Shower Head Maths: Flow, Spray & Water Use

Why a Shower Head Is a Mathematics Lesson

A shower head turns one stream inside a pipe into many jets in the air. That everyday change brings together flow rate, pressure, nozzle area, spray geometry, droplet motion, water use and user experience. Mathematics helps us compare those quantities without assuming that the largest flow, the smallest holes or the widest spray must be best.

This article explains why mathematics is important in everyday life through shower heads. It is about measurement and design reasoning, not plumbing work. Do not dismantle a pressurised fitting, alter a heater, block outlets, measure scalding water or stand on a wet surface for an experiment. A safe investigation uses cold or comfortably cool water, ordinary collection containers, published specifications, photographs, paper geometry and teacher-provided datasets.

Water and electricity can be a dangerous combination. Keep electronics away from wet areas, ask an adult to manage household fixtures and follow manufacturer instructions. If a shower has unsafe temperature, pressure, leakage or electrical behaviour, stop using it and seek qualified help. Mathematics should make a study safer and clearer, never more adventurous around hazards.

Flow Rate, Volume and Water Use

Flow is a rate, not a feeling

Volumetric flow is Q=V/t. If a container collects 6 litres in 120 seconds, average flow is 6/2=3 L/min. If it collects 7.5 litres in 90 seconds, flow is 5 L/min.

Users often describe a shower as “strong,” but that sensation can depend on jet speed, droplet size, spray pattern, distance and skin area as well as total flow. A high flow can feel weak if spread broadly; a lower flow can feel concentrated. The number and the experience answer related but different questions.

This distinction is a valuable numeracy habit. A measurement should be named precisely. “Five litres per minute” is reproducible under stated conditions. “Strong” is a human judgement that may still matter, but it needs its own method.

Measure with enough time to reduce relative error

Suppose a timing error is about ±1 second. For a 10-second collection, that is roughly ±10% of the interval. For a 60-second collection, it is roughly ±1.7%. A longer interval can reduce the relative effect of reaction time.

Container markings also have uncertainty. If volume can be read only to the nearest 0.1 L and the sample is 1.0 L, the resolution is 10% of the reading. Collecting 5.0 L reduces that ratio to 2%, provided the container is safe to handle.

Students should not chase precision by filling an awkwardly heavy container or working on a slippery floor. Safety sets the constraint. Several moderate, safe repeats are better than one risky “perfect” measurement.

Convert litres per minute into session volume

Water used is flow multiplied by time. At 7.0 L/min for eight minutes, volume is 56 L. At 5.5 L/min for six minutes, it is 33 L.

The first scenario uses 23 L more per shower. Over 30 showers, the simple difference is 690 L. This calculation assumes the same average flow throughout and ignores other household uses.

An arithmetic result does not tell a family how long a person should shower. Comfort, hygiene, health needs and household context vary. It shows the consequence of a chosen rate and duration.

Duration and flow can compensate

If a target session volume were 40 L, a 5 L/min flow corresponds to eight minutes, while 8 L/min corresponds to five minutes. The equation t=V/Q shows the inverse relationship.

However, equal water volume does not guarantee equal energy use, comfort or effectiveness. Water temperature and heater efficiency affect energy. Spray distribution affects how quickly a user rinses. A lower-flow device that encourages much longer use may save less water than its label suggests.

This is an example of behavioural rebound. A technical change affects behaviour, and behaviour affects the total. A fair evaluation measures the whole session rather than only the rated flow.

Cumulative plots reveal interruptions

Imagine cumulative water readings every minute: 0, 5.5, 11.0, 16.5, 16.5, 22.0 and 27.5 L. The flat interval shows the water was off for one minute. A single final total of 27.5 L over six minutes would give 4.58 L/min and hide the active-flow rate of 5.5 L/min.

Both numbers can be useful. Average session flow includes the off interval; active flow describes the device while running. The analyst should state which denominator is used.

This principle transfers to study time, machine utilisation and transport. “Per elapsed hour” and “per active hour” can differ greatly.

Did You Know? A small rate difference accumulates

A difference of 0.5 L/min seems modest. Across an eight-minute shower it is 4 L. Across two showers per day for a year, a simple calculation gives 0.5×8×2×365=2,920 L.

The calculation is a scenario, not a prediction. Actual duration, flow and occupancy change. Its value is to reveal how a rate becomes a total through multiplication by time and frequency.

