Agricultural Sprinkler Uniformity: Design and Run a Catch-Can Test
A sprinkler catch-can test measures how evenly water reaches a defined area under recorded operating conditions. Map the collectors, measure comparable application depths, and choose the uniformity calculation that matches the area each collector represents. Report both average depth and the named metric: Christiansen’s coefficient of uniformity (CU) and low-quarter distribution uniformity (DUlq) describe different features of the same measurements. Neither percentage alone proves adequate root-zone irrigation.
Field sprinkler irrigation machine and control cabinet. This equipment photo does not show the fixed-grid catch-can test or measured results used in the hypothetical example. Photo: Scott Bauer, USDA Agricultural Research Service.
This guide provides an original fixed-sprinkler worksheet and a separate area-weighting example. All example dimensions, catches and test durations are hypothetical teaching inputs, not results from an IrriNex installation or recommended acceptance limits. A qualified evaluator should select the field protocol for the actual sprinkler package, crop and test purpose.
Define the question and the area before setting out cans
Decide whether the test concerns an interior overlap area, a field boundary, a visibly dry strip or a whole operating block. Those questions require different sampling coverage. Label an interior result as an interior result; it cannot certify an untested corner. Keep deliberately targeted diagnostic catches separate from the representative sample so the final average does not quietly change its meaning.
Record the operating configuration: active zones, nozzle identities, sprinkler spacing, mounting height, pressure-control settings and any scheduled overlap from adjacent sets. A rotating sprinkler head can remain stationary in the field; “fixed system” here refers to the installation’s position during collection, not whether its nozzle rotates. For a full irrigation cycle, account for every contributing set using its actual duration. Do not average unlike set durations as though they were equal.
Agree on the evaluation area, method, missing-data rule and acceptance criteria before seeing the catches. A procurement requirement for CU cannot be checked by reporting DUlq without explanation. The NRCS sprinkler-irrigation handbook, printed page 11–50, distinguishes these metrics and explicitly relates the ordinary low-quarter calculation to equal represented areas.
Match the collector layout to the sprinkler system
For the worked example, define a 12 m ×12 m observation patch inside a fixed sprinkler installation. Divide it into 16 equal cells, each 3 m ×3 m and therefore 9 m². Place a collector at each cell centre, with coordinates 1.5, 4.5, 7.5 and 10.5 m along both axes. The represented area is 144 m². This is a transparent calculation geometry, not a universal can spacing, sample count or sprinkler-spacing recommendation.
Position the patch where the intended overlapping sprinkler patterns actually operate. Keep enough surrounding sprinklers running to reproduce normal contributions and hydraulic loading. Sampling only beside a nozzle or only along one wet line cannot establish the patch’s areal distribution. A larger field, irregular boundary or localized orchard pattern may require several separately defined patches and a different sampling density.
A center pivot needs a different interpretation: collectors spaced along a radius represent progressively larger areas toward the outer end. A linear-move machine can have equal-area strips under an appropriate layout, but the entire spray passage must be captured. A traveler needs its travel speed, lane spacing and adjacent-pass overlap considered. Do not reuse the stationary grid’s timer or unweighted arithmetic merely because the same cans are available.
Collect depth without confusing container size and field area
Use stable, clean collectors with known openings and readable measurements. Distinguish the collector’s opening area, used to convert volume to depth, from the field area that its location represents, used for weighting. Identical containers simplify conversion; they do not automatically make the sampling locations represent equal field areas. Set openings level at the sampling height selected for the intended measurement, and document canopy interception or obstructions.
For volume V in mL and collector opening area A in cm², depth in mm is d =10 ×V ÷A. A hypothetical opening of 100 cm² receiving 60 mL therefore records 6 mm. A different opening requires its own conversion. Reading the water height inside a tapered cup is not a reliable substitute for converting collected volume through the measured opening or using the cup’s appropriate calibration.
The NMSU catch-can audit guide explains container-area measurement, mapped collection and volume-based calculations. Its lawn-specific spacing, sample-count recommendations and rating bands are not agricultural design defaults. Select sufficient catch depth for the actual measuring resolution without allowing overflow, and record the actual collection interval.
During a water-only test, record operating pressure at defined locations and record wind speed, direction and changes during collection. Log nozzle condition, rainfall and any interruption. Compare measured pressure with the actual package requirement; one pump-gauge reading does not establish pressure at every sprinkler. Wind can distort the footprint and influence drift, while evaporation from collected water can affect the reading. A calm comparison test and a test under typical operating wind answer related but different questions.
Keep the raw catch map beside the calculation
Assume all 16 collectors in the equal-cell example receive their full catch over the same 30 min interval. The table shows depths already converted to mm. Letters identify map rows; columns retain their physical order. All readings are valid for this invented exercise. Do not rearrange the map itself when sorting a separate copy for the low-quarter calculation.
| Map row | Column 1 | Column 2 | Column 3 | Column 4 |
|---|---|---|---|---|
| A | 6 | 7 | 8 | 9 |
| B | 10 | 10 | 10 | 10 |
| C | 10 | 10 | 10 | 11 |
| D | 11 | 12 | 12 | 14 |
The sum is 160 mm across the 16 observations, so the mean depth is 160 ÷16 =10 mm. This sum is a statistical sum of readings, not a claim that the field received a 160 mm irrigation. The patch’s mean collected rate is 10 ×60 ÷30 =20 mm/h over this fixed collection interval. It is not the peak instantaneous application rate of a moving sprinkler.
