Hydro · Measuring head and flow
Two measurements decide whether a stream is worth developing, and both can be taken in an afternoon with a hose, a pressure gauge, a tape measure, and something that floats. Take them before you price a single component.
Start here
Head is the vertical distance water falls between the point where you take it out of the stream and the point where the turbine sits. Flow is how much water passes in a given time. Multiply them together and you have the energy available. That is the whole physics of itMicrohydropower System">[1].
The two are interchangeable in the arithmetic but not in practice. A great deal of head with very little water works well, because a small pipe carrying high pressure is cheap to build. Very little head with a great deal of water is expensive, because moving large volumes takes large pipe and large machinery to do the same job.
This is why steep ground beats flat ground even when the flat site has the bigger creek. Given a choice, buy head.
The Department of Energy defines microhydropower as systems producing up to 100 kilowatts, and notes that a 10 kilowatt system is generally enough for a large home, a small resort, or a hobby farm[2]. Most household systems land far below that. Continuous output is what makes a small number useful, so do not judge a site by its peak wattage alone.
Measuring head
A column of water 2.31 feet high exerts one pound per square inch at its base. That constant is the entire method. Fill a hose that runs from your intake down to your turbine site, read the pressure at the bottom, multiply by 2.31, and you have your head in feet[3].
Head (feet) = pressure (psi) × 2.31
Working the other way, each vertical foot of water produces 0.433 psi. A reading of 26 psi is 60 feet of head.
An air pocket anywhere in the hose reads as missing head. Run water through for several minutes before attaching the gauge, and route the hose so it declines steadily rather than rising over a hump partway down.
A gauge reads most accurately in the middle of its range. Measuring 12 psi on a dial that runs to 800 psi produces a number worth nothing. For low-head sites, use a gauge that tops out near 30 psi.
Head can be measured in segments down a long slope, but each reading carries its own error and the errors accumulate. One hose spanning the full drop gives one number to be wrong about instead of six.
Barometric altimeters and GPS elevation are fine for deciding whether a site is worth walking. They are not fine for design. Inexpensive barometric units have been documented in error by 150 feet or more even when calibrated[4].
What the hose measures with the water standing still is gross head, the true vertical drop. The moment water moves through a pipe, friction takes some of it back, and what reaches the turbine is net head. Net head is the number that makes power.
For a first estimate, Department of Energy guidance is to allow 5 to 10 percent for pipe friction[1]. The common design rule is to size the pipe so friction costs no more than 10 to 15 percent of gross head, and to stay at the low end of that range on low-head sites where there is less to give away.
Pipe diameter is the lever that controls this, and it is the single most consequential civil decision in a micro-hydro system. Sizing it properly is its own subject, covered in the intake and penstock guide.
Measuring flow
Flow is measured in gallons per minute for small systems, or cubic feet per second for anything a hydrologist wrote down. One cubic foot per second is about 449 gallons per minute, and you will need that conversion the moment you look up a public gauge record.
Temporarily dam the stream with boards or logs so that all of it runs through one outlet, then time how long a container of known volume takes to fill. A five gallon bucket filling in one minute is five gallons per minute[1].
It is the most accurate method available to a homeowner, and it only works where the entire flow can be captured. On anything larger than a spring outflow or a small creek in late summer, it stops being practical.
Flow equals cross-sectional area multiplied by velocity. Measure the area with a tape and a marked rod, then find velocity by timing a floating object over a measured distance.
The catch is that a float rides the surface, where water moves faster than it does through the whole cross-section. Montana's Department of Natural Resources and Conservation specifies multiplying surface velocity by a coefficient of 0.66 to 0.75 depending on channel depth to get the mean[5]. Skipping that correction overstates flow by roughly a third.
Their field procedure: pick a straight reach with uniform grade and minimal surface agitation, ideally at least three channel widths long. Time at least three runs and average them, with travel time over 20 seconds. Measure depth at regular intervals across the channel, five points at minimum and more on a wider stream.
A weir is a temporary barrier with a precisely shaped notch cut in it. All the water passes through the notch, and the depth of water flowing over it corresponds to a known flow rate read from a published weir table.
It takes more effort to set up than the other two and rewards it by making repeat measurement trivial. Once a weir and its depth stake are in place, checking the flow in April and again in September takes a minute and a glance, which is exactly the habit this whole page is trying to build.
Department of Energy guidance is explicit that the float method is not recommended where the stream is fast-flowing or deeper than your calves[1]. Moving water carries far more force than its depth suggests, and cold water takes coordination away quickly.
Measure at low water rather than at peak flow, which is when you want the number anyway. Wade only where you can see the bottom and stand comfortably, and do not do it alone.
When to measure
A stream is a season, not a number. The same channel that carries 1,900 gallons a minute during snowmelt may carry a fifth of that in late August, and a system designed on the spring figure spends every summer producing a fraction of what its owner expected.
Department of Energy guidance is direct about this. Unless you are building storage, use the lowest average flow of the year as the basis for system design[1]. Readers who came here from the solar section will recognize the discipline. It is the same worst-month logic that sizes an array against December rather than against the annual average.