Students should label projections clearly. A forecast based on fixed behaviour is not observed consumption.

Pressure, Nozzles and Flow Control

Pressure difference drives flow

Water moves through a shower head because upstream pressure exceeds the pressure at the outlet. The head, hose, valves and piping consume part of that pressure through losses.

For an ideal opening, a simplified relation is Q=C_d A√(2Δp/ρ), where C_d is a discharge coefficient, A is total open area, Δp is pressure difference and ρ is fluid density. The square root means doubling pressure difference does not double flow in this ideal model; it multiplies flow by √2, about 1.414.

Real shower heads can include regulators, elastic elements, swirl chambers and complex jets. The formula is a starting model, not a product certification.

Total nozzle area is the sum of many small areas

If there are n circular holes of diameter d, total ideal open area is nπd²/4. With 60 holes each 0.8 mm in diameter, one hole has area π(0.8²)/4≈0.503 mm², and total area is about 30.2 mm².

If diameter falls by 10% from 0.8 to 0.72 mm while the count stays fixed, area per hole falls by 0.72²/0.8²=0.81, or 19%. Area depends on the square of diameter.

That does not prove flow falls by 19%. Pressure, coefficient and regulation can change. Geometry identifies sensitivity; system behaviour determines the operating result.

Jet speed differs from total flow

Average speed through the ideal open area is v=Q/A. If flow is 6 L/min, that is 0.0001 m³/s. If total open area is 30 mm², or 30×10⁻⁶ m², average exit speed is about 3.33 m/s.

If the same flow passed through only 20 mm², ideal average speed would be 5 m/s. Concentrating area can raise speed while total water use remains the same.

The calculation helps explain why sensation is not determined by L/min alone. Yet actual jets contract, interact and break into droplets. Do not turn the simple quotient into a claim about skin comfort.

Pressure-compensating devices flatten the curve

A simple unregulated opening produces flow that changes with pressure. A pressure-compensating or flow-regulating design aims to maintain a more stable rate across a stated pressure range.

Suppose an unregulated head delivers 4.5, 6.0 and 7.4 L/min at 100, 200 and 300 kPa, while a regulated model delivers 5.4, 5.8 and 6.0 L/min. The regulated range is 0.6 L/min; the unregulated range is 2.9 L/min.

One stability metric is coefficient of variation: standard deviation divided by mean. Another is percentage change over the test range. The chosen metric and pressure points should be disclosed.

Stable flow is not the only performance goal. Spray coverage and force still matter. A regulator also has an operating range; outside it, behaviour may differ.

A flow curve should not be extrapolated casually

If data cover 100–300 kPa, predicting performance at 600 kPa with the same fitted curve may be unsafe and inaccurate. Components can deform, regulators can reach limits and product ratings matter.

Interpolation within a validated range usually requires fewer assumptions than extrapolation beyond it. Students should shade the measured domain on a graph and distinguish it from any hypothetical extension.

This is one of the most transferable benefits of learning mathematics: a line on paper does not automatically continue forever.

Blocked nozzles change distribution

If mineral deposits obstruct some holes, total open area falls and the remaining jets may carry more or less flow depending on regulation and supply. Spray coverage can become uneven even when total flow changes only modestly.

Suppose 6 of 60 identical holes are blocked. Open-hole count falls 10%. If total flow stayed fixed, average flow per open hole would rise by 60/54≈1.111, or 11.1%. If supply is unregulated, total flow may instead decline.

The two scenarios show why a missing mechanism matters. Equal-pressure and equal-total-flow assumptions lead to different answers.

Follow the approved cleaning guidance for a real fixture. Do not insert objects into outlets or use unapproved chemicals for an experiment.

Pressure drop is shared across the system

If supply pressure is 300 kPa and pipe, valve, hose and head losses are 40, 60, 30 and 120 kPa at one flow, total loss is 250 kPa, leaving 50 kPa in the simplified balance. Changing the hose affects the operating flow and therefore every flow-dependent loss.

Losses are not necessarily fixed numbers. Many increase roughly with Q² over a range. Doubling flow can make a quadratic loss four times larger.

This coupling is similar to water filters, flow rate and pressure drop: the actual operating point belongs to the whole system.

