsible for covering.
The tradeoff with catch-can testing
The test works. It's also slow enough that most districts only run it once a season, or after a repair, because it means driving out, setting cans by hand, running the pivot, collecting cans, and doing the math, all on a day calm enough that wind isn't skewing the catch. In the stretch between tests, a nozzle can fail or a gasket can start weeping and nobody catches it until the crop shows stress or someone notices a soggy ring driving the circuit.
That gap is the reason we built a weekly check against each pivot's own geometry instead of a once-a-season catch-can run: comparing index and thermal patterns to where water should be landing by design, so a dry sector or an over-watered one shows up as a flagged pattern before it's obvious from the road.
If you're already running catch-can tests and want something that catches problems in the weeks between them, that's the gap Irrigation Monitor is built to close.