A genuine dry collector remains a zero after confirming that it was correctly installed and exposed. A spilled, overturned or overflowing collector is invalid information, not zero. Retain the incident note and repeat the affected collection under comparable conditions or use the evaluator’s documented treatment. Do not remove low catches just because they lower the score, and do not silently replace them with the mean.
Calculate CU and DUlq from the same valid observations
For this equal-area sample, CU =100 ×[1 −sum of absolute deviations ÷(number of observations ×mean depth)]. The deviations from 10 mm total 20 mm: the low readings contribute 10 mm and the high readings contribute another 10 mm. Thus CU =100 ×[1 −20 ÷160] =87.5%. Use absolute differences; signed positive and negative differences would cancel and falsely suggest perfect uniformity.
For DUlq, sort the complete valid records by depth. The lowest quarter of 16 equal-area readings contains four observations: 6, 7, 8 and 9 mm. Their mean is 30 ÷4 =7.5 mm, giving DUlq =100 ×7.5 ÷10 =75%. UF/IFAS explains the low-quarter definition for sprinkler application. Calculate from your own unrounded data; do not copy a worked example’s rounding or use a CU-to-DU approximation in place of the actual records.
| Reported item | Result | What it describes |
|---|---|---|
| Mean depth | 10 mm | Average collected application over the sampled patch |
| Mean collected rate | 20 mm/h | Mean depth divided by the fixed test duration |
| CU | 87.5% | Absolute deviations across all observations relative to their mean |
| DUlq | 75% | Mean of the lowest quarter of represented area relative to the overall mean |
The different percentages are consistent. CU summarizes all absolute deviations; DUlq focuses on the lower tail. DUlq of 75% does not mean 25% of the water was wasted or that irrigation efficiency is 75%. The low-quarter mean is also not the minimum: the driest mapped cell received 6 mm. A uniformity score can remain high when every collector receives too little water.
Use area weights when observations represent unequal areas
Consider a separate hypothetical survey with four representative cells. Their areas are deliberately unequal; assume each cell’s listed depth adequately represents its whole area. This coarse example explains weighting and is not a sufficient field-sampling prescription. The weighted mean is the sum of area ×depth divided by total area.
| Cell | Represented field area | Depth | Area ×depth |
|---|---|---|---|
| W | 10 m² | 4 mm | 40 m²·mm |
| X | 20 m² | 8 mm | 160 m²·mm |
| Y | 30 m² | 10 mm | 300 m²·mm |
| Z | 40 m² | 12 mm | 480 m²·mm |
Total area is 100 m² and the weighted sum is 980 m²·mm, giving 980 ÷100 =9.8 mm. An unweighted average would be 34 ÷4 =8.5 mm and would overrepresent the small dry cell. Sort by depth, then accumulate represented area. The lowest quarter is 25 m²: all 10 m² of W plus 15 m² of X. Under the stated within-cell representation assumption, its mean is (10 ×4 +15 ×8) ÷25 =6.4 mm. Therefore the area-based DUlq is 100 ×6.4 ÷9.8 ≈65.31%.
The remaining area of X belongs outside the lowest quarter. Selecting only the driest one of four cans would confuse a quarter of observations with a quarter of area. For a radial pivot test, use the prescribed distance or annular-area weighting and boundary treatment for that layout. NRCS printed page 11–198 describes selection of the low quarter by represented area; this example makes fractional allocation explicit rather than claiming it is the complete pivot protocol.
Use the map, pressure and wind record to choose the next check
In the equal-area example, all four low-quarter catches occupy row A. Inspect that part of the patch and its contributing sprinklers first. A wind-aligned dry band, an overlap gap and an obstructed nozzle can produce different physical situations with a similar numerical score. Compare adjacent patterns, local operating pressure and the field observations before assigning a cause.
The UGA pivot-uniformity guide connects catch profiles with nozzle, pressure, movement and wind checks. Use its diagnostic approach without transferring its pivot rating bands into an unrelated fixed-system acceptance contract. If pressure needs investigation, the pressure-budget worksheet shows how to account for route losses; the sprinkler package still supplies the relevant operating requirement.
After confirming a hardware problem, consult the nozzle-replacement overview to identify the maintenance task and follow the actual equipment procedure. The sprinkler-head adjustment guide covers alignment and overlap checks. Record the changed part or setting and retest the same mapped area with comparable operating conditions. A better percentage from a different patch or wind direction is not, by itself, evidence that the repair caused an improvement.
Questions to settle before accepting the result
Will running longer correct poor uniformity?
If the spatial pattern remains unchanged and all catches scale proportionally, runtime changes depth but leaves both uniformity ratios unchanged. In the first example, multiplying duration by 4 ÷3 raises the low-quarter mean from 7.5 to 10 mm, but raises the overall mean to about 13.33 mm; the driest cell still receives only 8 mm. This is a mathematical illustration, not a runtime recommendation. Soil storage, runoff, crop demand and actual losses require a separate scheduling assessment.
Can I calculate an ordinary mean from a pivot transect?
You can describe the arithmetic mean of its cans, but you must not label it the field’s area-weighted mean unless the sampling geometry makes those equivalent. Preserve each catch’s distance, represented area and management-zone identity. Deliberately different application prescriptions also require separate evaluation of the relevant zones. Complete-pass collection and the appropriate moving-system method are essential.
Is a sprinkler test interchangeable with an emitter-flow test?
No. A sprinkler grid measures spatial application depth at collector locations; an isolated drip catch measures an outlet’s discharge over time. The drip-emitter catch-test worksheet addresses that separate measurement boundary. For an IrriNex sprinkler review, bring the raw map, opening measurements, operating-state and weather log, calculations, uncertainty notes and retest record. These make the result reviewable without pretending that one percentage describes the entire farm.