Spray Geometry, Coverage and Uniformity

A spray cone creates a growing footprint

In a simple cone model, spray radius at distance L is r=L tan(θ/2), where θ is the full cone angle. Footprint area is πr².

For θ=40° and L=0.50 m, r=0.50 tan20°≈0.182 m and area is about 0.104 m². At L=0.70 m, radius becomes about 0.255 m and area about 0.204 m², nearly double.

Real spray boundaries are not sharp circles. Individual jets, tilt, air interaction and droplet breakup create irregular patterns. The cone model is a geometric approximation.

Wider coverage dilutes average flux

If 6 L/min were spread uniformly across 0.10 m², mean areal flux would be 60 L/min/m². Across 0.20 m², it would be 30 L/min/m².

Uniformity is a strong assumption. A ring may receive more water than the centre. Two heads with equal footprint area can feel different because their distributions differ.

This introduces the idea of a field: flow is not just one total but a quantity distributed over space. Heat maps and grids are natural mathematical tools.

A catch grid estimates spatial distribution

In a controlled laboratory, identical collectors can be arranged on a grid. Water collected in each cup approximates local flux over its area. At home, this can be messy and slippery, so students should use a teacher-provided dataset or a dry paper simulation unless supervised in a safe facility.

Suppose nine equal cells collect 8, 10, 7, 12, 15, 11, 6, 9 and 7 units. The mean is 9.44. The centre cell is above average, and the corners are generally lower.

A simple uniformity measure is minimum divided by mean: 6/9.44=0.635. Another is coefficient of variation. Different indices emphasise different aspects, so the definition belongs beside the score.

Centroid reveals where the spray is biased

Assign each grid cell coordinates and use collected volume as a weight. The weighted centroid is x̄=Σw_ix_i/Σw_i and ȳ=Σw_iy_i/Σw_i.

If the centroid lies to the right of the geometric centre, more water was collected on the right. That could reflect head angle, asymmetric blockage or measurement setup.

The centroid does not explain the cause. It summarises the observed distribution. A second statistic, such as spread around the centroid, describes width.

Coverage threshold changes the answer

If any droplet counts as “covered,” a footprint may look wide. If a cell must receive at least 50% of mean flux, the effective coverage is smaller.

Suppose 25 grid cells include 21 with some water but only 14 above the chosen threshold. Coverage is 84% under the first rule and 56% under the second.

Neither number is inherently dishonest, but an unlabeled coverage claim is incomplete. A rigorous comparison states the threshold and grid method.

Overlap can improve or worsen uniformity

Multiple jets overlap. If two weak edges add, overlap can fill a gap. If strong centres overlap, it can create a hot spot.

For two ideal distributions f₁(x,y) and f₂(x,y), combined flux is f₁+f₂ if their contributions are independent. The sum is easy; predicting each field is the hard part.

This concept appears in lighting, irrigation, wireless signals and sound. Spatial superposition is a powerful reason mathematics matters in design.

Droplet size needs a distribution

A spray contains many droplet sizes. Reporting only an arithmetic mean can be misleading because a few large drops contribute much of the water volume.

For a spherical droplet, volume is proportional to diameter cubed. A 2 mm droplet has eight times the volume of a 1 mm droplet. Count-weighted and volume-weighted averages therefore differ.

Measuring droplet size accurately requires suitable instruments and methods. A household photograph is not enough for a precise claim. Students can learn the statistical principle using simulated diameters instead.

Angle measurement has geometric uncertainty

If spray radius is measured as 0.20±0.01 m at distance 0.50±0.01 m, the calculated angle has uncertainty. Using θ=2 arctan(r/L), the central estimate is about 43.6°.

Using plausible extremes gives approximately 2 arctan(0.19/0.51)=40.9° and 2 arctan(0.21/0.49)=46.4°. Reporting 43.600° would imply far more precision than the setup supports.

Uncertainty is not failure. It tells the reader what distinctions the measurement can justify.

Standards, Evidence and Fair Comparisons

Flow limits and performance should be read together

The US Environmental Protection Agency WaterSense showerheads page explains that labelled products must meet efficiency and performance criteria. The WaterSense Specification for Showerheads, Version 1.1 sets a maximum rated flow of 2.0 US gallons per minute and includes spray-force and spray-coverage requirements under stated tests.