Monthly if you can manage it, and certainly at the end of the driest stretch. A weir makes this cheap enough that there is no excuse for guessing.
The U.S. Geological Survey publishes streamflow records from gauges nationwide[6]. Your creek is almost certainly not gauged, but a nearby watershed shows the shape of a normal year and how bad a dry one gets.
People who have watched a creek for thirty years know which summers it stopped running. That is data no single season of measurement will give you, and it is free.
Worked example
A hillside property with a spring-fed creek running through it. The figures below are illustrative and chosen to demonstrate the sequence. This is the same property the rest of the hydro section returns to, so these numbers reappear in later guides.
Three joined garden hoses run from the proposed intake down to a flat spot near the house. Flushed, filled, and read on a 0 to 60 psi gauge: 26 psi. That is 26 × 2.31 = 60 feet of gross head.
Allowing 10 percent for pipe friction at this preliminary stage leaves 54 feet of net head. The real figure depends on pipe diameter and length, and a larger pipe buys some of it back.
Late August, float-area method. Average width 4.0 feet. Five depth readings across the channel of 0.2, 0.35, 0.5, 0.4 and 0.3 feet average to 0.35 feet, so the area is 4.0 × 0.35 = 1.40 square feet.
Three float runs over 50 feet take 66, 70 and 68 seconds, averaging 68. Surface velocity is 50 ÷ 68 = 0.74 feet per second. Corrected at 0.7 for a shallow channel: 0.52 feet per second.
Flow is 1.40 × 0.52 = 0.72 cubic feet per second, which at 449 gallons per cubic foot is roughly 320 gallons per minute in the whole creek.
Not all of it, for reasons both legal and ecological. This design takes 150 gallons per minute and leaves 170 in the channel, slightly more than half. The permissible fraction is set by state water law, not by preference.
54 × 150 ÷ 10 = 810 watts, continuous. Across a full day that is 19,440 watt-hours, and it arrives at the same rate at three in the morning in February as it does at noon in June.
The property worked through the solar section needed 9,100 watt-hours a day. This creek delivers more than twice that from a machine small enough to sit on a workbench, and it does it in December when the same property's fifteen solar panels are managing 9,600.
That is the case for water in one line. Modest instantaneous output, multiplied by 8,760 hours a year, beats impressive output multiplied by whenever the weather allows.
The divide-by-ten formula carries a typical system efficiency inside it, somewhere in the 50 to 70 percent range that the Department of Energy considers representative of micro-hydro[1]. The actual figure depends on which turbine you choose, how well it is matched to your head and flow, and how much head the penstock gives back. This number decides whether to keep going. It does not size equipment.
Before you plan around the number
A stream crossing your deed is not automatically yours to divert. In much of the western United States, surface water is allocated under prior appropriation, where senior rights holders downstream are entitled to their share before you take any. In most eastern states a riparian doctrine applies instead, allowing reasonable use that does not harm other riparian owners.
Either way, the quantity you may divert, the season you may divert it, and the minimum you must leave in the channel are set by state law and often by a permit. Diverting the whole flow is rarely legal and never good practice. Fish, insects and the plants along the bank all depend on water staying in the streambed.
Federal jurisdiction can also apply. The permits guide in this section covers the Federal Energy Regulatory Commission's licensing and exemption pathways, state water right filings, and where the Army Corps of Engineers becomes involved. Read it before committing money, not after.
Water rights, zoning and what a property actually permits, state by state.
Read the sectionThe other half of a property's relationship with moving water, including wells and storage.
Read the sectionWhether 810 continuous watts covers your household, worked from your device list.
Run the numbersCommon mistakes
The most expensive error on this page. A system sized on snowmelt is a system that underperforms for the two-thirds of the year that matters most.
Water at the surface moves faster than the average through the channel. Omitting the 0.66 to 0.75 coefficient inflates flow by around a third before any other error is made.
The hose reading is what the site offers. What the turbine receives is less, and the gap is set by a pipe that has not been bought yet.
A 0 to 800 psi dial cannot resolve 12 psi meaningfully. On low-head sites this single choice can move the head figure by a quarter.
The design flow is what you may lawfully divert while leaving the channel intact, not what passes the property. Those are usually very different numbers.
Divide-by-ten is a screening tool with an assumed efficiency baked in. Ordering a turbine against it, without matching the machine to the site, is how people buy the wrong runner.
Next
Head and flow between them determine nearly every decision that follows. They set which of the three system configurations suits the site, which turbine can use the water efficiently, how large the penstock has to be, and how much of the year the whole arrangement will actually run.
The run-of-river guide comes next in reading order. Every guide in this section starts from the numbers measured here, and the same creek runs through all eight.
The configuration most qualifying properties should build, and what the component chain costs.
Read the guideHead and flow read off against the machines that can use them.
Read the guideWhat you may lawfully divert, which is the flow these calculations should actually use.
Read the guideSources
Head, flow and power figures in the worked example are illustrative and chosen to demonstrate the method. Measure your own site across a full year before designing anything, and confirm what you may divert with your state water agency before spending money.