The specification is a useful example of multi-criteria engineering. Reducing a resource rate is not enough if the device fails its intended function. Performance criteria make the objective richer than “smallest flow wins.”

The US gallon and test conditions must not be silently transplanted into a different jurisdiction or product claim. Use local requirements and current official information for real purchasing decisions.

A threshold is not a universal comfort guarantee

The WaterSense specification includes a minimum spray-force requirement of 0.56 newton at 20 psi, or 140 kPa, and annular-ring coverage criteria. These are test thresholds, not promises that every user will prefer every labelled spray.

Human comfort is variable. Hair length, mobility, sensitivity, height and shower enclosure all affect experience. A standard can support consistent minimum performance while leaving room for preferences.

Mathematics helps by distinguishing conformance from satisfaction. A binary pass/fail result and a continuous preference score should not be merged.

Test repeatability matters

If three flow measurements are 5.8, 6.0 and 5.9 L/min, the mean is 5.9 L/min and the range is 0.2. If another device gives 5.7, 6.2 and 5.9, the same mean hides greater variation.

Report individual values or an honest spread. The number of repeats, pressure condition, water temperature and setup belong in the method.

Repeated household measurements do not reproduce a certification laboratory. They can teach measurement and reveal trends, but they should not be presented as a formal compliance test.

Compare at equivalent pressure and settings

Device A measured at 300 kPa and Device B at 100 kPa are not directly compared if pressure response matters. Similarly, a massage setting and a wide-spray setting may use the same product differently.

A fair table includes inlet pressure, setting, flow, distance, angle and method. If values come from different sources, identify them rather than averaging them together.

The habit transfers to digital multimeters, counts, ranges and uncertainty: a reading is inseparable from range, method and conditions.

Avoid a single opaque score

Suppose one head has lower flow, another wider coverage and a third more stable pressure response. A weighted score can rank them, but the weights express preferences.

If water use receives weight 0.5, coverage 0.3 and stability 0.2, changing water-use weight to 0.3 may change the winner. Sensitivity analysis exposes that dependence.

Families may reasonably choose different weights. A transparent table often communicates better than a “best shower head” score that hides values.

Correlation does not establish the mechanism

If lower-flow models in a small sample also receive lower user ratings, that does not prove low flow caused dissatisfaction. Price, spray pattern, installation, brand and expectations may differ.

A controlled comparison would change one factor at a time or use a design that accounts for confounders. Even then, preferences vary.

Mathematical literacy includes resisting causal stories that outrun the study design.

A Worked Shower-Use Example

Record a safe fictional week

Consider a household that records fictional shower durations for seven days: 6.0, 7.5, 5.5, 8.0, 6.5, 9.0 and 7.0 minutes. Mean duration is 7.07 minutes and total is 49.5 minutes.

The shower head’s measured mean flow in a safe cold-water bucket test is 6.2 L/min. Estimated weekly shower water for one person is 49.5×6.2=306.9 L.

This estimate assumes flow is on continuously. If water is paused, active-flow time should replace total bathroom time.

Compare a second head without promising savings

A second head averages 5.4 L/min under comparable household pressure. If behaviour stayed identical, estimated weekly use would be 49.5×5.4=267.3 L, a difference of 39.6 L.

If average duration increased to 8.0 minutes per day, weekly use would be 5.4×56=302.4 L, nearly the original amount. The technical rate fell, but behaviour offset much of the simple saving.

The honest conclusion is conditional: the lower measured flow would reduce volume for the same active-flow duration.

Add water-heating energy cautiously

The ideal heat needed is E=mcΔT. For 40 L of water, mass is approximately 40 kg. Raising it by 20°C with c≈4.186 kJ/(kg·°C) requires 40×4.186×20=3,348.8 kJ, or about 0.93 kWh of thermal energy.

Real heater input is larger because efficiency and heat losses matter. Mixing hot and cold water changes the source volumes. The calculation is a first-law model, not a utility-bill guarantee.

This connects to electric kettles, power, heating time and energy efficiency, where the same mass–specific-heat–temperature relation appears at a smaller scale.

Estimate an annual scenario range

At 5.4 L/min, durations from six to eight minutes correspond to 32.4–43.2 L per shower. Across 365 showers, the simple range is 11,826–15,768 L.

This is a scenario range, not a confidence interval. It assumes one shower per day and stable flow. The range makes duration sensitivity visible without pretending to predict every day.

If two people share the pattern, multiplying by two is appropriate only if their durations are similar. It may be better to model each person separately.

Compare cost without hiding tariff dependence

If a fictional combined water and heating cost were c dollars per litre-equivalent, total cost would be cV. Because tariffs and energy sources differ over time and place, this article does not insert a current price.

Students can plug in an authoritative local tariff for a dated assignment and label the date. They should separate water charges, energy charges and fixed charges.

The method remains useful even when the price changes. That is one advantage of symbolic mathematics.

Make the decision table multi-dimensional

Criterion	Head A	Head B	What the number does not prove
Mean flow at test condition	6.2 L/min	5.4 L/min	Actual annual use
Flow range across repeats	0.3 L/min	0.2 L/min	Formal certification
Modelled footprint at distance	0.16 m²	0.20 m²	Uniform coverage
Household preference score	7/10	8/10	Everyone’s preference

The table keeps evidence separate. Head B looks favourable in this fictional scenario, but the conclusion remains limited to the measures and users observed.

Check whether the result is sensitive to one unusual day

The nine-minute observation is the longest in the week. Removing it gives a six-day total of 40.5 minutes and a mean of 6.75 minutes, compared with 7.07 minutes for all seven days. The difference is 0.32 minute, or about nineteen seconds.

That calculation is not a reason to delete inconvenient data. It is a sensitivity check. If one point changes the conclusion dramatically, the report should say so and investigate whether the day was valid, unusual or recorded incorrectly.

Students can also use the median. For the seven durations, the median is 7.0 minutes, close to the mean. Agreement does not prove the distribution is normal; it simply shows that the single longest value is not dominating this small dataset.

Distinguish household totals from per-person averages

If three household members have weekly shower-water estimates of 280, 320 and 410 L, the household total is 1,010 L and the per-person mean is about 337 L. Neither number should replace the individual pattern.

A conservation plan based only on the mean may miss that one person has a medical, mobility or caregiving need. Mathematics helps reveal variation, while responsible interpretation keeps the people behind the data visible.

Fifteen Safe Student Investigations

1. Repeat a bucket-flow measurement

With adult supervision and a dry, stable setup, use cool water and a marked container. Repeat the same collection five times and report mean, range and method.

2. Compare short and long timing intervals

Estimate how a one-second timing uncertainty affects 10-, 30- and 60-second tests. Choose the safest useful interval.

3. Build a duration–flow matrix

Calculate session volume for several flows and durations. Highlight combinations that produce equal volume.

4. Graph cumulative volume

Use fictional data with pauses. Distinguish active-flow rate from elapsed-session average.

5. Convert gallons per minute

Convert product values to L/min while noting the gallon convention. Do not rank products with ambiguous units.

6. Explore nozzle-area sensitivity

Calculate total area for several hole counts and diameters. Explain why area changes with diameter squared.

7. Estimate ideal jet speed

Use Q/A with a fictional total opening area. State why real jets and regulators make this an approximation.

8. Compare pressure curves

Plot regulated and unregulated flow data over the same pressure range. Calculate percentage change and identify the tested domain.

9. Model blocked-hole scenarios

Compare fixed-total-flow and fixed-pressure assumptions. Show that the two models predict different per-hole behaviour.

10. Draw spray cones

Use paper geometry and r=L tan(θ/2). Compare how distance changes footprint area.

11. Analyse a catch-grid dataset

Calculate mean, minimum-to-mean ratio and coefficient of variation. Discuss what each uniformity measure rewards.

12. Find a weighted centroid

Use a synthetic 3×3 spray grid. Locate the water-weighted centre and compare it with the geometric centre.

13. Test coverage thresholds

Count grid cells above several thresholds. Explain why coverage depends on the chosen definition.

14. Run a rebound scenario

Reduce flow and increase duration by different amounts. Find when expected volume stops decreasing.

15. Write a fair recommendation

Name the user’s priorities, evidence, uncertainty and safety boundary. Avoid universal “best” language.

How Students Can Build Transferable Mathematical Skill

Treat units as part of the reasoning

L/min multiplied by min gives L. Pressure multiplied by area gives force. Diameter squared determines area. Every valid calculation tells a dimensional story.

When units fail to cancel or combine, pause before pressing a calculator. This single habit prevents many errors.

Define the denominator

Litres per active minute, litres per elapsed session, litres per person and litres per day are different metrics. State the denominator in words before comparing.

The same discipline improves comparisons of marks, prices, energy and travel speeds.

Use distributions when space matters

One total flow cannot describe where water lands. A grid, heat map, centroid and spread convert a spatial experience into analysable evidence.

Students learn that averages are not enemies; they are summaries with limits.

Keep performance and preference separate

Flow stability and coverage can be measured. Comfort can be surveyed. A standard can set a threshold. These forms of evidence should inform one another without being collapsed.

A design decision can respect both physics and people.

Ask how behaviour changes totals

A device rate is only one factor in annual use. Duration, frequency and pauses matter. Modelling them encourages systems thinking rather than label-only reasoning.

That perspective is useful in transport, food, electricity and personal study habits.

Guidance for Parents and Teachers

Make safety the first constraint

Use cool water, stable containers and dry calculation areas. Keep phones, laptops and mains electricity away from water. Do not ask students to dismantle fixtures or touch heaters.

If the environment cannot be made safely dry and stable, use provided datasets instead. The mathematics remains rich.

Ask for conditional language

Encourage “at the measured flow and same duration” rather than “this will save exactly.” Conditional wording is not weak; it accurately matches a model.

Students can be confident and careful at the same time.

Value user experience without turning it into physics

A preference score is real evidence about that participant. It is not a pressure reading. Let students design respectful surveys and report sample size.

This teaches that quantitative work can include human variation without pretending everyone is identical.

Prefer transparent tables to winner labels

A table of flow, spread, coverage and preference supports discussion. A single hidden score suppresses trade-offs.

Ask how changing the weights changes the ranking. That turns disagreement into a mathematical sensitivity question.

Questions Parents and Students Often Ask

Does a lower-flow shower always use less water?

It uses less per active minute under the compared condition. Total session use also depends on duration and pauses.

Is pressure the same as flow?

No. Pressure difference drives flow through resistance, but the same pressure can produce different flows through different heads.

Do smaller holes always save water?

No. Hole count, regulation, pressure and internal design matter. Smaller diameter reduces area, but the system may respond in other ways.

Why can a lower flow still feel strong?

The spray may concentrate water into faster or differently distributed jets. Sensation depends on more than total L/min.

What does pressure compensation do?

It aims to keep flow more stable across a stated pressure range. It does not make flow identical under every condition.

Is a wider spray always better?

Not necessarily. Wider coverage may reduce average flux and can be uneven. User preference and intended use matter.

Can a household bucket test prove certification?

No. It can estimate flow under household conditions and teach measurement. Certification uses specified equipment and procedures.

Why use a grid for spray?

Because one total cannot show spatial distribution. A grid reveals centre bias, gaps and unevenness.

How does shower water relate to energy?

Heating energy depends on water mass, temperature rise, heater efficiency and losses. Reducing hot-water volume can reduce required heat under comparable conditions.

What is the main mathematical lesson?

A design has several objectives. Flow, pressure response, geometry, uniformity, duration and preference must be defined before they can be compared fairly.

Useful Next Reading

Read the EPA WaterSense showerheads overview for current programme information.
Consult the WaterSense Specification for Showerheads, Version 1.1 for the actual flow and performance criteria discussed here.
Continue with Why Mathematics? | Water Filters, Flow Rate, Pressure Drop and Breakthrough Curves for another household flow system.
Read Why Mathematics? | Electric Kettles, Power, Heating Time and Energy Efficiency to explore water-heating energy.
Review Why Mathematics? | Comparing Percentages Fairly before interpreting savings claims.

A Final Encouraging Thought

The mathematics of a shower head begins with a simple ratio and opens into a design problem. Flow becomes volume when multiplied by time. Nozzle diameter becomes area through a square. A spray cone becomes a footprint through trigonometry. A total becomes a spatial distribution through a grid. A label becomes a household outcome only after behaviour is included.

Students do not need to become plumbers to learn from this system. They need careful units, safe methods, transparent assumptions and respect for human variation. Those habits make everyday decisions clearer and engineering questions more welcoming. That is a happy reason to learn mathematics: it helps us see how comfort, resources and evidence can be considered together.

Why Mathematics? | Shower Heads, Flow Rate, Spray Geometry and Water Use